Introduced new log files capturing detailed training and evaluation metrics for Qwen 2.5-7B and DeepSeek 14B models, including SFT, GRPO, and RLVR tasks. Enhanced logging to facilitate better tracking of command execution, runtime performance, and evaluation results, improving the analysis of model behavior during training and evaluation processes.
feat: add training and evaluation logs for Qwen models
Introduced new log files for various training and evaluation runs, including SFT and GRPO tasks for Qwen 2.5-7B and DeepSeek 14B models. Enhanced logging details to capture command execution, runtime metrics, and evaluation results, facilitating better tracking and analysis of model performance during training and evaluation processes.
