Qualcomm® AI HubAI Hub

Profile Job Results

Jobs
j5786l2np
Results Ready
Name
whisper_medium_quantized_decoder
Target Device
  • Snapdragon X2 Elite CRD
  • Windows 11
  • Snapdragon® X2 Elite | SC8480XP
Creator
ai-hub-support@qti.qualcomm.com
Input Specs
input_ids: int32[1, 1]
attention_mask: uint16[1, 1, 1, 200]
k_cache_self_0_in: uint8[16, 1, 64, 199]
v_cache_self_0_in: uint8[16, 1, 199, 64]
k_cache_self_1_in: uint8[16, 1, 64, 199]
v_cache_self_1_in: uint8[16, 1, 199, 64]
k_cache_self_2_in: uint8[16, 1, 64, 199]
v_cache_self_2_in: uint8[16, 1, 199, 64]
k_cache_self_3_in: uint8[16, 1, 64, 199]
v_cache_self_3_in: uint8[16, 1, 199, 64]
k_cache_self_4_in: uint8[16, 1, 64, 199]
v_cache_self_4_in: uint8[16, 1, 199, 64]
k_cache_self_5_in: uint8[16, 1, 64, 199]
v_cache_self_5_in: uint8[16, 1, 199, 64]
k_cache_self_6_in: uint8[16, 1, 64, 199]
v_cache_self_6_in: uint8[16, 1, 199, 64]
k_cache_self_7_in: uint8[16, 1, 64, 199]
v_cache_self_7_in: uint8[16, 1, 199, 64]
k_cache_self_8_in: uint8[16, 1, 64, 199]
v_cache_self_8_in: uint8[16, 1, 199, 64]
k_cache_self_9_in: uint8[16, 1, 64, 199]
v_cache_self_9_in: uint8[16, 1, 199, 64]
k_cache_self_10_in: uint8[16, 1, 64, 199]
v_cache_self_10_in: uint8[16, 1, 199, 64]
k_cache_self_11_in: uint8[16, 1, 64, 199]
v_cache_self_11_in: uint8[16, 1, 199, 64]
k_cache_self_12_in: uint8[16, 1, 64, 199]
v_cache_self_12_in: uint8[16, 1, 199, 64]
k_cache_self_13_in: uint8[16, 1, 64, 199]
v_cache_self_13_in: uint8[16, 1, 199, 64]
k_cache_self_14_in: uint8[16, 1, 64, 199]
v_cache_self_14_in: uint8[16, 1, 199, 64]
k_cache_self_15_in: uint8[16, 1, 64, 199]
v_cache_self_15_in: uint8[16, 1, 199, 64]
k_cache_self_16_in: uint8[16, 1, 64, 199]
v_cache_self_16_in: uint8[16, 1, 199, 64]
k_cache_self_17_in: uint8[16, 1, 64, 199]
v_cache_self_17_in: uint8[16, 1, 199, 64]
k_cache_self_18_in: uint8[16, 1, 64, 199]
v_cache_self_18_in: uint8[16, 1, 199, 64]
k_cache_self_19_in: uint8[16, 1, 64, 199]
v_cache_self_19_in: uint8[16, 1, 199, 64]
k_cache_self_20_in: uint8[16, 1, 64, 199]
v_cache_self_20_in: uint8[16, 1, 199, 64]
k_cache_self_21_in: uint8[16, 1, 64, 199]
v_cache_self_21_in: uint8[16, 1, 199, 64]
k_cache_self_22_in: uint8[16, 1, 64, 199]
v_cache_self_22_in: uint8[16, 1, 199, 64]
k_cache_self_23_in: uint8[16, 1, 64, 199]
v_cache_self_23_in: uint8[16, 1, 199, 64]
k_cache_cross_0: uint8[16, 1, 64, 1500]
v_cache_cross_0: uint8[16, 1, 1500, 64]
k_cache_cross_1: uint8[16, 1, 64, 1500]
v_cache_cross_1: uint8[16, 1, 1500, 64]
k_cache_cross_2: uint8[16, 1, 64, 1500]
v_cache_cross_2: uint8[16, 1, 1500, 64]
k_cache_cross_3: uint8[16, 1, 64, 1500]
v_cache_cross_3: uint8[16, 1, 1500, 64]
k_cache_cross_4: uint8[16, 1, 64, 1500]
v_cache_cross_4: uint8[16, 1, 1500, 64]
k_cache_cross_5: uint8[16, 1, 64, 1500]
v_cache_cross_5: uint8[16, 1, 1500, 64]
k_cache_cross_6: uint8[16, 1, 64, 1500]
v_cache_cross_6: uint8[16, 1, 1500, 64]
k_cache_cross_7: uint8[16, 1, 64, 1500]
v_cache_cross_7: uint8[16, 1, 1500, 64]
k_cache_cross_8: uint8[16, 1, 64, 1500]
v_cache_cross_8: uint8[16, 1, 1500, 64]
k_cache_cross_9: uint8[16, 1, 64, 1500]
v_cache_cross_9: uint8[16, 1, 1500, 64]
k_cache_cross_10: uint8[16, 1, 64, 1500]
v_cache_cross_10: uint8[16, 1, 1500, 64]
k_cache_cross_11: uint8[16, 1, 64, 1500]
v_cache_cross_11: uint8[16, 1, 1500, 64]
k_cache_cross_12: uint8[16, 1, 64, 1500]
v_cache_cross_12: uint8[16, 1, 1500, 64]
k_cache_cross_13: uint8[16, 1, 64, 1500]
v_cache_cross_13: uint8[16, 1, 1500, 64]
k_cache_cross_14: uint8[16, 1, 64, 1500]
v_cache_cross_14: uint8[16, 1, 1500, 64]
k_cache_cross_15: uint8[16, 1, 64, 1500]
v_cache_cross_15: uint8[16, 1, 1500, 64]
k_cache_cross_16: uint8[16, 1, 64, 1500]
v_cache_cross_16: uint8[16, 1, 1500, 64]
k_cache_cross_17: uint8[16, 1, 64, 1500]
v_cache_cross_17: uint8[16, 1, 1500, 64]
k_cache_cross_18: uint8[16, 1, 64, 1500]
v_cache_cross_18: uint8[16, 1, 1500, 64]
k_cache_cross_19: uint8[16, 1, 64, 1500]
v_cache_cross_19: uint8[16, 1, 1500, 64]
k_cache_cross_20: uint8[16, 1, 64, 1500]
v_cache_cross_20: uint8[16, 1, 1500, 64]
k_cache_cross_21: uint8[16, 1, 64, 1500]
v_cache_cross_21: uint8[16, 1, 1500, 64]
k_cache_cross_22: uint8[16, 1, 64, 1500]
v_cache_cross_22: uint8[16, 1, 1500, 64]
k_cache_cross_23: uint8[16, 1, 64, 1500]
v_cache_cross_23: uint8[16, 1, 1500, 64]
position_ids: int32[1]
Completion Time
11/14/2026, 11:03:55 AM
Options
--qairt_version latest
Versions
  • ONNX Runtime: 1.30.0
  • QNN Execution Provider: 2.5.0
  • QAIRT: v2.51.0.260929181845
  • Windows: Windows 11 (28000)
  • Build ID: APSS.WP_GL.1.0.r0-05000-SC8480XRELSFNWZA-4
  • AI Hub: aihub-2026.11.06.1
Estimated Inference Time
10.7 ms
Estimated Peak Memory Usage
80 MB
Compute Units
NPU
7425
StageTimeMemory
First App Load
503 ms32 MB
Subsequent App Load
430 ms31 MB
Inference
10.7 ms80 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