Mistral 7B The real killer Generative AI - ft. Mr Beast

Описание

Mistral 7B is a 7.3B parameter model that:
Outperforms Llama 2 13B on all benchmarks
Outperforms Llama 1 34B on many benchmarks
Approaches CodeLlama 7B performance on code, while remaining good at English tasks
Uses Grouped-query attention (GQA) for faster inference
Uses Sliding Window Attention (SWA) to handle longer sequences at smaller cost

In this video we will see the advantages of Mistral 7B model, how to run it on our local machines and will also build a sample python notebook to test it out.

0:00 Introduction
0:42 What is Mistral AI
1:34 Mistral AI API access
3:06 Test Mistral AI with Ollama
5:14 Testing Mitsral-7B model in Google colab
15:03 Advantages of Mistral AI open source models
16:10 Train your own Mistral AI model with your data
17:39 Checking the Mistral model response
18:25 Conclusion

Github link - https://github.com/ruchibahl18/mistral-ai/tree/main

Mistral AI - https://mistral.ai/
https://mistral.ai/news/announcing-mistral-7b/

Original Mistral AI model - https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1

Mistral AI 7B Quanatized model - https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF/tree/main

Try with Ollama - https://ollama.ai/library/mistral