michalk8 commited on
Commit
ba28fb8
1 Parent(s): ad6ac86

Upload files

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ aimv2_overview_light.png filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -1,3 +1,35 @@
1
- ---
2
- license: apple-ascl
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: apple-ascl
4
+ pipeline_tag: zero-shot-image-classification
5
+ tags:
6
+ - vision
7
+ - mlx
8
+ - pytorch
9
+ ---
10
+ # Introduction
11
+ [[`AIMv2 Paper`](#)] [[`BibTeX`](#citation)]
12
+
13
+ We introduce the AIMv2 family of vision models pre-trained with a multimodal autoregressive objective.
14
+ AIMv2 pre-training is simple and straightforward to train and to scale effectively. Some AIMv2 highlights include:
15
+
16
+ 1. Outperforms OAI CLIP and SigLIP on the majority of multimodal understanding benchmarks.
17
+ 2. Outperforms DINOv2 on open-vocabulary object detection and referring expression comprehension.
18
+ 3. Exhibits strong recognition performance with AIMv2-3B achieving *89.5% on ImageNet using a frozen trunk*.
19
+
20
+ <img src="aimv2_overview_light.png" alt="AIMv2 Overview"/>
21
+
22
+ ## Usage
23
+ Under construction. Please consider using the models in the [ml-aim](https://github.com/apple/ml-aim) repository.
24
+
25
+ ## Citation
26
+ If you find our work useful, please consider citing us as:
27
+ ```bibtex
28
+ @misc{fini2024multimodal,
29
+ title = {Multimodal Autoregressive Pre-training of Large Vision Encoders},
30
+ author = {Enrico Fini and Mustafa Shukor and Xiujun Li and Philipp Dufter and Michal Klein and David Haldimann and Sai Aitharaju and Victor Guilherme Turrisi da Costa and Louis Béthune and Zhe Gan and Alexander T Toshev and Marcin Eichner and Moin Nabi and Yinfei Yang and Joshua M. Susskind and Alaaeldin El-Nouby},
31
+ year = {2024},
32
+ archivePrefix = {arXiv},
33
+ primaryClass = {cs.CV},
34
+ }
35
+ ```
aimv2_overview_light.png ADDED

Git LFS Details

  • SHA256: 524b6eb5049fb4bac6303ecee386d0e885fa69a96756557d843084ba4caae08f
  • Pointer size: 131 Bytes
  • Size of remote file: 336 kB
flax_model.msgpack ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e9a0949dabf81c8c75891cef7cb624140b82570da0922cd190926dd623e62194
3
+ size 1746731764
mlx_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:503fbcc7e77654e026b0103d7d8d663362cba93176ec428e3405fbd24abd404f
3
+ size 1746752303
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c218dc23f746407cacb100506d83a1c5f169db11ff3cf141853ac28771af3222
3
+ size 1746752308