dverdu-freepik
commited on
Commit
•
e45409a
1
Parent(s):
abf16b5
fix: Update README.md
Browse files- README.md +20 -17
- sample_images/models_comparison.png +3 -0
README.md
CHANGED
@@ -19,25 +19,12 @@ We want to announce the alpha version of our new distilled Flux.1 Lite model, an
|
|
19 |
|
20 |
Our goal is to further reduce FLUX.1-dev transformer parameters up to 24Gb to make it compatible with most of GPU cards.
|
21 |
|
22 |
-
|
23 |
-
## News🔥🔥🔥
|
24 |
-
* Oct.18, 2024. Alpha 8B checkpoint and comparison demo 🤗 (i.e. [Flux.1 Lite](https://huggingface.co/spaces/Freepik/flux.1-lite)) is publicly available on [HuggingFace Repo](https://huggingface.co/Freepik/flux.1-lite-8B-alpha).
|
25 |
-
|
26 |
-
## Try our Hugging Face demos:
|
27 |
-
Flux.1 Lite demo host on [🤗 flux.1-lite](https://huggingface.co/spaces/Freepik/flux.1-lite)
|
28 |
-
|
29 |
-
## Introduction
|
30 |
-
|
31 |
-
Hyper-SD is one of the new State-of-the-Art diffusion model acceleration techniques.
|
32 |
-
In this repository, we release the models distilled from [FLUX.1-dev](https://huggingface.co/black-forest-labs/FLUX.1-dev), [SD3-Medium](https://huggingface.co/stabilityai/stable-diffusion-3-medium-diffusers), [SDXL Base 1.0](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0) and [Stable-Diffusion v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5)。
|
33 |
-
|
34 |
-
## Checkpoints
|
35 |
-
|
36 |
-
* `flux.1-lite-8B-alpha.safetensors`: Transformer checkpoint, in Flux original format.
|
37 |
|
38 |
## Text-to-Image Usage
|
39 |
|
40 |
-
|
|
|
41 |
```python
|
42 |
import torch
|
43 |
from diffusers import FluxPipeline
|
@@ -69,4 +56,20 @@ with torch.inference_mode():
|
|
69 |
width=1024,
|
70 |
).images[0]
|
71 |
image.save("output.png")
|
72 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
19 |
|
20 |
Our goal is to further reduce FLUX.1-dev transformer parameters up to 24Gb to make it compatible with most of GPU cards.
|
21 |
|
22 |
+
![Flux.1 Lite vs FLUX.1-dev](./sample_images/models_comparison.png)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
23 |
|
24 |
## Text-to-Image Usage
|
25 |
|
26 |
+
It is recommended to use a `guidance_scale` of 3.5 and a `n_steps` between 22 and 30 for best results.
|
27 |
+
|
28 |
```python
|
29 |
import torch
|
30 |
from diffusers import FluxPipeline
|
|
|
56 |
width=1024,
|
57 |
).images[0]
|
58 |
image.save("output.png")
|
59 |
+
```
|
60 |
+
|
61 |
+
## Checkpoints
|
62 |
+
|
63 |
+
* `flux.1-lite-8B-alpha.safetensors`: Transformer checkpoint, in Flux original format.
|
64 |
+
* `transformers/`: Contains distilled 8B transformer model, in diffusers format.
|
65 |
+
|
66 |
+
## Try our Hugging Face demos:
|
67 |
+
Flux.1 Lite demo host on [🤗 flux.1-lite](https://huggingface.co/spaces/Freepik/flux.1-lite)
|
68 |
+
|
69 |
+
## News🔥🔥🔥
|
70 |
+
* Oct.18, 2024. Alpha 8B checkpoint and comparison demo 🤗 (i.e. [Flux.1 Lite](https://huggingface.co/spaces/Freepik/flux.1-lite)) is publicly available on [HuggingFace Repo](https://huggingface.co/Freepik/flux.1-lite-8B-alpha).
|
71 |
+
|
72 |
+
|
73 |
+
|
74 |
+
|
75 |
+
|
sample_images/models_comparison.png
ADDED
Git LFS Details
|