diff --git a/examples/llm_finetune/nemotron/customizer_nemotron_nano_full_sft_chat.yaml b/examples/llm_finetune/nemotron/customizer_nemotron_nano_full_sft_chat.yaml index a0b34c4c10..0ef677a9b9 100644 --- a/examples/llm_finetune/nemotron/customizer_nemotron_nano_full_sft_chat.yaml +++ b/examples/llm_finetune/nemotron/customizer_nemotron_nano_full_sft_chat.yaml @@ -346,7 +346,7 @@ parallelizer: ci: time: "00:30:00" checkpoint_robustness: - hf_kl_threshold: 7e-2 + hf_kl_threshold: 1e-1 tokenizer_name: nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 no_check_resume: true experts_implementation: grouped_mm