Hugging Face's logo

jmajkutewicz
/
zephyr-7b-shp

PEFT
Safetensors
mistral
alignment-handbook
trl
dpo
Generated from Trainer
Model card Files Community
zephyr-7b-shp / README.md
Jan Majkutewicz
End of training
df491f0 verified
preview code
|
raw
history blame contribute delete
1.83 kB
metadata
base_model: alignment-handbook/zephyr-7b-sft-full
datasets:
  - SeniorKabanocci/SHP-binarized
library_name: peft
license: apache-2.0
tags:
  - alignment-handbook
  - trl
  - dpo
  - generated_from_trainer
model-index:
  - name: zephyr-7b-shp
    results: []

Visualize in Weights & Biases

zephyr-7b-shp

This model is a fine-tuned version of alignment-handbook/zephyr-7b-sft-full on the SeniorKabanocci/SHP-binarized dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4546
  • Rewards/chosen: -1.0035
  • Rewards/rejected: -2.3680
  • Rewards/accuracies: 0.7864
  • Rewards/margins: 1.3644
  • Logps/rejected: -260.2372
  • Logps/chosen: -378.4407
  • Logits/rejected: -2.7522
  • Logits/chosen: -2.7984

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
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 1

Training results

Framework versions

  • PEFT 0.10.0
  • Transformers 4.42.3
  • Pytorch 2.2.0
  • Datasets 2.20.0
  • Tokenizers 0.19.1