#!/usr/bin/env python
# coding=utf-8
import argparse
import logging
import math
import os
import random
import datasets
from datetime import timedelta
import torch
from functools import partial
from accelerate import Accelerator
from accelerate.logging import get_logger
from accelerate.utils import set_seed, InitProcessGroupKwargs
from datasets import load_dataset
from torch.utils.data import DataLoader
from tqdm.auto import tqdm
import deepspeed
import transformers
from transformers import (
AutoConfig,
AutoModelForCausalLM,
AutoTokenizer,
LlamaTokenizer,
LlamaTokenizerFast,
SchedulerType,
DataCollatorForSeq2Seq,
get_scheduler,
GPTNeoXTokenizerFast,
GPT2Tokenizer,
OPTForCausalLM,
BitsAndBytesConfig
)
from peft import LoraConfig, TaskType, get_peft_model, prepare_model_for_kbit_training
logger = get_logger(__name__)
def parse_args():
parser = argparse.ArgumentParser(description="Finetune a transformers model on a causal language modeling task")
parser.add_argument(
"--dataset_name",
type=str,
default=None,
help="The name of the dataset to use (via the datasets library).",
)
parser.add_argument(
"--dataset_config_name",
type=str,
default=None,
help="The configuration name of the dataset to use (via the datasets library).",
)
parser.add_argument(
"--train_file", type=str, default=None, help="A csv or a json file containing the training data."
)
parser.add_argument(
"--model_name_or_path",
type=str,
help="Path to pretrained model or model identifier from huggingface.co/models.",
required=False,
)
parser.add_argument(
"--config_name",
type=str,
default=None,
help="Pretrained config name or path if not the same as model_name",
)
parser.add_argument(
"--model_revision",
help="""If given, specifies a model revision (for HuggingFace models). This will
be applied to both the `model_name_or_path` and `config_name` args.""",
default="main",
required=False,
)
parser.add_argument(
"--use_lora",
action="store_true",
help="If passed, will use LORA (low-rank parameter-efficient training) to train the model.",
)
parser.add_argument(
"--lora_rank",
type=int,
default=64,
help="The rank of lora.",
)
parser.add_argument(
"--lora_alpha",
type=float,
default=16,
help="The alpha parameter of lora.",
)
parser.add_argument(
"--lora_dropout",
type=float,
default=0.1,
help="The dropout rate of lora modules.",
)
parser.add_argument(
"--use_flash_attn",
action="store_true",
help="If passed, will use flash attention to train the model.",
)
parser.add_argument(
"--tokenizer_name",
type=str,
default=None,
help="Pretrained tokenizer name or path if not the same as model_name",
)
parser.add_argument(
"--tokenizer_revision",
help="""Specifies a revision for the tokenizer. If not given, defaults
to the value of the `model_revision` arg. In most cases, the tokenizer
revision should be the same as the model revision and this flag shouldn't
be needed.""",
default=None,
required=False,
)
parser.add_argument(
"--use_slow_tokenizer",
action="store_true",
help="If passed, will use a slow tokenizer (not backed by the 🤗 Tokenizers library).",
)
parser.add_argument(
"--max_seq_length",
type=int,
default=512,
help="The maximum total sequence length (prompt+completion) of each training example.",
)
parser.add_argument(
"--per_device_train_batch_size",
type=int,
default=8,
help="Batch size (per device) for the training dataloader.",
)
parser.add_argument(
"--learning_rate",
type=float,
default=5e-5,
help="Initial learning rate (after the potential warmup period) to use.",
)
parser.add_argument("--weight_decay", type=float, default=0.0, help="Weight decay to use.")
parser.add_argument("--num_train_epochs", type=int, default=3, help="Total number of training epochs to perform.")
parser.add_argument(
"--max_train_steps",
type=int,
default=None,
help="Total number of training steps to perform. If provided, overrides num_train_epochs.",
)
parser.add_argument(
"--gradient_accumulation_steps",
type=int,
default=1,
help="Number of updates steps to accumulate before performing a backward/update pass.",
)
parser.add_argument(
"--lr_scheduler_type",
type=SchedulerType,
default="linear",
help="The scheduler type to use.",
choices=["linear", "cosine", "cosine_with_restarts", "polynomial", "constant", "constant_with_warmup"],
)
parser.add_argument(
"--warmup_ratio", type=float, default=0, help="Ratio of total training steps used for warmup."
)
parser.add_argument("--output_dir", type=str, default=None, help="Where to store the final model.")
parser.add_argument("--seed", type=int, default=42, help="A seed for reproducible training.")
parser.add_argument(
"--preprocessing_num_workers",
type=int,
default=None,
help="The number of processes to use for the preprocessing.",
)
parser.add_argument(
"--overwrite_cache", action="store_true", help="Overwrite the cached training and evaluation sets"
)
parser.add_argument(
"--checkpointing_steps",
type=str,
default=None,
help="Whether the various states should be saved at the end of every n steps, or 'epoch' for each epoch.",
)
parser.add_argument(
"--logging_steps",
type=int,
default=None,
help="Log the training loss and learning rate every logging_steps steps.",
)
parser.add_argument(
"--resume_from_checkpoint",
type=str,
default=None,
help="If the training should continue from a checkpoint folder.",
)
parser.add_argument(
"--with_tracking",
action="store_true",
help="Whether to enable experiment trackers for logging.",
)
parser.add_argument(
"--report_to",
type=str,
default="all",
help=(
'The integration to report the results and logs to. Supported platforms are `"tensorboard"`,'
' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.'
"Only applicable when `--with_tracking` is passed."
),
)
parser.add_argument(
"--low_cpu_mem_usage",
action="store_true",
help=(
"It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded."
"If passed, LLM loading time and RAM consumption will be benefited."
),
)
parser.add_argument(
"--gradient_checkpointing",
action="store_true",
help=(
"Turn on gradient checkpointing. Saves memory but slows training."
),
)
parser.add_argument(
"--use_qlora",
action="store_true",
help=(
"Use qLoRA training - main thing is initialising model in quantised form. Not compatible with deepspeed."
),
)
parser.add_argument(
'--clip_grad_norm',
type=float,
default=-1,
help='Clip gradient norm. Not compatible with deepspeed (use deepspeed config instead).',
)
parser.add_argument(
'--use_8bit_optimizer',
action='store_true',
help='Use 8bit optimizer from bitsandbytes. Not compatible with deepspeed (use deepspeed config instead).',
)
parser.add_argument(
'--add_bos',
action='store_true',
help='Forcibly add bos token to the beginning of the input sequence. Use only when tokenizer does not add bos token by default (e.g., olmo).',
)
parser.add_argument(
'--timeout',
type=int,
default=1800,
help='Timeout for the training process. Useful if tokenization process is long. Default is 1800 seconds (30 minutes).',
)
parser.add_argument(
'--trust_remote_code',
action='store_true',
help='Trust remote code when loading pretrained models and tokenizers. Use only when you trust the remote code.',
)
parser.add_argument(
'--reduce_loss',
default='mean',
choices=['mean', 'sum'],
help='How to reduce loss over tokens. Default is mean, but using sum can improve chat model performance.',
)
parser.add_argument(
'--wandb_entity',
type=str,
default=None,
help='Entity to use for logging to wandb.'
)
args = parser.parse_args()
# Sanity checks
if args.dataset_name is None and args.train_file is None:
raise ValueError("Need either a dataset name or a training file.")
else:
if args.train_file is not None:
extension = args.train_file.split(".")[-1]
assert extension in ["json", "jsonl"], "`train_file` should be a json/jsonl file."
return args
def encode_with_prompt_completion_format(example, tokenizer, max_seq_length, add_bos=False):
'''
Here we assume each example has 'prompt' and 'completion' fields.
We concatenate prompt and completion and tokenize them together because otherwise prompt will be padded/trancated
and it doesn't make sense to follow directly with the completion.
'''
# if prompt doesn't end with space and completion doesn't start with space, add space
if not example['prompt'].endswith((' ', '\n', '\t')) and not example['completion'].startswith((' ', '\n', '\t')):
example_text = example['prompt'] + ' ' + example['completion']
else:
example_text = example['prompt'] + example['completion']
example_text = example_text + tokenizer.eos_token
if add_bos:
example_text = tokenizer.bos_token + example_text
tokenized_example = tokenizer(example_text, return_tensors='pt', max_length=max_seq_length, truncation=True)
input_ids = tokenized_example.input_ids
labels = input_ids.clone()
tokenized_prompt = tokenizer(example['prompt'], return_tensors='pt', max_length=max_seq_length, truncation=True)
# mask the prompt part for avoiding loss
labels[:, :tokenized_prompt.input_ids.shape[1]] = -100
attention_mask = torch.ones_like(input_ids)
return {
'input_ids': input_ids.flatten(),
'labels': labels.flatten(),
'attention_mask': attention_mask.flatten(),
}
def encode_with_messages_format(example, tokenizer, max_seq_length, add_bos=False):
'''
Here we assume each example has a 'messages' field Each message is a dict with 'role' and 'content' fields.
We concatenate all messages with the roles as delimiters and tokenize them together.
'''
messages = example['messages']
if len(messages) == 0:
raise ValueError('messages field is empty.')
def _concat_messages(messages):
message_text = ""
for message in messages:
if message["role"] == "system":
message_text += "<|system|>\n" + message["content"].strip() + "\n"
elif message["role"] == "user":
message_text += "<|user|>\n" + message["content"].strip() + "\n"
elif message["role"] == "assistant":
message_text += "<|assistant|>\n" + message["content"].strip() + tokenizer.eos_token + "\n"
else:
raise ValueError("Invalid role: {}".format(message["role"]))
return message_text
example_text = _concat_messages(messages).strip()
if add_bos:
example_text = tokenizer.bos_token + example_text
tokenized_example = tokenizer(example_text, return_tensors='pt', max_length=max_seq_length, truncation=True)
input_ids = tokenized_example.input_ids
labels = input_ids.clone()
# mask the non-assistant part for avoiding loss
for message_idx, message in enumerate(messages):
if message["role"] != "assistant":
if message_idx == 0:
message_start_idx = 0
else:
message_start_idx = tokenizer(
_concat_messages(messages[:message_idx]), return_tensors='pt', max_length=max_seq_length, truncation=True
).input_ids.shape[1]
if message_idx < len(messages) - 1 and messages[message_idx+1]["role"] == "assistant":
# here we also ignore the role of the assistant
messages_so_far = _concat_messages(messages[:message_idx+1]) + "<|assistant|>\n"
else:
messages_so_far = _concat_messages(messages[:message_idx+1])
message_end_idx = tokenizer(
messages_so_far,
return_tensors='pt',
max_length=max_seq_length,
truncation=True
).input_ids.shape[1]
labels[:, message_start_idx:message_end_idx] = -100
if message_end_idx >= max_seq_length:
break
attention_mask = torch.ones_like(input_ids)
return {
'input_ids': input_ids.flatten(),
'labels': labels.flatten(),
'attention_mask': attention_mask.flatten(),
}
def save_with_accelerate(accelerator, model, tokenizer, output_dir, args):
# set the generation config to an empty setting to be safe.
# we usually do greedy decoding for generation, so this should be okay.
# otherwise, we get an error thrown at save time.
model.generation_config = transformers.GenerationConfig(
temperature=None,
top_p=None,
eos_token_id=tokenizer.eos_token_id,
bos_token_id=tokenizer.bos_token_id
)
unwrapped_model = accelerator.unwrap_model(model)
# When doing multi-gpu training, we need to use accelerator.get_state_dict(model) to get the state_dict.
# Otherwise, sometimes the model will be saved with only part of the parameters.
# Also, accelerator needs to use the wrapped model to get the state_dict.
state_dict = accelerator.get_state_dict(model)
if args.use_lora:
# When using lora, the unwrapped model is a PeftModel, which doesn't support the is_main_process
# and has its own save_pretrained function for only saving lora modules.
# We have to manually specify the is_main_process outside the save_pretrained function.
if accelerator.is_main_process:
unwrapped_model.save_pretrained(output_dir, state_dict=state_dict)
else:
# don't use safetensors for saving for now
unwrapped_model.save_pretrained(
output_dir, is_main_process=accelerator.is_main_process, save_function=accelerator.save, state_dict=state_dict,
safe_serialization=False
)
def main():
args = parse_args()
# Initialize the accelerator. We will let the accelerator handle device placement for us in this example.
# If we're using tracking, we also need to initialize it here and it will by default pick up all supported trackers
# in the environment
accelerator_log_kwargs = {}
if args.with_tracking:
accelerator_log_kwargs["log_with"] = args.report_to
accelerator_log_kwargs["project_dir"] = args.output_dir
# if you get timeouts (e.g. due to long tokenization) increase this.
timeout_kwargs = InitProcessGroupKwargs(timeout=timedelta(seconds=args.timeout))
accelerator = Accelerator(
gradient_accumulation_steps=args.gradient_accumulation_steps,
**accelerator_log_kwargs,
kwargs_handlers=[timeout_kwargs]
)
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
datefmt="%m/%d/%Y %H:%M:%S",
level=logging.INFO,
)
logger.info(accelerator.state, main_process_only=False)
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_info()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
# If passed along, set the training seed now.
if args.seed is not None:
set_seed(args.seed)
if accelerator.is_main_process:
if args.output_dir is not None:
os.makedirs(args.output_dir, exist_ok=True)
accelerator.wait_for_everyone()
if args.dataset_name is not None:
# Downloading and loading a dataset from the hub.
raw_datasets = load_dataset(
args.dataset_name,
args.dataset_config_name,
)
else:
data_files = {}
dataset_args = {}
if args.train_file is not None:
data_files["train"] = args.train_file
raw_datasets = load_dataset(
"json",
data_files=data_files,
**dataset_args,
)
# Load pretrained model and tokenizer
if args.config_name:
config = AutoConfig.from_pretrained(
args.config_name,
trust_remote_code=args.trust_remote_code,
revision=args.model_revision,
token=os.getenv("HF_TOKEN", None)
)
elif args.model_name_or_path:
config = AutoConfig.from_pretrained(
args.model_name_or_path,
trust_remote_code=args.trust_remote_code,
revision=args.model_revision,
token=os.getenv("HF_TOKEN", None)
)
else:
raise ValueError(
"You are instantiating a new config instance from scratch. This is not supported by this script."
)
tokenizer_revision = (
args.model_revision
if args.tokenizer_revision is None
else args.tokenizer_revision
)
if tokenizer_revision != args.model_revision:
# Warn user if tokenizer and model use different revisions; this is an unusual
# use case.
warning = f"""Requested tokenizer revision `{tokenizer_revision}` is different
from the model revision `{args.model_revision}`."""
logger.warn(warning)
if args.tokenizer_name:
tokenizer = AutoTokenizer.from_pretrained(
args.tokenizer_name,
trust_remote_code=args.trust_remote_code,
use_fast=not args.use_slow_tokenizer,
revision=tokenizer_revision,
token=os.getenv("HF_TOKEN", None)
)
elif args.model_name_or_path:
tokenizer = AutoTokenizer.from_pretrained(
args.model_name_or_path,
trust_remote_code=args.trust_remote_code,
use_fast=not args.use_slow_tokenizer,
revision=tokenizer_revision,
token=os.getenv("HF_TOKEN", None)
)
else:
raise ValueError(
"You are instantiating a new tokenizer from scratch. This is not supported by this script."
"You can do it from another script, save it, and load it from here, using --tokenizer_name."
)
if args.model_name_or_path:
if args.use_qlora:
bnb_config = BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_use_double_quant=True,
bnb_4bit_quant_type="nf4",
bnb_4bit_compute_dtype=torch.bfloat16,
)
device_index = accelerator.local_process_index
device_map = {"": device_index} # force data-parallel training.
model = AutoModelForCausalLM.from_pretrained(
args.model_name_or_path,
from_tf=bool(".ckpt" in args.model_name_or_path),
config=config,
load_in_4bit=True,
quantization_config=bnb_config,
device_map=device_map,
trust_remote_code=args.trust_remote_code,
torch_dtype=torch.bfloat16,
use_flash_attention_2=True if args.use_flash_attn else False,
revision=args.model_revision,
token=os.getenv("HF_TOKEN", None)
)
else:
model = AutoModelForCausalLM.from_pretrained(
args.model_name_or_path,
from_tf=bool(".ckpt" in args.model_name_or_path),
config=config,
trust_remote_code=args.trust_remote_code,
low_cpu_mem_usage=args.low_cpu_mem_usage,
use_flash_attention_2=True if args.use_flash_attn else False,
revision=args.model_revision,
token=os.getenv("HF_TOKEN", None)
)
else:
logger.info("Training new model from scratch")
model = AutoModelForCausalLM.from_config(config)
# no default pad token for llama!
# here we add all special tokens again, because the default ones are not in the special_tokens_map
if isinstance(tokenizer, LlamaTokenizer) or isinstance(tokenizer, LlamaTokenizerFast):
num_added_tokens = tokenizer.add_special_tokens({
"bos_token": "",
"eos_token": "",
"unk_token": "",
"pad_token": "",
})
assert num_added_tokens in [0, 1], "LlamaTokenizer should only add one special token - the pad_token, or no tokens if pad token present."
elif isinstance(tokenizer, GPTNeoXTokenizerFast):
# OLMo newer models use this tokenizer
if tokenizer.bos_token is None:
tokenizer.bos_token = tokenizer.eos_token
assert args.add_bos, "For OLMo with GPTNeoX, you must add bos token to the beginning of the input sequence."
# else, pythia / other models
else:
num_added_tokens = tokenizer.add_special_tokens({
"pad_token": "",
})
assert num_added_tokens == 1, "GPTNeoXTokenizer should only add one special token - the pad_token."
elif isinstance(tokenizer, GPT2Tokenizer) and isinstance(model, OPTForCausalLM):
num_added_tokens = tokenizer.add_special_tokens({'unk_token': ''})
elif isinstance(tokenizer, transformers.PreTrainedTokenizerFast) and tokenizer.pad_token is None:
num_added_tokens = tokenizer.add_special_tokens({'pad_token': ''})
assert num_added_tokens == 1, "We detected no padding token but add_special_tokens did not add one."
# We resize the embeddings only when necessary to avoid index errors. If you are creating a model from scratch
# gather deepspeed to get "real" embedding size
embeddings = model.get_input_embeddings()
with deepspeed.zero.GatheredParameters(embeddings.weight, modifier_rank=None):
embedding_size = embeddings.weight.shape[0]
# resize does its own gather
if len(tokenizer) > embedding_size:
# pad to multiple for tensor cores.
model.resize_token_embeddings(len(tokenizer), pad_to_multiple_of=8)
# update embedding size after resizing for sum loss
embeddings = model.get_input_embeddings()
with deepspeed.zero.GatheredParameters(embeddings.weight, modifier_rank=None):
embedding_size = embeddings.weight.shape[0]
# set the tokenizer chat template to the tulu format
# this makes evaluation/etc easier down the line.
tokenizer.chat_template = "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n' + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}" # noqa: E501
if args.add_bos:
# also add bos in the chat template
tokenizer.chat_template = "{{ bos_token }}" + tokenizer.chat_template
if args.use_lora:
if args.use_qlora:
model = prepare_model_for_kbit_training(model, use_gradient_checkpointing=args.gradient_checkpointing)
logger.info("Initializing LORA model...")
peft_config = LoraConfig(
task_type=TaskType.CAUSAL_LM,
inference_mode=False,
r=args.lora_rank,
lora_alpha=args.lora_alpha,
lora_dropout=args.lora_dropout,
target_modules=["q_proj", "o_proj", "v_proj", "k_proj", "gate_proj", "up_proj", "down_proj"]
)
model = get_peft_model(model, peft_config)
model.print_trainable_parameters()
elif args.gradient_checkpointing:
model.gradient_checkpointing_enable()
# Preprocessing the datasets.
if "prompt" in raw_datasets["train"].column_names and "completion" in raw_datasets["train"].column_names:
encode_function = partial(
encode_with_prompt_completion_format,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
add_bos=args.add_bos,
)
elif "messages" in raw_datasets["train"].column_names:
encode_function = partial(
encode_with_messages_format,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length,
add_bos=args.add_bos,
)
else:
raise ValueError("You need to have either 'prompt'&'completion' or 'messages' in your column names.")
with accelerator.main_process_first():
lm_datasets = raw_datasets.map(
encode_function,
batched=False,
num_proc=args.preprocessing_num_workers,
load_from_cache_file=not args.overwrite_cache,
remove_columns=[name for name in raw_datasets["train"].column_names if name not in ["input_ids", "labels", "attention_mask"]],
desc="Tokenizing and reformatting instruction data",
)
lm_datasets.set_format(type="pt")
lm_datasets = lm_datasets.filter(lambda example: (example['labels'] != -100).any())
train_dataset = lm_datasets["train"]
# Log a few random samples from the training set:
for index in random.sample(range(len(train_dataset)), 3):
logger.info(f"Sample {index} of the training set: {train_dataset[index]}.")
# DataLoaders creation:
train_dataloader = DataLoader(
train_dataset,
shuffle=True,
collate_fn=DataCollatorForSeq2Seq(tokenizer=tokenizer, model=model, padding="longest"),
batch_size=args.per_device_train_batch_size
)
# Optimizer
# Split weights in two groups, one with weight decay and the other not.
no_decay = ["bias", "layer_norm.weight"]
optimizer_grouped_parameters = [
{
"params": [p for n, p in model.named_parameters() if not any(nd in n for nd in no_decay)],
"weight_decay": args.weight_decay,
},
{
"params": [p for n, p in model.named_parameters() if any(nd in n for nd in no_decay)],
"weight_decay": 0.0,
},
]
if args.use_qlora:
from bitsandbytes.optim import AdamW
optimizer = AdamW(
optimizer_grouped_parameters,
lr=args.learning_rate,
optim_bits=8 if args.use_8bit_optimizer else 32,
is_paged=True
)
else:
optimizer = torch.optim.AdamW(optimizer_grouped_parameters, lr=args.learning_rate)
# Scheduler and math around the number of training steps.
overrode_max_train_steps = False
num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
if args.max_train_steps is None:
args.max_train_steps = args.num_train_epochs * num_update_steps_per_epoch
overrode_max_train_steps = True
# Create the learning rate scheduler.
# Note: the current accelerator.step() calls the .step() of the real scheduler for the `num_processes` times. This is because they assume
# the user initialize the scheduler with the entire training set. In the case of data parallel training, each process only
# sees a subset (1/num_processes) of the training set. So each time the process needs to update the lr multiple times so that the total
# number of updates in the end matches the num_training_steps here.
# Here we need to set the num_training_steps to either using the entire training set (when epochs is specified) or we need to multiply the
# num_training_steps by num_processes so that the total number of updates matches the num_training_steps.
num_training_steps_for_scheduler = args.max_train_steps if overrode_max_train_steps else args.max_train_steps * accelerator.num_processes
lr_scheduler = get_scheduler(
name=args.lr_scheduler_type,
optimizer=optimizer,
num_training_steps=num_training_steps_for_scheduler,
num_warmup_steps=int(num_training_steps_for_scheduler * args.warmup_ratio),
)
# Prepare everything with `accelerator`.
model, optimizer, train_dataloader, lr_scheduler = accelerator.prepare(
model, optimizer, train_dataloader, lr_scheduler
)
# We need to recalculate our total training steps as the size of the training dataloader may have changed.
num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
if overrode_max_train_steps:
args.max_train_steps = args.num_train_epochs * num_update_steps_per_epoch
# Afterwards we recalculate our number of training epochs
args.num_train_epochs = math.ceil(args.max_train_steps / num_update_steps_per_epoch)
# Figure out how many steps we should save the Accelerator states
checkpointing_steps = args.checkpointing_steps
if checkpointing_steps is not None and checkpointing_steps.isdigit():
checkpointing_steps = int(checkpointing_steps)
# We need to initialize the trackers we use, and also store our configuration.
# The trackers initializes automatically on the main process.
if args.with_tracking:
experiment_config = vars(args)
# TensorBoard cannot log Enums, need the raw value
experiment_config["lr_scheduler_type"] = experiment_config["lr_scheduler_type"].value
accelerator.init_trackers("open_instruct_sft",
experiment_config,
init_kwargs={"wandb": {"entity": args.wandb_entity}})
# Train!
total_batch_size = args.per_device_train_batch_size * accelerator.num_processes * args.gradient_accumulation_steps
logger.info("***** Running training *****")
logger.info(f" Num examples = {len(train_dataset)}")
logger.info(f" Num Epochs = {args.num_train_epochs}")
logger.info(f" Instantaneous batch size per device = {args.per_device_train_batch_size}")
logger.info(f" Total train batch size (w. parallel, distributed & accumulation) = {total_batch_size}")
logger.info(f" Gradient Accumulation steps = {args.gradient_accumulation_steps}")
logger.info(f" Total optimization steps = {args.max_train_steps}")
# Only show the progress bar once on each machine.
progress_bar = tqdm(range(args.max_train_steps), disable=not accelerator.is_local_main_process)
completed_steps = 0
starting_epoch = 0
# Potentially load in the weights and states from a previous save
if args.resume_from_checkpoint:
if args.resume_from_checkpoint is not None or args.resume_from_checkpoint != "":
checkpoint_path = args.resume_from_checkpoint
path = os.path.basename(args.resume_from_checkpoint)
else:
# Get the most recent checkpoint
dirs = [f.name for f in os.scandir(os.getcwd()) if f.is_dir()]
dirs.sort(key=os.path.getctime)
path = dirs[
-1
] # Sorts folders by date modified, most recent checkpoint is the last
checkpoint_path = path
path = os.path.basename(checkpoint_path)
accelerator.print(f"Resumed from checkpoint: {checkpoint_path}")
accelerator.load_state(path)
# Extract `epoch_{i}` or `step_{i}`
training_difference = os.path.splitext(path)[0]
if "epoch" in training_difference:
starting_epoch = int(training_difference.replace("epoch_", "")) + 1
resume_step = None
completed_steps = starting_epoch * num_update_steps_per_epoch
else:
# need to multiply `gradient_accumulation_steps` to reflect real steps
resume_step = (
int(training_difference.replace("step_", ""))
* args.gradient_accumulation_steps
)
starting_epoch = resume_step // len(train_dataloader)
completed_steps = resume_step // args.gradient_accumulation_steps
resume_step -= starting_epoch * len(train_dataloader)
# update the progress_bar if load from checkpoint
progress_bar.update(completed_steps)
for epoch in range(starting_epoch, args.num_train_epochs):
model.train()
total_loss = 0
if (
args.resume_from_checkpoint
and epoch == starting_epoch
and resume_step is not None
):
# We skip the first `n` batches in the dataloader when resuming from a checkpoint
active_dataloader = accelerator.skip_first_batches(
train_dataloader, resume_step
)
else:
active_dataloader = train_dataloader
for step, batch in enumerate(active_dataloader):
with accelerator.accumulate(model):
outputs = model(**batch, use_cache=False)
if args.reduce_loss == 'mean':
loss = outputs.loss
else:
# reduce loss is sum
# this ensures that we weight all tokens in the dataset equally,
# rather than weighting each overall example equally when
# using high amounts of gradient accumulation.
# this can result in > 5 point improvements in AlpacaEval
# see https://github.com/huggingface/transformers/issues/24725 for
# more discussion and details.
logits = outputs.logits
labels = batch["labels"]
# Shift so that tokens < n predict n
shift_logits = logits[..., :-1, :].contiguous()
shift_labels = labels[..., 1:].contiguous()
# Flatten the tokens
loss_fct = torch.nn.CrossEntropyLoss(reduction='sum')
shift_logits = shift_logits.view(-1, embedding_size)
shift_labels = shift_labels.view(-1)
# Enable model parallelism
shift_labels = shift_labels.to(shift_logits.device)
loss = loss_fct(shift_logits, shift_labels)
# We keep track of the loss at each logged step
total_loss += loss.detach().float()
accelerator.backward(loss)
# clip gradient norm. don't do this with deepspeed
if accelerator.sync_gradients and args.clip_grad_norm > 0:
accelerator.clip_grad_norm_(model.parameters(), args.clip_grad_norm)
optimizer.step()
optimizer.zero_grad()
lr_scheduler.step()
# Checks if the accelerator has performed an optimization step behind the scenes
if accelerator.sync_gradients:
progress_bar.update(1)
completed_steps += 1
if args.logging_steps and completed_steps % args.logging_steps == 0:
avg_loss = accelerator.gather(total_loss).mean().item() / args.gradient_accumulation_steps / args.logging_steps
logger.info(f" Step: {completed_steps}, LR: {lr_scheduler.get_last_lr()[0]}, Loss: {avg_loss}")
if args.with_tracking:
accelerator.log(
{
"learning_rate": lr_scheduler.get_last_lr()[0],
"train_loss": avg_loss,
},
step=completed_steps,
)
total_loss = 0
if isinstance(checkpointing_steps, int):
if completed_steps % checkpointing_steps == 0:
output_dir = f"step_{completed_steps}"
if args.output_dir is not None:
output_dir = os.path.join(args.output_dir, output_dir)
save_with_accelerate(accelerator, model, tokenizer, output_dir, args)
if completed_steps >= args.max_train_steps:
break
if args.checkpointing_steps == "epoch":
output_dir = f"epoch_{epoch}"
if args.output_dir is not None:
output_dir = os.path.join(args.output_dir, output_dir)
save_with_accelerate(accelerator, model, tokenizer, output_dir, args)
if args.output_dir is not None:
if accelerator.is_main_process:
tokenizer.save_pretrained(args.output_dir)
save_with_accelerate(accelerator, model, tokenizer, args.output_dir, args)
accelerator.wait_for_everyone()
if args.with_tracking:
accelerator.end_training()
if __name__ == "__main__":
main()