| { | |
| "architectures": [ | |
| "RWForCausalLM" | |
| ], | |
| "attention_dropout": 0.0, | |
| "auto_map": { | |
| "AutoConfig": "configuration_yalm.YalmConfig", | |
| "AutoModel": "modelling_yalm.YalmModel", | |
| "AutoModelForSequenceClassification": "modelling_yalm.RWForSequenceClassification", | |
| "AutoModelForCausalLM": "modelling_yalm.YalmForCausalLM" | |
| }, | |
| "padded_vocab_size": 128000, | |
| "embedding_size": 2048, | |
| "hidden_size": 10240, | |
| "intermediate_size": 27308, | |
| "num_layers": 80, | |
| "num_attention_heads": 128, | |
| "scale_attn_by_inverse_layer_idx": true, | |
| "activation_type": "geglu", | |
| "model_type": "YaLM", | |
| "max_position_embeddings": 1024, | |
| "apply_residual_connection_post_layernorm": false, | |
| "initializer_range": 0.02, | |
| "layernorm_epsilon": 1e-5, | |
| "torch_dtype": "float16", | |
| "transformers_version": "4.32.1" | |
| } | |