Calibration dataset: VMware Open Instruct, 4096
Dumb assistant, 🔥 author.
4-bit Examples with Alpaca
!!NSFW!! - 🔥Erotica Writing Example🔥 - !!NSFW!!
Thanks to Charles Goddard for the recipe.
The idea here is to "move" Iambe from being based on vanilla L2 to being based on sequelbox/DynamicFactor instead.
Because task_arithmetic uses the raw deltas, this should be similar to if the SFT had been done over DynamicFactor.
Recipe
merge_method: task_arithmetic
base_model: athirdpath/BigLlama-20b-v1.1 # Base model you want to "move out" from
models:
model: athirdpath/Iambe-20b-DARE-v2 # SFTd model you want to transfer
model: athirdpath/DoubleFactor-20b # Base model you want to "move in" to
parameters:
- weight: 1.0
dtype: bfloat16
- Downloads last month
- 12
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.