| { | |
| "_class_name": "StableDiffusionControlNetInpaintPipeline", | |
| "_diffusers_version": "0.21.0", | |
| "_name_or_path": "runwayml/stable-diffusion-v1-5", | |
| "model_type": "stable-diffusion", | |
| "pipeline_class": "StableDiffusionControlNetInpaintPipeline", | |
| "components": { | |
| "unet": { | |
| "_class_name": "UNet2DConditionModel", | |
| "_name_or_path": "runwayml/stable-diffusion-v1-5/unet", | |
| "sample_size": 64, | |
| "in_channels": 9, | |
| "out_channels": 4, | |
| "down_block_types": [ | |
| "CrossAttnDownBlock2D", | |
| "CrossAttnDownBlock2D", | |
| "CrossAttnDownBlock2D", | |
| "DownBlock2D" | |
| ], | |
| "up_block_types": [ | |
| "UpBlock2D", | |
| "CrossAttnUpBlock2D", | |
| "CrossAttnUpBlock2D", | |
| "CrossAttnUpBlock2D" | |
| ], | |
| "block_out_channels": [320, 640, 1280, 1280], | |
| "layers_per_block": 2, | |
| "attention_head_dim": 8, | |
| "norm_num_groups": 32, | |
| "cross_attention_dim": 768 | |
| }, | |
| "controlnet": { | |
| "_class_name": "ControlNetModel", | |
| "_name_or_path": "lllyasviel/sd-controlnet-openpose", | |
| "conditioning_channels": 3, | |
| "conditioning_embedding_out_channels": [16, 32, 96, 256], | |
| "controlnet_conditioning_channel_order": "rgb", | |
| "cross_attention_dim": 768, | |
| "down_block_types": [ | |
| "CrossAttnDownBlock2D", | |
| "CrossAttnDownBlock2D", | |
| "CrossAttnDownBlock2D", | |
| "DownBlock2D" | |
| ], | |
| "flip_sin_to_cos": true, | |
| "freq_shift": 0, | |
| "layers_per_block": 2, | |
| "mid_block_scale_factor": 1, | |
| "norm_num_groups": 32, | |
| "only_cross_attention": false, | |
| "use_linear_projection": false | |
| }, | |
| "vae": { | |
| "_class_name": "AutoencoderKL", | |
| "_name_or_path": "runwayml/stable-diffusion-v1-5/vae", | |
| "sample_size": 512, | |
| "in_channels": 3, | |
| "out_channels": 3, | |
| "down_block_types": ["DownEncoderBlock2D", "DownEncoderBlock2D", "DownEncoderBlock2D", "DownEncoderBlock2D"], | |
| "up_block_types": ["UpDecoderBlock2D", "UpDecoderBlock2D", "UpDecoderBlock2D", "UpDecoderBlock2D"], | |
| "block_out_channels": [128, 256, 512, 512], | |
| "latent_channels": 4, | |
| "layers_per_block": 2 | |
| }, | |
| "text_encoder": { | |
| "_class_name": "CLIPTextModel", | |
| "_name_or_path": "runwayml/stable-diffusion-v1-5/text_encoder", | |
| "vocab_size": 49408, | |
| "hidden_size": 768, | |
| "intermediate_size": 3072, | |
| "num_hidden_layers": 12, | |
| "num_attention_heads": 12, | |
| "max_position_embeddings": 77, | |
| "hidden_act": "quick_gelu", | |
| "layer_norm_eps": 1e-05, | |
| "attention_dropout": 0.0, | |
| "initializer_range": 0.02, | |
| "initializer_factor": 1.0 | |
| }, | |
| "tokenizer": { | |
| "_class_name": "CLIPTokenizer", | |
| "_name_or_path": "runwayml/stable-diffusion-v1-5/tokenizer", | |
| "vocab_size": 49408, | |
| "model_max_length": 77, | |
| "pad_token": "<|endoftext|>", | |
| "unk_token": "<|endoftext|>", | |
| "bos_token": "<|startoftext|>", | |
| "eos_token": "<|endoftext|>" | |
| }, | |
| "scheduler": { | |
| "_class_name": "PNDMScheduler", | |
| "_name_or_path": "runwayml/stable-diffusion-v1-5/scheduler", | |
| "num_train_timesteps": 1000, | |
| "beta_start": 0.00085, | |
| "beta_end": 0.012, | |
| "beta_schedule": "scaled_linear", | |
| "trained_betas": null, | |
| "skip_prk_steps": true, | |
| "set_alpha_to_one": false, | |
| "prediction_type": "epsilon", | |
| "timestep_spacing": "leading", | |
| "steps_offset": 1 | |
| } | |
| }, | |
| "custom_pipeline": { | |
| "pose_system": { | |
| "method": "mediapipe", | |
| "coverage": { | |
| "total": 25.3, | |
| "body": 11.4, | |
| "hands": 3.5, | |
| "face": 6.7, | |
| "feet": 8.1, | |
| "skeleton": 13.2 | |
| }, | |
| "channels": 5, | |
| "target_size": [512, 512] | |
| }, | |
| "safety_configuration": { | |
| "mode": "fashion_moderate", | |
| "enable_debug": false, | |
| "safety_override": false, | |
| "log_all_decisions": true | |
| }, | |
| "generation_parameters": { | |
| "num_inference_steps": 50, | |
| "guidance_scale": 15.0, | |
| "strength": 0.75, | |
| "controlnet_conditioning_scale": 1.0, | |
| "negative_prompt": "low quality, blurry, distorted, deformed, extra limbs, bad anatomy, extra hands, extra arms, malformed hands, poorly drawn hands" | |
| }, | |
| "specialized_features": { | |
| "hand_exclusion": true, | |
| "proportion_preservation": true, | |
| "coverage_analysis": true, | |
| "checkpoint_support": ["realisticVisionV60B1_v51HyperVAE.safetensors"], | |
| "face_preservation": true | |
| } | |
| }, | |
| "task": "image-inpainting", | |
| "tags": [ | |
| "stable-diffusion", | |
| "controlnet", | |
| "pose-conditioning", | |
| "fashion-inpainting", | |
| "clothing-generation", | |
| "body-proportions", | |
| "pose-alignment" | |
| ], | |
| "library_name": "diffusers", | |
| "license": "openrail", | |
| "model_description": "Pose-conditioned fashion inpainting system using Stable Diffusion 1.5 + ControlNet OpenPose. Features 25.3% pose coverage, hand exclusion processing, and safety-aware generation optimized for realistic clothing generation while preserving body proportions and pose alignment.", | |
| "architecture": { | |
| "base_model": "runwayml/stable-diffusion-v1-5", | |
| "controlnet_model": "lllyasviel/sd-controlnet-openpose", | |
| "pose_detector": "mediapipe", | |
| "image_size": 512, | |
| "precision": "fp16" | |
| }, | |
| "performance_metrics": { | |
| "single_person_success_rate": 0.33, | |
| "pose_coverage_improvement": "25.3% vs 2% baseline", | |
| "safety_false_positive_reduction": "60-80%", | |
| "memory_stable": true | |
| } | |
| } |