tanliboy/lambda-llama-3-8b-ipo-test

<!-- 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. -->

lambda-llama-3-8b-ipo-test

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the HuggingFaceH4/ultrafeedbackbinarized and the tanliboy/orcadpo_pairs datasets. It achieves the following results on the evaluation set:

• —Loss: 0.8931
• —Rewards/chosen: -0.3610
• —Rewards/rejected: -0.5883
• —Rewards/accuracies: 0.7922
• —Rewards/margins: 0.2272
• —Logps/rejected: -3.1373
• —Logps/chosen: -2.5334
• —Logits/rejected: -2.9939
• —Logits/chosen: -2.9244

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-07
• —trainbatchsize: 4
• —evalbatchsize: 4
• —seed: 42
• —distributed_type: multi-GPU
• —num_devices: 8
• —gradientaccumulationsteps: 4
• —totaltrainbatch_size: 128
• —totalevalbatch_size: 32
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: cosine
• —lrschedulerwarmup_ratio: 0.1
• —num_epochs: 1

Training results

Framework versions

• —Transformers 4.44.2
• —Pytorch 2.4.0+cu121
• —Datasets 2.19.1
• —Tokenizers 0.19.1