Lei Zhang
Lei Zhang
Is there any recent plan to implent tensor slicing?
wrong evaluation method!
Canoe A System for Collaborative Learning for Neural Nets https://arxiv.org/abs/2108.12124
``` @article{usmanova2021distillation, title={A distillation-based approach integrating continual learning and federated learning for pervasive services}, author={Usmanova, Anastasiia and Portet, Fran{\c{c}}ois and Lalanda, Philippe and Vega, German}, journal={arXiv preprint arXiv:2109.04197}, year={2021} }...
> cc: @H-Huang @wconstab > > > > I am guessing we should run more steps to see if the memory keeps increasing or not. For the loss issue, I...
the recommended methods work for me! many thanks!
Sure! Gemini is also important, and I will add support for it soon. Besides, I will also divide the model.py into separate items (e.g., models/openai.py, models/anthropic.py, etc.) to better organize...
> Fix has been merged, will be present in the upcoming Sysbox v0.6.5 release. > > Closing this issue. Hi, @ctalledo. This fix is of great significance and essentiality. Could...
Great to see it fixed! Btw, does it mean that LMCache can only run on the official Docker image or other dockerfile based on it? Perhaps rebuilding nixl from source...
It seems like FP8Linear could not run on inference mode. I have removed the `torch.inference_mode()`, but it still not works: ``` /usr/local/lib/python3.10/dist-packages/torchao/ops.py:12: FutureWarning: `torch.library.impl_abstract` was renamed to `torch.library.register_fake`. Please use...