| { | |
| "architectures": [ | |
| "CLIPModel" | |
| ], | |
| "initializer_factor": 1.0, | |
| "logit_scale_init_value": 2.6592, | |
| "model_type": "clip", | |
| "projection_dim": 1024, | |
| "text_config": { | |
| "hidden_size": 1024, | |
| "intermediate_size": 4096, | |
| "model_type": "clip_text_model", | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 24 | |
| }, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.34.0", | |
| "vision_config": { | |
| "hidden_size": 1280, | |
| "intermediate_size": 5120, | |
| "model_type": "clip_vision_model", | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 32, | |
| "patch_size": 14 | |
| } | |
| } | |