nayohan's picture
Create README.md
5bacb00
metadata
license: apache-2.0
datasets:
  - DILAB-HYU/KoQuality
language:
  - ko
pipeline_tag: text-generation
tags:
  - polyglot-ko
  - gpt-neox
  - KoQuality
base_model: EleutherAI/polyglot-ko-1.3b

This model is a instruct-tuned EleutherAI/polyglot-ko-1.3b model.

Training hyperparameters

  • learning_rate: 5e-5
  • train_batch_size: 1
  • seed: 42
  • distributed_type: multi-GPU (A30 24G) + CPU Offloading (384GB)
  • num_devices: 2
  • gradient_accumulation_steps: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 2.0

Framework versions

  • Transformers 4.34.1
  • Pytorch 2.0.1+cu117
  • Datasets 2.11.0
  • deepspeed 0.9.5