| { | |
| "_commit_hash": null, | |
| "_name_or_path": "runs/387m/checkpoint-5000", | |
| "activation_function": "gelu_new", | |
| "architectures": [ | |
| "GiaModel" | |
| ], | |
| "attention_dropout": 0.0, | |
| "attention_layers": [ | |
| "global", | |
| "global", | |
| "global", | |
| "global", | |
| "global", | |
| "global", | |
| "global", | |
| "global", | |
| "global", | |
| "global", | |
| "global", | |
| "global" | |
| ], | |
| "bos_token_id": 50256, | |
| "classifier_dropout": 0.1, | |
| "embed_dropout": 0.0, | |
| "eos_token_id": 50256, | |
| "hidden_size": 1536, | |
| "image_vocab_size": 128, | |
| "initializer_range": 0.02, | |
| "intermediate_size": null, | |
| "layer_norm_epsilon": 1e-05, | |
| "max_local_position": 512, | |
| "model_type": "gia", | |
| "num_groups": 32, | |
| "num_heads": 12, | |
| "num_layers": 12, | |
| "num_res_channels": 64, | |
| "patch_size": 16, | |
| "resid_dropout": 0.0, | |
| "torch_dtype": "float32", | |
| "transformers_version": null, | |
| "use_cache": true, | |
| "vocab_size": 31025 | |
| } | |