colab使用本地数据集微调llama3-8b模型

        在Google的Colab上面采用unsloth,trl等库，训练数据集来自Google的云端硬盘，微调llama3-8b模型，进行推理验证模型的微调效果。

        保存模型到Google的云端硬盘可以下载到本地供其它使用。

准备工作：将训练数据集上传到google的云端硬盘根目录下，文件名就叫做train.json

train.json里面的数据格式如下：

[
  {
    "instruction": "你好",
    "output": "你好，我是智能助手胖胖"
  },
  {
    "instruction": "hello",
    "output": "Hello! I am 智能助手胖胖, an AI assistant developed by 丹宇码农. How can I assist you ?"
  }

......

]

采用unsloth库、trl库、transformers等库。

直接上代码：

%%capture
# Installs Unsloth, Xformers (Flash Attention) and all other packages!
!pip install "unsloth[colab-new] @ git+https://github.com/unslothai/unsloth.git"
!pip install --no-deps "xformers<0.0.26" trl peft accelerate bitsandbytes

from unsloth import FastLanguageModel
import torch
max_seq_length = 2048 # Choose any! We auto support RoPE Scaling internally!
dtype = None # None for auto detection. Float16 for Tesla T4, V100, Bfloat16 for Ampere+
load_in_4bit = True # Use 4bit quantization to reduce memory usage. Can be False.

# 4bit pre quantized models we support for 4x faster downloading + no OOMs.
fourbit_models = [
    "unsloth/mistral-7b-bnb-4bit",
    "unsloth/mistral-7b-instruct-v0.2-bnb-4bit",
    "unsloth/llama-2-7b-bnb-4bit",
    "unsloth/gemma-7b-bnb-4bit",
    "unsloth/gemma-7b-it-bnb-4bit", # Instruct version of Gemma 7b
    "unsloth/gemma-2b-bnb-4bit",
    "unsloth/gemma-2b-it-bnb-4bit", # Instruct version of Gemma 2b
    "unsloth/llama-3-8b-bnb-4bit", # [NEW] 15 Trillion token Llama-3
] # More models at https://huggingface.co/unsloth

model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "unsloth/llama-3-8b-bnb-4bit",
    max_seq_length = max_seq_length,
    dtype = dtype,
    load_in_4bit = load_in_4bit,
    # token = "hf_...", # use one if using gated models like meta-llama/Llama-2-7b-hf
)

model = FastLanguageModel.get_peft_model(
    model,
    r = 16, # Choose any number > 0 ! Suggested 8, 16, 32, 64, 128
    target_modules = ["q_proj", "k_proj", "v_proj", "o_proj",
                      "gate_proj", "up_proj", "down_proj",],
    lora_alpha = 16,
    lora_dropout = 0, # Supports any, but = 0 is optimized
    bias = "none",    # Supports any, but = "none" is optimized
    # [NEW] "unsloth" uses 30% less VRAM, fits 2x larger batch sizes!
    use_gradient_checkpointing = "unsloth", # True or "unsloth" for very long context
    random_state = 3407,
    use_rslora = False,  # We support rank stabilized LoRA
    loftq_config = None, # And LoftQ
)


alpaca_prompt = """Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
{}

### Input:
{}

### Response:
{}"""

EOS_TOKEN = tokenizer.eos_token # Must add EOS_TOKEN
def formatting_prompts_func(examples):
    instructions = examples["instruction"]
    outputs      = examples["output"]
    texts = []
    for instruction, output in zip(instructions, outputs):
        input = ""
        # Must add EOS_TOKEN, otherwise your generation will go on forever!
        text = alpaca_prompt.format(instruction, input, output) + EOS_TOKEN
        texts.append(text)
    return { "text" : texts, }
pass

from datasets import load_dataset
#dataset = load_dataset("yahma/alpaca-cleaned", split = "train")
#dataset = dataset.map(formatting_prompts_func, batched = True,)
from google.colab import drive
# 挂载云端硬盘，加载成功后，在左边的文件树中将会多一个 /content/drive/MyDrive/ 目录
drive.mount('/content/drive')


# 加载本地数据集：
# 有instruction和output,input为空字符串
from datasets import load_dataset

data_home = r"/content/drive/MyDrive/"
data_dict = {
    "train": os.path.join(data_home, "train.json"),
    #"validation": os.path.join(data_home, "dev.json"),
}
dataset = load_dataset("json", data_files=data_dict, split = "train")
print(dataset[0])
dataset = dataset.map(formatting_prompts_func, batched = True,)


from trl import SFTTrainer
from transformers import TrainingArguments

trainer = SFTTrainer(
    model = model,
    tokenizer = tokenizer,
    train_dataset = dataset,
    dataset_text_field = "text",
    max_seq_length = max_seq_length,
    dataset_num_proc = 2,
    packing = False, # Can make training 5x faster for short sequences.
    args = TrainingArguments(
        per_device_train_batch_size = 2,
        gradient_accumulation_steps = 4,
        warmup_steps = 5,
        max_steps = 60,
        learning_rate = 2e-4,
        fp16 = not torch.cuda.is_bf16_supported(),
        bf16 = torch.cuda.is_bf16_supported(),
        logging_steps = 1,
        optim = "adamw_8bit",
        weight_decay = 0.01,
        lr_scheduler_type = "linear",
        seed = 3407,
        output_dir = "outputs",
    ),
)

# 开始微调训练
trainer_stats = trainer.train()

#推理
# alpaca_prompt = Copied from above
FastLanguageModel.for_inference(model) # Enable native 2x faster inference
inputs = tokenizer(
[
    alpaca_prompt.format(
        "你是谁？", # instruction
        "", # input
        "", # output - leave this blank for generation!
    )
], return_tensors = "pt").to("cuda")

outputs = model.generate(**inputs, max_new_tokens = 64, use_cache = True)
tokenizer.batch_decode(outputs)

#此处输出的答案，能明显看到就是自己训练的数据，而不是原来模型的输出。说明微调起作用了


# 保存模型，改成挂接的云硬盘目录也可以保存到google的个人云存储空间，然后打开个人云存储空间下载到本地
model.save_pretrained("lora_model") # Local saving
tokenizer.save_pretrained("lora_model")

# Merge to 16bit
if True: model.save_pretrained_merged("model", tokenizer, save_method = "merged_16bit",)

其实可以将.ipynb文件上传到个人云存储空间，双击这个文件就会打开colab，然后依次执行代码即可，随时可以增加、删除、修改，特别方便，还能免费使用GPU、CPU等资源，真的是广大AI爱好者的不错选择。