Senthil Nayagam’s Post

#tinyml #tflite #tf.keras #esp32 *** Machine learning on MCU *** Steps to deploy tinyml (tensorflow lite) on esp32: - Using transfer learning from a pre-trained network - Fine-tuning the model to achieve the best result - Optional: Quantization - Convert the model to CC code - Implement inference code on esp32 Here is a demo that deploys MobileNetV1 on ESP32

Take note of the following: 1. If your cross-entropy loss remains consistently high during SFT 2. If your LoRA fine-tuning fails to meet its objectives 3. If you wish to enhance your models' capabilities even further Consider trying my implementation of the LinearMoE, which I use to replace all the linear layers. If you have any improvements in mind, feel free to share them in the comments. Many thanks for any suggestions! Link: https://lnkd.in/d-yWjPdb

How is #machinelearning beating best #video #compression algorithms H.264 and H.265. An independent developper did the #benchmark by using pretrained #diffusion #models. https://lnkd.in/e8T48uAX Imagine a compression algorithm that knows how to compress necause it can predict the rest od the movie. That's what Extreme Video Compression tries to do. At low bit rate the diffusion model performs astoundingly and remains extremely competitive in higher bit rates.

Record your screen + dev logs in a click Engineers get full devtools pane and can copy as cURL to replay the bug locally

Replay network requests on your local machine & debug faster w/ Jam logs. Just copy as cURL.

letfn is a special form that allows you to define a group of mutually recursive functions. It's useful when you have multiple functions that depend on each other and need to refer to each other within their bodies. https://lnkd.in/exJU_xDk

Test post with link and settings, personalized https://ow.ly/5cNN50S9t4T

Interested in learning more about HTJ2K, the new high performance image compression codec added to the DICOM standard? Check out this set of resources

I've done another video where I explain how you can create your own ingestion pipeline in Kernel Memory with customized steps. https://lnkd.in/dr-seQgU

SAVE THE DATE and Scan the CODE for you input