Update on the development branch

[kaiyux]

Hi,

The TensorRT-LLM team is pleased to announce that we have pushed an update to the development branch (and the Triton backend) this Oct 22, 2024.

This update includes:

• Model Support
  • Added support for Llama 3.2 model.
• Features
  • Added support for Qwen2ForSequenceClassification model architecture.
• API
  • [BREAKING CHANGE] Modified defaults for BuildConfig class so that they are aligned with the trtllm-build command.
• Bug fixes
  • Supported passing gpt_variant argument to the model conversion when provided by the user of examples/gpt/convert_checkpoint.py, thanks to the contribution from @tonylek in Passing gpt_variant to model conversion #2352.
• Documentation
  • Added a copy button for code snippets in the documentation. (need a copy code widget to be able to copy code snippets #2288)

Thanks,
The TensorRT-LLM Engineering Team