AI/ML●●●Banger
Orion – Native Training LLMs on the Apple Neural Engine Without CoreML
Direct ANE access bypasses CoreML to enable training—genuinely novel Apple Silicon unlock.
WizardryZero to OneBig Brain
mechramc
213mo ago
Back to browse
GitHub Repository
On-device LLM fine-tuning for Apple Silicon (ANE)
8 starsSwift
NeuralForge – Fine-Tune LLMs on Your Mac Using Apple Neural Engine
by khaeldur·Mar 12, 2026·1 point·1 comment
AI Analysis
●●SolidWizardryBig BrainNiche Gem
Fine-tune LLMs on Apple Neural Engine using reverse-engineered private frameworks — genuinely novel approach.
Strengths
- •Direct ANE access via reverse-engineered AppleNeuralEngine.framework — not just Metal GPU
- •Distributed training across multiple Macs via Bonjour with gradient aggregation
- •356 unit tests and 19 SwiftUI views show serious engineering investment
Weaknesses
- •13 commits but claims enterprise audit, cloud sync, webhooks — feature list feels inflated
- •Zero forks, zero stars — untested in real workflows despite ambitious scope
Category
Target Audience
ML engineers, macOS developers, privacy-focused AI practitioners
Similar To
MLX · llama.cpp · Petals
Similar Projects
Developer Tools●●Solid
M-Courtyard – Fine-tune LLMs on your Mac with zero code
Wraps mlx-lm fine-tuning into a guided desktop UI, but local LLM tools are crowded.
Niche GemShip It
tuwenbo0120
113mo ago
Developer Tools●●●Banger
LoRA gradients on Apple's Neural Engine at 2.8W
First LoRA gradients on ANE; matmul doesn't work, conv-as-matmul does.
WizardryBig BrainDark Horse
jmanhype
613mo ago
Developer Tools●●Solid
Courtyard – Open-source macOS app for local MLX fine-tuning Text
Tauri GUI wrapper around mlx-lm—useful for Mac users, but local fine-tuning UIs already exist.
Solve My ProblemShip It
tuwenbo0120
103mo ago
AI/ML●●●Banger
On-device meeting transcription for your Mac
CoreML models on Apple Neural Engine beat Whisper.cpp for on-device transcription.
WizardryDark Horse
paynedigital
602mo ago
AI/ML●●Solid
I fine-tuned Qwen 3.5 (0.8B–4B) on a Mac for text-to-SQL – 2B beats 12B
Unified memory trick lets a 2B model beat 12B; trains on MacBook with zero cloud costs.
Ship ItNiche GemBig Brain
sciences44
713mo ago