ArmoRM-Llama3-8B-v0.1

10.8K
182
8.0B
llama
by
RLHFlow
Language Model
OTHER
8B params
Fair
11K downloads
Community-tested
Edge AI:
Mobile
Laptop
Server
18GB+ RAM
Mobile
Laptop
Server
Quick Summary

License: llama3. Authors include Haoxiang Wang, Wei Xiong, Tengyang Xie, Han Zhao, and Tong Zhang.

Device Compatibility

Mobile
4-6GB RAM
Laptop
16GB RAM
Server
GPU
Minimum Recommended
8GB+ RAM

Training Data Analysis

🟡 Average (4.8/10)

Researched training datasets used by ArmoRM-Llama3-8B-v0.1 with quality assessment

Specialized For

general
science
multilingual
reasoning

Training Datasets (4)

common crawl
🔴 2.5/10
general
science
Key Strengths
  • Scale and Accessibility: At 9.5+ petabytes, Common Crawl provides unprecedented scale for training d...
  • Diversity: The dataset captures billions of web pages across multiple domains and content types, ena...
  • Comprehensive Coverage: Despite limitations, Common Crawl attempts to represent the broader web acro...
Considerations
  • Biased Coverage: The crawling process prioritizes frequently linked domains, making content from dig...
  • Large-Scale Problematic Content: Contains significant amounts of hate speech, pornography, violent c...
c4
🔵 6/10
general
multilingual
Key Strengths
  • Scale and Accessibility: 750GB of publicly available, filtered text
  • Systematic Filtering: Documented heuristics enable reproducibility
  • Language Diversity: Despite English-only, captures diverse writing styles
Considerations
  • English-Only: Limits multilingual applications
  • Filtering Limitations: Offensive content and low-quality text remain despite filtering
wikipedia
🟡 5/10
science
multilingual
Key Strengths
  • High-Quality Content: Wikipedia articles are subject to community review, fact-checking, and citatio...
  • Multilingual Coverage: Available in 300+ languages, enabling training of models that understand and ...
  • Structured Knowledge: Articles follow consistent formatting with clear sections, allowing models to ...
Considerations
  • Language Inequality: Low-resource language editions have significantly lower quality, fewer articles...
  • Biased Coverage: Reflects biases in contributor demographics; topics related to Western culture and ...
arxiv
🟡 5.5/10
science
reasoning
Key Strengths
  • Scientific Authority: Peer-reviewed content from established repository
  • Domain-Specific: Specialized vocabulary and concepts
  • Mathematical Content: Includes complex equations and notation
Considerations
  • Specialized: Primarily technical and mathematical content
  • English-Heavy: Predominantly English-language papers

Explore our comprehensive training dataset analysis

View All Datasets

Code Examples

[2.78125 2.859375 3.484375 1.3847656 1.296875 ]pythontransformers
from typing import Dict, List
import torch
from transformers import AutoModelForSequenceClassification, AutoTokenizer


class ArmoRMPipeline:
    def __init__(self, model_id, device_map="auto", torch_dtype=torch.bfloat16, truncation=True, trust_remote_code=False, max_length=4096):
        self.model = AutoModelForSequenceClassification.from_pretrained(
            model_id,
            device_map=device_map,
            trust_remote_code=trust_remote_code,
            torch_dtype=torch_dtype,
        )
        self.tokenizer = AutoTokenizer.from_pretrained(
            model_id,
            use_fast=True,
        )
        self.truncation = truncation
        self.device = self.model.device
        self.max_length = max_length

    def __call__(self, messages: List[Dict[str, str]]) -> Dict[str, float]:
        """
        messages: OpenAI chat messages to be scored
        Note: no batching since due to length differences, the model will have to pad to the max length which is not efficient
        Returns: a dictionary with the score between 0 and 1
        """
        input_ids = self.tokenizer.apply_chat_template(
            messages,
            return_tensors="pt",
            padding=True,
            truncation=self.truncation,
            max_length=self.max_length,
        ).to(self.device)
        with torch.no_grad():
            output = self.model(input_ids)
            score = output.score.float().item()
        return {"score": score}

# Create Reward Model Pipeline 
prompt = 'What are some synonyms for the word "beautiful"?'
rm = ArmoRMPipeline("RLHFlow/ArmoRM-Llama3-8B-v0.1", trust_remote_code=True)
# score the messages
response1 = 'Nicely, Beautifully, Handsome, Stunning, Wonderful, Gorgeous, Pretty, Stunning, Elegant'
score1 = rm([{"role": "user", "content": prompt}, {"role": "assistant", "content": response1}])
print(score1)

response2 = '''Certainly! Here are some synonyms for the word "beautiful":

1. Gorgeous
2. Lovely
3. Stunning
4. Attractive
5. Pretty
6. Elegant
7. Exquisite
8. Handsome
9. Charming
10. Alluring
11. Radiant
12. Magnificent
13. Graceful
14. Enchanting
15. Dazzling

These synonyms can be used in various contexts to convey the idea of beauty.'''
score2 = rm([{"role": "user", "content": prompt}, {"role": "assistant", "content": response2}])
print(score2)

response3 = 'Sorry i cannot answer this.'
score3 = rm([{"role": "user", "content": prompt}, {"role": "assistant", "content": response3}])
print(score3)

Deploy This Model

Production-ready deployment in minutes

Together.ai

Instant API access to this model

Fastest API

Production-ready inference API. Start free, scale to millions.

Try Free API

Replicate

One-click model deployment

Easiest Setup

Run models in the cloud with simple API. No DevOps required.

Deploy Now

Disclosure: We may earn a commission from these partners. This helps keep LLMYourWay free.