RTX PRO 6000 Blackwell Server Edition · 96 GB
granite-speech-5.0-470m-turboctc
Open a report for complete measurements.
- Task
- Speech recognition · Audio → Text
- Workload
- Audio → Text · asr_prompts · Test duration 16 s
- Batch size
- Batch 256
- Format · Engine
- Native · transformers-ctc · PyTorch 2.13.0
- Unit
- audio-sec/sec
- Mean throughput
- 8,343audio-sec/sec
- Observed GPU occupancy
- 20.1 GiB
- P95 latency
- Not recorded
- Reports combined
- 1
Launch this model
This engine does not yet have an executable serving recipe.
Individual tests
Audio → Text · Speech recognition · Native · transformers-ctc · asr · audio-sec/sec
Sortable table
| Test date | Throughput | Concurrency / batch | P95 (s) | TTFT (ms) | Observed occupancy (GiB) | GPU power (W) | Decode (tok/s) | Uncached prefill (tok/s) | Parameters | View report |
|---|---|---|---|---|---|---|---|---|---|---|
| 10/19/2026, 8:09:36 AM | 8,343 audio-sec/sec | 256 | — | — | 20.09 | 477.57 | — | Uncached prefill unavailableView report | Parameters ↗ | 3be7f67e ↗ |
Recorded test parameters
Configuration 1 · Parameters
- model
- ibm-granite/granite-speech-5.0-470m-turboctc
- engine
- transformers-ctc
- quant
- Native
- method
- audio-native-v1
- task
- asr
- dataset
- asr_prompts
- context
- asr
- input_tokens
- 0
- output_tokens
- 0
- duration
- 16
- concurrency
- 256
- unit
- audio-sec/sec
- software
- { "torch": "2.13.0", "transformers": "5.20.1", "torchaudio": "2.11.0", "OMP_NUM_THREADS": "8", "MKL_NUM_THREADS": "8", "OPENBLAS_NUM_THREADS": "8", "NUMEXPR_NUM_THREADS": "8" }
- engine_provenance
- —
- kernel_settings
- —
- precision
- —
- resolution
- —
- performance_task
- —
- performance_scope
- —
- batch_size
- 256
- dimensions
- —
- weight_quantization
- —
- max_sequence_length
- —
- model_scope
- —
- performance_dataset
- —
- workload_mode
- —
- graph_mode
- —
- model_revision
- 947f59af40db9791170a0628cf0f3f4812d720f1
- client_version
- python-native-bench-1.1.3