metadata

license: mit
library_name: peft
tags:
  - trl
  - kto
  - KTO
  - WeniGPT
  - generated_from_trainer
base_model: HuggingFaceH4/zephyr-7b-beta
model-index:
  - name: WeniGPT-Agents-Zephyr-1.0.17-KTO
    results: []

WeniGPT-Agents-Zephyr-1.0.17-KTO

This model is a fine-tuned version of HuggingFaceH4/zephyr-7b-beta on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.4733
Eval/rewards/chosen: -221.5663
Eval/logps/chosen: -2499.3209
Eval/rewards/rejected: -201.6566
Eval/logps/rejected: -2280.9316
Eval/rewards/margins: -19.9096
Eval/kl: 0.0

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: 0.0002
train_batch_size: 4
eval_batch_size: 4
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: linear
lr_scheduler_warmup_ratio: 0.03
training_steps: 145
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.4844	0.34	50	0.4733	0.0
0.5031	0.68	100	0.4733	0.0

Framework versions

PEFT 0.10.0
Transformers 4.39.1
Pytorch 2.1.0+cu118
Datasets 2.18.0
Tokenizers 0.15.2