Instructions to use yusufizzetmurat/fed-pulse-forecaster with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use yusufizzetmurat/fed-pulse-forecaster with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("yusufizzetmurat/fed-pulse-forecaster", dtype="auto") - Notebooks
- Google Colab
- Kaggle
yusufizzetmurat/fed-pulse-forecaster
Artefact for the fed-pulse FOMC text analytics project.
- Source code: https://github.com/yusufizzetmurat/fed-pulse
- Live demo: https://fedpulse.yusufizzetmurat.com
- License:
mit
Training corpus
Multi-head forecaster trained on the canonical FOMC event-window dataset over an expanding walk-forward fold protocol. Multi-target rates heads and the vol-regime classifier share the cross-bank DAPT encoder backbone.
Training command
python -m app.train_forecaster --training-package-id <tp-id> --seed 11
Attribution
- FOMC statements, minutes, and press conferences: scraped from the Board of Governors of the Federal Reserve System (US-government public domain), with supervised labels merged from gtfintechlab/fomc_communication and the historical drlexus/fed-statements-and-minutes Kaggle archive.
- Treasury yields and macro series: FRED (Federal Reserve Bank of St. Louis).
- Equity / volatility prices: Yahoo Finance via
yfinance.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support