ghlili

> @StomachCold 请问如何复现使用 checkpoints_pooling在mini-imagenet上5-shot的结果？ --num_shots 设5

> We know that, unlike previous methods, here the images are resized to 224 x 224, not 84 x 84. > > So what about the **original JPEG image size**?...

`Traceback (most recent call last): File "/root/LLM-Tuning/chatglm_lora_tuning.py", line 141, in main() File "/root/LLM-Tuning/chatglm_lora_tuning.py", line 134, in main trainer.train() File "/root/anaconda3/envs/llmtune/lib/python3.11/site-packages/transformers/trainer.py", line 1633, in train return inner_training_loop( ^^^^^^^^^^^^^^^^^^^^ File "/root/anaconda3/envs/llmtune/lib/python3.11/site-packages/transformers/trainer.py", line...

`RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0! (when checking argument for argument index in method wrapper_CUDA__index_select)` 那请问有没有遇到过在tensor在两张gpu上的情况

我输出了model，发现都在cuda:0上，好奇怪

输出的{'': 0} 我没有改过你的代码 不知道是不是数据加载问题，以前用DDP只存在过存在于cpu和cuda上，没遇到过存在于两个cuda的 [https://discuss.huggingface.co/t/runtimeerror-expected-all-tensors-to-be-on-the-same-device-but-found-at-least-two-devices-cuda-1-and-cuda-0/39548/3](url) 我在这里看到有人遇到同样问题，但是好像没解决问题