MeshToad V3 turns Linux systems into Meshtastic nodes
https://fed.brid.gy/r/https://linuxgizmos.com/meshtoad-v3-turns-linux-systems-into-meshtastic-nodes/
MeshToad V3 turns Linux systems into Meshtastic nodes
https://fed.brid.gy/r/https://linuxgizmos.com/meshtoad-v3-turns-linux-systems-into-meshtastic-nodes/
Krea (@krea_ai)
Krea 2에 LoRA 기반 파인튜닝 기능(beta)을 도입. 사용자가 자신의 스타일, 객체, 캐릭터에 맞춰 Krea 2를 정밀하게 학습시킬 수 있는 시스템으로, 이미지 생성/커스텀 모델 워크플로우에서 실무 활용성이 높은 업데이트다.
In these pleasant (relatively) days of late spring, it's probably a good idea to follow the EFF's advice and figure out how to set up emergency communications for when things go very wrong, and there's no internet, and no cell phone service.
I've done a lot of thinking about this the last year or so, and actually set up, tested, and used three different approaches, and here I share that with you:
#EmergencyPrep #EmergencyPreparedness #radio #GMRS #FRS #AmateurRadio #LoRa #MeshCore #Meshtastic #HamRadio #cascadia #seattle #oregon
💫 Call for Hardware (for MeshCore dev)
Der Entwickler von Kiekr (ziemlich coole Meshcore app) sucht nach alten Android-Smartphones für ein dauerhaftes Test-Setup. Falls du alte/obsolete Hardware rumliegen hast und die Weiterentwicklung von Meshcore unterstützen möchtest, melde dich: https://kiekr.app
Akku: heil oder kaputt
Display: zerkratzt/gesprungen, aber sollte funktionieren.
android: >= 8.0.
#MeshCore #LoRa #HanseMesh #Mesh #Notfunk #EmCom
@meshcore
@blackforestmesh @voyagermesh
@Quantensalat @scott @k8vsy @KC8JC
Should have searched first. Here are a few findings:
https://github.com/tapparelj/gr-lora_sdr
https://events.gnuradio.org/event/24/contributions/641/attachments/192/477/slides_tapparel.pdf
Now off to buy a few “properly licensed” modules to help Scott’s stock return. :-)

This is the fully-functional GNU Radio software-defined radio (SDR) implementation of a LoRa transceiver with all the necessary receiver components to operate correctly even at very low SNRs. This ...
Epiphany: I entered the #TBeamSupreme arena to field-test #LoRa and #Reticulum mesh behavior. That led to building a Bluetooth interface.
The surprise: Bluetooth may let me shrink the test range. Instead of sending people blocks away to force LoRa gaps, I can use walls, houses, and barriers on my own property to create repeatable mesh-resiliency tests.
[Перевод] Нехватка CUDA-памяти при обучении с GRPO: как перестать гадать и начать считать
Ошибка CUDA out of memory при обучении LLM обычно превращается в бесконечный цикл случайных правок: уменьшили batch size, урезали sequence length, снизили LoRA rank — и всё равно где-то снова падает. Особенно весело становится в RL-сценариях с GRPO, vLLM и генерацией нескольких ответов на один промпт. Поговорим о том, как перестать гадать и начать считать потребление GPU-памяти: от чтения самого текста ошибки до оценки вклада vLLM, активаций и параметров обучения. С формулами, реальными конфигами и объяснением, какие настройки действительно дают эффект, а какие только создают иллюзию оптимизации. Оптимизировать LLM
https://habr.com/ru/companies/otus/articles/1037332/
#NLP #LLM #GRPO #обучение_с_подкреплением #CUDA_out_of_memory #vLLM #оптимизация_GPU_памяти #дообучение_моделей #LoRA #PyTorch
A small local model can know plenty and still fail one tiny agent contract.
This tutorial uses Quarkus and MLX LM on a Mac to teach a model a private devcard command with LoRA, then measure the before-and-after difference.