Profile Job Results
Jobs
jp164w7k5
Results Ready
Name
pi05_backbone
Target Device
- QCS8550 (Proxy)
- Android 12
- Qualcomm® Dragonwing™ QCS8550
Creator
ai-hub-support@qti.qualcomm.com
Input Specs
hidden_state: float32[1, 968, 2048]prefix_att_2d_masks: float32[1, 1, 968, 968]rope_emb_sin: float32[1, 968, 1, 128]rope_emb_cos: float32[1, 968, 1, 128]Completion Time
7/25/2026, 9:23:05 AM
Versions
- ONNX Runtime: 1.27.1
- QAIRT: v2.45.0.260326154327
- Android: 14 (UP1A.231005.007)
- AI Hub: aihub-2026.07.17.0
Estimated Inference Time
462 ms
Estimated Peak Memory Usage
73 ‑ 75 MB
Compute Units
NPU
2473
| Stage | Time | Memory |
|---|---|---|
First App Load | 3.52 s | 924‑925 MB |
Subsequent App Load | 3.19 s | 935‑937 MB |
Inference | 462 ms | 73‑75 MB |
| ONNX Runtime | Value |
|---|---|
| execution_mode | SEQUENTIAL |
| intra_op_num_threads | 0 |
| inter_op_num_threads | 0 |
| enable_memory_pattern | false |
| enable_cpu_memory_arena | false |
| graph_optimization_level | ENABLE_ALL |
| QNN Execution Provider | Value |
|---|---|
| htp_performance_mode | "burst" |
| htp_graph_finalization_optimization_mode | "3" |
| enable_htp_fp16_precision | "1" |
| capture_network_visualizations | false |
| context_priority | "normal" |
| offload_graph_io_quantization | "1" |
Sign up to run this model on a hosted Qualcomm® device!
Run on device