tensorblock/Phi-3.5-MoE-instruct-GGUF

Feedback and support: TensorBlock's Twitter/X, Telegram Group and Discord server

microsoft/Phi-3.5-MoE-instruct - GGUF

This repo contains GGUF format model files for microsoft/Phi-3.5-MoE-instruct.

The files were quantized using machines provided by TensorBlock, and they are compatible with llama.cpp as of commit ec7f3ac.

Run them on the TensorBlock client using your local machine ↗

Prompt template

<|system|>
{system_prompt}<|end|>
<|user|>
{prompt}<|end|>
<|assistant|>

Model file specification

Filename	Quant type	File Size	Description
Phi-3.5-MoE-instruct-Q2_K.gguf	Q2_K	15.265 GB	smallest, significant quality loss - not recommended for most purposes
Phi-3.5-MoE-instruct-Q3_K_S.gguf	Q3_K_S	18.055 GB	very small, high quality loss
Phi-3.5-MoE-instruct-Q3_K_M.gguf	Q3_K_M	20.033 GB	very small, high quality loss
Phi-3.5-MoE-instruct-Q3_K_L.gguf	Q3_K_L	21.688 GB	small, substantial quality loss
Phi-3.5-MoE-instruct-Q4_0.gguf	Q4_0	23.599 GB	legacy; small, very high quality loss - prefer using Q3_K_M
Phi-3.5-MoE-instruct-Q4_K_S.gguf	Q4_K_S	23.810 GB	small, greater quality loss
Phi-3.5-MoE-instruct-Q4_K_M.gguf	Q4_K_M	25.346 GB	medium, balanced quality - recommended
Phi-3.5-MoE-instruct-Q5_0.gguf	Q5_0	28.816 GB	legacy; medium, balanced quality - prefer using Q4_K_M
Phi-3.5-MoE-instruct-Q5_K_S.gguf	Q5_K_S	28.816 GB	large, low quality loss - recommended
Phi-3.5-MoE-instruct-Q5_K_M.gguf	Q5_K_M	29.716 GB	large, very low quality loss - recommended
Phi-3.5-MoE-instruct-Q6_K.gguf	Q6_K	34.359 GB	very large, extremely low quality loss
Phi-3.5-MoE-instruct-Q8_0.gguf	Q8_0	44.500 GB	very large, extremely low quality loss - not recommended

Downloading instruction

Command line

Firstly, install Huggingface Client

pip install -U "huggingface_hub[cli]"

Then, downoad the individual model file the a local directory

huggingface-cli download tensorblock/Phi-3.5-MoE-instruct-GGUF --include "Phi-3.5-MoE-instruct-Q2_K.gguf" --local-dir MY_LOCAL_DIR

If you wanna download multiple model files with a pattern (e.g., *Q4_K*gguf), you can try:

huggingface-cli download tensorblock/Phi-3.5-MoE-instruct-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'