A hands-on log of running Qwen-Scope's Sparse Autoencoder locally on M1 Max 64GB with Qwen3-8B-Base, extracting feature IDs that discriminate between Japanese, English, code, and Chinese from a single middle layer.
Pruna AI's FP8 speedup needs compute capability 8.9, so Apple Silicon is out. Measured what M1 Max 64GB actually does with MLX-based mflux and antirez's iris.c: install traps, real generation times, and a wrapper kit to skip the setup.
After Xiaomi MiMo-V2.5's weights went public, I checked whether it runs on Mac/ROCm or on cloud GPU (RunPod/GCE). It's still rough on local hardware, but RunPod's 4x H200 runs it for ~$14/hr and GCE Spot H100 brings it down to ~$1.6/hr.
SeeSee21/Z-Anime is a full fine-tune of Z-Image Base (S3-DiT, 6B) — not a LoRA merge. Ran the AIO build on local ComfyUI (M1 Max 64GB): working t2i and i2i settings for Base vs Distill 8/4-step, GGUF options, and what NSFW prompts actually pass.
A verification log for converting color anime-style AI illustrations to manga-style monochrome. AI re-generation approaches lean to either color leakage or face drift, and pure deterministic local processing looks mechanical. Frames the next directions to try: putting a grayscale-only LoRA on Anima, and using See-through for part decomposition before mechanical composition.
Hands-on running inclusionAI Ling-flash-2.0 (100B / 6.1B active, MXFP4 quant, 54.7GB) on SwiftLM via mlx-swift-lm on an M1 Max 64GB. Covers bailing_moe + MXFP4 support check in mlx-swift, the startup surprise, and what --stream-experts actually saves.
WAI-Illustrious SDXL v17 tested on M1 Max 64GB ComfyUI against v16 with the same seed. Hires fix now auto-corrects hands and feet, the four rating tags (general/sensitive/nsfw/explicit) still drive NSFW output, and v16-trained LoRAs mostly carry over — with one case where they don't.
A hands-on build and run of the Swift-based LLM inference server SwiftLM on an M1 Max 64GB. Covers Qwen3.6-35B-A3B and Qwen3.5-122B-A10B, with the same BST, BBS, and persona tests used in the existing Ollama and MLX-lm write-ups.
The NotebookLM clone open-notebook assumes Docker and cloud APIs by default. I installed SurrealDB natively, ran four processes in tmux, and wired everything through Ollama's qwen3.6:35b and bge-m3. I fed it the Qwen3.6 benchmark article I wrote this morning, and it answered with the correct numbers.
Tried Qwen3.6-27B on both Ollama and MLX. Ollama couldn't load the VL-projector-embedded GGUF, MLX ran it at 11 tok/s. On the side, running 35B-A3B under MLX was roughly 2× faster than the Ollama GGUF. Also had both models build a BBS to gauge intent handling.
Hands-on Qwen3.6-35B-A3B (23GB 4bit GGUF) on M1 Max 64GB via Ollama 0.20.6. Generation speed stays at 27 tok/s — same as Qwen3.5-35B-A3B — but the same prompt produces 13× more thinking tokens. Multi-turn behavior, persona handling, and a three-tier NSFW probe included.
Z-Image has its own pixel art LoRAs, but can they actually convert photos to pixel art via i2i? Tested Z-Image Turbo, base model, and compared with Illustrious on M1 Max 64GB.
