DongfuJiang/prm_version3_subsample_hf

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

prmversion3hf

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the prmconversationsprmversion3math+webinstructsub-mcq+webinstructsub-oe+apps+gsmmixrefsubsamplehf dataset. It achieves the following results on the evaluation set:

• —Loss: 0.1262

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• —learning_rate: 5e-06
• —trainbatchsize: 1
• —evalbatchsize: 1
• —seed: 42
• —distributed_type: multi-GPU
• —num_devices: 8
• —gradientaccumulationsteps: 8
• —totaltrainbatch_size: 64
• —totalevalbatch_size: 8
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: cosine
• —lrschedulerwarmup_ratio: 0.05
• —num_epochs: 1

Training results

Framework versions

• —Transformers 4.45.0
• —Pytorch 2.5.1+cu124
• —Datasets 3.1.0
• —Tokenizers 0.20.3