Missing files
#1
by
gghfez
- opened
Hey mate, would you be able to upload the actual model files?
I lost my copy of this model, and I see this one didn't come back with your new account.
@gghfez If i have time ill try to reupload the files but ive been really busy lately. If you want to recreate it with mergekit its pretty easy though, here is the config:
models:
- model: Qwen2.5-14b-instruct
parameters:
weight: 1
merge_method: ties
base_model: Qwen2.5-14b
parameters:
normalize: true
int8_mask: true
dtype: bfloat16
That was the original config. Although I havnt tested it with every model, I think the bellow config works better.
models:
- model: Qwen2.5-14B-Instruct
parameters:
weight: 1
density: 1
merge_method: ties
base_model: Qwen2.5-14B
parameters:
weight: 1
density: 1
normalize: true
int8_mask: true
dtype: bfloat16
Thanks. I re-created it using the first config since that's like the model I had earlier.