Profile Job Results
Jobs
j5wed40m5
Results Ready
Name
whisper_tiny_en_WhisperDecoder
Target Device
- SA8295P ADP
- Android 14
- Qualcomm® SA8295P
Creator
ai-hub-support@qti.qualcomm.com
Target Model
Input Specs
x
: int32[1, 1]index
: int32[1, 1]k_cache_cross
: float32[4, 6, 64, 1500]v_cache_cross
: float32[4, 6, 1500, 64]k_cache_self
: float32[4, 6, 64, 224]v_cache_self
: float32[4, 6, 224, 64]Completion Time
11/10/2024, 6:16:51 AM
Versions
- TensorFlow Lite: 2.17.0
- QNN TfLite Delegate: v2.27.0.240926142112_100894-auto
- Android: 14 (UQ1A.240205.002)
- Build ID: SA8295P.HQX.4.5.6.0-00006-STD.PROD-1
- AI Hub: aihub-2024.10.28.0
Estimated Inference Time
4.79 ms
Estimated Peak Memory Usage
3 ‑ 29 MB
Compute Units
NPU
557
Stage | Time | Memory |
---|---|---|
First App Load | 2.95 s | 559‑565 MB |
Subsequent App Load | 267 ms | 94‑121 MB |
Inference | 4.79 ms | 3‑29 MB |
TensorFlow Lite | Value |
---|---|
number_of_threads | 4 |
QNN Delegate | Value |
---|---|
backend_type | kHtpBackend |
log_level | kLogLevelWarn |
htp_options.precision | kHtpFp16 |
htp_options.useConvHmx | true |
Sign up to run this model on a hosted Qualcomm® device!
Run on device