4 ms·
Not the parent, but I started using Llama 3.1 8b and it's very good. I'd say it's as good as or better than GPT 3.5 based on my usage. Some benchmarks: https:/
by rkwz 2y ago
Not the parent, but I started using Llama 3.1 8b and it's very good.
I'd say it's as good as or better than GPT 3.5 based on my usage. Some benchmarks: https://ai.meta.com/blog/meta-llama-3-1/ https://ai.meta.com/blog/meta-llama-3-1/
Looking forward to try other models like Qwen and Phi in near future.
- milleramp 2y agoI found it to not be as good in my case for code generation and suggestions. I am using a quantized version maybe that's the difference.