liuhaotian commited on
Commit
6ba60f0
1 Parent(s): 65d618b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +44 -0
README.md CHANGED
@@ -1,3 +1,47 @@
1
  ---
2
  license: apache-2.0
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: apache-2.0
3
+ inference: false
4
  ---
5
+
6
+ **NOTE: This "delta model" cannot be used directly.**
7
+ Users have to apply it on top of the original LLaMA weights to get actual LLaVA weights.
8
+ See https://github.com/haotian-liu/LLaVA#llava-weights for instructions.
9
+
10
+ <br>
11
+ <br>
12
+
13
+ # LLaVA Model Card
14
+
15
+ ## Model details
16
+
17
+ **Model type:**
18
+ LLaVA is an open-source chatbot trained by fine-tuning LLaMA/Vicuna on GPT-generated multimodal instruction-following data.
19
+ It is an auto-regressive language model, based on the transformer architecture.
20
+
21
+ **Model date:**
22
+ LLaVA-Lightning was trained in May 2023.
23
+
24
+ **Paper or resources for more information:**
25
+ https://llava-vl.github.io/
26
+
27
+ **License:**
28
+ Apache License 2.0
29
+
30
+ **Where to send questions or comments about the model:**
31
+ https://github.com/haotian-liu/LLaVA/issues
32
+
33
+ ## Intended use
34
+ **Primary intended uses:**
35
+ The primary use of LLaVA is research on large multimodal models and chatbots.
36
+
37
+ **Primary intended users:**
38
+ The primary intended users of the model are researchers and hobbyists in computer vision, natural language processing, machine learning, and artificial intelligence.
39
+
40
+ ## Training dataset
41
+ 595K filtered image-text pairs from CC3M.
42
+ 150K GPT-generated multimodal instruction-following data.
43
+
44
+ ## Evaluation dataset
45
+ A preliminary evaluation of the model quality is conducted by creating a set of 90 visual reasoning questions from 30 unique images randomly sampled from COCO val 2014 and each is associated with three types of questions: conversational, detailed description, and complex reasoning. We utilize GPT-4 to judge the model outputs.
46
+ We also evaluate our model on the ScienceQA dataset. Our synergy with GPT-4 sets a new state-of-the-art on the dataset.
47
+ See https://llava-vl.github.io/ for more details.