Title: Gemma 4 Multimodal Fine-Tuning Tool: Optimized for Apple ...
URL Source: https://www.bestblogs.dev/status/2041646415831822622
Published Time: 2026-04-07 22:36:52
Markdown Content: Skip to main content Toggle navigation menu Toggle navigation menuArticlesPodcastsVideosTweetsSourcesNewsletters
⌘K
Change language Switch ThemeSign In
Narrow Mode
Gemma 4 Multimodal Fine-Tuning Tool: Optimized for Apple Silicon
Gemma 4 Multimodal Fine-Tuning Tool: Optimized for Apple Silicon
 ### Berryxia.AI@berryxia
Gemma 4 多模态微调工具终于来了！专为 Apple Silicon 打造！
最新开源项目：
✅ 本地 LoRA 微调 Gemma 4（PyTorch + Metal）
✅ 支持文本 + 图像 + 音频
✅ 从 Google Cloud 直接流式拉数据（不用下载几个 TB）
✅ 超简单 CLI ，一步一步带你走
再也不用租 H100、也不用把数据全拷到笔记本了！
Mac 用户直接本地玩转多模态微调🤯
PS：目前还不支持24b、32B 模型，仅支持e4/2b模型。Show More
!Image 3: Matt Mireles | Accelerate or Die
#### Matt Mireles | Accelerate or Die
@mattmireles · 4h ago
Introducing...
Gemma 4 Multimodal Fine-Tuner for Apple Silicon
- LoRA fine-tunning toolkit for Gemma LLM
- runs locally on macOS via PyTorch and Metal
- streams data from Google Cloud to your machine
- fine-tune on audio, image and text
- easy-to-use CLI wizard
3
15
99
6,664
Apr 7, 2026, 10:36 PM View on X
1 Replies
5 Retweets
6 Likes
822 Views  Berryxia.AI @berryxia
One Sentence Summary
An open-source project enables local LoRA fine-tuning of Gemma 4 on Mac, eliminating the need for expensive GPU rentals.
Summary
This is an open-source fine-tuning toolkit optimized for Apple Silicon (M-series chips). It supports local LoRA fine-tuning of Gemma 4 models using PyTorch and Metal, covering text, image, and audio multimodal data. A key feature is its ability to stream data from Google Cloud, avoiding the need for massive local storage. It currently supports 4B and 2B scale models.
AI Score
86
Influence Score 4
Published At Yesterday
Language
Chinese
Tags
Gemma 4
Apple Silicon
LoRA Fine-tuning
Multimodal
Mac Development HomeArticlesPodcastsVideosTweets