ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

support FlashAttention2

Browse Source 
Former-commit-id: 23e56c5554b948d4f08ad87849b261eafd2c7890
This commit is contained in:
hiyouga
2023-09-10 20:43:56 +08:00
parent b481ad58e6
commit a402161631
9 changed files with 875 additions and 115 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
src/llmtuner/tuner/core/loader.py
View File

@@ -4,6 +4,7 @@ import torch
from types import MethodType
from typing import TYPE_CHECKING, Literal, Optional, Tuple
import transformers
from transformers import (
AutoConfig,
AutoModelForCausalLM,
@@ -84,7 +85,8 @@ def load_model_and_tokenizer(
config = AutoConfig.from_pretrained(model_to_load, **config_kwargs)
if is_trainable and hasattr(config, "fp16") and hasattr(config, "bf16"): # fix Qwen config
# Fix config (for Qwen)
if is_trainable and hasattr(config, "fp16") and hasattr(config, "bf16"):
if model_args.compute_dtype == torch.bfloat16:
setattr(config, "bf16", True)
else:
@@ -105,6 +107,7 @@ def load_model_and_tokenizer(
if is_trainable:
if model_args.rope_scaling == "dynamic":
assert not model_args.flash_attn, "Flash attention does not support dynamic rope scaling."
logger.warning(
"Dynamic NTK may not work well with fine-tuning. "
"See: https://github.com/huggingface/transformers/pull/24653"
@@ -127,6 +130,15 @@ def load_model_and_tokenizer(
else:
logger.warning("Current model does not support RoPE scaling.")
# Set flash attention
if model_args.flash_attn and getattr(config, "model_type", None) == "llama":
from llmtuner.extras.models.flash_llama import LlamaForCausalLM
transformers.models.llama.modeling_llama.LlamaForCausalLM = LlamaForCausalLM
if not hasattr(config, "num_key_value_heads"):
setattr(config, "num_key_value_heads", getattr(config, "num_attention_heads"))
if getattr(config, "pretraining_tp", 1) != 1:
setattr(config, "pretraining_tp", 1)
# Quantization configurations (using bitsandbytes library).
is_mergeable = True
if model_args.quantization_bit is not None:

45
src/llmtuner/tuner/sft/trainer.py
View File

@@ -33,27 +33,28 @@ class Seq2SeqPeftTrainer(PeftTrainer):
Subclass and override to inject custom behavior.
"""
prompt_len, label_len = inputs["input_ids"].size(-1), inputs["labels"].size(-1)
if prompt_len > label_len:
inputs["labels"] = self._pad_tensors_to_target_len(inputs["labels"], inputs["input_ids"])
if label_len > prompt_len:
inputs["input_ids"] = self._pad_tensors_to_target_len(inputs["input_ids"], inputs["labels"])
if "attention_mask" in inputs:
inputs["attention_mask"] = self._pad_tensors_to_target_len(
inputs["attention_mask"], inputs["labels"], pad_token_id=0
)
if "position_ids" in inputs:
inputs["position_ids"] = self._pad_tensors_to_target_len(
inputs["position_ids"], inputs["labels"], pad_token_id=0
)
if self.args.predict_with_generate:
assert self.tokenizer.padding_side == "left", "This method only accepts left-padded tensor."
assert self.tokenizer.pad_token_id is not None, "Pad token is required."
prompt_len, label_len = inputs["input_ids"].size(-1), inputs["labels"].size(-1)
if prompt_len > label_len:
inputs["labels"] = self._pad_tensors_to_target_len(inputs["labels"], inputs["input_ids"])
if label_len > prompt_len:
inputs["input_ids"] = self._pad_tensors_to_target_len(inputs["input_ids"], inputs["labels"])
if "attention_mask" in inputs:
inputs["attention_mask"] = self._pad_tensors_to_target_len(
inputs["attention_mask"], inputs["labels"], pad_token_id=0
)
if "position_ids" in inputs:
inputs["position_ids"] = self._pad_tensors_to_target_len(
inputs["position_ids"], inputs["labels"], pad_token_id=0
)
loss, generated_tokens, labels = super().prediction_step(
model, inputs, prediction_loss_only=prediction_loss_only, ignore_keys=ignore_keys
)
if generated_tokens is not None:
generated_tokens[:, :max(prompt_len, label_len)] = (
self.tokenizer.pad_token_id * torch.ones_like(generated_tokens[:, :max(prompt_len, label_len)])
)
if generated_tokens is not None and self.args.predict_with_generate:
generated_tokens[:, :max(prompt_len, label_len)] = self.tokenizer.pad_token_id
generated_tokens = generated_tokens.contiguous()
return loss, generated_tokens, labels
@@ -66,16 +67,8 @@ class Seq2SeqPeftTrainer(PeftTrainer):
) -> torch.Tensor:
r"""
Pads the tensor to the same length as the target tensor.
Should only be called when predict_with_generate=True.
"""
if pad_token_id is None:
if self.tokenizer is not None and hasattr(self.tokenizer, "pad_token_id"):
assert self.tokenizer.padding_side == "left", "This method only accepts left-padded tensor."
pad_token_id = self.tokenizer.pad_token_id
else:
raise ValueError("PAD token is required.")
pad_token_id = pad_token_id if pad_token_id is not None else self.tokenizer.pad_token_id
padded_tensor = pad_token_id * torch.ones_like(tgt_tensor)
padded_tensor[:, -src_tensor.shape[-1]:] = src_tensor # adopt left-padding
return padded_tensor.contiguous() # in contiguous memory