config.json

{
  "_name_or_path": "/models/glm-4-9b-chat",
  "add_bias_linear": false,
  "add_qkv_bias": true,
  "apply_query_key_layer_scaling": true,
  "apply_residual_connection_post_layernorm": false,
  "architectures": [
    "ChatGLMForConditionalGeneration"
  ],
  "attention_dropout": 0.0,
  "attention_softmax_in_fp32": true,
  "auto_map": {
    "AutoConfig": "configuration_chatglm.ChatGLMConfig",
    "AutoModel": "modeling_chatglm.ChatGLMForConditionalGeneration",
    "AutoModelForCausalLM": "modeling_chatglm.ChatGLMForConditionalGeneration",
    "AutoModelForSeq2SeqLM": "modeling_chatglm.ChatGLMForConditionalGeneration",
    "AutoModelForSequenceClassification": "modeling_chatglm.ChatGLMForSequenceClassification"
  },
  "bias_dropout_fusion": true,
  "classifier_dropout": null,
  "eos_token_id": [
    151329,
    151336,
    151338
  ],
  "ffn_hidden_size": 13696,
  "fp32_residual_connection": false,
  "hidden_dropout": 0.0,
  "hidden_size": 4096,
  "kv_channels": 128,
  "layernorm_epsilon": 1.5625e-07,
  "model_type": "chatglm",
  "multi_query_attention": true,
  "multi_query_group_num": 2,
  "num_attention_heads": 32,
  "num_hidden_layers": 40,
  "num_layers": 40,
  "original_rope": true,
  "pad_token_id": 151329,
  "padded_vocab_size": 151552,
  "post_layer_norm": true,
  "quantization_config": {
    "amp": true,
    "autoround_version": "0.4.0.dev",
    "backend": "auto_round:gptq:exllamav2",
    "batch_size": 8,
    "bits": 4,
    "data_type": "int",
    "dataset": "NeelNanda/pile-10k",
    "enable_minmax_tuning": true,
    "enable_norm_bias_tuning": false,
    "enable_quanted_input": true,
    "gradient_accumulate_steps": 1,
    "group_size": 128,
    "iters": 1000,
    "low_gpu_mem_usage": false,
    "lr": 0.001,
    "minmax_lr": 0.001,
    "nsamples": 512,
    "quant_block_list": null,
    "quant_method": "intel/auto-round",
    "scale_dtype": "torch.float16",
    "seqlen": 2048,
    "sym": true
  },
  "rmsnorm": true,
  "rope_ratio": 500,
  "seq_length": 131072,
  "tie_word_embeddings": false,
  "torch_dtype": "float16",
  "transformers_version": "4.46.1",
  "use_cache": true,
  "vocab_size": 151552
}