Transformer Cache Memory
Transformer cache (KV cache, past_key_values) memory grows with sequence length and can cause OOM at inference.
Transformer cache (KV cache, past_key_values) memory grows with sequence length and can cause OOM at inference.
What this failure is
Transformer Cache Memory is a Memory failure seen during ML training runs. Transformer cache (KV cache, past_key_values) memory grows with sequence length and can cause OOM at inference. Common tags: Kv Cache, Transformers, Llm, Inference.
Is this what broke your run? Paste your log.
You're reading about Transformer Cache Memory. 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)
Standard MHA KV cache: 2 * layers * hidden * seq * 2 bytes (FP16). Cache not cleared between requests. GQA/MQA not used. Sliding window not applied. 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
- OOM at inference with long context
- KV cache uses too much memory
- transformers.TransformerCache OOM
Common symptoms and what they mean
| Symptom | Why it happens |
|---|---|
| past_key_values length matches input length | Standard MHA KV cache: 2 * layers * hidden * seq * 2 bytes (FP16) |
| Memory grows linearly with generated tokens | Cache not cleared between requests |
| OOM with 100k+ context | GQA/MQA not used |
Which systems are affected
- LLM inference
- Long-form generation
- Multi-turn conversation with long history
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: past_key_values length matches input length
- ✓Verified signal present: Memory grows linearly with generated tokens
- ✓Verified signal present: OOM with 100k+ context
- ✓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
- Standard MHA KV cache: 2 * layers * hidden * seq * 2 bytes (FP16)
- Cache not cleared between requests
- GQA/MQA not used
- Sliding window not applied
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.