## This script is used to finetune the model on the entity extraction task using LoRA
## This script is adapted from the original script in the LIT repository: https://github.com/Lightning-AI/lit-gpt
import os
import sys
import time
from pathlib import Path
from typing import Dict, List, Literal, Optional, Tuple
import lightning as L
import torch
from lightning.fabric.loggers import CSVLogger
from lightning.fabric.plugins import BitsandbytesPrecision
from lightning.fabric.strategies import FSDPStrategy
from lightning.fabric.utilities import ThroughputMonitor
# support running without installing as a package
wd = Path(__file__).parent.parent.resolve()
sys.path.append(str(wd))
from generate.base import generate
from lit_gpt.lora import GPT, Block, Config, lora_filter, mark_only_lora_as_trainable
from lit_gpt.tokenizer import Tokenizer
from lit_gpt.utils import (
check_valid_checkpoint_dir,
chunked_cross_entropy,
get_default_supported_precision,
load_checkpoint,
num_parameters,
)
from scripts.prepare_entity_extraction_data import generate_prompt
eval_interval = 100
save_interval = 100
eval_iters = 100
eval_max_new_tokens = 38
log_interval = 1
devices = 1
# Hyperparameters
learning_rate = 3e-4
batch_size = 16
micro_batch_size = 4
gradient_accumulation_iters = batch_size // micro_batch_size
assert gradient_accumulation_iters > 0
max_seq_length = None # assign value to truncate
max_iters = 700 # train dataset size
weight_decay = 0.01
lora_r = 8
lora_alpha = 16
lora_dropout = 0.05
lora_query = True
lora_key = False
lora_value = True
lora_projection = False
lora_mlp = False
lora_head = False
warmup_steps = 100
hparams = {k: v for k, v in locals().items() if isinstance(v, (int, float, str)) and not k.startswith("_")}
def setup(
data_dir: Path = Path("data/entity_extraction"),
checkpoint_dir: Path = Path("checkpoints/stabilityai/stablelm-base-alpha-3b"),
out_dir: Path = Path("out/lora/Stable-LM/entity_extraction"),
precision: Optional[str] = None,
quantize: Optional[Literal["bnb.nf4", "bnb.nf4-dq", "bnb.fp4", "bnb.fp4-dq", "bnb.int8-training"]] = None,
) -> None:
"""
This script is used to finetune the model on the entity extraction task using LoRA
Args:
data_dir: Path to the data directory containing the train.pt and test.pt files
checkpoint_dir: Path to the directory containing the pre-trained model checkpoint
out_dir: Path to the directory where the finetuned model will be saved
precision: Precision to use for training
quantize: Quantization to use for training
Returns:
None
"""
precision = precision or get_default_supported_precision(training=True)
plugins = None
if quantize is not None and quantize.startswith("bnb."):
if "mixed" in precision:
raise ValueError("Quantization and mixed precision is not supported.")
dtype = {"16-true": torch.float16, "bf16-true": torch.bfloat16, "32-true": torch.float32}[precision]
plugins = BitsandbytesPrecision(quantize[4:], dtype)
precision = None
if devices > 1:
if quantize:
raise NotImplementedError(
"Quantization is currently not supported for multi-GPU training. Please set devices=1 when using the"
" --quantize flag."
)
strategy = FSDPStrategy(
auto_wrap_policy={Block},
activation_checkpointing_policy={Block},
state_dict_type="full",
limit_all_gathers=True,
cpu_offload=False,
)
else:
strategy = "auto"
logger = CSVLogger(out_dir.parent, out_dir.name, flush_logs_every_n_steps=log_interval)
fabric = L.Fabric(devices=devices, strategy=strategy, precision=precision, loggers=logger, plugins=plugins)
fabric.print(hparams)
fabric.launch(main, data_dir, checkpoint_dir, out_dir)
def main(fabric: L.Fabric, data_dir: Path, checkpoint_dir: Path, out_dir: Path) -> None:
"""
Main function for finetuning the model on the entity extraction task using LoRA
Args:
fabric: Lightning Fabric object
data_dir: Path to the data directory containing the train.pt and test.pt files
checkpoint_dir: Path to the directory containing the pre-trained model checkpoint
out_dir: Path to the directory where the finetuned model will be saved
Returns:
None
"""
check_valid_checkpoint_dir(checkpoint_dir)
fabric.seed_everything(1337) # same seed for every process to init model (FSDP)
if fabric.global_rank == 0:
os.makedirs(out_dir, exist_ok=True)
train_data = torch.load(data_dir / "train.pt")
val_data = torch.load(data_dir / "test.pt")
if not any((lora_query, lora_key, lora_value, lora_projection, lora_mlp, lora_head)):
fabric.print("Warning: all LoRA layers are disabled!")
config = Config.from_name(
name=checkpoint_dir.name,
r=lora_r,
alpha=lora_alpha,
dropout=lora_dropout,
to_query=lora_query,
to_key=lora_key,
to_value=lora_value,
to_projection=lora_projection,
to_mlp=lora_mlp,
to_head=lora_head,
)
checkpoint_path = checkpoint_dir / "lit_model.pth"
fabric.print(f"Loading model {str(checkpoint_path)!r} with {config.__dict__}")
with fabric.init_module(empty_init=(devices > 1)):
model = GPT(config)
mark_only_lora_as_trainable(model)
fabric.print(f"Number of trainable parameters: {num_parameters(model, requires_grad=True):,}")
fabric.print(f"Number of non trainable parameters: {num_parameters(model, requires_grad=False):,}")
model = fabric.setup_module(model)
trainable_params = [p for p in model.parameters() if p.requires_grad]
if isinstance(fabric.strategy.precision, BitsandbytesPrecision):
import bitsandbytes as bnb
optimizer = bnb.optim.PagedAdamW(trainable_params, lr=learning_rate, weight_decay=weight_decay)
else:
optimizer = torch.optim.AdamW(trainable_params, lr=learning_rate, weight_decay=weight_decay)
optimizer = fabric.setup_optimizers(optimizer)
scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer, T_max=max_iters // batch_size)
# strict=False because missing keys due to LoRA weights not contained in state dict
load_checkpoint(fabric, model, checkpoint_path, strict=False)
fabric.seed_everything(1337 + fabric.global_rank)
train_time = time.perf_counter()
train(fabric, model, optimizer, scheduler, train_data, val_data, checkpoint_dir, out_dir)
fabric.print(f"Training time: {(time.perf_counter()-train_time):.2f}s")
if fabric.device.type == "cuda":
fabric.print(f"Memory used: {torch.cuda.max_memory_allocated() / 1e9:.02f} GB")
# Save the final LoRA checkpoint at the end of training
save_path = out_dir / "lit_model_lora_finetuned.pth"
save_lora_checkpoint(fabric, model, save_path)
def train(
fabric: L.Fabric,
model: GPT,
optimizer: torch.optim.Optimizer,
scheduler: torch.optim.lr_scheduler,
train_data: List[Dict],
val_data: List[Dict],
checkpoint_dir: Path,
out_dir: Path,
) -> None:
"""
Function for training the model on the entity extraction task using LoRA
Args:
fabric: Lightning Fabric object
model: GPT model
optimizer: Optimizer for training
scheduler: Learning rate scheduler
train_data: Training data
val_data: Validation data
checkpoint_dir: Path to the directory containing the pre-trained model checkpoint
out_dir: Path to the directory where the finetuned model will be saved
Returns:
None
"""
tokenizer = Tokenizer(checkpoint_dir)
longest_seq_length, longest_seq_ix = get_longest_seq_length(train_data)
model.max_seq_length = min(longest_seq_length, max_seq_length or float("inf"))
fabric.print(
f"The longest sequence length in the train data is {longest_seq_length}, the model's maximum sequence length is"
f" {model.max_seq_length} and context length is {model.config.block_size}"
)
validate(fabric, model, val_data, tokenizer, max_iters=2) # sanity check
throughput = ThroughputMonitor(fabric, window_size=50)
step_count = 0
total_lengths = 0
total_t0 = time.perf_counter()
for iter_num in range(1, max_iters + 1):
if step_count <= warmup_steps:
# linear warmup
lr = learning_rate * step_count / warmup_steps
for param_group in optimizer.param_groups:
param_group["lr"] = lr
iter_t0 = time.perf_counter()
input_ids, targets = get_batch(fabric, train_data, longest_seq_ix if iter_num == 1 else None)
is_accumulating = iter_num % gradient_accumulation_iters != 0
with fabric.no_backward_sync(model, enabled=is_accumulating):
logits = model(input_ids, lm_head_chunk_size=128)
# shift the targets such that output n predicts token n+1
logits[-1] = logits[-1][..., :-1, :]
loss = chunked_cross_entropy(logits, targets[..., 1:])
fabric.backward(loss / gradient_accumulation_iters)
if not is_accumulating:
optimizer.step()
optimizer.zero_grad()
if step_count > warmup_steps:
scheduler.step()
step_count += 1
total_lengths += input_ids.numel()
if iter_num % log_interval == 0:
loss_item = loss.item() # expensive device-to-host synchronization
t1 = time.perf_counter()
throughput.update(
time=t1 - total_t0, batches=iter_num, samples=iter_num * micro_batch_size, lengths=total_lengths
)
throughput.compute_and_log(step=iter_num)
fabric.print(
f"iter {iter_num} step {step_count}: loss {loss_item:.4f}, iter time:"
f" {(t1 - iter_t0) * 1000:.2f}ms{' (optimizer.step)' if not is_accumulating else ''}"
)
if not is_accumulating and step_count % eval_interval == 0:
t0 = time.perf_counter()
val_loss = validate(fabric, model, val_data, tokenizer, max_iters=eval_iters)
t1 = time.perf_counter() - t0
fabric.print(f"step {iter_num}: val loss {val_loss.item():.4f}, val time: {t1 * 1000:.2f}ms")
fabric.barrier()
if not is_accumulating and step_count % save_interval == 0:
checkpoint_path = out_dir / f"iter-{iter_num:06d}-ckpt.pth"
save_lora_checkpoint(fabric, model, checkpoint_path)
# FSDP has issues with `inference_mode`
@torch.no_grad()
def validate(fabric: L.Fabric, model: GPT, val_data: List[Dict], tokenizer: Tokenizer, max_iters: int) -> torch.Tensor:
"""
Function for validating the model on the entity extraction task using LoRA
Args:
fabric: Lightning Fabric object
model: GPT model
val_data: Validation data
tokenizer: Tokenizer
max_iters: Maximum number of iterations
Returns:
val_loss: Validation loss
"""
fabric.print("Validating ...")
model.eval()
losses = torch.zeros(max_iters)
for k in range(max_iters):
input_ids, targets = get_batch(fabric, val_data)
logits = model(input_ids)
losses[k] = chunked_cross_entropy(logits[..., :-1, :], targets[..., 1:], chunk_size=0)
val_loss = losses.mean()
# produce an example:
sample = {"input": "Robert Johnson\nrobert.johnson@email.com\n789 Maple Lane, Chicago, IL 60601\n555-234-5678, United States\n\nRelationship to XYZ Pharma Inc.: Patient\nReason for contacting: Adverse Event\n\nMessage: I've been on Onglyza for a while, and I've noticed that I'm experiencing frequent painful urination. Is this a known side effect?"}
prompt = generate_prompt(sample)
encoded = tokenizer.encode(prompt, device=fabric.device)
with fabric.init_tensor():
# do not set `max_seq_length=max_returned_token` because memory is not a concern here
model.set_kv_cache(batch_size=1)
output = generate(model, encoded, max_returned_tokens=len(encoded) + eval_max_new_tokens, temperature=0.8)
model.clear_kv_cache()
output = tokenizer.decode(output)
fabric.print(output)
model.train()
return val_loss
def get_batch(
fabric: L.Fabric, data: List[Dict], longest_seq_ix: Optional[int] = None
) -> Tuple[torch.Tensor, torch.Tensor]:
"""
Function for getting a batch of data
Args:
fabric: Lightning Fabric object
data: Data
longest_seq_ix: Index of the longest sequence
Returns:
x: Input IDs
y: Targets
"""
ix = torch.randint(len(data), (micro_batch_size,))
if longest_seq_ix is not None:
# force the longest sample at the beginning so potential OOMs happen right away
ix[0] = longest_seq_ix
input_ids = [data[i]["input_ids"].type(torch.int64) for i in ix]
labels = [data[i]["labels"].type(torch.int64) for i in ix]
# this could be `longest_seq_length` to have a fixed size for all batches
max_len = max(len(s) for s in input_ids)
def pad_right(x, pad_id):
# pad right based on the longest sequence
n = max_len - len(x)
return torch.cat((x, torch.full((n,), pad_id, dtype=x.dtype)))
x = torch.stack([pad_right(x, pad_id=0) for x in input_ids])
y = torch.stack([pad_right(x, pad_id=-1) for x in labels])
# Truncate if needed
if max_seq_length:
x = x[:, :max_seq_length]
y = y[:, :max_seq_length]
if fabric.device.type == "cuda" and x.device.type == "cpu":
x, y = fabric.to_device((x.pin_memory(), y.pin_memory()))
else:
x, y = fabric.to_device((x, y))
return x, y
def get_longest_seq_length(data: List[Dict]) -> Tuple[int, int]:
"""
Function for getting the longest sequence length
Args:
data: Data
Returns:
longest_seq_length: Longest sequence length
longest_seq_ix: Index of the longest sequence
"""
# find out the minimum max_seq_length required during fine-tuning (saves memory!)
lengths = [len(d["input_ids"]) for d in data]
longest_seq_length = max(lengths)
longest_seq_ix = lengths.index(longest_seq_length)
return longest_seq_length, longest_seq_ix
def save_lora_checkpoint(fabric: L.Fabric, model: torch.nn.Module, file_path: Path) -> None:
"""
Function for saving the LoRA checkpoint
Args:
fabric: Lightning Fabric object
model: GPT model
file_path: Path to the LoRA checkpoint file
Returns:
None
"""
fabric.print(f"Saving LoRA weights to {str(file_path)!r}")
fabric.save(file_path, {"model": model}, filter={"model": lora_filter})
if __name__ == "__main__":
torch.set_float32_matmul_precision("high")
from jsonargparse import CLI
CLI(setup)
