Repository navigation
Conversation
Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
…M62D setup Signed-off-by: JuneChul Roh <jroh@ti.com>
Audio AI docs prefix every shell command with "$ "; configure copybutton to strip it so copied text is directly runnable. Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
…tion. Signed-off-by: JuneChul Roh <jroh@ti.com>
…l zoo URL Rename source/devices/AM62DX/audioai to edgeai and the AM62DX_audioai_* config files to AM62DX_edgeai_*. Point __MODELZOO_REPO_URL__ and the model zoo clone command to TexasInstruments/audioai-modelzoo. Signed-off-by: JuneChul Roh <jroh@ti.com>
Split into "Model Compile and Evaluate" (model_compile_evaluate.rst) and "Model Inspect" (model_inspect.rst). Update the toctree, the AM62DX edgeai toc config, labels, and cross-references in the other pages. Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
Signed-off-by: JuneChul Roh <jroh@ti.com>
Trim the Vale vocabulary to the entries Vale needs, fix heading adornments and non-ASCII text, use isonum.txt and #. autonumbered steps, and replace the removed replacement variables with literal values. Signed-off-by: JuneChul Roh <jroh@ti.com>
…mentation Add comprehensive documentation for AM62D EdgeAI audio AI applications: - overview.rst: Platform overview, hardware architecture, and key features - getting_started_platform.rst: Hardware setup, network connection, and audio device configuration - getting_started_gstreamer.rst: GStreamer pipeline examples and BYOM guide - linux_audio_inference_app.rst: C++ reference application build and deployment - audio_ai_model_workflow.rst: Renamed from introduction.rst for clarity Add supporting images for GUI demos and SDK architecture diagrams. Update table of contents with new sections. Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
Add comprehensive documentation for the Custom Audio GStreamer Pipeline demo, including step-by-step guides for: - GCRN-based speech enhancement pipeline with visualization - YAMNet audio classification pipeline with top-3 predictions - VGGish audio classification pipeline with class detection Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
Add comprehensive guide for extending GStreamer plugins to support new model architectures. Includes detailed code snippets from actual plugin source files: - gstdspkernel.cpp: Adding new properties to tidspkernel element - gsttitvm.cpp: Adding new properties to titvm element Sections added: - Supporting New Model Architectures in the Plugins - Adding a New GStreamer Property to tidspkernel - Adding a New Property to titvm - Automatic Model Detection from Path - Example: Using a New Model with Extended Plugin Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
Improve clarity and technical accuracy in Bring-Your-Own Model (BYOM) sections All changes improve readability without altering technical content. Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
and check page build Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
The deploy workflow needs pages:write and id-token:write permissions at the root level to properly authenticate with GitHub Pages API. Previously these were only set on the deploy job, which wasn't sufficient. This should fix the 'Missing environment' error when creating the deployment. Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
…rding Add a new high-level guide (adding_new_model.rst) walking through the steps needed to integrate a new model into the AM62DX Edge AI inference chain: TVM compilation, DSP firmware signal-chain wiring, the Linux application, and the GStreamer plugin. Link it into the edgeai/index.rst toctree after model_inspect. Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
… link Add linux/index_Edge_AI to the AM62DX linux TOC and device index toctree so the generic Edge AI landing page is reachable from the AM62DX linux SDK guide. Add an AM62DX_family ifconfig block pointing to the published AM62DX EdgeAI overview page (12_01_00_00), matching the pattern used for AM62AX and the Jacinto7 variants. Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
…improved organization Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
…nd fix RST hierarchy Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
v-singh1
reviewed
Oct 23, 2026
| :scale: 75 | ||
| :align: center | ||
|
|
||
| GStreamer data-flow for VGGish audio classification: input WAV → STFT on DSP → VGGish model inference on TVM → console output |
v-singh1
reviewed
Oct 23, 2026
| Processor SDK Linux Edge AI Software Developer’s Guide | ||
| ###################################################### | ||
|
|
||
| .. rubric:: Welcome to the AM62D EdgeAI/DL Software Developer's Guide |
v-singh1
reviewed
Oct 23, 2026
| of the `rpmsg-dma <https://github.com/TexasInstruments/rpmsg-dma>`__ repository. The | ||
| application enables: | ||
|
|
||
| - **Real-time Audio Processing** — Process live ALSA audio devices or WAV file inputs |
Contributor
There was a problem hiding this comment.
Process live ALSA audio devices -> Process live input or...
v-singh1
reviewed
Oct 23, 2026
| application enables: | ||
|
|
||
| - **Real-time Audio Processing** — Process live ALSA audio devices or WAV file inputs | ||
| - **DSP Offload** — Offload STFT, ISTFT and deinterleave/interleave to the C7x DSP over RPMsg-DMA |
Contributor
There was a problem hiding this comment.
Offload STFT, ISTFT and deinterleave/interleave to the C7x DSP over RPMsg-DMA
Offload DSP signalchains (ex :STFT, ISTFT and deinterleave/interleave) to the C7x DSP over RPMsg-DMA
v-singh1
reviewed
Oct 23, 2026
| - **TVM Inference** — Run TVM-compiled deep learning models on the C7x NPU | ||
| - **Flexible I/O** — WAV files, live ALSA capture, raw PCM streams, or binary tensors | ||
| - **JSON Pipeline Configuration** — Define processing chains via JSON files | ||
| - **Model Daemon Mode** — Persistent TVM model daemon to avoid repeated artifact loading |
Contributor
There was a problem hiding this comment.
A model daemon hot-reloads artifacts without an application rebuild
v-singh1
reviewed
Oct 23, 2026
| - Build system | ||
| * - ``make``, ``git`` | ||
| - Build utilities | ||
| * - Neo-TVM headers (``TVM_ROOT``) |
v-singh1
reviewed
Oct 23, 2026
| - ``build/example/edge-ai/rpmsg_inference_example`` — main inference application | ||
| - ``build/example/edge-ai/tvm_model_daemon`` — persistent model daemon | ||
|
|
||
| #. Install to a staging directory and transfer to the AM62D EVM: |
Contributor
There was a problem hiding this comment.
.service file copying to EVM is missing
v-singh1
reviewed
Oct 23, 2026
| GCRN based Speech Enhancement | ||
| ***************************** | ||
|
|
||
| Speech enhancement removes background noise from a recording while keeping the speech |
Contributor
There was a problem hiding this comment.
Refer GST plugin comments and update in this file as well
v-singh1
reviewed
Oct 23, 2026
|
|
||
| $ ./devices/setup_am62d.sh | ||
|
|
||
| It downloads the ARM GCC 15.2 and C7000 CGT 5.0.0.LTS cross-toolchains into |
Contributor
There was a problem hiding this comment.
C& compiler version need update, Please check with Junchul
v-singh1
reviewed
Oct 23, 2026
| .. _am62dx_edgeai_oob_demo: | ||
|
|
||
| ========================================= | ||
| Getting Started with the OOB Web GUI Demo |
Contributor
There was a problem hiding this comment.
Getting Started with the OOB Web GUI Demo
Getting Started with the OOB Demo using Web GUI
v-singh1
reviewed
Oct 23, 2026
| any additional setup. | ||
|
|
||
| The portal showcases the three core EdgeAI pipelines — home screen, speech | ||
| enhancement, and audio classification — running on the C7x DSP via the TVM runtime. |
Contributor
There was a problem hiding this comment.
Not only TVM runtime, so remove via the TVM runtime
v-singh1
reviewed
Oct 23, 2026
|
|
||
| AM62DX out-of-box web GUI home screen | ||
|
|
||
| This is the Sitara AM62D Edge AI Portal, a web-based user interface for managing and |
Contributor
There was a problem hiding this comment.
Correct the name it is DSP + EDGE AI Portal, please rerun the latest and check for correctness
v-singh1
reviewed
Oct 23, 2026
| This is the Sitara AM62D Edge AI Portal, a web-based user interface for managing and | ||
| evaluating Texas Instruments' AM62D Evaluation Module. The dashboard allows developers | ||
| to launch real-time audio analytics and DSP compute demos, inspect AI models via the | ||
| TIDL, and monitor live system telemetry like CPU load, memory footprints, and SDK software |
v-singh1
reviewed
Oct 23, 2026
| **GET STARTED** button that jumps straight to the Audio Intelligence demo. Below it, an | ||
| **Explore Demos And Tools** panel repeats the same three entry points (Audio Intelligence, DSP | ||
| Acceleration, AI Model Inspector) as clickable cards with a one-line description each, so a | ||
| first-time visitor does not need to use the sidebar at all to start exploring. |
Contributor
There was a problem hiding this comment.
repeats the same -> Opens or shows Audio Intelligence, DSP Acceleration, AI Model Inspector
remove first-time visitor parts
v-singh1
reviewed
Oct 23, 2026
| - **Device Info panel** — ``Device``, ``Board``, ``SoC`` (AM62Dx quad Cortex-A53 @ 1.4GHz + | ||
| C7x DSP), ``IP Address``, and system ``Uptime``. | ||
| - **Runtime Status panel** — live ``CPU Load``, ``CPU Avg``, ``CPU Peak``, and ``RAM | ||
| Used/Free``, refreshed continuously while the portal is open. |
v-singh1
reviewed
Oct 23, 2026
|
|
||
| Under ``Demo Selection`` on the left, three options are available: | ||
|
|
||
| - **Speech Enhancement** — C7x DSP noise reduction + TIDL. |
v-singh1
reviewed
Oct 23, 2026
| Processor SDK Linux Edge AI for AM62DX feature overview | ||
|
|
||
| The SDK enables an interplay of open-source components such as GStreamer and | ||
| deep learning runtimes such as TVM-RT. The SDK supports edit-build-debug |
Contributor
There was a problem hiding this comment.
Currently it supports only TVM RT, rephrase the sentence
v-singh1
reviewed
Oct 23, 2026
|
|
||
| The SDK enables an interplay of open-source components such as GStreamer and | ||
| deep learning runtimes such as TVM-RT. The SDK supports edit-build-debug | ||
| cycles directly on the target and also on a host PC to cross-compile and |
Contributor
There was a problem hiding this comment.
what is edit-build-debug cycles ?
v-singh1
reviewed
Oct 23, 2026
|
|
||
| - **Edge AI application stack** — Linux-side applications and GStreamer pipelines | ||
| that drive audio inference using TVM-RT and open-source components. | ||
| - **Foundational Linux components** — OS, U-Boot, kernel, filesystem, and Linux |
Add a new doc for EdgeAI and new model info. Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
…eline documentation - Add AM62D EVM hardware setup details with boot mode diagrams and new images - Update GStreamer pipeline documentation with improved clarity and trademark formatting - Refresh pipeline diagrams and optimize image sizes - Add missing user guide references and detailed system verification steps - Standardize RST formatting and reference anchors across EdgeAI documentation Signed-off-by: Pratham Deshmukh <p-deshmukh@ti.com>
v-singh1
reviewed
Oct 23, 2026
| that drive audio inference using TVM-RT and open-source components. | ||
| - **Foundational Linux components** — OS, U-Boot, kernel, filesystem, and Linux | ||
| drivers that form the base platform. | ||
| - **DSP firmware** — C7x DSP firmware that runs the signal chain (STFT/ISTFT) and |
Contributor
There was a problem hiding this comment.
signal chain (ex: STFT/ISTFT)
v-singh1
reviewed
Oct 23, 2026
| - **Foundational Linux components** — OS, U-Boot, kernel, filesystem, and Linux | ||
| drivers that form the base platform. | ||
| - **DSP firmware** — C7x DSP firmware that runs the signal chain (STFT/ISTFT) and | ||
| TVM model inference on the dedicated DSP hardware. |
Contributor
There was a problem hiding this comment.
what is dedicated DSP hardware.
v-singh1
reviewed
Oct 23, 2026
| | | | Version** | | Linux** | RT Linux** | RTOS** | | ||
| +--------------+-----------------+-----------+-----------------------+-------------------+------------------+------------------+ | ||
| | AM62D-EVM | AUDIO-AM62D-EVM | E2 | Hardware User's Guide | Y | Y | Y | | ||
| | AM62D-EVM | AM62D EVM | E2 | Hardware User's Guide | Y | Y | Y | |
Contributor
There was a problem hiding this comment.
why rename AUDIO-AM62D-EVM ?
|
New warnings found with rstcheck: |
|
New unreachable files found with check_files.py: |
v-singh1
reviewed
Oct 23, 2026
| ********************** | ||
|
|
||
| - `AM62D-EVM <https://www.ti.com/tool/AUDIO-AM62D-EVM>`__ | ||
| - `AM62D-EVM <https://www.ti.com/tool/AM62D EVM>`__ |
v-singh1
reviewed
Oct 23, 2026
|
|
||
| The SDK provides software and tools to let users run high-performance audio deep learning inference on Texas Instruments' AM62DX processor. We offer a practical embedded EdgeAI solution for smart speakers, industrial monitoring, voice interfaces, and acoustic event detection applications. The AM62DX integrates a high-performance A53 ARM processor with a dedicated C7x DSP optimized for neural network inference and digital signal processing. With a few simple steps one can run real-time EdgeAI pipelines such as speech enhancement and audio classification on the C7x DSP. | ||
|
|
||
| The SDK enables an interplay of open-source components such as GStreamer and deep learning runtimes such as TVM-RT. The reference applications showcase EdgeAI examples such as speech enhancement and audio classification in C++ variants. The SDK supports edit-build-debug cycles directly on the target and also on a host PC to cross compile and build the applications. |
Contributor
There was a problem hiding this comment.
one only RT, rephrase the words
Contributor
There was a problem hiding this comment.
what is in C++ variants ?
Contributor
There was a problem hiding this comment.
edit-build-debug cycles ?
v-singh1
suggested changes
Oct 23, 2026
|
|
||
| The SDK provides software and tools to let users run high-performance audio deep learning inference on Texas Instruments' AM62DX processor. We offer a practical embedded EdgeAI solution for smart speakers, industrial monitoring, voice interfaces, and acoustic event detection applications. The AM62DX integrates a high-performance A53 ARM processor with a dedicated C7x DSP optimized for neural network inference and digital signal processing. With a few simple steps one can run real-time EdgeAI pipelines such as speech enhancement and audio classification on the C7x DSP. | ||
|
|
||
| The SDK enables an interplay of open-source components such as GStreamer and deep learning runtimes such as TVM-RT. The reference applications showcase EdgeAI examples such as speech enhancement and audio classification in C++ variants. The SDK supports edit-build-debug cycles directly on the target and also on a host PC to cross compile and build the applications. |
Contributor
There was a problem hiding this comment.
what is in C++ variants ?
|
|
||
| The SDK provides software and tools to let users run high-performance audio deep learning inference on Texas Instruments' AM62DX processor. We offer a practical embedded EdgeAI solution for smart speakers, industrial monitoring, voice interfaces, and acoustic event detection applications. The AM62DX integrates a high-performance A53 ARM processor with a dedicated C7x DSP optimized for neural network inference and digital signal processing. With a few simple steps one can run real-time EdgeAI pipelines such as speech enhancement and audio classification on the C7x DSP. | ||
|
|
||
| The SDK enables an interplay of open-source components such as GStreamer and deep learning runtimes such as TVM-RT. The reference applications showcase EdgeAI examples such as speech enhancement and audio classification in C++ variants. The SDK supports edit-build-debug cycles directly on the target and also on a host PC to cross compile and build the applications. |
Contributor
There was a problem hiding this comment.
edit-build-debug cycles ?
cshilwant
reviewed
Oct 23, 2026
| * - ``libti_rpmsg_dma`` | ||
| - RPMsg and DMA client library for inter-processor communication | ||
| * - ``libtvm_runtime.so`` | ||
| - Neo-TVM ARM64 runtime for model inference |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR introduces a complete, production-ready documentation suite for AM62DX Edge AI audio processing, featuring new plugin extension guides, visual GStreamer pipeline diagrams, comprehensive build and deployment instructions, and an out-of-box web demo guide. The documentation supports both reference implementations and custom Bring-Your-Own-Model (BYOM) workflows.