Geng Zhang

I've tried the official merlin-pytorch image, and met another bug. Please see: [https://github.com/rapidsai/dask-cuda/issues/974](https://github.com/rapidsai/dask-cuda/issues/974)

> > 请问下有适合做aspect-based sentiment analysis的中文数据集吗？我在网上找到的似乎都是句子级别的数据。 > > 你好，请问句子级别的中文数据集，是用的哪个呢？ https://github.com/SophonPlus/ChineseNlpCorpus 还有CSDN上有一些，基本都是来源不明的那种

I've tested solution 3, and discovered that, in order to equivalently load a global batch in each rank, I must also set `worker_count=1` when invoking `make_batch_reader`, otherwise the batch in...

I'm trying to build a DLRM model containing both model parallel layers & data parallel layers. In the model parallel part, I must make sure each rank consumes the same...

I'm too busy to handle dataloader recently:dizzy_face: that's why I resort to petastorm. I may have a look into the source code in the future, but I would recommend fixing...

@liuqingpu no

另外，计算rho有时候会报exp溢出，这种数值问题您有遇到过吗？

> @zxgx please read the following Contributor License Agreement(CLA). If you agree with the CLA, please reply with the following information. > > ``` > @microsoft-github-policy-service agree [company="{your company}"] >...

# Before fixing the server side reports the above error and exits. The client side reports error and hangs after server exiting: This is returned from the [request](https://github.com/microsoft/agent-lightning/blob/d00cc3e4aab227207520c080e30003c5597a988c/agentlightning/verl/daemon.py#L211) # After...

# Clarification In my debugging code, as all generated samples exceed the length limitation, there is no valid samples in each batch, and wandb log is empty. However, the skipping...