| { | |
| "_name_or_path": "core42/jais-family-256m", | |
| "activation_function": "swiglu", | |
| "alibi_scaling": null, | |
| "architectures": [ | |
| "JAISLMHeadModel" | |
| ], | |
| "attn_pdrop": 0.0, | |
| "auto_map": { | |
| "AutoConfig": "configuration_jais.JAISConfig", | |
| "AutoModel": "modeling_jais.JAISModel", | |
| "AutoModelForCausalLM": "modeling_jais.JAISLMHeadModel", | |
| "AutoModelForQuestionAnswering": "modeling_jais.JAISForQuestionAnswering", | |
| "AutoModelForSequenceClassification": "modeling_jais.JAISForSequenceClassification", | |
| "AutoModelForTokenClassification": "modeling_jais.JAISForTokenClassification" | |
| }, | |
| "bos_token_id": 0, | |
| "embd_pdrop": 0.0, | |
| "eos_token_id": 0, | |
| "initializer_range": 0.02, | |
| "layer_norm_epsilon": 1e-05, | |
| "model_type": "jais", | |
| "mup_embeddings_scale": 9.1705785388303, | |
| "mup_output_alpha": 1.09518349815769, | |
| "mup_scale_qk_dot_by_d": true, | |
| "mup_width_scale": 0.23529411764705882, | |
| "n_embd": 1088, | |
| "n_head": 17, | |
| "n_inner": 2912, | |
| "n_layer": 14, | |
| "n_positions": 2048, | |
| "pad_token_id": 0, | |
| "position_embedding_type": "alibi", | |
| "reorder_and_upcast_attn": false, | |
| "resid_pdrop": 0.0, | |
| "rotary_dim": null, | |
| "scale_attn_by_inverse_layer_idx": false, | |
| "scale_attn_weights": true, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.40.1", | |
| "use_cache": true, | |
| "vocab_size": 84992 | |
| } | |