NaN Detection and Skip
NaN detection and skipping prevents corrupted gradients from propagating but requires careful implementation.
NaN detection and skipping prevents corrupted gradients from propagating but requires careful implementation.
What this failure is
NaN Detection and Skip is a Training Stability failure seen during ML training runs. NaN detection and skipping prevents corrupted gradients from propagating but requires careful implementation. Common tags: Nan Detection, Gradscaler, Mixed Precision, Skip Update.
Is this what broke your run? Paste your log.
You're reading about NaN Detection and Skip. Paste your own crash log or traceback below and get the real root cause for YOUR run, not this generic entry. No account, no card. Logs are masked at ingress and never saved to account history.
Want 14 days on the Scale plan?
Request an evaluation code. A verified workplace organization activates up to 50 diagnoses a day, alerts, history, and follow-up questions. No credit card or automatic subscription.
Why it happens (the mechanism)
Gradient scaler detects NaN and skips update. NaN detected in loss but optimizer step is skipped. NaN in one rank doesn't propagate to others. Mixed precision underflow causes NaN. Taken together, these mechanisms explain why the failure is reproducible, why it tends to surface on specific workloads or scales, and why generic mitigation attempts often fall short without addressing the underlying cause.
What you'll observe
- Training continues after NaN occurs
- Loss is NaN for some steps but recovers
- NaN detection prevents training crash
Common symptoms and what they mean
| Symptom | Why it happens |
|---|---|
| Loss is NaN for one step but then recovers | Gradient scaler detects NaN and skips update |
| Gradient norm reports NaN | NaN detected in loss but optimizer step is skipped |
| Parameter updates are skipped due to NaN | NaN in one rank doesn't propagate to others |
Which systems are affected
- Training with mixed precision and gradient scaler
- Training with noisy data
- Training with unstable loss
How to confirm this is the problem
Use this checklist to test the hypothesis against a small reproduction. No single line proves the root cause, so preserve the preceding events and compare one variable at a time.
- ✓Reproduce the failure from a clean checkpoint/seed: the symptom must appear without warm-up state from a previous run.
- ✓Verified signal present: Loss is NaN for one step but then recovers
- ✓Verified signal present: Gradient norm reports NaN
- ✓Verified signal present: Parameter updates are skipped due to NaN
- ✓A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.
The fix and the prevention pattern
The root cause is on this page and stays free. A free account adds the exact remediation steps, saved history, and the fix on every entry in the encyclopedia.
Sign up free. Unlock the full analysisNo credit card. Daily allowance follows verified trust tier. Instant access.
Diagnose this failure in VS Code
Select the traceback or open the failed terminal, then run Denpex locally to see the initiating rank, collateral failures, exact fix, and verification command without uploading the log.
Install the free VS Code extensionRelated failures to investigate next
Root cause
- Gradient scaler detects NaN and skips update
- NaN detected in loss but optimizer step is skipped
- NaN in one rank doesn't propagate to others
- Mixed precision underflow causes NaN
The fix and how to prevent it
Evaluate Denpex on your own logs
Request a Scale evaluation code. A verified workplace organization activates 14 days with up to 50 diagnoses a day. Every account keeps its current diagnosis allowance and gets a verification path. No card or automatic subscription.
Don't just read the fix, diagnose your run
The encyclopedia tells you what went wrong. Denpex tells you what went wrong in YOUR training run. With your logs, your config, and your stack.