One
One
**Describe the bug** I encountered an issue when using DeepSpeed 0.12.4 with the [OpenChat trainer](https://github.com/imoneoi/openchat), where checkpointing failed and raised an NCCL error. However, the checkpoints work fine when using...
### Describe the bug When running a training loop multiple times and calling `wandb.finish()` after each run, although it shows that all data is uploaded, the program is still stuck...
Brax cannot be installed with pip for Python 3.11. The dependency pytinyrenderer is not compatible with the latest Python. pytinyrenderer seems like a course project and hasn't been updated for...
Can we download the original (supposedly bfloat16) weights for fine-tuning? The checkpoint is int8 quantized.
Why is OmniIsaacGymEnvs (Omniverse Isaac Sim) version about 6x slower, compared with IsaacGymEnvs (IsaacGym Preview) version, with same hardware? Tested on Humanoid environment, headless, with 8192 environments (RTX 3080 Laptop)...
When I try to create "flan2021_submix", the process fails with a wrong checksum on the AESLC dataset. ``` Downloading and preparing dataset 11.10 MiB (download: 11.10 MiB, generated: Unknown size,...
I can't create "t0_submix", says no training split found in "adversarial_qa_dbert_answer_the_following_q_template_0to10_no_opt_x_shot" ``` ERROR:absl:Failed to load task 't0_task_adaptation:adversarial_qa_dbert_answer_the_following_q_template_0to10_no_opt_x_shot' as part of mixture 't0_submix' Traceback (most recent call last): File "/home/one/anaconda3/envs/flan/lib/python3.10/runpy.py", line...
### 🚀 The feature, motivation and pitch It'd be great to have a fused linear and cross-entropy function in PyTorch, for example, `torch.nn.functional.linear_cross_entropy`. This function acts as a fused linear...