facebookresearch dino issues

facebookresearch / dino

PyTorch code for Vision Transformers training with the Self-Supervised learning method DINO

Apache License 2.0

6.04k stars 884 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Cannot reproduce KNN performance for vanilla ViT-S training

#280 wangh09 opened 4 days ago
1
Choice of out_dim

#279 JunboShen opened 5 days ago
0
Can I use dino model to match images with pixels? Just like with the clip model you can match pixels with text.

#278 dajiebi opened 1 month ago
0
add memory-efficient attention

#277 Andy-Ko-0620 opened 1 month ago
2
How to convert to onnx

#276 Ochre-amber opened 2 months ago
1
get cls block

#275 ShovalMishal opened 2 months ago
1
Why Teacher network perform bettert than Student one during training?

#274 Hongbo-Z opened 2 months ago
0
Loss is not decresing

#273 risingClouds opened 3 months ago
3
Best features for image similarity

#272 fbliman opened 3 months ago
1
mis pull req

#271 RyuAmakaze closed 4 months ago
1
x

#270 philippfeigl closed 4 months ago
2
a solution to solve memory issues (but slows down training a bit)

#269 fawazsammani opened 6 months ago
0
Update eval_image_retrieval.py

#268 anonymousdouble opened 6 months ago
1
Why do we skip cases where the student and teacher operate on the same view? If they are operating on different views, why should they produce similar results to calculate the cross-entropy loss?

#267 jinghere11 opened 6 months ago
1
Why not setting correct img_size when building the student network?

#266 LichunZhang closed 6 months ago
1
For large batches (256), there is a problem of loss non convergence

#265 Elijah-Yi opened 6 months ago
0
Difference between DINO and DINOv2

#264 Tulsishah opened 7 months ago
0
Loss Nan Error

#263 Elijah-Yi opened 7 months ago
0
Changes to support pytorch2.0

#262 Tulsishah opened 7 months ago
2
Why the patchembedding defaut img size is not equal to the image size in visualize attention?

#261 LWShowTime opened 7 months ago
1
Facing Accuracy Drop while running in QDQ onnx runtime

#260 akarshit15092000 opened 7 months ago
0
Intermediate checkpoints?

#259 ghost opened 7 months ago
0
Prefix python modules

#258 SomeoneSerge opened 8 months ago
0
Detached

#257 punims closed 8 months ago
1
issue with the ViT-S/8 full checkpoint

#256 melikaqh closed 8 months ago
0
Hyperparameters for ViT-B/16 + ImageNet pretraining

#255 kalyani7195 opened 9 months ago
0
The averaged learning rate in log.txt

#254 haohang96 opened 9 months ago
0
starting from checkpoint

#253 hyunjung00 closed 9 months ago
0
Image resolution at intermediate layers?

#252 yousafe007 opened 9 months ago
3
problem when trying to finetune a model, using resnet50 backbone

#251 saeidrazavi opened 9 months ago
1
Question regarding patch retrieval demo

#250 crypdick closed 2 months ago
2
Linear Evaluation LR Sweep Values

#249 ferreirafabio opened 11 months ago
1
dinohead weight normalized

#248 kffeng opened 12 months ago
0
Faster training

#247 rbareja25 opened 1 year ago
2
Loss doesn't decrease

#246 AwayFromKeyboardwarrior opened 1 year ago
6
How to use teacher weights?

#245 zahrasa opened 1 year ago
0
A question about DINOHead

#244 wenhaoli-xmu opened 1 year ago
1
torch hubconf.py is missing a function

#243 nkondapa opened 1 year ago
3
RuntimeError: CUDA error: CUBLAS_STATUS_EXECUTION_FAILED when calling `cublasSgemm( handle, opa, opb, m, n, k, &alpha, a, lda, b, ldb, &beta, c, ldc)`

#242 PapaMadeleine2022 opened 1 year ago
0
when extracting features from image, i see that the program use the teacher network, can I use the student network, if using the sutdent network, weather the performance will be reduced?What's the difference between the feature of student network and the feature of teacher network？

#241 yyy809643573 opened 1 year ago
0
how to train DINO with MSE loss.

#240 Backdrop9019 opened 1 year ago
0
How to get the feature map by using pretrained model.

#239 JiuqingDong opened 1 year ago
0
Point to DINOv2

#238 patricklabatut closed 1 year ago
1
How to concatenate attention maps from different heads

#237 LLL-YUE opened 1 year ago
1
Could you tell me how to visualize PCA 3-channel image just like the showcase in your README

#236 shoutOutYangJie closed 1 year ago
0
How does a zero KLD indicate a collapse?

#235 YoojLee opened 1 year ago
0
normalize knn probabilities

#234 clemsgrs opened 1 year ago
0
sock.connect(sa)TimeoutError : raise err[Errno 110] Connection timed out

#233 wdc233 opened 1 year ago
1
Memory Occupancy

#232 develop-productivity opened 1 year ago
0
Training on uncurated/unbound datasets

#231 ajkailash opened 1 year ago
0