Tulu 3 8B Reward
A reward model based on Llama 3.1 8B
Deploy Tulu 3 8B Reward behind an API endpoint in seconds.
Example usage
allenai/Llama-3.1-Tulu-3-8B-RM is a text-classification model, used to classify a text into a category.
It is frequently used in sentiment analysis, spam detection, and more. It's also used for deployment of chat rating models, e.g. RLHF reward models or toxicity detection models.
Input
1import requests
2import os
3
4headers = {
5 f"Authorization": f"Api-Key {os.environ['BASETEN_API_KEY']}"
6}
7
8requests.post(
9 headers=headers,
10 url="https://model-xxxxxx.api.baseten.co/environments/production/sync/predict",
11 json={
12 "inputs": [["Baseten is a fast inference provider"], ["classify this separately."]],
13 "raw_scores": True,
14 "truncate": True,
15 "truncation_direction": "Right"
16 }
17)JSON output
1[
2 [
3 {
4 "label": "excitement",
5 "score": 0.99
6 }
7 ],
8 [
9 {
10 "label": "excitement",
11 "score": 0.01
12 }
13 ]
14]
