0
0
Budget: 5000 UAH Deadline: 3 days
3 years of experience deploying systems on servers with 4GB of RAM. Experience working with YOLO on servers, quantizing models, and optimizing code.
2
2
Budget: 10000 UAH Deadline: 7 days
Greetings! I have experience in developing my own AI models (GAN, RNN, LSTM). I have used Tensorflow, PyTorch. I am ready to consider more details.
14
0
Budget: 2000 UAH Deadline: 1 day
The water is fine, everything is ok 🙂
Regarding the task, I see that the main problems right now are due to asynchronicity and resources, plus frame drops, which is especially critical for Yolo and DQN. If the modular code worked but not perfectly, then the separation of files somewhere disrupted the connections between components or caused additional delays. We need to look at how data is transmitted between modules, whether there are bottlenecks in video stream processing.
Regarding optimization – we can dig into multithreading, quantum optimization (if we really need to squeeze the maximum), as well as simply properly setting up the pipeline. Let's then figure out what we have and what we want to achieve.
Regarding the stages – it makes complete sense, because you can't avoid testing in such cases. It's better to identify critical points right away to avoid redoing everything unnecessarily.
Question: what is this all running on right now? What hardware, what libraries? Because this greatly affects which optimization approaches make sense.
Regarding hours – I will write in private to discuss it in more detail, as it depends on the complexity of the tasks that arise along the way.
