Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import subprocess
|
3 |
+
from huggingface_hub import snapshot_download
|
4 |
+
|
5 |
+
def set_accelerate_default_config():
|
6 |
+
try:
|
7 |
+
subprocess.run(["accelerate", "config", "default"], check=True)
|
8 |
+
print("Accelerate default config set successfully!")
|
9 |
+
except subprocess.CalledProcessError as e:
|
10 |
+
print(f"An error occurred: {e}")
|
11 |
+
|
12 |
+
def train_dreambooth_lora_sdxl(instance_data_dir):
|
13 |
+
|
14 |
+
script_filename = "train_dreambooth_lora_sdxl.py" # Assuming it's in the same folder
|
15 |
+
|
16 |
+
command = [
|
17 |
+
"accelerate",
|
18 |
+
"launch",
|
19 |
+
script_filename, # Use the local script
|
20 |
+
"--pretrained_model_name_or_path=stabilityai/stable-diffusion-xl-base-1.0",
|
21 |
+
"--pretrained_vae_model_name_or_path=madebyollin/sdxl-vae-fp16-fix",
|
22 |
+
f"--instance_data_dir={instance_data_dir}",
|
23 |
+
"--output_dir=lora-trained-xl-colab",
|
24 |
+
"--mixed_precision=fp16",
|
25 |
+
"--instance_prompt=egnestl",
|
26 |
+
"--resolution=1024",
|
27 |
+
"--train_batch_size=2",
|
28 |
+
"--gradient_accumulation_steps=2",
|
29 |
+
"--gradient_checkpointing",
|
30 |
+
"--learning_rate=1e-4",
|
31 |
+
"--lr_scheduler=constant",
|
32 |
+
"--lr_warmup_steps=0",
|
33 |
+
"--enable_xformers_memory_efficient_attention",
|
34 |
+
"--mixed_precision=fp16",
|
35 |
+
"--use_8bit_adam",
|
36 |
+
"--max_train_steps=25",
|
37 |
+
"--checkpointing_steps=717",
|
38 |
+
"--seed=0",
|
39 |
+
"--push_to_hub"
|
40 |
+
]
|
41 |
+
|
42 |
+
try:
|
43 |
+
subprocess.run(command, check=True)
|
44 |
+
print("Training is finished!")
|
45 |
+
except subprocess.CalledProcessError as e:
|
46 |
+
print(f"An error occurred: {e}")
|
47 |
+
|
48 |
+
def main(dataset_url):
|
49 |
+
|
50 |
+
dataset_repo = dataset_url
|
51 |
+
|
52 |
+
# Automatically set local_dir based on the last part of dataset_repo
|
53 |
+
repo_parts = dataset_repo.split("/")
|
54 |
+
local_dir = f"./{repo_parts[-1]}" # Use the last part of the split
|
55 |
+
|
56 |
+
gr.Info("Downloading dataset ...")
|
57 |
+
|
58 |
+
snapshot_download(
|
59 |
+
dataset_repo,
|
60 |
+
local_dir=local_dir,
|
61 |
+
repo_type="dataset",
|
62 |
+
ignore_patterns=".gitattributes",
|
63 |
+
)
|
64 |
+
|
65 |
+
set_accelerate_default_config()
|
66 |
+
|
67 |
+
gr.Info("Training begins ...")
|
68 |
+
train_dreambooth_lora_sdxl(instance_data_dir=repo_parts[-1])
|
69 |
+
|
70 |
+
return "Done"
|
71 |
+
|
72 |
+
with gr.Blocks() as demo:
|
73 |
+
with gr.Column():
|
74 |
+
dataset_id = gr.Textbox(label="Dataset ID")
|
75 |
+
train_button = gr.Button("Train !")
|
76 |
+
status = gr.Textbox(labe="Training status")
|
77 |
+
|
78 |
+
train_button.click(
|
79 |
+
fn = main,
|
80 |
+
inputs = [dataset_id],
|
81 |
+
outputs = [status]
|
82 |
+
)
|
83 |
+
|
84 |
+
demo.queue().launch()
|