---
title: Lens-3.8B-bf16
canonical_url: "https://www.modelscope.cn/models/mlx-community/Lens-3.8B-bf16"
md_url: "https://www.modelscope.cn/models/mlx-community/Lens-3.8B-bf16.md"
repository: mlx-community/Lens-3.8B-bf16
last_updated: 2026-06-21
license: mit
pipeline_tag: text-to-image-synthesis
tasks:
  - text-to-image-synthesis
base_model:
  - microsoft/Lens
base_model_relation: finetune
parameters: 4.1B
tensor_type:
  - BF16
library_name:
  - mlx
  - lora
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 8
stars: 0
tags:
  - mlx
  - text-to-image
  - diffusion
  - lens
  - apple-silicon
---

# Lens-3.8B-bf16

> Lens-3.8B-bf16 - mlx-community 在 ModelScope 开源的模型。Lens-3.8B-bf16 (MLX)

mlx-community/Lens-3.8B-bf16 是 ModelScope 魔搭社区上的 4.1B 参数text-to-image-synthesis模型，采用 mit 许可，基于 microsoft/Lens 构建。

- **Repository**: mlx-community/Lens-3.8B-bf16
- **License**: mit
- **Tasks**: text-to-image-synthesis
- **Parameters**: 4.1B
- **Base model**: microsoft/Lens
- **Tags**: mlx, text-to-image, diffusion, lens, apple-silicon
- **Downloads**: 8
- **Stars**: 0
- **Last updated**: 2026-06-21

Source: https://www.modelscope.cn/models/mlx-community/Lens-3.8B-bf16

---

# Lens-3.8B-bf16 (MLX)

Apple **MLX** conversion of the denoising transformer (DiT) from
[`microsoft/Lens`](https://huggingface.co/microsoft/Lens) — a 3.8B-parameter foundational
text-to-image model — for fast inference on Apple Silicon. **bf16**, full precision.

This repo contains the **DiT only** (MIT-licensed). The full pipeline also uses the
GPT-OSS-20B text encoder (Apache-2.0) and the FLUX.2 semantic VAE, which the loader pulls
from their own sources rather than re-hosting here (see **License** below).

| component | parity vs PyTorch reference |
|---|---|
| GPT-OSS text features | per-layer cosine ≈ 0.998 |
| **Lens DiT (this repo)** | **cosine 0.999999** |
| FLUX.2 VAE decode | PSNR 57.65 dB |
| full end-to-end image | PSNR 45.26 dB |

Generates a 1024×1024 image in ~33 s on Apple Silicon (20 steps, ~39 GB peak).

![sample](sample.png)

## Usage

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

# `base` = a microsoft/Lens snapshot providing the tokenizer, GPT-OSS encoder, and FLUX.2 VAE.
pipe = LensPipeline.from_pretrained(base, dit_repo="mlx-community/Lens-3.8B-bf16")
img = pipe("A serene lake below snow-capped mountains, golden hour.",
           height=1024, width=1024, num_inference_steps=20, seed=42)
img.save("out.png")
```

## Conversion

Converted from `microsoft/Lens` with `recipes/convert_lens.py` (lens-mlx). The DiT is pure
Linear + RMSNorm; weights map 1:1 (no transpose) and every tensor is materialized before
save. Layer-by-layer parity against the PyTorch reference is in the lens-mlx test suite.

## License

- **DiT weights (this repo):** MIT, inherited from `microsoft/Lens`.
- **GPT-OSS-20B encoder:** Apache-2.0 (not included; reuse the mlx-community MXFP4 repo).
- **FLUX.2 VAE:** governed by its own (FLUX.2-dev) terms — **not re-hosted here**; the
  loader fetches it from source. Verify the VAE license for your use case.

## Citation

Upstream: [microsoft/Lens](https://huggingface.co/microsoft/Lens) ·
MLX port: [xocialize/lens-mlx](https://github.com/xocialize/lens-mlx)