Instructions to use kashif/weathernext2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use kashif/weathernext2 with Transformers:
# Load model directly from transformers import WeatherNext2ForWeatherForecasting model = WeatherNext2ForWeatherForecasting.from_pretrained("kashif/weathernext2", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Store the NaN fill value used by the original NaNCleaner
Browse files
model2/preprocessor_config.json
CHANGED
|
@@ -691,9 +691,9 @@
|
|
| 691 |
"year_progress_cos": -0.004568948580583765,
|
| 692 |
"year_progress_sin": 0.0006449962704664358
|
| 693 |
},
|
| 694 |
-
"
|
| 695 |
-
"sea_surface_temperature"
|
| 696 |
-
|
| 697 |
"num_input_timesteps": 2,
|
| 698 |
"pressure_levels": [
|
| 699 |
50,
|
|
@@ -1080,4 +1080,4 @@
|
|
| 1080 |
"cyclone_usa_pres_disc"
|
| 1081 |
],
|
| 1082 |
"time_step_hours": 6
|
| 1083 |
-
}
|
|
|
|
| 691 |
"year_progress_cos": -0.004568948580583765,
|
| 692 |
"year_progress_sin": 0.0006449962704664358
|
| 693 |
},
|
| 694 |
+
"nan_fill_values": {
|
| 695 |
+
"sea_surface_temperature": 269.48291015625
|
| 696 |
+
},
|
| 697 |
"num_input_timesteps": 2,
|
| 698 |
"pressure_levels": [
|
| 699 |
50,
|
|
|
|
| 1080 |
"cyclone_usa_pres_disc"
|
| 1081 |
],
|
| 1082 |
"time_step_hours": 6
|
| 1083 |
+
}
|