CALISTA-INDUSTRY/llama-3.2-3B-reasoning-en-ft-v1

Model Summary

CALISTA-INDUSTRY/llama-3.2-3B-reasoning-en-ft-v1 is a fine-tuned version of Meta's LLaMA 3 3B model, optimized for English-language reasoning tasks. This model has been adapted to enhance performance in logical reasoning, problem-solving, and conversational understanding.

Model Details

• Developed by: Mohammad Yani & Rizky Sulaeman, Politeknik Negeri Indramayu
• Model type: Decoder-only transformer (LLaMA 3 architecture)
• Parameter count: 3.21 billion
• Quantization formats: 4-bit (Q4_K_M), 5-bit (Q5_K_M), 8-bit (Q8_0)
• Training data: [Specify datasets or data sources used]
• License: Apache License 2.0
• Base model: meta-llama/Llama-3-3B

Intended Uses & Limitations

Intended Uses

• Applications:
  • Logical reasoning tasks
  • Conversational agents requiring enhanced reasoning capabilities
  • Educational tools focusing on critical thinking
• Users:
  • Researchers in natural language processing
  • Developers building AI-driven applications
  • Educators and students in AI-related fields

Limitations

• The model's performance may degrade on tasks outside its fine-tuned domain.
• Not suitable for real-time applications without further optimization.
• May produce incorrect or nonsensical answers; outputs should be verified in critical applications.

How to Use

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="CALISTA-INDUSTRY/llama-3.2-3B-reasoning-en-ft-v1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)