Profile Job Results

Jobs
jgz7mxz4p
Results Ready
Name
opus_mt_zh_en_OpusMTDecoder
Target Device
  • Dragonwing IQ-9075 EVK
  • Qualcomm Linux 1.7
  • Qualcomm® Dragonwing™ IQ-9075 | QCS9075
Creator
ai-hub-support@qti.qualcomm.com
Target Model
Input Specs
input_ids: int32[1, 1]
encoder_attention_mask: int32[1, 256]
position: int32[1]
block_0_past_self_key_states: float32[1, 8, 255, 64]
block_0_past_self_value_states: float32[1, 8, 255, 64]
block_0_cross_key_states: float32[1, 8, 256, 64]
block_0_cross_value_states: float32[1, 8, 256, 64]
block_1_past_self_key_states: float32[1, 8, 255, 64]
block_1_past_self_value_states: float32[1, 8, 255, 64]
block_1_cross_key_states: float32[1, 8, 256, 64]
block_1_cross_value_states: float32[1, 8, 256, 64]
block_2_past_self_key_states: float32[1, 8, 255, 64]
block_2_past_self_value_states: float32[1, 8, 255, 64]
block_2_cross_key_states: float32[1, 8, 256, 64]
block_2_cross_value_states: float32[1, 8, 256, 64]
block_3_past_self_key_states: float32[1, 8, 255, 64]
block_3_past_self_value_states: float32[1, 8, 255, 64]
block_3_cross_key_states: float32[1, 8, 256, 64]
block_3_cross_value_states: float32[1, 8, 256, 64]
block_4_past_self_key_states: float32[1, 8, 255, 64]
block_4_past_self_value_states: float32[1, 8, 255, 64]
block_4_cross_key_states: float32[1, 8, 256, 64]
block_4_cross_value_states: float32[1, 8, 256, 64]
block_5_past_self_key_states: float32[1, 8, 255, 64]
block_5_past_self_value_states: float32[1, 8, 255, 64]
block_5_cross_key_states: float32[1, 8, 256, 64]
block_5_cross_value_states: float32[1, 8, 256, 64]
Completion Time
2/22/2026, 3:43:00 AM
Versions
  • TensorFlow Lite: 2.17.0
  • QAIRT: v2.43.0.260127150333_193827
  • QNN TfLite Delegate: v2.43.0.260127150333_193827
  • Qualcomm Linux: 1.7-ver.1.1
  • AI Hub: aihub-2026.02.14.0
Estimated Inference Time
4.21 ms
Estimated Peak Memory Usage
0 ‑ 186 MB
Compute Units
NPU
418
StageTimeMemory
First App Load
7.06 s564‑565 MB
Subsequent App Load
299 ms173‑174 MB
Inference
4.21 ms0‑186 MB
TensorFlow LiteValue
number_of_threads4
QNN DelegateValue
backend_typekHtpBackend
log_levelkLogLevelWarn
htp_options.performance_modekHtpBurst
htp_options.precisionkHtpFp16
htp_options.optimization_strategykHtpOptimizeForInference
htp_options.useConvHmxtrue

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

Run on device