checkpointing, dataset, evaluation, final_adapter_path, lora, model, optimization, prompt_style, runtime, sdpo, sequence, total_steps

amr-fma/amr-fma-Mistral-7B-Instruct-v0.3-lora_sdpo-math-p1_sdpo_math_l3plus-s42

Method: lora_sdpo
Base model: mistralai/Mistral-7B-Instruct-v0.3
Dataset: DigitalLearningGmbH/MATH-lighteval (slug: math )
Seed: 42
Git commit: 8b979a30de6dfbf3b5a1052e42d8c0453b214d3f
Exp name: p1_sdpo_math_l3plus
WandB run: 02olkt1m

amr-fma/amr-fma-Mistral-7B-Instruct-v0.3-lora_sdpo-math-p1_sdpo_math_l3plus-s42

Tags

Checkpoints (branches)

Hyperparameter sections

amr-fma/
amr-fma-Mistral-7B-Instruct-v0.3-lora_sdpo-math-p1_sdpo_math_l3plus-s42