StarCoder私有代码库微调
在本文中,我们将展示如何在私有代码库上微调StarCoder,以增强其上下文感知能力并提高模型对组织需求的实用性。

公开可用的代码 LLM(例如 Codex、StarCoder 和 Code Llama)非常适合生成遵循一般编程原则和语法的代码,但它们可能不符合组织的内部惯例,或者不了解专有库。
在本文中,我们将展示如何在私有代码库上微调StarCoder,以增强其上下文感知能力并提高模型对组织需求的实用性。由于代码 LLM 非常大,以传统方式对其进行微调可能会耗费资源。不用担心!我们将展示如何优化微调以适应单个 GPU。
1、数据集
对于此示例,我们选择了 GitHub 上排名前 10 的 Hugging Face 公共存储库。我们从数据中排除了非代码文件,例如图像、音频文件、演示文稿等。对于 Jupyter 笔记本,我们只保留了包含代码的单元格。生成的代码存储为数据集,你可以在 Hugging Face Hub 的 smangrul/hf-stack-v1 下找到它。它包含 repo id、文件路径和文件内容。
2、模型
我们将微调 bigcode/starcoderbase-1b,这是一个使用 80 多种编程语言训练的 1B 参数模型。这是一个门控模型,因此如果你计划使用此确切模型运行此笔记本,则需要在模型页面上访问它。登录你的 Hugging Face 帐户以执行此操作:
from huggingface_hub import notebook_login
notebook_login()
首先,让我们安装所有必要的库。如你所见,除了转换器和数据集之外,我们还将使用 peft、bitsandbytes 和 flash-attn 来优化训练。
通过采用参数高效的训练技术,我们可以在单个 A100 高 RAM GPU 上运行此笔记本:
!pip install -q transformers datasets peft bitsandbytes flash-attn
现在让我们定义一些变量。请随意使用这些变量。
MODEL = "bigcode/starcoderbase-1b" # Model checkpoint on the Hugging Face Hub
DATASET = "smangrul/hf-stack-v1" # Dataset on the Hugging Face Hub
DATA_COLUMN = "content" # Column name containing the code content
SEQ_LENGTH = 2048 # Sequence length
# Training arguments
MAX_STEPS = 2000 # max_steps
BATCH_SIZE = 16 # batch_size
GR_ACC_STEPS = 1 # gradient_accumulation_steps
LR = 5e-4 # learning_rate
LR_SCHEDULER_TYPE = "cosine" # lr_scheduler_type
WEIGHT_DECAY = 0.01 # weight_decay
NUM_WARMUP_STEPS = 30 # num_warmup_steps
EVAL_FREQ = 100 # eval_freq
SAVE_FREQ = 100 # save_freq
LOG_FREQ = 25 # log_freq
OUTPUT_DIR = "peft-starcoder-lora-a100" # output_dir
BF16 = True # bf16
FP16 = False # no_fp16
# FIM trasformations arguments
FIM_RATE = 0.5 # fim_rate
FIM_SPM_RATE = 0.5 # fim_spm_rate
# LORA
LORA_R = 8 # lora_r
LORA_ALPHA = 32 # lora_alpha
LORA_DROPOUT = 0.0 # lora_dropout
LORA_TARGET_MODULES = "c_proj,c_attn,q_attn,c_fc,c_proj" # lora_target_modules
# bitsandbytes config
USE_NESTED_QUANT = True # use_nested_quant
BNB_4BIT_COMPUTE_DTYPE = "bfloat16" # bnb_4bit_compute_dtype
SEED = 0
3、准备数据
首先加载数据。由于数据集可能非常大,请确保启用流式传输模式。流式传输允许我们在迭代数据集时逐步加载数据,而不是一次下载整个数据集。
我们将保留前 4000 个示例作为验证集,其余所有内容将是训练数据。
from datasets import load_dataset
import torch
from tqdm import tqdm
dataset = load_dataset(
DATASET,
data_dir="data",
split="train",
streaming=True,
)
valid_data = dataset.take(4000)
train_data = dataset.skip(4000)
train_data = train_data.shuffle(buffer_size=5000, seed=SEED)
在此步骤中,数据集仍包含具有任意长度代码的原始数据。对于训练,我们需要固定长度的输入。让我们创建一个可迭代数据集,它将从文本文件流中返回恒定长度的标记块。
首先,让我们估计数据集中每个标记的平均字符数,这将有助于我们稍后估计文本缓冲区中的标记数。默认情况下,我们只从数据集中获取 400 个示例(nb_examples)。仅使用整个数据集的子集将降低计算成本,同时仍提供整体字符与标记比率的合理估计。
tokenizer = AutoTokenizer.from_pretrained(MODEL, trust_remote_code=True)
def chars_token_ratio(dataset, tokenizer, data_column, nb_examples=400):
"""
Estimate the average number of characters per token in the dataset.
"""
total_characters, total_tokens = 0, 0
for _, example in tqdm(zip(range(nb_examples), iter(dataset)), total=nb_examples):
total_characters += len(example[data_column])
total_tokens += len(tokenizer(example[data_column]).tokens())
return total_characters / total_tokens
chars_per_token = chars_token_ratio(train_data, tokenizer, DATA_COLUMN)
print(f"The character to token ratio of the dataset is: {chars_per_token:.2f}")
The character to token ratio of the dataset is: 2.43
字符与标记比率也可以用作文本标记化质量的指标。例如,字符与标记比率为 1.0 意味着每个字符都用一个标记表示,这没有什么意义。这表明标记化效果不佳。在标准英语文本中,一个标记通常相当于大约四个字符,这意味着字符与标记比率约为 4.0。我们可以预期代码数据集中的比率会更低,但一般来说,2.0 到 3.5 之间的数字可以被认为足够好。
4、可选的 FIM 转换
自回归语言模型通常从左到右生成序列。通过应用 FIM 转换,模型还可以学习填充文本。查看“有效训练语言模型以填充中间”论文以了解有关该技术的更多信息。我们将在此处定义 FIM 转换,并在创建可迭代数据集时使用它们。但是,如果你想省略转换,请随意将 fim_rate
设置为 0。
import functools
import numpy as np
# Helper function to get token ids of the special tokens for prefix, suffix and middle for FIM transformations.
@functools.lru_cache(maxsize=None)
def get_fim_token_ids(tokenizer):
try:
FIM_PREFIX, FIM_MIDDLE, FIM_SUFFIX, FIM_PAD = tokenizer.special_tokens_map["additional_special_tokens"][1:5]
suffix_tok_id, prefix_tok_id, middle_tok_id, pad_tok_id = (
tokenizer.vocab[tok] for tok in [FIM_SUFFIX, FIM_PREFIX, FIM_MIDDLE, FIM_PAD]
)
except KeyError:
suffix_tok_id, prefix_tok_id, middle_tok_id, pad_tok_id = None, None, None, None
return suffix_tok_id, prefix_tok_id, middle_tok_id, pad_tok_id
## Adapted from https://github.com/bigcode-project/Megatron-LM/blob/6c4bf908df8fd86b4977f54bf5b8bd4b521003d1/megatron/data/gpt_dataset.py
def permute(
sample,
np_rng,
suffix_tok_id,
prefix_tok_id,
middle_tok_id,
pad_tok_id,
fim_rate=0.5,
fim_spm_rate=0.5,
truncate_or_pad=False,
):
"""
Take in a sample (list of tokens) and perform a FIM transformation on it with a probability of fim_rate, using two FIM modes:
PSM and SPM (with a probability of fim_spm_rate).
"""
# The if condition will trigger with the probability of fim_rate
# This means FIM transformations will apply to samples with a probability of fim_rate
if np_rng.binomial(1, fim_rate):
# Split the sample into prefix, middle, and suffix, based on randomly generated indices stored in the boundaries list.
boundaries = list(np_rng.randint(low=0, high=len(sample) + 1, size=2))
boundaries.sort()
prefix = np.array(sample[: boundaries[0]], dtype=np.int64)
middle = np.array(sample[boundaries[0] : boundaries[1]], dtype=np.int64)
suffix = np.array(sample[boundaries[1] :], dtype=np.int64)
if truncate_or_pad:
# calculate the new total length of the sample, taking into account tokens indicating prefix, middle, and suffix
new_length = suffix.shape[0] + prefix.shape[0] + middle.shape[0] + 3
diff = new_length - len(sample)
# trancate or pad if there's a difference in length between the new length and the original
if diff > 0:
if suffix.shape[0] <= diff:
return sample, np_rng
suffix = suffix[: suffix.shape[0] - diff]
elif diff < 0:
suffix = np.concatenate([suffix, np.full((-1 * diff), pad_tok_id)])
# With the probability of fim_spm_rateapply SPM variant of FIM transformations
# SPM: suffix, prefix, middle
if np_rng.binomial(1, fim_spm_rate):
new_sample = np.concatenate(
[
[prefix_tok_id, suffix_tok_id],
suffix,
[middle_tok_id],
prefix,
middle,
]
)
# Otherwise, apply the PSM variant of FIM transformations
# PSM: prefix, suffix, middle
else:
new_sample = np.concatenate(
[
[prefix_tok_id],
prefix,
[suffix_tok_id],
suffix,
[middle_tok_id],
middle,
]
)
else:
# don't apply FIM transformations
new_sample = sample
return list(new_sample), np_rng
让我们定义 ConstantLengthDataset
,这是一个可迭代数据集,它将返回恒定长度的标记块。为此,我们将从原始数据集中读取文本缓冲区,直到达到大小限制,然后应用标记器将原始文本转换为标记化输入。可选的,我们仅对某些序列执行 FIM 转换(受影响的序列比例由 fim_rate
控制)。
定义后,我们可以从训练和验证数据中创建 ConstantLengthDataset
的实例。
from torch.utils.data import IterableDataset
from torch.utils.data.dataloader import DataLoader
import random
# Create an Iterable dataset that returns constant-length chunks of tokens from a stream of text files.
class ConstantLengthDataset(IterableDataset):
"""
Iterable dataset that returns constant length chunks of tokens from stream of text files.
Args:
tokenizer (Tokenizer): The processor used for proccessing the data.
dataset (dataset.Dataset): Dataset with text files.
infinite (bool): If True the iterator is reset after dataset reaches end else stops.
seq_length (int): Length of token sequences to return.
num_of_sequences (int): Number of token sequences to keep in buffer.
chars_per_token (int): Number of characters per token used to estimate number of tokens in text buffer.
fim_rate (float): Rate (0.0 to 1.0) that sample will be permuted with FIM.
fim_spm_rate (float): Rate (0.0 to 1.0) of FIM permuations that will use SPM.
seed (int): Seed for random number generator.
"""
def __init__(
self,
tokenizer,
dataset,
infinite=False,
seq_length=1024,
num_of_sequences=1024,
chars_per_token=3.6,
content_field="content",
fim_rate=0.5,
fim_spm_rate=0.5,
seed=0,
):
self.tokenizer = tokenizer
self.concat_token_id = tokenizer.eos_token_id
self.dataset = dataset
self.seq_length = seq_length
self.infinite = infinite
self.current_size = 0
self.max_buffer_size = seq_length * chars_per_token * num_of_sequences
self.content_field = content_field
self.fim_rate = fim_rate
self.fim_spm_rate = fim_spm_rate
self.seed = seed
(
self.suffix_tok_id,
self.prefix_tok_id,
self.middle_tok_id,
self.pad_tok_id,
) = get_fim_token_ids(self.tokenizer)
if not self.suffix_tok_id and self.fim_rate > 0:
print("FIM is not supported by tokenizer, disabling FIM")
self.fim_rate = 0
def __iter__(self):
iterator = iter(self.dataset)
more_examples = True
np_rng = np.random.RandomState(seed=self.seed)
while more_examples:
buffer, buffer_len = [], 0
while True:
if buffer_len >= self.max_buffer_size:
break
try:
buffer.append(next(iterator)[self.content_field])
buffer_len += len(buffer[-1])
except StopIteration:
if self.infinite:
iterator = iter(self.dataset)
else:
more_examples = False
break
tokenized_inputs = self.tokenizer(buffer, truncation=False)["input_ids"]
all_token_ids = []
for tokenized_input in tokenized_inputs:
# optionally do FIM permutations
if self.fim_rate > 0:
tokenized_input, np_rng = permute(
tokenized_input,
np_rng,
self.suffix_tok_id,
self.prefix_tok_id,
self.middle_tok_id,
self.pad_tok_id,
fim_rate=self.fim_rate,
fim_spm_rate=self.fim_spm_rate,
truncate_or_pad=False,
)
all_token_ids.extend(tokenized_input + [self.concat_token_id])
examples = []
for i in range(0, len(all_token_ids), self.seq_length):
input_ids = all_token_ids[i : i + self.seq_length]
if len(input_ids) == self.seq_length:
examples.append(input_ids)
random.shuffle(examples)
for example in examples:
self.current_size += 1
yield {
"input_ids": torch.LongTensor(example),
"labels": torch.LongTensor(example),
}
train_dataset = ConstantLengthDataset(
tokenizer,
train_data,
infinite=True,
seq_length=SEQ_LENGTH,
chars_per_token=chars_per_token,
content_field=DATA_COLUMN,
fim_rate=FIM_RATE,
fim_spm_rate=FIM_SPM_RATE,
seed=SEED,
)
eval_dataset = ConstantLengthDataset(
tokenizer,
valid_data,
infinite=False,
seq_length=SEQ_LENGTH,
chars_per_token=chars_per_token,
content_field=DATA_COLUMN,
fim_rate=FIM_RATE,
fim_spm_rate=FIM_SPM_RATE,
seed=SEED,
)
5、准备模型
数据准备好后,就可以加载模型了!我们将加载模型的量化版本。
这将使我们能够减少内存使用量,因为量化表示的数据位数更少。我们将使用 bitsandbytes
库来量化模型,因为它与 transformers
有很好的集成。我们需要做的就是定义一个 bitsandbytes
配置,然后在加载模型时使用它。
4 位量化有不同的变体,但通常,我们建议使用 NF4 量化以获得更好的性能(bnb_4bit_quant_type="nf4")。
bnb_4bit_use_double_quant
选项在第一个量化之后添加了第二个量化,以为每个参数节省额外的 0.4 位。
要了解有关量化的更多信息,请查看“使用 bitsandbytes、4 位量化和 QLoRA 使 LLM 更易于访问”博客文章。
定义完成后,将配置传递给 from_pretrained
方法以加载模型的量化版本。
from peft import LoraConfig, get_peft_model, prepare_model_for_kbit_training
from peft.tuners.lora import LoraLayer
load_in_8bit = False
# 4-bit quantization
compute_dtype = getattr(torch, BNB_4BIT_COMPUTE_DTYPE)
bnb_config = BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_quant_type="nf4",
bnb_4bit_compute_dtype=compute_dtype,
bnb_4bit_use_double_quant=USE_NESTED_QUANT,
)
device_map = {"": 0}
model = AutoModelForCausalLM.from_pretrained(
MODEL,
load_in_8bit=load_in_8bit,
quantization_config=bnb_config,
device_map=device_map,
use_cache=False, # We will be using gradient checkpointing
trust_remote_code=True,
use_flash_attention_2=True,
)
使用量化模型进行训练时,需要调用 prepare_model_for_kbit_training()
函数对量化模型进行预处理以进行训练。
model = prepare_model_for_kbit_training(model)
现在量化模型已准备就绪,我们可以设置 LoRA 配置。LoRA 通过大幅减少可训练参数的数量使微调更加高效。
要使用 LoRA 技术训练模型,我们需要将基础模型包装为 PeftModel。这涉及使用 LoraConfig 定义 LoRA 配置,并使用 LoraConfig 使用 get_peft_model() 包装原始模型。
要了解有关 LoRA 及其参数的更多信息,请参阅 PEFT 文档。
# Set up lora
peft_config = LoraConfig(
lora_alpha=LORA_ALPHA,
lora_dropout=LORA_DROPOUT,
r=LORA_R,
bias="none",
task_type="CAUSAL_LM",
target_modules=LORA_TARGET_MODULES.split(","),
)
model = get_peft_model(model, peft_config)
model.print_trainable_parameters()
trainable params: 5,554,176 || all params: 1,142,761,472 || trainable%: 0.4860310866343243
如你所见,通过应用 LoRA 技术,我们现在需要训练不到 1% 的参数。
6、训练模型
现在我们已经准备好数据并优化了模型,我们准备将所有内容整合在一起开始训练。
要实例化 Trainer
,你需要定义训练配置。最重要的是 TrainingArguments
,它是一个包含配置训练的所有属性的类。
这些与你可能运行的任何其他类型的模型训练类似,因此我们在此不再赘述。
train_data.start_iteration = 0
training_args = TrainingArguments(
output_dir=f"Your_HF_username/{OUTPUT_DIR}",
dataloader_drop_last=True,
evaluation_strategy="steps",
save_strategy="steps",
max_steps=MAX_STEPS,
eval_steps=EVAL_FREQ,
save_steps=SAVE_FREQ,
logging_steps=LOG_FREQ,
per_device_train_batch_size=BATCH_SIZE,
per_device_eval_batch_size=BATCH_SIZE,
learning_rate=LR,
lr_scheduler_type=LR_SCHEDULER_TYPE,
warmup_steps=NUM_WARMUP_STEPS,
gradient_accumulation_steps=GR_ACC_STEPS,
gradient_checkpointing=True,
fp16=FP16,
bf16=BF16,
weight_decay=WEIGHT_DECAY,
push_to_hub=True,
include_tokens_per_second=True,
)
最后一步,实例化 Trainer 并调用 train
方法。
trainer = Trainer(model=model, args=training_args, train_dataset=train_dataset, eval_dataset=eval_dataset)
print("Training...")
trainer.train()
Training...
最后,你可以将微调后的模型推送到 Hub 存储库,以便与你的团队共享。
trainer.push_to_hub()
7、推理
将模型上传到 Hub 后,我们可以使用它进行推理。为此,我们首先初始化原始基础模型及其标记器。接下来,我们需要将微调后的权重与基础模型合并。
from peft import PeftModel
import torch
# load the original model first
tokenizer = AutoTokenizer.from_pretrained(MODEL, trust_remote_code=True)
base_model = AutoModelForCausalLM.from_pretrained(
MODEL,
quantization_config=None,
device_map=None,
trust_remote_code=True,
torch_dtype=torch.bfloat16,
).cuda()
# merge fine-tuned weights with the base model
peft_model_id = f"Your_HF_username/{OUTPUT_DIR}"
model = PeftModel.from_pretrained(base_model, peft_model_id)
model.merge_and_unload()
现在我们可以使用合并后的模型进行推理。为方便起见,我们将定义一个 get_code_completion
- 请随意尝试文本生成参数!
def get_code_completion(prefix, suffix):
text = prompt = f"""<fim_prefix>{prefix}<fim_suffix>{suffix}<fim_middle>"""
model.eval()
outputs = model.generate(
input_ids=tokenizer(text, return_tensors="pt").input_ids.cuda(),
max_new_tokens=128,
temperature=0.2,
top_k=50,
top_p=0.95,
do_sample=True,
repetition_penalty=1.0,
)
return tokenizer.batch_decode(outputs, skip_special_tokens=True)[0]
现在,我们要完成代码补全,只需调用 get_code_complete
函数,并将我们想要补全的前几行作为前缀,将空字符串作为后缀传递即可。
prefix = """from peft import LoraConfig, TaskType, get_peft_model
from transformers import AutoModelForCausalLM
peft_config = LoraConfig(
"""
suffix = """"""
print(get_code_completion(prefix, suffix))
作为刚刚在此笔记本中使用过 PEFT 库的人,你可以看到创建 LoraConfig 生成的结果相当不错!
如果你返回到我们实例化模型进行推理的单元格,并注释掉我们合并微调权重的行,你可以看到原始模型会为完全相同的前缀生成什么:
prefix = """from peft import LoraConfig, TaskType, get_peft_model
from transformers import AutoModelForCausalLM
peft_config = LoraConfig(
"""
suffix = """"""
print(get_code_completion(prefix, suffix))
虽然它是 Python 语法,但你可以看到原始模型不了解 LoraConfig
应该做什么。
原文链接:Fine-tuning a Code LLM on Custom Code on a single GPU
汇智网翻译整理,转载请标明出处