P

Powerlm 3b

Developed by ibm-research
PowerLM-3B is a small language model with 3 billion parameters, trained using the Power learning rate scheduler, and demonstrates excellent performance across multiple benchmarks including natural language multiple-choice, code generation, and mathematical reasoning.
Downloads 11.07k
Release Time : 8/14/2024

Model Overview

PowerLM-3B is an advanced small language model trained on a mix of open-source and proprietary datasets, suitable for tasks such as text generation, code generation, and mathematical reasoning.

Model Features

Efficient Training
Trained using the Power learning rate scheduler to optimize training efficiency.
Excellent Multi-Task Performance
Outperforms models of similar scale across multiple benchmarks including natural language multiple-choice, code generation, and mathematical reasoning.
Compact and Efficient
Designed with a compact 3 billion parameters, making it suitable for deployment in resource-limited environments.

Model Capabilities

Text Generation
Code Generation
Mathematical Reasoning
Natural Language Understanding

Use Cases

Programming Assistance
Code Generation
Generate code snippets based on natural language descriptions.
Achieved 26.8% pass@1 on the HumanEval benchmark.
Code Completion
Assist developers in completing code writing.
Achieved 33.6% pass@1 on the MBPP benchmark.
Education
Math Problem Solving
Solve mathematical reasoning problems.
Achieved 34.9% accuracy on the GSM8k benchmark.
Knowledge Q&A
Answer various knowledge-based questions.
Achieved 49.2% accuracy on the MMLU benchmark.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase