rabum

文本向量的存储和读取

1

长文本向量化后需要存储，节省之后每次计算的时间，请问官方有没有像gensim的save()那样保存词向量的方法呢

image_demo运行起来太慢了，有没有连续快速图片识别的方法

13

有没有能秒出结果的办法？

How do I turn off the default output

1

When I use results = model(img), it will print img size, all the detection results and time used automatically, how do I turn it off?

Yolo-pose output

1

I see those occluded keypoints are set to [0,0], is it a postprocess? If so can I get their original xy the model predicted? 

Internvideo2.5显存占用情况

8

我想在一张4090 24g上跑InternVideo2_5_Chat_8B，报段错误(核心已转储)，是不是显存不够？

No module named 'torch' when pip install -e .

8

When I tried to install the project with line pip install -e . there seems to be a problem: DEPRECATION: Legacy editable install of detectron2==0.6 from file:///home/yunzheng/devit (setup.py develop) is...

I'm trying to rebuild the prototypes in config yamls with the code provided, but it seems the keys are not the same as the .pkl and .pth files provided in...

how to improve inference time

3

My code is: ``` def extract_global_vector(backbone, img_tensor, device): x = img_tensor.unsqueeze(0).to(device) with torch.inference_mode(): feats = backbone.get_intermediate_layers(x, n=range(12), reshape=True, norm=False) last = feats[-1].squeeze(0) if last.ndim == 3: vec = last.view(last.shape[0], -1).mean(dim=1)...

If I open the mod manager offline, it will not respond and lead to a crash, I guess it was looking up for update no matter what? Hope this problem...