🔄 In a Training Loop

Barney Greenway

McG-221

·

AI & ML interests

LLMs on Apple Silicon

Recent Activity

liked a model about 2 hours ago

prism-ml/Ternary-Bonsai-2-9B-gguf

liked a model 1 day ago

unsloth/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-GGUF

liked a model 3 days ago

unsloth/LFM2.5-24B-A2B-GGUF

View all activity

Organizations

upvoted a collection 12 days ago

GSQ-RCO-GGUF

Non-uniform GGUF quantizations via GSQ + RCO: per-tensor mixed precision in standard GGUF form • 3 items • Updated 20 days ago • 62

upvoted a collection about 1 month ago

DFlash 2

Keep Drafting Parallel • 4 items • Updated Sep 12 • 31

upvoted a collection about 2 months ago

Unsloth Dynamic 3.0 Quants

Introducing Dynamic V3.0 quants, our new SOTA quantization methodology. • 4 items • Updated 1 day ago • 83

upvoted 3 collections 2 months ago

Qwen3.8

4 items • Updated Aug 13 • 585

Qwen3.8

Qwen3.8 models converted to MLX format for Apple silicon. • 12 items • Updated Aug 14 • 23

Recent medium models

84 items • Updated about 7 hours ago • 18

upvoted 2 collections 3 months ago

Bonsai 27B

10 items • Updated Jul 15 • 251

NVIDIA NVFP4 → MLX

NVIDIA post-trained NVFP4 checkpoints (ModelOpt PTQ) transcoded to MLX for Apple Silicon / mlx-node, keeping NVIDIA's per-layer bit allocation. • 1 item • Updated Jul 21 • 2

upvoted 6 collections 4 months ago

Ornith-1.0

Ornith-1.0 is a family of open-source LLMs specialized for agentic coding. • 8 items • Updated 27 days ago • 396

Ornith 1.0

MLX versions of Ornith 1.0 • 6 items • Updated Jun 27 • 4

Gemma 4 QAT

Gemma 4 QAT (Quantization-Aware Training) for 3x less memory use and near original accuracy. • 16 items • Updated Aug 26 • 124

Qwen 3.x MTP

MLX MTP drafter checkpoints for Qwen 3.x speculative decoding with mlx-vlm. • 12 items • Updated Jun 1 • 9

Gemma 4 QAT Q4_0

19 items • Updated Jul 21 • 166

Gemma 4 QAT

40 items • Updated Jun 5 • 12

upvoted 4 collections 5 months ago

Gemma-4-31B-IT-unsloth-mlx

Gemma-4-31B-IT (dense vision-language) quantized for Apple Silicon (MLX) — Unsloth Dynamic 2.0 with AWQ imatrix pre-scaling. • 9 items • Updated May 16 • 1

Gemma-4-26B-IT-unsloth-mlx

Gemma-4-26B-A4B-IT MoE quantized for Apple Silicon (MLX) — Unsloth Dynamic 2.0 with AWQ imatrix pre-scaling. • 8 items • Updated May 16 • 1

Magistry

Q1 2026 merges based on Mistral 24B • 2 items • Updated Mar 22 • 2

Qwen-3.6-unsloth-mlx

AWQ-style pre-scaling using Unsloth's imatrix calibration data, then 3-6-bit affine quantization with the Unsloth mixed-precision recipe via MLX • 18 items • Updated May 15 • 19

upvoted 2 collections 6 months ago

gemma-4-abliterated

17 items • Updated 18 days ago • 40

Uncensored Gemma 4 MLX

10 items • Updated Apr 16 • 2