Just finished this local tts system called yaptts ๐
its nice for low res laptops with no gpu ๐
it is made to be simple to use self healing concepts so anyone can use it
hopfully can work in both linux (i think) and windows ๐ค
need internet only in the first run to download and cache the models then wont care or ask for internet to work ๐
have amharic text to tone converter so you can test amharic text in any language mode withn the app and hear them แฒแฎแแฐแ ๐๐
https://github.com/laladigitalsystem-oss/yaptts
This is a hobby project i do to learn in depth about tts and AI in general while keeping it reacabel and grounded for Ethiopic enviroment.
#yap #my_product #project
its nice for low res laptops with no gpu ๐
it is made to be simple to use self healing concepts so anyone can use it
hopfully can work in both linux (i think) and windows ๐ค
need internet only in the first run to download and cache the models then wont care or ask for internet to work ๐
have amharic text to tone converter so you can test amharic text in any language mode withn the app and hear them แฒแฎแแฐแ ๐๐
https://github.com/laladigitalsystem-oss/yaptts
This is a hobby project i do to learn in depth about tts and AI in general while keeping it reacabel and grounded for Ethiopic enviroment.
#yap #my_product #project
โค3
Yared's log ๐ซ
Just finished this local tts system called yaptts ๐ its nice for low res laptops with no gpu ๐ it is made to be simple to use self healing concepts so anyone can use it hopfully can work in both linux (i think) and windows ๐ค need internet only in theโฆ
Audio
แแฎแแฐแ แญแแณแ แฅแแฒ แแ
Yared's log ๐ซ
Just finished this local tts system called yaptts ๐ its nice for low res laptops with no gpu ๐ it is made to be simple to use self healing concepts so anyone can use it hopfully can work in both linux (i think) and windows ๐ค need internet only in theโฆ
I dont know why peopls hype opus i have tried both opus and gpt 5.2 and 5.4 in this project and in all case gpt out perform opus
AI Post โ Artificial Intelligence
OpenAI developing real-time conversational audio
OpenAI is building a new bidirectional (BiDi) audio model designed to make AI conversations feel far more natural. Unlike todayโs turn-based voice systems - where the AI waits for you to finish and canโt adjust mid-response - BiDi continuously processes speech, allowing it to adapt instantly if users interrupt or change direction.
OpenAI is building a new bidirectional (BiDi) audio model designed to make AI conversations feel far more natural. Unlike todayโs turn-based voice systems - where the AI waits for you to finish and canโt adjust mid-response - BiDi continuously processes speech, allowing it to adapt instantly if users interrupt or change direction.
Her is trying to come in real world
GTB
Morningggg ๐
แแฝแแ แฒแซแญ
แแแแ แจแแญแ แแฅ แแญแญ แแ
แแแแ แจแแญแ แแฅ แแญแญ แแ
Forwarded from แแแต's Journey (Freab)
Media is too big
VIEW IN TELEGRAM
The Qwen team just shipped their official agent framework and it has everything.
No stitching together third-party libraries. No fighting abstractions.
Qwen-Agent gives you:
โ Native function calling built directly into the framework
โ Secure code interpreter sandbox out of the box
โ RAG and MCP support included
โ Chrome extension for browser-native agent workflows
Built by the team that built the model. So it just works.
100% open source and completely free.
No stitching together third-party libraries. No fighting abstractions.
Qwen-Agent gives you:
โ Native function calling built directly into the framework
โ Secure code interpreter sandbox out of the box
โ RAG and MCP support included
โ Chrome extension for browser-native agent workflows
Built by the team that built the model. So it just works.
100% open source and completely free.
โค2๐ฅ2๐ซก1
Yared's log ๐ซ
There are few good news comming on local llms I will share them soon Am also preparing tools to make it easy for anyone to use smaller models on cpu
Well yaptts is one of those tools
there is another one i will share too
And there are more news to follow around Qwen 3.5 light models
I am planing to test how much of the Hallucinations can be removed with Wikipedia as a local RAG to keep them grounded in low resource local llm infrance
there is another one i will share too
And there are more news to follow around Qwen 3.5 light models
I am planing to test how much of the Hallucinations can be removed with Wikipedia as a local RAG to keep them grounded in low resource local llm infrance
GitHub
GitHub - laladigitalsystem-oss/yaptts: Production-ready offline TTS reader with local web UI, resilient chunked synthesis, punctuationโฆ
Production-ready offline TTS reader with local web UI, resilient chunked synthesis, punctuation-aware playback, and portable project-local caching. - laladigitalsystem-oss/yaptts
๐ฅ2
Yared's log ๐ซ
Photo
If you notice both curent theoretically passable and practically used capabilities collapse in physical ai and i dont think they are reporting global reality in physical ai
It's more like they are reporting what they observe in their ai usage
And since anthropic is not involved that much in physical ai that part of the graph don't make sense
Other than that everything else is well discribed
Tesla FSD 14 and other impressive models are not considered
#my_thought #reports #anthropic #physical_ai
It's more like they are reporting what they observe in their ai usage
And since anthropic is not involved that much in physical ai that part of the graph don't make sense
Other than that everything else is well discribed
Tesla FSD 14 and other impressive models are not considered
#my_thought #reports #anthropic #physical_ai