WisdomShell/Shell-7B-Chat

1{2  "_name_or_path": "WisdomShell/Shell-7B-Chat",3  "activation_function": "gelu_pytorch_tanh",4  "architectures": [5    "CodeShellForCausalLM"6  ],7  "attention_softmax_in_fp32": true,8  "attn_pdrop": 0.1,9  "auto_map": {10    "AutoConfig": "configuration_codeshell.CodeShellConfig",11    "AutoModelForCausalLM": "modeling_codeshell.CodeShellForCausalLM"12  },13  "bos_token_id": 70000,14  "embd_pdrop": 0.1,15  "eos_token_id": 70000,16  "group_query_attention": true,17  "inference_runner": 0,18  "initializer_range": 0.02,19  "layer_norm_epsilon": 1e-05,20  "max_batch_size": null,21  "max_sequence_length": null,22  "model_type": "codeshell",23  "n_embd": 4096,24  "n_head": 32,25  "n_inner": 16384,26  "n_layer": 42,27  "n_positions": 8192,28  "num_query_groups": 8,29  "pad_key_length": true,30  "position_embedding_type": "rope",31  "pre_allocate_kv_cache": false,32  "resid_pdrop": 0.1,33  "rope_scaling": null,34  "scale_attention_softmax_in_fp32": true,35  "scale_attn_weights": true,36  "summary_activation": null,37  "summary_first_dropout": 0.1,38  "summary_proj_to_labels": true,39  "summary_type": "cls_index",40  "summary_use_proj": true,41  "torch_dtype": "bfloat16",42  "transformers_version": "4.36.2",43  "use_cache": true,44  "validate_runner_input": true,45  "vocab_size": 7014446}47