lmzheng commited on
Commit
846d1ae
1 Parent(s): 7b3236e

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +45 -0
README.md ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ inference: false
3
+ license: llama2
4
+ ---
5
+
6
+ # Vicuna Model Card
7
+
8
+ ## Model Details
9
+
10
+ Vicuna is a chat assistant trained by fine-tuning Llama 2 on user-shared conversations collected from ShareGPT.
11
+
12
+ - **Developed by:** [LMSYS](https://lmsys.org/)
13
+ - **Model type:** An auto-regressive language model based on the transformer architecture
14
+ - **License:** Non-commercial license
15
+ - **Finetuned from model:** [Llama 2](https://arxiv.org/abs/2307.09288)
16
+
17
+ ### Model Sources
18
+
19
+ - **Repository:** https://github.com/lm-sys/FastChat
20
+ - **Blog:** https://lmsys.org/blog/2023-03-30-vicuna/
21
+ - **Paper:** https://arxiv.org/abs/2306.05685
22
+ - **Demo:** https://chat.lmsys.org/
23
+
24
+ ## Uses
25
+
26
+ The primary use of Vicuna is research on large language models and chatbots.
27
+ The primary intended users of the model are researchers and hobbyists in natural language processing, machine learning, and artificial intelligence.
28
+
29
+ ## How to Get Started with the Model
30
+
31
+ - Command line interface: https://github.com/lm-sys/FastChat#vicuna-weights
32
+ - APIs (OpenAI API, Huggingface API): https://github.com/lm-sys/FastChat/tree/main#api
33
+
34
+ ## Training Details
35
+
36
+ Vicuna v1.5 (16k) is fine-tuned from Llama 2 with supervised instruction fine-tuning and linear RoPE scaling.
37
+ The training data is around 125K conversations collected from ShareGPT.com. These conversations are packed into sequences that contain 16K tokens each.
38
+ See more details in the "Training Details of Vicuna Models" section in the appendix of this [paper](https://arxiv.org/pdf/2306.05685.pdf).
39
+
40
+ ## Evaluation
41
+
42
+ Vicuna is evaluated with standard benchmarks, human preference, and LLM-as-a-judge. See more details in this [paper](https://arxiv.org/pdf/2306.05685.pdf) and [leaderboard](https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard).
43
+
44
+ ## Difference between different versions of Vicuna
45
+ See [vicuna_weights_version.md](https://github.com/lm-sys/FastChat/blob/main/docs/vicuna_weights_version.md)