| { | |
| "architectures": [ | |
| "GenerannoForMaskedLM" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_generanno.GenerannoConfig", | |
| "AutoModel": "modeling_generanno.GenerannoModel", | |
| "AutoModelForMaskedLM": "modeling_generanno.GenerannoForMaskedLM", | |
| "AutoModelForSequenceClassification": "modeling_generanno.GenerannoForSequenceClassification", | |
| "AutoModelForTokenClassification": "modeling_generanno.GenerannoForTokenClassification" | |
| }, | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "hidden_act": "silu", | |
| "hidden_size": 1280, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 3520, | |
| "mask_token_id": 4, | |
| "max_position_embeddings": 8192, | |
| "mlp_bias": false, | |
| "model_type": "generanno", | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 28, | |
| "num_key_value_heads": 4, | |
| "pad_token_id": 3, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": null, | |
| "rope_theta": 500000.0, | |
| "tie_word_embeddings": false, | |
| "dtype": "float32", | |
| "transformers_version": "4.44.0", | |
| "vocab_size": 64 | |
| } |