| { | |
| "activation_dropout": 0.1, | |
| "architectures": [ | |
| "RetNetForCausalLM" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_retnet.RetNetConfig", | |
| "FlaxAutoModel": "modeling_flax_retnet.FlaxRetNetModel", | |
| "FlaxAutoModelForCausalLM": "modeling_flax_retnet.FlaxRetNetForCausalLM" | |
| }, | |
| "attention_type": "parallel", | |
| "dropout": 0.1, | |
| "hidden_act": "gelu", | |
| "hidden_size": 768, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 1536, | |
| "layer_norm_eps": 1e-05, | |
| "max_position_embeddings": 512, | |
| "model_type": "retnet", | |
| "normalize_before": false, | |
| "num_hidden_layers": 12, | |
| "num_rettention_heads": 4, | |
| "output_retentions": false, | |
| "recurrent_chunk_size": 512, | |
| "transformers_version": "4.29.2", | |
| "vocab_size": 50432 | |
| } | |