mkl-debug-ml-training
/mkl-debug-ml-training
Bada konkretne awarie trenowania w PyTorch, Lightning lub TensorFlow/Keras: NaN-y, brakujące gradienty, mylące straty i niepowtarzalne uruchomienia.
- Autor
- 00200200 ↗
- Repozytorium
- maintainer-skills-lab ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 24.10.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 1
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 2
Brak sygnałów. Repozytorium zmienione 24.10.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Bada konkretne awarie trenowania w PyTorch, Lightning lub TensorFlow/Keras: NaN-y, brakujące gradienty, mylące straty i niepowtarzalne uruchomienia.
Oryginalny opis (EN)
Investigate training failures, NaNs, missing gradients, misleading losses, and non-reproducible runs in PyTorch, Lightning, or TensorFlow/Keras. Use for concrete training bugs, not open-ended architecture or hyperparameter search.
Jak działa
- Establish the failing boundary
- Framework-specific controls
- Compare baseline and fix
Podgląd SKILL.md
---
name: "mkl-debug-ml-training"
description: "Investigate training failures, NaNs, missing gradients, misleading losses, and non-reproducible runs in PyTorch, Lightning, or TensorFlow/Keras. Use for concrete training bugs, not open-ended architecture or hyperparameter search."
globs: ["*.py", "*.ipynb", "models/**", "*.pt", "*.pth", "*.ckpt"]
---
# Debug a training run
Turn the reported symptom into an observable failure on a bounded input. Preserve the user's framework, backend, and intended training semantics. A CPU or eager-mode reproduction can isolate a bug, but does not establish what happens on the original accelerator, compiled graph, or distributed setup.
## Establish the failing boundary
Record the command, revision, Python/framework versions, imported package paths, device, dtype, precision, and execution mode. For resume or reproducibility issues, include data order, seed handling, checkpoint, optimizer, scheduler, and scaler state. Classify dependency failures separately from training bugs.
Reduce to a fixed batch while retaining shapes, label semantics, masks, and value ranges. Inspect tensors entering the loss: class indices, probability targets, and scalar targets have distinct contracts. Avoid blanket reshapes or squeezes that mask errors.
Example: scalar predictions of shape `[N, 1]` minus labels `[N]` produce `[N, N]` residuals via broadcasting. Aligning scalar labels to `[N, 1]` fixes the residual calculation.
Trace finite values through inputs, outputs, loss, gradients, and parameters. Find the first non-finite value or disconnected gradient before altering learning rates or clipping.
## Framework-specific controls
### PyTorch
- Distinguish `model.eval()` (affects dropout/batchnorm) from `torch.no_grad()` (disables autograd).
- Inspect unexpected `.detach()`, `.item()`, or unlisted optimizer parameters when gradients are missing.
- Pinpoint internal dispatcher tracebacks using the trace minifier:
```bash
python3 tools/minify_torch_trace.py traceback.log
```
- Guard against non-finite values before and after backward:
```python
if not torch.isfinite(loss).all():
raise FloatingPointError(f"Non-finite loss at step {step}: {loss.item()}")
loss.backward()
for name, param in model.named_parameters():
if param.grad is not None and not torch.isfinite(param.grad).all():
raise FloatingPointError(f"Non-finite gradient in {name}")
```
- Use `with torch.autograd.set_detect_anomaly(True):` to trace backward NaN crashes to forward origin ops.
- See [PyTorch reproducibility](https://docs.pytorch.org/docs/stable/notes/randomness.html) for deterministic settings (`CUBLAS_WORKSPACE_CONFIG`, `torch.use_deterministic_algorithms`).
### Lightning
- `Trainer(fast_dev_run=True)` tests loop plumbing, but disables loggers/callbacks. To test checkpoint or callback bugs, use `limit_train_batches` and `limit_val_batches` instead.
- Identify automatic vs. manual optimization before altering backward/step hooks.
- Check [Lightning debugging controls](https://lightning.ai/docs/pytorch/stable/debug/debugging_basic.html).
### TensorFlow / Keras
- Distinguish `tf.keras` from standalone `keras` and check backend engine.
- Verify forward operations execute inside `tf.GradientTape` and variables are watched without intermediate NumPy conversions.
- Use `compile(run_eagerly=True)` to inspect intermediate tensor values during debugging, then re-enable graph execution to confirm.
- For determinism, review `tf.keras.utils.set_random_seed` and `tf.config.experimental.enable_op_determinism()`.
## Compare baseline and fix
Start each comparison from equivalent model and optimizer states with identical batch and behavioral assertions. Check intermediate results and updates, not merely zero exit codes.
Report minimal command, environment, baseline failure, candidate outcome, and modes exercised. For stochastic symptoms, use bounded repeated comparisons rather than picking favorable single runs.
Wersje
| Repozytorium | Wynik | Licencja |
|---|---|---|
| Ta strona · 00200200/maintainer-skills-lab | 100 | MIT |
| 00200200/maintainer-skills-lab | 100 | MIT |
| 00200200/maintainer-skills-lab | 100 | MIT |
Identyczna treść występuje jeszcze w 2 innych folderach na GitHubie. Tu pokazujemy jedną, najlepiej ocenioną kopię.
Migawka bazy z 13.11.2026 · wersja paczki 7f319c4a45e9 (skrót SHA-256 całej paczki)
Pobrania w ostatnich 30 dniach: 0 · ranking
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 100keras
Wspiera pracę z Keras 3: wykonanie na wielu backendach (JAX, PyTorch, TensorFlow), API Sequential i Functional oraz własne pętle treningowe modeli głębokiego uczenia. - 100nvflare-convert-lightning
Konwertuje kod treningowy PyTorch Lightning na federacyjne zadanie NVFLARE przez patch Lightning Client API, z lokalną walidacją i eksportem zadania. - 100tensorflow
Opisuje TensorFlow z Keras 3 do budowy, trenowania i wdrażania sieci neuronowych: trening w Keras, własna pętla treningowa, instalacja z GPU, SavedModel dla TF Serving i konwersja do TFLite. - 100tensorflow-deep-learning
Podpowiada praktyki budowy, trenowania, oceny i wdrażania sieci w TensorFlow i Keras: potoki tf.data, callbacki checkpointów i early stopping oraz eksport modelu do serwowania. - 100matlab-import-external-ai-model
Importuje modele głębokiego uczenia z PyTorch, ONNX, Keras 3 i TensorFlow do MATLAB-a jako obiekty dlnetwork, z walidacją numeryczną względem oryginału. - 100ray-train
Orkiestruje rozproszone trenowanie PyTorch, TensorFlow i HuggingFace na klastrach, z Ray Tune, odpornością na awarie i elastycznym skalowaniem.