search

wusize / CLIPSelf

[ICLR2024 Spotlight] Code Release of CLIPSelf: Vision Transformer Distills Itself for Open-Vocabulary Dense Prediction
https://arxiv.org/abs/2310.01403
Other
170 stars 9 forks source link

Problem happend when perform "test_eva_vitb16_macc_boxes_masks.sh " #20

Closed JuanJia closed 7 months ago

JuanJia commented 7 months ago

Hi, thanks for your great work.

When I try to reproduce the results using command below test_eva_vitb16_macc_boxes_masks.sh I ran into the errorAttributeError: 'NoneType' object has no attribute 'memory_efficient_attention, and refers to this line in forward x = xops.memory_efficient_attention( CLIPSelf/src/open_clip/eva_clip/eva_vit_model.py", line 211 Is there any bug in this code or what should I do to avoid this bug?

Thanks.

———————————————————————————————————————— loading annotations into memory... Done (t=0.31s) creating index... index created! 2024-04-10,08:09:54 | INFO | Evaluate before training 2024-04-10,08:09:54 | INFO | Region classifier 0%| | 0/5000 [00:00<?, ?it/s]2024-04-10,08:10:40 | INFO | Add a new rope freq of shape: torch.Size([4096, 64]) Add a new rope freq of shape: torch.Size([4096, 64]) 0%| | 0/5000 [00:46<?, ?it/s] Traceback (most recent call last): File "/home/nkd-434/anaconda3/envs/clipself/lib/python3.9/runpy.py", line 197, in _run_module_as_main return _run_code(code, main_globals, None, File "/home/nkd-434/anaconda3/envs/clipself/lib/python3.9/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/training/main.py", line 346, in main(sys.argv[1:]) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/training/main.py", line 267, in main evaluate(model, data, start_epoch, args) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/training/train.py", line 172, in evaluate zero_shot_metrics = zero_shot_eval(model, data, epoch, args) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/training/zero_shot.py", line 188, in zero_shot_eval all_box_sizes, all_is_thing, all_cls_labels = run(model, data['val'].dataloader, args) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/training/zero_shot.py", line 72, in run roi_features = roi_extractor(images, rois, normalize=True, File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/open_clip/eva_clip/model.py", line 337, in encode_pseudo_boxes features = self.visual.extract_roi_features(image, normed_boxes, extract_type=extract_type) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/open_clip/eva_clip/eva_vit_model.py", line 626, in extract_roi_features x = self.encode_dense(x, keep_shape=True) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/open_clip/eva_clip/eva_vit_model.py", line 614, in encode_dense x = blk(x, rel_pos_bias=rel_pos_bias) File "/home/nkd-434/anaconda3/envs/clipself/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl return self._call_impl(*args, kwargs) File "/home/nkd-434/anaconda3/envs/clipself/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl return forward_call(*args, *kwargs) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/open_clip/eva_clip/eva_vit_model.py", line 306, in forward x = x + self.drop_path(self.attn(self.norm1(x), rel_pos_bias=rel_pos_bias, attn_mask=attn_mask)) File "/home/nkd-434/anaconda3/envs/clipself/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl return self._call_impl(args, kwargs) File "/home/nkd-434/anaconda3/envs/clipself/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl return forward_call(*args, **kwargs) File "/home/nkd-434/Documents/jjy/MLC/CLIPSelf/src/open_clip/eva_clip/eva_vit_model.py", line 211, in forward x = xops.memory_efficient_attention( AttributeError: 'NoneType' object has no attribute 'memory_efficient_attention'

zhiyustar commented 7 months ago

hello,have you solved this? I met this problem too.

Valeria-MaW commented 4 months ago

hello,have you solved this? I met this problem too.