mGPT-1.3B-tuvan / README.md
ai-forever's picture
Upload README.md with huggingface_hub
3508d84
metadata
language:
  - tyv
  - en
  - ru
license: mit
tags:
  - gpt3
  - transformers
  - mgpt

๐ŸŽ Tuvan mGPT 1.3B

Language model for Tuvan. Model has 1.3B parameters as you can guess from it's name.

Tuvan belongs to Turkic language family. It's a very evocative language with approximately 0.3 million speakers. Here are some facts about it:

  1. Spoken in the Tuva Republic in Russia, located in southern Siberia.
  2. It is famous for its throat singing or khoomei tradition.
  3. The Tuvan language has been influenced by Mongolic languages, reflecting the historical ties between the Tuvans and Mongols.

Technical details

It's one of the models derived from the base mGPT-XL (1.3B) model (see the list below) which was originally trained on the 61 languages from 25 language families using Wikipedia and C4 corpus.

We've found additional data for 23 languages most of which are considered as minor and decided to further tune the base model. Tuvan mGPT 1.3B was trained for another 2000 steps with batch_size=4 and context window of 2048 tokens on 1 A100.

Final perplexity for this model on validation is 40.84.

Chart of the training loss and perplexity:

Other mGPT-1.3B models

Feedback

If you'll found a bug of have additional data to train model on your language โ€” please, give us feedback.

Model will be improved over time. Stay tuned!