Andrew Lapp

I see a 2% reduction in runtime with the configuration below, but it complicates the model quite a bit and GQA has absurd compile times. I'll think about this some...

@Mikeriess I don't think it would be a substantial effort. We pinned to numpy 1 due to - https://github.com/outlines-dev/outlines/issues/976 - https://github.com/outlines-dev/outlines/pull/977 The imports would need to be fixed. Unless there's...

>Got it. Have had some compatibility issues with outlines and the newest llama.cpp version due to this, so if this is a trivial change that is great news 👍 Is...

It's working now! Thanks for fixing quickly @Wauplin, great job!

Hi, I'm seeing this issue again @XciD @severo It seems that - Tensorboards aren't being populated with updated when provided new logs as of ~48 hours ago. [Example](https://huggingface.co/distily/distily_validate_extra_grad_stats/tensorboard) - ~~**all...

Appears to be resolved, thanks!

Re-occurring.

Here's the hacky script I'm using to render a hub tensorboard locally. `python3 run.py distily/distily_dataset_sweep` It retrieves all tfevent files from a repo, puts them in a temporary directory, and...

Error re-occurring starting some time in the past few days.

>I can't unfortunately share the model weights or lora adapter, but am happy to help replicate this on my end and share any specific diagnostics! Are you able to share...