{ "architectures": [ "CLIPModel" ], "initializer_factor": 1.0, "logit_scale_init_value": 2.6592, "model_type": "clip", "projection_dim": 512, "text_config": { "hidden_act": "gelu", "hidden_size": 384, "intermediate_size": 1536, "model_type": "clip_text_model", "num_attention_heads": 6 }, "torch_dtype": "float32", "transformers_version": "4.44.0", "vision_config": { "hidden_act": "gelu", "hidden_size": 192, "intermediate_size": 768, "model_type": "clip_vision_model", "num_attention_heads": 3, "patch_size": 16 } }