reciprocate's picture
Create README.md
41f353d
|
raw
history blame
692 Bytes
metadata
license: apache-2.0
language:
  - en

OpenLLama-13B for reward modeling

Usage:

from transformers import AutoModelForSequenceClassification, AutoTokenizer

ckpt = "reciprocate/openllama-13b_rm_oasst-hh"
model = AutoModelForSequenceClassification.from_pretrained(ckpt, load_in_4bit=True)
tokenizer = AutoTokenizer.from_pretrained(ckpt)

model(**tokenizer("ASSISTANT: This sentence is a lie.", return_tensors="pt"))[0].item()

Output:

-1.626953125