WisdomShell/CodeShell-7B

1{2  "_name_or_path": "WisdomShell/CodeShell",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  "group_query_attention": true,14  "num_query_groups": 8,15  "position_embedding_type": "rope",16  "bos_token_id": 70000,17  "eos_token_id": 70000,18  "vocab_size": 70144,19  "embd_pdrop": 0.1,20  "inference_runner": 0,21  "initializer_range": 0.02,22  "layer_norm_epsilon": 1e-05,23  "max_batch_size": null,24  "max_sequence_length": null,25  "model_type": "kclgpt",26  "n_layer": 42,27  "n_embd": 4096,28  "n_inner": 16384,29  "n_head": 32,30  "n_positions": 8192,31  "pad_key_length": true,32  "resid_pdrop": 0.1,33  "rope_scaling": null,34  "pre_allocate_kv_cache": false,35  "scale_attention_softmax_in_fp32": true,36  "scale_attn_weights": true,37  "summary_activation": null,38  "summary_first_dropout": 0.1,39  "summary_proj_to_labels": true,40  "summary_type": "cls_index",41  "summary_use_proj": true,42  "torch_dtype": "bfloat16",43  "transformers_version": "4.31.0",44  "use_cache": true,45  "validate_runner_input": true46}47