4bit-quantization

DPO fine-tuning of Phi-3-mini for safety alignment using QLoRA, runnable end-to-end on a single 16GB T4 GPU. Includes training health monitoring, data validation, and a toxicity/reasoning eval harness.

pytorch ai-safety fine-tuning preference-learning dpo huggingface-transformers rlhf qlora llm-fine-tuning phi-3 preference-optimization llm-alignment 4bit-quantization

• Updated Jul 16, 2026
• Python

A practical guide for local LLM deployment with 4-bit quantization. / 4-bit量化本地大模型部署实战指南

python cuda quantization gradio huggingface llm local-llm bitsandbytes qwen deployment-guide 4bit-quantization

• Updated May 6, 2026
• Python

Benchmarking NER on Naamapadam across 11 Indic languages. EDA + model training using mBERT, XLM-R, T5, FlanT5, mT5 + LLM fine-tuning (TinyLlama, Llama-3.2, Gemma, Qwen, Mistral) + 0–5 shot inference on 9 generative models.

pytorch named-entity-recognition lora bert indic-languages peft few-shot-learning xlm-roberta huggingface-transformers multilingual-nlp llm ai4bharat hindi-nlp naamapadam 4bit-quantization 8bit-quantization

• Updated Sep 7, 2026
• Jupyter Notebook