---
title: "18MB, 70x Real-Time: The TTS Model tiny-voca Runs Without a GPU｜Local AI Lab"
description: "On November 16, 2026, the local model radar dug up a real beast: tiny-voca is only 18.2MB, and officially runs at 70x real-time on CPU, no GPU needed at all. The other lead is FLUX.1-schnell with 12B parameters, released under Apache-2.0 for open commercial use and LoRA training. In a moment, I'll tell you how to pick between these two."
canonical_url: "https://blog.markkulab.net/en/local-ai-lab/local-ai-lab-2026-11-16"
author: "Mark Ku"
author_url: "https://blog.markkulab.net/en/author/mark-ku"
site: "Mark Ku's Tech Notes"
date_published: "2026-11-16T02:00:00.000Z"
category: "Local AI Lab"
tags: ["local-ai", "podcast", "地端模型", "lora", "開源模型", "tiny-voca", "FLUX.1-schnell", "TTS語音模型", "生圖模型", "LoRA訓練", "ApacheLicense"]
language: "en"
license: "CC BY 4.0"
license_url: "https://creativecommons.org/licenses/by/4.0/"
attribution: "when reusing or quoting, credit the author and link back to the original"
---

# 18MB, 70x Real-Time: The TTS Model tiny-voca Runs Without a GPU｜Local AI Lab

## Opening

In the November 16, 2026 scan of the local model radar, we dug up a serious contender: tiny-voca, weighing in at just 18.2MB, with the team claiming it can run at 70x real-time on CPU alone, no GPU required. The other headliner is FLUX.1-schnell, a 12B-parameter model released under the Apache-2.0 license for open commercial use and LoRA training. Let's dig into how to pick between these two.

## This Episode's Picks

### 1. tiny-voca: A Speech Model Small Enough to Bundle Right into Your Repo

- **What it is**: n4ze3m/tiny-voca is a text-to-speech (TTS) model with only 4.5M parameters and an extremely small weight file, released under the Apache-2.0 license with no restrictions on personal or commercial use. Compared to typical TTS models that start at hundreds of megabytes, this one is a genuine outlier in the lightweight category.

- **Where it fits**:
  - If you're building a small browser-based tool or something for an edge/IoT device and want to add voice feedback without complicating your deployment with a GPU server, this model is small enough to bundle directly into your repo and ship as-is, and the Apache-2.0 license means you don't have to worry about hitting commercial-use roadblocks.
  - If you're adding an English voice alert or announcement feature to a side project, say, a timer that says "time's up," or a small robot that talks back, you can solve it with a model under 20MB that doesn't need to run in the cloud, without renting extra GPU capacity for one small feature.
  - When you need a lightweight baseline for comparing TTS model quality, this one comes with publicly verifiable WER rankings, which makes it a more credible reference point than just grabbing a random small model off the shelf.

- **Will it actually run**: The official model card states "runs about 70x faster than real time on four CPU threads," with no GPU required. The model file is only 18.2MB (PyTorch format) or 21.4MB (ONNX format), and GGUF is also available as a format option. Whether you're running PyTorch, ONNX Runtime, or another lightweight inference environment, there's a format that fits.

- **Current buzz**: HF popularity score of 135, 860 downloads, 143 likes. The numbers aren't jaw-dropping, but for a recently released 4.5M-parameter model, there's already a solid group of people using it.

- **How it stacks up**: According to the Open TTS Leaderboard (November 2026), tiny-voca's M2 voice "currently ranks 8th by average WER," and its F2 voice "ranks 13th." It's rare to find a model at this 4.5M-parameter scale with verifiable third-party benchmark rankings, rather than just the model's own self-reported claims.

- **How to get it**: The official docs provide an install command, `pip install git+https://github.com/n4ze3m/dizione git+https://github.com/n4ze3m/tiny-voca`, after which running `python -m tiny_voca.infer --checkpoint tiny-voca/tiny-voca.pt --voice F2 --text "Hello"` generates speech. An ONNX version is also available, loadable via onnxruntime.

- **Link**: [huggingface.co/n4ze3m/tiny-voca](https://huggingface.co/n4ze3m/tiny-voca)

### 2. FLUX.1-schnell: An Image-Generation and LoRA Training Base Without Licensing Headaches

- **What it is**: black-forest-labs/FLUX.1-schnell is a 12B-parameter rectified flow transformer for image generation, released under the Apache-2.0 license. According to the official docs, it was trained using latent adversarial diffusion distillation, allowing it to generate images in just 1 to 4 inference steps, significantly faster than traditional diffusion models that typically require 20 to 50 steps.

- **Where it fits**:
  - If you want to set up an image-generation service on your own machine, whether through local ComfyUI or your own custom pipeline, and don't want to get tripped up by licensing terms, the Apache-2.0 schnell version places no restrictions on personal or commercial use, so you can adopt it directly as your base model.
  - If you want to train your own style LoRA, say, for brand visuals or a specific illustration style, FLUX.1-schnell is currently one of the most commonly used free base models in the community for LoRA training, making it relatively easy to find existing tutorials and community resources.
  - When you need to prototype images quickly, for example, generating a few concept images for colleagues before a design review, and don't want to wait through a full round of traditional diffusion steps, the official example only requires setting `num_inference_steps=4` to produce output, cutting wait time considerably.

- **Will it actually run**: The official model card doesn't specify a clear VRAM threshold, but the Diffusers official example provides an option, `enable_model_cpu_offload()`, that lets users "save VRAM by offloading the model to CPU," suggesting the team also anticipated this model being used in resource-constrained environments. However, the exact amount of VRAM needed for smooth operation isn't explicitly stated. Safetensors is the available format.

- **Current buzz**: HF popularity score of 55, 280.1k downloads, 6.8k likes. Among this batch of picks, the download count here is particularly solid.

- **How it stacks up**: The official docs state the model was trained using latent adversarial diffusion distillation, enabling image generation in 1 to 4 inference steps, and the model card describes its performance as reaching "the performance of closed-source alternatives." However, it doesn't provide direct benchmark comparisons against FLUX.1-dev or SDXL, so this part is based solely on the official description, without a quantitative comparison table.

- **How to get it**: The official example loads it via Diffusers: `from diffusers import FluxPipeline` combined with `FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=torch.bfloat16)`, and calling it with `num_inference_steps=4, guidance_scale=0.0` generates an image.

- **Link**: [huggingface.co/black-forest-labs/FLUX.1-schnell](https://huggingface.co/black-forest-labs/FLUX.1-schnell)

## Closing

Today's two picks sit at opposite ends of the local model ecosystem: tiny-voca proves that speech models don't need a GPU to work, with its minuscule footprint, while FLUX.1-schnell uses a permissive license to clear the path for image generation and LoRA training. If all you've got is a laptop without a GPU, give tiny-voca a try first. If you've got a graphics card and want to train your own style LoRA, FLUX.1-schnell is worth putting on your shortlist. Same time tomorrow, we'll keep digging through the pile of local models. See you in the next episode.

---

## About this article and its author

Originally published on [Mark Ku's Tech Notes](https://blog.markkulab.net/en/local-ai-lab/local-ai-lab-2026-11-16)

License: [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) — when reusing or quoting, credit the author and link back to the original

### About the author

**[Mark Ku](https://blog.markkulab.net/en/author/mark-ku)** — Software engineer

- 10+ years as a software engineer
- Built North-American e-commerce and AI SaaS subscription billing

### Free tools built by the author

All of these are free to use:

- [Free PDF Sign Tool](https://blog.markkulab.net/en/tools/pdf-sign): Online PDF sign tool — draw, type, or upload a signature, then drag, resize, and download. Everything runs in your browser; nothing is uploaded.
- [VS Code Refactory](https://blog.markkulab.net/en/tools/refactory): Refactory is a VS Code refactoring extension: 34 actions plus a 37-rule code-smell inspection layer with a Code Health dashboard, across 18 languages, backed by 534 tests. It learns your repo's conventions: where interfaces live, where DI is registered, whether 'use client' belongs. It ranks files by git churn × complexity so you know what to fix first, and hands any smell to the Claude Code already on your machine. Free to use, and your source never leaves your computer.
- [DB-Kit Database Manager](https://blog.markkulab.net/en/tools/db-kit): DB-Kit is a lightweight, cross-platform database manager built with Tauri + Rust + React. Manage MySQL, MariaDB, PostgreSQL, SQL Server, Oracle, SQLite, MongoDB, Redis, Kafka, Elasticsearch and RabbitMQ from one consistent interface: passwords encrypted in the OS keychain, SSH tunnels, full CRUD, a visual query builder, stacked multi-statement result sets, cross-connection data transfer, schema & data compare with sync SQL and schema snapshots, Excel / CSV import & export, visualized execution plans, ER diagrams, scheduled backups, SQL stress testing with p50–p99 latency percentiles, a 15-rule SQL review engine, Review & Run (AI review plus per-statement backups and an auto-generated rollback script), Kafka message browsing with monitoring & alerts, an SSH terminal with SFTP / FTP, Docker / Kubernetes, remote desktop (RDP / VNC / RustDesk), file / folder compare, stored procedure integration tests, six UI languages, a built-in AI assistant (a local CLI or any Anthropic / OpenAI-compatible API; natural-language SQL, AI review and tuning advice) and the dbk CLI. Free and open source (MIT), with installers for Windows, macOS and Linux.
- [VS Code Super Mermaid](https://blog.markkulab.net/en/tools/super-mermaid): Super Mermaid is a VS Code extension for beautiful Mermaid diagrams out of the box: auto-colored live preview, mouse pan & zoom, high-res PNG / SVG export, 21 templates and multiple themes. Free and open source (MIT).
- [React Super Mermaid](https://blog.markkulab.net/en/tools/react-super-mermaid): react-super-mermaid is an open-source React component library: render beautiful Mermaid diagrams with a single <MermaidViewer>, with built-in colorful / sketch themes, pan & zoom, in-diagram search, and high-res SVG / PNG export. Lightweight, SSR-safe, fully typed. Free and open source (MIT).
- [Jira / Confluence Super Mermaid](https://blog.markkulab.net/en/tools/jira-super-mermaid): An Atlassian Forge app: write Mermaid syntax directly inside a Jira issue or a Confluence page and get flowcharts, sequence diagrams, state machines and Gantt charts. 11 diagram types, SVG / PNG export, light and dark themes, full CJK support. Runs on Atlassian: your diagrams live in your own site and the app calls no third-party service. Free, coming soon to the Atlassian Marketplace.
- [Mermaid Live Preview](https://blog.markkulab.net/en/tools/mermaid-preview): Write Mermaid in your browser, see it render instantly, and share the whole diagram as a single link. No sign-up, nothing uploaded to a server, and mermaid.live share links work as-is.
- [React Intl Phone Number](https://blog.markkulab.net/en/tools/react-intl-phone-number): react-intl-phone-number is an open-source React component: framework-agnostic and antd-free, with E.164 in/out, a searchable flag / country-code dropdown, configurable validation levels (strict / mobile-strict / loose), themeable CSS, and i18n — phone logic powered by google-libphonenumber. Lightweight and fully typed. Free and open source (MIT).
- [Uptime Kuma Cluster](https://blog.markkulab.net/en/tools/uptime-kuma-cluster): Turn single-node Uptime Kuma into a highly available cluster: OpenResty + Lua smart load balancing, shared MariaDB state, health checks and automatic failover, plus cluster-management REST APIs. One Docker Compose command to start. Free and open source (MIT).
- [AI Podcast Cut](https://blog.markkulab.net/en/tools/ai-podcast-cut): Drop in a recording and it removes fillers and stutters, levels loudness segment by segment, and sends a second agent to review every cut. Cut points snap to word boundaries and zero crossings, every splice gets a fade, and sentence-end breaths are preserved. Desktop app for Windows, macOS and Linux. MIT licensed; the Windows installer bundles ffmpeg.
- [AI Video Cut](https://blog.markkulab.net/en/tools/ai-video-cut): Pick something in your video: type it (face, license plate, phone screen, logo), click points or drag a box, or let Claude Code / Codex pick it. AI Video Cut tracks it frame by frame, then mosaics or blurs it, recolors it, pins stickers and text to it, swaps a screen or poster for your own image or video, or removes it using background that other frames actually captured. Plus sequence editing, local captions and an AI assistant. Free, open-source (MIT) Windows desktop app (macOS / Linux experimental) that runs locally, nothing uploaded.
- [open-pos restaurant POS](https://blog.markkulab.net/en/tools/open-pos): One computer and one receipt printer is enough to open the shop. Your data lives on your own disk, no subscription, no lock-in, MIT licensed. Money is integer New Taiwan dollars with tax split by the statutory formula, so sales + tax always equals the total. Printing goes straight over ESC/POS on TCP 9100, with no vendor driver. Tauri + Rust + SQLite desktop app, v1.0 in development.
- [.NET e-commerce API template](https://blog.markkulab.net/en/tools/dotnet-ecommerce-api): An e-commerce backend API template built on .NET 10: Controller → Service → Repository layering, Autofac convention-based registration, SqlSugar for MySQL / MariaDB and SQL Server, Redis cache and message queue, Hangfire jobs, JWT auth, and an extensible payment provider module (ECPay, Newebpay). Apache 2.0.
- [Open Token Monitor team AI usage](https://blog.markkulab.net/en/tools/open-token-monitor): Collect token usage and equivalent cost from every employee's AI coding tools into a hub you host yourself: PostgreSQL, org roster, dashboards compared by company and department, and a reports API. Electron and Rust/Tauri clients upload every 30 minutes. MIT licensed, one Docker command.
- [ai-security-monitor defensive security](https://blog.markkulab.net/en/tools/ai-security-monitor): Your website is live. Who checks its certificates, security settings and known vulnerabilities? A self-hosted Security Hub brings website scans and host monitoring together, deduplicates findings and compares scans after fixes.

### Daily podcasts

- [Mark's Tech Insights — Daily AI News](https://blog.markkulab.net/en/category/tech-news): Daily curated AI and tech trends. Catch the latest developments via audio summaries — covering AI applications, software architecture, DevOps, and engineering practice. — RSS: https://blog.markkulab.net/feed.xml
- [AI股市蝦聊](https://blog.markkulab.net/en/category/ai-stock-chat): Every trading day, an AI-analyzed take on the Taiwan stock market, delivered as a two-host conversation covering the session and the next-day outlook. — RSS: https://blog.markkulab.net/ai-stock-chat/feed.xml
- [開源好物日報](https://blog.markkulab.net/en/category/open-source-weekly): A daily two-host pick of free open-source tools surfaced from real Hacker News, GitHub, and Reddit buzz — what pain they solve and the fastest way to get started. — RSS: https://blog.markkulab.net/open-source-weekly/feed.xml
- [AI 運動日報](https://blog.markkulab.net/en/category/sports-weekly): Two hosts talk NBA, MLB and world sport every day — scores, records and the stories behind them, from a Taiwanese fan perspective. — RSS: https://blog.markkulab.net/sports-weekly/feed.xml
- [AI 國際新聞快報](https://blog.markkulab.net/en/category/world-news): A daily two-host briefing that makes sense of the past 24 hours in world news — geopolitics, the global economy, conflict and security, disasters and climate — from a Taiwanese perspective, neutral and fully sourced. — RSS: https://blog.markkulab.net/world-news/feed.xml
- [地端 AI 實驗室](https://blog.markkulab.net/en/category/local-ai-lab): Every day, a two-host look at open-weight models and LoRAs you can actually run on your own machine: what they are for, whether your GPU can handle them, and how they compare — sourced from official model cards. — RSS: https://blog.markkulab.net/local-ai-lab/feed.xml
- [AI 行銷日報](https://blog.markkulab.net/en/category/ai-marketing): Every day, two hosts break down the latest in AI for marketing, advertising, SEO/GEO and content marketing: practical, never sponsored, every story sourced. — RSS: https://blog.markkulab.net/ai-marketing/feed.xml
- [AI 資安日報](https://blog.markkulab.net/en/category/ai-security): Every day, two hosts cover the latest where AI meets cybersecurity: LLM and AI agent vulnerabilities, prompt injection, model supply-chain risks and AI-powered attacks, from a defense-first angle with every story sourced. — RSS: https://blog.markkulab.net/ai-security/feed.xml

### Newsletter

[Subscribe to the newsletter](https://blog.markkulab.net/en/subscribe) — Be the first to know about new posts. No spam, unsubscribe anytime.
