1. 多卡训练:
# 设定使用的设备,如果有多张显卡,用逗号分隔
os.environ["CUDA_VISIBLE_DEVICES"] = "0,1"
# 使用所有可见的设备
DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# 检查设备数量
num_gpus = torch.cuda.device_count()
print(f"Number of available GPUs: {num_gpus}")
def model_to_device(model):
if num_gpus > 1:
model = torch.nn.DataParallel(model)
return model.to(DEVICE)
def torch_gc():
if torch.cuda.is_available():
torch.cuda.empty_cache()
torch.cuda.ipc_collect()
2. Windows安装bitsandbytes库:
pip install https://github.com/jllllll/bitsandbytes-windows-webui/releases/download/wheels/bitsandbytes-0.40.0-py3-none-win_amd64.whl
3. NameError: name 'round_up' is not defined
删除Tesseract相关目录,可能是有冲突。
4. error @achrinza/node-ipc@9.2.2: The engine "node" is incompatible with this
yarn config set ignore-engines true