DongChenMSRA commited on
Commit
3fa9e97
·
verified ·
1 Parent(s): b88e057

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +6 -6
README.md CHANGED
@@ -33,7 +33,7 @@
33
  <p>
34
  <a href="https://arxiv.org/abs/PLACEHOLDER"><img alt="arXiv" src="https://img.shields.io/badge/arXiv-Paper-b31b1b?logo=arxiv&logoColor=white" height="22" /></a>
35
  &nbsp;
36
- <a href="https://huggingface.co/Jinjing713/Lens"><img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97-Models-yellow" height="22" /></a>
37
  &nbsp;
38
  <a href="LICENSE"><img alt="License: MIT" src="https://img.shields.io/badge/License-MIT-green.svg" height="22" /></a>
39
  </p>
@@ -307,9 +307,9 @@ The default GPT-OSS encoder and FLUX.2 VAE are loaded from Hugging Face. Make su
307
 
308
  | Repo | Description | Recommended Steps | CFG |
309
  | :--- | :--- | :---: | :---: |
310
- | [`Jinjing713/Lens`](https://huggingface.co/Jinjing713/Lens) | **Default.** RL-tuned for visual quality | 20 | 5.0 |
311
- | [`Jinjing713/Lens-Turbo`](https://huggingface.co/Jinjing713/Lens-Turbo) | Distilled from the RL model for fast 4-step sampling | 4 | 1.0 |
312
- | [`Jinjing713/Lens-Base`](https://huggingface.co/Jinjing713/Lens-Base) | Supervised base model (no RL, no distillation) | 50 | 5.0 |
313
 
314
  Pick a variant by passing its repo id to `--repo_id` (CLI) or `LensPipeline.from_pretrained(...)` (Python).
315
 
@@ -324,7 +324,7 @@ import torch
324
  from lens import LensPipeline
325
 
326
  pipe = LensPipeline.from_pretrained(
327
- "Jinjing713/Lens", torch_dtype=torch.bfloat16
328
  ).to("cuda")
329
 
330
  image = pipe(
@@ -366,7 +366,7 @@ python inference.py --prompt "a cat" --disable_mxfp4 --offload
366
 
367
  | Flag | Description | Default |
368
  | :--- | :--- | :--- |
369
- | `--repo_id` | HF repo id (or local path) of the assembled Lens pipeline | `Jinjing713/Lens` |
370
  | `--base_resolution` | `1024` or `1440` | `1440` |
371
  | `--aspect_ratio` | `1:2`, `9:16`, `2:3`, `3:4`, `1:1`, `4:3`, `3:2`, `16:9`, `2:1` | `1:1` |
372
  | `--steps` | Number of denoising steps | `20` |
 
33
  <p>
34
  <a href="https://arxiv.org/abs/PLACEHOLDER"><img alt="arXiv" src="https://img.shields.io/badge/arXiv-Paper-b31b1b?logo=arxiv&logoColor=white" height="22" /></a>
35
  &nbsp;
36
+ <a href="https://huggingface.co/microsoft/Lens"><img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97-Models-yellow" height="22" /></a>
37
  &nbsp;
38
  <a href="LICENSE"><img alt="License: MIT" src="https://img.shields.io/badge/License-MIT-green.svg" height="22" /></a>
39
  </p>
 
307
 
308
  | Repo | Description | Recommended Steps | CFG |
309
  | :--- | :--- | :---: | :---: |
310
+ | [`microsoft/Lens`](https://huggingface.co/microsoft/Lens) | **Default.** RL-tuned for visual quality | 20 | 5.0 |
311
+ | [`microsoft/Lens-Turbo`](https://huggingface.co/microsoft/Lens-Turbo) | Distilled from the RL model for fast 4-step sampling | 4 | 1.0 |
312
+ | [`microsoft/Lens-Base`](https://huggingface.co/microsoft/Lens-Base) | Supervised base model (no RL, no distillation) | 50 | 5.0 |
313
 
314
  Pick a variant by passing its repo id to `--repo_id` (CLI) or `LensPipeline.from_pretrained(...)` (Python).
315
 
 
324
  from lens import LensPipeline
325
 
326
  pipe = LensPipeline.from_pretrained(
327
+ "microsoft/Lens", torch_dtype=torch.bfloat16
328
  ).to("cuda")
329
 
330
  image = pipe(
 
366
 
367
  | Flag | Description | Default |
368
  | :--- | :--- | :--- |
369
+ | `--repo_id` | HF repo id (or local path) of the assembled Lens pipeline | `microsoft/Lens` |
370
  | `--base_resolution` | `1024` or `1440` | `1440` |
371
  | `--aspect_ratio` | `1:2`, `9:16`, `2:3`, `3:4`, `1:1`, `4:3`, `3:2`, `16:9`, `2:1` | `1:1` |
372
  | `--steps` | Number of denoising steps | `20` |