Matthias Brunel
Matthias Brunel
## Description There used to be some kind of tutorial about how to use them. But not anymore, so it might be difficult to set up the proper environment ##...
Hi there, I benchmarked resnet50 in onnx format against onnxruntime. I found out on an Azuredsv3 server that the inference ran in 200ms for glow with AOT against 50ms for...
Reorganization and update of the docs required by the mithril cloud launch. This change only concerns the docs.
The pull_from_pretrained method should be improved to match the original implementation. See #1 for details about the problems it causes
Goal is to add a ActionBuilderStore and an ActionPreprocessor to improve the accuracy the actions.
Quotes and yaml syntax might crash lavague, because the llm might output it as well, and then it's not parsable anymore. So we should sanytize the instructions.
on some interactive search bar like this one https://huggingface.co/nvidia/bigvgan_v2_24khz_100band_256x you have to wait for results to load before pressing enter. Or the search doesn't work
When the new tab is opened, we stay on the old one, and so the screenshot doesn't work. Happens on https://www.lavague.ai/ (which is funny). Probably not too hard to fix