Download conf/config.yaml from OneScience-Group/ACE2: direct link, hf CLI and curl.
- Browser
- Download file 629 Bytes
-
https://huggingface.co/OneScience-Group/ACE2/resolve/main/conf/config.yaml
- Command line
-
hf download hf://OneScience-Group/ACE2/conf/config.yaml
-
curl -L -o config.yaml https://huggingface.co/OneScience-Group/ACE2/resolve/main/conf/config.yaml
629 Bytes
| seed: 17 | |
| data: | |
| format_version: ace2-full-grid-v1 | |
| path: data/ace2_tiny.npz | |
| samples: 4 | |
| time_steps: 6 | |
| channels: 50 | |
| height: 180 | |
| width: 360 | |
| levels: 8 | |
| dt_hours: 6 | |
| model: | |
| width: 4 | |
| depth: 8 | |
| modes_lat: 4 | |
| modes_lon: 4 | |
| forcing_channels: 4 | |
| paper_model: | |
| embedding_dim: 384 | |
| parameters: 450000000 | |
| autoregressive_training_steps: 2 | |
| train: | |
| epochs: 1 | |
| batch_size: 1 | |
| learning_rate: 0.001 | |
| two_step_weight: 0.5 | |
| checkpoint: result/checkpoints/ace2.pt | |
| inference: | |
| steps: 3 | |
| output: result/output/predictions.npz | |
| evaluation: | |
| output: result/evaluation/metrics.json | |
| figure: result/evaluation/comparison.png | |