Fewfeed - V2

project/ │ ├─ data/ │ └─ imdb_reviews.jsonl # "text": "...", "label": "pos/neg" │ ├─ prompts/ │ └─ sentiment.yaml # hierarchical prompt spec │ ├─ config/ │ └─ run.yaml # master configuration │ └─ main.py # driver script

| Setting | Standard ProtoNet | V2 FewFeed | |---------|------------------|-------------| | 5-way 1-shot | 49.2% | 53.7% | | 5-way 5-shot | 68.5% | 71.3% | | Training time (hours) | 3.2 | 2.1 | | Inference memory (MB) | 512 | 187 | v2 fewfeed

Metrics averaged over 600 episodes. Gains attributed to adaptive feeding and prototype caching. project/ │ ├─ data/ │ └─ imdb_reviews

name: SentimentAnalysis description: Few‑shot sentiment classification for movie reviews. v2 fewfeed