Running a 180B-Parameter MoE Model on a Gaming Laptop: VIDRAFT's POCKET-Darwin-180B-GGUF

·4 views

Running a 180B-Parameter MoE Model on a Gaming Laptop: VIDRAFT's POCKET-Darwin-180B-GGUF TL;DR: VIDRAFT has released POCKET-Darwin-180B-GGUF, a 4-bit quantized, GGUF-format version of their 180B Mixture-of-Experts model that runs on consumer hardware — including a gaming laptop with 8 GB VRAM and 32 GB RAM. By combining MoE sparse activation with llama.cpp-based SSD streaming, only ~3B parameters are computed per token at inference time, slashing hardware requirements by roughly 250× compared to the full-precision server setup. Developers working on air-gapped or on-device deployments should p

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Using SCP on a Custom Port (and Avoiding the -p vs -P Mix-Up)

If SSH listens on a nonstandard port, SCP needs to connect to that same port. The easy-to-miss detail is capitalization: use uppercase -P with scp, not lowercase -p. SCP doesn't have a separate network port. It connects through SSH, which uses TCP port 22 by default. If a server's SSH service listens on port 2222, SCP must connect to 2222 too.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Contract-First Engineering in Distributed Core Banking

In enterprise core banking modernizations, microservice decomposition often creates an unexpected governance failure: Semantic Drift. When dozens of distributed product teams build microservices independently, subtle discrepancies emerge between architectural specifications and real-world implementations. A field defined as ISO4217 currency string in a central domain model might be implemented as a raw string in one service, a numeric currency code in another, or omitted entirely in an asynchronous event payload. Relying on a "Code-First" development approach—where engineers write Java control

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How to Build a Podcast Intro Generator

What You’ll Build By the end of this post you’ll have a tiny but powerful “Podcast Intro Generator” that: Accepts a podcast title, host name, and a short tagline. Dynamically stitches those strings together into a polished intro script. Sends that script to a text‑to‑speech engine. Returns a high‑quality audio clip you can drop straight into your podcast editing workflow. We’ll use ElevenLabs for the heavy lifting—its API gives us instant, human‑like voices and even voice‑cloning capabilities.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Create AI Voice Responses for Slack Bots

Introduction If you’ve ever wished your Slack bot could talk back to you, you’re not alone. Adding voice responses turns a plain text interaction into a more engaging, accessible experience—perfect for status updates, alerts, or just a little fun. In this article we’ll walk through building a Slack bot that speaks using modern text‑to‑speech (TTS) and voice‑cloning technology. We’ll use ElevenLabs as the TTS engine (the affiliate link is included below) and glue everything together with a few lines of Python. TL;DR – By the end of this guide you’ll have a Slack bot that receives a message, gen

0 comments Read more at DEV Community