TGTGInsighttelegram intelligenceLIVE / telegram public index
← GitHub Trends

TGINSIGHT SIMILAR POSTS

Find similar content

Source channel @githubtrending · Post #15141 · Sep 13

#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

Results

2 similar posts found

Search: #beechat

当前筛选 #beechat清除筛选
Libreware

@libreware · Post #1007 · 11/05/2021, 12:54 AM

We just pushed the V1.2 of the Beechat Crank to our GitHub @beechatnetwork2 Beechat Crank is a human-powered, open-hardware, fully-waterproof, electricity generator. The project was born out of the need to power electronics in all off-grid scenarios. We needed a way to power our devices in all weather conditions, and making a human-powered generator meant that problem could be solved. When we looked for a solution, we didn't find what we were looking for, so we decided to make it, and make it open source! https://github.com/BeechatNetworkSystemsLtd/BeechatCrank/ Features High power (3A) USB-C IPX8 output 18650 3600 mAh battery (LG M36) Charging time of 4 hrs 24 min. to full charge with 10% efficiency loss at 4.5 Watts. Overcharge protection Parts Beechat Crank is composed of the main device, which we call the Heart. It contains the battery, the generator and all the electronics (the handle is detachable) The Pedal Case is an assembly which the Heart can be attached in. which dramatically speeds up the charging from 2.5 hours to 25-30 minutes. This is achieved with reduction gears to, with the same revolutions per minute, generate more electricity. Generator part We have chosen the 35L048B1B motor with a mass of 88 grams, giving us us 450 mA for each phase of the coil and works at 5V. The wire leads are connected via female Molex connectors to the charging PCB board. New modifications: * Added a Molex type connector to the PCB to connect the motor. * The Zener diodes were changed, by means of a rectifier bridge. * A new and improved Texas Instruments brand battery management chip was selected. * Added a boost converter to allow different voltage batteries. * New Portescap brand two-phase generator with 90% efficiency and peak power of 4.5 Watts was selected. #crank#beechat

Libreware

@libreware · Post #1295 · 06/21/2024, 11:55 AM

Last weekend we presented the Beechat Rune device for the very first time, as part of the Volla Community days in Germany. Featuring 4 simultaneous streams, capable of audio and video streams with next-gen CRYSTALS cryptography, the Beechat Rune is a perfect addition to privacy-focused systems. @beechatnetwork2 #beechat#mesh#p2p