ludis's picture
Update README.md
e626c3d
|
raw
history blame
637 Bytes
metadata
datasets:
  - ewof/koishi-instruct-metharme
  - PygmalionAI/PIPPA

just testing for now, qlora merge, several things different between this and the 7b

https://rentry.org/v43eo updated the rentry slightly

image/png

NousResearch/Llama-2-13b-hf tuned on koishi data (without code subsets) for 1 epoch

then tuned on pippa for 1 epoch

then tuned on gpt4 rp data from whocars proxy for 1 epoch

then tuned on limarp (without ponyville, lolicit, all the fallen, and eka's portal subsets) for 2 epochs

all metharme format