alekseycalvin/anna_akhmatova_flux 🖼️🔢❓📝✓ → 🖼️

▶️ 24 runs 📅 Oct 2024 ⚙️ Cog 0.11.1
lora lora-famous-person lora-person text-to-image

About

Flux LoRa evoking the visage of poet Anna Akhmatova (1889-1966), whose voice emblematizes one self's capacity to universalize, harnessing, the pain of all, refracting it back dignified unto transcendence...

Example Output

Prompt:

"poet Anna AKHMATOVA in 1910, sitting at an outdoors lounge table and angrily throwing Tarot cards from a balcony, close up on her detailed face, moderately blemished lined skin texture, extremely photorealistic, cards fly down from balcony, scattering everywhere"

Output

Example output

Performance Metrics

13.97s Prediction Time
41.37s Total Time
All Input Parameters 
{
  "model": "dev",
  "prompt": "poet Anna AKHMATOVA in 1910, sitting at an outdoors lounge table and angrily throwing Tarot cards from a balcony, close up on her detailed face, moderately blemished lined skin texture, extremely photorealistic, cards fly down from balcony, scattering everywhere",
  "lora_scale": 1.09,
  "num_outputs": 1,
  "aspect_ratio": "1:1",
  "output_format": "png",
  "guidance_scale": 2.97,
  "output_quality": 100,
  "prompt_strength": 1,
  "extra_lora_scale": 1,
  "num_inference_steps": 29
}
Input Parameters
mask Type: string
Image mask for image inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.
seed Type: integer
Random seed. Set for reproducible generation
image Type: string
Input image for image to image or inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.
model Default: dev
Which model to run inference with. The dev model performs best with around 28 inference steps but the schnell model only needs 4 steps.
width Type: integerRange: 256 - 1440
Width of generated image. Only works if `aspect_ratio` is set to custom. Will be rounded to nearest multiple of 16. Incompatible with fast generation
height Type: integerRange: 256 - 1440
Height of generated image. Only works if `aspect_ratio` is set to custom. Will be rounded to nearest multiple of 16. Incompatible with fast generation
prompt (required) Type: string
Prompt for generated image. If you include the `trigger_word` used in the training process you are more likely to activate the trained object, style, or concept in the resulting image.
go_fast Type: booleanDefault: false
Run faster predictions with model optimized for speed (currently fp8 quantized); disable to run in original bf16
extra_lora Type: string
Load LoRA weights. Supports Replicate models in the format <owner>/<username> or <owner>/<username>/<version>, HuggingFace URLs in the format huggingface.co/<owner>/<model-name>, CivitAI URLs in the format civitai.com/models/<id>[/<model-name>], or arbitrary .safetensors URLs from the Internet. For example, 'fofr/flux-pixar-cars'
lora_scale Type: numberDefault: 1Range: -1 - 3
Determines how strongly the main LoRA should be applied. Sane results between 0 and 1 for base inference. For go_fast we apply a 1.5x multiplier to this value; we've generally seen good performance when scaling the base value by that amount. You may still need to experiment to find the best value for your particular lora.
megapixels Default: 1
Approximate number of megapixels for generated image
num_outputs Type: integerDefault: 1Range: 1 - 4
Number of outputs to generate
aspect_ratio Default: 1:1
Aspect ratio for the generated image. If custom is selected, uses height and width below & will run in bf16 mode
output_format Default: webp
Format of the output images
guidance_scale Type: numberDefault: 3Range: 0 - 10
Guidance scale for the diffusion process. Lower values can give more realistic images. Good values to try are 2, 2.5, 3 and 3.5
output_quality Type: integerDefault: 80Range: 0 - 100
Quality when saving the output images, from 0 to 100. 100 is best quality, 0 is lowest quality. Not relevant for .png outputs
prompt_strength Type: numberDefault: 0.8Range: 0 - 1
Prompt strength when using img2img. 1.0 corresponds to full destruction of information in image
extra_lora_scale Type: numberDefault: 1Range: -1 - 3
Determines how strongly the extra LoRA should be applied. Sane results between 0 and 1 for base inference. For go_fast we apply a 1.5x multiplier to this value; we've generally seen good performance when scaling the base value by that amount. You may still need to experiment to find the best value for your particular lora.
replicate_weights Type: string
Load LoRA weights. Supports Replicate models in the format <owner>/<username> or <owner>/<username>/<version>, HuggingFace URLs in the format huggingface.co/<owner>/<model-name>, CivitAI URLs in the format civitai.com/models/<id>[/<model-name>], or arbitrary .safetensors URLs from the Internet. For example, 'fofr/flux-pixar-cars'
num_inference_steps Type: integerDefault: 28Range: 1 - 50
Number of denoising steps. More steps can give more detailed images, but take longer.
disable_safety_checker Type: booleanDefault: false
Disable safety checker for generated images.
Output Schema

Output

Type: array • Items Type: string • Items Format: uri

Example Execution Logs 
Using seed: 45793
Prompt: poet Anna AKHMATOVA in 1910, sitting at an outdoors lounge table and angrily throwing Tarot cards from a balcony, close up on her detailed face, moderately blemished lined skin texture, extremely photorealistic, cards fly down from balcony, scattering everywhere
[!] txt2img mode
Using dev model
free=29579398946816
Downloading weights
2024-11-16T19:32:18Z | INFO  | [ Initiating ] chunk_size=150M dest=/tmp/tmpsiz6z5jx/weights url=https://replicate.delivery/xezq/TMzxS7q27ybVAZ0qDb5L6tLX8CAuekZDt0Fyo8VWfc08dQxTA/trained_model.tar
2024-11-16T19:32:22Z | INFO  | [ Complete ] dest=/tmp/tmpsiz6z5jx/weights size="258 MB" total_elapsed=3.396s url=https://replicate.delivery/xezq/TMzxS7q27ybVAZ0qDb5L6tLX8CAuekZDt0Fyo8VWfc08dQxTA/trained_model.tar
Downloaded weights in 3.42s
Loaded LoRAs in 4.02s
  0%|          | 0/29 [00:00<?, ?it/s]
  3%|▎         | 1/29 [00:00<00:09,  3.00it/s]
  7%|▋         | 2/29 [00:00<00:08,  3.31it/s]
 10%|█         | 3/29 [00:00<00:08,  3.19it/s]
 14%|█▍        | 4/29 [00:01<00:07,  3.14it/s]
 17%|█▋        | 5/29 [00:01<00:07,  3.11it/s]
 21%|██        | 6/29 [00:01<00:07,  3.09it/s]
 24%|██▍       | 7/29 [00:02<00:07,  3.08it/s]
 28%|██▊       | 8/29 [00:02<00:06,  3.07it/s]
 31%|███       | 9/29 [00:02<00:06,  3.07it/s]
 34%|███▍      | 10/29 [00:03<00:06,  3.06it/s]
 38%|███▊      | 11/29 [00:03<00:05,  3.06it/s]
 41%|████▏     | 12/29 [00:03<00:05,  3.06it/s]
 45%|████▍     | 13/29 [00:04<00:05,  3.06it/s]
 48%|████▊     | 14/29 [00:04<00:04,  3.06it/s]
 52%|█████▏    | 15/29 [00:04<00:04,  3.06it/s]
 55%|█████▌    | 16/29 [00:05<00:04,  3.06it/s]
 59%|█████▊    | 17/29 [00:05<00:03,  3.06it/s]
 62%|██████▏   | 18/29 [00:05<00:03,  3.05it/s]
 66%|██████▌   | 19/29 [00:06<00:03,  3.06it/s]
 69%|██████▉   | 20/29 [00:06<00:02,  3.06it/s]
 72%|███████▏  | 21/29 [00:06<00:02,  3.06it/s]
 76%|███████▌  | 22/29 [00:07<00:02,  3.06it/s]
 79%|███████▉  | 23/29 [00:07<00:01,  3.06it/s]
 83%|████████▎ | 24/29 [00:07<00:01,  3.06it/s]
 86%|████████▌ | 25/29 [00:08<00:01,  3.06it/s]
 90%|████████▉ | 26/29 [00:08<00:00,  3.06it/s]
 93%|█████████▎| 27/29 [00:08<00:00,  3.06it/s]
 97%|█████████▋| 28/29 [00:09<00:00,  3.06it/s]
100%|██████████| 29/29 [00:09<00:00,  3.06it/s]
100%|██████████| 29/29 [00:09<00:00,  3.07it/s]
Version Details
Version ID
4c391eb0dac0ff0446292edf718af5caf9064d5cb6dfcc934d213ee665af7e36
Version Created
November 15, 2024
Run on Replicate →