The difference with normal quantizations is that I quantize the output and embed tensors to f16. and the other tensors to 15_k,q6_k or q8_0. This creates models that are little or not degraded at all and have a smaller size. They run at about 3-6 t/sec on CPU only using llama.cpp And obviously faster on computers with potent GPUs

the fast cat at ZeroWw/Phi-3-song-lyrics-1.0-GGUF

Model Description

This model is a specialized model, specifically designed for writing song lyrics. Phi-3 has a built-in mechanism to reduce rejection, so please use the default_lm_studio_windows in LM Studio.
The generated lyrics may be a bit long, so please modify them yourself. If you don't modify them, I'll assume you're using the modified version, and the generated content is not related to the model :p
lease handle the generated content yourself, as the model does not guarantee anything about - the generated content. If any copyright issues arise, please handle them yourself.
Thanks to aifeifei799 for creating the lora
I just integrated aifeifei799's LORA, so if you have any issues, please contact aifeifei799 :p
The model strictly follows the MIT license.

Model Set

download GGUF file
set LM Studio https://lmstudio.ai/

Model Test Prompt

Text

"Please write song lyrics with the following structure: Intro-Verse-Chorus-Verse-Chorus-Bridge-Chorus-Outro.Rewrite an English love song in English, give me the title, complete lyrics, and style of the song, without translation or phonetic notation, with emotional and heartfelt lyrics. Here's a breakdown of each section:
song name:
Style of Music:
Intro:
Verse 1:
Chorus:
Verse 2:
Chorus:
Bridge:
Chorus:
Outro:

Ocean82/Phi-3-song-lyrics-1.0

Special Thanks:

Lewdiculous's superb gguf version, thank you for your conscientious and responsible dedication.

mradermacher's superb gguf version, thank you for your conscientious and responsible dedication.

These are my own quantizations (updated almost daily).

Model Description

Model Set

Model Test Prompt

Ocean82/
Phi-3-song-lyrics-1.0