Commit
·
04b7e5f
1
Parent(s):
a389f19
fix: Update README
Browse files
README.md
CHANGED
@@ -14,25 +14,28 @@ tags:
|
|
14 |
![Flux.1 Lite](sample_images/flux1-lite-8B_sample.png)
|
15 |
|
16 |
# Flux.1 Lite
|
|
|
17 |
|
18 |
-
|
|
|
|
|
|
|
19 |
|
20 |
![Flux.1 Lite vs FLUX.1-dev](sample_images/models_comparison.png)
|
21 |
|
22 |
## Text-to-Image
|
23 |
|
24 |
-
Flux.1 Lite is ready to unleash your creativity! For the best results, we strongly **recommend using a `guidance_scale`
|
25 |
|
26 |
```python
|
27 |
import torch
|
28 |
from diffusers import FluxPipeline
|
29 |
|
30 |
-
base_model_id = "Freepik/flux.1-lite-8B-alpha"
|
31 |
torch_dtype = torch.bfloat16
|
32 |
device = "cuda"
|
33 |
|
34 |
# Load the pipe
|
35 |
-
model_id = "Freepik/flux.1-lite-8B
|
36 |
pipe = FluxPipeline.from_pretrained(
|
37 |
model_id, torch_dtype=torch_dtype
|
38 |
).to(device)
|
@@ -40,7 +43,7 @@ pipe = FluxPipeline.from_pretrained(
|
|
40 |
# Inference
|
41 |
prompt = "A close-up image of a green alien with fluorescent skin in the middle of a dark purple forest"
|
42 |
|
43 |
-
guidance_scale = 3.5
|
44 |
n_steps = 28
|
45 |
seed = 11
|
46 |
|
@@ -78,14 +81,14 @@ Stay tuned! Our goal is to distill FLUX.1-dev further until it can run smoothly
|
|
78 |
We've also crafted a ComfyUI workflow to make using Flux.1 Lite even more seamless! Find it in `comfy/flux.1-lite_workflow.json`.
|
79 |
![ComfyUI workflow](comfy/flux.1-lite_workflow.png)
|
80 |
|
81 |
-
The safetensors checkpoint is available here: [flux.1-lite-8B
|
82 |
|
83 |
## Try it out at Freepik!
|
84 |
|
85 |
Our [AI generator](https://www.freepik.com/pikaso/ai-image-generator) is now powered by Flux.1 Lite!
|
86 |
|
87 |
## 🔥 News 🔥
|
88 |
-
|
89 |
* Oct 23, 2024. Alpha 8B checkpoint is publicly available on [HuggingFace Repo](https://huggingface.co/Freepik/flux.1-lite-8B-alpha).
|
90 |
|
91 |
## Citation
|
|
|
14 |
![Flux.1 Lite](sample_images/flux1-lite-8B_sample.png)
|
15 |
|
16 |
# Flux.1 Lite
|
17 |
+
We are thrilled to announce the release of Flux.1 Lite, an 8B parameter transformer model distilled from the FLUX.1-dev model. This version uses 7 GB less RAM and runs 23% faster while maintaining the same precision (bfloat16) as the original model.
|
18 |
|
19 |
+
`🔥 UPDATE 🔥`: We have released a new version of Flux.1 Lite 8B. This version is trained with a new dataset and achieves better results than the previous alpha version. The main changes include:
|
20 |
+
- Distillation for a broader range of guidance values (2.0-5.0)
|
21 |
+
- Distillation for a broader range of number of steps (20-32)
|
22 |
+
- More diverse dataset with longer prompts
|
23 |
|
24 |
![Flux.1 Lite vs FLUX.1-dev](sample_images/models_comparison.png)
|
25 |
|
26 |
## Text-to-Image
|
27 |
|
28 |
+
Flux.1 Lite is ready to unleash your creativity! For the best results, we strongly **recommend using a `guidance_scale` between 2.0 and 5.0 and setting `n_steps` between 20 and 32**.
|
29 |
|
30 |
```python
|
31 |
import torch
|
32 |
from diffusers import FluxPipeline
|
33 |
|
|
|
34 |
torch_dtype = torch.bfloat16
|
35 |
device = "cuda"
|
36 |
|
37 |
# Load the pipe
|
38 |
+
model_id = "Freepik/flux.1-lite-8B"
|
39 |
pipe = FluxPipeline.from_pretrained(
|
40 |
model_id, torch_dtype=torch_dtype
|
41 |
).to(device)
|
|
|
43 |
# Inference
|
44 |
prompt = "A close-up image of a green alien with fluorescent skin in the middle of a dark purple forest"
|
45 |
|
46 |
+
guidance_scale = 3.5
|
47 |
n_steps = 28
|
48 |
seed = 11
|
49 |
|
|
|
81 |
We've also crafted a ComfyUI workflow to make using Flux.1 Lite even more seamless! Find it in `comfy/flux.1-lite_workflow.json`.
|
82 |
![ComfyUI workflow](comfy/flux.1-lite_workflow.png)
|
83 |
|
84 |
+
The safetensors checkpoint is available here: [flux.1-lite-8B.safetensors](flux.1-lite-8B.safetensors)
|
85 |
|
86 |
## Try it out at Freepik!
|
87 |
|
88 |
Our [AI generator](https://www.freepik.com/pikaso/ai-image-generator) is now powered by Flux.1 Lite!
|
89 |
|
90 |
## 🔥 News 🔥
|
91 |
+
* Dec 30, 2024. Flux.1 Lite 8B new trained model is publicly available on [HuggingFace Repo](https://huggingface.co/Freepik/flux.1-lite-8B).
|
92 |
* Oct 23, 2024. Alpha 8B checkpoint is publicly available on [HuggingFace Repo](https://huggingface.co/Freepik/flux.1-lite-8B-alpha).
|
93 |
|
94 |
## Citation
|