Waliu Ayuba

23 days ago

Harverz🔥🔥

DailyAFC

@DailyAFC

23 days ago

Kai Havertz puts Arsenal up 1-0! 😮‍💨

6

692

53

31

53K

0

23

Upcoming Billionaire Illogically Logical, Unreasonably flamboyant, Rare being, Paper Chaser, #GraceAmbassador, #TeamArsenal #Capricon

24 days ago

@now_arsenaI COYG❤️ we can do it💪

0

16

ayuwal12 retweeted

Claude

@claudeai

25 days ago

Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors. Available today at the same price.

claudeai's tweet photo. Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors.

Available today at the same price. https://t.co/EufxL7T1kb

4K

67K

9K

8K

15M

Who to follow

Riley 69(Omo Pastor)

@subzer089

#MASTER HUN 🧔

@ik_eddie

The only man born in Canada 🇨🇦, grew up in Japan 🇯🇵 and was brought to Nigeria 🇳🇬 for breastfeeding 🤷‍♂️ #liverpuldian till the end

Adamu Alex

@AneruAlex

Digital Marketing | Public Relations | @ManUtd lover

25 days ago

@premierleague @Arsenal Impressive🔥

0

9

ayuwal12 retweeted

27 days ago

Your TTS prototype doesn’t need a cloud API for every sentence. Supertonic 3 is an on-device text-to-speech system, and supertonic-py is its Python package for running synthesis locally via ONNX. It helps you add speech to apps faster by giving you a pip-installable SDK, a CLI, and a local HTTP server around the same engine. Key features: • On-device inference – runs locally through ONNX Runtime, with no GPU or network dependency required after setup • Python + CLI quick start – install with pip, synthesize text, and save WAV output from code or the terminal • Multilingual synthesis – supports 31 language codes plus an `na` fallback for unknown or unsupported input language • Voice styles – includes 10 built-in voices and supports Voice Builder JSON exports for custom styles • Local HTTP server – exposes native TTS and OpenAI-compatible speech endpoints, with batch synthesis support It’s open-source (MIT license). Link in the reply 👇

DanKornas's tweet photo. Your TTS prototype doesn’t need a cloud API for every sentence.

Supertonic 3 is an on-device text-to-speech system, and supertonic-py is its Python package for running synthesis locally via ONNX.

It helps you add speech to apps faster by giving you a pip-installable SDK, a CLI, and a local HTTP server around the same engine.

Key features:

• On-device inference – runs locally through ONNX Runtime, with no GPU or network dependency required after setup
• Python + CLI quick start – install with pip, synthesize text, and save WAV output from code or the terminal
• Multilingual synthesis – supports 31 language codes plus an `na` fallback for unknown or unsupported input language
• Voice styles – includes 10 built-in voices and supports Voice Builder JSON exports for custom styles
• Local HTTP server – exposes native TTS and OpenAI-compatible speech endpoints, with batch synthesis support

It’s open-source (MIT license).

Link in the reply 👇

4

10

3

12

910

ayuwal12 retweeted

27 days ago

Your coding agent doesn’t need more vibes. It needs a stricter engineering loop. ROACH PI is an extension suite for the pi coding agent that adds disciplined planning, orchestration, review, memory, code intelligence, MCP access, and faster search. It helps you move from fuzzy request to verified implementation by forcing clarify → plan → worker → validator style workflows, with inspectable TypeScript and Markdown commands, tools, hooks, agents, and skills inside the repo. Key features: • Clarify + plan workflow – turns vague requests into context briefs, task plans, verification commands, and success criteria • Subagent orchestration – delegates single, parallel, chain, or async work to separate pi processes • Review pipelines – supports quick `/review` and deeper `/ultrareview` flows with multiple reviewers and synthesis • Developer tooling – adds FFF search, LSP code intelligence, MCP adapter access, and workspace memory • Installable pi extension – installs from GitHub, then runs `/setup` to configure the startup flow It’s open-source (MIT license). Link in the reply 👇

DanKornas's tweet photo. Your coding agent doesn’t need more vibes. It needs a stricter engineering loop.

ROACH PI is an extension suite for the pi coding agent that adds disciplined planning, orchestration, review, memory, code intelligence, MCP access, and faster search.

It helps you move from fuzzy request to verified implementation by forcing clarify → plan → worker → validator style workflows, with inspectable TypeScript and Markdown commands, tools, hooks, agents, and skills inside the repo.

Key features:

• Clarify + plan workflow – turns vague requests into context briefs, task plans, verification commands, and success criteria
• Subagent orchestration – delegates single, parallel, chain, or async work to separate pi processes
• Review pipelines – supports quick `/review` and deeper `/ultrareview` flows with multiple reviewers and synthesis
• Developer tooling – adds FFF search, LSP code intelligence, MCP adapter access, and workspace memory
• Installable pi extension – installs from GitHub, then runs `/setup` to configure the startup flow

It’s open-source (MIT license).

Link in the reply 👇

1

11

9

14

1K

ayuwal12 retweeted

28 days ago

LLM demos are easy. The last mile is not. LLM-engineer-handbook is a curated GitHub handbook of LLM frameworks, tools, tutorials, and learning resources for engineers building real LLM applications. It helps you move beyond random tutorials by mapping resources across the LLM lifecycle: application frameworks, pretraining, fine-tuning, serving, prompt management, datasets, benchmarks, LLMOps, fundamentals, books, and communities. Key features: • Lifecycle map – resources span building apps, model training, serving, fine-tuning, prompt optimization, and LLMOps • Tool index – includes frameworks and tools like DSPy, LlamaIndex, LangChain, Haystack, vLLM, Ollama, llama.cpp, Opik, and Agenta • Data + eval sections – points to datasets, synthetic data tooling, fine-tuning datasets, benchmarks, and evaluation libraries • Learning path included – collects courses, agent guides, modeling resources, fundamentals, books, and optimization material • Community layer – lists social accounts and communities to follow for LLM engineering and ML systems updates It’s open-source (MIT license). Link in the reply 👇

DanKornas's tweet photo. LLM demos are easy. The last mile is not.

LLM-engineer-handbook is a curated GitHub handbook of LLM frameworks, tools, tutorials, and learning resources for engineers building real LLM applications.

It helps you move beyond random tutorials by mapping resources across the LLM lifecycle: application frameworks, pretraining, fine-tuning, serving, prompt management, datasets, benchmarks, LLMOps, fundamentals, books, and communities.

Key features:

• Lifecycle map – resources span building apps, model training, serving, fine-tuning, prompt optimization, and LLMOps
• Tool index – includes frameworks and tools like DSPy, LlamaIndex, LangChain, Haystack, vLLM, Ollama, llama.cpp, Opik, and Agenta
• Data + eval sections – points to datasets, synthetic data tooling, fine-tuning datasets, benchmarks, and evaluation libraries
• Learning path included – collects courses, agent guides, modeling resources, fundamentals, books, and optimization material
• Community layer – lists social accounts and communities to follow for LLM engineering and ML systems updates

It’s open-source (MIT license).

Link in the reply 👇

4

87

19

129

4K

29 days ago

Job well done @Arsenal

Premier League

@premierleague

29 days ago

It's done.

339

54K

11K

2K

511K

0

25

ayuwal12 retweeted