Bangla Language Model Fine-tuning
15 repos
Fine-tuning and optimization techniques for large language models adapted to Bengali language tasks. The cluster centers on adapting base models like Llama and Mistral for Bangla through LoRA, quantization, and other parameter-efficient methods, leveraging frameworks like TRL, Unsloth, and Hugging Face Transformers. Repositories span training implementations, model checkpoints, and inference optimization for low-resource deployment of Bengali-capable LLMs.
Jupyter Notebook · 2
Python · 1
unsloth ·190
trl ·190
transformers ·110
safetensors ·110
text-generation-inference ·106
endpoints_compatible ·106
huggingface ·102
en ·97
conversational ·93
qwen2 ·93
