kaanakan commited on
Commit
989e91d
·
verified ·
1 Parent(s): 3f95087

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip 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
 
 
 
33
  *.zip 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
+ example_edit.png filter=lfs diff=lfs merge=lfs -text
37
+ example_t2i.png filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: other
5
+ license_name: flux-dev-non-commercial-license
6
+ license_link: https://huggingface.co/black-forest-labs/FLUX.2-dev/blob/main/LICENSE.txt
7
+ base_model: black-forest-labs/FLUX.2-dev
8
+ tags:
9
+ - image-generation
10
+ - flux
11
+ - lora
12
+ - distillation
13
+ - turbo
14
+ pipeline_tag: text-to-image
15
+ library_name: diffusers
16
+ ---
17
+
18
+ # FLUX.2 [dev] Turbo LoRA
19
+
20
+ **FLUX.2 [dev] Turbo** is a distilled LoRA adapter for [FLUX.2 [dev]](https://huggingface.co/black-forest-labs/FLUX.2-dev) that enables high-quality image generation in just **8 inference steps**.
21
+
22
+ ![Example Text-to-Image](./example_t2i.png)
23
+ ---
24
+
25
+ ---
26
+
27
+ ![Example Image Editing](./example_edit.png)
28
+
29
+ ## Key Features
30
+
31
+ - ⚡ **8-step inference** — 6x faster than the base model's typical 50 steps
32
+ - 🎨 **Quality preserved** — Matches or surpasses the original FLUX.2 [dev] quality
33
+ - 🔌 **LoRA adapter** — Lightweight, easy to integrate with existing FLUX.2 workflows
34
+ - 🖼️ **Multiple modes** — Supports both text-to-image and image editing
35
+
36
+
37
+ ## Hosted API Endpoints
38
+
39
+ FLUX.2 [dev] Turbo is available through fal.ai hosted endpoints:
40
+ - [Text-to-Image](https://fal.ai/models/fal-ai/flux-2/turbo)
41
+ - [Image Editing](https://fal.ai/models/fal-ai/flux-2/turbo/edit)
42
+
43
+
44
+ ## Usage
45
+
46
+ ```python
47
+ import torch
48
+ from diffusers import Flux2Pipeline
49
+
50
+ # Pre-shifted custom sigmas for 8-step turbo inference
51
+ TURBO_SIGMAS = [1.0, 0.6509, 0.4374, 0.2932, 0.1893, 0.1108, 0.0495, 0.00031]
52
+
53
+ pipe = Flux2Pipeline.from_pretrained(
54
+ "black-forest-labs/FLUX.2-dev",
55
+ torch_dtype=torch.bfloat16
56
+ ).to("cuda")
57
+
58
+ pipe.load_lora_weights(
59
+ "fal/FLUX.2-Turbo",
60
+ weight_name="flux.2-turbo-lora.safetensors"
61
+ )
62
+
63
+ prompt = "Industrial product shot of a chrome turbocharger with glowing hot exhaust manifold, engraved text 'FLUX.2 [dev] Turbo by fal' on the compressor housing and 'fal' on the turbine wheel, gradient heat glow from orange to electric blue , studio lighting with dramatic shadows, shallow depth of field, engineering blueprint pattern in background."
64
+
65
+ image = pipe(
66
+ prompt=prompt,
67
+ sigmas=TURBO_SIGMAS,
68
+ guidance_scale=2.5,
69
+ height=1024,
70
+ width=1024,
71
+ num_inference_steps=8,
72
+ ).images[0]
73
+
74
+ image.save("output.png")
75
+ ```
76
+
77
+ ## License
78
+
79
+ This model inherits the [FLUX [dev] Non-Commercial License](https://huggingface.co/black-forest-labs/FLUX.2-dev/blob/main/LICENSE.txt) from the base model.
example_edit.png ADDED

Git LFS Details

  • SHA256: 0372a83c07acc7bba7e7d92d08b5e6a8d96771e46962d59f99f107a62498c0f1
  • Pointer size: 132 Bytes
  • Size of remote file: 5.01 MB
example_t2i.png ADDED

Git LFS Details

  • SHA256: 025e69e8a70a27813e0f756ad9125e8e0b80c624e0f2bc8179aad5a6210cf67e
  • Pointer size: 132 Bytes
  • Size of remote file: 6.57 MB
flux.2-turbo-lora.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f76cf9c2cc546ddca878799136434a1098477af3f4b0adff2cfd79f2ebe4aa01
3
+ size 2760818216