3 ms·
https://huggingface.co/collections/microsoft/phi-3-6626e15e9585a200d2d761e3 https://huggingface.co/collections/microsoft/phi-3-6626e15e9..., all of these models
by GaggiX 2y ago
https://huggingface.co/collections/microsoft/phi-3-6626e15e9585a200d2d761e3 https://huggingface.co/collections/microsoft/phi-3-6626e15e9..., all of these models except Phi-3 mini are new.
- free_bip 2y agoLooking at the benchmarks, it seems like Phi-3 Small (7B) marginally beats out Llama3-8B on most tasks, which is pretty exciting!
- ashu1461 2y agoIt also seems to be comparable to gpt 3.5 turbo which I feel hard to believe. People have obviously found out a way around these benchmarks.
- Filligree 2y agoHave you found either model to be good enough to do anything interesting, reliably?
- ukuina 2y agoLlama3-8B is adequate for non-technical summarization or simple categorization.
- zone411 2y agoOn my benchmark (NYT Connections), Phi-3 Small performs well (8.4) but Llama 3 8B Instruct is still better (12.3). Phi-3 Medium 4k is disappointing and often fails to properly follow the output format.