Add generic prompt experiment structure
This commit is contained in:
@@ -0,0 +1,58 @@
|
||||
# Person to clothing flat lay
|
||||
|
||||
## Purpose
|
||||
|
||||
Use a reference image containing a person and reconstruct every visible item
|
||||
of clothing and accessory as a premium fashion product flat lay. The output
|
||||
must contain **no person, body parts, mannequin, model, hanger, or human
|
||||
silhouette**—only the garments and accessories arranged as a coherent set.
|
||||
|
||||
## Baseline
|
||||
|
||||
This prompt starts from the current best Qwen Image 2.1 I2I workflow:
|
||||
|
||||
- Source baseline: [`optimized/I2I/qwen/`](../../optimized/I2I/qwen/)
|
||||
- Model: Qwen Image 2.1 INT8 ConvRot
|
||||
- Sampler: Euler CFG++
|
||||
- Scheduler: beta
|
||||
- Steps: 50
|
||||
- CFG: 3.5
|
||||
- Output: portrait 1080×1920 baseline, adjustable for composition
|
||||
- Reference input: one image containing the person and outfit
|
||||
|
||||
The copied workflow files preserve this starting point. The optimized Qwen
|
||||
workflow remains the source of truth for future baseline changes.
|
||||
|
||||
## Expected result
|
||||
|
||||
The image should look like a carefully art-directed fashion editorial or
|
||||
premium e-commerce catalogue photograph. It should identify and lay out every
|
||||
visible garment and accessory, preserve material, colour, pattern, construction
|
||||
details and relative styling, and keep each item separated enough to be read.
|
||||
|
||||
Suggested arrangement: a clean neutral surface, garments folded or naturally
|
||||
laid flat in neat groups, accessories beside the relevant outfit, soft
|
||||
directional studio lighting, realistic fabric texture, subtle contact shadows,
|
||||
and a balanced composition with no clutter.
|
||||
|
||||
## Input assumptions
|
||||
|
||||
- Clothing and accessories should be inferred only from what is visible.
|
||||
- Hidden garments, unseen backs, and speculative details should not be invented.
|
||||
- If multiple people are present, show their visible clothing as separate sets
|
||||
while keeping the output people-free.
|
||||
|
||||
## Iteration process
|
||||
|
||||
1. Replace the reference image in the Qwen workflow.
|
||||
2. Use `prompt.txt` as the positive prompt and keep baseline settings fixed.
|
||||
3. Save each generated image and exact metadata in `results/`.
|
||||
4. Record outfit coverage, human-removal quality, material fidelity, and layout.
|
||||
5. Change the prompt only when the visual result shows a repeatable weakness.
|
||||
6. Mark the accepted prompt revision and current best result here.
|
||||
|
||||
## Current status
|
||||
|
||||
- Status: scaffolded; first generation pending.
|
||||
- Current best result: none yet.
|
||||
- First test: baseline Qwen workflow with `prompt.txt`.
|
||||
@@ -0,0 +1 @@
|
||||
Transform the supplied reference image into a premium fashion product flat-lay photograph of every visible clothing item and accessory worn by the person. Carefully identify the garments, footwear, bags, jewellery, eyewear, belts, hats and other visible accessories, preserving their exact colours, materials, textures, patterns, proportions, construction details and overall styling. Arrange the complete outfit as a coherent editorial fashion catalogue composition on a clean neutral surface: garments neatly folded or naturally laid flat in separate, readable groups, accessories placed beside the relevant garments, balanced spacing, tasteful overlap only where it improves the styling, realistic fabric drape, crisp seams and believable contact shadows. Use soft directional studio lighting, subtle tonal contrast, accurate material response and high-end product photography detail. The result must contain no person, no model, no mannequin, no body parts, no hands, no head, no face, no limbs, no human silhouette, no hanger and no human reflection. Show only the clothing and accessories, with a refined fashion editorial layout and no text, logos, labels, watermark or clutter.
|
||||
@@ -0,0 +1,12 @@
|
||||
# Results
|
||||
|
||||
Store generated images and matching metadata records here.
|
||||
|
||||
Use this naming pattern:
|
||||
|
||||
- `<prompt-id>-<iteration>.png`
|
||||
- `<prompt-id>-<iteration>.json`
|
||||
|
||||
Each JSON record should include the source image name or hash, exact prompt,
|
||||
negative prompt, seed, dimensions, sampler, scheduler, CFG, steps, runtime,
|
||||
and a short visual verdict.
|
||||
@@ -0,0 +1,28 @@
|
||||
{
|
||||
"id": "qwen-image-2.1-photorealistic-i2i-ui",
|
||||
"revision": 0,
|
||||
"last_node_id": 15,
|
||||
"last_link_id": 18,
|
||||
"nodes": [
|
||||
{"id":1,"type":"UNETLoader","pos":[-900,0],"size":[360,110],"flags":{},"order":0,"mode":0,"inputs":[],"outputs":[{"name":"MODEL","type":"MODEL","links":[1]}],"properties":{"Node name for S&R":"UNETLoader"},"widgets_values":["qwen_image_2.1_int8_convrot.safetensors","default"]},
|
||||
{"id":2,"type":"CLIPLoader","pos":[-900,180],"size":[360,150],"flags":{},"order":1,"mode":0,"inputs":[],"outputs":[{"name":"CLIP","type":"CLIP","links":[2]}],"properties":{"Node name for S&R":"CLIPLoader"},"widgets_values":["qwen3vl_8b_w4a8.safetensors","qwen_image","default"]},
|
||||
{"id":3,"type":"VAELoader","pos":[-900,390],"size":[360,110],"flags":{},"order":2,"mode":0,"inputs":[],"outputs":[{"name":"VAE","type":"VAE","links":[3,4]}],"properties":{"Node name for S&R":"VAELoader"},"widgets_values":["qwen_image_2.1_vae_bf16.safetensors"]},
|
||||
{"id":4,"type":"LoadImage","pos":[-900,600],"size":[360,420],"flags":{},"order":3,"mode":0,"inputs":[],"outputs":[{"name":"IMAGE","type":"IMAGE","links":[5]}],"properties":{"Node name for S&R":"LoadImage"},"widgets_values":["character_reference.png"]},
|
||||
{"id":5,"type":"ResizeAndPadImage","pos":[-420,620],"size":[320,190],"flags":{},"order":4,"mode":0,"inputs":[{"name":"image","type":"IMAGE","link":5}],"outputs":[{"name":"IMAGE","type":"IMAGE","links":[6]}],"properties":{"Node name for S&R":"ResizeAndPadImage"},"widgets_values":[576,1024,"white","lanczos"]},
|
||||
{"id":6,"type":"QwenImage21Cache","pos":[-420,0],"size":[300,130],"flags":{},"order":5,"mode":0,"inputs":[{"name":"model","type":"MODEL","link":1}],"outputs":[{"name":"MODEL","type":"MODEL","links":[7,8]}],"properties":{"Node name for S&R":"QwenImage21Cache"},"widgets_values":["auto","int8"]},
|
||||
{"id":7,"type":"TextEncodeQwenImage21","pos":[0,220],"size":[650,470],"flags":{},"order":6,"mode":0,"inputs":[{"name":"clip","type":"CLIP","link":2},{"name":"vae","type":"VAE","link":3},{"name":"image_1","type":"IMAGE","link":6}],"outputs":[{"name":"positive","type":"CONDITIONING","links":[9]},{"name":"negative","type":"CONDITIONING","links":[10]},{"name":"latent","type":"LATENT","links":[11]}],"properties":{"Node name for S&R":"TextEncodeQwenImage21"},"widgets_values":["Transform the supplied reference image into a photorealistic image of the same subject. Infer identity, facial features, skin tone, hairstyle, clothing, accessories, colors, pose, body proportions, composition, and framing directly from the reference. Preserve those visual attributes faithfully while replacing the original rendering style with natural human anatomy, realistic skin and hair texture, physically accurate materials, believable fabric and surface detail, natural hands and fingers, cinematic soft light, realistic shadows, coherent perspective, premium editorial photography, high dynamic range, photorealistic, highly detailed. Do not redesign, restyle, substitute, add, remove, mirror, or crop the subject. No text or logos.","different person, changed identity, changed face, changed hairstyle, changed clothing, changed colors, altered accessories, changed pose, different composition, extra person, duplicate person, extra limbs, malformed hands, fused fingers, distorted anatomy, cropped head, cropped feet, mirrored image, cartoon, illustration, anime, CGI, plastic skin, waxy skin, blurry, low detail, text, logo, watermark",0]},
|
||||
{"id":8,"type":"KSamplerSelect","pos":[760,0],"size":[280,90],"flags":{},"order":7,"mode":0,"inputs":[],"outputs":[{"name":"SAMPLER","type":"SAMPLER","links":[12]}],"properties":{"Node name for S&R":"KSamplerSelect"},"widgets_values":["euler_cfg_pp"]},
|
||||
{"id":9,"type":"BasicScheduler","pos":[760,150],"size":[300,150],"flags":{},"order":8,"mode":0,"inputs":[{"name":"model","type":"MODEL","link":7}],"outputs":[{"name":"SIGMAS","type":"SIGMAS","links":[13]}],"properties":{"Node name for S&R":"BasicScheduler"},"widgets_values":["beta",50,1]},
|
||||
{"id":10,"type":"RandomNoise","pos":[760,360],"size":[280,90],"flags":{},"order":9,"mode":0,"inputs":[],"outputs":[{"name":"NOISE","type":"NOISE","links":[14]}],"properties":{"Node name for S&R":"RandomNoise"},"widgets_values":[20260921]},
|
||||
{"id":11,"type":"CFGGuider","pos":[1120,120],"size":[300,150],"flags":{},"order":10,"mode":0,"inputs":[{"name":"model","type":"MODEL","link":8},{"name":"positive","type":"CONDITIONING","link":9},{"name":"negative","type":"CONDITIONING","link":10}],"outputs":[{"name":"GUIDER","type":"GUIDER","links":[15]}],"properties":{"Node name for S&R":"CFGGuider"},"widgets_values":[3.5]},
|
||||
{"id":12,"type":"SamplerCustomAdvanced","pos":[1480,120],"size":[320,220],"flags":{},"order":11,"mode":0,"inputs":[{"name":"noise","type":"NOISE","link":14},{"name":"guider","type":"GUIDER","link":15},{"name":"sampler","type":"SAMPLER","link":12},{"name":"sigmas","type":"SIGMAS","link":13},{"name":"latent_image","type":"LATENT","link":11}],"outputs":[{"name":"output","type":"LATENT","links":[16]}],"properties":{"Node name for S&R":"SamplerCustomAdvanced"},"widgets_values":[]},
|
||||
{"id":13,"type":"VAEDecode","pos":[1850,120],"size":[260,110],"flags":{},"order":12,"mode":0,"inputs":[{"name":"samples","type":"LATENT","link":16},{"name":"vae","type":"VAE","link":4}],"outputs":[{"name":"IMAGE","type":"IMAGE","links":[17]}],"properties":{"Node name for S&R":"VAEDecode"},"widgets_values":[]},
|
||||
{"id":14,"type":"ResizeAndPadImage","pos":[2160,120],"size":[320,190],"flags":{},"order":13,"mode":0,"inputs":[{"name":"image","type":"IMAGE","link":17}],"outputs":[{"name":"IMAGE","type":"IMAGE","links":[18]}],"properties":{"Node name for S&R":"ResizeAndPadImage"},"widgets_values":[1080,1920,"white","lanczos"]},
|
||||
{"id":15,"type":"SaveImage","pos":[2540,120],"size":[320,110],"flags":{},"order":14,"mode":0,"inputs":[{"name":"images","type":"IMAGE","link":18}],"outputs":[],"properties":{"Node name for S&R":"SaveImage"},"widgets_values":["fluxExperiments/imageEdit/qwen_i2i"]}
|
||||
],
|
||||
"links": [[1,1,0,6,0,"MODEL"],[2,2,0,7,0,"CLIP"],[3,3,0,7,1,"VAE"],[4,3,0,13,1,"VAE"],[5,4,0,5,0,"IMAGE"],[6,5,0,7,2,"IMAGE"],[7,6,0,9,0,"MODEL"],[8,6,0,11,0,"MODEL"],[9,7,0,11,1,"CONDITIONING"],[10,7,1,11,2,"CONDITIONING"],[11,7,2,12,4,"LATENT"],[12,8,0,12,2,"SAMPLER"],[13,9,0,12,3,"SIGMAS"],[14,10,0,12,0,"NOISE"],[15,11,0,12,1,"GUIDER"],[16,12,0,13,0,"LATENT"],[17,13,0,14,0,"IMAGE"],[18,14,0,15,0,"IMAGE"]],
|
||||
"groups": [],
|
||||
"config": {},
|
||||
"extra": {"note":"Qwen Image 2.1 photorealistic I2I; current best Q-I2I-34; 1080x1920 output"},
|
||||
"version": 0.4
|
||||
}
|
||||
@@ -0,0 +1,17 @@
|
||||
{
|
||||
"1": {"class_type":"UNETLoader","inputs":{"unet_name":"qwen_image_2.1_int8_convrot.safetensors","weight_dtype":"default"}},
|
||||
"2": {"class_type":"CLIPLoader","inputs":{"clip_name":"qwen3vl_8b_w4a8.safetensors","type":"qwen_image","device":"default"}},
|
||||
"3": {"class_type":"VAELoader","inputs":{"vae_name":"qwen_image_2.1_vae_bf16.safetensors"}},
|
||||
"4": {"class_type":"LoadImage","inputs":{"image":"character_reference.png"}},
|
||||
"5": {"class_type":"ResizeAndPadImage","inputs":{"image":["4",0],"target_width":576,"target_height":1024,"padding_color":"white","interpolation":"lanczos"}},
|
||||
"6": {"class_type":"QwenImage21Cache","inputs":{"model":["1",0],"device":"auto","dtype":"int8"}},
|
||||
"7": {"class_type":"TextEncodeQwenImage21","inputs":{"clip":["2",0],"prompt":"Transform the supplied reference image into a photorealistic image of the same subject. Infer identity, facial features, skin tone, hairstyle, clothing, accessories, colors, pose, body proportions, composition, and framing directly from the reference. Preserve those visual attributes faithfully while replacing the original rendering style with natural human anatomy, realistic skin and hair texture, physically accurate materials, believable fabric and surface detail, natural hands and fingers, cinematic soft light, realistic shadows, coherent perspective, premium editorial photography, high dynamic range, photorealistic, highly detailed. Do not redesign, restyle, substitute, add, remove, mirror, or crop the subject. No text or logos.","negative_prompt":"different person, changed identity, changed face, changed hairstyle, changed clothing, changed colors, altered accessories, changed pose, different composition, extra person, duplicate person, extra limbs, malformed hands, fused fingers, distorted anatomy, cropped head, cropped feet, mirrored image, cartoon, illustration, anime, CGI, plastic skin, waxy skin, blurry, low detail, text, logo, watermark","resolution":0,"images.image_1":["5",0],"vae":["3",0]}},
|
||||
"8": {"class_type":"KSamplerSelect","inputs":{"sampler_name":"euler_cfg_pp"}},
|
||||
"9": {"class_type":"BasicScheduler","inputs":{"model":["6",0],"scheduler":"beta","steps":50,"denoise":1}},
|
||||
"10": {"class_type":"RandomNoise","inputs":{"noise_seed":20260921}},
|
||||
"11": {"class_type":"CFGGuider","inputs":{"model":["6",0],"positive":["7",0],"negative":["7",1],"cfg":3.5}},
|
||||
"12": {"class_type":"SamplerCustomAdvanced","inputs":{"noise":["10",0],"guider":["11",0],"sampler":["8",0],"sigmas":["9",0],"latent_image":["7",2]}},
|
||||
"13": {"class_type":"VAEDecode","inputs":{"samples":["12",0],"vae":["3",0]}},
|
||||
"14": {"class_type":"ResizeAndPadImage","inputs":{"image":["13",0],"target_width":1080,"target_height":1920,"padding_color":"white","interpolation":"lanczos"}},
|
||||
"15": {"class_type":"SaveImage","inputs":{"images":["14",0],"filename_prefix":"fluxExperiments/imageEdit/qwen_i2i"}}
|
||||
}
|
||||
Reference in New Issue
Block a user