How syeddaniyalg/blissnet-era5-inference works
This repository implements BLISSNet to reconstruct ERA5 temperature fields, progressing from data preparation through two-stage training to cross-region inference. The model’s internal stages include a coordinate-based SIREN trunk, a full-grid coefficient branch, and a sparse-sensor branch. The README documents the notebook workflows and evaluation; notebook source was not sampled, so those connections are based on its explicit descriptions rather than inspected calls.
Main entry points: blissnet.py BLISSNet orchestration · training.ipynb Stage 1 training · inference.ipynb Cross-region evaluation · datasets Regional NetCDF
Components 18
Data preparation
- Prepare data
dataset_and_visualization.ipynb - Regional NetCDF
datasets
BLISSNet model
- BLISSNet orchestration
blissnet/blissnet.py - SIREN trunk
blissnet/blissnet.py - Full-grid branch
blissnet/blissnet.py - Attention U-Net
blissnet/blissnet.py - Sparse-sensor branch
blissnet/blissnet.py - Sensor transformer
blissnet/blissnet.py - Cross-attention
blissnet/blissnet.py - Attention pool
blissnet/blissnet.py - Coefficient decoder
blissnet/blissnet.py
Model training
- Stage 1 training
training.ipynb - Stage 2 training
training.ipynb - Model checkpoints
Inference evaluation
- Cross-region evaluation
inference.ipynb - Evaluation outputs
Other components
- Researcher
- Raw ERA5 data
Connections 24
- Researcher → Prepare data: prepares data
README.md - Raw ERA5 data → Prepare data: provides GRIB
README.md - Prepare data → Regional NetCDF: exports NetCDF
README.md - Researcher → Stage 1 training: starts training
README.md - Regional NetCDF → Stage 1 training: trains on grids
README.md - Stage 1 training → BLISSNet orchestration: trains model
README.md - BLISSNet orchestration → SIREN trunk: evaluates basis
blissnet/blissnet.py - BLISSNet orchestration → Full-grid branch: routes full grids
blissnet/blissnet.py - Full-grid branch → Attention U-Net: extracts features
blissnet/blissnet.py - Researcher → Stage 2 training: starts training
README.md - Stage 1 training → Stage 2 training: freezes components
README.md - Stage 2 training → BLISSNet orchestration: trains sparse path
README.md - BLISSNet orchestration → Sparse-sensor branch: routes sensors
blissnet/blissnet.py - Sparse-sensor branch → Sensor transformer: encodes sensors
blissnet/blissnet.py - Sparse-sensor branch → Cross-attention: attends to latents
blissnet/blissnet.py - Sparse-sensor branch → Attention pool: pools features
blissnet/blissnet.py - Sparse-sensor branch → Coefficient decoder: decodes coefficients
blissnet/blissnet.py - Stage 1 training → Model checkpoints: saves checkpoints
README.md - Stage 2 training → Model checkpoints: saves checkpoints
README.md - Researcher → Cross-region evaluation: runs evaluation
README.md - Model checkpoints → Cross-region evaluation: loads checkpoints
README.md - Regional NetCDF → Cross-region evaluation: supplies test data
README.md - Cross-region evaluation → BLISSNet orchestration: evaluates model
README.md - Cross-region evaluation → Evaluation outputs: records metrics
README.md