怎么样做一个网站,苏州网页制作培训班,计算机作业网页设计代码,优秀软文范例800字文章目录一、量化加速sdxl模型1#xff09;涉及模型2#xff09;环境安装3#xff09;转换模型safetensor to pytorch文件4#xff09;tensorRT的环境准备#xff08;1#xff09;下载tensorRT 10.10#xff08;2#xff09;下载cuda一、量化加速sdxl模型
1#xff0…
文章目录一、量化加速sdxl模型1涉及模型2环境安装3转换模型safetensor to pytorch文件4tensorRT的环境准备1下载tensorRT 10.102下载cuda一、量化加速sdxl模型
1涉及模型
涉及模型
waiNSFWIllustrious_v120.safetensors
anyloraCheckpoint_bakedvaeBlessedFp16.safetensors2环境安装
torch1.9.0
safetensors0.3.0
numpy2.2.63转换模型safetensor to pytorch文件
转换指令
python3 safetensor_to_pytorch_converter.py代码
#!/usr/bin/env python3Safetensor to PyTorch Converter
支持将safetensor格式的模型文件转换为PyTorch格式
import os
import sys
import argparse
from pathlib import Path
import torch
from safetensors.torch import load_file, save_file
import logging# 设置日志
logging.basicConfig(levellogging.INFO,format%(asctime)s - %(levelname)s - %(message)s
)
logger logging.getLogger(__name__)def convert_safetensor_to_pytorch(safetensor_path, output_pathNone):将safetensor文件转换为PyTorch格式Args:safetensor_path (str): safetensor文件路径output_path (str, optional): 输出文件路径如果为None则自动生成Returns:str: 输出文件路径try:# 检查输入文件是否存在if not os.path.exists(safetensor_path):raise FileNotFoundError(f文件不存在: {safetensor_path})# 如果未指定输出路径自动生成if output_path is None:input_path Path(safetensor_path)output_path input_path.with_suffix(.pth)logger.info(f正在加载safetensor文件: {safetensor_path})# 加载safetensor文件state_dict load_file(safetensor_path)logger.info(f模型包含 {len(state_dict)} 个参数)# 保存为PyTorch格式torch.save(state_dict, output_path)logger.info(f转换完成保存到: {output_path})# 显示文件大小信息input_size os.path.getsize(safetensor_path) / (1024 * 1024) # MBoutput_size os.path.getsize(output_path) / (1024 * 1024) # MBlogger.info(f输入文件大小: {input_size:.2f} MB)logger.info(f输出文件大小: {output_size:.2f} MB)return output_pathexcept Exception as e:logger.error(f转换失败: {str(e)})raisedef convert_specific_models():转换指定的两个模型文件 models_to_convert [/home/fun/tensor_pic/safe_models/waiNSFWIllustrious_v120.safetensors,/home/fun/tensor_pic/safe_models/anyloraCheckpoint_bakedvaeBlessedFp16.safetensors]converted_files []for model_name in models_to_convert:if os.path.exists(model_name):logger.info(f开始转换: {model_name})try:output_file convert_safetensor_to_pytorch(model_name)converted_files.append(output_file)logger.info(f✅ {model_name} 转换成功)except Exception as e:logger.error(f❌ {model_name} 转换失败: {str(e)})else:logger.warning(f⚠️ 文件不存在: {model_name})return converted_filesdef main():parser argparse.ArgumentParser(description将safetensor文件转换为PyTorch格式)parser.add_argument(--input, -i, typestr, help输入的safetensor文件路径)parser.add_argument(--output, -o, typestr, help输出的PyTorch文件路径)parser.add_argument(--convert-specific, actionstore_true, help转换指定的两个模型文件)args parser.parse_args()try:if args.convert_specific:logger.info(开始转换指定的模型文件...)converted_files convert_specific_models()if converted_files:logger.info(f成功转换 {len(converted_files)} 个文件:)for file in converted_files:logger.info(f - {file})else:logger.warning(没有成功转换任何文件)elif args.input:# 转换单个文件output_file convert_safetensor_to_pytorch(args.input, args.output)logger.info(f转换完成: {output_file})else:# 默认转换指定模型logger.info(未指定参数开始转换指定的模型文件...)convert_specific_models()except KeyboardInterrupt:logger.info(用户中断操作)sys.exit(1)except Exception as e:logger.error(f程序执行失败: {str(e)})sys.exit(1)if __name__ __main__:main() 输出
(tensor_sd_xl) funfun:~/tensor_pic/safe2pytorch_tool$ python3 safetensor_to_pytorch_converter.py
2025-07-31 13:02:27,927 - INFO - 未指定参数开始转换指定的模型文件...
2025-07-31 13:02:27,927 - INFO - 开始转换: /home/fun/tensor_pic/safe_models/waiNSFWIllustrious_v120.safetensors
2025-07-31 13:02:27,927 - INFO - 正在加载safetensor文件: /home/fun/tensor_pic/safe_models/waiNSFWIllustrious_v120.safetensors
2025-07-31 13:02:27,945 - INFO - 模型包含 2515 个参数
2025-07-31 13:02:31,083 - INFO - 转换完成保存到: /home/fun/tensor_pic/safe_models/waiNSFWIllustrious_v120.pth
2025-07-31 13:02:31,084 - INFO - 输入文件大小: 6616.63 MB
2025-07-31 13:02:31,084 - INFO - 输出文件大小: 6617.17 MB
2025-07-31 13:02:31,084 - INFO - ✅ /home/fun/tensor_pic/safe_models/waiNSFWIllustrious_v120.safetensors 转换成功
2025-07-31 13:02:31,084 - INFO - 开始转换: /home/fun/tensor_pic/safe_models/anyloraCheckpoint_bakedvaeBlessedFp16.safetensors
2025-07-31 13:02:31,085 - INFO - 正在加载safetensor文件: /home/fun/tensor_pic/safe_models/anyloraCheckpoint_bakedvaeBlessedFp16.safetensors
2025-07-31 13:02:31,092 - INFO - 模型包含 1143 个参数
2025-07-31 13:02:32,022 - INFO - 转换完成保存到: /home/fun/tensor_pic/safe_models/anyloraCheckpoint_bakedvaeBlessedFp16.pth
2025-07-31 13:02:32,022 - INFO - 输入文件大小: 2033.85 MB
2025-07-31 13:02:32,022 - INFO - 输出文件大小: 2034.11 MB
2025-07-31 13:02:32,023 - INFO - ✅ /home/fun/tensor_pic/safe_models/anyloraCheckpoint_bakedvaeBlessedFp16.safetensors 转换成功
(tensor_sd_xl) funfun:~/tensor_pic/safe2pytorch_tool$ 4tensorRT的环境准备
1下载tensorRT 10.10
[链接] 命令安装这个包
1、安装包
sudo dpkg -i nv-tensorrt-local-repo-ubuntu2204-10.10.0-cuda-12.9_1.0-1_amd64.deb
2、查看 TensorRT 包是否可用如果能看到 libnvinfer、libnvinfer-dev 等相关包说明源添加成功
apt-cache search tensorrt
如果没有输出说明还没安装。
3、安装 TensorRT
sudo apt update
sudo apt install libnvinfer-dev libnvinfer-plugin-dev python3-tensorrt
4、安装完成后再用第2步命令检查
apt-cache search tensorrt
验证 Python TensorRT 是否可用
python3 -c import tensorrt; print(tensorrt.__version__)
如果能输出版本号说明 Python TensorRT 安装成功。2下载cuda
12.9 cuda链接
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-ubuntu2204.pin
sudo mv cuda-ubuntu2204.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/12.9.1/local_installers/cuda-repo-ubuntu2204-12-9-local_12.9.1-575.57.08-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu2204-12-9-local_12.9.1-575.57.08-1_amd64.deb
sudo cp /var/cuda-repo-ubuntu2204-12-9-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-9
