| { | |
| "_name_or_path": "/content/drive/MyDrive/models_mlm/Llama_Custom-40M/checkpoint-5165", | |
| "architectures": [ | |
| "CustomLlamaForMaskedLM" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "modeling_custom_llama.CustomLlamaConfig", | |
| "AutoModelForMaskedLM": "modeling_custom_llama.CustomLlamaForMaskedLM" | |
| }, | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "head_dim": 48, | |
| "hidden_act": "silu", | |
| "hidden_size": 384, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 1280, | |
| "max_position_embeddings": 256, | |
| "mlp_bias": false, | |
| "mode": "mlm", | |
| "model_type": "custom_llama", | |
| "num_attention_heads": 8, | |
| "num_hidden_layers": 12, | |
| "num_key_value_heads": 8, | |
| "pad_token_id": 0, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "rope_theta": 10000.0, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.48.3", | |
| "use_cache": true, | |
| "vocab_size": 16002 | |
| } | |