YeungNLP/firefly-llama2-13b
The YeungNLP/firefly-llama2-13b is a 13 billion parameter English-only language model based on the Llama 2 architecture. It achieves a competitive score of 62 on the Open LLM leaderboard, ranking third among all 13B models. This model is notable for its efficient training using QLoRA technology, requiring significantly fewer resources than other top-ranked models, making it accessible for training on GPUs with 24GB VRAM.
Loading preview...
YeungNLP/firefly-llama2-13b Overview
The YeungNLP/firefly-llama2-13b is a 13 billion parameter English language model built upon the Llama 2 architecture. It has demonstrated strong performance on the Open LLM leaderboard, securing the third position among all 13B models with a score of 62, trailing the top model by only 0.5 points.
Key Characteristics
- English-only Model: This model was trained exclusively on English data and its vocabulary has not been expanded for Chinese or other languages.
- Efficient Training: A significant feature is its training methodology, utilizing QLoRA technology. This approach drastically reduces the computational resources needed for training, enabling the fine-tuning of this 13B parameter model on GPUs with as little as 24GB of VRAM.
Use Cases
This model is particularly well-suited for applications requiring a high-performing 13B English language model where resource efficiency during training or fine-tuning is a critical factor. Its competitive benchmark performance suggests its applicability in various English NLP tasks.