Dear all,
I am using OpenVINO toolkit release 2020.1 in Ubuntu 16. I got the runtime error as below while try to inference a LSTM-based model in MYRIAD plugin:
Traceback (most recent call last):
File "inference.py", line 54, in <module>
sys.exit(main() or 0)
File "inference.py", line 34, in main
exec_net, input_blob = load_network(args)
File "/home/majinbuu/Documents/asr-ncs/asr/utils_inference.py", line 45, in load_network
exec_net = ie.load_network(network=net, device_name=args.device)
File "ie_api.pyx", line 134, in openvino.inference_engine.ie_api.IECore.load_network
File "ie_api.pyx", line 141, in openvino.inference_engine.ie_api.IECore.load_network
RuntimeError: Check (currentLoopAttributes == nullptr) failed: Nested loops are not supported yet
Best regards,
HA Minh Quyet
Link Copied
Hi HA Minh Quyet,
Can you please provide your model so we can take a look at it. Also, what framework was your model trained in? Is it a custom or pre-trained model? Can you provide the commands you used to convert your model and the commands you used to deploy the model in the Inference Engine?
Regards,
Randall B.
Dear Randall,
I used Tensorflow to train my model. It is a custom model based on LSTM cell. Is there any private way to upload the model? Because I am not allowed to public the model. The command used to convert my model as follows:
python /opt/intel/openvino/deployment_tools/model_optimizer/mo_tf.py --input_model model/tf_model/asr_horovod_nchw-12.pb --output_dir model/ir_model/ --model_name asr_VPU_ir_r3_12 --data_type FP16 --log_level DEBUG --input_shape [1,1,1000,161] --disable_nhwc_to_nchw
Best regards,
HA Minh Quyet
HA Minh Quyet,
I have sent you a PM, please attach your model there and we can take a look at it.
Best Regards,
Sahira
For more complete information about compiler optimizations, see our Optimization Notice.