Categories — API Model Output
Generated by Claude Sonnet (claude-sonnet-4-20250514) via the Anthropic API. Each category summary is a full recommendation article (~600 words) with top products, common complaints, worst product, and buying advice. Total cost: $0.43 USD for 39 API calls. Temperature: 0.3 for consistent, factual output.
Categories — Local Model Output
Generated by Flan-T5-base (google/flan-t5-base) — a 250M parameter seq2seq transformer running locally on GPU (NVIDIA RTX 4050). Unlike the API-based summaries, this model runs entirely offline with no API costs. The smaller model size results in shorter, more factual outputs compared to the Claude Sonnet API approach.
All Products
65 Amazon products with review stats and AI summaries
| Product | Category | Reviews | Rating | Sentiment |
|---|
Review Explorer
Browse sample reviews with model predictions
About This Project
Project 3: NLP Business Case — Group 4, IronHack AI Engineering Bootcamp