🔄 In a Training Loop

Barney Greenway

McG-221

·

AI & ML interests

LLMs on Apple Silicon

Recent Activity

liked a model about 1 hour ago

openbmb/MiniCPM-V-4.7-35B-A3B

liked a model 1 day ago

Jundot/Qwen3.8-Flash-Next-oQ2.5e-mtp

liked a model 3 days ago

ukisai/Swift1.5-Qwen3.8-Flash-Next

View all activity

Organizations

upvoted a collection 8 days ago

GSQ-RCO-GGUF

Non-uniform GGUF quantizations via GSQ + RCO: per-tensor mixed precision in standard GGUF form • 3 items • Updated 17 days ago • 57

upvoted a collection about 1 month ago

DFlash 2

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

upvoted 3 collections about 2 months ago

Unsloth Dynamic 3.0 Quants

Introducing Dynamic V3.0 quants, our new SOTA quantization methodology. • 1 item • Updated Aug 26 • 78

Qwen3.8

4 items • Updated Aug 13 • 575

Qwen3.8

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

upvoted a collection 2 months ago

Recent medium models

81 items • Updated 2 days ago • 17

upvoted 2 collections 3 months ago

Bonsai 27B

10 items • Updated Jul 15 • 250

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 24 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 • 164

Gemma 4 QAT

40 items • Updated Jun 5 • 12

upvoted 5 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

gemma-4-abliterated

17 items • Updated 14 days ago • 40

upvoted a collection 6 months ago

Uncensored Gemma 4 MLX

10 items • Updated Apr 16 • 2