EMA Decay Misconfiguration
Exponential Moving Average (EMA) decay misconfiguration causes poor model averaging and training instability.
Exponential Moving Average (EMA) decay misconfiguration causes poor model averaging and training instability.
What this failure is
EMA Decay Misconfiguration is a Training Stability failure seen during ML training runs. Exponential Moving Average (EMA) decay misconfiguration causes poor model averaging and training instability. Common tags: Ema, Model Averaging, Stability, Swa.
Is this what broke your run? Paste your log.
You're reading about EMA Decay Misconfiguration. 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)
EMA decay rate too high (no averaging) or too low (too slow update). EMA buffer on CPU but model on GPU. EMA update during gradient accumulation. EMA model not in eval mode for validation. 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
- EMA model is worse than training model
- EMA model has inconsistent behavior
- EMA model not improving during training
Common symptoms and what they mean
| Symptom | Why it happens |
|---|---|
| Validation accuracy is lower for EMA than online model | EMA decay rate too high (no averaging) or too low (too slow update) |
| EMA model loss is higher | EMA buffer on CPU but model on GPU |
| EMA decay not updating properly | EMA update during gradient accumulation |
Which systems are affected
- Training with EMA for stable model
- Diffusion model training
- GAN generator training with EMA
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: Validation accuracy is lower for EMA than online model
- ✓Verified signal present: EMA model loss is higher
- ✓Verified signal present: EMA decay not updating properly
- ✓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
- EMA decay rate too high (no averaging) or too low (too slow update)
- EMA buffer on CPU but model on GPU
- EMA update during gradient accumulation
- EMA model not in eval mode for validation
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.