qwen3-0.6b-sft-best

Qwen3-0.6B SFT LoRA Adapter (Best Trial)

Best SFT (Supervised Fine-Tuning) LoRA adapter for Qwen/Qwen3-0.6B-Base, trained on the Databricks Dolly 15k dataset.

Model Details

• Developed by: Abdullah Iqbal (26904), Anushe Ali (26418)
• Model type: LoRA Adapter (PEFT)
• Language: English
• Base model: Qwen/Qwen3-0.6B-Base
• Training dataset: databricks/databricks-dolly-15k
• Training regime: fp16 mixed precision
• Hardware: Google Colab T4 GPU

Training Hyperparameters

Evaluation Results

Evaluated on 10 manually curated prompts with gold-reference answers.

How to Use

Prerequisites

pip install --upgrade torchao transformers peft accelerate

Note: You may encounter ImportError: Found an incompatible version of torchao if your torchao version is below 0.16.0. Run pip install --upgrade torchao to fix this.

Inference Code

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-0.6B-Base", torch_dtype="auto", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("Abdullah121212/qwen3-0.6b-sft-best")

model = PeftModel.from_pretrained(base_model, "Abdullah121212/qwen3-0.6b-sft-best")
model.eval()

prompt = "Explain the difference between supervised and unsupervised learning in simple terms."
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=256, do_sample=True, temperature=0.7)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Pipeline

Baseline (Qwen3-0.6B-Base)
  -> SFT (5 LoRA trials, this is trial 3 = best)
    -> DPO (see Abdullah121212/qwen3-0.6b-dpo-best)

Framework Versions

• PEFT 0.19.1
• Transformers 4.x
• TRL 0.x
• PyTorch 2.x