DaertML/WizardCoder-15B-V1.0-nf4-8bit-doublequant-BNB

1{2  "_name_or_path": "WizardLM/WizardCoder-15B-V1.0",3  "activation_function": "gelu",4  "architectures": [5    "GPTBigCodeForCausalLM"6  ],7  "attention_softmax_in_fp32": true,8  "attn_pdrop": 0.1,9  "bos_token_id": 0,10  "embd_pdrop": 0.1,11  "eos_token_id": 0,12  "inference_runner": 0,13  "initializer_range": 0.02,14  "layer_norm_epsilon": 1e-05,15  "max_batch_size": null,16  "max_sequence_length": null,17  "model_type": "gpt_bigcode",18  "multi_query": true,19  "n_embd": 6144,20  "n_head": 48,21  "n_inner": 24576,22  "n_layer": 40,23  "n_positions": 8192,24  "pad_key_length": true,25  "pre_allocate_kv_cache": false,26  "quantization_config": {27    "bnb_4bit_compute_dtype": "bfloat16",28    "bnb_4bit_quant_type": "nf4",29    "bnb_4bit_use_double_quant": true,30    "llm_int8_enable_fp32_cpu_offload": false,31    "llm_int8_has_fp16_weight": false,32    "llm_int8_skip_modules": null,33    "llm_int8_threshold": 6.0,34    "load_in_4bit": false,35    "load_in_8bit": true,36    "quant_method": "bitsandbytes"37  },38  "resid_pdrop": 0.1,39  "scale_attention_softmax_in_fp32": true,40  "scale_attn_weights": true,41  "summary_activation": null,42  "summary_first_dropout": 0.1,43  "summary_proj_to_labels": true,44  "summary_type": "cls_index",45  "summary_use_proj": true,46  "torch_dtype": "float16",47  "transformers_version": "4.33.1",48  "use_cache": true,49  "validate_runner_input": true,50  "vocab_size": 4915351}52