model_hh_shp4_200

This model is a fine-tuned version of meta-llama/Llama-2-7b-chat-hf on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.4445
  • Rewards/chosen: -0.1401
  • Rewards/rejected: -1.3796
  • Rewards/accuracies: 0.6300
  • Rewards/margins: 1.2395
  • Logps/rejected: -230.1749
  • Logps/chosen: -224.4940
  • Logits/rejected: -0.7701
  • Logits/chosen: -0.7769

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: 0.0005
  • train_batch_size: 4
  • eval_batch_size: 1
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 100
  • training_steps: 1000

Training results

Training Loss Epoch Step Validation Loss Rewards/chosen Rewards/rejected Rewards/accuracies Rewards/margins Logps/rejected Logps/chosen Logits/rejected Logits/chosen
0.0 8.0 100 1.4330 -0.1080 -1.3964 0.6200 1.2883 -230.1935 -224.4584 -0.7684 -0.7753
0.0 16.0 200 1.4371 -0.0911 -1.3887 0.6400 1.2976 -230.1849 -224.4396 -0.7692 -0.7762
0.0 24.0 300 1.4477 -0.1125 -1.3921 0.6200 1.2795 -230.1887 -224.4634 -0.7693 -0.7763
0.0 32.0 400 1.4521 -0.1143 -1.4167 0.6200 1.3024 -230.2161 -224.4653 -0.7696 -0.7763
0.0 40.0 500 1.4631 -0.1153 -1.3806 0.6200 1.2653 -230.1759 -224.4665 -0.7701 -0.7771
0.0 48.0 600 1.4455 -0.1180 -1.3970 0.6300 1.2791 -230.1942 -224.4695 -0.7698 -0.7769
0.0 56.0 700 1.4292 -0.0800 -1.3720 0.6100 1.2920 -230.1664 -224.4273 -0.7704 -0.7775
0.0 64.0 800 1.4434 -0.0943 -1.3739 0.6200 1.2796 -230.1686 -224.4432 -0.7703 -0.7773
0.0 72.0 900 1.4493 -0.1016 -1.4044 0.6100 1.3028 -230.2024 -224.4513 -0.7704 -0.7773
0.0 80.0 1000 1.4445 -0.1401 -1.3796 0.6300 1.2395 -230.1749 -224.4940 -0.7701 -0.7769

Framework versions

  • PEFT 0.10.0
  • Transformers 4.39.1
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
11
Inference API
Unable to determine this model’s pipeline type. Check the docs .

Model tree for guoyu-zhang/model_hh_shp4_200

Adapter
(1079)
this model