L'électron rare
f55093d6fe
ESP-IDF CI / Host Tests (Unity) (push) Successful in 1m8s
CI / firmware-native (push) Successful in 2m57s
Rust Protection Tests / Cargo test (host) (push) Failing after 3m21s
ESP-IDF CI / ESP-IDF Build (v5.4) (push) Failing after 6m55s
ESP-IDF CI / Memory Budget Gate (push) Has been skipped
qa-cicd-environments / qa-kxkm-s3-build (push) Successful in 8m53s
qa-cicd-environments / qa-sim-host (push) Successful in 2m2s
qa-cicd-environments / qa-kxkm-s3-memory-budget (push) Successful in 11m17s
Context: the project archive (KXKM_Batterie_Parallelator-main) had no git history locally; a fresh repository is needed to host it on git.saillant.cc (electron/KXKM_Batterie_Parallelator). Approach: initialize a new repo on branch main, stage the archive content, and harden .gitignore before the first commit. Changes: - Import the full project tree: firmware/, firmware-idf/, firmware-rs/, iosApp/, kxkm-bmu-app/, kxkm-api/, hardware/, docs/, specs/, scripts/, models/, tests/ - Keep project dotfiles tracked despite the trailing '.*' ignore rule: .github/, .claude/, .superpowers/, .gitattributes, .markdownlint.json - Extend .gitignore: firmware/src/credentials.h (local secrets, template kept), kxkm-bmu-app/**/build/ (66 MB compiled iOS framework), .remember/ (session data) Impact: the project can now be maintained on the self-hosted Gitea forge with a clean, secret-free initial history.
33 lines
2.3 KiB
Plaintext
33 lines
2.3 KiB
Plaintext
22:48:11 INFO Torch device: cpu
|
|
22:48:11 INFO Loading models/features_adapted.parquet
|
|
22:48:11 INFO Raw shape: (15272, 20)
|
|
22:48:11 INFO After dropna: (11887, 21)
|
|
22:48:11 INFO Devices found: ['gocab', 'k-led1', 'k-led2', 'tender']
|
|
22:48:11 INFO Split — train: ['gocab'] val: ['k-led1'] test: ['k-led2']
|
|
22:48:11 INFO Train: 4567 Val: 2801 Test: 2427
|
|
22:48:11 INFO Input features: 13
|
|
22:48:11 INFO FPNN — degree=2 hidden=64 params=6785 poly_features=104
|
|
22:48:12 INFO Epoch 1/50 train_loss=0.574156 val_loss=0.765727 (0.2s)
|
|
22:48:13 INFO Epoch 10/50 train_loss=0.339697 val_loss=0.369873 (0.6s)
|
|
22:48:13 INFO Epoch 20/50 train_loss=0.322594 val_loss=0.361639 (1.0s)
|
|
22:48:14 INFO Epoch 30/50 train_loss=0.319076 val_loss=0.361201 (1.5s)
|
|
22:48:14 INFO Epoch 40/50 train_loss=0.317846 val_loss=0.360959 (2.1s)
|
|
22:48:15 INFO Epoch 50/50 train_loss=0.317077 val_loss=0.360825 (2.5s)
|
|
22:48:15 INFO Best epoch: 46 best val_loss: 0.360760
|
|
22:48:15 INFO Test metrics — MAPE: 6.15% RMSE: 0.428533 R2: 0.0986
|
|
22:48:15 INFO Saved PyTorch model: models/fpnn_soh.pt (31.6 KB)
|
|
/Users/electron/Documents/Lelectron_rare/1-KXKM/KXKM_Batterie_Parallelator/scripts/ml/train_fpnn.py:421: DeprecationWarning: You are using the legacy TorchScript-based ONNX export. Starting in PyTorch 2.9, the new torch.export-based ONNX exporter has become the default. Learn more about the new export logic: https://docs.pytorch.org/docs/stable/onnx_export.html. For exporting control flow: https://pytorch.org/tutorials/beginner/onnx/export_control_flow_model_to_onnx_tutorial.html
|
|
torch.onnx.export(
|
|
22:48:15 INFO Saved ONNX model: models/fpnn_soh.onnx (29.1 KB)
|
|
22:48:15 INFO ============================================================
|
|
22:48:15 INFO FPNN Training Summary
|
|
22:48:15 INFO Features: 13 (degree-2 expansion -> 104)
|
|
22:48:15 INFO Parameters: 6785
|
|
22:48:15 INFO Best epoch: 46 / 50
|
|
22:48:15 INFO Val MSE: 0.360760
|
|
22:48:15 INFO Test MAPE: 6.15%
|
|
22:48:15 INFO Test RMSE: 0.428533
|
|
22:48:15 INFO Test R2: 0.0986
|
|
22:48:15 INFO Model size: 31.6 KB (pt) 29.1 KB (onnx)
|
|
22:48:15 INFO ============================================================
|