iTPN
iTPN copied to clipboard
sunsmarterjie
Integrally Pre-Trained Transformer Pyramid Networks
很棒的工作，恭喜！ 有几个问题想请教一下： 1. 在进行itpn_clip中的预训练时，即执行以下指令时，--blr与--beta参数似乎并不存在；此外，在engine_for_pretraining.py的第62行中，计算loss_1 = loss_fn(input=outputs[0], target=labels)时，有报错"RuntimeError: The size of tensor a (512) must match the size of tensor b (768) at non-singleton dimension 2" ` python -m torch.distributed.launch --nproc_per_node=8...
Congratulations for you great work ! It seems that the Google Drive link for the checkpoint iTPN-B-pixel is wrong because it actually downloads a iTPN-L-pixel. Would it be possible to...
Hello, when will the code related to masked feature modeling be released?
采用如下命令安装，后面跟着网址。 pip install torch==1.7.1+cu110 torchvision==0.8.2+cu110 torchaudio==0.7.2 -f https://download.pytorch.org/whl/torch_stable.html 另外，start那里，如果pip版本大于等于23.1，那么用另一个命令。 git clone https://github.com/NVIDIA/apex cd apex # if pip >= 23.1 (ref: https://pip.pypa.io/en/stable/news/#v23-1) which supports multiple `--config-settings` with the same key... pip...
Hello, it is a great job for iTPN. We are very interested in it. Currently, the inference delay of visual foundation models is hindering the applications in industry. I am...
I find that some Scholars study crowd counting on basis of PVT (Pyramid Vision Transformer). So can I use iTPN to study crowd counting?
hello！I directly run the main_finetue.py, but miss the module named 'jit_analysis' . So my operation is wrong？
Hello, thanks for your work and congrats for the ICRA2023 acceptation, it was really interesting to read. I am experimenting with iTPN for self-supervised learning, but the checkpoint provided (`itpn_base_pixel.pth`...
Hello! I think there are some files missing in CLIP_as_supervision. When I want to run pretrain task of iTPN with CLIP as supervision, I find "iTPN_clip_distill" is missing. Can you...
Metadata
Owner
Metadata
Integrally Pre-Trained Transformer Pyramid Networks