README_en.md

llm-jp-eval-inference

[ English | 日本語 ]

In this repository, we primarily release implementations of fast batch inference processing for llm-jp-eval using the following libraries: For installation and inference execution, please refer to the README.md within each module.

• vLLM
• TensorRT-LLM
• Hugging Face Transformers (baseline)

In addition, a tool for run management using Weights & Biases are published in wandb_run_management.

Inference and Evaluation Execution Methods

Please refer to the Inference Execution Method and Evaluation Method in llm-jp-eval.