Public Launch links
Burmese GPT was publicly announced on January 5, 2024.
Technical overview
Burmese GPT is an open-source, 1.42B-parameter fine-tuned Burmese language model. At launch, it was described as the largest fine-tuned Burmese model. Its training work used a Burmese corpus of more than 15,000 texts, with about 80% of the model weights further pre-trained.
Data and development
A central part of the project was assembling open Burmese text from varied sources into a usable dataset. The corpus and dataset contribution are intended to support academic, research, and open-community work.
Intended use cases
- Text generation and completion in Burmese.
- A foundation model for task-specific Burmese fine-tunes, including question-and-answer and chat-style systems.
Difference vs other models
- Burmese GPT provides general language modeling capacity for Burmese.
- Burmese-Coder-4B focuses on coding and technical prompt support for developers.
Use Burmese GPT
Open Burmese GPT on Hugging Face
The model repository includes the published model files and reference details. Install Transformers and load the model directly with the example below.
!pip install transformers
# Loading the Model:
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("WYNN747/Burmese-GPT")
model = AutoModelForCausalLM.from_pretrained("WYNN747/Burmese-GPT")
input_text = "မီးထွန်းပွဲတော်သည် သီ"
input_ids = tokenizer.encode(input_text, return_tensors='pt')
output = model.generate(input_ids, max_length=50)
print(tokenizer.decode(output[0], skip_special_tokens=True))
# Completion: မီးထွန်းပွဲတော်သည် သီတင်းကျွတ်လပြည့်နေ့တွင် ကျင်းပသော ရိုးရာပွဲတော်တစ်ခု ဖြစ်သည်။