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 Sep 24, 2024.

This update includes:

• Model Support
  • Supported Deepseek-v1, see examples/deepseek_v1/README.md.
• Features
  • Added __repr__ methods for class Module, thanks to the contribution from @1ytic in Add module __repr__ methods #2191.
• API
  • [BREAKING CHANGE] The default max_batch_size of trtllm-build command is set to 2048.
• Bug fixes
  • Enabled share_embedding for the models that have no lm_head in legacy checkpoint conversion path, thanks to the contribution from @lkm2835 in Fix check_share_embedding #2232.

Thanks,
The TensorRT-LLM Engineering Team