qlora_finetuning_cpu.py

#
# Copyright 2016 The BigDL Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#

import torch
import os

import transformers
from transformers import LlamaTokenizer

from transformers import BitsAndBytesConfig
from ipex_llm.transformers.qlora import get_peft_model, prepare_model_for_kbit_training, LoraConfig
from ipex_llm.transformers import AutoModelForCausalLM
from datasets import load_dataset
import argparse
from ipex_llm.utils.isa_checker import ISAChecker

current_dir = os.path.dirname(os.path.realpath(__file__))
common_util_path = os.path.join(current_dir, '..', '..', 'GPU', 'LLM-Finetuning')
import sys
sys.path.append(common_util_path)
from common.utils import Prompter, get_train_val_data

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='Predict Tokens using `generate()` API for Llama2 model')
    parser.add_argument('--repo-id-or-model-path', type=str, default="meta-llama/Llama-2-7b-hf",
                        help='The huggingface repo id for the Llama2 (e.g. `meta-llama/Llama-2-7b-hf` and `meta-llama/Llama-2-13b-chat-hf`) to be downloaded'
                             ', or the path to the huggingface checkpoint folder')
    parser.add_argument('--dataset', type=str, default="yahma/alpaca-cleaned")

    args = parser.parse_args()
    model_path = args.repo_id_or_model_path
    dataset_path = args.dataset
    tokenizer = LlamaTokenizer.from_pretrained(model_path, trust_remote_code=True)

    if dataset_path.endswith(".json") or dataset_path.endswith(".jsonl"):
        data = load_dataset("json", data_files=dataset_path)
    else:
        data = load_dataset(dataset_path)

    # For illustration purpose, only use part of data to train
    data = data["train"].train_test_split(train_size=0.1, shuffle=False)

    # Data processing
    prompter = Prompter("alpaca")
    train_data, _ = get_train_val_data(data, tokenizer, prompter, train_on_inputs=True,
                                       add_eos_token=False, cutoff_len=256, val_set_size=0, seed=42)


    bnb_config = BitsAndBytesConfig(
        load_in_4bit=True,
        bnb_4bit_use_double_quant=False,
        bnb_4bit_quant_type="int4",
        bnb_4bit_compute_dtype=torch.bfloat16
    )
    model = AutoModelForCausalLM.from_pretrained(model_path,
                                                 quantization_config=bnb_config, )

    # below is also supported
    # model = AutoModelForCausalLM.from_pretrained(model_path,
    #                                              # nf4 not supported on cpu yet
    #                                              load_in_low_bit="sym_int4",
    #                                              optimize_model=False,
    #                                              torch_dtype=torch.bfloat16,
    #                                              modules_to_not_convert=["lm_head"], )

    model = model.to('cpu')
    model = prepare_model_for_kbit_training(model, use_gradient_checkpointing=False)
    model.enable_input_require_grads()
    config = LoraConfig(
        r=8,
        lora_alpha=32,
        target_modules=["q_proj", "k_proj", "v_proj"],
        lora_dropout=0.05,
        bias="none",
        task_type="CAUSAL_LM"
    )
    model = get_peft_model(model, config)
    tokenizer.pad_token_id = 0
    tokenizer.padding_side = "left"
    
    # To avoid only one core is used on client CPU
    isa_checker = ISAChecker()
    bf16_flag = isa_checker.check_avx512()
    
    trainer = transformers.Trainer(
        model=model,
        train_dataset=train_data,
        args=transformers.TrainingArguments(
            per_device_train_batch_size=4,
            gradient_accumulation_steps=1,
            warmup_steps=20,
            max_steps=200,
            learning_rate=2e-4,
            save_steps=100,
            bf16=bf16_flag,
            logging_steps=20,
            output_dir="outputs",
            optim="adamw_hf",  # paged_adamw_8bit is not supported yet
            # gradient_checkpointing=True, # can further reduce memory but slower
        ),
        # Inputs are dynamically padded to the maximum length of a batch
        data_collator=transformers.DataCollatorForSeq2Seq(
            tokenizer, pad_to_multiple_of=8, return_tensors="pt", padding=True
        ),
    )
    model.config.use_cache = False  # silence the warnings. Please re-enable for inference!
    result = trainer.train()
    print(result)