Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16

Training loss became NaN. This entry explains how to confirm the cause, apply the fix, and separate it from adjacent numerics failures.

Quick answer

Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16 means Training loss became NaN. Preserve the first preceding error, then run the targeted control below.

Symptom: Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16
Root cause: Training loss became NaN. The decisive evidence is the first log line that precedes "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" and differs from a healthy run. A nearby failure remains a competing hypothesis until a control separates configuration, capacity, transport, and hardware causes.
Recommended fix: Enable gradient clipping, lower learning rate 5-10x, check for bad data batch.
How Denpex helps: Denpex investigates Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16 using the evidence you provide or your connected workload collects. Earlier rank, host or application evidence is needed to distinguish an initiating failure from a downstream report.

Training Stability#numerics#loss#nan#bf16#transformer#gradient

What this failure is

The literal signature is "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16". It is a training stability failure associated with mixed precision, gradient scaling, and numerical kernels. The line identifies the failing operation or subsystem, while the surrounding evidence decides whether it is the initiating fault or a downstream symptom.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16. Paste your own traceback and relevant evidence for an investigation of your workload, with a next action or a specific missing fact. A reference entry does not establish your cause. No account or card for the free diagnosis. Review data handling before submitting sensitive logs.

Before uploading, review cloud data handling and local options.

training_logs.txt

No log to hand? Try one:

3 free diagnoses/day

Why it happens (the mechanism)

Training loss became NaN. The failure becomes visible at this call site because the operation first requires the missing resource, valid state, healthy peer, or correct result. Earlier log lines and a known-good control carry more causal value than the final wrapper exception.

What you'll observe

The workload stops or loses forward progress after emitting "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16".
A retry on the same configuration reproduces the failure because the causal state has not changed.
The outer framework exception can hide the rank, node, allocation, or dependency that failed first.
Increasing timeouts or reducing workload size can suppress the symptom without correcting the cause.

Common symptoms and what they mean

Symptom	Why it happens
Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16	Training loss became NaN.
The same operation fails at a consistent stage of mixed precision, gradient scaling, and numerical kernels.	The decisive evidence is the first log line that precedes "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" and differs from a healthy run.
The first related warning appears before the final exception and names the causal subsystem.	A nearby failure remains a competing hypothesis until a control separates configuration, capacity, transport, and hardware causes.
A known-good control changes one variable and either reproduces or clears the failure.	Training loss became NaN.

Which systems are affected

mixed precision, gradient scaling, and numerical kernels
production-shaped multi-accelerator workloads
containerized and bare-metal deployments of the same stack

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.

✓Find the first occurrence of "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" and preserve at least 100 lines before it.
✓Identify which rank, node, device, or process emitted the first related warning.
✓Compare the failing rank, node, input, or configuration with one known-good control.
✓Repeat the same input after the targeted change and require the signature to disappear.
✓Resume from last checkpoint before divergence only after the control passes.

Root cause

Training loss became NaN.
The decisive evidence is the first log line that precedes "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" and differs from a healthy run.
A nearby failure remains a competing hypothesis until a control separates configuration, capacity, transport, and hardware causes.

The fix and how to prevent it

Searchable error signature

search key

Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16

Use this text as a lookup key in logs and upstream issue trackers. It is not presented as a captured customer log. Confirm the cause from your own preceding events, versions, configuration and the cited references.

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.

No credit card. Daily allowance follows verified trust tier. Instant access.

Why the recommended fix works

Enable gradient clipping, lower learning rate 5-10x, check for bad data batch. This changes the condition in the causal diagnosis instead of hiding the outer exception. The repeated control proves ownership before recovery from last checkpoint before divergence.

Code examples

snippet

# Preserve evidence before restarting
rg -n -i 'error|exception|timeout|failed' <log-file>
nvidia-smi
python -m torch.utils.collect_env

# Find the exact signature in the complete log
rg -n -F -- "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" <log-file>

Adapt the snippet to your framework. The same pattern holds for PyTorch Lightning, Hugging Face Trainer, DeepSpeed, Megatron-LM, and vLLM training wrappers. Where the wrapper exposes a config flag (for examplelr_scheduler_type in Trainer), prefer the flag over the imperative API to keep the schedule declarative and reproducible.

Best practices by model family

Model / Stack	Recommendation	Notes
First response	Preserve the first failure	Keep the context before "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" so aggregation does not erase causality.
Confirmation	Change one variable	Use a known-good node, rank, input, or configuration as the control.
Recovery	Resume from last checkpoint before divergence	Resume only after the literal signature no longer appears in the same control.

With the fix vs without the fix

Dimension	With the fix	Without the fix
Evidence	First preceding error and one controlled comparison	Only the final aggregated exception
Fix	Enable gradient clipping, lower learning rate 5-10x, check for bad data batch.	Retrying the unchanged workload
Exit criterion	"Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" is absent in the repeated control	The job happened to run once

Diagnostic note

“Treat "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" as a search key and an investigation checkpoint, not as proof of every cause associated with the phrase. The high-value evidence is what changed immediately before it and whether the failure follows the workload, node, or configuration.”

Visual fingerprint

Decision path for Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16

literal error captured
        |
        v
find first preceding failure
        |
        v
run one known-good control
        |
        +-- follows workload --> inspect input or configuration
        +-- follows node ------> inspect hardware or platform
        +-- disappears --------> validate the targeted fix

The control separates workload, configuration, and node ownership before recovery from last checkpoint before divergence.

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 extension

Related failures to investigate next

Frequently asked questions

Questions engineers and on-call staff commonly ask about this failure.

What does "Loss is NaN gradients contained NaN in transformer layer mlp c_proj bf16" mean?

Training loss became NaN.

Is this line always the root cause?

No. It can be the direct failure or the point where an earlier failure becomes visible. The first preceding error and a controlled comparison decide which.

What should I collect before restarting?

Collect complete log context, the emitting rank or node, component versions, resolved configuration, and the diagnostic output shown above.

What is the fastest confirmation?

Compare the failing rank, node, input, or configuration with one known-good control.

How do I prevent it from recurring?

Apply the smallest causal change and repeat the same workload.

References

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.

Diagnose your logs free Compare a solved incident Browse all failures

Related Training Stability errors

NaN Detection and Skip

Training Stability · high

LR Without Warmup or Decay

Training Stability · high

Learning Rate Finder Result Misuse

Training Stability · medium

Gradient Accumulation BatchNorm Issue

Training Stability · high