| { | |
| "architectures": [ | |
| "Qwen3VLVisionModel" | |
| ], | |
| "deepstack_visual_indexes": [ | |
| 0, | |
| 1 | |
| ], | |
| "depth": 2, | |
| "dtype": "float32", | |
| "hidden_act": "silu", | |
| "hidden_size": 32, | |
| "in_channels": 3, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 32, | |
| "model_type": "qwen3_vl_vision", | |
| "num_heads": 4, | |
| "num_position_embeddings": 16, | |
| "out_hidden_size": 32, | |
| "patch_size": 16, | |
| "spatial_merge_size": 1, | |
| "temporal_patch_size": 2, | |
| "transformers_version": "5.16.0.dev0" | |
| } | |