prithivMLmods
commited on
Commit
•
1179177
1
Parent(s):
8b4dd1d
Update README.md
Browse files
README.md
CHANGED
@@ -22,12 +22,54 @@ license: creativeml-openrail-m
|
|
22 |
|
23 |
<Gallery />
|
24 |
|
|
|
25 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
## Trigger words
|
27 |
|
28 |
You should use `Minimal Futuristic` to trigger the image generation.
|
29 |
|
30 |
-
|
31 |
## Download model
|
32 |
|
33 |
Weights for this model are available in Safetensors format.
|
|
|
22 |
|
23 |
<Gallery />
|
24 |
|
25 |
+
**The model is still in the training phase. This is not the final version and may contain artifacts and perform poorly in some cases.**
|
26 |
|
27 |
+
## Model description
|
28 |
+
|
29 |
+
**prithivMLmods/Minimal-Futuristic-Flux-LoRA**
|
30 |
+
|
31 |
+
Image Processing Parameters
|
32 |
+
|
33 |
+
| Parameter | Value | Parameter | Value |
|
34 |
+
|---------------------------|--------|---------------------------|--------|
|
35 |
+
| LR Scheduler | constant | Noise Offset | 0.03 |
|
36 |
+
| Optimizer | AdamW | Multires Noise Discount | 0.1 |
|
37 |
+
| Network Dim | 64 | Multires Noise Iterations | 10 |
|
38 |
+
| Network Alpha | 32 | Repeat & Steps | 25 & 2.7K|
|
39 |
+
| Epoch | 14 | Save Every N Epochs | 1 |
|
40 |
+
|
41 |
+
Labeling: florence2-en(natural language & English)
|
42 |
+
|
43 |
+
Total Images Used for Training : 20 [ Hi-RES ]
|
44 |
+
|
45 |
+
## Best Dimensions
|
46 |
+
|
47 |
+
- 1024 x 1024 (Default)
|
48 |
+
|
49 |
+
## Setting Up
|
50 |
+
```
|
51 |
+
import torch
|
52 |
+
from pipelines import DiffusionPipeline
|
53 |
+
|
54 |
+
base_model = "black-forest-labs/FLUX.1-dev"
|
55 |
+
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
|
56 |
+
|
57 |
+
lora_repo = "prithivMLmods/Minimal-Futuristic-Flux-LoRA"
|
58 |
+
trigger_word = "Minimal Futuristic"
|
59 |
+
pipe.load_lora_weights(lora_repo)
|
60 |
+
|
61 |
+
device = torch.device("cuda")
|
62 |
+
pipe.to(device)
|
63 |
+
```
|
64 |
+
|
65 |
+
## Data source
|
66 |
+
|
67 |
+
- https://playground.com/
|
68 |
+
|
69 |
## Trigger words
|
70 |
|
71 |
You should use `Minimal Futuristic` to trigger the image generation.
|
72 |
|
|
|
73 |
## Download model
|
74 |
|
75 |
Weights for this model are available in Safetensors format.
|