models/LLaMA3/LLaMA3-8B-Instruct Lora.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "de53995b-32ed-4722-8cac-ba104c8efacb",
   "metadata": {},
   "source": [
    "# 导入环境"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "52fac949-4150-4091-b0c3-2968ab5e385c",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "from datasets import Dataset\n",
    "import pandas as pd\n",
    "from transformers import AutoTokenizer, AutoModelForCausalLM, DataCollatorForSeq2Seq, TrainingArguments, Trainer, GenerationConfig"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e098d9eb",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# 将JSON文件转换为CSV文件\n",
    "df = pd.read_json('./huanhuan.json')\n",
    "ds = Dataset.from_pandas(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8ac92d42-efae-49b1-a00e-ccaa75b98938",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'instruction': ['小姐，别的秀女都在求中选，唯有咱们小姐想被撂牌子，菩萨一定记得真真儿的——',\n",
       "  '这个温太医啊，也是古怪，谁不知太医不得皇命不能为皇族以外的人请脉诊病，他倒好，十天半月便往咱们府里跑。',\n",
       "  '嬛妹妹，刚刚我去府上请脉，听甄伯母说你来这里进香了。'],\n",
       " 'input': ['', '', ''],\n",
       " 'output': ['嘘——都说许愿说破是不灵的。', '你们俩话太多了，我该和温太医要一剂药，好好治治你们。', '出来走走，也是散心。']}"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ds[:3]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "51d05e5d-d14e-4f03-92be-9a9677d41918",
   "metadata": {},
   "source": [
    "# 处理数据集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "74ee5a67-2e55-4974-b90e-cbf492de500a",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.\n"
     ]
    }
   ],
   "source": [
    "tokenizer = AutoTokenizer.from_pretrained('/root/autodl-tmp/LLM-Research/Meta-Llama-3-8B-Instruct', use_fast=False, trust_remote_code=True)\n",
    "tokenizer.pad_token = tokenizer.eos_token"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "60590653",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('<|end_of_text|>', 128001, 128001)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tokenizer.pad_token, tokenizer.pad_token_id, tokenizer.eos_token_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2503a5fa-9621-4495-9035-8e7ef6525691",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def process_func(example):\n",
    "    MAX_LENGTH = 384    # Llama分词器会将一个中文字切分为多个token，因此需要放开一些最大长度，保证数据的完整性\n",
    "    input_ids, attention_mask, labels = [], [], []\n",
    "    instruction = tokenizer(f\"<|start_header_id|>user<|end_header_id|>\\n\\n{example['instruction'] + example['input']}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\\n\\n\", add_special_tokens=False)  # add_special_tokens 不在开头加 special_tokens\n",
    "    response = tokenizer(f\"{example['output']}<|eot_id|>\", add_special_tokens=False)\n",
    "    input_ids = instruction[\"input_ids\"] + response[\"input_ids\"] + [tokenizer.pad_token_id]\n",
    "    attention_mask = instruction[\"attention_mask\"] + response[\"attention_mask\"] + [1]  # 因为eos token咱们也是要关注的所以 补充为1\n",
    "    labels = [-100] * len(instruction[\"input_ids\"]) + response[\"input_ids\"] + [tokenizer.pad_token_id]  \n",
    "    if len(input_ids) > MAX_LENGTH:  # 做一个截断\n",
    "        input_ids = input_ids[:MAX_LENGTH]\n",
    "        attention_mask = attention_mask[:MAX_LENGTH]\n",
    "        labels = labels[:MAX_LENGTH]\n",
    "    return {\n",
    "        \"input_ids\": input_ids,\n",
    "        \"attention_mask\": attention_mask,\n",
    "        \"labels\": labels\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "84f870d6-73a9-4b0f-8abf-687b32224ad8",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Map: 100%|██████████| 3729/3729 [00:00<00:00, 4484.79 examples/s]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Dataset({\n",
       "    features: ['input_ids', 'attention_mask', 'labels'],\n",
       "    num_rows: 3729\n",
       "})"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tokenized_id = ds.map(process_func, remove_columns=ds.column_names)\n",
    "tokenized_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "1f7e15a0-4d9a-4935-9861-00cc472654b1",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<|start_header_id|>user<|end_header_id|>\n",
      "\n",
      "小姐，别的秀女都在求中选，唯有咱们小姐想被撂牌子，菩萨一定记得真真儿的——<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n",
      "\n",
      "嘘——都说许愿说破是不灵的。<|eot_id|><|end_of_text|>\n"
     ]
    }
   ],
   "source": [
    "print(tokenizer.decode(tokenized_id[0]['input_ids']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "97f16f66-324a-454f-8cc3-ef23b100ecff",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'你们俩话太多了，我该和温太医要一剂药，好好治治你们。<|eot_id|><|end_of_text|>'"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tokenizer.decode(list(filter(lambda x: x != -100, tokenized_id[1][\"labels\"])))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "424823a8-ed0d-4309-83c8-3f6b1cdf274c",
   "metadata": {},
   "source": [
    "# 创建模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "170764e5-d899-4ef4-8c53-36f6dec0d198",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Loading checkpoint shards: 100%|██████████| 4/4 [00:28<00:00,  7.04s/it]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LlamaForCausalLM(\n",
       "  (model): LlamaModel(\n",
       "    (embed_tokens): Embedding(128256, 4096)\n",
       "    (layers): ModuleList(\n",
       "      (0-31): 32 x LlamaDecoderLayer(\n",
       "        (self_attn): LlamaAttention(\n",
       "          (q_proj): Linear(in_features=4096, out_features=4096, bias=False)\n",
       "          (k_proj): Linear(in_features=4096, out_features=1024, bias=False)\n",
       "          (v_proj): Linear(in_features=4096, out_features=1024, bias=False)\n",
       "          (o_proj): Linear(in_features=4096, out_features=4096, bias=False)\n",
       "          (rotary_emb): LlamaRotaryEmbedding()\n",
       "        )\n",
       "        (mlp): LlamaMLP(\n",
       "          (gate_proj): Linear(in_features=4096, out_features=14336, bias=False)\n",
       "          (up_proj): Linear(in_features=4096, out_features=14336, bias=False)\n",
       "          (down_proj): Linear(in_features=14336, out_features=4096, bias=False)\n",
       "          (act_fn): SiLU()\n",
       "        )\n",
       "        (input_layernorm): LlamaRMSNorm()\n",
       "        (post_attention_layernorm): LlamaRMSNorm()\n",
       "      )\n",
       "    )\n",
       "    (norm): LlamaRMSNorm()\n",
       "  )\n",
       "  (lm_head): Linear(in_features=4096, out_features=128256, bias=False)\n",
       ")"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "\n",
    "model = AutoModelForCausalLM.from_pretrained('/root/autodl-tmp/LLM-Research/Meta-Llama-3-8B-Instruct', device_map=\"auto\",torch_dtype=torch.bfloat16)\n",
    "model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "2323eac7-37d5-4288-8bc5-79fac7113402",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "model.enable_input_require_grads() # 开启梯度检查点时，要执行该方法"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "f808b05c-f2cb-48cf-a80d-0c42be6051c7",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.bfloat16"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.dtype"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "13d71257-3c1c-4303-8ff8-af161ebc2cf1",
   "metadata": {},
   "source": [
    "# lora "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "2d304ae2-ab60-4080-a80d-19cac2e3ade3",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LoraConfig(peft_type=<PeftType.LORA: 'LORA'>, auto_mapping=None, base_model_name_or_path=None, revision=None, task_type=<TaskType.CAUSAL_LM: 'CAUSAL_LM'>, inference_mode=False, r=8, target_modules={'o_proj', 'up_proj', 'k_proj', 'down_proj', 'v_proj', 'gate_proj', 'q_proj'}, lora_alpha=32, lora_dropout=0.1, fan_in_fan_out=False, bias='none', use_rslora=False, modules_to_save=None, init_lora_weights=True, layers_to_transform=None, layers_pattern=None, rank_pattern={}, alpha_pattern={}, megatron_config=None, megatron_core='megatron.core', loftq_config={}, use_dora=False, layer_replication=None)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from peft import LoraConfig, TaskType, get_peft_model\n",
    "\n",
    "config = LoraConfig(\n",
    "    task_type=TaskType.CAUSAL_LM, \n",
    "    target_modules=[\"q_proj\", \"k_proj\", \"v_proj\", \"o_proj\", \"gate_proj\", \"up_proj\", \"down_proj\"],\n",
    "    inference_mode=False, # 训练模式\n",
    "    r=8, # Lora 秩\n",
    "    lora_alpha=32, # Lora alaph，具体作用参见 Lora 原理\n",
    "    lora_dropout=0.1# Dropout 比例\n",
    ")\n",
    "config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "2c2489c5-eaab-4e1f-b06a-c3f914b4bf8e",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LoraConfig(peft_type=<PeftType.LORA: 'LORA'>, auto_mapping=None, base_model_name_or_path='/root/model/LLM-Research/Meta-Llama-3-8B-Instruct', revision=None, task_type=<TaskType.CAUSAL_LM: 'CAUSAL_LM'>, inference_mode=False, r=8, target_modules={'o_proj', 'up_proj', 'k_proj', 'down_proj', 'v_proj', 'gate_proj', 'q_proj'}, lora_alpha=32, lora_dropout=0.1, fan_in_fan_out=False, bias='none', use_rslora=False, modules_to_save=None, init_lora_weights=True, layers_to_transform=None, layers_pattern=None, rank_pattern={}, alpha_pattern={}, megatron_config=None, megatron_core='megatron.core', loftq_config={}, use_dora=False, layer_replication=None)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = get_peft_model(model, config)\n",
    "config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "ebf5482b-fab9-4eb3-ad88-c116def4be12",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "trainable params: 20,971,520 || all params: 8,051,232,768 || trainable%: 0.26047588741133265\n"
     ]
    }
   ],
   "source": [
    "model.print_trainable_parameters()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ca055683-837f-4865-9c57-9164ba60c00f",
   "metadata": {},
   "source": [
    "# 配置训练参数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "7e76bbff-15fd-4995-a61d-8364dc5e9ea0",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "args = TrainingArguments(\n",
    "    output_dir=\"./output/llama3\",\n",
    "    per_device_train_batch_size=4,\n",
    "    gradient_accumulation_steps=4,\n",
    "    logging_steps=10,\n",
    "    num_train_epochs=3,\n",
    "    save_steps=100,\n",
    "    learning_rate=1e-4,\n",
    "    save_on_each_node=True,\n",
    "    gradient_checkpointing=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "f142cb9c-ad99-48e6-ba86-6df198f9ed96",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "trainer = Trainer(\n",
    "    model=model,\n",
    "    args=args,\n",
    "    train_dataset=tokenized_id,\n",
    "    data_collator=DataCollatorForSeq2Seq(tokenizer=tokenizer, padding=True),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "aec9bc36-b297-45af-99e1-d4c4d82be081",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "    <div>\n",
       "      \n",
       "      <progress value='608' max='699' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
       "      [608/699 14:49 < 02:13, 0.68 it/s, Epoch 2.60/3]\n",
       "    </div>\n",
       "    <table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       " <tr style=\"text-align: left;\">\n",
       "      <th>Step</th>\n",
       "      <th>Training Loss</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>4.002100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>3.122300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>30</td>\n",
       "      <td>2.904600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>40</td>\n",
       "      <td>2.787900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50</td>\n",
       "      <td>2.741800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>60</td>\n",
       "      <td>2.749500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>70</td>\n",
       "      <td>2.693400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>80</td>\n",
       "      <td>2.743500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>90</td>\n",
       "      <td>2.741500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>100</td>\n",
       "      <td>2.641100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>110</td>\n",
       "      <td>2.635300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>120</td>\n",
       "      <td>2.743800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>130</td>\n",
       "      <td>2.690700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>140</td>\n",
       "      <td>2.620300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>150</td>\n",
       "      <td>2.709000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>160</td>\n",
       "      <td>2.695500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>170</td>\n",
       "      <td>2.677900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>180</td>\n",
       "      <td>2.523100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>190</td>\n",
       "      <td>2.548900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>200</td>\n",
       "      <td>2.609700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>210</td>\n",
       "      <td>2.544200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>220</td>\n",
       "      <td>2.579900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>230</td>\n",
       "      <td>2.611900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>240</td>\n",
       "      <td>2.456900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>250</td>\n",
       "      <td>2.277900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>260</td>\n",
       "      <td>2.157000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>270</td>\n",
       "      <td>2.175000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>280</td>\n",
       "      <td>2.254900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>290</td>\n",
       "      <td>2.252100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>300</td>\n",
       "      <td>2.210500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>310</td>\n",
       "      <td>2.209400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>320</td>\n",
       "      <td>2.285200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>330</td>\n",
       "      <td>2.252100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>340</td>\n",
       "      <td>2.202300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>350</td>\n",
       "      <td>2.248000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>360</td>\n",
       "      <td>2.253100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>370</td>\n",
       "      <td>2.220800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>380</td>\n",
       "      <td>2.174100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>390</td>\n",
       "      <td>2.249600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>400</td>\n",
       "      <td>2.179000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>410</td>\n",
       "      <td>2.288600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>420</td>\n",
       "      <td>2.138200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>430</td>\n",
       "      <td>2.234700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>440</td>\n",
       "      <td>2.173500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>450</td>\n",
       "      <td>2.164100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>460</td>\n",
       "      <td>2.215300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>470</td>\n",
       "      <td>2.029100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>480</td>\n",
       "      <td>1.837400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>490</td>\n",
       "      <td>1.845700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>500</td>\n",
       "      <td>1.876300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>510</td>\n",
       "      <td>1.787900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>520</td>\n",
       "      <td>1.803700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>530</td>\n",
       "      <td>1.816700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>540</td>\n",
       "      <td>1.922900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>550</td>\n",
       "      <td>1.918100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>560</td>\n",
       "      <td>1.913900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>570</td>\n",
       "      <td>1.872900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>580</td>\n",
       "      <td>1.806500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>590</td>\n",
       "      <td>1.813900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>600</td>\n",
       "      <td>1.796800</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table><p>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/peft/utils/save_and_load.py:154: UserWarning: Could not find a config file in /root/model/LLM-Research/Meta-Llama-3-8B-Instruct - will assume that the vocabulary was not modified.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/peft/utils/save_and_load.py:154: UserWarning: Could not find a config file in /root/model/LLM-Research/Meta-Llama-3-8B-Instruct - will assume that the vocabulary was not modified.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/peft/utils/save_and_load.py:154: UserWarning: Could not find a config file in /root/model/LLM-Research/Meta-Llama-3-8B-Instruct - will assume that the vocabulary was not modified.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/peft/utils/save_and_load.py:154: UserWarning: Could not find a config file in /root/model/LLM-Research/Meta-Llama-3-8B-Instruct - will assume that the vocabulary was not modified.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/peft/utils/save_and_load.py:154: UserWarning: Could not find a config file in /root/model/LLM-Research/Meta-Llama-3-8B-Instruct - will assume that the vocabulary was not modified.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/peft/utils/save_and_load.py:154: UserWarning: Could not find a config file in /root/model/LLM-Research/Meta-Llama-3-8B-Instruct - will assume that the vocabulary was not modified.\n",
      "  warnings.warn(\n",
      "/root/.conda/envs/llama3/lib/python3.10/site-packages/torch/utils/checkpoint.py:429: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "trainer.train()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f93a86ec",
   "metadata": {},
   "source": [
    "# 保存 LoRA 和 tokenizer 结果\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4e376229",
   "metadata": {},
   "outputs": [],
   "source": [
    "peft_model_id=\"./llama3_lora\"\n",
    "trainer.model.save_pretrained(peft_model_id)\n",
    "tokenizer.save_pretrained(peft_model_id)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9823e3c7",
   "metadata": {},
   "source": [
    "# 加载 lora 权重推理"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "12dad881",
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import AutoModelForCausalLM, AutoTokenizer\n",
    "import torch\n",
    "from peft import PeftModel\n",
    "\n",
    "mode_path = '/root/autodl-tmp/LLM-Research/Meta-Llama-3-8B-Instruct'\n",
    "lora_path = './llama3_lora'\n",
    "\n",
    "# 加载tokenizer\n",
    "tokenizer = AutoTokenizer.from_pretrained(mode_path)\n",
    "\n",
    "# 加载模型\n",
    "model = AutoModelForCausalLM.from_pretrained(mode_path, device_map=\"auto\",torch_dtype=torch.bfloat16)\n",
    "\n",
    "# 加载lora权重\n",
    "model = PeftModel.from_pretrained(model, model_id=lora_path, config=config)\n",
    "\n",
    "prompt = \"你是谁？\"\n",
    "messages = [\n",
    "    # {\"role\": \"system\", \"content\": \"现在你要扮演皇帝身边的女人--甄嬛\"},\n",
    "    {\"role\": \"user\", \"content\": prompt}\n",
    "]\n",
    "\n",
    "text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)\n",
    "\n",
    "model_inputs = tokenizer([text], return_tensors=\"pt\").to('cuda')\n",
    "\n",
    "generated_ids = model.generate(\n",
    "    model_inputs.input_ids,\n",
    "    max_new_tokens=512,\n",
    "    do_sample=True,\n",
    "    top_p=0.9, \n",
    "    temperature=0.5, \n",
    "    repetition_penalty=1.1,\n",
    "    eos_token_id=tokenizer.encode('<|eot_id|>')[0],\n",
    ")\n",
    "generated_ids = [\n",
    "    output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)\n",
    "]\n",
    "\n",
    "response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]\n",
    "\n",
    "print(response)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}