Create README.md
Browse files![00021.jpg](https://s3.amazonaws.com/moonup/production/uploads/1667941162243-6305d083df993a789e61126d.jpeg)
![00022.jpg](https://s3.amazonaws.com/moonup/production/uploads/1667941163855-6305d083df993a789e61126d.jpeg)
![00023.jpg](https://s3.amazonaws.com/moonup/production/uploads/1667941162648-6305d083df993a789e61126d.jpeg)
![00024.jpg](https://s3.amazonaws.com/moonup/production/uploads/1667941165190-6305d083df993a789e61126d.jpeg)
README.md
ADDED
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
## Model description
|
2 |
+
This model trained based on Stable Diffusion 1.5 model to create isometric cities, venues, etc more precisely.
|
3 |
+
|
4 |
+
## Intended uses & limitations
|
5 |
+
Do not use for commercial purposes.
|
6 |
+
|
7 |
+
## Training procedure
|
8 |
+
Trained on a WSL2 Ubuntu machine with DreamBooth.
|
9 |
+
|
10 |
+
### Training hyperparameters
|
11 |
+
The following hyperparameters were used during training:
|
12 |
+
--with_prior_preservation --prior_loss_weight=1.0 \
|
13 |
+
--instance_prompt="an illustration of an isometric city, digital art, highly detailed" \
|
14 |
+
--class_prompt="isometric city" \
|
15 |
+
--resolution=512 \
|
16 |
+
--train_batch_size=1 \
|
17 |
+
--mixed_precision="fp16" \
|
18 |
+
--gradient_accumulation_steps=1 --gradient_checkpointing \
|
19 |
+
--use_8bit_adam \
|
20 |
+
--learning_rate=5e-6 \
|
21 |
+
--lr_scheduler="constant" \
|
22 |
+
--lr_warmup_steps=0 \
|
23 |
+
--num_class_images=50 \
|
24 |
+
--max_train_steps=1500
|
25 |
+
|
26 |
+
### Training results
|