Chris Kerwell Gresla

Results 7 issues of Chris Kerwell Gresla

Loved the quite elegant themes! I am somewhat of a newbie when it comes to Linux and was wondering what browser you were using in the [hermine](https://github.com/kiddae/dotfiles/blob/main/hermine/screenshot3.png) Screenshot -- i...

**Describe the bug** A clear and concise description of what the bug is. The issue I am facing is that of the assertion on [L316 of partitioned_param_coordinator.py](https://github.com/microsoft/DeepSpeed/blob/08e0733e4ad0e2a4a4c1014e393967b36a55daf1/deepspeed/runtime/zero/partitioned_param_coordinator.py#L316) is being raised...

bug
training

I added the functionality to measure when we receive the first chars response from models and then display the time to that moment in either milliseconds (ms) or seconds (s)

Hello! Awesome [X announcement](https://x.com/angli_ai/status/1937179078014226856) that you folks put out for [Simular](https://www.simular.ai/) and kudos on the [Agent S](https://arxiv.org/abs/2410.08164) & [Agent S2](https://arxiv.org/abs/2504.00906) papers. I was curious about the performance of the Agent...

Hello! Thank you folks for releasing this package -- its a brilliant addition to the `transformers` & `sentence-transformers` ecosystem. I was trying to build the package today with [flash-attention](https://github.com/Dao-AILab/flash-attention/releases/tag/v2.8.3), however...

please fix mate, else [slim dusty](https://www.wikiwand.com/en/articles/Slim_Dusty) will be reincarnated.