You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
ColossalAI/examples/inference/build_smoothquant_weight.py

60 lines
1.7 KiB

import argparse
import os
import torch
from datasets import load_dataset
from transformers import LlamaTokenizer
from colossalai.inference.quant.smoothquant.models.llama import SmoothLlamaForCausalLM
def build_model_and_tokenizer(model_name):
tokenizer = LlamaTokenizer.from_pretrained(model_name, model_max_length=512)
kwargs = {"torch_dtype": torch.float16, "device_map": "sequential"}
model = SmoothLlamaForCausalLM.from_pretrained(model_name, **kwargs)
model = model.to(torch.float32)
return model, tokenizer
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument("--model-name", type=str, help="model name")
parser.add_argument(
"--output-path",
type=str,
help="where to save the checkpoint",
)
parser.add_argument(
"--dataset-path",
type=str,
help="location of the calibration dataset",
)
parser.add_argument("--num-samples", type=int, default=10)
parser.add_argument("--seq-len", type=int, default=512)
args = parser.parse_args()
return args
@torch.no_grad()
def main():
args = parse_args()
model_path = args.model_name
dataset_path = args.dataset_path
output_path = args.output_path
num_samples = args.num_samples
seq_len = args.seq_len
model, tokenizer = build_model_and_tokenizer(model_path)
if not os.path.exists(dataset_path):
raise FileNotFoundError(f"Cannot find the dataset at {args.dataset_path}")
dataset = load_dataset("json", data_files=dataset_path, split="train")
model.quantized(tokenizer, dataset, num_samples=num_samples, seq_len=seq_len)
model = model.cuda()
model.save_quantized(output_path, model_basename="llama-7b")
if __name__ == "__main__":
main()