AM5728: TIDL conversion issues

Hi,

Thank you for your patience.  We are looking into your question.

Regards,
Mike

Hi,

Thank you very much for your kindness.

I really appreciate your help.

We are waiting for the answer.

Best regards,

Hello - I'm currently working on replicating on my EVM. Did you do the training and setup on a Windows or Linux machine? If Linux, which distribution? Just making sure that I'm replicating correctly.

Cheers,

Caleb

Hi,

Thank you very much for your kindness.

I really appreciate your help.

I use Linux(ubuntu-18.04.2).

We are waiting for the answer/guide to solve the problem.

Best regards,

Hi,

Sorry, it seemed to lack of information.

I'd like to send the detail again on my previous post.

   - I also use windows10 anaconda (Anaconda Navigator 1.9.7) in python(3.5.6) operation

Here is detail:

   windows10 (anaconda)> activate tf1.1_env

      Download from https://github.com/tensorflow/models

   windows10 (anaconda)>cd inception\inception

   windows10 (anaconda)>python freeze_inception_net_v1.py ( copy from python_freeze_graph.zip )

   (copy from ..Lib\site-packages\tensorflow\python\tools\optimize_for_inference.py )

   windows10 (anaconda)>python optimize_for_inference.py --input=inception_v1.pb --output=inceptionnet_v1.pb --input_names=input --output_name=softmax/Softmax

Then, convert to tidl files on TMDSEVM572X

I use Linux(ubuntu-18.04.2) with TFTP-boot

  evm> tidl_model_import.out.exe tidl_import_inceptionNetv1.txt

Best regards,

Hi,

Thank you very much for your kindness.

I really appreciate your help.

I understand you also found the same situation on your side.

And it seems to be used TIDL-unsupport operation such as L2 (L2's require square root operations).

Question:

  Could you tell us the successful procedure with below E2E-thread?

  https://e2e.ti.com/support/processors/f/791/t/735984

We would appreciate if you fix the TIDL conversion detail procedure of below E2E-thread.

  https://e2e.ti.com/support/processors/f/791/t/735984

Download inception_v1.ckpt from :

•Reset Models repo to :5eab06018c51d0ec8bf85919bf50438e2c4056a3 - MobileNet V1 commit (#1551)

•Download attached script freeze_inception_net_v1.py and copy it to the directory pointed by the environment variable %TF_TOOLS_1.1%

•Open the script freeze_inception_net_v1.py and verify that the variable models_slim_dir is set to os.getenv('TF_MODELS')+"/slim" .

•Open an Anaconda Prompt

•Type: activate tf1.1_env

•Type: cd %TF_MODELS%\inception\inception

•Type: python "%TF_TOOLS_1.1%\freeze_inception_net_v1.py"

•Type: python "%TF_TOOLS_1.1%\optimize_for_inference.py" --input=inception_v1.pb --output=inceptionnet_v1.pb --input_names=input --output_name=softmax/Softmax

•Type: xcopy inceptionnet_v1.pb %TI_DL_IMPORT%\test\testvecs\config\tensorflow_models\inceptionnet_V1\inceptionnet_v1.pb

•Type: cd %TI_DL_IMPORT%\utils\tidlModelImport

•Type: tidl_model_import.out.exe ..\..\test\testvecs\config\import\tidl_import_inceptionNetv1.txt

•Verify the presence of files tidl_inception_v1_net.bin and tidl_inception_v1_param.bin and their timestamps in %TI_DL_IMPORT%\test\testvecs\config\tidl_models

•Copy these files into %TI_DL% directory: xcopy %TI_DL_IMPORT%\test\testvecs\config\tidl_models\tidl_inception_v1_* %TI_DL%\test\testvecs\config\tidl_models\

•Test the model by running the inference in host emulation, using the default configuration for inception net. For that edit the first 2 lines %TI_DL%\test\testvecs\config\config_list.txt:

1 ..\..\test\testvecs\config\infer\tidl_config_inceptionNetv1.txt

0

•Type: cd %TI_DL%\test\out

•Type: dsp_test_dl_algo.out.exe

Verify that the log finishes with: Layer 82 :-------Max Index 896 : 254 ------- TSC Cycles = 0.00

Best regards,