vicuna-7b-v1.5#
README(From Huggingface)#
inference: false license: llama2#
Vicuna Model Card#
Model Details#
Vicuna is a chat assistant trained by fine-tuning Llama 2 on user-shared conversations collected from ShareGPT.
Developed by: LMSYS
Model type: An auto-regressive language model based on the transformer architecture
License: Llama 2 Community License Agreement
Finetuned from model: Llama 2
Model Sources#
Repository: https://github.com/lm-sys/FastChat
Blog: https://lmsys.org/blog/2023-03-30-vicuna/
Paper: https://arxiv.org/abs/2306.05685
Demo: https://chat.lmsys.org/
Uses#
The primary use of Vicuna is research on large language models and chatbots. The primary intended users of the model are researchers and hobbyists in natural language processing, machine learning, and artificial intelligence.
How to Get Started with the Model#
Command line interface: https://github.com/lm-sys/FastChat#vicuna-weights
APIs (OpenAI API, Huggingface API): https://github.com/lm-sys/FastChat/tree/main#api
Training Details#
Vicuna v1.5 is fine-tuned from Llama 2 with supervised instruction fine-tuning. The training data is around 125K conversations collected from ShareGPT.com. See more details in the "Training Details of Vicuna Models" section in the appendix of this paper.
Evaluation#
Vicuna is evaluated with standard benchmarks, human preference, and LLM-as-a-judge. See more details in this paper and leaderboard.
Difference between different versions of Vicuna#
Model Files#
.gitattributes (1.5 KB)
README.md (1.9 KB)
config.json (1.2 KB)
generation_config.json (162.0 B)
model_state.pdparams (12.6 GB)
processor/eval/preprocessor_config.json (584.0 B)
processor/train/preprocessor_config.json (584.0 B)
sentencepiece.bpe.model (488.0 KB)
special_tokens_map.json (438.0 B)
tokenizer_config.json (749.0 B)