where to get merges.txt and vocab.json from?

#34
by bigcheese11 - opened

Hi,
i tried the model converter to play around in some tool with different models (.safetensor), but i wonder where i can get the merges.txt file and vocab.json from...
is there an option in the model converter which generates them maybe?

thnx

@bigcheese11
The final stage of conversion will be automatically downloaded from https://huggingface.co/openai/clip-vit-base-patch32/tree/main

It will be saved in TextEncoder.mlmodelc

hm, ok... so for now i use the direct link to huggingface to download the model of interest (juggernaut xl 5)... lets see if this works :) thnx

@bigcheese11
I have tested juggernaut xl v5. Although the number of training steps displayed may be the highest at present, the effect is actually not good.
I recommend you to try ProtoVision, NightVision, CrystalClear. Texture details can beat all SDXL, and there is no need for Refiner at all.

Sign up or log in to comment