Image Recognition with LLaVa in Python
ฝัง
- เผยแพร่เมื่อ 26 มิ.ย. 2024
- In this video we learn how to easily do image recognition and labeling in Python using LLaVa and Ollama locally.
Ollama: ollama.com
◾◾◾◾◾◾◾◾◾◾◾◾◾◾◾◾◾
📚 Programming Books & Merch 📚
🐍 The Python Bible Book: www.neuralnine.com/books/
💻 The Algorithm Bible Book: www.neuralnine.com/books/
👕 Programming Merch: www.neuralnine.com/shop
💼 Services 💼
💻 Freelancing & Tutoring: www.neuralnine.com/services
🌐 Social Media & Contact 🌐
📱 Website: www.neuralnine.com/
📷 Instagram: / neuralnine
🐦 Twitter: / neuralnine
🤵 LinkedIn: / neuralnine
📁 GitHub: github.com/NeuralNine
🎙 Discord: / discord - วิทยาศาสตร์และเทคโนโลยี
Riding the awesomeness wave again!
Nice, very helpful!
Is it possible to create embeddings of pictures with the model?
Thanks for the video, how to make sure that I install Ollama on the GPU not on the CPU?
Thanks :) Is it possible to use this model as an ocr alternativ to get for example informationen from a jpeg image which is an id-card ?
This will be too much heavy for just that
Instead considering yolo would be a better option
@@sumukhas5418 Thanks for the answer :) Actually I am trying pytesseract to read id-card information, which are photographed by a phone and the results are not very good :/ Do you have some ideas, how I could get some better results?
If my local ram is 8 gb, which ollama model would you recommend to use?
deepseek-coder ❤
deepseek-coder ❤
What a nice vid. Can I do a ai without using open ai ?
Hi could you please explain why someone has to refresh to get some reasonable output at least what we expect !? Why does something like this happens with such ai models ?
Each Input for the model comes with a random seed, that's why you don't get the same responses with the same input every time, because of that you could try multiple times until you get a reasonable response.
First comment 😊😊😊
How much RAM and VRAM needed ?!
With 4-bit quantization, for LLaVA-1.5-7B, it uses less than 8GB VRAM on a single GPU, typically the 7B model can run with a GPU with less than 24GB memory, and the 13B model requires ~32 GB memory. You can use multiple 24-GB GPUs to run 13B model
oh im too fast