ErtasAI-Qwen2.5-3B-Instruct

ErtasAI — Fine-Tuned Qwen2.5-3B-Instruct

Overview

ErtasAI is a fine-tuned Large Language Model based on Qwen2.5-3B-Instruct.

The model was fine-tuned using QLoRA and Unsloth to adapt the base model for the ErtasAI project's domain-specific instruction following and text generation requirements.

This repository contains the trained LoRA adapter weights.

---

Model Details

---

Training Details

The model was trained for 3 epochs with a total of 1,545 training steps.

Training completed successfully.

---

Fine-Tuning Approach

The model was fine-tuned using QLoRA (Quantized Low-Rank Adaptation).

QLoRA enables efficient fine-tuning by training a small number of additional adapter parameters while keeping the base model quantized.

The training workflow was implemented using Unsloth for memory-efficient and accelerated fine-tuning.

---

Repository Contents

.
├── adapter_model.safetensors
├── adapter_config.json
├── tokenizer.json
├── tokenizer_config.json
├── chat_template.jinja
├── training_args.bin
└── README.md