ideprado commited on
Commit
d7f9ea0
·
1 Parent(s): 54108d3

Model Card v1

Browse files
Files changed (3) hide show
  1. .gitattributes +1 -0
  2. README.md +45 -0
  3. output_tight_mosaic.jpeg +3 -0
.gitattributes CHANGED
@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
  test_image.png filter=lfs diff=lfs merge=lfs -text
 
 
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
  test_image.png filter=lfs diff=lfs merge=lfs -text
37
+ *.jpeg filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ # F-Lite Model Card
3
+
4
+ ![F-Lite mosaic](output_tight_mosaic.jpeg)
5
+
6
+ F-Lite is a 10B parameter diffusion model created by [Freepik](https://www.freepik.com) and [Fal](https://fal.ai), trained exclusively on copyright-safe and SFW content. The model was trained on Freepik's internal dataset comprising approximately 80 million copyright-safe images, making it the first publicly available model of this scale trained exclusively on legally compliant and SFW content.
7
+
8
+ ## Usage
9
+
10
+ Experience F-Lite instantly through our [interactive demo](https://huggingface.co/spaces/Freepik/F-Lite) on Hugging Face.
11
+
12
+ F-Lite works with both the `diffusers` library and [ComfyUI](https://www.comfy.org/). For details, see the [F-Lite GitHub repository](https://github.com/Freepik/F-Lite).
13
+
14
+ # Technical Report
15
+
16
+ Read the [technical report](XXXXX) to learn more about the model details.
17
+
18
+ # Limitations and Bias
19
+ * The models can generate malformations.
20
+ * The text capabilities of the model are limited.
21
+ * The model can be subject to biases, although we think we have a good balance given the quality and variety of the Freepik's dataset.
22
+
23
+ # Recommendations
24
+ * Use long prompts to generate better results. Short prompts may result in low-quality images.
25
+ * Generate images above the megapixel. Smaller images will result in low-quality images.
26
+
27
+ ## Acknowledgements
28
+
29
+ This model uses [T5 XXL](https://huggingface.co/google/t5-v1_1-xxl)and [Flux Schnell VAE](https://huggingface.co/black-forest-labs/FLUX.1-schnell)
30
+
31
+ ## License
32
+
33
+ The F-Lite weights are licensed under the permissive CreativeML Open RAIL-M license. The T5 XXL and Flux Schnell VAE are licensed under Apache 2.0.
34
+
35
+ ## Citation
36
+
37
+ If you find our work helpful, please cite it!
38
+
39
+ ```
40
+ @article{ryu2025flite,
41
+ title={F-Lite Technical Report},
42
+ author={Ryu, Simo and Pengqi, Lu and Mart\'in Juan, Javier and de Prado Alonso, Iv\'an},
43
+ year={2025}
44
+ }
45
+ ```
output_tight_mosaic.jpeg ADDED

Git LFS Details

  • SHA256: 80c5ec0d5692767f76f2880b4aa9ec8086485d7e5ffa71ab70b786cad8cbd67a
  • Pointer size: 132 Bytes
  • Size of remote file: 1.01 MB