replicate/dreambooth 🔢✓🖼️📝❓ → 🖼️

▶️ 296.2K runs 📅 Nov 2022 ⚙️ Cog v0.7.0-beta7+dev 🔗 GitHub 📄 Paper ⚖️ License
dreambooth image-model-training stable-diffusion

About

Train your own custom Stable Diffusion model using a small set of images

Example Output

Output

Example output

Performance Metrics

992.31s Prediction Time
1046.67s Total Time
All Input Parameters
{
  "class_prompt": "a photo of a man",
  "instance_data": "https://replicate.delivery/pbxt/HoUeWsrtTTCJEpKGdLKqIYTfo8nbUTSNs565MkGxEstjfwKt/data.zip",
  "instance_prompt": "a photo of bfirsh",
  "max_train_steps": 2000
}
Input Parameters
seed Type: integerDefault: 1337
A seed for reproducible training
scale_lr Type: booleanDefault: false
Scale the learning rate by the number of GPUs, gradient accumulation steps, and batch size.
ckpt_base (required) Type: string
A ckpt file with existing training base
adam_beta1 Type: numberDefault: 0.9
The beta1 parameter for the Adam optimizer.
adam_beta2 Type: numberDefault: 0.999
The beta2 parameter for the Adam optimizer.
class_data Type: string
An optional ZIP file containing the training data of class images. This corresponds to `class_prompt` above, also with the purpose of keeping the model generalizable. By default, the pretrained stable-diffusion model will generate N images (determined by the `num_class_images` you set) based on the `class_prompt` provided. But to save time or to have your preferred specific set of `class_data`, you can also provide them in a ZIP file.
pad_tokens Type: booleanDefault: false
Flag to pad tokens to length 77.
resolution Type: integerDefault: 512
The resolution for input images. All the images in the train/validation dataset will be resized to this resolution.
center_crop Type: booleanDefault: false
Whether to center crop images before resizing to resolution
adam_epsilon Type: numberDefault: 1e-8
Epsilon value for the Adam optimizer
class_prompt (required) Type: string
The prompt or description of the coarse class of your training images, in the format of `a [class noun]`, optionally with some extra description. `class_prompt` is used to alleviate overfitting to your customised images (the trained model should still keep the learnt prior so that it can still generate different dogs when the `[identifier]` is not in the prompt). Corresponding to the examples of the `instant_prompt` above, the `class_prompt` can be `a dog` or `a photo of a dog`.
lr_scheduler Default: constant
The scheduler type to use
instance_data (required) Type: string
A ZIP file containing your training images (JPG, PNG, etc. size not restricted). These images contain your 'subject' that you want the trained model to embed in the output domain for later generating customized scenes beyond the training images. For best results, use images without noise or unrelated objects in the background.
learning_rate Type: numberDefault: 0.000001
Initial learning rate (after the potential warmup period) to use.
max_grad_norm Type: numberDefault: 1
Max gradient norm.
n_save_sample Type: integerDefault: 4
The number of samples to save.
use_8bit_adam Type: booleanDefault: false
Whether or not to use 8-bit Adam from bitsandbytes.
instance_prompt (required) Type: string
The prompt you use to describe your training images, in the format: `a [identifier] [class noun]`, where the `[identifier]` should be a rare token. Relatively short sequences with 1-3 letters work the best (e.g. `sks`, `xjy`). `[class noun]` is a coarse class descriptor of the subject (e.g. cat, dog, watch, etc.). For example, your `instance_prompt` can be: `a sks dog`, or with some extra description `a photo of a sks dog`. The trained model will learn to bind a unique identifier with your specific subject in the `instance_data`.
lr_warmup_steps Type: integerDefault: 0
Number of steps for the warmup in the lr scheduler.
max_train_steps Type: integerDefault: 2000
Total number of training steps to perform. If provided, overrides num_train_epochs.
num_class_images Type: integerDefault: 50
Minimal class images for prior preservation loss. If not enough images are provided in class_data, additional images will be sampled with class_prompt.
num_train_epochs Type: integerDefault: 1
save_infer_steps Type: integerDefault: 50
The number of inference steps for save sample.
train_batch_size Type: integerDefault: 1
Batch size (per device) for the training dataloader.
adam_weight_decay Type: numberDefault: 0.01
Weight decay to use
prior_loss_weight Type: numberDefault: 1
Weight of prior preservation loss.
sample_batch_size Type: integerDefault: 4
Batch size (per device) for sampling images.
save_sample_prompt Type: string
The prompt used to generate sample outputs to save.
train_text_encoder Type: booleanDefault: true
Whether to train the text encoder
save_guidance_scale Type: numberDefault: 7.5
CFG for save sample.
gradient_checkpointing Type: booleanDefault: false
Whether or not to use gradient checkpointing to save memory at the expense of slower backward pass.
with_prior_preservation Type: booleanDefault: true
Flag to add prior preservation loss.
gradient_accumulation_steps Type: integerDefault: 1
Number of updates steps to accumulate before performing a backward/update pass.
save_sample_negative_prompt Type: string
The negative prompt used to generate sample outputs to save.
Output Schema

Output

Type: stringFormat: uri

Example Execution Logs
/root/.pyenv/versions/3.10.8/lib/python3.10/site-packages/accelerate/accelerator.py:179: UserWarning: `log_with=tensorboard` was passed but no supported trackers are currently installed.
warnings.warn(f"`log_with={log_with}` was passed but no supported trackers are currently installed.")
You have passed `None` for safety_checker to disable its functionality in <class 'diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline'>. Note that this might lead to problems when using <class 'diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline'> and is not recommended.
Generating class images:   0%|          | 0/13 [00:00<?, ?it/s]
Generating class images:   8%|▊         | 1/13 [00:42<08:28, 42.41s/it]
Generating class images:  15%|█▌        | 2/13 [00:51<04:13, 23.06s/it]
Generating class images:  23%|██▎       | 3/13 [01:01<02:48, 16.87s/it]
Generating class images:  31%|███       | 4/13 [01:10<02:05, 13.97s/it]
Generating class images:  38%|███▊      | 5/13 [01:20<01:39, 12.43s/it]
Generating class images:  46%|████▌     | 6/13 [01:30<01:20, 11.43s/it]
Generating class images:  54%|█████▍    | 7/13 [01:39<01:04, 10.80s/it]
Generating class images:  62%|██████▏   | 8/13 [01:49<00:51, 10.38s/it]
Generating class images:  69%|██████▉   | 9/13 [01:58<00:40, 10.10s/it]
Generating class images:  77%|███████▋  | 10/13 [02:08<00:29,  9.91s/it]
Generating class images:  85%|████████▍ | 11/13 [02:17<00:19,  9.78s/it]
Generating class images:  92%|█████████▏| 12/13 [02:27<00:09,  9.69s/it]
Generating class images: 100%|██████████| 13/13 [02:41<00:00, 11.19s/it]
Generating class images: 100%|██████████| 13/13 [02:41<00:00, 12.44s/it]
Caching latents:   0%|          | 0/50 [00:00<?, ?it/s]
Caching latents:   2%|▏         | 1/50 [00:01<00:49,  1.00s/it]
Caching latents:   4%|▍         | 2/50 [00:01<00:24,  1.94it/s]
Caching latents:  10%|█         | 5/50 [00:01<00:07,  5.69it/s]
Caching latents:  16%|█▌        | 8/50 [00:01<00:04,  9.45it/s]
Caching latents:  22%|██▏       | 11/50 [00:01<00:03, 12.71it/s]
Caching latents:  28%|██▊       | 14/50 [00:01<00:02, 15.95it/s]
Caching latents:  34%|███▍      | 17/50 [00:01<00:02, 14.63it/s]
Caching latents:  40%|████      | 20/50 [00:01<00:01, 17.54it/s]
Caching latents:  46%|████▌     | 23/50 [00:02<00:01, 15.88it/s]
Caching latents:  52%|█████▏    | 26/50 [00:02<00:01, 18.53it/s]
Caching latents:  58%|█████▊    | 29/50 [00:02<00:01, 20.86it/s]
Caching latents:  64%|██████▍   | 32/50 [00:02<00:01, 17.87it/s]
Caching latents:  70%|███████   | 35/50 [00:02<00:00, 20.24it/s]
Caching latents:  76%|███████▌  | 38/50 [00:02<00:00, 21.71it/s]
Caching latents:  82%|████████▏ | 41/50 [00:03<00:00, 18.45it/s]
Caching latents:  88%|████████▊ | 44/50 [00:03<00:00, 16.64it/s]
Caching latents:  92%|█████████▏| 46/50 [00:03<00:00, 12.44it/s]
Caching latents:  98%|█████████▊| 49/50 [00:03<00:00, 15.27it/s]
Caching latents: 100%|██████████| 50/50 [00:03<00:00, 13.38it/s]
0%|          | 0/2000 [00:00<?, ?it/s]
Steps:   0%|          | 0/2000 [00:00<?, ?it/s]
Steps:   0%|          | 0/2000 [00:10<?, ?it/s, loss=0.966, lr=1e-6]
Steps:   0%|          | 1/2000 [00:10<5:42:49, 10.29s/it, loss=0.966, lr=1e-6]
Steps:   0%|          | 2/2000 [00:10<2:28:45,  4.47s/it, loss=0.966, lr=1e-6]
Steps:   0%|          | 3/2000 [00:11<1:26:03,  2.59s/it, loss=0.966, lr=1e-6]
Steps:   0%|          | 4/2000 [00:11<56:48,  1.71s/it, loss=0.966, lr=1e-6]  
Steps:   0%|          | 5/2000 [00:11<40:39,  1.22s/it, loss=0.966, lr=1e-6]
Steps:   0%|          | 6/2000 [00:12<30:47,  1.08it/s, loss=0.966, lr=1e-6]
Steps:   0%|          | 7/2000 [00:12<24:33,  1.35it/s, loss=0.966, lr=1e-6]
Steps:   0%|          | 8/2000 [00:12<20:31,  1.62it/s, loss=0.966, lr=1e-6]
Steps:   0%|          | 9/2000 [00:13<17:44,  1.87it/s, loss=0.966, lr=1e-6]
Steps:   0%|          | 10/2000 [00:13<16:29,  2.01it/s, loss=0.966, lr=1e-6]
Steps:   0%|          | 10/2000 [00:13<16:29,  2.01it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 11/2000 [00:13<15:12,  2.18it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 12/2000 [00:14<14:11,  2.33it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 13/2000 [00:14<13:29,  2.46it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 14/2000 [00:15<12:53,  2.57it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 15/2000 [00:15<12:34,  2.63it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 16/2000 [00:15<12:21,  2.68it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 17/2000 [00:16<12:09,  2.72it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 18/2000 [00:16<12:00,  2.75it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 19/2000 [00:16<11:51,  2.78it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 20/2000 [00:17<11:43,  2.82it/s, loss=0.305, lr=1e-6]
Steps:   1%|          | 20/2000 [00:17<11:43,  2.82it/s, loss=0.268, lr=1e-6]
Steps:   1%|          | 21/2000 [00:17<11:54,  2.77it/s, loss=0.268, lr=1e-6]
Steps:   1%|          | 22/2000 [00:17<11:46,  2.80it/s, loss=0.268, lr=1e-6]
Steps:   1%|          | 23/2000 [00:18<11:50,  2.78it/s, loss=0.268, lr=1e-6]
Steps:   1%|          | 24/2000 [00:18<11:45,  2.80it/s, loss=0.268, lr=1e-6]
Steps:   1%|▏         | 25/2000 [00:18<11:39,  2.83it/s, loss=0.268, lr=1e-6]
Steps:   1%|▏         | 26/2000 [00:19<11:38,  2.83it/s, loss=0.268, lr=1e-6]
Steps:   1%|▏         | 27/2000 [00:19<11:43,  2.80it/s, loss=0.268, lr=1e-6]
Steps:   1%|▏         | 28/2000 [00:19<11:35,  2.84it/s, loss=0.268, lr=1e-6]
Steps:   1%|▏         | 29/2000 [00:20<11:38,  2.82it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 30/2000 [00:20<11:39,  2.82it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 30/2000 [00:21<11:39,  2.82it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 31/2000 [00:21<11:42,  2.80it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 32/2000 [00:21<11:48,  2.78it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 33/2000 [00:21<11:47,  2.78it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 34/2000 [00:22<11:42,  2.80it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 35/2000 [00:22<11:39,  2.81it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 36/2000 [00:22<11:36,  2.82it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 37/2000 [00:23<11:36,  2.82it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 38/2000 [00:23<11:34,  2.82it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 39/2000 [00:23<11:32,  2.83it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 40/2000 [00:24<11:30,  2.84it/s, loss=0.268, lr=1e-6]
Steps:   2%|▏         | 40/2000 [00:24<11:30,  2.84it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 41/2000 [00:24<11:36,  2.81it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 42/2000 [00:24<11:34,  2.82it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 43/2000 [00:25<11:27,  2.84it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 44/2000 [00:25<11:32,  2.82it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 45/2000 [00:26<11:32,  2.82it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 46/2000 [00:26<11:28,  2.84it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 47/2000 [00:26<11:34,  2.81it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 48/2000 [00:27<11:31,  2.82it/s, loss=0.28, lr=1e-6]
Steps:   2%|▏         | 49/2000 [00:27<11:30,  2.82it/s, loss=0.28, lr=1e-6]
Steps:   2%|▎         | 50/2000 [00:27<11:30,  2.82it/s, loss=0.28, lr=1e-6]
Steps:   2%|▎         | 50/2000 [00:28<11:30,  2.82it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 51/2000 [00:28<11:32,  2.82it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 52/2000 [00:28<11:29,  2.82it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 53/2000 [00:28<11:31,  2.81it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 54/2000 [00:29<11:31,  2.81it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 55/2000 [00:29<11:24,  2.84it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 56/2000 [00:29<11:27,  2.83it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 57/2000 [00:30<11:30,  2.81it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 58/2000 [00:30<11:26,  2.83it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 59/2000 [00:30<11:26,  2.83it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 60/2000 [00:31<11:22,  2.84it/s, loss=0.293, lr=1e-6]
Steps:   3%|▎         | 60/2000 [00:31<11:22,  2.84it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 61/2000 [00:31<11:31,  2.81it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 62/2000 [00:32<11:28,  2.81it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 63/2000 [00:32<11:23,  2.83it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 64/2000 [00:32<11:25,  2.82it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 65/2000 [00:33<11:24,  2.83it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 66/2000 [00:33<11:31,  2.80it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 67/2000 [00:33<11:30,  2.80it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 68/2000 [00:34<11:29,  2.80it/s, loss=0.308, lr=1e-6]
Steps:   3%|▎         | 69/2000 [00:34<11:27,  2.81it/s, loss=0.308, lr=1e-6]
Steps:   4%|▎         | 70/2000 [00:34<11:28,  2.80it/s, loss=0.308, lr=1e-6]
Steps:   4%|▎         | 70/2000 [00:35<11:28,  2.80it/s, loss=0.311, lr=1e-6]
Steps:   4%|▎         | 71/2000 [00:35<11:26,  2.81it/s, loss=0.311, lr=1e-6]
Steps:   4%|▎         | 72/2000 [00:35<11:24,  2.82it/s, loss=0.311, lr=1e-6]
Steps:   4%|▎         | 73/2000 [00:35<11:22,  2.82it/s, loss=0.311, lr=1e-6]
Steps:   4%|▎         | 74/2000 [00:36<11:25,  2.81it/s, loss=0.311, lr=1e-6]
Steps:   4%|▍         | 75/2000 [00:36<11:19,  2.83it/s, loss=0.311, lr=1e-6]
Steps:   4%|▍         | 76/2000 [00:37<11:20,  2.83it/s, loss=0.311, lr=1e-6]
Steps:   4%|▍         | 77/2000 [00:37<11:17,  2.84it/s, loss=0.311, lr=1e-6]
Steps:   4%|▍         | 78/2000 [00:37<11:12,  2.86it/s, loss=0.311, lr=1e-6]
Steps:   4%|▍         | 79/2000 [00:38<11:16,  2.84it/s, loss=0.311, lr=1e-6]
Steps:   4%|▍         | 80/2000 [00:38<11:21,  2.82it/s, loss=0.311, lr=1e-6]
Steps:   4%|▍         | 80/2000 [00:38<11:21,  2.82it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 81/2000 [00:38<11:22,  2.81it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 82/2000 [00:39<11:27,  2.79it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 83/2000 [00:39<11:20,  2.82it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 84/2000 [00:39<11:14,  2.84it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 85/2000 [00:40<11:13,  2.84it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 86/2000 [00:40<11:11,  2.85it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 87/2000 [00:40<11:14,  2.84it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 88/2000 [00:41<11:17,  2.82it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 89/2000 [00:41<11:31,  2.76it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 90/2000 [00:41<11:24,  2.79it/s, loss=0.306, lr=1e-6]
Steps:   4%|▍         | 90/2000 [00:42<11:24,  2.79it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 91/2000 [00:42<11:15,  2.83it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 92/2000 [00:42<11:17,  2.82it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 93/2000 [00:43<11:17,  2.81it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 94/2000 [00:43<11:21,  2.79it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 95/2000 [00:43<11:20,  2.80it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 96/2000 [00:44<11:19,  2.80it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 97/2000 [00:44<11:38,  2.72it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 98/2000 [00:44<11:28,  2.76it/s, loss=0.3, lr=1e-6]
Steps:   5%|▍         | 99/2000 [00:45<11:28,  2.76it/s, loss=0.3, lr=1e-6]
Steps:   5%|▌         | 100/2000 [00:45<11:30,  2.75it/s, loss=0.3, lr=1e-6]
Steps:   5%|▌         | 100/2000 [00:45<11:30,  2.75it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 101/2000 [00:45<11:24,  2.77it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 102/2000 [00:46<11:29,  2.75it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 103/2000 [00:46<11:21,  2.78it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 104/2000 [00:47<11:14,  2.81it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 105/2000 [00:47<11:24,  2.77it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 106/2000 [00:47<11:16,  2.80it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 107/2000 [00:48<11:22,  2.77it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 108/2000 [00:48<11:32,  2.73it/s, loss=0.298, lr=1e-6]
Steps:   5%|▌         | 109/2000 [00:48<11:21,  2.77it/s, loss=0.298, lr=1e-6]
Steps:   6%|▌         | 110/2000 [00:49<11:26,  2.75it/s, loss=0.298, lr=1e-6]
Steps:   6%|▌         | 110/2000 [00:49<11:26,  2.75it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 111/2000 [00:49<11:21,  2.77it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 112/2000 [00:49<11:14,  2.80it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 113/2000 [00:50<11:18,  2.78it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 114/2000 [00:50<11:13,  2.80it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 115/2000 [00:50<11:15,  2.79it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 116/2000 [00:51<11:11,  2.80it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 117/2000 [00:51<11:19,  2.77it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 118/2000 [00:52<11:22,  2.76it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 119/2000 [00:52<11:21,  2.76it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 120/2000 [00:52<11:21,  2.76it/s, loss=0.29, lr=1e-6]
Steps:   6%|▌         | 120/2000 [00:53<11:21,  2.76it/s, loss=0.283, lr=1e-6]
Steps:   6%|▌         | 121/2000 [00:53<11:20,  2.76it/s, loss=0.283, lr=1e-6]
Steps:   6%|▌         | 122/2000 [00:53<11:16,  2.78it/s, loss=0.283, lr=1e-6]
Steps:   6%|▌         | 123/2000 [00:53<11:21,  2.75it/s, loss=0.283, lr=1e-6]
Steps:   6%|▌         | 124/2000 [00:54<11:22,  2.75it/s, loss=0.283, lr=1e-6]
Steps:   6%|▋         | 125/2000 [00:54<11:37,  2.69it/s, loss=0.283, lr=1e-6]
Steps:   6%|▋         | 126/2000 [00:54<11:22,  2.74it/s, loss=0.283, lr=1e-6]
Steps:   6%|▋         | 127/2000 [00:55<11:15,  2.77it/s, loss=0.283, lr=1e-6]
Steps:   6%|▋         | 128/2000 [00:55<11:11,  2.79it/s, loss=0.283, lr=1e-6]
Steps:   6%|▋         | 129/2000 [00:56<11:05,  2.81it/s, loss=0.283, lr=1e-6]
Steps:   6%|▋         | 130/2000 [00:56<11:26,  2.72it/s, loss=0.283, lr=1e-6]
Steps:   6%|▋         | 130/2000 [00:56<11:26,  2.72it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 131/2000 [00:56<11:18,  2.75it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 132/2000 [00:57<11:10,  2.79it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 133/2000 [00:57<11:22,  2.74it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 134/2000 [00:57<11:10,  2.78it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 135/2000 [00:58<11:05,  2.80it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 136/2000 [00:58<11:12,  2.77it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 137/2000 [00:58<11:04,  2.80it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 138/2000 [00:59<11:23,  2.73it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 139/2000 [00:59<11:13,  2.76it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 140/2000 [01:00<11:03,  2.80it/s, loss=0.281, lr=1e-6]
Steps:   7%|▋         | 140/2000 [01:00<11:03,  2.80it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 141/2000 [01:00<11:04,  2.80it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 142/2000 [01:00<10:54,  2.84it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 143/2000 [01:01<10:58,  2.82it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 144/2000 [01:01<10:56,  2.83it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 145/2000 [01:01<10:55,  2.83it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 146/2000 [01:02<11:00,  2.81it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 147/2000 [01:02<10:55,  2.83it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 148/2000 [01:02<10:50,  2.84it/s, loss=0.275, lr=1e-6]
Steps:   7%|▋         | 149/2000 [01:03<11:02,  2.79it/s, loss=0.275, lr=1e-6]
Steps:   8%|▊         | 150/2000 [01:03<11:01,  2.80it/s, loss=0.275, lr=1e-6]
Steps:   8%|▊         | 150/2000 [01:03<11:01,  2.80it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 151/2000 [01:03<11:00,  2.80it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 152/2000 [01:04<10:57,  2.81it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 153/2000 [01:04<10:51,  2.84it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 154/2000 [01:04<10:57,  2.81it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 155/2000 [01:05<10:59,  2.80it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 156/2000 [01:05<10:55,  2.81it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 157/2000 [01:06<10:57,  2.80it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 158/2000 [01:06<10:53,  2.82it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 159/2000 [01:06<10:52,  2.82it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 160/2000 [01:07<10:51,  2.82it/s, loss=0.279, lr=1e-6]
Steps:   8%|▊         | 160/2000 [01:07<10:51,  2.82it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 161/2000 [01:07<11:08,  2.75it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 162/2000 [01:07<11:00,  2.78it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 163/2000 [01:08<10:55,  2.80it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 164/2000 [01:08<10:53,  2.81it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 165/2000 [01:08<10:49,  2.82it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 166/2000 [01:09<11:03,  2.77it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 167/2000 [01:09<10:55,  2.80it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 168/2000 [01:09<10:52,  2.81it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 169/2000 [01:10<10:45,  2.84it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 170/2000 [01:10<10:43,  2.85it/s, loss=0.271, lr=1e-6]
Steps:   8%|▊         | 170/2000 [01:11<10:43,  2.85it/s, loss=0.273, lr=1e-6]
Steps:   9%|▊         | 171/2000 [01:11<10:36,  2.87it/s, loss=0.273, lr=1e-6]
Steps:   9%|▊         | 172/2000 [01:11<10:37,  2.87it/s, loss=0.273, lr=1e-6]
Steps:   9%|▊         | 173/2000 [01:11<10:44,  2.84it/s, loss=0.273, lr=1e-6]
Steps:   9%|▊         | 174/2000 [01:12<10:42,  2.84it/s, loss=0.273, lr=1e-6]
Steps:   9%|▉         | 175/2000 [01:12<10:40,  2.85it/s, loss=0.273, lr=1e-6]
Steps:   9%|▉         | 176/2000 [01:12<10:42,  2.84it/s, loss=0.273, lr=1e-6]
Steps:   9%|▉         | 177/2000 [01:13<10:41,  2.84it/s, loss=0.273, lr=1e-6]
Steps:   9%|▉         | 178/2000 [01:13<10:40,  2.84it/s, loss=0.273, lr=1e-6]
Steps:   9%|▉         | 179/2000 [01:13<10:44,  2.82it/s, loss=0.273, lr=1e-6]
Steps:   9%|▉         | 180/2000 [01:14<10:41,  2.84it/s, loss=0.273, lr=1e-6]
Steps:   9%|▉         | 180/2000 [01:14<10:41,  2.84it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 181/2000 [01:14<10:48,  2.81it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 182/2000 [01:14<10:47,  2.81it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 183/2000 [01:15<10:44,  2.82it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 184/2000 [01:15<10:44,  2.82it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 185/2000 [01:15<10:40,  2.83it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 186/2000 [01:16<10:42,  2.83it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 187/2000 [01:16<10:43,  2.82it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 188/2000 [01:17<10:43,  2.82it/s, loss=0.27, lr=1e-6]
Steps:   9%|▉         | 189/2000 [01:17<10:42,  2.82it/s, loss=0.27, lr=1e-6]
Steps:  10%|▉         | 190/2000 [01:17<10:42,  2.82it/s, loss=0.27, lr=1e-6]
Steps:  10%|▉         | 190/2000 [01:18<10:42,  2.82it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 191/2000 [01:18<10:36,  2.84it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 192/2000 [01:18<10:30,  2.87it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 193/2000 [01:18<10:30,  2.87it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 194/2000 [01:19<10:27,  2.88it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 195/2000 [01:19<10:28,  2.87it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 196/2000 [01:19<10:27,  2.88it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 197/2000 [01:20<10:36,  2.83it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 198/2000 [01:20<10:35,  2.83it/s, loss=0.269, lr=1e-6]
Steps:  10%|▉         | 199/2000 [01:20<10:30,  2.86it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 200/2000 [01:21<10:33,  2.84it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 200/2000 [01:21<10:33,  2.84it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 201/2000 [01:21<10:51,  2.76it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 202/2000 [01:21<10:43,  2.80it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 203/2000 [01:22<10:39,  2.81it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 204/2000 [01:22<10:34,  2.83it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 205/2000 [01:23<10:30,  2.85it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 206/2000 [01:23<10:30,  2.84it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 207/2000 [01:23<10:30,  2.84it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 208/2000 [01:24<10:28,  2.85it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 209/2000 [01:24<10:26,  2.86it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 210/2000 [01:24<10:26,  2.86it/s, loss=0.269, lr=1e-6]
Steps:  10%|█         | 210/2000 [01:25<10:26,  2.86it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 211/2000 [01:25<10:27,  2.85it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 212/2000 [01:25<10:26,  2.86it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 213/2000 [01:25<10:23,  2.87it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 214/2000 [01:26<10:25,  2.86it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 215/2000 [01:26<10:23,  2.86it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 216/2000 [01:26<10:20,  2.87it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 217/2000 [01:27<10:28,  2.84it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 218/2000 [01:27<10:23,  2.86it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 219/2000 [01:27<10:19,  2.87it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 220/2000 [01:28<10:17,  2.88it/s, loss=0.262, lr=1e-6]
Steps:  11%|█         | 220/2000 [01:28<10:17,  2.88it/s, loss=0.268, lr=1e-6]
Steps:  11%|█         | 221/2000 [01:28<10:16,  2.89it/s, loss=0.268, lr=1e-6]
Steps:  11%|█         | 222/2000 [01:28<10:15,  2.89it/s, loss=0.268, lr=1e-6]
Steps:  11%|█         | 223/2000 [01:29<10:22,  2.85it/s, loss=0.268, lr=1e-6]
Steps:  11%|█         | 224/2000 [01:29<10:27,  2.83it/s, loss=0.268, lr=1e-6]
Steps:  11%|█▏        | 225/2000 [01:30<10:22,  2.85it/s, loss=0.268, lr=1e-6]
Steps:  11%|█▏        | 226/2000 [01:30<10:25,  2.84it/s, loss=0.268, lr=1e-6]
Steps:  11%|█▏        | 227/2000 [01:30<10:22,  2.85it/s, loss=0.268, lr=1e-6]
Steps:  11%|█▏        | 228/2000 [01:31<10:17,  2.87it/s, loss=0.268, lr=1e-6]
Steps:  11%|█▏        | 229/2000 [01:31<10:20,  2.86it/s, loss=0.268, lr=1e-6]
Steps:  12%|█▏        | 230/2000 [01:31<10:25,  2.83it/s, loss=0.268, lr=1e-6]
Steps:  12%|█▏        | 230/2000 [01:32<10:25,  2.83it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 231/2000 [01:32<10:20,  2.85it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 232/2000 [01:32<10:16,  2.87it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 233/2000 [01:32<10:14,  2.88it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 234/2000 [01:33<10:11,  2.89it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 235/2000 [01:33<10:13,  2.88it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 236/2000 [01:33<10:13,  2.88it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 237/2000 [01:34<10:12,  2.88it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 238/2000 [01:34<10:19,  2.85it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 239/2000 [01:34<10:15,  2.86it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 240/2000 [01:35<10:18,  2.85it/s, loss=0.266, lr=1e-6]
Steps:  12%|█▏        | 240/2000 [01:35<10:18,  2.85it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 241/2000 [01:35<10:19,  2.84it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 242/2000 [01:35<10:18,  2.84it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 243/2000 [01:36<10:16,  2.85it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 244/2000 [01:36<10:16,  2.85it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 245/2000 [01:37<10:15,  2.85it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 246/2000 [01:37<10:14,  2.85it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 247/2000 [01:37<10:13,  2.86it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 248/2000 [01:38<10:17,  2.84it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▏        | 249/2000 [01:38<10:15,  2.85it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▎        | 250/2000 [01:38<10:13,  2.85it/s, loss=0.27, lr=1e-6]
Steps:  12%|█▎        | 250/2000 [01:39<10:13,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 251/2000 [01:39<10:12,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 252/2000 [01:39<10:12,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 253/2000 [01:39<10:09,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 254/2000 [01:40<10:10,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 255/2000 [01:40<10:14,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 256/2000 [01:40<10:12,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 257/2000 [01:41<10:10,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 258/2000 [01:41<10:24,  2.79it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 259/2000 [01:41<10:17,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 260/2000 [01:42<10:29,  2.76it/s, loss=0.275, lr=1e-6]
Steps:  13%|█▎        | 260/2000 [01:42<10:29,  2.76it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 261/2000 [01:42<10:30,  2.76it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 262/2000 [01:43<10:35,  2.74it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 263/2000 [01:43<10:34,  2.74it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 264/2000 [01:43<10:36,  2.73it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 265/2000 [01:44<10:30,  2.75it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 266/2000 [01:44<10:38,  2.72it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 267/2000 [01:44<10:31,  2.74it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 268/2000 [01:45<10:32,  2.74it/s, loss=0.278, lr=1e-6]
Steps:  13%|█▎        | 269/2000 [01:45<10:26,  2.76it/s, loss=0.278, lr=1e-6]
Steps:  14%|█▎        | 270/2000 [01:45<10:16,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  14%|█▎        | 270/2000 [01:46<10:16,  2.81it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▎        | 271/2000 [01:46<10:21,  2.78it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▎        | 272/2000 [01:46<10:18,  2.79it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▎        | 273/2000 [01:47<10:17,  2.80it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▎        | 274/2000 [01:47<10:15,  2.80it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 275/2000 [01:47<10:13,  2.81it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 276/2000 [01:48<10:08,  2.83it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 277/2000 [01:48<10:09,  2.83it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 278/2000 [01:48<10:09,  2.82it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 279/2000 [01:49<10:05,  2.84it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 280/2000 [01:49<10:03,  2.85it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 280/2000 [01:49<10:03,  2.85it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 281/2000 [01:49<10:06,  2.83it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 282/2000 [01:50<10:02,  2.85it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 283/2000 [01:50<10:02,  2.85it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 284/2000 [01:50<09:58,  2.87it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 285/2000 [01:51<09:54,  2.88it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 286/2000 [01:51<09:58,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 287/2000 [01:51<09:58,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 288/2000 [01:52<09:56,  2.87it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 289/2000 [01:52<09:58,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 290/2000 [01:52<09:56,  2.87it/s, loss=0.281, lr=1e-6]
Steps:  14%|█▍        | 290/2000 [01:53<09:56,  2.87it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 291/2000 [01:53<09:58,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 292/2000 [01:53<09:58,  2.85it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 293/2000 [01:54<10:00,  2.84it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 294/2000 [01:54<10:04,  2.82it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 295/2000 [01:54<10:08,  2.80it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 296/2000 [01:55<10:06,  2.81it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 297/2000 [01:55<09:59,  2.84it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 298/2000 [01:55<09:57,  2.85it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▍        | 299/2000 [01:56<09:54,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▌        | 300/2000 [01:56<09:51,  2.88it/s, loss=0.28, lr=1e-6]
Steps:  15%|█▌        | 300/2000 [01:56<09:51,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 301/2000 [01:56<09:54,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 302/2000 [01:57<09:51,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 303/2000 [01:57<09:49,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 304/2000 [01:57<09:53,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 305/2000 [01:58<10:05,  2.80it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 306/2000 [01:58<09:58,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 307/2000 [01:58<09:54,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 308/2000 [01:59<09:49,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  15%|█▌        | 309/2000 [01:59<09:52,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  16%|█▌        | 310/2000 [02:00<09:50,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  16%|█▌        | 310/2000 [02:00<09:50,  2.86it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 311/2000 [02:00<09:51,  2.86it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 312/2000 [02:00<09:47,  2.87it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 313/2000 [02:01<09:52,  2.85it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 314/2000 [02:01<09:54,  2.84it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 315/2000 [02:01<09:54,  2.83it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 316/2000 [02:02<09:55,  2.83it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 317/2000 [02:02<09:52,  2.84it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 318/2000 [02:02<09:49,  2.85it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 319/2000 [02:03<09:46,  2.87it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 320/2000 [02:03<09:45,  2.87it/s, loss=0.283, lr=1e-6]
Steps:  16%|█▌        | 320/2000 [02:03<09:45,  2.87it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▌        | 321/2000 [02:03<09:44,  2.87it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▌        | 322/2000 [02:04<09:45,  2.87it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▌        | 323/2000 [02:04<09:49,  2.85it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▌        | 324/2000 [02:04<09:50,  2.84it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▋        | 325/2000 [02:05<09:50,  2.84it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▋        | 326/2000 [02:05<09:46,  2.85it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▋        | 327/2000 [02:05<09:43,  2.86it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▋        | 328/2000 [02:06<09:40,  2.88it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▋        | 329/2000 [02:06<09:40,  2.88it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▋        | 330/2000 [02:07<09:39,  2.88it/s, loss=0.286, lr=1e-6]
Steps:  16%|█▋        | 330/2000 [02:07<09:39,  2.88it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 331/2000 [02:07<09:40,  2.88it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 332/2000 [02:07<09:42,  2.86it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 333/2000 [02:08<09:41,  2.86it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 334/2000 [02:08<09:45,  2.85it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 335/2000 [02:08<09:48,  2.83it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 336/2000 [02:09<09:55,  2.79it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 337/2000 [02:09<09:59,  2.77it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 338/2000 [02:09<10:01,  2.76it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 339/2000 [02:10<10:00,  2.77it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 340/2000 [02:10<10:00,  2.76it/s, loss=0.287, lr=1e-6]
Steps:  17%|█▋        | 340/2000 [02:10<10:00,  2.76it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 341/2000 [02:10<09:54,  2.79it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 342/2000 [02:11<09:55,  2.79it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 343/2000 [02:11<10:01,  2.76it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 344/2000 [02:12<09:57,  2.77it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 345/2000 [02:12<09:51,  2.80it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 346/2000 [02:12<09:49,  2.80it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 347/2000 [02:13<09:46,  2.82it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 348/2000 [02:13<09:50,  2.80it/s, loss=0.286, lr=1e-6]
Steps:  17%|█▋        | 349/2000 [02:13<09:46,  2.81it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 350/2000 [02:14<09:48,  2.80it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 350/2000 [02:14<09:48,  2.80it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 351/2000 [02:14<09:51,  2.79it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 352/2000 [02:14<09:46,  2.81it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 353/2000 [02:15<09:45,  2.81it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 354/2000 [02:15<09:42,  2.83it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 355/2000 [02:15<09:39,  2.84it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 356/2000 [02:16<09:38,  2.84it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 357/2000 [02:16<09:37,  2.85it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 358/2000 [02:16<09:35,  2.85it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 359/2000 [02:17<09:41,  2.82it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 360/2000 [02:17<09:38,  2.83it/s, loss=0.285, lr=1e-6]
Steps:  18%|█▊        | 360/2000 [02:18<09:38,  2.83it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 361/2000 [02:18<09:45,  2.80it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 362/2000 [02:18<09:41,  2.82it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 363/2000 [02:18<09:40,  2.82it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 364/2000 [02:19<09:39,  2.82it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 365/2000 [02:19<09:36,  2.83it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 366/2000 [02:19<09:37,  2.83it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 367/2000 [02:20<09:35,  2.84it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 368/2000 [02:20<09:36,  2.83it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 369/2000 [02:20<09:31,  2.85it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 370/2000 [02:21<09:33,  2.84it/s, loss=0.286, lr=1e-6]
Steps:  18%|█▊        | 370/2000 [02:21<09:33,  2.84it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▊        | 371/2000 [02:21<09:39,  2.81it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▊        | 372/2000 [02:21<09:35,  2.83it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▊        | 373/2000 [02:22<09:35,  2.83it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▊        | 374/2000 [02:22<09:33,  2.83it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▉        | 375/2000 [02:22<09:29,  2.85it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▉        | 376/2000 [02:23<09:32,  2.84it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▉        | 377/2000 [02:23<09:31,  2.84it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▉        | 378/2000 [02:24<09:34,  2.82it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▉        | 379/2000 [02:24<09:36,  2.81it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▉        | 380/2000 [02:24<09:35,  2.81it/s, loss=0.284, lr=1e-6]
Steps:  19%|█▉        | 380/2000 [02:25<09:35,  2.81it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 381/2000 [02:25<09:34,  2.82it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 382/2000 [02:25<09:38,  2.80it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 383/2000 [02:25<09:34,  2.82it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 384/2000 [02:26<09:33,  2.82it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 385/2000 [02:26<09:29,  2.84it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 386/2000 [02:26<09:26,  2.85it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 387/2000 [02:27<09:26,  2.85it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 388/2000 [02:27<09:30,  2.83it/s, loss=0.281, lr=1e-6]
Steps:  19%|█▉        | 389/2000 [02:27<09:29,  2.83it/s, loss=0.281, lr=1e-6]
Steps:  20%|█▉        | 390/2000 [02:28<09:27,  2.84it/s, loss=0.281, lr=1e-6]
Steps:  20%|█▉        | 390/2000 [02:28<09:27,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 391/2000 [02:28<09:27,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 392/2000 [02:29<09:29,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 393/2000 [02:29<09:23,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 394/2000 [02:29<09:20,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 395/2000 [02:30<09:23,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 396/2000 [02:30<09:22,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 397/2000 [02:30<09:20,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 398/2000 [02:31<09:18,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  20%|█▉        | 399/2000 [02:31<09:27,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  20%|██        | 400/2000 [02:31<09:30,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  20%|██        | 400/2000 [02:32<09:30,  2.81it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 401/2000 [02:32<09:27,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 402/2000 [02:32<09:23,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 403/2000 [02:32<09:19,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 404/2000 [02:33<09:24,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 405/2000 [02:33<09:20,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 406/2000 [02:33<09:18,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 407/2000 [02:34<09:21,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 408/2000 [02:34<09:24,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 409/2000 [02:34<09:19,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 410/2000 [02:35<09:16,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  20%|██        | 410/2000 [02:35<09:16,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 411/2000 [02:35<09:16,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 412/2000 [02:36<09:17,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 413/2000 [02:36<09:20,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 414/2000 [02:36<09:28,  2.79it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 415/2000 [02:37<09:23,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 416/2000 [02:37<09:23,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 417/2000 [02:37<09:19,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 418/2000 [02:38<09:18,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 419/2000 [02:38<09:15,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 420/2000 [02:38<09:13,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 420/2000 [02:39<09:13,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 421/2000 [02:39<09:14,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 422/2000 [02:39<09:17,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 423/2000 [02:39<09:20,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  21%|██        | 424/2000 [02:40<09:19,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  21%|██▏       | 425/2000 [02:40<09:16,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  21%|██▏       | 426/2000 [02:40<09:14,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  21%|██▏       | 427/2000 [02:41<09:14,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  21%|██▏       | 428/2000 [02:41<09:16,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  21%|██▏       | 429/2000 [02:42<09:14,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  22%|██▏       | 430/2000 [02:42<09:14,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  22%|██▏       | 430/2000 [02:42<09:14,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 431/2000 [02:42<09:15,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 432/2000 [02:43<09:11,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 433/2000 [02:43<09:06,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 434/2000 [02:43<09:07,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 435/2000 [02:44<09:08,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 436/2000 [02:44<09:18,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 437/2000 [02:44<09:12,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 438/2000 [02:45<09:09,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 439/2000 [02:45<09:07,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 440/2000 [02:45<09:02,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  22%|██▏       | 440/2000 [02:46<09:02,  2.87it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 441/2000 [02:46<09:02,  2.88it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 442/2000 [02:46<08:59,  2.89it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 443/2000 [02:46<09:02,  2.87it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 444/2000 [02:47<09:04,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 445/2000 [02:47<09:04,  2.85it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 446/2000 [02:48<09:03,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 447/2000 [02:48<09:04,  2.85it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 448/2000 [02:48<09:03,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▏       | 449/2000 [02:49<09:02,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▎       | 450/2000 [02:49<09:08,  2.82it/s, loss=0.28, lr=1e-6]
Steps:  22%|██▎       | 450/2000 [02:49<09:08,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 451/2000 [02:49<09:08,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 452/2000 [02:50<09:05,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 453/2000 [02:50<09:08,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 454/2000 [02:50<09:06,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 455/2000 [02:51<09:04,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 456/2000 [02:51<09:12,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 457/2000 [02:51<09:09,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 458/2000 [02:52<09:15,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 459/2000 [02:52<09:10,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 460/2000 [02:52<09:07,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  23%|██▎       | 460/2000 [02:53<09:07,  2.81it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 461/2000 [02:53<09:12,  2.78it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 462/2000 [02:53<09:08,  2.81it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 463/2000 [02:54<09:04,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 464/2000 [02:54<09:05,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 465/2000 [02:54<09:08,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 466/2000 [02:55<09:07,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 467/2000 [02:55<09:07,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 468/2000 [02:55<09:06,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  23%|██▎       | 469/2000 [02:56<09:05,  2.81it/s, loss=0.279, lr=1e-6]
Steps:  24%|██▎       | 470/2000 [02:56<08:59,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  24%|██▎       | 470/2000 [02:56<08:59,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▎       | 471/2000 [02:56<08:55,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▎       | 472/2000 [02:57<08:52,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▎       | 473/2000 [02:57<08:51,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▎       | 474/2000 [02:57<08:48,  2.89it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▍       | 475/2000 [02:58<08:54,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▍       | 476/2000 [02:58<08:56,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▍       | 477/2000 [02:58<08:53,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▍       | 478/2000 [02:59<09:00,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▍       | 479/2000 [02:59<09:00,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▍       | 480/2000 [03:00<08:55,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  24%|██▍       | 480/2000 [03:00<08:55,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 481/2000 [03:00<09:00,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 482/2000 [03:00<08:55,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 483/2000 [03:01<08:56,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 484/2000 [03:01<09:00,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 485/2000 [03:01<08:59,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 486/2000 [03:02<08:55,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 487/2000 [03:02<08:54,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 488/2000 [03:02<08:52,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 489/2000 [03:03<08:52,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 490/2000 [03:03<08:50,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  24%|██▍       | 490/2000 [03:03<08:50,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 491/2000 [03:03<08:48,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 492/2000 [03:04<08:50,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 493/2000 [03:04<08:51,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 494/2000 [03:04<08:48,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 495/2000 [03:05<08:45,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 496/2000 [03:05<08:42,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 497/2000 [03:06<08:39,  2.90it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 498/2000 [03:06<08:35,  2.91it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▍       | 499/2000 [03:06<08:35,  2.91it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▌       | 500/2000 [03:07<08:35,  2.91it/s, loss=0.276, lr=1e-6]
Steps:  25%|██▌       | 500/2000 [03:07<08:35,  2.91it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 501/2000 [03:07<08:38,  2.89it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 502/2000 [03:07<08:39,  2.89it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 503/2000 [03:08<08:41,  2.87it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 504/2000 [03:08<08:39,  2.88it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 505/2000 [03:08<08:39,  2.88it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 506/2000 [03:09<08:38,  2.88it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 507/2000 [03:09<08:35,  2.90it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 508/2000 [03:09<08:38,  2.88it/s, loss=0.274, lr=1e-6]
Steps:  25%|██▌       | 509/2000 [03:10<08:39,  2.87it/s, loss=0.274, lr=1e-6]
Steps:  26%|██▌       | 510/2000 [03:10<08:38,  2.87it/s, loss=0.274, lr=1e-6]
Steps:  26%|██▌       | 510/2000 [03:10<08:38,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 511/2000 [03:10<08:38,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 512/2000 [03:11<08:37,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 513/2000 [03:11<08:48,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 514/2000 [03:11<08:43,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 515/2000 [03:12<08:38,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 516/2000 [03:12<08:38,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 517/2000 [03:12<08:37,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 518/2000 [03:13<08:48,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 519/2000 [03:13<08:45,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 520/2000 [03:14<08:41,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  26%|██▌       | 520/2000 [03:14<08:41,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▌       | 521/2000 [03:14<08:41,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▌       | 522/2000 [03:14<08:41,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▌       | 523/2000 [03:15<08:40,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▌       | 524/2000 [03:15<08:41,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▋       | 525/2000 [03:15<08:40,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▋       | 526/2000 [03:16<08:38,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▋       | 527/2000 [03:16<08:39,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▋       | 528/2000 [03:16<08:36,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▋       | 529/2000 [03:17<08:33,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▋       | 530/2000 [03:17<08:32,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  26%|██▋       | 530/2000 [03:17<08:32,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 531/2000 [03:17<08:32,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 532/2000 [03:18<08:33,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 533/2000 [03:18<08:33,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 534/2000 [03:18<08:34,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 535/2000 [03:19<08:33,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 536/2000 [03:19<08:35,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 537/2000 [03:20<08:33,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 538/2000 [03:20<08:29,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 539/2000 [03:20<08:31,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 540/2000 [03:21<08:36,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  27%|██▋       | 540/2000 [03:21<08:36,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 541/2000 [03:21<08:41,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 542/2000 [03:21<08:38,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 543/2000 [03:22<08:36,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 544/2000 [03:22<08:33,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 545/2000 [03:22<08:33,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 546/2000 [03:23<08:31,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 547/2000 [03:23<08:29,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 548/2000 [03:23<08:31,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  27%|██▋       | 549/2000 [03:24<08:29,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  28%|██▊       | 550/2000 [03:24<08:32,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  28%|██▊       | 550/2000 [03:24<08:32,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 551/2000 [03:24<08:33,  2.82it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 552/2000 [03:25<08:28,  2.85it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 553/2000 [03:25<08:27,  2.85it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 554/2000 [03:26<08:26,  2.85it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 555/2000 [03:26<08:25,  2.86it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 556/2000 [03:26<08:24,  2.86it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 557/2000 [03:27<08:27,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 558/2000 [03:27<08:28,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 559/2000 [03:27<08:27,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 560/2000 [03:28<08:24,  2.86it/s, loss=0.274, lr=1e-6]
Steps:  28%|██▊       | 560/2000 [03:28<08:24,  2.86it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 561/2000 [03:28<08:26,  2.84it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 562/2000 [03:28<08:24,  2.85it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 563/2000 [03:29<08:22,  2.86it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 564/2000 [03:29<08:19,  2.87it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 565/2000 [03:29<08:17,  2.89it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 566/2000 [03:30<08:19,  2.87it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 567/2000 [03:30<08:18,  2.87it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 568/2000 [03:30<08:17,  2.88it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 569/2000 [03:31<08:22,  2.85it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 570/2000 [03:31<08:34,  2.78it/s, loss=0.271, lr=1e-6]
Steps:  28%|██▊       | 570/2000 [03:31<08:34,  2.78it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▊       | 571/2000 [03:31<08:29,  2.81it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▊       | 572/2000 [03:32<08:27,  2.81it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▊       | 573/2000 [03:32<08:26,  2.82it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▊       | 574/2000 [03:33<08:23,  2.83it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▉       | 575/2000 [03:33<08:21,  2.84it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▉       | 576/2000 [03:33<08:21,  2.84it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▉       | 577/2000 [03:34<08:24,  2.82it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▉       | 578/2000 [03:34<08:28,  2.80it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▉       | 579/2000 [03:34<08:31,  2.78it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▉       | 580/2000 [03:35<08:30,  2.78it/s, loss=0.273, lr=1e-6]
Steps:  29%|██▉       | 580/2000 [03:35<08:30,  2.78it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 581/2000 [03:35<08:28,  2.79it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 582/2000 [03:35<08:23,  2.81it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 583/2000 [03:36<08:24,  2.81it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 584/2000 [03:36<08:24,  2.81it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 585/2000 [03:36<08:20,  2.83it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 586/2000 [03:37<08:22,  2.81it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 587/2000 [03:37<08:20,  2.82it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 588/2000 [03:38<08:18,  2.83it/s, loss=0.272, lr=1e-6]
Steps:  29%|██▉       | 589/2000 [03:38<08:18,  2.83it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 590/2000 [03:38<08:18,  2.83it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 590/2000 [03:39<08:18,  2.83it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 591/2000 [03:39<08:17,  2.83it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 592/2000 [03:39<08:29,  2.76it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 593/2000 [03:39<08:27,  2.77it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 594/2000 [03:40<08:21,  2.80it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 595/2000 [03:40<08:17,  2.83it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 596/2000 [03:40<08:13,  2.84it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 597/2000 [03:41<08:12,  2.85it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 598/2000 [03:41<08:18,  2.81it/s, loss=0.272, lr=1e-6]
Steps:  30%|██▉       | 599/2000 [03:41<08:12,  2.84it/s, loss=0.272, lr=1e-6]
Steps:  30%|███       | 600/2000 [03:42<08:09,  2.86it/s, loss=0.272, lr=1e-6]
Steps:  30%|███       | 600/2000 [03:42<08:09,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 601/2000 [03:42<08:10,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 602/2000 [03:42<08:09,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 603/2000 [03:43<08:08,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 604/2000 [03:43<08:07,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 605/2000 [03:44<08:05,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 606/2000 [03:44<08:05,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 607/2000 [03:44<08:07,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 608/2000 [03:45<08:07,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 609/2000 [03:45<08:12,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 610/2000 [03:45<08:08,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  30%|███       | 610/2000 [03:46<08:08,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 611/2000 [03:46<08:08,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 612/2000 [03:46<08:10,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 613/2000 [03:46<08:13,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 614/2000 [03:47<08:09,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 615/2000 [03:47<08:09,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 616/2000 [03:47<08:04,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 617/2000 [03:48<08:03,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 618/2000 [03:48<08:02,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 619/2000 [03:48<08:02,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 620/2000 [03:49<08:01,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  31%|███       | 620/2000 [03:49<08:01,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  31%|███       | 621/2000 [03:49<08:06,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  31%|███       | 622/2000 [03:49<08:01,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  31%|███       | 623/2000 [03:50<08:02,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  31%|███       | 624/2000 [03:50<08:03,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  31%|███▏      | 625/2000 [03:51<08:02,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  31%|███▏      | 626/2000 [03:51<08:10,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  31%|███▏      | 627/2000 [03:51<08:07,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  31%|███▏      | 628/2000 [03:52<08:02,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  31%|███▏      | 629/2000 [03:52<08:12,  2.79it/s, loss=0.279, lr=1e-6]
Steps:  32%|███▏      | 630/2000 [03:52<08:06,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  32%|███▏      | 630/2000 [03:53<08:06,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 631/2000 [03:53<08:04,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 632/2000 [03:53<08:03,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 633/2000 [03:53<07:58,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 634/2000 [03:54<07:58,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 635/2000 [03:54<08:01,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 636/2000 [03:54<07:57,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 637/2000 [03:55<07:55,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 638/2000 [03:55<07:55,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 639/2000 [03:55<07:51,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 640/2000 [03:56<07:51,  2.89it/s, loss=0.277, lr=1e-6]
Steps:  32%|███▏      | 640/2000 [03:56<07:51,  2.89it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 641/2000 [03:56<07:55,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 642/2000 [03:57<07:53,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 643/2000 [03:57<07:49,  2.89it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 644/2000 [03:57<07:49,  2.89it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 645/2000 [03:58<07:47,  2.90it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 646/2000 [03:58<07:48,  2.89it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 647/2000 [03:58<07:49,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 648/2000 [03:59<07:48,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▏      | 649/2000 [03:59<07:52,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▎      | 650/2000 [03:59<07:49,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  32%|███▎      | 650/2000 [04:00<07:49,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 651/2000 [04:00<07:48,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 652/2000 [04:00<07:53,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 653/2000 [04:00<07:50,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 654/2000 [04:01<07:56,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 655/2000 [04:01<07:59,  2.80it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 656/2000 [04:01<07:55,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 657/2000 [04:02<07:54,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 658/2000 [04:02<07:51,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 659/2000 [04:02<07:53,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 660/2000 [04:03<07:54,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  33%|███▎      | 660/2000 [04:03<07:54,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 661/2000 [04:03<07:51,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 662/2000 [04:04<07:47,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 663/2000 [04:04<07:51,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 664/2000 [04:04<07:54,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 665/2000 [04:05<07:56,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 666/2000 [04:05<07:57,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 667/2000 [04:05<08:14,  2.70it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 668/2000 [04:06<08:07,  2.73it/s, loss=0.276, lr=1e-6]
Steps:  33%|███▎      | 669/2000 [04:06<08:06,  2.74it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▎      | 670/2000 [04:06<07:58,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▎      | 670/2000 [04:07<07:58,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▎      | 671/2000 [04:07<07:54,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▎      | 672/2000 [04:07<07:49,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▎      | 673/2000 [04:07<07:47,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▎      | 674/2000 [04:08<07:43,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▍      | 675/2000 [04:08<07:43,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▍      | 676/2000 [04:09<07:43,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▍      | 677/2000 [04:09<07:41,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▍      | 678/2000 [04:09<07:39,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▍      | 679/2000 [04:10<07:39,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▍      | 680/2000 [04:10<07:35,  2.90it/s, loss=0.276, lr=1e-6]
Steps:  34%|███▍      | 680/2000 [04:10<07:35,  2.90it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 681/2000 [04:10<07:35,  2.89it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 682/2000 [04:11<07:34,  2.90it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 683/2000 [04:11<07:42,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 684/2000 [04:11<07:39,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 685/2000 [04:12<07:37,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 686/2000 [04:12<07:37,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 687/2000 [04:12<07:38,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 688/2000 [04:13<07:42,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 689/2000 [04:13<07:39,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 690/2000 [04:13<07:37,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  34%|███▍      | 690/2000 [04:14<07:37,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 691/2000 [04:14<07:43,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 692/2000 [04:14<07:42,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 693/2000 [04:14<07:42,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 694/2000 [04:15<07:39,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 695/2000 [04:15<07:37,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 696/2000 [04:15<07:35,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 697/2000 [04:16<07:40,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 698/2000 [04:16<07:37,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▍      | 699/2000 [04:17<07:37,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 700/2000 [04:17<07:35,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 700/2000 [04:17<07:35,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 701/2000 [04:17<07:35,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 702/2000 [04:18<07:34,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 703/2000 [04:18<07:33,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 704/2000 [04:18<07:33,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 705/2000 [04:19<07:33,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 706/2000 [04:19<07:37,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 707/2000 [04:19<07:35,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 708/2000 [04:20<07:32,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  35%|███▌      | 709/2000 [04:20<07:35,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▌      | 710/2000 [04:20<07:32,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▌      | 710/2000 [04:21<07:32,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 711/2000 [04:21<07:37,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 712/2000 [04:21<07:40,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 713/2000 [04:21<07:34,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 714/2000 [04:22<07:41,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 715/2000 [04:22<07:36,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 716/2000 [04:23<07:35,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 717/2000 [04:23<07:33,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 718/2000 [04:23<07:32,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 719/2000 [04:24<07:30,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 720/2000 [04:24<07:34,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  36%|███▌      | 720/2000 [04:24<07:34,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▌      | 721/2000 [04:24<07:31,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▌      | 722/2000 [04:25<07:27,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▌      | 723/2000 [04:25<07:31,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▌      | 724/2000 [04:25<07:27,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▋      | 725/2000 [04:26<07:27,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▋      | 726/2000 [04:26<07:24,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▋      | 727/2000 [04:26<07:26,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▋      | 728/2000 [04:27<07:27,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▋      | 729/2000 [04:27<07:29,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▋      | 730/2000 [04:28<07:34,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  36%|███▋      | 730/2000 [04:28<07:34,  2.79it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 731/2000 [04:28<07:33,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 732/2000 [04:28<07:32,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 733/2000 [04:29<07:30,  2.81it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 734/2000 [04:29<07:29,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 735/2000 [04:29<07:26,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 736/2000 [04:30<07:24,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 737/2000 [04:30<07:23,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 738/2000 [04:30<07:24,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 739/2000 [04:31<07:22,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 740/2000 [04:31<07:32,  2.79it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 740/2000 [04:31<07:32,  2.79it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 741/2000 [04:31<07:29,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 742/2000 [04:32<07:27,  2.81it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 743/2000 [04:32<07:29,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 744/2000 [04:32<07:28,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 745/2000 [04:33<07:26,  2.81it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 746/2000 [04:33<07:24,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 747/2000 [04:34<07:19,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 748/2000 [04:34<07:15,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  37%|███▋      | 749/2000 [04:34<07:17,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 750/2000 [04:35<07:21,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 750/2000 [04:35<07:21,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 751/2000 [04:35<07:19,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 752/2000 [04:35<07:17,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 753/2000 [04:36<07:15,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 754/2000 [04:36<07:19,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 755/2000 [04:36<07:17,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 756/2000 [04:37<07:16,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 757/2000 [04:37<07:16,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 758/2000 [04:37<07:14,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 759/2000 [04:38<07:16,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 760/2000 [04:38<07:13,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 760/2000 [04:38<07:13,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 761/2000 [04:38<07:10,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 762/2000 [04:39<07:12,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 763/2000 [04:39<07:11,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 764/2000 [04:39<07:10,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 765/2000 [04:40<07:06,  2.90it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 766/2000 [04:40<07:07,  2.89it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 767/2000 [04:40<07:07,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 768/2000 [04:41<07:07,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 769/2000 [04:41<07:09,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 770/2000 [04:42<07:07,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  38%|███▊      | 770/2000 [04:42<07:07,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▊      | 771/2000 [04:42<07:06,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▊      | 772/2000 [04:42<07:06,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▊      | 773/2000 [04:43<07:04,  2.89it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▊      | 774/2000 [04:43<07:02,  2.90it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 775/2000 [04:43<07:02,  2.90it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 776/2000 [04:44<07:00,  2.91it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 777/2000 [04:44<07:10,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 778/2000 [04:44<07:13,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 779/2000 [04:45<07:19,  2.78it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 780/2000 [04:45<07:18,  2.78it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 780/2000 [04:45<07:18,  2.78it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 781/2000 [04:45<07:12,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 782/2000 [04:46<07:16,  2.79it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 783/2000 [04:46<07:10,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 784/2000 [04:46<07:09,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 785/2000 [04:47<07:08,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 786/2000 [04:47<07:09,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 787/2000 [04:48<07:07,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 788/2000 [04:48<07:13,  2.79it/s, loss=0.279, lr=1e-6]
Steps:  39%|███▉      | 789/2000 [04:48<07:10,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  40%|███▉      | 790/2000 [04:49<07:07,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  40%|███▉      | 790/2000 [04:49<07:07,  2.83it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 791/2000 [04:49<07:06,  2.83it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 792/2000 [04:49<07:06,  2.83it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 793/2000 [04:50<07:03,  2.85it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 794/2000 [04:50<07:00,  2.87it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 795/2000 [04:50<06:58,  2.88it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 796/2000 [04:51<06:58,  2.88it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 797/2000 [04:51<07:08,  2.81it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 798/2000 [04:51<07:06,  2.82it/s, loss=0.28, lr=1e-6]
Steps:  40%|███▉      | 799/2000 [04:52<07:07,  2.81it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 800/2000 [04:52<07:06,  2.82it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 800/2000 [04:52<07:06,  2.82it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 801/2000 [04:52<07:06,  2.81it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 802/2000 [04:53<07:05,  2.82it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 803/2000 [04:53<07:01,  2.84it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 804/2000 [04:54<06:57,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 805/2000 [04:54<06:57,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 806/2000 [04:54<06:55,  2.87it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 807/2000 [04:55<07:03,  2.82it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 808/2000 [04:55<07:01,  2.82it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 809/2000 [04:55<06:57,  2.85it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 810/2000 [04:56<06:56,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  40%|████      | 810/2000 [04:56<06:56,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 811/2000 [04:56<06:53,  2.88it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 812/2000 [04:56<06:53,  2.87it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 813/2000 [04:57<06:52,  2.88it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 814/2000 [04:57<07:02,  2.80it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 815/2000 [04:57<06:58,  2.83it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 816/2000 [04:58<06:56,  2.84it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 817/2000 [04:58<06:55,  2.84it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 818/2000 [04:58<06:55,  2.84it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 819/2000 [04:59<06:52,  2.87it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 820/2000 [04:59<06:52,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 820/2000 [04:59<06:52,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 821/2000 [04:59<06:51,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 822/2000 [05:00<06:51,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 823/2000 [05:00<06:51,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  41%|████      | 824/2000 [05:01<06:49,  2.87it/s, loss=0.281, lr=1e-6]
Steps:  41%|████▏     | 825/2000 [05:01<06:51,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  41%|████▏     | 826/2000 [05:01<06:52,  2.85it/s, loss=0.281, lr=1e-6]
Steps:  41%|████▏     | 827/2000 [05:02<06:50,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  41%|████▏     | 828/2000 [05:02<06:48,  2.87it/s, loss=0.281, lr=1e-6]
Steps:  41%|████▏     | 829/2000 [05:02<06:49,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  42%|████▏     | 830/2000 [05:03<06:49,  2.86it/s, loss=0.281, lr=1e-6]
Steps:  42%|████▏     | 830/2000 [05:03<06:49,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 831/2000 [05:03<06:53,  2.83it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 832/2000 [05:03<06:50,  2.84it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 833/2000 [05:04<06:48,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 834/2000 [05:04<06:47,  2.86it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 835/2000 [05:04<06:50,  2.84it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 836/2000 [05:05<06:51,  2.83it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 837/2000 [05:05<06:48,  2.85it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 838/2000 [05:05<06:45,  2.87it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 839/2000 [05:06<06:44,  2.87it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 840/2000 [05:06<06:43,  2.87it/s, loss=0.28, lr=1e-6]
Steps:  42%|████▏     | 840/2000 [05:06<06:43,  2.87it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 841/2000 [05:06<06:45,  2.86it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 842/2000 [05:07<06:44,  2.86it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 843/2000 [05:07<06:47,  2.84it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 844/2000 [05:08<06:45,  2.85it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 845/2000 [05:08<06:43,  2.86it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 846/2000 [05:08<06:42,  2.87it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 847/2000 [05:09<06:43,  2.86it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 848/2000 [05:09<06:41,  2.87it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▏     | 849/2000 [05:09<06:40,  2.88it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▎     | 850/2000 [05:10<06:40,  2.87it/s, loss=0.282, lr=1e-6]
Steps:  42%|████▎     | 850/2000 [05:10<06:40,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 851/2000 [05:10<06:43,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 852/2000 [05:10<06:42,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 853/2000 [05:11<06:40,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 854/2000 [05:11<06:46,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 855/2000 [05:11<06:43,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 856/2000 [05:12<06:42,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 857/2000 [05:12<06:42,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 858/2000 [05:12<06:43,  2.83it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 859/2000 [05:13<06:44,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 860/2000 [05:13<06:46,  2.81it/s, loss=0.279, lr=1e-6]
Steps:  43%|████▎     | 860/2000 [05:14<06:46,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 861/2000 [05:14<06:42,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 862/2000 [05:14<06:43,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 863/2000 [05:14<06:43,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 864/2000 [05:15<06:45,  2.80it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 865/2000 [05:15<06:46,  2.79it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 866/2000 [05:15<06:45,  2.80it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 867/2000 [05:16<06:40,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 868/2000 [05:16<06:43,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  43%|████▎     | 869/2000 [05:16<06:40,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  44%|████▎     | 870/2000 [05:17<06:37,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  44%|████▎     | 870/2000 [05:17<06:37,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▎     | 871/2000 [05:17<06:37,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▎     | 872/2000 [05:17<06:33,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▎     | 873/2000 [05:18<06:33,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▎     | 874/2000 [05:18<06:32,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 875/2000 [05:18<06:31,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 876/2000 [05:19<06:31,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 877/2000 [05:19<06:32,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 878/2000 [05:20<06:30,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 879/2000 [05:20<06:29,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 880/2000 [05:20<06:29,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 880/2000 [05:21<06:29,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 881/2000 [05:21<06:32,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 882/2000 [05:21<06:34,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 883/2000 [05:21<06:35,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 884/2000 [05:22<06:34,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 885/2000 [05:22<06:33,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 886/2000 [05:22<06:31,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 887/2000 [05:23<06:31,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 888/2000 [05:23<06:30,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 889/2000 [05:23<06:29,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 890/2000 [05:24<06:29,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  44%|████▍     | 890/2000 [05:24<06:29,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 891/2000 [05:24<06:28,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 892/2000 [05:24<06:27,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 893/2000 [05:25<06:30,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 894/2000 [05:25<06:28,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 895/2000 [05:25<06:26,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 896/2000 [05:26<06:27,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 897/2000 [05:26<06:26,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 898/2000 [05:27<06:24,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▍     | 899/2000 [05:27<06:27,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▌     | 900/2000 [05:27<06:27,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  45%|████▌     | 900/2000 [05:28<06:27,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 901/2000 [05:28<06:27,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 902/2000 [05:28<06:23,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 903/2000 [05:28<06:22,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 904/2000 [05:29<06:23,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 905/2000 [05:29<06:21,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 906/2000 [05:29<06:20,  2.88it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 907/2000 [05:30<06:21,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 908/2000 [05:30<06:20,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  45%|████▌     | 909/2000 [05:30<06:19,  2.88it/s, loss=0.275, lr=1e-6]
Steps:  46%|████▌     | 910/2000 [05:31<06:19,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  46%|████▌     | 910/2000 [05:31<06:19,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 911/2000 [05:31<06:26,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 912/2000 [05:31<06:25,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 913/2000 [05:32<06:24,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 914/2000 [05:32<06:25,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 915/2000 [05:33<06:24,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 916/2000 [05:33<06:23,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 917/2000 [05:33<06:21,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 918/2000 [05:34<06:21,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 919/2000 [05:34<06:31,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 920/2000 [05:34<06:30,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  46%|████▌     | 920/2000 [05:35<06:30,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▌     | 921/2000 [05:35<06:26,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▌     | 922/2000 [05:35<06:24,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▌     | 923/2000 [05:35<06:20,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▌     | 924/2000 [05:36<06:19,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▋     | 925/2000 [05:36<06:16,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▋     | 926/2000 [05:36<06:14,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▋     | 927/2000 [05:37<06:14,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▋     | 928/2000 [05:37<06:18,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▋     | 929/2000 [05:37<06:16,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▋     | 930/2000 [05:38<06:17,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  46%|████▋     | 930/2000 [05:38<06:17,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 931/2000 [05:38<06:15,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 932/2000 [05:39<06:14,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 933/2000 [05:39<06:11,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 934/2000 [05:39<06:11,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 935/2000 [05:40<06:11,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 936/2000 [05:40<06:09,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 937/2000 [05:40<06:15,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 938/2000 [05:41<06:15,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 939/2000 [05:41<06:15,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 940/2000 [05:41<06:13,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  47%|████▋     | 940/2000 [05:42<06:13,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 941/2000 [05:42<06:12,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 942/2000 [05:42<06:12,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 943/2000 [05:42<06:09,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 944/2000 [05:43<06:12,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 945/2000 [05:43<06:09,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 946/2000 [05:43<06:07,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 947/2000 [05:44<06:06,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 948/2000 [05:44<06:06,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  47%|████▋     | 949/2000 [05:44<06:05,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  48%|████▊     | 950/2000 [05:45<06:10,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  48%|████▊     | 950/2000 [05:45<06:10,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 951/2000 [05:45<06:11,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 952/2000 [05:46<06:07,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 953/2000 [05:46<06:06,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 954/2000 [05:46<06:05,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 955/2000 [05:47<06:06,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 956/2000 [05:47<06:05,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 957/2000 [05:47<06:04,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 958/2000 [05:48<06:05,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 959/2000 [05:48<06:03,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 960/2000 [05:48<06:03,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 960/2000 [05:49<06:03,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 961/2000 [05:49<06:03,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 962/2000 [05:49<06:05,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 963/2000 [05:49<06:03,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 964/2000 [05:50<06:04,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 965/2000 [05:50<06:10,  2.79it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 966/2000 [05:50<06:06,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 967/2000 [05:51<06:06,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 968/2000 [05:51<06:14,  2.76it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 969/2000 [05:52<06:10,  2.78it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 970/2000 [05:52<06:07,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  48%|████▊     | 970/2000 [05:52<06:07,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▊     | 971/2000 [05:52<06:06,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▊     | 972/2000 [05:53<06:05,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▊     | 973/2000 [05:53<06:05,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▊     | 974/2000 [05:53<06:03,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 975/2000 [05:54<06:03,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 976/2000 [05:54<06:07,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 977/2000 [05:54<06:03,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 978/2000 [05:55<06:02,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 979/2000 [05:55<06:02,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 980/2000 [05:55<06:00,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 980/2000 [05:56<06:00,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 981/2000 [05:56<06:02,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 982/2000 [05:56<06:02,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 983/2000 [05:56<05:59,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 984/2000 [05:57<06:01,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 985/2000 [05:57<06:00,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 986/2000 [05:58<05:58,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 987/2000 [05:58<06:00,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 988/2000 [05:58<05:58,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  49%|████▉     | 989/2000 [05:59<05:57,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 990/2000 [05:59<06:01,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 990/2000 [05:59<06:01,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 991/2000 [05:59<05:57,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 992/2000 [06:00<05:55,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 993/2000 [06:00<05:55,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 994/2000 [06:00<05:54,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 995/2000 [06:01<05:59,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 996/2000 [06:01<06:03,  2.76it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 997/2000 [06:01<05:58,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 998/2000 [06:02<05:58,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  50%|████▉     | 999/2000 [06:02<05:57,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1000/2000 [06:03<05:56,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1000/2000 [06:03<05:56,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1001/2000 [06:03<05:56,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1002/2000 [06:03<05:53,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1003/2000 [06:04<05:50,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1004/2000 [06:04<05:50,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1005/2000 [06:04<05:52,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1006/2000 [06:05<05:53,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1007/2000 [06:05<05:57,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1008/2000 [06:05<05:53,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1009/2000 [06:06<05:52,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1010/2000 [06:06<05:52,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  50%|█████     | 1010/2000 [06:06<05:52,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1011/2000 [06:06<05:51,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1012/2000 [06:07<05:48,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1013/2000 [06:07<05:46,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1014/2000 [06:07<05:44,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1015/2000 [06:08<05:44,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1016/2000 [06:08<05:45,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1017/2000 [06:09<05:44,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1018/2000 [06:09<05:45,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1019/2000 [06:09<05:47,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1020/2000 [06:10<05:44,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  51%|█████     | 1020/2000 [06:10<05:44,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████     | 1021/2000 [06:10<05:43,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████     | 1022/2000 [06:10<05:45,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████     | 1023/2000 [06:11<05:43,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████     | 1024/2000 [06:11<05:47,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████▏    | 1025/2000 [06:11<05:43,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████▏    | 1026/2000 [06:12<05:42,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████▏    | 1027/2000 [06:12<05:43,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████▏    | 1028/2000 [06:12<05:41,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  51%|█████▏    | 1029/2000 [06:13<05:40,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  52%|█████▏    | 1030/2000 [06:13<05:39,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  52%|█████▏    | 1030/2000 [06:13<05:39,  2.86it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1031/2000 [06:13<05:38,  2.86it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1032/2000 [06:14<05:41,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1033/2000 [06:14<05:40,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1034/2000 [06:15<05:41,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1035/2000 [06:15<05:40,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1036/2000 [06:15<05:40,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1037/2000 [06:16<05:40,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1038/2000 [06:16<05:39,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1039/2000 [06:16<05:38,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1040/2000 [06:17<05:38,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1040/2000 [06:17<05:38,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1041/2000 [06:17<05:37,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1042/2000 [06:17<05:39,  2.82it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1043/2000 [06:18<05:47,  2.75it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1044/2000 [06:18<05:52,  2.71it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1045/2000 [06:18<05:45,  2.77it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1046/2000 [06:19<05:50,  2.72it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1047/2000 [06:19<05:46,  2.75it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1048/2000 [06:20<05:41,  2.79it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▏    | 1049/2000 [06:20<05:38,  2.81it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▎    | 1050/2000 [06:20<05:36,  2.82it/s, loss=0.274, lr=1e-6]
Steps:  52%|█████▎    | 1050/2000 [06:21<05:36,  2.82it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1051/2000 [06:21<05:38,  2.80it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1052/2000 [06:21<05:42,  2.77it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1053/2000 [06:21<05:37,  2.81it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1054/2000 [06:22<05:35,  2.82it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1055/2000 [06:22<05:33,  2.83it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1056/2000 [06:22<05:31,  2.85it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1057/2000 [06:23<05:31,  2.85it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1058/2000 [06:23<05:28,  2.87it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1059/2000 [06:23<05:26,  2.88it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1060/2000 [06:24<05:26,  2.88it/s, loss=0.273, lr=1e-6]
Steps:  53%|█████▎    | 1060/2000 [06:24<05:26,  2.88it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1061/2000 [06:24<05:32,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1062/2000 [06:24<05:31,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1063/2000 [06:25<05:29,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1064/2000 [06:25<05:30,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1065/2000 [06:26<05:30,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1066/2000 [06:26<05:30,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1067/2000 [06:26<05:29,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1068/2000 [06:27<05:28,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  53%|█████▎    | 1069/2000 [06:27<05:28,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  54%|█████▎    | 1070/2000 [06:27<05:26,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  54%|█████▎    | 1070/2000 [06:28<05:26,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▎    | 1071/2000 [06:28<05:27,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▎    | 1072/2000 [06:28<05:26,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▎    | 1073/2000 [06:28<05:24,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▎    | 1074/2000 [06:29<05:24,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1075/2000 [06:29<05:22,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1076/2000 [06:29<05:19,  2.89it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1077/2000 [06:30<05:19,  2.89it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1078/2000 [06:30<05:19,  2.89it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1079/2000 [06:30<05:19,  2.88it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1080/2000 [06:31<05:22,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1080/2000 [06:31<05:22,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1081/2000 [06:31<05:29,  2.79it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1082/2000 [06:32<05:26,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1083/2000 [06:32<05:24,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1084/2000 [06:32<05:23,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1085/2000 [06:33<05:21,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1086/2000 [06:33<05:22,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1087/2000 [06:33<05:21,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1088/2000 [06:34<05:19,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  54%|█████▍    | 1089/2000 [06:34<05:22,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1090/2000 [06:34<05:21,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1090/2000 [06:35<05:21,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1091/2000 [06:35<05:21,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1092/2000 [06:35<05:19,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1093/2000 [06:35<05:17,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1094/2000 [06:36<05:15,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1095/2000 [06:36<05:16,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1096/2000 [06:36<05:16,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1097/2000 [06:37<05:17,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1098/2000 [06:37<05:17,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▍    | 1099/2000 [06:37<05:15,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1100/2000 [06:38<05:21,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1100/2000 [06:38<05:21,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1101/2000 [06:38<05:21,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1102/2000 [06:39<05:18,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1103/2000 [06:39<05:18,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1104/2000 [06:39<05:16,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1105/2000 [06:40<05:15,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1106/2000 [06:40<05:14,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1107/2000 [06:40<05:15,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1108/2000 [06:41<05:14,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  55%|█████▌    | 1109/2000 [06:41<05:20,  2.78it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1110/2000 [06:41<05:17,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1110/2000 [06:42<05:17,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1111/2000 [06:42<05:18,  2.79it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1112/2000 [06:42<05:17,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1113/2000 [06:42<05:16,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1114/2000 [06:43<05:18,  2.78it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1115/2000 [06:43<05:18,  2.78it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1116/2000 [06:44<05:15,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1117/2000 [06:44<05:17,  2.78it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1118/2000 [06:44<05:15,  2.79it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1119/2000 [06:45<05:12,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1120/2000 [06:45<05:11,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1120/2000 [06:45<05:11,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1121/2000 [06:45<05:09,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1122/2000 [06:46<05:07,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1123/2000 [06:46<05:07,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▌    | 1124/2000 [06:46<05:06,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▋    | 1125/2000 [06:47<05:05,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▋    | 1126/2000 [06:47<05:15,  2.77it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▋    | 1127/2000 [06:47<05:11,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▋    | 1128/2000 [06:48<05:13,  2.78it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▋    | 1129/2000 [06:48<05:10,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▋    | 1130/2000 [06:49<05:06,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  56%|█████▋    | 1130/2000 [06:49<05:06,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1131/2000 [06:49<05:04,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1132/2000 [06:49<05:04,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1133/2000 [06:50<05:02,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1134/2000 [06:50<05:01,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1135/2000 [06:50<05:02,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1136/2000 [06:51<05:02,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1137/2000 [06:51<05:03,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1138/2000 [06:51<05:01,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1139/2000 [06:52<04:59,  2.88it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1140/2000 [06:52<04:59,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1140/2000 [06:52<04:59,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1141/2000 [06:52<04:59,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1142/2000 [06:53<05:00,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1143/2000 [06:53<04:58,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1144/2000 [06:53<04:57,  2.88it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1145/2000 [06:54<04:56,  2.88it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1146/2000 [06:54<04:59,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1147/2000 [06:54<04:58,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1148/2000 [06:55<04:57,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▋    | 1149/2000 [06:55<04:57,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▊    | 1150/2000 [06:55<04:56,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  57%|█████▊    | 1150/2000 [06:56<04:56,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1151/2000 [06:56<04:57,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1152/2000 [06:56<04:56,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1153/2000 [06:57<04:56,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1154/2000 [06:57<04:54,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1155/2000 [06:57<04:55,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1156/2000 [06:58<04:54,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1157/2000 [06:58<04:52,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1158/2000 [06:58<04:52,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1159/2000 [06:59<04:51,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1160/2000 [06:59<04:50,  2.89it/s, loss=0.276, lr=1e-6]
Steps:  58%|█████▊    | 1160/2000 [06:59<04:50,  2.89it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1161/2000 [06:59<04:49,  2.90it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1162/2000 [07:00<04:51,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1163/2000 [07:00<04:53,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1164/2000 [07:00<04:51,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1165/2000 [07:01<04:52,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1166/2000 [07:01<05:00,  2.78it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1167/2000 [07:01<04:57,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1168/2000 [07:02<04:54,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1169/2000 [07:02<04:53,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1170/2000 [07:03<04:53,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  58%|█████▊    | 1170/2000 [07:03<04:53,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▊    | 1171/2000 [07:03<04:52,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▊    | 1172/2000 [07:03<04:51,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▊    | 1173/2000 [07:04<04:51,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▊    | 1174/2000 [07:04<04:50,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1175/2000 [07:04<04:52,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1176/2000 [07:05<04:57,  2.77it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1177/2000 [07:05<04:56,  2.78it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1178/2000 [07:05<04:52,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1179/2000 [07:06<04:49,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1180/2000 [07:06<04:48,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1180/2000 [07:06<04:48,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1181/2000 [07:06<04:48,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1182/2000 [07:07<04:47,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1183/2000 [07:07<04:45,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1184/2000 [07:07<04:45,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1185/2000 [07:08<04:49,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1186/2000 [07:08<04:47,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1187/2000 [07:09<04:46,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1188/2000 [07:09<04:45,  2.85it/s, loss=0.275, lr=1e-6]
Steps:  59%|█████▉    | 1189/2000 [07:09<04:42,  2.87it/s, loss=0.275, lr=1e-6]
Steps:  60%|█████▉    | 1190/2000 [07:10<04:43,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  60%|█████▉    | 1190/2000 [07:10<04:43,  2.86it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1191/2000 [07:10<04:41,  2.88it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1192/2000 [07:10<04:41,  2.87it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1193/2000 [07:11<04:39,  2.89it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1194/2000 [07:11<04:39,  2.89it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1195/2000 [07:11<04:40,  2.87it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1196/2000 [07:12<04:41,  2.86it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1197/2000 [07:12<04:39,  2.88it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1198/2000 [07:12<04:37,  2.89it/s, loss=0.274, lr=1e-6]
Steps:  60%|█████▉    | 1199/2000 [07:13<04:42,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1200/2000 [07:13<04:41,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1200/2000 [07:13<04:41,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1201/2000 [07:13<04:40,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1202/2000 [07:14<04:40,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1203/2000 [07:14<04:41,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1204/2000 [07:14<04:40,  2.84it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1205/2000 [07:15<04:43,  2.81it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1206/2000 [07:15<04:43,  2.80it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1207/2000 [07:16<04:41,  2.82it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1208/2000 [07:16<04:41,  2.81it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1209/2000 [07:16<04:39,  2.83it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1210/2000 [07:17<04:36,  2.86it/s, loss=0.274, lr=1e-6]
Steps:  60%|██████    | 1210/2000 [07:17<04:36,  2.86it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1211/2000 [07:17<04:39,  2.82it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1212/2000 [07:17<04:37,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1213/2000 [07:18<04:36,  2.84it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1214/2000 [07:18<04:41,  2.79it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1215/2000 [07:18<04:41,  2.79it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1216/2000 [07:19<04:41,  2.79it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1217/2000 [07:19<04:38,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1218/2000 [07:19<04:39,  2.80it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1219/2000 [07:20<04:37,  2.81it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1220/2000 [07:20<04:35,  2.83it/s, loss=0.275, lr=1e-6]
Steps:  61%|██████    | 1220/2000 [07:20<04:35,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████    | 1221/2000 [07:20<04:35,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████    | 1222/2000 [07:21<04:39,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████    | 1223/2000 [07:21<04:39,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████    | 1224/2000 [07:22<04:36,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████▏   | 1225/2000 [07:22<04:36,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████▏   | 1226/2000 [07:22<04:35,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████▏   | 1227/2000 [07:23<04:32,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████▏   | 1228/2000 [07:23<04:34,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  61%|██████▏   | 1229/2000 [07:23<04:33,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1230/2000 [07:24<04:33,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1230/2000 [07:24<04:33,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1231/2000 [07:24<04:35,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1232/2000 [07:24<04:34,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1233/2000 [07:25<04:33,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1234/2000 [07:25<04:31,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1235/2000 [07:25<04:31,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1236/2000 [07:26<04:30,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1237/2000 [07:26<04:29,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1238/2000 [07:27<04:28,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1239/2000 [07:27<04:27,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1240/2000 [07:27<04:27,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1240/2000 [07:28<04:27,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1241/2000 [07:28<04:25,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1242/2000 [07:28<04:23,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1243/2000 [07:28<04:24,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1244/2000 [07:29<04:23,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1245/2000 [07:29<04:24,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1246/2000 [07:29<04:22,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1247/2000 [07:30<04:22,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1248/2000 [07:30<04:21,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▏   | 1249/2000 [07:30<04:19,  2.89it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▎   | 1250/2000 [07:31<04:19,  2.89it/s, loss=0.276, lr=1e-6]
Steps:  62%|██████▎   | 1250/2000 [07:31<04:19,  2.89it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1251/2000 [07:31<04:26,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1252/2000 [07:31<04:24,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1253/2000 [07:32<04:27,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1254/2000 [07:32<04:25,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1255/2000 [07:33<04:22,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1256/2000 [07:33<04:22,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1257/2000 [07:33<04:22,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1258/2000 [07:34<04:20,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1259/2000 [07:34<04:21,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1260/2000 [07:34<04:21,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  63%|██████▎   | 1260/2000 [07:35<04:21,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1261/2000 [07:35<04:24,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1262/2000 [07:35<04:29,  2.74it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1263/2000 [07:35<04:26,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1264/2000 [07:36<04:22,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1265/2000 [07:36<04:20,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1266/2000 [07:36<04:18,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1267/2000 [07:37<04:17,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1268/2000 [07:37<04:16,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  63%|██████▎   | 1269/2000 [07:37<04:15,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▎   | 1270/2000 [07:38<04:13,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▎   | 1270/2000 [07:38<04:13,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▎   | 1271/2000 [07:38<04:14,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▎   | 1272/2000 [07:39<04:14,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▎   | 1273/2000 [07:39<04:13,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▎   | 1274/2000 [07:39<04:12,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▍   | 1275/2000 [07:40<04:15,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▍   | 1276/2000 [07:40<04:13,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▍   | 1277/2000 [07:40<04:13,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▍   | 1278/2000 [07:41<04:11,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▍   | 1279/2000 [07:41<04:15,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▍   | 1280/2000 [07:41<04:13,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  64%|██████▍   | 1280/2000 [07:42<04:13,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1281/2000 [07:42<04:12,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1282/2000 [07:42<04:11,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1283/2000 [07:42<04:10,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1284/2000 [07:43<04:10,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1285/2000 [07:43<04:10,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1286/2000 [07:43<04:09,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1287/2000 [07:44<04:10,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1288/2000 [07:44<04:12,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1289/2000 [07:44<04:13,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1290/2000 [07:45<04:14,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  64%|██████▍   | 1290/2000 [07:45<04:14,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1291/2000 [07:45<04:16,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1292/2000 [07:46<04:18,  2.74it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1293/2000 [07:46<04:21,  2.70it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1294/2000 [07:46<04:21,  2.69it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1295/2000 [07:47<04:21,  2.69it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1296/2000 [07:47<04:19,  2.71it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1297/2000 [07:47<04:17,  2.73it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1298/2000 [07:48<04:16,  2.74it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▍   | 1299/2000 [07:48<04:11,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1300/2000 [07:49<04:09,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1300/2000 [07:49<04:09,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1301/2000 [07:49<04:08,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1302/2000 [07:49<04:06,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1303/2000 [07:50<04:06,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1304/2000 [07:50<04:06,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1305/2000 [07:50<04:05,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1306/2000 [07:51<04:05,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1307/2000 [07:51<04:07,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1308/2000 [07:51<04:07,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  65%|██████▌   | 1309/2000 [07:52<04:06,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1310/2000 [07:52<04:05,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1310/2000 [07:52<04:05,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1311/2000 [07:52<04:04,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1312/2000 [07:53<04:04,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1313/2000 [07:53<04:03,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1314/2000 [07:53<04:01,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1315/2000 [07:54<04:01,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1316/2000 [07:54<04:02,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1317/2000 [07:55<04:00,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1318/2000 [07:55<04:01,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1319/2000 [07:55<04:05,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1320/2000 [07:56<04:05,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1320/2000 [07:56<04:05,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1321/2000 [07:56<04:02,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1322/2000 [07:56<04:00,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1323/2000 [07:57<03:58,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▌   | 1324/2000 [07:57<03:57,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▋   | 1325/2000 [07:57<03:56,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▋   | 1326/2000 [07:58<03:54,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▋   | 1327/2000 [07:58<03:56,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▋   | 1328/2000 [07:58<03:55,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▋   | 1329/2000 [07:59<03:55,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▋   | 1330/2000 [07:59<03:56,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  66%|██████▋   | 1330/2000 [07:59<03:56,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1331/2000 [07:59<03:55,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1332/2000 [08:00<03:56,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1333/2000 [08:00<03:58,  2.79it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1334/2000 [08:01<03:58,  2.79it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1335/2000 [08:01<04:00,  2.77it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1336/2000 [08:01<04:00,  2.76it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1337/2000 [08:02<03:59,  2.77it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1338/2000 [08:02<03:59,  2.77it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1339/2000 [08:02<03:58,  2.77it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1340/2000 [08:03<03:58,  2.77it/s, loss=0.278, lr=1e-6]
Steps:  67%|██████▋   | 1340/2000 [08:03<03:58,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1341/2000 [08:03<03:56,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1342/2000 [08:03<03:54,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1343/2000 [08:04<03:53,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1344/2000 [08:04<03:53,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1345/2000 [08:04<03:51,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1346/2000 [08:05<03:50,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1347/2000 [08:05<03:49,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1348/2000 [08:06<03:48,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  67%|██████▋   | 1349/2000 [08:06<03:48,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1350/2000 [08:06<03:48,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1350/2000 [08:07<03:48,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1351/2000 [08:07<03:48,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1352/2000 [08:07<03:47,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1353/2000 [08:07<03:46,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1354/2000 [08:08<03:46,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1355/2000 [08:08<03:46,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1356/2000 [08:08<03:45,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1357/2000 [08:09<03:43,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1358/2000 [08:09<03:43,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1359/2000 [08:09<03:43,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1360/2000 [08:10<03:43,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1360/2000 [08:10<03:43,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1361/2000 [08:10<03:43,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1362/2000 [08:10<03:42,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1363/2000 [08:11<03:41,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1364/2000 [08:11<03:44,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1365/2000 [08:11<03:43,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1366/2000 [08:12<03:43,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1367/2000 [08:12<03:42,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1368/2000 [08:13<03:42,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1369/2000 [08:13<03:43,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1370/2000 [08:13<03:42,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  68%|██████▊   | 1370/2000 [08:14<03:42,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▊   | 1371/2000 [08:14<03:40,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▊   | 1372/2000 [08:14<03:41,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▊   | 1373/2000 [08:14<03:40,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▊   | 1374/2000 [08:15<03:42,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1375/2000 [08:15<03:43,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1376/2000 [08:15<03:40,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1377/2000 [08:16<03:40,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1378/2000 [08:16<03:42,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1379/2000 [08:16<03:40,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1380/2000 [08:17<03:44,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1380/2000 [08:17<03:44,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1381/2000 [08:17<03:43,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1382/2000 [08:18<03:41,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1383/2000 [08:18<03:41,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1384/2000 [08:18<03:40,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1385/2000 [08:19<03:39,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1386/2000 [08:19<03:36,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1387/2000 [08:19<03:39,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1388/2000 [08:20<03:37,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  69%|██████▉   | 1389/2000 [08:20<03:36,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  70%|██████▉   | 1390/2000 [08:20<03:35,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  70%|██████▉   | 1390/2000 [08:21<03:35,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1391/2000 [08:21<03:37,  2.80it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1392/2000 [08:21<03:40,  2.76it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1393/2000 [08:21<03:37,  2.79it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1394/2000 [08:22<03:34,  2.82it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1395/2000 [08:22<03:33,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1396/2000 [08:22<03:31,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1397/2000 [08:23<03:30,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1398/2000 [08:23<03:32,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  70%|██████▉   | 1399/2000 [08:24<03:31,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  70%|███████   | 1400/2000 [08:24<03:32,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  70%|███████   | 1400/2000 [08:24<03:32,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1401/2000 [08:24<03:33,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1402/2000 [08:25<03:32,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1403/2000 [08:25<03:31,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1404/2000 [08:25<03:31,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1405/2000 [08:26<03:31,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1406/2000 [08:26<03:30,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1407/2000 [08:26<03:29,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1408/2000 [08:27<03:31,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1409/2000 [08:27<03:29,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1410/2000 [08:27<03:27,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  70%|███████   | 1410/2000 [08:28<03:27,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1411/2000 [08:28<03:27,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1412/2000 [08:28<03:38,  2.69it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1413/2000 [08:29<03:35,  2.72it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1414/2000 [08:29<03:31,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1415/2000 [08:29<03:28,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1416/2000 [08:30<03:27,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1417/2000 [08:30<03:27,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1418/2000 [08:30<03:27,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1419/2000 [08:31<03:26,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1420/2000 [08:31<03:28,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1420/2000 [08:31<03:28,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1421/2000 [08:31<03:25,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1422/2000 [08:32<03:23,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1423/2000 [08:32<03:25,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████   | 1424/2000 [08:32<03:23,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████▏  | 1425/2000 [08:33<03:23,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████▏  | 1426/2000 [08:33<03:22,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████▏  | 1427/2000 [08:34<03:21,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████▏  | 1428/2000 [08:34<03:20,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  71%|███████▏  | 1429/2000 [08:34<03:20,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1430/2000 [08:35<03:20,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1430/2000 [08:35<03:20,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1431/2000 [08:35<03:20,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1432/2000 [08:35<03:19,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1433/2000 [08:36<03:20,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1434/2000 [08:36<03:22,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1435/2000 [08:36<03:21,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1436/2000 [08:37<03:20,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1437/2000 [08:37<03:20,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1438/2000 [08:37<03:19,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1439/2000 [08:38<03:18,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1440/2000 [08:38<03:17,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1440/2000 [08:38<03:17,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1441/2000 [08:38<03:16,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1442/2000 [08:39<03:16,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1443/2000 [08:39<03:16,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1444/2000 [08:40<03:15,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1445/2000 [08:40<03:16,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1446/2000 [08:40<03:15,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1447/2000 [08:41<03:14,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1448/2000 [08:41<03:18,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▏  | 1449/2000 [08:41<03:17,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▎  | 1450/2000 [08:42<03:16,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  72%|███████▎  | 1450/2000 [08:42<03:16,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1451/2000 [08:42<03:16,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1452/2000 [08:42<03:14,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1453/2000 [08:43<03:13,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1454/2000 [08:43<03:12,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1455/2000 [08:43<03:10,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1456/2000 [08:44<03:10,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1457/2000 [08:44<03:11,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1458/2000 [08:44<03:10,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1459/2000 [08:45<03:09,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1460/2000 [08:45<03:08,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1460/2000 [08:46<03:08,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1461/2000 [08:46<03:07,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1462/2000 [08:46<03:08,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1463/2000 [08:46<03:07,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1464/2000 [08:47<03:06,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1465/2000 [08:47<03:06,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1466/2000 [08:47<03:06,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1467/2000 [08:48<03:05,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1468/2000 [08:48<03:05,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  73%|███████▎  | 1469/2000 [08:48<03:04,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▎  | 1470/2000 [08:49<03:06,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▎  | 1470/2000 [08:49<03:06,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▎  | 1471/2000 [08:49<03:06,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▎  | 1472/2000 [08:49<03:06,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▎  | 1473/2000 [08:50<03:05,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▎  | 1474/2000 [08:50<03:09,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1475/2000 [08:50<03:07,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1476/2000 [08:51<03:06,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1477/2000 [08:51<03:07,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1478/2000 [08:52<03:05,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1479/2000 [08:52<03:03,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1480/2000 [08:52<03:01,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1480/2000 [08:53<03:01,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1481/2000 [08:53<03:05,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1482/2000 [08:53<03:07,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1483/2000 [08:53<03:04,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1484/2000 [08:54<03:04,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1485/2000 [08:54<03:06,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1486/2000 [08:54<03:03,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1487/2000 [08:55<03:02,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1488/2000 [08:55<02:59,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1489/2000 [08:55<02:57,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1490/2000 [08:56<02:57,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  74%|███████▍  | 1490/2000 [08:56<02:57,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1491/2000 [08:56<02:57,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1492/2000 [08:56<02:57,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1493/2000 [08:57<02:56,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1494/2000 [08:57<02:56,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1495/2000 [08:58<02:55,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1496/2000 [08:58<02:56,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1497/2000 [08:58<02:56,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1498/2000 [08:59<02:55,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▍  | 1499/2000 [08:59<02:54,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1500/2000 [08:59<02:54,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1500/2000 [09:00<02:54,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1501/2000 [09:00<02:53,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1502/2000 [09:00<02:52,  2.89it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1503/2000 [09:00<02:53,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1504/2000 [09:01<02:52,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1505/2000 [09:01<02:57,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1506/2000 [09:01<02:56,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1507/2000 [09:02<02:57,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1508/2000 [09:02<02:55,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  75%|███████▌  | 1509/2000 [09:02<02:53,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1510/2000 [09:03<02:54,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1510/2000 [09:03<02:54,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1511/2000 [09:03<02:52,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1512/2000 [09:04<02:51,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1513/2000 [09:04<02:51,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1514/2000 [09:04<02:51,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1515/2000 [09:05<02:51,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1516/2000 [09:05<02:52,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1517/2000 [09:05<02:51,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1518/2000 [09:06<02:49,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1519/2000 [09:06<02:49,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1520/2000 [09:06<02:48,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  76%|███████▌  | 1520/2000 [09:07<02:48,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▌  | 1521/2000 [09:07<02:50,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▌  | 1522/2000 [09:07<02:48,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▌  | 1523/2000 [09:07<02:47,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▌  | 1524/2000 [09:08<02:46,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▋  | 1525/2000 [09:08<02:45,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▋  | 1526/2000 [09:08<02:45,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▋  | 1527/2000 [09:09<02:44,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▋  | 1528/2000 [09:09<02:47,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▋  | 1529/2000 [09:09<02:46,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▋  | 1530/2000 [09:10<02:50,  2.76it/s, loss=0.276, lr=1e-6]
Steps:  76%|███████▋  | 1530/2000 [09:10<02:50,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1531/2000 [09:10<02:48,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1532/2000 [09:11<02:46,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1533/2000 [09:11<02:44,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1534/2000 [09:11<02:43,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1535/2000 [09:12<02:43,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1536/2000 [09:12<02:42,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1537/2000 [09:12<02:41,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1538/2000 [09:13<02:42,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1539/2000 [09:13<02:41,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1540/2000 [09:13<02:41,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  77%|███████▋  | 1540/2000 [09:14<02:41,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1541/2000 [09:14<02:41,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1542/2000 [09:14<02:40,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1543/2000 [09:14<02:39,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1544/2000 [09:15<02:40,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1545/2000 [09:15<02:39,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1546/2000 [09:15<02:38,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1547/2000 [09:16<02:37,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1548/2000 [09:16<02:37,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  77%|███████▋  | 1549/2000 [09:17<02:36,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1550/2000 [09:17<02:38,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1550/2000 [09:17<02:38,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1551/2000 [09:17<02:37,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1552/2000 [09:18<02:36,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1553/2000 [09:18<02:37,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1554/2000 [09:18<02:36,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1555/2000 [09:19<02:38,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1556/2000 [09:19<02:37,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1557/2000 [09:19<02:37,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1558/2000 [09:20<02:35,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1559/2000 [09:20<02:36,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1560/2000 [09:20<02:36,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1560/2000 [09:21<02:36,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1561/2000 [09:21<02:36,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1562/2000 [09:21<02:37,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1563/2000 [09:21<02:35,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1564/2000 [09:22<02:40,  2.71it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1565/2000 [09:22<02:39,  2.72it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1566/2000 [09:23<02:38,  2.74it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1567/2000 [09:23<02:36,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1568/2000 [09:23<02:34,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1569/2000 [09:24<02:33,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1570/2000 [09:24<02:32,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  78%|███████▊  | 1570/2000 [09:24<02:32,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▊  | 1571/2000 [09:24<02:32,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▊  | 1572/2000 [09:25<02:31,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▊  | 1573/2000 [09:25<02:33,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▊  | 1574/2000 [09:25<02:31,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1575/2000 [09:26<02:30,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1576/2000 [09:26<02:30,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1577/2000 [09:27<02:28,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1578/2000 [09:27<02:29,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1579/2000 [09:27<02:28,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1580/2000 [09:28<02:26,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1580/2000 [09:28<02:26,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1581/2000 [09:28<02:26,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1582/2000 [09:28<02:26,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1583/2000 [09:29<02:26,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1584/2000 [09:29<02:28,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1585/2000 [09:29<02:26,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1586/2000 [09:30<02:25,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1587/2000 [09:30<02:25,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1588/2000 [09:30<02:24,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  79%|███████▉  | 1589/2000 [09:31<02:23,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1590/2000 [09:31<02:24,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1590/2000 [09:31<02:24,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1591/2000 [09:31<02:23,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1592/2000 [09:32<02:22,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1593/2000 [09:32<02:23,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1594/2000 [09:32<02:21,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1595/2000 [09:33<02:20,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1596/2000 [09:33<02:21,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1597/2000 [09:34<02:20,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1598/2000 [09:34<02:21,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  80%|███████▉  | 1599/2000 [09:34<02:22,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1600/2000 [09:35<02:21,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1600/2000 [09:35<02:21,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1601/2000 [09:35<02:21,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1602/2000 [09:35<02:20,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1603/2000 [09:36<02:20,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1604/2000 [09:36<02:21,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1605/2000 [09:36<02:20,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1606/2000 [09:37<02:19,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1607/2000 [09:37<02:25,  2.71it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1608/2000 [09:37<02:24,  2.70it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1609/2000 [09:38<02:22,  2.74it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1610/2000 [09:38<02:20,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  80%|████████  | 1610/2000 [09:39<02:20,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1611/2000 [09:39<02:19,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1612/2000 [09:39<02:18,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1613/2000 [09:39<02:17,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1614/2000 [09:40<02:17,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1615/2000 [09:40<02:17,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1616/2000 [09:40<02:16,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1617/2000 [09:41<02:15,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1618/2000 [09:41<02:17,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1619/2000 [09:41<02:15,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1620/2000 [09:42<02:15,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  81%|████████  | 1620/2000 [09:42<02:15,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████  | 1621/2000 [09:42<02:15,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████  | 1622/2000 [09:42<02:14,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████  | 1623/2000 [09:43<02:14,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████  | 1624/2000 [09:43<02:13,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████▏ | 1625/2000 [09:44<02:12,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████▏ | 1626/2000 [09:44<02:13,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████▏ | 1627/2000 [09:44<02:13,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████▏ | 1628/2000 [09:45<02:14,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  81%|████████▏ | 1629/2000 [09:45<02:12,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1630/2000 [09:45<02:11,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1630/2000 [09:46<02:11,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1631/2000 [09:46<02:11,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1632/2000 [09:46<02:11,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1633/2000 [09:46<02:11,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1634/2000 [09:47<02:10,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1635/2000 [09:47<02:09,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1636/2000 [09:47<02:08,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1637/2000 [09:48<02:07,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1638/2000 [09:48<02:08,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1639/2000 [09:49<02:06,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1640/2000 [09:49<02:05,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1640/2000 [09:49<02:05,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1641/2000 [09:49<02:05,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1642/2000 [09:50<02:05,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1643/2000 [09:50<02:05,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1644/2000 [09:50<02:04,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1645/2000 [09:51<02:05,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1646/2000 [09:51<02:04,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1647/2000 [09:51<02:04,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1648/2000 [09:52<02:03,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▏ | 1649/2000 [09:52<02:04,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▎ | 1650/2000 [09:52<02:04,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  82%|████████▎ | 1650/2000 [09:53<02:04,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1651/2000 [09:53<02:04,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1652/2000 [09:53<02:03,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1653/2000 [09:53<02:02,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1654/2000 [09:54<02:04,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1655/2000 [09:54<02:03,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1656/2000 [09:55<02:03,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1657/2000 [09:55<02:06,  2.71it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1658/2000 [09:55<02:03,  2.76it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1659/2000 [09:56<02:01,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1660/2000 [09:56<02:00,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  83%|████████▎ | 1660/2000 [09:56<02:00,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1661/2000 [09:56<02:00,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1662/2000 [09:57<01:59,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1663/2000 [09:57<01:59,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1664/2000 [09:57<01:58,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1665/2000 [09:58<01:57,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1666/2000 [09:58<01:58,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1667/2000 [09:58<01:57,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1668/2000 [09:59<01:57,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  83%|████████▎ | 1669/2000 [09:59<01:55,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▎ | 1670/2000 [09:59<01:54,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▎ | 1670/2000 [10:00<01:54,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▎ | 1671/2000 [10:00<01:54,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▎ | 1672/2000 [10:00<01:54,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▎ | 1673/2000 [10:01<01:53,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▎ | 1674/2000 [10:01<01:54,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▍ | 1675/2000 [10:01<01:54,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▍ | 1676/2000 [10:02<01:53,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▍ | 1677/2000 [10:02<01:53,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▍ | 1678/2000 [10:02<01:53,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▍ | 1679/2000 [10:03<01:52,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▍ | 1680/2000 [10:03<01:54,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  84%|████████▍ | 1680/2000 [10:03<01:54,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1681/2000 [10:03<01:52,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1682/2000 [10:04<01:51,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1683/2000 [10:04<01:53,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1684/2000 [10:04<01:52,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1685/2000 [10:05<01:51,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1686/2000 [10:05<01:50,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1687/2000 [10:05<01:49,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1688/2000 [10:06<01:52,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1689/2000 [10:06<01:51,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1690/2000 [10:07<01:50,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  84%|████████▍ | 1690/2000 [10:07<01:50,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1691/2000 [10:07<01:48,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1692/2000 [10:07<01:47,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1693/2000 [10:08<01:47,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1694/2000 [10:08<01:47,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1695/2000 [10:08<01:47,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1696/2000 [10:09<01:47,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1697/2000 [10:09<01:46,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1698/2000 [10:09<01:45,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▍ | 1699/2000 [10:10<01:45,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1700/2000 [10:10<01:45,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1700/2000 [10:10<01:45,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1701/2000 [10:10<01:44,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1702/2000 [10:11<01:44,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1703/2000 [10:11<01:46,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1704/2000 [10:11<01:44,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1705/2000 [10:12<01:45,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1706/2000 [10:12<01:44,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1707/2000 [10:13<01:43,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1708/2000 [10:13<01:41,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  85%|████████▌ | 1709/2000 [10:13<01:42,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1710/2000 [10:14<01:41,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1710/2000 [10:14<01:41,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1711/2000 [10:14<01:42,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1712/2000 [10:14<01:41,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1713/2000 [10:15<01:40,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1714/2000 [10:15<01:40,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1715/2000 [10:15<01:39,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1716/2000 [10:16<01:38,  2.89it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1717/2000 [10:16<01:40,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1718/2000 [10:16<01:40,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1719/2000 [10:17<01:39,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1720/2000 [10:17<01:38,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  86%|████████▌ | 1720/2000 [10:17<01:38,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▌ | 1721/2000 [10:17<01:37,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▌ | 1722/2000 [10:18<01:36,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▌ | 1723/2000 [10:18<01:36,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▌ | 1724/2000 [10:18<01:35,  2.89it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▋ | 1725/2000 [10:19<01:36,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▋ | 1726/2000 [10:19<01:36,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▋ | 1727/2000 [10:20<01:35,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▋ | 1728/2000 [10:20<01:36,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▋ | 1729/2000 [10:20<01:35,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▋ | 1730/2000 [10:21<01:34,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  86%|████████▋ | 1730/2000 [10:21<01:34,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1731/2000 [10:21<01:35,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1732/2000 [10:21<01:34,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1733/2000 [10:22<01:33,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1734/2000 [10:22<01:34,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1735/2000 [10:22<01:33,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1736/2000 [10:23<01:33,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1737/2000 [10:23<01:32,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1738/2000 [10:23<01:31,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1739/2000 [10:24<01:32,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1740/2000 [10:24<01:32,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1740/2000 [10:24<01:32,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1741/2000 [10:24<01:31,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1742/2000 [10:25<01:31,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1743/2000 [10:25<01:30,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1744/2000 [10:26<01:29,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1745/2000 [10:26<01:29,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1746/2000 [10:26<01:29,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1747/2000 [10:27<01:28,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1748/2000 [10:27<01:27,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  87%|████████▋ | 1749/2000 [10:27<01:27,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1750/2000 [10:28<01:27,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1750/2000 [10:28<01:27,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1751/2000 [10:28<01:29,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1752/2000 [10:28<01:28,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1753/2000 [10:29<01:27,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1754/2000 [10:29<01:27,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1755/2000 [10:29<01:26,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1756/2000 [10:30<01:25,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1757/2000 [10:30<01:25,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1758/2000 [10:30<01:24,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1759/2000 [10:31<01:24,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1760/2000 [10:31<01:25,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1760/2000 [10:32<01:25,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1761/2000 [10:32<01:24,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1762/2000 [10:32<01:23,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1763/2000 [10:32<01:23,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1764/2000 [10:33<01:23,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1765/2000 [10:33<01:22,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1766/2000 [10:33<01:22,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1767/2000 [10:34<01:21,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1768/2000 [10:34<01:22,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1769/2000 [10:34<01:21,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1770/2000 [10:35<01:20,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  88%|████████▊ | 1770/2000 [10:35<01:20,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▊ | 1771/2000 [10:35<01:20,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▊ | 1772/2000 [10:35<01:19,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▊ | 1773/2000 [10:36<01:18,  2.89it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▊ | 1774/2000 [10:36<01:17,  2.90it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1775/2000 [10:36<01:18,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1776/2000 [10:37<01:17,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1777/2000 [10:37<01:17,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1778/2000 [10:37<01:17,  2.88it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1779/2000 [10:38<01:16,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1780/2000 [10:38<01:16,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1780/2000 [10:39<01:16,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1781/2000 [10:39<01:16,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1782/2000 [10:39<01:16,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1783/2000 [10:39<01:15,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1784/2000 [10:40<01:15,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1785/2000 [10:40<01:15,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1786/2000 [10:40<01:14,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1787/2000 [10:41<01:15,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1788/2000 [10:41<01:16,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  89%|████████▉ | 1789/2000 [10:41<01:15,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1790/2000 [10:42<01:14,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1790/2000 [10:42<01:14,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1791/2000 [10:42<01:13,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1792/2000 [10:42<01:12,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1793/2000 [10:43<01:12,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1794/2000 [10:43<01:12,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1795/2000 [10:43<01:12,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1796/2000 [10:44<01:12,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1797/2000 [10:44<01:11,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1798/2000 [10:45<01:10,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  90%|████████▉ | 1799/2000 [10:45<01:11,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1800/2000 [10:45<01:11,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1800/2000 [10:46<01:11,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1801/2000 [10:46<01:10,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1802/2000 [10:46<01:09,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1803/2000 [10:46<01:09,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1804/2000 [10:47<01:09,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1805/2000 [10:47<01:09,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1806/2000 [10:47<01:08,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1807/2000 [10:48<01:07,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1808/2000 [10:48<01:07,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1809/2000 [10:48<01:07,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1810/2000 [10:49<01:07,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  90%|█████████ | 1810/2000 [10:49<01:07,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1811/2000 [10:49<01:07,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1812/2000 [10:49<01:06,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1813/2000 [10:50<01:05,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1814/2000 [10:50<01:05,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1815/2000 [10:51<01:05,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1816/2000 [10:51<01:04,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1817/2000 [10:51<01:04,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1818/2000 [10:52<01:05,  2.76it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1819/2000 [10:52<01:05,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1820/2000 [10:52<01:04,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1820/2000 [10:53<01:04,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1821/2000 [10:53<01:03,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1822/2000 [10:53<01:03,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1823/2000 [10:53<01:02,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████ | 1824/2000 [10:54<01:02,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████▏| 1825/2000 [10:54<01:01,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████▏| 1826/2000 [10:54<01:01,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████▏| 1827/2000 [10:55<01:01,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████▏| 1828/2000 [10:55<01:00,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  91%|█████████▏| 1829/2000 [10:56<00:59,  2.87it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1830/2000 [10:56<01:00,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1830/2000 [10:56<01:00,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1831/2000 [10:56<01:00,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1832/2000 [10:57<00:59,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1833/2000 [10:57<00:59,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1834/2000 [10:57<00:58,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1835/2000 [10:58<00:57,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1836/2000 [10:58<00:57,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1837/2000 [10:58<00:57,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1838/2000 [10:59<00:56,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1839/2000 [10:59<00:57,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1840/2000 [10:59<00:56,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1840/2000 [11:00<00:56,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1841/2000 [11:00<00:55,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1842/2000 [11:00<00:55,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1843/2000 [11:00<00:55,  2.85it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1844/2000 [11:01<00:54,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1845/2000 [11:01<00:54,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1846/2000 [11:02<00:55,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1847/2000 [11:02<00:54,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1848/2000 [11:02<00:54,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▏| 1849/2000 [11:03<00:53,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▎| 1850/2000 [11:03<00:53,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  92%|█████████▎| 1850/2000 [11:03<00:53,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1851/2000 [11:03<00:52,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1852/2000 [11:04<00:52,  2.82it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1853/2000 [11:04<00:53,  2.75it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1854/2000 [11:04<00:52,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1855/2000 [11:05<00:52,  2.75it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1856/2000 [11:05<00:52,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1857/2000 [11:05<00:50,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1858/2000 [11:06<00:51,  2.76it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1859/2000 [11:06<00:50,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1860/2000 [11:07<00:49,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1860/2000 [11:07<00:49,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1861/2000 [11:07<00:48,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1862/2000 [11:07<00:49,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1863/2000 [11:08<00:49,  2.75it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1864/2000 [11:08<00:49,  2.74it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1865/2000 [11:08<00:48,  2.76it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1866/2000 [11:09<00:48,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1867/2000 [11:09<00:47,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1868/2000 [11:09<00:46,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  93%|█████████▎| 1869/2000 [11:10<00:46,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▎| 1870/2000 [11:10<00:45,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▎| 1870/2000 [11:10<00:45,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▎| 1871/2000 [11:10<00:45,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▎| 1872/2000 [11:11<00:44,  2.86it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▎| 1873/2000 [11:11<00:44,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▎| 1874/2000 [11:12<00:44,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1875/2000 [11:12<00:44,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1876/2000 [11:12<00:43,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1877/2000 [11:13<00:43,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1878/2000 [11:13<00:43,  2.80it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1879/2000 [11:13<00:43,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1880/2000 [11:14<00:43,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1880/2000 [11:14<00:43,  2.78it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1881/2000 [11:14<00:43,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1882/2000 [11:14<00:43,  2.72it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1883/2000 [11:15<00:42,  2.74it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1884/2000 [11:15<00:41,  2.77it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1885/2000 [11:15<00:41,  2.79it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1886/2000 [11:16<00:40,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1887/2000 [11:16<00:40,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1888/2000 [11:17<00:39,  2.81it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1889/2000 [11:17<00:39,  2.83it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1890/2000 [11:17<00:38,  2.84it/s, loss=0.276, lr=1e-6]
Steps:  94%|█████████▍| 1890/2000 [11:18<00:38,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1891/2000 [11:18<00:38,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1892/2000 [11:18<00:37,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1893/2000 [11:18<00:37,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1894/2000 [11:19<00:37,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1895/2000 [11:19<00:37,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1896/2000 [11:19<00:37,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1897/2000 [11:20<00:37,  2.75it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1898/2000 [11:20<00:36,  2.77it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▍| 1899/2000 [11:20<00:36,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1900/2000 [11:21<00:35,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1900/2000 [11:21<00:35,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1901/2000 [11:21<00:35,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1902/2000 [11:22<00:34,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1903/2000 [11:22<00:34,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1904/2000 [11:22<00:33,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1905/2000 [11:23<00:33,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1906/2000 [11:23<00:33,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1907/2000 [11:23<00:33,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1908/2000 [11:24<00:32,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  95%|█████████▌| 1909/2000 [11:24<00:32,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1910/2000 [11:24<00:31,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1910/2000 [11:25<00:31,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1911/2000 [11:25<00:32,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1912/2000 [11:25<00:31,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1913/2000 [11:25<00:30,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1914/2000 [11:26<00:30,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1915/2000 [11:26<00:30,  2.83it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1916/2000 [11:26<00:29,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1917/2000 [11:27<00:29,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1918/2000 [11:27<00:29,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1919/2000 [11:28<00:28,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1920/2000 [11:28<00:28,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1920/2000 [11:28<00:28,  2.84it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1921/2000 [11:28<00:27,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1922/2000 [11:29<00:27,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1923/2000 [11:29<00:26,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▌| 1924/2000 [11:29<00:26,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▋| 1925/2000 [11:30<00:26,  2.88it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▋| 1926/2000 [11:30<00:26,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▋| 1927/2000 [11:30<00:25,  2.81it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▋| 1928/2000 [11:31<00:25,  2.79it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▋| 1929/2000 [11:31<00:25,  2.75it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▋| 1930/2000 [11:31<00:25,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  96%|█████████▋| 1930/2000 [11:32<00:25,  2.78it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1931/2000 [11:32<00:24,  2.80it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1932/2000 [11:32<00:24,  2.82it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1933/2000 [11:32<00:23,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1934/2000 [11:33<00:23,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1935/2000 [11:33<00:22,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1936/2000 [11:34<00:22,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1937/2000 [11:34<00:22,  2.86it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1938/2000 [11:34<00:21,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1939/2000 [11:35<00:21,  2.87it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1940/2000 [11:35<00:21,  2.85it/s, loss=0.277, lr=1e-6]
Steps:  97%|█████████▋| 1940/2000 [11:35<00:21,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1941/2000 [11:35<00:20,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1942/2000 [11:36<00:20,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1943/2000 [11:36<00:19,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1944/2000 [11:36<00:19,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1945/2000 [11:37<00:19,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1946/2000 [11:37<00:19,  2.81it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1947/2000 [11:37<00:18,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1948/2000 [11:38<00:18,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  97%|█████████▋| 1949/2000 [11:38<00:17,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1950/2000 [11:38<00:17,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1950/2000 [11:39<00:17,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1951/2000 [11:39<00:17,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1952/2000 [11:39<00:16,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1953/2000 [11:39<00:16,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1954/2000 [11:40<00:16,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1955/2000 [11:40<00:15,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1956/2000 [11:41<00:15,  2.89it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1957/2000 [11:41<00:15,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1958/2000 [11:41<00:14,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1959/2000 [11:42<00:14,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1960/2000 [11:42<00:13,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  98%|█████████▊| 1960/2000 [11:42<00:13,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1961/2000 [11:42<00:13,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1962/2000 [11:43<00:13,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1963/2000 [11:43<00:12,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1964/2000 [11:43<00:12,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1965/2000 [11:44<00:12,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1966/2000 [11:44<00:11,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1967/2000 [11:44<00:11,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1968/2000 [11:45<00:11,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1969/2000 [11:45<00:10,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1970/2000 [11:45<00:10,  2.88it/s, loss=0.279, lr=1e-6]
Steps:  98%|█████████▊| 1970/2000 [11:46<00:10,  2.88it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▊| 1971/2000 [11:46<00:10,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▊| 1972/2000 [11:46<00:09,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▊| 1973/2000 [11:46<00:09,  2.86it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▊| 1974/2000 [11:47<00:09,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▉| 1975/2000 [11:47<00:08,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▉| 1976/2000 [11:48<00:08,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▉| 1977/2000 [11:48<00:08,  2.83it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▉| 1978/2000 [11:48<00:07,  2.84it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▉| 1979/2000 [11:49<00:07,  2.85it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▉| 1980/2000 [11:49<00:06,  2.87it/s, loss=0.278, lr=1e-6]
Steps:  99%|█████████▉| 1980/2000 [11:49<00:06,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1981/2000 [11:49<00:06,  2.86it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1982/2000 [11:50<00:06,  2.87it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1983/2000 [11:50<00:05,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1984/2000 [11:50<00:05,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1985/2000 [11:51<00:05,  2.85it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1986/2000 [11:51<00:04,  2.80it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1987/2000 [11:51<00:04,  2.82it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1988/2000 [11:52<00:04,  2.84it/s, loss=0.279, lr=1e-6]
Steps:  99%|█████████▉| 1989/2000 [11:52<00:03,  2.86it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1990/2000 [11:52<00:03,  2.86it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1990/2000 [11:53<00:03,  2.86it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1991/2000 [11:53<00:03,  2.86it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1992/2000 [11:53<00:02,  2.86it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1993/2000 [11:54<00:02,  2.87it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1994/2000 [11:54<00:02,  2.88it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1995/2000 [11:54<00:01,  2.86it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1996/2000 [11:55<00:01,  2.87it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1997/2000 [11:55<00:01,  2.88it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1998/2000 [11:55<00:00,  2.87it/s, loss=0.279, lr=1e-6]
Steps: 100%|█████████▉| 1999/2000 [11:56<00:00,  2.88it/s, loss=0.279, lr=1e-6]
Steps: 100%|██████████| 2000/2000 [11:56<00:00,  2.89it/s, loss=0.279, lr=1e-6]You have passed `None` for safety_checker to disable its functionality in <class 'diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline'>. Note that this might lead to problems when using <class 'diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline'> and is not recommended.
[*] Weights saved at checkpoints
Steps: 100%|██████████| 2000/2000 [12:01<00:00,  2.77it/s, loss=0.279, lr=1e-6]
Sun Nov 20 21:13:41 2022
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 510.47.03    Driver Version: 510.47.03    CUDA Version: 11.6     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  NVIDIA A100-SXM...  Off  | 00000000:00:04.0 Off |                    0 |
| N/A   38C    P0    50W / 400W |   2122MiB / 40960MiB |     11%      Default |
|                               |                      |             Disabled |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|    0   N/A  N/A    343420      C                                    2119MiB |
+-----------------------------------------------------------------------------+
checkpoints/tokenizer
checkpoints/unet
checkpoints/vae
checkpoints/text_encoder
checkpoints/feature_extractor
checkpoints/args.json
checkpoints/model_index.json
checkpoints/scheduler
checkpoints/tokenizer/vocab.json
checkpoints/tokenizer/special_tokens_map.json
checkpoints/tokenizer/merges.txt
checkpoints/tokenizer/tokenizer_config.json
checkpoints/unet/diffusion_pytorch_model.bin
checkpoints/unet/config.json
checkpoints/vae/diffusion_pytorch_model.bin
checkpoints/vae/config.json
checkpoints/text_encoder/pytorch_model.bin
checkpoints/text_encoder/config.json
checkpoints/feature_extractor/preprocessor_config.json
checkpoints/scheduler/scheduler_config.json
Version Details
Version ID
a8ba568da0313951a6b311b43b1ea3bf9f2ef7b9fd97ed94cebd7ffd2da66654
Version Created
April 21, 2023
Run on Replicate →