Files
WeClone/settings.json
T

63 lines
2.0 KiB
JSON

{
"train_pt_args": {
"stage": "pt",
"dataset": "wechat-pt",
"dataset_dir": "./data/res_csv/pt",
"lora_target": "query_key_value",
"lora_rank": 2,
"lora_dropout": 0.1,
"output_dir": "model_output",
"overwrite_cache": true,
"per_device_train_batch_size": 1,
"gradient_accumulation_steps": 1,
"lr_scheduler_type": "cosine",
"logging_steps": 10,
"save_steps": 1000,
"learning_rate": 0.001,
"num_train_epochs": 30,
"plot_loss": true,
"fp16": true
},
"train_sft_args": {
"stage": "sft",
"dataset": "wechat-sft",
"dataset_dir": "./data/res_csv/sft",
"lora_target": "query_key_value",
"lora_rank": 4,
"lora_dropout": 0.5,
"overwrite_cache": true,
"per_device_train_batch_size": 4,
"gradient_accumulation_steps": 8,
"lr_scheduler_type": "cosine",
"logging_steps": 10,
"save_steps": 150,
"learning_rate": 0.0001,
"num_train_epochs": 3,
"plot_loss": true,
"fp16": true
},
"infer_args": {
"repetition_penalty": 1.2,
"temperature": 0.5,
"max_length": 50,
"top_p": 0.65
},
"make_dataset_args": {
// "enable_vision_model": false,//后续实现
// "include_type": [
// "文本"
// ],
"single_combine_strategy": "time_window", // 单人组成单句策略
"qa_match_strategy": "time_window", // 组成qa策略
"single_combine_time_window": 2, // 单人组成单句时间窗口(分钟),
"qa_match_time_window": 5, // 组成qa时间窗口(分钟),
"prompt_with_history": false // 是否在prompt中包含历史对话
},
"common_args": {
"model_name_or_path": "./chatglm3-6b",
"adapter_name_or_path": "./model_output", //同时做为train_sft_args的output_dir
"template": "chatglm3-weclone",
"finetuning_type": "lora",
"trust_remote_code": true
}
}