metadata

license: mit
base_model: TheBloke/zephyr-7B-alpha-GPTQ
tags:
  - generated_from_trainer
model-index:
  - name: helpdesk-zephyr-7B-alpha
    results: []

helpdesk-zephyr-7B-alpha

This model is a fine-tuned version of TheBloke/zephyr-7B-alpha-GPTQ on the None dataset.

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: 8
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
num_epochs: 2

Training results

Framework versions

Transformers 4.34.0
Pytorch 2.0.0
Datasets 2.11.0
Tokenizers 0.14.1