Update voice assistant LP with multimodal functionality #2368
Conversation
| KleidiAI simplifies development by abstracting away low-level optimization: developers can write high-level code while the KleidiAI library selects the most efficient implementation at runtime based on the target hardware. This is possible thanks to its deeply optimized micro-kernels tailored for Arm architectures. | ||
|
|
||
| As newer versions of the architecture become available, KleidiAI becomes even more powerful: simply updating the library allows applications like the Voice Assistant to take advantage of the latest architectural improvements - such as SME2 — without requiring any code changes. This means better performance on newer devices with no additional effort from developers. | ||
| As newer versions of the architecture become available, KleidiAI becomes even more powerful: simply updating the library allows applications like the Voice Assistant to take advantage of the latest architectural improvements - such as SME2 — without requiring any code changes. This means better performance on newer devices with no additional effort from developers. |
There was a problem hiding this comment.
Like the "multi-modal Voice Assistant" - just rephrasing things to add the latest enablement to multi-modal
| - Kotlin | ||
| - C++ | ||
| operatingsystems: | ||
| - Android |
There was a problem hiding this comment.
Should we say that the multi-modal voice assist example is based on Android, but for benchmarking purposes of the different individual building blocks, we support multi-platform operating systems (and we list them)
| https://gitlab.arm.com/kleidi/kleidi-examples/large-language-models | ||
| ``` | ||
|
|
||
| and build for various platforms: |
There was a problem hiding this comment.
And build for various platforms to test the individual building blocks. For the full pipeline, the example runs on Android OS
There was a problem hiding this comment.
thanks Gemma, I've updated this in 2 places for STT and LLM module and updated to say can build for various platforms to benchmark the functionality independently, also updated the last line in this page to mention that the example is based on android, hopefully this clarifies things more but happy to add more changes if unclear.
* added updated to multi-modal use * update to learning objectives and overview pages to specify platforms supported
|
All good on my side. Thanks Nina! |
6 participants
Before submitting a pull request for a new Learning Path, please review Create a Learning Path
Please do not include any confidential information in your contribution. This includes confidential microarchitecture details and unannounced product information.
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of the Creative Commons Attribution 4.0 International License.