Duan Zhichao

编译时提示 > 无法打开包括文件："vulkan/vulkan.h": No such file or directory" 解决办法：将 vulkan 安装目录下 Include 文件夹中的 vulkan 文件夹整个复制到 lite.ai.toolkit 目录下（就像复制 onnxruntime 的 include 文件夹那样）。

> Thanks for your interest in our work. As we addressed in the Section 4.1 of our [paper](https://arxiv.org/pdf/2312.02145.pdf), we use gradient accumulation to account for the memory pressure during training:...

> @kingpeter2015 thanks，i have done it. @1234269 How did you solve it? please help me. I have a lot of difficulties, please see here #6

We just added support for batch inference. See the parameter `batch_size` in `mono/tools/test_scale_cano.py`. We have only implemented batch data parallelization for model inference, but not for other pre/post processing due...

Sure. The torchhub model only supports batch inference. I use batchsize=1 in the example: https://github.com/YvanYin/Metric3D/blob/fd90d0dad05d913c4f45ed7a9789b8f1dfde22e0/hubconf.py#L179

Essentially the same, first record the camera intrinsic, then infer the predicted depth, and finally convert the predicted depth from the canonical camera to the real intrinsic camera.

We provide an example of a json file here: https://github.com/YvanYin/Metric3D/blob/main/training/data_server_info/annos_test_normal_nyu_example.json. But for some larger datasets, like Matterport3D, we store the metadata for each sample as a `.pkl`, which includes something...

Can I reproduce the results of the paper by using the default options? https://github.com/woshidandan/TANet/blob/d911967eede3fc8a526d5a43bf5359aece187a1c/code/AVA/option.py#L3-L28 I apologize for not knowing much about NNI, my previous understanding of it was: "find a...

I see, thank you!

Hi, I think I really understand nni now, it's just normal hyperparameter search. nni uses different hyperparameters to conduct experiments, and supervises the experimental process and results to get better...