| { |
| "architectures": [ |
| "RavenForImageForensics" |
| ], |
| "model_type": "raven", |
| "name": "Raven", |
| "version": "0.1.0", |
| "backbone": "dinov2_vits14", |
| "hidden_size": 384, |
| "feature_dims": { |
| "semantic": 384, |
| "camera": 52, |
| "spectral": 47, |
| "patch": 36, |
| "specialist": 8, |
| "quality": 12 |
| }, |
| "ensemble_size": 5, |
| "thresholds": { |
| "real": 0.28, |
| "ai": 0.72 |
| }, |
| "num_labels": 2, |
| "id2label": { |
| "0": "REAL", |
| "1": "AI" |
| }, |
| "label2id": { |
| "REAL": 0, |
| "AI": 1 |
| }, |
| "checkpoint_structure": { |
| "version": 5, |
| "name": "Raven", |
| "config": { |
| "project": { |
| "name": "Raven F2K", |
| "version": "0.1.0" |
| }, |
| "hardware": { |
| "torch_threads": 6, |
| "interop_threads": 1, |
| "dataloader_workers": 0, |
| "preprocess_workers": 2, |
| "train_device": "cpu", |
| "foundation_device": "auto", |
| "directml_preferred_for_onnx": true |
| }, |
| "foundation": { |
| "backend": "dinov2_vits14", |
| "image_size": 224, |
| "accelerator": "auto", |
| "directml_batch_size": 4, |
| "openvino_batch_size": 16, |
| "torch_batch_size": 8, |
| "runtime_cache_dir": "cache/foundation_runtime", |
| "openvino_compress_fp16": true, |
| "reuse_semantic_for_augmented_views": false, |
| "augment_max_per_class": 6000, |
| "train_coreset": true, |
| "train_cap_per_class": 0, |
| "cache_semantic_float16": true, |
| "compress_feature_cache": false, |
| "fallback_backend": "torchvision_convnext_tiny", |
| "allow_fallback": true |
| }, |
| "real_specialists": { |
| "map_size": 64, |
| "latent_dim": 48, |
| "max_real_images": 14010, |
| "precompute_workers": 2, |
| "epochs": 12, |
| "batch_size": 64, |
| "lr": 0.0016, |
| "weight_decay": 0.0001, |
| "mask_fraction": 0.27, |
| "camera_noise_std": 0.016, |
| "early_stop_patience": 3 |
| }, |
| "features": { |
| "patch_count": 8, |
| "patch_fraction": 0.32, |
| "spectral_radial_bins": 32, |
| "spectral_angular_bins": 8, |
| "camera_feature_dim_hint": 56, |
| "quality_dim_hint": 12, |
| "resolution_debias": true, |
| "resolution_policy": "fixed_square_analysis_grid" |
| }, |
| "training": { |
| "ensemble_members": 5, |
| "epochs": 60, |
| "batch_size": 256, |
| "lr": 0.00135, |
| "weight_decay": 0.01, |
| "early_stop_patience": 7, |
| "label_smoothing": 0.02, |
| "vib_beta": 0.0015, |
| "branch_loss_weight": 0.12, |
| "consistency_weight": 0.08, |
| "contrastive_weight": 0.028, |
| "branch_dropout": 0.12, |
| "hard_refine_epochs": 6, |
| "hard_refine_lr_multiplier": 0.22, |
| "hard_replay_boost": 4.0, |
| "adversarial_weight_ai": 0.035, |
| "adversarial_weight_real": 0.02, |
| "source_dropout_per_member": 1, |
| "max_minority_oversample_factor": 2.0, |
| "feature_noise_semantic": 0.012, |
| "feature_noise_forensic": 0.02, |
| "seed": 1337 |
| }, |
| "anchors": { |
| "pca_max_components": 48, |
| "eps": 1e-06 |
| }, |
| "calibration": { |
| "bootstrap_members": 15, |
| "target_false_ai_rate": 0.015, |
| "target_false_real_rate": 0.04, |
| "minimum_ai_threshold": 0.72, |
| "maximum_real_threshold": 0.28 |
| }, |
| "manifest": { |
| "val_fraction": 0.15, |
| "real_test_fraction": 0.15, |
| "dedupe_hamming_distance": 3, |
| "ai_source_name": "gpt-image-2" |
| }, |
| "budget": { |
| "target_hours": 4.5, |
| "minimum_hours": 3.0, |
| "maximum_hours": 6.0, |
| "benchmark_runtime": "openvino", |
| "foundation_images_per_second": 14.9871539016036 |
| } |
| }, |
| "dims": { |
| "semantic": 384, |
| "camera": 52, |
| "spectral": 47, |
| "patch": 36, |
| "specialist": 8, |
| "quality": 12 |
| }, |
| "foundation_backend": "dinov2_vits14", |
| "foundation_dim": 384, |
| "foundation_state": { |
| "backend": "dinov2_vits14", |
| "dim": 384, |
| "state_dict": { |
| "cls_token": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.cls_token", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 1, |
| 384 |
| ] |
| }, |
| "pos_embed": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.pos_embed", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 1370, |
| 384 |
| ] |
| }, |
| "mask_token": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.mask_token", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 384 |
| ] |
| }, |
| "patch_embed.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.patch_embed.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 3, |
| 14, |
| 14 |
| ] |
| }, |
| "patch_embed.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.patch_embed.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.0.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.0.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.0.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.0.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.0.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.0.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.0.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.0.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.0.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.0.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.0.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.0.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.0.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.0.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.0.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.1.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.1.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.1.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.1.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.1.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.1.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.1.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.1.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.1.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.1.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.1.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.1.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.1.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.1.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.1.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.2.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.2.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.2.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.2.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.2.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.2.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.2.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.2.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.2.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.2.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.2.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.2.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.2.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.2.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.2.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.3.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.3.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.3.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.3.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.3.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.3.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.3.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.3.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.3.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.3.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.3.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.3.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.3.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.3.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.3.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.4.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.4.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.4.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.4.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.4.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.4.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.4.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.4.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.4.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.4.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.4.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.4.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.4.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.4.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.4.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.5.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.5.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.5.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.5.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.5.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.5.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.5.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.5.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.5.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.5.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.5.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.5.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.5.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.5.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.5.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.6.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.6.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.6.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.6.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.6.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.6.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.6.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.6.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.6.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.6.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.6.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.6.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.6.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.6.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.6.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.7.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.7.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.7.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.7.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.7.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.7.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.7.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.7.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.7.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.7.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.7.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.7.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.7.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.7.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.7.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.8.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.8.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.8.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.8.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.8.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.8.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.8.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.8.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.8.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.8.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.8.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.8.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.8.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.8.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.8.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.9.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.9.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.9.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.9.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.9.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.9.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.9.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.9.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.9.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.9.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.9.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.9.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.9.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.9.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.9.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.10.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.10.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.10.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.10.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.10.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.10.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.10.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.10.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.10.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.10.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.10.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.10.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.10.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.10.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.10.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.11.norm1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.norm1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.11.norm1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.norm1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.11.attn.qkv.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.attn.qkv.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152, |
| 384 |
| ] |
| }, |
| "blocks.11.attn.qkv.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.attn.qkv.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1152 |
| ] |
| }, |
| "blocks.11.attn.proj.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.attn.proj.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 384 |
| ] |
| }, |
| "blocks.11.attn.proj.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.attn.proj.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.11.ls1.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.ls1.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.11.norm2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.norm2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.11.norm2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.norm2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.11.mlp.fc1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.mlp.fc1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536, |
| 384 |
| ] |
| }, |
| "blocks.11.mlp.fc1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.mlp.fc1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1536 |
| ] |
| }, |
| "blocks.11.mlp.fc2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.mlp.fc2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384, |
| 1536 |
| ] |
| }, |
| "blocks.11.mlp.fc2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.mlp.fc2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "blocks.11.ls2.gamma": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.blocks.11.ls2.gamma", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "norm.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.norm.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "norm.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.foundation_state.state_dict.norm.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| } |
| } |
| }, |
| "normalizer": { |
| "semantic": { |
| "mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.semantic.mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "std": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.semantic.std", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ], |
| "numpy_dtype": "float32" |
| } |
| }, |
| "camera": { |
| "mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.camera.mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "std": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.camera.std", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ], |
| "numpy_dtype": "float32" |
| } |
| }, |
| "spectral": { |
| "mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.spectral.mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "std": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.spectral.std", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ], |
| "numpy_dtype": "float32" |
| } |
| }, |
| "patch": { |
| "mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.patch.mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "std": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.patch.std", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ], |
| "numpy_dtype": "float32" |
| } |
| }, |
| "specialist": { |
| "mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.specialist.mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "std": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.normalizer.specialist.std", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ], |
| "numpy_dtype": "float32" |
| } |
| } |
| }, |
| "anchors": { |
| "semantic": { |
| "scaler_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.semantic.scaler_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "scaler_scale": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.semantic.scaler_scale", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.semantic.pca_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_components": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.semantic.pca_components", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 384 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "location": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.semantic.location", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "precision": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.semantic.precision", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "dist_mean": 6.791811456942514, |
| "dist_std": 1.3277029621489218 |
| }, |
| "camera": { |
| "scaler_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.camera.scaler_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "scaler_scale": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.camera.scaler_scale", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.camera.pca_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_components": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.camera.pca_components", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 52 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "location": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.camera.location", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "precision": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.camera.precision", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "dist_mean": 5.520394421018968, |
| "dist_std": 2.4500221910948587 |
| }, |
| "spectral": { |
| "scaler_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.spectral.scaler_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "scaler_scale": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.spectral.scaler_scale", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.spectral.pca_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_components": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.spectral.pca_components", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47, |
| 47 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "location": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.spectral.location", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "precision": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.spectral.precision", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47, |
| 47 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "dist_mean": 6.245673063484355, |
| "dist_std": 2.164304630474745 |
| }, |
| "patch": { |
| "scaler_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.patch.scaler_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "scaler_scale": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.patch.scaler_scale", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.patch.pca_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_components": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.patch.pca_components", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36, |
| 36 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "location": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.patch.location", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "precision": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.patch.precision", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36, |
| 36 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "dist_mean": 5.390552832973847, |
| "dist_std": 2.0362408822655875 |
| }, |
| "specialist": { |
| "scaler_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.specialist.scaler_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "scaler_scale": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.specialist.scaler_scale", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.specialist.pca_mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "pca_components": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.specialist.pca_components", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8, |
| 8 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "location": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.specialist.location", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "precision": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.anchors.specialist.precision", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8, |
| 8 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "dist_mean": 2.2865919167065503, |
| "dist_std": 1.598617687764203 |
| } |
| }, |
| "ai_source_map": { |
| "gpt-image-2": 0 |
| }, |
| "real_source_map": { |
| "real:images": 0 |
| }, |
| "ensemble_states": { |
| "__raven_type__": "list", |
| "items": [ |
| { |
| "towers.semantic.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.semantic.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.semantic.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.semantic.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.semantic.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.semantic.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.semantic.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.camera.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.camera.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.camera.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.camera.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.camera.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.camera.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.camera.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.spectral.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.spectral.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.spectral.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.spectral.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.spectral.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.spectral.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.patch.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.patch.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.patch.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.patch.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.patch.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.patch.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.specialist.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.specialist.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.specialist.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.specialist.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.specialist.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.specialist.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.towers.specialist.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "branch_heads.semantic.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.semantic.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 96 |
| ] |
| }, |
| "branch_heads.semantic.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.semantic.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.camera.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.camera.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.camera.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.camera.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.spectral.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.spectral.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.spectral.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.spectral.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.patch.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.patch.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.patch.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.patch.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.specialist.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.specialist.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "branch_heads.specialist.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.branch_heads.specialist.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "pre_vib.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.pre_vib.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.pre_vib.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.pre_vib.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128, |
| 264 |
| ] |
| }, |
| "pre_vib.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.pre_vib.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128 |
| ] |
| }, |
| "mu.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.mu.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "mu.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.mu.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "logvar.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.logvar.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "logvar.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.logvar.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "router.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.router.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.router.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.router.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 12 |
| ] |
| }, |
| "router.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.router.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "router.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.router.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5, |
| 48 |
| ] |
| }, |
| "router.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.router.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.anomaly_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.anomaly_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.anomaly_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 5 |
| ] |
| }, |
| "anomaly_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.anomaly_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "anomaly_head.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.anomaly_head.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "anomaly_head.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.anomaly_head.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "global_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.global_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.global_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.global_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 64 |
| ] |
| }, |
| "global_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.global_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "global_head.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.global_head.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "global_head.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.global_head.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "mixer.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.mixer.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 3 |
| ] |
| }, |
| "mixer.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.0.mixer.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| } |
| }, |
| { |
| "towers.semantic.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.semantic.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.semantic.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.semantic.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.semantic.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.semantic.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.semantic.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.camera.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.camera.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.camera.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.camera.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.camera.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.camera.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.camera.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.spectral.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.spectral.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.spectral.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.spectral.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.spectral.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.spectral.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.patch.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.patch.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.patch.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.patch.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.patch.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.patch.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.specialist.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.specialist.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.specialist.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.specialist.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.specialist.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.specialist.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.towers.specialist.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "branch_heads.semantic.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.semantic.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 96 |
| ] |
| }, |
| "branch_heads.semantic.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.semantic.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.camera.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.camera.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.camera.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.camera.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.spectral.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.spectral.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.spectral.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.spectral.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.patch.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.patch.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.patch.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.patch.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.specialist.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.specialist.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "branch_heads.specialist.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.branch_heads.specialist.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "pre_vib.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.pre_vib.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.pre_vib.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.pre_vib.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128, |
| 264 |
| ] |
| }, |
| "pre_vib.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.pre_vib.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128 |
| ] |
| }, |
| "mu.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.mu.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "mu.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.mu.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "logvar.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.logvar.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "logvar.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.logvar.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "router.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.router.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.router.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.router.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 12 |
| ] |
| }, |
| "router.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.router.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "router.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.router.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5, |
| 48 |
| ] |
| }, |
| "router.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.router.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.anomaly_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.anomaly_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.anomaly_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 5 |
| ] |
| }, |
| "anomaly_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.anomaly_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "anomaly_head.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.anomaly_head.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "anomaly_head.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.anomaly_head.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "global_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.global_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.global_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.global_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 64 |
| ] |
| }, |
| "global_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.global_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "global_head.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.global_head.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "global_head.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.global_head.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "mixer.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.mixer.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 3 |
| ] |
| }, |
| "mixer.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.1.mixer.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| } |
| }, |
| { |
| "towers.semantic.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.semantic.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.semantic.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.semantic.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.semantic.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.semantic.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.semantic.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.camera.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.camera.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.camera.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.camera.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.camera.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.camera.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.camera.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.spectral.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.spectral.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.spectral.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.spectral.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.spectral.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.spectral.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.patch.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.patch.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.patch.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.patch.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.patch.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.patch.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.specialist.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.specialist.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.specialist.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.specialist.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.specialist.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.specialist.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.towers.specialist.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "branch_heads.semantic.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.semantic.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 96 |
| ] |
| }, |
| "branch_heads.semantic.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.semantic.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.camera.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.camera.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.camera.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.camera.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.spectral.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.spectral.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.spectral.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.spectral.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.patch.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.patch.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.patch.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.patch.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.specialist.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.specialist.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "branch_heads.specialist.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.branch_heads.specialist.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "pre_vib.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.pre_vib.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.pre_vib.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.pre_vib.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128, |
| 264 |
| ] |
| }, |
| "pre_vib.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.pre_vib.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128 |
| ] |
| }, |
| "mu.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.mu.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "mu.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.mu.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "logvar.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.logvar.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "logvar.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.logvar.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "router.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.router.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.router.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.router.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 12 |
| ] |
| }, |
| "router.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.router.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "router.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.router.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5, |
| 48 |
| ] |
| }, |
| "router.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.router.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.anomaly_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.anomaly_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.anomaly_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 5 |
| ] |
| }, |
| "anomaly_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.anomaly_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "anomaly_head.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.anomaly_head.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "anomaly_head.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.anomaly_head.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "global_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.global_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.global_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.global_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 64 |
| ] |
| }, |
| "global_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.global_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "global_head.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.global_head.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "global_head.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.global_head.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "mixer.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.mixer.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 3 |
| ] |
| }, |
| "mixer.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.2.mixer.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| } |
| }, |
| { |
| "towers.semantic.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.semantic.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.semantic.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.semantic.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.semantic.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.semantic.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.semantic.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.camera.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.camera.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.camera.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.camera.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.camera.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.camera.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.camera.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.spectral.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.spectral.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.spectral.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.spectral.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.spectral.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.spectral.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.patch.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.patch.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.patch.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.patch.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.patch.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.patch.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.specialist.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.specialist.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.specialist.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.specialist.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.specialist.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.specialist.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.towers.specialist.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "branch_heads.semantic.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.semantic.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 96 |
| ] |
| }, |
| "branch_heads.semantic.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.semantic.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.camera.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.camera.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.camera.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.camera.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.spectral.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.spectral.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.spectral.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.spectral.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.patch.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.patch.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.patch.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.patch.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.specialist.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.specialist.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "branch_heads.specialist.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.branch_heads.specialist.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "pre_vib.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.pre_vib.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.pre_vib.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.pre_vib.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128, |
| 264 |
| ] |
| }, |
| "pre_vib.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.pre_vib.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128 |
| ] |
| }, |
| "mu.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.mu.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "mu.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.mu.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "logvar.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.logvar.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "logvar.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.logvar.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "router.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.router.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.router.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.router.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 12 |
| ] |
| }, |
| "router.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.router.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "router.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.router.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5, |
| 48 |
| ] |
| }, |
| "router.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.router.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.anomaly_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.anomaly_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.anomaly_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 5 |
| ] |
| }, |
| "anomaly_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.anomaly_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "anomaly_head.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.anomaly_head.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "anomaly_head.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.anomaly_head.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "global_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.global_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.global_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.global_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 64 |
| ] |
| }, |
| "global_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.global_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "global_head.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.global_head.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "global_head.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.global_head.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "mixer.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.mixer.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 3 |
| ] |
| }, |
| "mixer.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.3.mixer.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| } |
| }, |
| { |
| "towers.semantic.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.semantic.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.semantic.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.semantic.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 384 |
| ] |
| }, |
| "towers.semantic.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.semantic.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.semantic.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96, |
| 96 |
| ] |
| }, |
| "towers.semantic.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.semantic.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 96 |
| ] |
| }, |
| "towers.camera.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.camera.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.camera.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.camera.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 52 |
| ] |
| }, |
| "towers.camera.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.camera.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.camera.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.camera.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.camera.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.spectral.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.spectral.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.spectral.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 47 |
| ] |
| }, |
| "towers.spectral.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.spectral.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.spectral.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.spectral.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.spectral.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.patch.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.patch.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.patch.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 36 |
| ] |
| }, |
| "towers.patch.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.patch.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.patch.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48 |
| ] |
| }, |
| "towers.patch.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.patch.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "towers.specialist.net.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.specialist.net.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.specialist.net.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.specialist.net.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 8 |
| ] |
| }, |
| "towers.specialist.net.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.specialist.net.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.specialist.net.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 24 |
| ] |
| }, |
| "towers.specialist.net.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.towers.specialist.net.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "branch_heads.semantic.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.semantic.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 96 |
| ] |
| }, |
| "branch_heads.semantic.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.semantic.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.camera.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.camera.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.camera.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.camera.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.spectral.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.spectral.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.spectral.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.spectral.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.patch.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.patch.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "branch_heads.patch.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.patch.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "branch_heads.specialist.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.specialist.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "branch_heads.specialist.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.branch_heads.specialist.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "pre_vib.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.pre_vib.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.pre_vib.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 264 |
| ] |
| }, |
| "pre_vib.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.pre_vib.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128, |
| 264 |
| ] |
| }, |
| "pre_vib.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.pre_vib.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 128 |
| ] |
| }, |
| "mu.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.mu.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "mu.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.mu.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "logvar.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.logvar.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64, |
| 128 |
| ] |
| }, |
| "logvar.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.logvar.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "router.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.router.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.router.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 12 |
| ] |
| }, |
| "router.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.router.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 12 |
| ] |
| }, |
| "router.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.router.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "router.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.router.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5, |
| 48 |
| ] |
| }, |
| "router.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.router.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.anomaly_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.anomaly_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 5 |
| ] |
| }, |
| "anomaly_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.anomaly_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24, |
| 5 |
| ] |
| }, |
| "anomaly_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.anomaly_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 24 |
| ] |
| }, |
| "anomaly_head.3.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.anomaly_head.3.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 24 |
| ] |
| }, |
| "anomaly_head.3.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.anomaly_head.3.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "global_head.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.global_head.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.global_head.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 64 |
| ] |
| }, |
| "global_head.1.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.global_head.1.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 64 |
| ] |
| }, |
| "global_head.1.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.global_head.1.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "global_head.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.global_head.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 48 |
| ] |
| }, |
| "global_head.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.global_head.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| }, |
| "mixer.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.mixer.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1, |
| 3 |
| ] |
| }, |
| "mixer.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.ensemble_states.4.mixer.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| } |
| } |
| ] |
| }, |
| "members": { |
| "__raven_type__": "list", |
| "items": [ |
| { |
| "best_auc": 0.9976531080825559, |
| "excluded_ai_sources": { |
| "__raven_type__": "list", |
| "items": [] |
| } |
| }, |
| { |
| "best_auc": 0.9976878381786357, |
| "excluded_ai_sources": { |
| "__raven_type__": "list", |
| "items": [] |
| } |
| }, |
| { |
| "best_auc": 0.997580015984924, |
| "excluded_ai_sources": { |
| "__raven_type__": "list", |
| "items": [] |
| } |
| }, |
| { |
| "best_auc": 0.997248604610568, |
| "excluded_ai_sources": { |
| "__raven_type__": "list", |
| "items": [] |
| } |
| }, |
| { |
| "best_auc": 0.9975754761030834, |
| "excluded_ai_sources": { |
| "__raven_type__": "list", |
| "items": [] |
| } |
| } |
| ] |
| }, |
| "calibrator": { |
| "version": 2, |
| "prior": "balanced_0.5_0.5", |
| "coef": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.calibrator.coef", |
| "dtype": "torch.float32", |
| "shape": [ |
| 15, |
| 25 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "intercept": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.calibrator.intercept", |
| "dtype": "torch.float32", |
| "shape": [ |
| 15 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "mean": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.calibrator.mean", |
| "dtype": "torch.float32", |
| "shape": [ |
| 25 |
| ], |
| "numpy_dtype": "float32" |
| }, |
| "std": { |
| "__raven_type__": "numpy_array", |
| "key": "raven.calibrator.std", |
| "dtype": "torch.float32", |
| "shape": [ |
| 25 |
| ], |
| "numpy_dtype": "float32" |
| } |
| }, |
| "real_threshold": 0.28, |
| "ai_threshold": 0.72, |
| "validation_metrics": { |
| "n": 4470, |
| "accuracy@0.5_raw": 0.9863534675615212, |
| "balanced_accuracy@0.5": 0.9856595347392894, |
| "AUROC": 0.9982546424264216, |
| "AP_raw_prevalence_sensitive": 0.9971892233721003, |
| "AP_balanced": 0.9984715280833314, |
| "balanced_Brier": 0.011402727910489683, |
| "coverage": 0.9805369127516779, |
| "selective_accuracy_raw": 0.9926990645676478, |
| "false_AI_rate_REAL_to_AI": 0.005994005994005994, |
| "false_REAL_rate_AI_to_REAL": 0.00954328561690525, |
| "AI_confirmed_recall": 0.9700068166325835, |
| "REAL_confirmed_recall": 0.975024975024975, |
| "uncertain_rate": 0.019463087248322148, |
| "ECE_balanced": 0.00557215572591252, |
| "real_threshold": 0.28, |
| "ai_threshold": 0.72, |
| "selective_accuracy_balanced": 0.9920737506147539 |
| }, |
| "specialists_checkpoint": { |
| "version": 4, |
| "map_size": 64, |
| "latent_dim": 48, |
| "spectral_state": { |
| "enc.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.enc.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 16, |
| 1, |
| 3, |
| 3 |
| ] |
| }, |
| "enc.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.enc.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 16 |
| ] |
| }, |
| "enc.2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.enc.2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 32, |
| 16, |
| 3, |
| 3 |
| ] |
| }, |
| "enc.2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.enc.2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 32 |
| ] |
| }, |
| "enc.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.enc.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 32, |
| 3, |
| 3 |
| ] |
| }, |
| "enc.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.enc.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "enc.6.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.enc.6.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48, |
| 3, |
| 3 |
| ] |
| }, |
| "enc.6.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.enc.6.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "dec.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.dec.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48, |
| 4, |
| 4 |
| ] |
| }, |
| "dec.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.dec.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "dec.2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.dec.2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 32, |
| 4, |
| 4 |
| ] |
| }, |
| "dec.2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.dec.2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 32 |
| ] |
| }, |
| "dec.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.dec.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 32, |
| 16, |
| 4, |
| 4 |
| ] |
| }, |
| "dec.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.dec.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 16 |
| ] |
| }, |
| "dec.6.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.dec.6.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 16, |
| 1, |
| 4, |
| 4 |
| ] |
| }, |
| "dec.6.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.spectral_state.dec.6.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| } |
| }, |
| "camera_state": { |
| "enc.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.enc.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 16, |
| 1, |
| 3, |
| 3 |
| ] |
| }, |
| "enc.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.enc.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 16 |
| ] |
| }, |
| "enc.2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.enc.2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 32, |
| 16, |
| 3, |
| 3 |
| ] |
| }, |
| "enc.2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.enc.2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 32 |
| ] |
| }, |
| "enc.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.enc.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 32, |
| 3, |
| 3 |
| ] |
| }, |
| "enc.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.enc.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "enc.6.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.enc.6.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48, |
| 3, |
| 3 |
| ] |
| }, |
| "enc.6.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.enc.6.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "dec.0.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.dec.0.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 48, |
| 4, |
| 4 |
| ] |
| }, |
| "dec.0.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.dec.0.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48 |
| ] |
| }, |
| "dec.2.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.dec.2.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 48, |
| 32, |
| 4, |
| 4 |
| ] |
| }, |
| "dec.2.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.dec.2.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 32 |
| ] |
| }, |
| "dec.4.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.dec.4.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 32, |
| 16, |
| 4, |
| 4 |
| ] |
| }, |
| "dec.4.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.dec.4.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 16 |
| ] |
| }, |
| "dec.6.weight": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.dec.6.weight", |
| "dtype": "torch.float32", |
| "shape": [ |
| 16, |
| 1, |
| 4, |
| 4 |
| ] |
| }, |
| "dec.6.bias": { |
| "__raven_type__": "torch_tensor", |
| "key": "raven.specialists_checkpoint.camera_state.dec.6.bias", |
| "dtype": "torch.float32", |
| "shape": [ |
| 1 |
| ] |
| } |
| }, |
| "validation_loss": 0.0503488639369607, |
| "capped_real_images": 14010 |
| } |
| } |
| } |
|
|
