#python#large_language_models#machine_learning_systems#natural_language_processing
Flash Linear Attention (FLA) is a fast, memory-efficient library for advanced linear attention models used in transformers, written in PyTorch and Triton, and compatible with NVIDIA, AMD, and Intel GPUs. It offers many state-of-the-art linear attention models and fused modules that speed up training and reduce memory use. You can easily replace standard attention layers in your models with FLA’s efficient versions, improving training and inference speed, especially for long sequences. FLA supports hybrid models mixing linear and standard attention, and integrates with Hugging Face Transformers for easy use and evaluation. This helps you train and run large language models faster and with less memory, making your AI projects more efficient and scalable.
https://github.com/fla-org/flash-linear-attention
DataBackup
An easy-to-use backup Libre application for Android, however your phone needs to be rooted , unless a ROM were to include it as a system app
Support for #multi-user/double backup: same backup/restore on multiple partitions!
Cloud: fully supports #Rclone, which can perform local mounting of network drives from various service providers.
100% data integrity: all data will be retained and there is no need to reconnect or download additional packages.
Complete: Split Apk, Arm32, Arm64, x86, x86_64, Android9+.
Fast: Support: tar lz4 zstd (default).
GitHub - XayahSuSu/Android-DataBackup: 数据备份 DataBackup for Android - https://github.com/XayahSuSuSu/Android-DataBackup
Reminder : to install it via F-Droid first add the Izzy repository to F-Droid ( settings - repositories/sources)
IzzyOnDroid F-Droid Repository -
https://apt.izzysoft.de/fdroid/
#Backup #Android
Rclone Remount
Remount cloud storage locally during boot via rclone & fusermount directly on your Android powered smart device.
Virtually limitless storage expansion with support for dozens of cloud providers including Dropbox, GDrive, OneDrive, SFTP & many more. Extremely useful for devices without physical storage expansion capabilities. Also great for streaming large media files without need for full caching. Binaries compiled using Termux.
Features:
Support for arm, arm64, & x86
Huge list of supported cloud storage providers
Apps with ability to specify paths can access /mnt/cloud/
Most file explorers work just fine (issue #9)
Mount points use names of remote(s) in rclone.conf
Specify custom rclone params for each remote via /sdcard/.rclone/.REMOTE.param
Access remotes via http://127.0.0.1:38762
Access remotes via ftp://127.0.0.1:38763
Mount bind to /sdcard/ (see issue #5)
Support for Work Profiles
https://github.com/Magisk-Modules-Repo/com.piyushgarg.rclone
#rclone#remount#cloud#alternatives
RCX
Android GUI for rclone. Manage files across many cloud services in a simple way under one interface.
https://x0b.github.io
https://f-droid.org/packages/io.github.x0b.rcx
https://github.com/x0b/rcx
@nogoolag
#rcx#rclone#cloud#alternatives