| { | |
| "_class_name": "StableDiffusionPipeline", | |
| "_diffusers_version": "0.21.0", | |
| "_name_or_path": "OpenTrouter/Trouter-Imagine-1", | |
| "model_type": "stable-diffusion", | |
| "architecture": "latent-diffusion", | |
| "model_config": { | |
| "base_model": "stable-diffusion-v1-5", | |
| "fine_tuned": true, | |
| "training_resolution": 512, | |
| "max_resolution": 1024, | |
| "aspect_ratios_supported": [ | |
| "1:1", | |
| "3:2", | |
| "2:3", | |
| "4:3", | |
| "3:4", | |
| "16:9", | |
| "9:16" | |
| ], | |
| "optimal_resolutions": [ | |
| {"width": 512, "height": 512, "description": "Standard square, fastest"}, | |
| {"width": 768, "height": 768, "description": "High quality square"}, | |
| {"width": 1024, "height": 1024, "description": "Maximum quality square"}, | |
| {"width": 768, "height": 512, "description": "Landscape standard"}, | |
| {"width": 512, "height": 768, "description": "Portrait standard"}, | |
| {"width": 1024, "height": 768, "description": "Landscape HD"}, | |
| {"width": 768, "height": 1024, "description": "Portrait HD"} | |
| ] | |
| }, | |
| "vae_config": { | |
| "_class_name": "AutoencoderKL", | |
| "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, | |
| "norm_num_groups": 32, | |
| "scaling_factor": 0.18215 | |
| }, | |
| "unet_config": { | |
| "_class_name": "UNet2DConditionModel", | |
| "sample_size": 64, | |
| "in_channels": 4, | |
| "out_channels": 4, | |
| "center_input_sample": false, | |
| "flip_sin_to_cos": true, | |
| "freq_shift": 0, | |
| "down_block_types": [ | |
| "CrossAttnDownBlock2D", | |
| "CrossAttnDownBlock2D", | |
| "CrossAttnDownBlock2D", | |
| "DownBlock2D" | |
| ], | |
| "up_block_types": [ | |
| "UpBlock2D", | |
| "CrossAttnUpBlock2D", | |
| "CrossAttnUpBlock2D", | |
| "CrossAttnUpBlock2D" | |
| ], | |
| "only_cross_attention": false, | |
| "block_out_channels": [320, 640, 1280, 1280], | |
| "layers_per_block": 2, | |
| "downsample_padding": 1, | |
| "mid_block_scale_factor": 1, | |
| "act_fn": "silu", | |
| "norm_num_groups": 32, | |
| "norm_eps": 1e-05, | |
| "cross_attention_dim": 768, | |
| "attention_head_dim": 8, | |
| "dual_cross_attention": false, | |
| "use_linear_projection": false, | |
| "class_embed_type": null, | |
| "num_class_embeds": null, | |
| "upcast_attention": false, | |
| "resnet_time_scale_shift": "default" | |
| }, | |
| "text_encoder_config": { | |
| "_class_name": "CLIPTextModel", | |
| "architectures": ["CLIPTextModel"], | |
| "model_type": "clip_text_model", | |
| "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, | |
| "dropout": 0.0, | |
| "attention_dropout": 0.0, | |
| "initializer_range": 0.02, | |
| "initializer_factor": 1.0, | |
| "pad_token_id": 1, | |
| "bos_token_id": 0, | |
| "eos_token_id": 2 | |
| }, | |
| "tokenizer_config": { | |
| "_class_name": "CLIPTokenizer", | |
| "add_prefix_space": false, | |
| "model_max_length": 77, | |
| "pad_token": "<|endoftext|>", | |
| "unk_token": "<|endoftext|>", | |
| "bos_token": "<|startoftext|>", | |
| "eos_token": "<|endoftext|>" | |
| }, | |
| "scheduler_config": { | |
| "_class_name": "PNDMScheduler", | |
| "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 | |
| }, | |
| "inference_defaults": { | |
| "num_inference_steps": 30, | |
| "guidance_scale": 7.5, | |
| "width": 512, | |
| "height": 512, | |
| "eta": 0.0, | |
| "clip_skip": 0, | |
| "description": "Recommended default parameters for balanced quality and speed" | |
| }, | |
| "quality_presets": { | |
| "draft": { | |
| "num_inference_steps": 15, | |
| "guidance_scale": 6.0, | |
| "width": 512, | |
| "height": 512, | |
| "description": "Fast generation for previews and iterations" | |
| }, | |
| "balanced": { | |
| "num_inference_steps": 30, | |
| "guidance_scale": 7.5, | |
| "width": 512, | |
| "height": 512, | |
| "description": "Good balance of quality and speed" | |
| }, | |
| "quality": { | |
| "num_inference_steps": 50, | |
| "guidance_scale": 8.0, | |
| "width": 768, | |
| "height": 768, | |
| "description": "High quality output, slower generation" | |
| }, | |
| "maximum": { | |
| "num_inference_steps": 75, | |
| "guidance_scale": 9.0, | |
| "width": 1024, | |
| "height": 1024, | |
| "description": "Maximum quality, slowest generation" | |
| } | |
| }, | |
| "style_presets": { | |
| "photorealistic": { | |
| "guidance_scale": 8.5, | |
| "negative_prompt": "cartoon, anime, painting, drawing, illustration, low quality, blurry, distorted", | |
| "recommended_keywords": ["photography", "photorealistic", "realistic", "4k", "high detail", "professional photo"] | |
| }, | |
| "artistic": { | |
| "guidance_scale": 7.0, | |
| "negative_prompt": "low quality, blurry, distorted, bad anatomy", | |
| "recommended_keywords": ["oil painting", "digital art", "concept art", "illustration", "artistic"] | |
| }, | |
| "anime": { | |
| "guidance_scale": 7.5, | |
| "negative_prompt": "realistic, 3d, photograph, low quality, blurry", | |
| "recommended_keywords": ["anime", "manga", "anime style", "cel shaded", "highly detailed anime"] | |
| }, | |
| "cinematic": { | |
| "guidance_scale": 8.0, | |
| "negative_prompt": "low quality, amateur, home video, poor lighting", | |
| "recommended_keywords": ["cinematic", "movie still", "film grain", "dramatic lighting", "bokeh"] | |
| }, | |
| "fantasy": { | |
| "guidance_scale": 7.5, | |
| "negative_prompt": "modern, contemporary, realistic, photograph", | |
| "recommended_keywords": ["fantasy art", "magical", "ethereal", "dreamlike", "mystical"] | |
| }, | |
| "scifi": { | |
| "guidance_scale": 7.5, | |
| "negative_prompt": "medieval, fantasy, low tech, primitive", | |
| "recommended_keywords": ["sci-fi", "futuristic", "cyberpunk", "tech", "advanced technology"] | |
| } | |
| }, | |
| "negative_prompt_presets": { | |
| "general": "blurry, low quality, distorted, deformed, ugly, bad anatomy, extra limbs, mutation, disfigured, bad proportions, watermark, signature, text", | |
| "people": "blurry, low quality, distorted, deformed, ugly, bad anatomy, extra limbs, mutation, disfigured, bad proportions, extra fingers, missing fingers, fused fingers, too many fingers, extra arms, missing arms, extra legs, missing legs, bad hands, bad feet, malformed limbs, disconnected limbs, floating limbs", | |
| "faces": "blurry, low quality, distorted, deformed, ugly, bad anatomy, asymmetric eyes, crossed eyes, lazy eye, bad teeth, open mouth, cloned face, duplicate, two faces, malformed features, distorted features", | |
| "animals": "blurry, low quality, distorted, deformed, ugly, bad anatomy, extra limbs, mutation, disfigured, bad proportions, extra legs, missing legs, fused limbs, incorrect anatomy", | |
| "landscapes": "blurry, low quality, distorted, oversaturated, underexposed, overexposed, ugly, poor composition, watermark, signature, text", | |
| "architecture": "blurry, low quality, distorted, bad perspective, warped, crooked, asymmetric, poor composition, watermark, signature, text" | |
| }, | |
| "hardware_recommendations": { | |
| "minimum": { | |
| "gpu": "NVIDIA RTX 2060 (6GB VRAM)", | |
| "ram": "16GB", | |
| "storage": "10GB SSD", | |
| "max_resolution": "512x512", | |
| "max_batch_size": 1, | |
| "notes": "Enable all memory optimizations, use DPM++ scheduler" | |
| }, | |
| "recommended": { | |
| "gpu": "NVIDIA RTX 3080 (10GB VRAM)", | |
| "ram": "32GB", | |
| "storage": "20GB SSD", | |
| "max_resolution": "768x768", | |
| "max_batch_size": 2, | |
| "notes": "Good balance for most users, enables higher quality settings" | |
| }, | |
| "optimal": { | |
| "gpu": "NVIDIA RTX 4090 (24GB VRAM) or A100", | |
| "ram": "64GB", | |
| "storage": "50GB NVMe SSD", | |
| "max_resolution": "1024x1024", | |
| "max_batch_size": 4, | |
| "notes": "Maximum quality and speed, can generate multiple high-res images" | |
| }, | |
| "apple_silicon": { | |
| "chip": "M1 Pro/Max/Ultra or M2 Pro/Max/Ultra", | |
| "ram": "32GB+ unified memory", | |
| "storage": "20GB SSD", | |
| "max_resolution": "768x768", | |
| "max_batch_size": 1, | |
| "notes": "Use MPS backend, enable attention slicing" | |
| } | |
| }, | |
| "scheduler_recommendations": { | |
| "DPMSolverMultistep": { | |
| "speed": "fast", | |
| "quality": "high", | |
| "recommended_steps": "20-25", | |
| "best_for": "General purpose, fast iterations", | |
| "description": "Fast and high-quality, best default choice" | |
| }, | |
| "EulerAncestralDiscrete": { | |
| "speed": "medium", | |
| "quality": "high", | |
| "recommended_steps": "30-40", | |
| "best_for": "Artistic and creative outputs", | |
| "description": "More creative and varied results" | |
| }, | |
| "DDIM": { | |
| "speed": "medium", | |
| "quality": "good", | |
| "recommended_steps": "30-50", | |
| "best_for": "Reproducible results", | |
| "description": "Deterministic sampler, good for reproducibility" | |
| }, | |
| "PNDM": { | |
| "speed": "slow", | |
| "quality": "high", | |
| "recommended_steps": "40-50", | |
| "best_for": "Maximum quality", | |
| "description": "Original scheduler, very stable" | |
| } | |
| }, | |
| "training_info": { | |
| "base_model": "stable-diffusion-v1-5", | |
| "dataset_size": "5M+ images", | |
| "training_steps": 500000, | |
| "batch_size": 256, | |
| "learning_rate": 1e-5, | |
| "optimizer": "AdamW", | |
| "mixed_precision": "fp16", | |
| "gradient_accumulation_steps": 4, | |
| "ema_decay": 0.9999, | |
| "training_hardware": "8x A100 80GB", | |
| "training_duration": "~3 weeks", | |
| "checkpoint_frequency": 10000, | |
| "validation_frequency": 5000 | |
| }, | |
| "performance_metrics": { | |
| "fid_score": 12.3, | |
| "inception_score": 28.5, | |
| "clip_score": 0.31, | |
| "lpips_score": 0.45, | |
| "generation_time_512x512": { | |
| "RTX_3080": "3-5 seconds", | |
| "RTX_4090": "1-2 seconds", | |
| "A100": "1-2 seconds", | |
| "M1_Max": "8-12 seconds" | |
| }, | |
| "generation_time_1024x1024": { | |
| "RTX_3080": "12-18 seconds", | |
| "RTX_4090": "4-6 seconds", | |
| "A100": "4-6 seconds", | |
| "M1_Max": "30-45 seconds" | |
| } | |
| }, | |
| "supported_features": { | |
| "text_to_image": true, | |
| "image_to_image": true, | |
| "inpainting": false, | |
| "outpainting": false, | |
| "controlnet": false, | |
| "lora": true, | |
| "textual_inversion": true, | |
| "attention_slicing": true, | |
| "vae_slicing": true, | |
| "xformers": true, | |
| "safety_checker": true, | |
| "watermarking": false | |
| }, | |
| "api_compatibility": { | |
| "diffusers": ">=0.21.0", | |
| "transformers": ">=4.25.0", | |
| "torch": ">=2.0.0", | |
| "cuda": ">=11.7", | |
| "python": ">=3.8" | |
| }, | |
| "example_prompts": { | |
| "landscapes": [ | |
| "a serene mountain landscape at golden hour, misty valleys, dramatic clouds, professional landscape photography, highly detailed", | |
| "tropical beach at sunset, crystal clear water, palm trees, vibrant colors, paradise scene, 4k photography", | |
| "autumn forest path, colorful leaves, soft sunlight filtering through trees, peaceful atmosphere, oil painting style" | |
| ], | |
| "portraits": [ | |
| "portrait of a woman with flowing red hair, green eyes, natural lighting, professional photography, highly detailed, 4k", | |
| "close-up portrait of an elderly man with wise expression, wrinkled face, dramatic lighting, black and white photography", | |
| "fantasy portrait of an elf warrior, silver hair, armor, forest background, digital art, highly detailed" | |
| ], | |
| "architecture": [ | |
| "modern glass skyscraper reflecting sunset clouds, urban architecture, professional photography, ultra detailed", | |
| "ancient japanese temple surrounded by cherry blossoms, spring scene, traditional architecture, cinematic lighting", | |
| "futuristic cyberpunk city street at night, neon lights, rain-slicked roads, flying cars, highly detailed sci-fi art" | |
| ], | |
| "fantasy": [ | |
| "majestic dragon perched on mountain peak, wings spread, breathing fire, epic fantasy art, highly detailed", | |
| "magical library with floating books, glowing crystals, ethereal atmosphere, fantasy concept art", | |
| "enchanted forest with bioluminescent plants, mystical creatures, moonlight, dreamlike atmosphere, digital painting" | |
| ], | |
| "scifi": [ | |
| "massive spaceship orbiting alien planet, stars in background, sci-fi concept art, highly detailed", | |
| "robot with glowing blue eyes in dark laboratory, cyberpunk style, dramatic lighting, cinematic", | |
| "futuristic megacity with flying vehicles, holographic advertisements, night scene, blade runner style" | |
| ], | |
| "animals": [ | |
| "majestic lion with golden mane, intense gaze, african savanna background, wildlife photography, 4k", | |
| "colorful parrot in tropical rainforest, vibrant feathers, natural habitat, professional wildlife photography", | |
| "pod of dolphins jumping through ocean waves at sunset, dynamic movement, nature photography" | |
| ] | |
| }, | |
| "license": "Apache-2.0", | |
| "version": "1.0.0", | |
| "release_date": "2025-11-02", | |
| "maintainer": "OpenTrouter Team", | |
| "repository": "https://huggingface.co/OpenTrouter/Trouter-Imagine-1" | |
| } |