Learn More →
創造一個 conda for PyTorch 環境
conda create -n pytorch python=3.7
切進環境
conda activate pytorch
Learn More →
Learn More →
conda activate pytorch
conda install pytorch torchvision torchaudio cudatoolkit=10.2 -c pytorch
安裝後驗證
conda activate pytorch
python
import torch
Learn More →
import torchvision
torchvision.__version__
Learn More →
驗證 CUDA 是否可用
torch.cuda.is_available()
torch.cuda.get_device_name()
Learn More →
torch.cuda.device_count()
Learn More →
CUDA ID
torch.cuda.current_device()
Learn More →
CUDA 裝置名稱
torch.cuda.get_device_name(0)
Learn More →
pip install jupyterlab
conda install -c conda-forge matplotlib -n pytorch
conda install -c anaconda pandas -n pytorch
pip install torchsummary
pip install torch
pip install opencv-python
pip install opencv-contrib-python
setup
, Python
# README.md 來源
Jan 21, 2024NTUT VoTT Main Features of NTUT VoTT NTUT_VoTT tool is a modified version of Microsoft VoTT (based on Version: V2.2.0) open-source labeling software. It has the following new features: Automatic tracking after labeling, to reduce the tedious work of re-labeling each frame Multi-core tracking Track FPS for different scenarios :::info The below video shows NTUT VoTT how to save your time for labeling objects, you don't need to label those objects each frame, and this tool supports multi-core tracking which can speed up while you tracking more objects. :::
Dec 28, 2023[TOC] NTUT VoTT Main Features of NTUT VoTT NTUT_VoTT tool is a modified version of Microsoft VoTT (based on Version: V2.2.0) open-source labeling software. It has the following new features: Automatic tracking after labeling, to reduce the tedious work of re-labeling each frame Multi-core tracking Track FPS for different scenarios :::info
Dec 14, 2022PaperV2 FeedBack [TOC] Reviewer2 Reviewer3 Question1 You mentioned that some of tools like DarkLabel cannot track more than 100 objects in one frame (Fig10). What is your performance in that kind of frame? What is the FPS? Answer1
Jul 11, 2022or
By clicking below, you agree to our terms of service.
New to HackMD? Sign up