Rebiha

Home › Datasets › E-commerce & Product Descriptions

E-commerce & Product Descriptions Fine-Tuning Dataset

Product listings, descriptions, and selling copy for online stores.

Fine-tune a model on this dataset

$29 for the dataset. Training is priced separately by model, starting at $47.

What is in this dataset

948 instruction-response pairs covering product listings, descriptions, and selling copy for online stores. Each pair is a realistic question with a full written answer, in the JSONL format used for supervised fine-tuning. No synthetic filler, no duplicated rows.

Example pairs

Real rows from the dataset, shown in full.

Write a product title for a high-end Bluetooth speaker that ranks well on Amazon and encourages clicks.

Start with the most searched keyword: 'Bluetooth speaker'. Then add key features that differentiate it: 'waterproof', '360° sound', '30-hour battery'. Include brand name and a benefit: 'Ultimate Ears Boom 3 - Portable Waterproof Bluetooth Speaker with 360° Immersive Sound, 30-Hour Battery, and Deep Bass for Outdoor Adventures'. Keep it under 150 characters, use pipes or hyphens for separators, and avoid ALL CAPS for every word—just capitalize the first letter of each major word. Test with Amazon's search term report to confirm top-performing keywords are included.

Write a product description for a women's cashmere sweater that feels luxurious and drives sales.

Open with an emotional hook: 'Wrap yourself in the buttery-soft embrace of 100% Grade-A Mongolian cashmere—designed for days when comfort meets elegance.' Then describe the fit and feel: 'This classic crewneck features a relaxed silhouette that drapes effortlessly, with ribbed cuffs and hem to hold its shape wash after wash.' Add sensory details: 'Each fiber is hand-selected and lightly brushed for a halo of warmth without the weight.' Close with styling tips and care: 'Pair with tailored trousers for the office or jeans for weekend coffee. Hand wash cold, lay flat to dry—this sweater only gets softer with time.' Include a size guide table and a…

Write a product description for a wireless noise-cancelling headphone that appeals to tech-savvy buyers.

Lead with the primary benefit: 'Experience studio-grade silence anywhere with the SoundShield Pro—adaptive noise cancellation that learns your environment in real time.' Break down specs in an easy-to-scan format: '40mm custom drivers deliver rich, balanced audio from deep bass to crystalline highs. Bluetooth 5.3 ensures lag-free streaming up to 100 feet. Battery life: 60 hours with ANC on, 80 hours with ANC off.' Address pain points: 'Memory foam ear cups with breathable leatherette stay comfortable during all-day wear. The folding hinge fits into the included hard case for travel.' End with social proof: 'Rated 4.8 stars by 2,500+ audiophiles—backed by a…

Who this is for

Online stores and listing tools that need product copy at volume.

Which models this works with

The dataset is plain JSONL, so it works with any instruction-tuned open model. On Rebiha you can train it directly on:

Training uses LoRA, or QLoRA on a 4-bit base for larger models. Your base model's weights are never modified — training produces a separate adapter.

How it works

  1. Pick this dataset and a base model.
  2. Rebiha runs the job on a managed GPU. You set up nothing.
  3. Download your model: a ready-to-run GGUF, plus developer assets (adapter, tokenizer, configs).

Fine-tuning or RAG?

Use RAG when the model needs facts that change — your prices, your documents, today's inventory. Use fine-tuning when you want consistent style and structure without prompting for it every time. This dataset teaches the shape of a good answer in this domain, not facts to look up.

Related datasets

Start fine-tuning