YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

FLAN-T5-Small Fine-Tuned on Red Hat Documentation

Overview

Fine-tuned FLAN-T5-Small for question-answering on Red Hat documentation using LoRA and 4-bit quantization. Trained on redhat-docs_dataset (55,741 rows).

Model Details

Base Model: google/flan-t5-small
Fine-Tuning: LoRA (r=8, alpha=32, target_modules=["q", "v"])
Quantization: 4-bit (nf4)

Dataset

Fields: title, content, command, url
Artifacts: data/redhat-docs_dataset.jsonl, data/formatted_dataset.jsonl, data/tokenized_dataset.jsonl

Training

Hardware: T4 GPU, CUDA 11.8
Epochs: 2
Batch Size: 32 (4 per-device, 8 gradient accumulation)

Usage

Load with peft and transformers for Red Hat queries.

License

MIT License. Verify dataset licensing.

Contact

See GitHub.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support