Llama-3.1-Minitron4B: Nvidia's Compressed Language Model
🦙
Llama-3.1-Minitron4B
Nvidia's compressed version of Llama3, designed for resource-constrained devices
Key Features
- ✅ Only 400 million parameters
- ✅ Efficient training and deployment
- ✅ Comparable performance to larger models
Technologies Used
- 🔪 Model Pruning (Deep & Wide)
- 🧠 Knowledge Distillation
- 🔬 NeMo-Aligner Fine-tuning
Training Efficiency
40x
Less training data
→
16%
Performance improvement
Model Capabilities
📝
Instruction Following
🎭
Role Playing
🔍
Retrieval-Augmented Generation
⚙️
Function Calling
@AIbase
Tags：Llama-3.1-Minitron4B Nvidia language model compressed language model resource-constrained devices model pruning knowledge distillation AI model deployment Hugging Face commercial use