Text-to-Image
Diffusers
Safetensors
lora
onurxtasar commited on
Commit
12c5a4d
1 Parent(s): b6ddb55

Update README.md

Browse files

- Added training details & metrics.
- Added Eyal's name.

Files changed (1) hide show
  1. README.md +5 -2
README.md CHANGED
@@ -10,7 +10,7 @@ inference: False
10
  # ⚡ FlashDiffusion: FlashPixart ⚡
11
 
12
 
13
- Flash Diffusion is a diffusion distillation method proposed in [ADD ARXIV]() *by Clément Chadebec, Onur Tasar and Benjamin Aubin.*
14
  This model is a **66.5M** LoRA distilled version of Pixart-α model that is able to generate 1024x1024 images in **4 steps**. See our [live demo](https://huggingface.co/spaces/jasperai/FlashPixart).
15
 
16
 
@@ -63,8 +63,11 @@ image = pipe(prompt, num_inference_steps=4, guidance_scale=0).images[0]
63
  </p>
64
 
65
  # Training Details
66
- The model was trained for 40k iterations on 4 H100 GPUs. Please refer to the [paper]() for further parameters details.
67
 
 
 
 
68
 
69
  ## License
70
  This model is released under the the Creative Commons BY-NC license.
 
10
  # ⚡ FlashDiffusion: FlashPixart ⚡
11
 
12
 
13
+ Flash Diffusion is a diffusion distillation method proposed in [ADD ARXIV]() *by Clément Chadebec, Onur Tasar, Eyal Benaroche, and Benjamin Aubin.*
14
  This model is a **66.5M** LoRA distilled version of Pixart-α model that is able to generate 1024x1024 images in **4 steps**. See our [live demo](https://huggingface.co/spaces/jasperai/FlashPixart).
15
 
16
 
 
63
  </p>
64
 
65
  # Training Details
66
+ The model was trained for 40k iterations on 4 H100 GPUs (representing approximately 188 hours of training). Please refer to the [paper]() for further parameters details.
67
 
68
+ **Metrics on COCO 2014 validation (Table 4)**
69
+ - FID-10k: 29.30 (4 NFE)
70
+ - CLIP Score: 0.303 (4 NFE)
71
 
72
  ## License
73
  This model is released under the the Creative Commons BY-NC license.