| { | |
| "architectures": [ | |
| "MotifVisionModel" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "modeling_motif_vision_encoder.MotifVisionConfig", | |
| "AutoModel": "modeling_motif_vision_encoder.MotifVisionModel" | |
| }, | |
| "depth": 40, | |
| "drop_path_rate": 0.0, | |
| "dtype": "bfloat16", | |
| "embed_dim": 4096, | |
| "ffn_bias": true, | |
| "ffn_layer": "swiglu64", | |
| "ffn_ratio": 3.0, | |
| "gated_attention": "elementwise", | |
| "img_size": 512, | |
| "in_chans": 3, | |
| "layerscale_init": 1e-05, | |
| "mask_k_bias": true, | |
| "model_type": "motif_vision", | |
| "n_storage_tokens": 4, | |
| "norm_layer": "layernormbf16", | |
| "num_frames": 1, | |
| "num_heads": 32, | |
| "patch_size": 16, | |
| "pos_embed_rope_base": 100.0, | |
| "pos_embed_rope_rescale_coords": 2.0, | |
| "proj_bias": true, | |
| "qk_norm": true, | |
| "qkv_bias": false, | |
| "transformers_version": "5.8.1", | |
| "tubelet_size": 2, | |
| "untie_cls_and_patch_norms": false, | |
| "untie_global_and_local_cls_norm": true | |
| } | |