Download _backup/logs/train_ca.log from TrendeX90/ca-cube: direct link, hf CLI and curl.
- Browser
- Download file 2.58 kB
-
https://huggingface.co/datasets/TrendeX90/ca-cube/resolve/main/_backup/logs/train_ca.log
- Command line
-
hf download hf://datasets/TrendeX90/ca-cube/_backup/logs/train_ca.log
-
curl -L -o train_ca.log https://huggingface.co/datasets/TrendeX90/ca-cube/resolve/main/_backup/logs/train_ca.log
2.58 kB
| [] loaded 2001 | |
| [] loaded 2002 | |
| [] loaded 2003 | |
| [] loaded 2004 | |
| [] loaded 2005 | |
| [] loaded 2006 | |
| [] loaded 2007 | |
| [] loaded 2008 | |
| [] loaded 2009 | |
| [] loaded 2010 | |
| [] loaded 2011 | |
| [] loaded 2012 | |
| [] loaded 2013 | |
| [] loaded 2014 | |
| [] loaded 2015 | |
| [] loaded 2016 | |
| [] loaded 2017 | |
| [] loaded 2018 | |
| [] loaded 2019 | |
| [] loaded 2020 | |
| [] loaded 2021 | |
| [] loaded 2022 | |
| [] loaded 2023 | |
| [] loaded 2024 | |
| [] loaded 2025 | |
| /root/trendex-ca/scripts/30_train_ca.py:67: RuntimeWarning: Mean of empty slice | |
| anom = lambda x: x - np.nanmean(x, axis=1, keepdims=True) | |
| /root/trendex-ca/scripts/30_train_ca.py:76: RuntimeWarning: Mean of empty slice | |
| {f"{k}_mean": np.nanmean(v[:, :, j], axis=1).astype("float32") for k, v in raw_m.items()} | | |
| /root/trendex-ca/scripts/30_train_ca.py:77: RuntimeWarning: Mean of empty slice | |
| {f"{k}_mean": np.nanmean(v, axis=1).astype("float32") for k, v in raw_y.items()}) | |
| [] eligible cells 3432538, sample 150000 | |
| /opt/conda/lib/python3.11/site-packages/xgboost/core.py:751: UserWarning: [17:41:03] WARNING: /__w/xgboost/xgboost/src/common/error_msg.cc:62: Falling back to prediction using DMatrix due to mismatched devices. This might lead to higher memory usage and slower performance. XGBoost is running on: cuda:0, while the input data is on: cpu. | |
| Potential solutions: | |
| - Use a data structure that matches the device ordinal in the booster. | |
| - Set the device for booster before call to inplace_predict. | |
| This warning will only be shown once. | |
| return func(**kwargs) | |
| [] tuning 20 trials, xgb 0.0519 vs ridge 0.0524, n_est 189 | |
| [] temporal skill vs ridge {'point': 0.05526524782180786, 'ci_lo': 0.03597217798233032, 'ci_hi': 0.07688862085342407, 'units': 6} | |
| [] spatial skill vs ridge {'point': 0.11963582038879395, 'ci_lo': 0.11683060228824615, 'ci_hi': 0.12261748313903809, 'units': 4115} | |
| [] GATE passed=True -> xgb | |
| /root/trendex-ca/scripts/30_train_ca.py:276: RuntimeWarning: Mean of empty slice | |
| clim_all = np.nanmean(f["ndvi"], axis=1, keepdims=True) | |
| /root/trendex-ca/scripts/30_train_ca.py:297: RuntimeWarning: Mean of empty slice | |
| summer_res[chunk[:, None], yi[None, :]] = np.nanmean(res[:, :, summer_idx], axis=2) | |
| [] oof block 2001-2005 done | |
| [] oof block 2006-2010 done | |
| [] oof block 2011-2015 done | |
| [] oof block 2016-2020 done | |
| [] oof block 2021-2025 done | |
| [] TRAIN_CA_DONE | |