Zhe Chen
# I will summarize common issues here. ## 1. Multi-GPU Inference - RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and...
Hello, thank you for contributing to our repository. We will proceed with the merge after conducting some tests. Thanks again!
Hello, if you want to get input_embeddings for the language model, you can use `model.language_model.get_input_embeddings()`.
Hello, this result was tested by the author of Video-MME. I am not clear about some of the details, but I can reproduce this result recently when I tested the...
这个报错是因为链接不到controller，建议您检查一下
可以先跳过SD worker，那个是可选的
Yes, the small image will be resized to multiple of 448 in the preprocessing.
InternVL 1.5 does not support object detection because we only used object detection data during the pre-training stage and did not utilize it during the SFT stage. Now, the new...
我遇到过一样的问题，你是不是没有安装flash attention
我测试下来，只要安装了flash attention，tp就能正常使用，暂时还不清楚内部是什么原因