Japanese-opt-2.7b Model

Disclaimer: This model is a work in progress!

This model is a fine-tuned version of facebook/opt-2.7b on the japanese wikipedia dataset.

Quick start

from transformers import pipeline

generator = pipeline('text-generation', model="tensorcat/japanese-opt-2.7b" , device=0, use_fast=False)
generator("今日は", min_length=80, max_length=200, 
    do_sample=True, early_stopping=True, temperature=.98, top_k=50, top_p=1.0)

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: 5e-05
• train_batch_size: 1
• eval_batch_size: 1
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 4
• total_eval_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3.0

Training results

Framework versions

• Pytorch 1.13.0+cu116