• State Resolved
  • Locked Locked
  • Replies 1 reply
  • Subscribers 97 subscribers
  • Views 498 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Original question:

TDA2EXEVM: TIDL power layer

TDA2EXEVM: Support for LSTM and RNN to enable Speech Recognition usingTIDL

SagarK
Intellectual 705 points

Part Number: TDA2EXEVM

Hi Kumar,

We have TD2X EVM board. And till now we were working on vision side and we understood how one can use any existing Deep Learning model for object detection like face detection, car detection etc using TIDL.

Now, I would like to know that if i want to add ASR (Automatic Speech Recognition) on TDA2X .What should be the approach? I understand that TIDL should support LSTM, RNN layer. Because, these layers are at core of NLP and speech recognition.

Somewhere in documentation, I read that support for these layers is future scope (http://software-dl.ti.com/processor-sdk-linux/esd/docs/latest/linux/Foundational_Components_TIDL.html)

1. Can you comment on the support for LSTM and RNN in TIDL ?

2. Is there any other existing approach of speech recognition which we can use on TDA2X ?

3. Any examples or demos available ?

4. Can you suggest any speech recognition approach other that ML or DL, for quick prototyping on TDA2X ?

Regards,

Sagar

  • 0
    TI__Genius 17580 points

    Hi Sagar,

    1. As of now, we don't have any plans for supporting LSTM and RNN in TIDL on TDA2.

    2. No, we don’t have solution using TIDL.

    3. No, we don't have any demos.

    4. User can use tensorflow Lite or ARM NN in the Linux running on A15

    Thanks,

    Praveen