Recherche avancée

Sur d’autres sites (10724)

• libavfi/dnn : add LibTorch as one of DNN backend

  15 mars 2024, par Wenbin Chen

  libavfi/dnn : add LibTorch as one of DNN backend
  PyTorch is an open source machine learning framework that accelerates
  
  the path from research prototyping to production deployment. Official
  
  website : https://pytorch.org/. We call the C++ library of PyTorch as
  
  LibTorch, the same below.
  To build FFmpeg with LibTorch, please take following steps as
  
  reference :
  
  1. download LibTorch C++ library in
  
   https://pytorch.org/get-started/locally/,
  
  please select C++/Java for language, and other options as your need.
  
  Please download cxx11 ABI version :
  
   (libtorch-cxx11-abi-shared-with-deps-*.zip).
  
  2. unzip the file to your own dir, with command
  
  unzip libtorch-shared-with-deps-latest.zip -d your_dir
  
  3. export libtorch_root/libtorch/include and
  
  libtorch_root/libtorch/include/torch/csrc/api/include to $PATH
  
  export libtorch_root/libtorch/lib/ to $LD_LIBRARY_PATH
  
  4. config FFmpeg with ../configure —enable-libtorch \
  
   —extra-cflag=-I/libtorch_root/libtorch/include \
  
   —extra-cflag=-I/libtorch_root/libtorch/include/torch/csrc/api/include \
  
   —extra-ldflags=-L/libtorch_root/libtorch/lib/
  
  5. make
  To run FFmpeg DNN inference with LibTorch backend :
  
  ./ffmpeg -i input.jpg -vf \
  
  dnn_processing=dnn_backend=torch:model=LibTorch_model.pt -y output.jpg
  The LibTorch_model.pt can be generated by Python with torch.jit.script()
  
  api. https://pytorch.org/tutorials/advanced/cpp_export.html. This is
  
  pytorch official guide about how to convert and load torchscript model.
  
  Please note, torch.jit.trace() is not recommanded, since it does
  
  not support ambiguous input size.
  Signed-off-by : Ting Fu <ting.fu@intel.com>
  
  Signed-off-by : Wenbin Chen <wenbin.chen@intel.com>
  
  Reviewed-by : Guo Yejun <yejun.guo@intel.com>
  

  • [DH] configure
  • [DH] libavfilter/dnn/Makefile
  • [DH] libavfilter/dnn/dnn_backend_torch.cpp
  • [DH] libavfilter/dnn/dnn_interface.c
  • [DH] libavfilter/dnn_filter_common.c
  • [DH] libavfilter/dnn_interface.h
  • [DH] libavfilter/vf_dnn_processing.c

• Matomo will now pay researchers 5,000 USD for a critical security vulnerability

  7 mai 2020, par Matomo Core Team
  Matomo Analytics is the leading open-source web analytics solution, designed to give you conclusive insights while respecting your user’s privacy, and keeping your data secure. We’re so proud Matomo is trusted with the analytics data of more than 1 million sites worldwide.

  Although we have had an excellent security track record so far, we recognise security is an ongoing challenge and requires constant vigilance. With this announcement we’re showing our commitment to reward those who help us maintain the highest security in Matomo.

  New bounty of 5,000 USD for a CRITICAL security issue responsibly disclosed to us

  We’re now paying 5,000 USD or 4,700 EUR for each critical vulnerability found, and responsibly disclosed to us. (Previously this bounty was less than 1,000USD.) 

  A Critical Issue in Matomo means an issue in our latest official release at : builds.matomo.org/latest.zip as installed on a typical server (and possibly using any of our official plugins by Matomo or InnoCraft from the Marketplace).

  If you can gain remote code execution on the server (i.e. RCE), or if you’re able to delete data with an HTTPS request (i.e. SQL Injection), this may qualify as a Critical Issue. Please report it on Hackerone.

  Matomo keeps your data secure

  The Matomo team has always been committed to achieving the highest standard of security. For example, Matomo was one of the first open-source projects in the world to launch a public bug bounty in January 2011. Every year many researchers, users and customers review the Matomo source code, and overall we’ve rewarded dozens of researchers over the years for their work in keeping Matomo data safe.

  How to make your Matomo server even more secure ?

  Check out our recommendations in How to configure Matomo for Security

   

• Displaying an AVFrame on the screen with SDL 2.0

  22 septembre 2013, par jsp99

  I am working on some code with the help of this tutorial and using the latest development libraries of ffmpeg and SDL. I am stuck at the point where I have to display the decoded frame (AVFrame) on the screen. I am inclined to do the above task i.e, Displaying the AVFrame on screen using the latest API of SDL 2.0 (Using Renderers and Textures alongside the usage of SDL_Window). Frankly speaking, I am not an expert in SDL_Renderer, SDL_Texture and the functions associated with them. But I am reading the documentation in the official site of SDL 2.0 and working my way through them.

  Is there a way to do the following using SDL 2.0 API :

  • Convert the native frame format to a flavour of YUV and display it.

    (OR)

  • If it is possible, display the frame without having to convert it from native format.

  I want to do the above using Renderers and Textures. There doesn't seem to be an easy way to work with them.

  Can anyone guide me through the steps to do the above tasks ?

  PS : Though I have not explicitly tried it, I came across some ways to display AVFrame on the screen by converting the AVFrame format(native) to RGB. But I do not want the native frame format (which is mostly YUV) to RGB conversion, as it is computationally expensive.

  Converting between formats is done by sws_scale()