DeepSpeed fp16: Current loss scale already at minimum
Persistent fp16 overflow drives the dynamic loss scale down to its minimum and DeepSpeed aborts. Typically a bf16-pretrained model being fine-tuned in fp16 on hardware without bf16.
Persistent fp16 overflow drives the dynamic loss scale down to its minimum and DeepSpeed aborts. Typically a bf16-pretrained model being fine-tuned in fp16 on hardware without bf16.
What this failure is
DeepSpeed fp16: Current loss scale already at minimum is a Training Stability failure seen during ML training runs. Persistent fp16 overflow drives the dynamic loss scale down to its minimum and DeepSpeed aborts. Typically a bf16-pretrained model being fine-tuned in fp16 on hardware without bf16. Common tags: Deepspeed, Fp16, Loss Scale, Overflow.
Is this what broke your run? Paste your log.
You're reading about DeepSpeed fp16: Current loss scale already at minimum. 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)
Continuous gradient overflow in fp16 (narrow exponent range). Loss-scale hysteresis cannot recover under constant overflow. Model weights/activations exceed fp16 dynamic range. 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
- Loss scale decreases every step until training exits
- Common fine-tuning bf16-pretrained models (e.g. LLaMA2) in fp16
- On V100-class hardware without bf16
Common symptoms and what they mean
| Symptom | Why it happens |
|---|---|
| Exception: Current loss scale already at minimum - cannot decrease scale anymore | Continuous gradient overflow in fp16 (narrow exponent range) |
| overflow counter increments every step | Loss-scale hysteresis cannot recover under constant overflow |
| Loss scale monotonically drops from 2^16 to 1 | Model weights/activations exceed fp16 dynamic range |
Which systems are affected
- DeepSpeed fp16 mixed precision
- bf16-pretrained models fine-tuned in fp16
- V100 / hardware lacking bf16
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: Exception: Current loss scale already at minimum - cannot decrease scale anymore
- ✓Verified signal present: overflow counter increments every step
- ✓Verified signal present: Loss scale monotonically drops from 2^16 to 1
- ✓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 extensionDeepSpeed errors in context
DeepSpeed changes when parameters, gradients and optimizer state are created, partitioned, gathered and offloaded. The hub separates ZeRO, memory, checkpoint and pipeline failures by lifecycle phase.
Compare every deepspeed error side by sideRelated failures to investigate next
Root cause
- Continuous gradient overflow in fp16 (narrow exponent range)
- Loss-scale hysteresis cannot recover under constant overflow
- Model weights/activations exceed fp16 dynamic range
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.