If Moonlight returns 521 error, update not only the driver itself, but also update the Geforce Experience as well.
| <?xml version="1.0" encoding="UTF-8"?> | |
| <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd"> | |
| <plist version="1.0"> | |
| <dict> | |
| <key>bundleIdentifier</key> | |
| <string>com.miHoYo.hkrpg</string> | |
| <key>buttonModels</key> | |
| <array> | |
| <dict> | |
| <key>keyCode</key> |
Since macOS 10.15, PPTP support has been dropped and no longer available. But I have to connect to a server throught VPN that only supports PPTP or L2TP without IPSec (No shared key or credentials). I tried the following ways:
- There is no selection of PPTP in the Settings.
- vpn-pptp-macos but PPTP.ppp has already been removed.
- VPN Tracker 365 is expensive.
- Activate VPN on the router if possible.
- Use a Windows machine as the mobile hotspot. First start the mobile hotspot, connect the macbook to this hotspot. Then start the VPN on the Windows machine. Then in the router settings, in the VPN Connection Properities, enable Internet connection sharing to allow other network users to connect through this computer's Internet connection, select the adapter corresponding to the hotspot. You may need to repeat this process when restarting the Windows
To install Tensorboard on Mac M1, you need to install grpcio from grpcio-mac-arm-build first.
The official guide talks a little about how to setup TVMC (a simple interface) over RPC, there are some pitfalls if you are not careful.
You need to install the full tvm on the host machine by make -j8. On the target machine, you need to install the runtime by make runtime.
Then, setup the python environment properly.
The cache folder of huggingface/transformers in in ~/.cache/huggingface. Clear it when the downloading procedure is down.
- Install QEMU.
- Since the recent Raspberry Pi OS has removed the preset user, use a
legacy version like
2022-01-28will work. Download the Raspberry Pi OS and extract these files from the image:
bcm2710-rpi-3-b.dtbkernel8.img
- Resize the image to 8G to avoid the disk error. Since the disk space has to be the square of 2 and the image itself occupies 3.8G already. If you want to make more actions, please allocate a larger space.
qemu-img resize 2022-01-28-raspios-bullseye-arm64.img 8GTVM is an open source machine learning compiler framework. The official guide gives some information about how to install it for macOS M1 chip. But there are still some pitfalls when installing. The setup for this tutorial is on MacBook Pro M1, macOS Ventura (13.0).
It is not recommended to just use pip to install tvm since the version is not compiled for apple chip. To maximize the availability, compile the source and then install it. You could download the released version of TVM source from download page, I use 0.10.0 in this tutorial.
Though in theory, you could git clone the source code from the offical GitHub repository. But it is tedious to download the submodules of this project -- they are just too many.
- Use Raspberry Pi Imager to install the Raspberry Pi OS (maybe 64-bit version). Change the setting before burning:
- Use password to login with a new user. You have to setup a new user in the newer version of Raspberry Pi OS. Otherwise, SSH will refuse to connect.
- Setup WiFi connection: SSID, password and region (e.g. 'CN'). Warning, use 2.4GHz band instead of 5GHz band for the most of the time.
- Open SSH. This is closed by default.
- Set the hostname.
raspberrypi.localby default.
- In a computer that connects to the same router, use command
ping [hostname](e.g.ping raspberrypi.local) to get the IP address of the Raspberry Pi. - Connect to the Raspberry Pi by SSH.
- Use command
ssh [username]@[IP]. - Add new host to the local computer. Sometimes, you reinstall the OS on the Raspberry Pi, the key validation might get broken. In this scenario, you should delete the previous known host first in your
~/.ssh/known_hosts. - Enter the passw