Released — December 2023
MyanmarGPT
128M Params Open Source First in Myanmar

The very first and largest usable Burmese language GPT in Myanmar. Lightweight, accurate, and easy to run on any device. Trained on private property datasets manually cleaned by Min Si Thu. Supports Burmese and international languages.

  • Free to use and open-source
  • Lightweight & accurate
  • Burmese + International Languages (61 total)
  • Runs on CPU — no GPU required
Released — December 2023
MyanmarGPT-Big
1.42B Params 61 Languages Enterprise

An enterprise-level multilingual LLM with Burmese as its primary language, supporting 61 languages. The largest Burmese language model released openly in Myanmar. Suitable for advanced NLP pipelines, RAG systems, and research applications.

Released — January 28, 2024
MyanmarGPT-Chat
128M Params Chat / QA

Fine-tuned on MyanmarGPT, MyanmarGPT-Chat is a question-answering model for the Burmese language with embedded knowledge of "A Brief History of the World." Ideal for conversational AI, chatbots, and interactive Burmese language applications.

Released — February 23, 2024
MyanmarGPTX
Faster Multiplatform Lightweight

Fine-tuned on MyanmarGPT-Chat, MyanmarGPTX is the fastest, lightest, and most portable model in the family. Designed for multiplatform deployment — mobile, edge devices, and low-resource environments. The best choice for production applications needing speed.

Quickstart

pip install transformers

MyanmarGPT — Code Examples

Using Pipeline
Python
# Using Pipeline
from transformers import pipeline

pipe = pipeline("text-generation", model="jojo-ai-mst/MyanmarGPT")
outputs = pipe("အီတလီ", do_sample=False)
print(outputs)
Using AutoTokenizer and CausalLM
Python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("jojo-ai-mst/MyanmarGPT")
model = AutoModelForCausalLM.from_pretrained("jojo-ai-mst/MyanmarGPT")

input_ids = tokenizer.encode("ချစ်သား", return_tensors='pt')
output = model.generate(input_ids, max_length=50)
print(tokenizer.decode(output[0], skip_special_tokens=True))

MyanmarGPT-Big — Code Examples

Using Pipeline
Python
# Using Pipeline
from transformers import pipeline

pipe = pipeline("text-generation", model="jojo-ai-mst/MyanmarGPT-Big")
outputs = pipe("အီတလီ", do_sample=False)
print(outputs)
Using AutoTokenizer and CausalLM
Python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("jojo-ai-mst/MyanmarGPT-Big")
model = AutoModelForCausalLM.from_pretrained("jojo-ai-mst/MyanmarGPT-Big")

input_ids = tokenizer.encode("ချစ်သား", return_tensors='pt')
output = model.generate(input_ids, max_length=50)
print(tokenizer.decode(output[0], skip_special_tokens=True))

DVB Interview

Chat GPT (AI) ကို မြန်မာလို သုံးစွဲနိုင်တော့မလား
DVB Youth Voice

Contributors

M
Min Si Thu
Creator · Founder · Lead Researcher