vincetheeleventh/flux-gramps01 🖼️🔢❓📝✓ → 🖼️

▶️ 154 runs 📅 Feb 2025 ⚙️ Cog 0.13.7
lora lora-person text-to-image

About

A fine-tuned FLUX.1 model

Example Output

Prompt:

"

An ink and watercolor illustration in the style of E.H. Shepard, featuring a warm and expressive scene. The linework is delicate and sketchy, with a flowing, organic quality, full of charm and gentle humor. The background is lightly detailed with soft, airy strokes, evoking a sense of nostalgia. The composition is balanced but informal, capturing a cozy, storybook atmosphere. The overall color palette is muted, with natural tones and delicate shading, resembling classic early 20th-century children's book illustrations.

illustration of an elderly chinese man ukj with stocky build wearing a leather jacket, playing tea party with cheerful granddaughter age 5 and stuffed animals

"

Output

Example output

Performance Metrics

4.20s Prediction Time
4.24s Total Time
All Input Parameters
{
  "model": "dev",
  "prompt": "An ink and watercolor illustration in the style of E.H. Shepard, featuring a warm and expressive scene. The linework is delicate and sketchy, with a flowing, organic quality, full of charm and gentle humor. The background is lightly detailed with soft, airy strokes, evoking a sense of nostalgia. The composition is balanced but informal, capturing a cozy, storybook atmosphere. The overall color palette is muted, with natural tones and delicate shading, resembling classic early 20th-century children's book illustrations.\n\nillustration of an elderly chinese man ukj with stocky build wearing a leather jacket, playing tea party with cheerful granddaughter age 5 and stuffed animals ",
  "go_fast": true,
  "lora_scale": 0.5,
  "megapixels": "1",
  "num_outputs": 1,
  "aspect_ratio": "1:1",
  "output_format": "webp",
  "guidance_scale": 5,
  "output_quality": 80,
  "prompt_strength": 1,
  "extra_lora_scale": 1,
  "num_inference_steps": 40
}
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: arrayItems Type: stringItems Format: uri

Example Execution Logs
2025-02-06 06:33:18.420 | INFO     | fp8.lora_loading:restore_clones:592 - Unloaded 304 layers
2025-02-06 06:33:18.422 | SUCCESS  | fp8.lora_loading:unload_loras:563 - LoRAs unloaded in 0.023s
free=28896009428992
Downloading weights
2025-02-06T06:33:18Z | INFO  | [ Initiating ] chunk_size=150M dest=/tmp/tmp5d61i51m/weights url=https://replicate.delivery/xezq/12RA57I3M5rHIRlgvWDQxaDmngFbPO11krJYSftviEJW7NGKA/trained_model.tar
2025-02-06T06:33:18Z | INFO  | [ Complete ] dest=/tmp/tmp5d61i51m/weights size="4.7 MB" total_elapsed=0.095s url=https://replicate.delivery/xezq/12RA57I3M5rHIRlgvWDQxaDmngFbPO11krJYSftviEJW7NGKA/trained_model.tar
Downloaded weights in 0.11s
2025-02-06 06:33:18.536 | INFO     | fp8.lora_loading:convert_lora_weights:502 - Loading LoRA weights for /src/weights-cache/138282184bb37d7a
2025-02-06 06:33:18.538 | INFO     | fp8.lora_loading:convert_lora_weights:523 - LoRA weights loaded
2025-02-06 06:33:18.538 | DEBUG    | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:602 - Extracting keys
2025-02-06 06:33:18.538 | DEBUG    | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:609 - Keys extracted
Applying LoRA:   0%|          | 0/4 [00:00<?, ?it/s]
Applying LoRA: 100%|██████████| 4/4 [00:00<00:00, 2553.61it/s]
2025-02-06 06:33:18.540 | INFO     | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:661 - Loading LoRA in fp8
2025-02-06 06:33:18.540 | SUCCESS  | fp8.lora_loading:load_lora:542 - LoRA applied in 0.004s
running quantized prediction
Using seed: 2734417738
  0%|          | 0/40 [00:00<?, ?it/s]
  5%|▌         | 2/40 [00:00<00:02, 17.07it/s]
 10%|█         | 4/40 [00:00<00:02, 12.59it/s]
 15%|█▌        | 6/40 [00:00<00:02, 11.58it/s]
 20%|██        | 8/40 [00:00<00:02, 11.16it/s]
 25%|██▌       | 10/40 [00:00<00:02, 10.86it/s]
 30%|███       | 12/40 [00:01<00:02, 10.46it/s]
 35%|███▌      | 14/40 [00:01<00:02, 10.49it/s]
 40%|████      | 16/40 [00:01<00:02, 10.51it/s]
 45%|████▌     | 18/40 [00:01<00:02, 10.53it/s]
 50%|█████     | 20/40 [00:01<00:01, 10.49it/s]
 55%|█████▌    | 22/40 [00:02<00:01, 10.38it/s]
 60%|██████    | 24/40 [00:02<00:01, 10.35it/s]
 65%|██████▌   | 26/40 [00:02<00:01, 10.41it/s]
 70%|███████   | 28/40 [00:02<00:01, 10.42it/s]
 75%|███████▌  | 30/40 [00:02<00:00, 10.43it/s]
 80%|████████  | 32/40 [00:03<00:00, 10.36it/s]
 85%|████████▌ | 34/40 [00:03<00:00, 10.38it/s]
 90%|█████████ | 36/40 [00:03<00:00, 10.40it/s]
 95%|█████████▌| 38/40 [00:03<00:00, 10.45it/s]
100%|██████████| 40/40 [00:03<00:00, 10.47it/s]
100%|██████████| 40/40 [00:03<00:00, 10.62it/s]
Total safe images: 1 out of 1
Version Details
Version ID
3c5446ab672252a9ff78cbf5ebbec5f6772b926833513f00349a719be559b414
Version Created
March 1, 2025
Run on Replicate →