Видео

Very powerful and spiring. waiting for more videos keep going

Perfect video, I'll use it to showcase the platform itself.

Is there videos that show more of a starting point? It's seems like those making these videos are too familiar with it and leave out a lot of background details. Sorry. It's seem to jump in a lot further ahead than it should.

Guess it's ahead of where I'm at currently. Thanks. Maybe after I do some more research first.

Very interesting. Thank you!

thanks, it's very useful for many projects.

👏👏🔥

All I want to do is detect people in the cameras fov, and get the location of them, with the camera being located within the height of a person, probably 6 ft away. This is for a Halloween project. I want to use the location value 'X' as an input to other functions. I want to run this on an inexpensive device such as a esp32 or rPi0. It doesn't have to classify the people, I just want the continuous locations of people in the camera frame. Just like on cctv security systems. Only challenges is it's outside in low light at night and cars may be going by in the camera frame that I want to ignore. Is this project feasible with edge impulse?

Im having an error says "no test available" after training in object detection. someone help me fix it

is this project available for cloning on edgeimpulse platform?? is it's model and code available on github?

Is it possible to detect human and not human by this process? And If I do not want to use Pi, what is other best option to do that?

You are using for testing the same video than for fine tuning your smaller model. What about a video with toys in another environment ?

Very interesting

Very clear explanation on undefitting and overfitting

A lot of information but very good

it didn't work at all I wasted 1 hour trying.

During the akida conversion, does it only handle only square kernels?

what eval system compatible with edge impulse can i start with to easily extract ADC samples for low resolution and kSPS rates? i would like to plug in an USB interface to the eval unit and hook up a couple of analog input sensors. What suggestions should i consider for an quick start? Thx

Very good explanation

We are so far but still at the beginning . Letting students and independent researchers to use it as free plan , make me to believe in beautiful future and give you a sub .

but this is a paid feature...

Is python programming related anywhere for embedded jobs?

I get /opt/homebrew/opt/python@3.12/bin/python3.12: No module named esptool

Whe I try to "re-balance" it gave me an error: Failed to perform train / test split. The solution was jus put this text in the box: perform split. It worked for me. Thank you

OpenAI still has everyone tricked that gpts are 'multimodal', they're not, not in any deep way. GPT-4o has no idea how to find boxes in a picture for instance, it can't 'see', it has no idea how to interpret audio data other than TTS, all they've done is bolt on crappy OCR and TTS in a somewhat integrated way.

Anomaly detection with supervised models is really bad, you need to support more unsupervised models they are way more efficient and you don t need anomaly data for training

*Pleas help me* hello sir, is it possible to do something like open cv video stream with a bounding box and as far as I know edge impulse has a centeroid output candidate for use in esp32 cam and edge impulse how to do it🤯

Great Job bro, Thanks a lot!

This company is seriously so awesome. It's been quite some time since I enjoyed some niche website as much as this one

the video is good but need to edit path in system variable not in user variable . in my case when i add path into user variable ,terminal wont recognize but when i add path into system variable it run

👍👍👍nice

In this video I talk about a plugin that I made in Edge Impulse. The general approach is as follows - use a foundational audio classifier to look at your unlabeled audio dataset! This approach is two-fold: first - we give the model a few sampels of an event or sound type we are looking for and determine to which class from the ones the model knows it belongs. It can be though of as encoding our sound events in the model "language". After that we give larger samples that may or may not contain these events and tell the model to only react when those classes identified earlier are identified. This way we can explore large audio datasets quickly. There is a good chance that some classifications are not exactly right - but it gives you a subset of you audio to actually take a look at - instead of having to listen though all of it

I found this library image_util.h

very cool!! Keep it up team!

Great job! And nice seeing you on edge impulse channel 🎉

Nice project! Which languages are supported?

It is very helpful

Thank you Soo much Bro❤❤🎉

Meu herói : )

This is amazing.

how to reconnect?

how to detect objects using esp32 deployment with live esp32 cam classification?

Your explanation of edge AI is excellent. I found your video by chance, but it made the concept very easy to understand.

Using video frames for training is really effective for small domains. Not only for Yes/No answers but also generative models. Before the release of GPT3, I was playing with the original Diffusion code (available on github), and giving it frames from a very short dashcam drive-around-town. The results were amazing. And this is for unlabeled data. I used this same data to train a VAE, and that just sucked, definitely some magic going on in the diffusion model, even for very limited data sets.

I paid for Enterprise but the Pipeline isn't working. I've 1178 data images in training, but none of them are being labelled because the Pipeline is registering 0 samples. How do I fix this?

I appreciate the demo, and I enjoyed it thanks. I guess if you distill this, you're paying ChatGPT (or the model of your choice) to label your data. I think the concern is that to do this for a variety of specific small models, it could cost a lot and Edge Impulse prices will always fluctuate according to OpenAI's pricing (or whoever is running the model). You would probably save some money and lose accuracy on running the LLM locally, can Edge Impulse use locally run models? And my other perspective to offer is that it'll only be good for labelling common things (things the teaching model was taught on), so some limiting of options there. I really did enjoy your video a lot though, thanks.

show the how not the what 🤦‍♂️👎

Pronounced Dow

Bawal

Stop smacking your lips dude..