| { | |
| "_name_or_path": "pszemraj/mega-ar-350m-L3t-v0.07-cosmo_webmath_py", | |
| "activation": "silu", | |
| "add_lm_hidden_dense_layer": false, | |
| "add_token_type_embeddings": false, | |
| "architectures": [ | |
| "MegaForCausalLM" | |
| ], | |
| "attention_activation": "softmax", | |
| "attention_probs_dropout_prob": 0, | |
| "bidirectional": false, | |
| "bos_token_id": 128000, | |
| "chunk_size": -1, | |
| "classifier_dropout": null, | |
| "dropout_prob": 0.0, | |
| "ema_beta_range": 0.02, | |
| "ema_delta_alpha_range": 0.2, | |
| "ema_gamma_omega_range": 1.0, | |
| "ema_projection_size": 16, | |
| "eos_token_id": 128001, | |
| "hidden_dropout_prob": 0.0, | |
| "hidden_size": 1024, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 3072, | |
| "is_decoder": true, | |
| "max_positions": 4096, | |
| "model_type": "mega", | |
| "nffn_activation_dropout_prob": 0.0, | |
| "nffn_hidden_size": 3072, | |
| "norm_affine": true, | |
| "normalization_type": "layernorm", | |
| "normalize_before_ffn": true, | |
| "normalize_before_mega": true, | |
| "num_attention_heads": 1, | |
| "num_hidden_layers": 12, | |
| "pad_token_id": null, | |
| "relative_positional_bias": "rotary", | |
| "shared_representation_size": 256, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.40.1", | |
| "truncation": null, | |
| "type_vocab_size": 1, | |
| "use_cache": true, | |
| "use_chunking": false, | |
| "use_feature_dropout": false, | |
| "use_normalized_ffn": true, | |
| "vocab_size": 128256 | |
| } | |