This is a quantized GGUF version of Microsoft Phi-2 to 4_0, 8_0 bits and the converted 16 FP model.

(link to the original model : https://huggingface.co/microsoft/phi-2)

Disclamer : make sure to have the latest version of llama.cpp after commit b9e74f9bca5fdf7d0a22ed25e7a9626335fdfa48

Downloads last month
479
GGUF
Model size
2.78B params
Architecture
phi2

4-bit

8-bit

16-bit

Inference Examples
Inference API (serverless) has been turned off for this model.

Space using kroonen/phi-2-GGUF 1