Discover and read the best of Twitter Threads about #tinyml
Most recents (2)
@mjrovai @seeedstudio my $15 #xiaoEsp32S3 arrived all three of them. Canadian Import tax was a bit high at $25 CDN not sure if that is fixed or increases if I purchased more.
Excited to test the camera, mic and WiFi to see if it works with robotics, @EdgeImpulse
#ml and #tinyML
Excited to test the camera, mic and WiFi to see if it works with robotics, @EdgeImpulse
#ml and #tinyML
WiFi antennae was hard to push on, not sure why, I have done a ton before. Camera clipped in easy.
@seeedstudio #xiaoEsp32S3
@seeedstudio #xiaoEsp32S3
Off course my setup is not standard so installing the board is a pain. Great instructions at wiki.seeedstudio.com/xiao_esp32s3_g…
#Highlights2021 for me: our #survey on efficient processing of #sparse and compressed tensors of #ML/#DNN models on #hardware accelerators published in @ProceedingsIEEE.
Paper: dx.doi.org/10.1109/JPROC.…
arXiv: arxiv.org/abs/2007.00864
RT/sharing appreciated. 🧵
Paper: dx.doi.org/10.1109/JPROC.…
arXiv: arxiv.org/abs/2007.00864
RT/sharing appreciated. 🧵
Context: Tensors of ML/DNN are compressed by leveraging #sparsity, #quantization, shape reduction. We summarize several such sources of sparsity & compression (§3). Sparsity is induced in structure while pruning & it is unstructured inherently for various applications or sources.
Likewise, leveraging value similarity or approximate operations could yield irregularity in processing. Also, techniques for size-reduction make tensors asymmetric-shaped. Hence, special mechanisms can be required for efficient processing of sparse and irregular computations.
