File size: 1,217 Bytes
f7887d4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 |
EXL3 quants of [Phi-4-mini-instruct](https://huggingface.co/microsoft/Phi-4-mini-instruct)
At the moment these are all converted with 8-bpw output layers. Currently investigating why there's a small-but-noticeable drop in accuracy at 6-bpw. Likely it has to do with the tied embeddings.
[2.00 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/2.0bpw)
[2.25 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/2.25bpw)
[2.50 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/2.5bpw)
[3.00 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/3.0bpw)
[3.50 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/3.5bpw)
[4.00 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/4.0bpw)
[5.00 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/5.0bpw)
[6.00 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/6.0bpw)
[8.00 bits per weight / H8](https://huggingface.co/turboderp/Phi-4-mini-instruct-exl3/tree/8.0bpw)
|