Changyao commited on
Commit
c0fea73
1 Parent(s): d20b571

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +13 -21
README.md CHANGED
@@ -1,6 +1,3 @@
1
- ---
2
- license: apache-2.0
3
- ---
4
  # ADDP
5
  The official implementation of the [paper](https://arxiv.org/abs/2306.05423) "ADDP: Learning General Representations for Image Recognition and Generation with Alternating Denoising Diffusion Process" (ICLR 2024).
6
 
@@ -18,21 +15,6 @@ Image recognition and generation have long been developed independently of each
18
 
19
 
20
 
21
- ## Citation
22
-
23
- If this work is helpful for your research, please consider citing the following BibTeX entry.
24
-
25
- ```
26
- @article{tian2023addp,
27
- title={Addp: Learning general representations for image recognition and generation with alternating denoising diffusion process},
28
- author={Tian, Changyao and Tao, Chenxin and Dai, Jifeng and Li, Hao and Li, Ziheng and Lu, Lewei and Wang, Xiaogang and Li, Hongsheng and Huang, Gao and Zhu, Xizhou},
29
- journal={arXiv preprint arXiv:2306.05423},
30
- year={2023}
31
- }
32
- ```
33
-
34
-
35
-
36
  ## Setup
37
 
38
  Step 1, download [ImageNet](http://image-net.org/download) dataset, and place it in your `IMAGENET_DIR`.
@@ -59,7 +41,7 @@ The following table provides the performance and weights of the pre-trained chec
59
 
60
  | | ViT-Large | ViT-Base |
61
  | ---------------------------------- | ---- | -------- |
62
- | Checkpoint | | |
63
  | Class-unconditional Generation FID | 7.6 | 8.9 |
64
  | Class-unconditional Generation IS | 105.1 | 95.3 |
65
  | Fine-tuning Top-1 Accuracy | 85.9 | 83.9 |
@@ -102,7 +84,6 @@ bash configs/release/base/pretrain_addp_base_1600ep.sh ${GPUS} ${GPUS_PER_NODE}
102
  ## finetune
103
  bash configs/release/base/finetune_addp_base_100ep.sh ${GPUS} ${GPUS_PER_NODE} ${JOB_NAME} ${QUOTATYPE} ${PARATITION} exp/release/addp-vit-base-16.pth
104
 
105
-
106
  ## generate
107
  bash configs/release/base/generate_addp_base_steps20.sh ${GPUS} ${GPUS_PER_NODE} ${JOB_NAME} ${QUOTATYPE} ${PARATITION} exp/release/addp-vit-base-16.pth
108
  ```
@@ -127,7 +108,18 @@ Then use the above package to evaluate FID/IS of the images generated by our mod
127
  fidelity --isc --fid --input1 ${GENERATED_IMAGES_DIR} --input2 ${IMAGENET256X256_DIR}
128
  ```
129
 
 
130
 
 
 
 
 
 
 
 
 
 
 
131
 
132
 
133
  ## License
@@ -136,4 +128,4 @@ This repository is released under the Apache 2.0 license as found in the [LICENS
136
 
137
  ## Contact
138
 
139
- If you have any questions, feel free to contact me through email (tcyhost@link.cuhk.edu.hk) directly.
 
 
 
 
1
  # ADDP
2
  The official implementation of the [paper](https://arxiv.org/abs/2306.05423) "ADDP: Learning General Representations for Image Recognition and Generation with Alternating Denoising Diffusion Process" (ICLR 2024).
3
 
 
15
 
16
 
17
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
18
  ## Setup
19
 
20
  Step 1, download [ImageNet](http://image-net.org/download) dataset, and place it in your `IMAGENET_DIR`.
 
41
 
42
  | | ViT-Large | ViT-Base |
43
  | ---------------------------------- | ---- | -------- |
44
+ | Checkpoint | <a href="https://huggingface.co/Changyao/ADDP/blob/main/addp-vit-large-16.pth">this link</a> | <a href="https://huggingface.co/Changyao/ADDP/blob/main/addp-vit-base-16.pth">this link</a> |
45
  | Class-unconditional Generation FID | 7.6 | 8.9 |
46
  | Class-unconditional Generation IS | 105.1 | 95.3 |
47
  | Fine-tuning Top-1 Accuracy | 85.9 | 83.9 |
 
84
  ## finetune
85
  bash configs/release/base/finetune_addp_base_100ep.sh ${GPUS} ${GPUS_PER_NODE} ${JOB_NAME} ${QUOTATYPE} ${PARATITION} exp/release/addp-vit-base-16.pth
86
 
 
87
  ## generate
88
  bash configs/release/base/generate_addp_base_steps20.sh ${GPUS} ${GPUS_PER_NODE} ${JOB_NAME} ${QUOTATYPE} ${PARATITION} exp/release/addp-vit-base-16.pth
89
  ```
 
108
  fidelity --isc --fid --input1 ${GENERATED_IMAGES_DIR} --input2 ${IMAGENET256X256_DIR}
109
  ```
110
 
111
+ ## Citation
112
 
113
+ If this work is helpful for your research, please consider citing the following BibTeX entry.
114
+
115
+ ```
116
+ @article{tian2023addp,
117
+ title={Addp: Learning general representations for image recognition and generation with alternating denoising diffusion process},
118
+ author={Tian, Changyao and Tao, Chenxin and Dai, Jifeng and Li, Hao and Li, Ziheng and Lu, Lewei and Wang, Xiaogang and Li, Hongsheng and Huang, Gao and Zhu, Xizhou},
119
+ journal={arXiv preprint arXiv:2306.05423},
120
+ year={2023}
121
+ }
122
+ ```
123
 
124
 
125
  ## License
 
128
 
129
  ## Contact
130
 
131
+ If you have any questions, feel free to contact me through email (tcyhost@link.cuhk.edu.hk) directly.