{ "_name_or_path": "checkpoint-19000", "activation_function": "gelu_new", "architectures": [ "GPT2ALMHeadModel" ], "attn_bias": false, "attn_pdrop": 0.0, "auto_map": { "AutoConfig": "configuration_gpt2a.GPT2AConfig", "AutoModelForCausalLM": "modeling_gpt2a.GPT2ALMHeadModel" }, "bos_token_id": 50256, "embd_pdrop": 0.0, "eos_token_id": 50256, "initializer_range": 0.02, "layer_norm_epsilon": 1e-06, "mlp_bias": false, "model_type": "gpt2a", "n_embd": 768, "n_head": 6, "n_inner": 1920, "n_layer": 6, "n_positions": 4096, "reorder_and_upcast_attn": false, "resid_pdrop": 0.0, "scale_attn_by_inverse_layer_idx": false, "scale_attn_weights": true, "summary_activation": null, "summary_first_dropout": 0.1, "summary_proj_to_labels": true, "summary_type": "cls_index", "summary_use_proj": true, "torch_dtype": "float32", "transformers_version": "4.34.1", "use_cache": true, "vocab_size": 50257 }