CUDA out of memory during sentence_transformers encode sequence length batch size

Sentence Transformers exceeded VRAM during encoding because the token count, batch size, output handling, and other GPU residents exceeded the model memory budget. Sequence length can dominate even when the document count looks small. This entry explains how to confirm the cause, apply the fix, and separate it from adjacent vectordb failures.

The literal signature is "CUDA out of memory during sentence_transformers encode sequence length batch size". It is a memory failure associated with FAISS, Milvus, and sentence-transformer GPU indexes. The line identifies the failing operation or subsystem, while the surrounding evidence decides whether it is the initiating fault or a downstream symptom.

Sentence Transformers exceeded VRAM during encoding because the token count, batch size, output handling, and other GPU residents exceeded the model memory budget. Sequence length can dominate even when the document count looks small. 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.

• The workload stops or loses forward progress after emitting "CUDA out of memory during sentence_transformers encode sequence length batch size".
• 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.

• FAISS, Milvus, and sentence-transformer GPU indexes
• production-shaped multi-accelerator workloads
• containerized and bare-metal deployments of the same stack

Apply the following checklist to a small reproduction: each box below is a positive signal that you are looking at this exact failure rather than a sibling in the same taxonomy.

• ✓Find the first occurrence of "CUDA out of memory during sentence_transformers encode sequence length batch size" and preserve at least 100 lines before it.
• ✓Identify which rank, node, device, or process emitted the first related warning.
• ✓log token lengths and peak allocated memory per batch. Find whether one outlier document or cumulative retention of output tensors drives the peak.
• ✓Repeat the same input after the targeted change and require the signature to disappear.
• ✓Resume from the first uncommitted embedding batch after memory bounds are corrected only after the control passes.

CUDA out of memory during sentence_transformers encode sequence length batch size

Timestamps and exact values vary across runs, but the pattern. An info-level start, an early WARN, an ERROR carrying the symptom. Is the actual fingerprint you should alert on. The Denpex platform flags this combination automatically.

reduce encode batch size and cap or bucket sequence length, then repeat the same documents with no unrelated GPU process resident. This changes the condition in the causal diagnosis instead of hiding the outer exception. The repeated control proves ownership before recovery from the first uncommitted embedding batch after memory bounds are corrected.

# 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 -- "CUDA out of memory during sentence_transformers encode sequence length batch size" <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.

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.

CUDA reports errors asynchronously, so the traceback usually points at whatever line synchronised next rather than the one at fault. The hub covers every common CUDA error and how to make it report honestly.

• Sentence Transformers exceeded VRAM during encoding because the token count, batch size, output handling, and other GPU residents exceeded the model memory budget. Sequence length can dominate even when the document count looks small.
• The decisive evidence is the first log line that precedes "CUDA out of memory during sentence_transformers encode sequence length batch size" and differs from a healthy run.
• A nearby failure remains a competing hypothesis until a control separates configuration, capacity, transport, and hardware causes.