Anvil Playground

o1: TF-IDF text matching - fast vendor prediction

0
Predictions Run
0%
Top-1 Accuracy
0%
Top-10 Accuracy
0ms
Avg Inference
Anvil o1 uses traditional NLP techniques to match solicitation text against historical contract awards. The model tokenizes input text, applies TF-IDF weighting, and classifies against 1,847 known vendors. Fast inference (~5ms) but limited to vendors seen during training.
Architecture TF-IDF Vectorizer + Logistic Regression
Training Data 47K solicitations from SAM.gov (FY2020-2024)
Input Solicitation text + PSC code
Output Top 10 predicted vendors with confidence scores
Solicitation Details
Sample Solicitations
Loading samples...