Lens-Turbo-3.8B-4bit (MLX)

Apple MLX conversion of microsoft/Lens-Turbo — the distilled 4-step sibling of Lens (identical 3.8B DiT architecture; sample at 4 steps, guidance 1.0). int4 (group_size 64, keeping in/out/time at bf16), ~2.35 GB. DiT-only (MIT); the GPT-OSS-20B encoder (Apache-2.0) and FLUX.2 VAE load from source. Architecture is byte-identical to base Lens, which is parity-locked vs the PT reference (DiT cosine 0.999999); this variant inherits that port.

Usage

from lens_mlx.pipeline_mlx import LensPipeline   # github.com/xocialize/lens-mlx

# `base` = a microsoft/Lens snapshot (tokenizer + GPT-OSS encoder + FLUX.2 VAE).
pipe = LensPipeline.from_pretrained(base, dit_repo="mlx-community/Lens-Turbo-3.8B-4bit")
img = pipe("A serene lake below snow-capped mountains, golden hour.",
           height=1024, width=1024, num_inference_steps=4, guidance_scale=1.0, seed=42)
img.save("out.png")

Tip: page weights into memory before the first forward (mx.eval the params) when loading from slow/external storage, to avoid a Metal command-buffer watchdog timeout at large sizes.

License

DiT weights MIT (from microsoft/Lens-Turbo) · GPT-OSS-20B encoder Apache-2.0 (not re-hosted) · FLUX.2 VAE under its own terms (not re-hosted). Upstream: microsoft/Lens-Turbo.