How sivagoshikadevi76-cell/h2s_ltm_project works
A proof-of-concept H₂S cumulative-dose forecaster, exposed both as a Streamlit interface and as a standalone script. Both generate synthetic worker data, train an LSTM on seven-day dose/temperature/humidity windows, and compare predictions with actual dose; the Streamlit app also displays a next-day forecast and charts. This is not a validated safety tool. The app excerpt omits some wiring, so only clear calls and documented workflow links are shown.
Main entry points: app.py Streamlit app · lstm_dose_predictor.py Standalone runner
Components 14
User interface
- Streamlit app
app.py - Forecast charts
app.py - Standalone runner
lstm_dose_predictor.py - Demo predictions
lstm_dose_predictor.py
Synthetic data
- Worker generator
app.py - Training dataset
app.py - Synthetic dataset
lstm_dose_predictor.py
Forecasting
- App training
app.py - App LSTM
app.py - App forecasts
app.py - Dose LSTM
lstm_dose_predictor.py - Script training
lstm_dose_predictor.py
Other components
- User
- Saved weights
Connections 15
- User → Streamlit app: uses
- Streamlit app → App training: starts training
- App training → Training dataset: builds examples
- Training dataset → Worker generator: generates workers
- App training → App LSTM: trains
- Streamlit app → App forecasts: runs forecast
- App forecasts → App LSTM: predicts with
- Streamlit app → Forecast charts: displays results
- User → Standalone runner: runs
- Standalone runner → Script training: calls
- Script training → Synthetic dataset: builds data
- Synthetic dataset → Dose LSTM: trains
- Standalone runner → Demo predictions: runs demo
- Demo predictions → Dose LSTM: predicts with
- Standalone runner → Saved weights: saves weights