license, language, pipeline_tag, library_name, base_model, tags

license	language	pipeline_tag	library_name	base_model	tags
apache-2.0	en	text-generation	transformers	Qwen/Qwen2.5-0.5B-Instruct	qwen llm sft conversational transformers pytorch

Supervised Fine-Tuned Qwen2.5-0.5B-Instruct (SFT)

This repository contains a Supervised Fine-Tuned (SFT) version of Qwen2.5-0.5B-Instruct. The model has been fine-tuned on a custom instruction-following conversational dataset to answer questions about Vishnu in a natural and helpful manner.

Model Details

• Base Model: Qwen/Qwen2.5-0.5B-Instruct
• Training Method: Supervised Fine-Tuning (SFT)
• Framework: Hugging Face Transformers
• Task: Conversational Text Generation

---

Installation

pip install transformers accelerate torch

---

Loading the Model

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

MODEL_ID = "vishnuamarapu/Full-Fine-Tuning-Qwen-2.5-0.5B-instruct-sft"

tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)

model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
    device_map="auto"
)

---

Example Inference

messages = [
    {
        "role": "system",
        "content": (
            "You are Vishnu's personal AI assistant. "
            "Answer questions about Vishnu using the provided information."
        )
    },
    {
        "role": "user",
        "content": "Tell me about Vishnu."
    }
]

text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)

inputs = tokenizer(
    text,
    return_tensors="pt"
).to(model.device)

outputs = model.generate(
    **inputs,
    max_new_tokens=256
)

response = tokenizer.decode(
    outputs[0][inputs.input_ids.shape[-1]:],
    skip_special_tokens=True
)

print(response)

---

Gradio Demo

import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

MODEL_ID = "vishnuamarapu/Full-Fine-Tuning-Qwen-2.5-0.5B-instruct-sft"

tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)

model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
    device_map="auto"
)

def chat(message, history):

    messages = [
        {
            "role": "system",
            "content": (
                "You are Vishnu's personal AI assistant. "
                "Answer questions about Vishnu using the provided information."
            )
        }
    ]

    for user, assistant in history:
        messages.append({"role": "user", "content": user})
        messages.append({"role": "assistant", "content": assistant})

    messages.append({"role": "user", "content": message})

    text = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )

    inputs = tokenizer(
        text,
        return_tensors="pt"
    ).to(model.device)

    outputs = model.generate(
        **inputs,
        max_new_tokens=256
    )

    response = tokenizer.decode(
        outputs[0][inputs.input_ids.shape[-1]:],
        skip_special_tokens=True
    )

    return response

gr.ChatInterface(chat).launch()

---

Generation Parameters

Parameter	Value
max_new_tokens	256
device_map	auto
torch_dtype	float16 (GPU) / float32 (CPU)

---

Repository Structure

config.json
generation_config.json
model.safetensors
tokenizer.json
tokenizer_config.json
chat_template.jinja
training_args.bin

---

Training Overview

This model was fine-tuned using Supervised Fine-Tuning (SFT) on a custom instruction-response dataset.

The training process included:

• Instruction-response formatting using the Qwen chat template
• Hugging Face Transformers
• TRL SFTTrainer
• PyTorch
• Custom conversational dataset

---

Citation

If you use this model in your work, please cite this repository.

@misc{vishnu_qwen25_sft,
  author = {Vishnu Amarapu},
  title = {Supervised Fine-Tuned Qwen2.5-0.5B-Instruct},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/vishnuamarapu/Full-Fine-Tuning-Qwen-2.5-0.5B-instruct-sft}}
}

---

Acknowledgements

• Alibaba Cloud Qwen Team for the base model.
• Hugging Face Transformers.
• TRL (Transformer Reinforcement Learning).
• PyTorch.