File size: 3,730 Bytes
baa4c14
 
 
 
 
 
 
 
 
 
 
 
87990e7
baa4c14
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
87990e7
baa4c14
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0b65bf8
caa5aa4
baa4c14
 
 
 
953315f
baa4c14
 
953315f
baa4c14
 
 
 
 
953315f
baa4c14
 
 
 
 
 
 
 
 
 
398b1c0
 
953315f
caa5aa4
398b1c0
 
 
 
 
baa4c14
 
953315f
d2f5ad4
baa4c14
 
 
 
6a2ecf6
baa4c14
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
---
license: apache-2.0
base_model: "fal/AuraFlow"
tags:
  - auraflow
  - auraflow-diffusers
  - text-to-image
  - image-to-image
  - diffusers
  - simpletuner
  - not-for-all-audiences
  - lora

  - template:sd-lora
  - standard
pipeline_tag: text-to-image
inference: true
widget:
- text: 'unconditional (blank prompt)'
  parameters:
    negative_prompt: 'ugly, cropped, blurry, low-quality, mediocre average'
  output:
    url: ./assets/image_0_0.png
- text: 'An domokun running through a field with flowers all around him.'
  parameters:
    negative_prompt: 'ugly, cropped, blurry, low-quality, mediocre average'
  output:
    url: ./assets/image_1_0.png
---

# Auraflow-DomoKun-LoRA-rank8

This is a PEFT LoRA derived from [fal/AuraFlow](https://huggingface.co/fal/AuraFlow).

The main validation prompt used during training was:
```
An domokun running through a field with flowers all around him.
```


## Validation settings
- CFG: `4.0`
- CFG Rescale: `0.0`
- Steps: `30`
- Sampler: `FlowMatchEulerDiscreteScheduler`
- Seed: `42`
- Resolution: `512x512`


Note: The validation settings are not necessarily the same as the [training settings](#training-settings).

You can find some example images in the following gallery:


<Gallery />

The text encoder **was not** trained.
You may reuse the base model text encoder for inference.


## Training settings

- Training epochs: 1
- Training steps: 1000
- Learning rate: 0.0001
  - Learning rate schedule: constant
  - Warmup steps: 100
- Max grad value: 0.01
- Effective batch size: 1
  - Micro-batch size: 1
  - Gradient accumulation steps: 1
  - Number of GPUs: 1
- Gradient checkpointing: True
- Prediction type: flow_matching (extra parameters=['shift=3'])
- Optimizer: adamw_bf16
- Trainable parameter precision: Pure BF16
- Base model precision: `no_change`
- Caption dropout probability: 0.1%


- LoRA Rank: 8
- LoRA Alpha: 8.0
- LoRA Dropout: 0.1
- LoRA initialisation style: default
    

## Datasets

### domokun-cropped-512-NonReg
- Repeats: 10
- Total number of images: 27
- Total number of aspect buckets: 3
- Resolution: 0.262144 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
### domokun-cropped-512
- Repeats: 10
- Total number of images: 27
- Total number of aspect buckets: 7
- Resolution: 0.262144 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: Yes


## Inference


```python
import torch
from diffusers import DiffusionPipeline

model_id = 'fal/AuraFlow'
adapter_id = 'bghira/Auraflow-DomoKun-LoRA-rank8'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "An domokun running through a field with flowers all around him."
negative_prompt = 'ugly, cropped, blurry, low-quality, mediocre average'

## Optional: quantise the model to save on vram.
## Note: The model was not quantised during training, so it is not necessary to quantise it during inference time.
#from optimum.quanto import quantize, freeze, qint8
#quantize(pipeline.transformer, weights=qint8)
#freeze(pipeline.transformer)
    
pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu') # the pipeline is already in its target precision level
model_output = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    num_inference_steps=30,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(42),
    width=512,
    height=512,
    guidance_scale=4.0,
).images[0]

model_output.save("output.png", format="PNG")

```