Zhonghao Yan
Zhonghao Yan
I think that my label is a little bit small that model only get bg after randomcrop. Do you think this is possible?
Yes, I didn't change the code on your github, just the path and name
When I use the demo, SAM gives more correct segmentation results, but using the end-to-end method, it doesn't work well, is there a solution?
@25benjaminli You're right. But I don't see how he implements it in the code, can you point me to the exact code location please?
@25benjaminli Thanks! I'll try it!
Why is the required shape[0] equal to `bs` (batch size)?
When I use `conda install tsnecuda -c conda-forge` The terminal displays the following message ``` Collecting package metadata (current_repodata.json): | WARNING conda.models.version:get_matcher(546): Using .* with relational operator is superfluous and...
So, does version 0.4.0 of `verl` solve this issue? I'm using `transformers==4.52.4` and `vllm==0.8.5`, and I’m still getting similar error messages.
你好，想请问一下您是如何配置环境的，仓库中似乎并没有明确说明需要的 torch 的版本。如果方便的话，可以分享一下您的环境吗？我的邮箱是 [email protected]