Qualcomm® AI HubAI Hub

Profile Job Results

Jobs
jgo9mm31g
Results Ready
Name
whisper_medium_decoder
Target Device
  • Snapdragon X Elite CRD
  • Windows 11
  • Snapdragon® X Elite | SC8380XP
Creator
ai-hub-support@qti.qualcomm.com
Input Specs
input_ids: int32[1, 1]
position_ids: int32[1]
k_cache_self_0_in: float16[16, 1, 64, 199]
attention_mask: float16[1, 1, 1, 200]
v_cache_self_0_in: float16[16, 1, 199, 64]
k_cache_cross_0: float16[16, 1, 64, 1500]
v_cache_cross_0: float16[16, 1, 1500, 64]
k_cache_self_1_in: float16[16, 1, 64, 199]
v_cache_self_1_in: float16[16, 1, 199, 64]
k_cache_cross_1: float16[16, 1, 64, 1500]
v_cache_cross_1: float16[16, 1, 1500, 64]
k_cache_self_2_in: float16[16, 1, 64, 199]
v_cache_self_2_in: float16[16, 1, 199, 64]
k_cache_cross_2: float16[16, 1, 64, 1500]
v_cache_cross_2: float16[16, 1, 1500, 64]
k_cache_self_3_in: float16[16, 1, 64, 199]
v_cache_self_3_in: float16[16, 1, 199, 64]
k_cache_cross_3: float16[16, 1, 64, 1500]
v_cache_cross_3: float16[16, 1, 1500, 64]
k_cache_self_4_in: float16[16, 1, 64, 199]
v_cache_self_4_in: float16[16, 1, 199, 64]
k_cache_cross_4: float16[16, 1, 64, 1500]
v_cache_cross_4: float16[16, 1, 1500, 64]
k_cache_self_5_in: float16[16, 1, 64, 199]
v_cache_self_5_in: float16[16, 1, 199, 64]
k_cache_cross_5: float16[16, 1, 64, 1500]
v_cache_cross_5: float16[16, 1, 1500, 64]
k_cache_self_6_in: float16[16, 1, 64, 199]
v_cache_self_6_in: float16[16, 1, 199, 64]
k_cache_cross_6: float16[16, 1, 64, 1500]
v_cache_cross_6: float16[16, 1, 1500, 64]
k_cache_self_7_in: float16[16, 1, 64, 199]
v_cache_self_7_in: float16[16, 1, 199, 64]
k_cache_cross_7: float16[16, 1, 64, 1500]
v_cache_cross_7: float16[16, 1, 1500, 64]
k_cache_self_8_in: float16[16, 1, 64, 199]
v_cache_self_8_in: float16[16, 1, 199, 64]
k_cache_cross_8: float16[16, 1, 64, 1500]
v_cache_cross_8: float16[16, 1, 1500, 64]
k_cache_self_9_in: float16[16, 1, 64, 199]
v_cache_self_9_in: float16[16, 1, 199, 64]
k_cache_cross_9: float16[16, 1, 64, 1500]
v_cache_cross_9: float16[16, 1, 1500, 64]
k_cache_self_10_in: float16[16, 1, 64, 199]
v_cache_self_10_in: float16[16, 1, 199, 64]
k_cache_cross_10: float16[16, 1, 64, 1500]
v_cache_cross_10: float16[16, 1, 1500, 64]
k_cache_self_11_in: float16[16, 1, 64, 199]
v_cache_self_11_in: float16[16, 1, 199, 64]
k_cache_cross_11: float16[16, 1, 64, 1500]
v_cache_cross_11: float16[16, 1, 1500, 64]
k_cache_self_12_in: float16[16, 1, 64, 199]
v_cache_self_12_in: float16[16, 1, 199, 64]
k_cache_cross_12: float16[16, 1, 64, 1500]
v_cache_cross_12: float16[16, 1, 1500, 64]
k_cache_self_13_in: float16[16, 1, 64, 199]
v_cache_self_13_in: float16[16, 1, 199, 64]
k_cache_cross_13: float16[16, 1, 64, 1500]
v_cache_cross_13: float16[16, 1, 1500, 64]
k_cache_self_14_in: float16[16, 1, 64, 199]
v_cache_self_14_in: float16[16, 1, 199, 64]
k_cache_cross_14: float16[16, 1, 64, 1500]
v_cache_cross_14: float16[16, 1, 1500, 64]
k_cache_self_15_in: float16[16, 1, 64, 199]
v_cache_self_15_in: float16[16, 1, 199, 64]
k_cache_cross_15: float16[16, 1, 64, 1500]
v_cache_cross_15: float16[16, 1, 1500, 64]
k_cache_self_16_in: float16[16, 1, 64, 199]
v_cache_self_16_in: float16[16, 1, 199, 64]
k_cache_cross_16: float16[16, 1, 64, 1500]
v_cache_cross_16: float16[16, 1, 1500, 64]
k_cache_self_17_in: float16[16, 1, 64, 199]
v_cache_self_17_in: float16[16, 1, 199, 64]
k_cache_cross_17: float16[16, 1, 64, 1500]
v_cache_cross_17: float16[16, 1, 1500, 64]
k_cache_self_18_in: float16[16, 1, 64, 199]
v_cache_self_18_in: float16[16, 1, 199, 64]
k_cache_cross_18: float16[16, 1, 64, 1500]
v_cache_cross_18: float16[16, 1, 1500, 64]
k_cache_self_19_in: float16[16, 1, 64, 199]
v_cache_self_19_in: float16[16, 1, 199, 64]
k_cache_cross_19: float16[16, 1, 64, 1500]
v_cache_cross_19: float16[16, 1, 1500, 64]
k_cache_self_20_in: float16[16, 1, 64, 199]
v_cache_self_20_in: float16[16, 1, 199, 64]
k_cache_cross_20: float16[16, 1, 64, 1500]
v_cache_cross_20: float16[16, 1, 1500, 64]
k_cache_self_21_in: float16[16, 1, 64, 199]
v_cache_self_21_in: float16[16, 1, 199, 64]
k_cache_cross_21: float16[16, 1, 64, 1500]
v_cache_cross_21: float16[16, 1, 1500, 64]
k_cache_self_22_in: float16[16, 1, 64, 199]
v_cache_self_22_in: float16[16, 1, 199, 64]
k_cache_cross_22: float16[16, 1, 64, 1500]
v_cache_cross_22: float16[16, 1, 1500, 64]
k_cache_self_23_in: float16[16, 1, 64, 199]
v_cache_self_23_in: float16[16, 1, 199, 64]
k_cache_cross_23: float16[16, 1, 64, 1500]
v_cache_cross_23: float16[16, 1, 1500, 64]
Completion Time
11/14/2026, 3:35:16 PM
Options
--qairt_version latest
Versions
  • ONNX Runtime: 1.30.0
  • QNN Execution Provider: 2.5.0
  • QAIRT: v2.51.0.260929181845
  • Windows: Windows 11 (26200)
  • Build ID: APSS.WP_HA.1.0-09400-SC8380XRELSFNWZA-8
  • AI Hub: aihub-2026.11.06.1
Estimated Inference Time
29.0 ms
Estimated Peak Memory Usage
726 MB
Compute Units
NPU
6011
StageTimeMemory
First App Load
1.76 s914 MB
Subsequent App Load
1.57 s913 MB
Inference
29.0 ms726 MB
ONNX RuntimeValue
execution_modeSEQUENTIAL
intra_op_num_threads0
inter_op_num_threads0
enable_memory_patternfalse
enable_cpu_memory_arenafalse
graph_optimization_levelENABLE_ALL
QNN Execution ProviderValue
htp_performance_mode"burst"
htp_graph_finalization_optimization_mode"3"
enable_htp_fp16_precision"1"
capture_network_visualizationsfalse
context_priority"normal"
offload_graph_io_quantization"1"
num_graph_prepare_threads"1"

Sign up to run this model on a hosted Qualcomm® device!

Run on device