Tool/software: TI-RTOS
Hi,
I'm trying to convert TensorFlow-slim based MobileNet model(ver1) following TIDLDeepLearningLibrary.pdf section 3.6.5.
I downloaded mobilenet_v1_1.0_224.ckpt.data-00000-of-00001, mobilenet_v1_1.0_224.ckpt.index, mobilenet_v1_1.0_224.ckpt.meta, mobilenet_v1_1.0_224.pb.
And i ran the command to optimize mobilenet_v1_1.0_224.pb.
However, these warning messages were shown.
--------------------------------------------------------------------
WARNING:tensorflow:From /usr/local/lib/python3.4/dist-packages/tensorflow/python/tools/strip_unused_lib.py:86: extract_sub_graph (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.compat.v1.graph_util.extract_sub_graph
WARNING:tensorflow:From /usr/local/lib/python3.4/dist-packages/tensorflow/python/tools/optimize_for_inference_lib.py:113: remove_training_nodes (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.compat.v1.graph_util.remove_training_nodes
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_1_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_2_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_3_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_4_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_5_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_6_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_7_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_8_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_9_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_10_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_11_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_12_depthwise/BatchNorm/FusedBatchNorm'
WARNING:tensorflow:Didn't find expected Conv2D input to 'MobilenetV1/MobilenetV1/Conv2d_13_depthwise/BatchNorm/FusedBatchNorm'
--------------------------------------------------------------------
Anyway, i ignored that messages and used result pb file to convert to TIDL model.
After i executed tidl_model_import.out.exe D:\TI_TDA2x\ti_dl\test\testvecs\config\import\tidl_import_mobileNet1.txt, this error message was shown.
--------------------------------------------------------------------
TF Model File : ..\..\test\testvecs\config\tensorflow_models\mobilenet\mobilenet_1_224.pb
Could not find the requested input Data : MobilenetV1/MobilenetV1/Conv2d_0/Relu6 !!
--------------------------------------------------------------------
Doesn't TIDL importing tool support Relu6 layer?
If then, could you let me know how could you validated TF MobileNet V1 model?
Thanks in advance.
Yoo.
