ByteDream / workflow_complete.py
Enzo8930302's picture
Upload workflow_complete.py with huggingface_hub
d3516f4 verified
Raw
History Blame Contribute Delete
5.54 kB
"""
Complete Workflow Script
Creates dataset, trains model, and uploads to Hugging Face
"""
import subprocess
import sys
from pathlib import Path
def print_banner(text):
print("\n" + "="*60)
print(text.center(60))
print("="*60 + "\n")
def check_dataset():
"""Check if dataset exists"""
dataset_dir = Path("./dataset")
if not dataset_dir.exists():
print("❌ Dataset directory not found!")
return False
# Check for images
images = list(dataset_dir.glob("*.jpg")) + list(dataset_dir.glob("*.png"))
if len(images) == 0:
print("⚠ No images found in dataset directory!")
return False
print(f"✓ Dataset found with {len(images)} images")
return True
def create_dataset():
"""Create test dataset"""
print_banner("STEP 1: CREATE DATASET")
if check_dataset():
print("\nDataset already exists!")
response = input("Overwrite? (y/n): ").strip().lower()
if response != 'y':
print("Skipping dataset creation...")
return True
print("\nCreating test dataset...")
try:
from create_test_dataset import create_test_dataset
create_test_dataset(output_dir="./dataset", num_images=10)
print("\n✅ Dataset created successfully!")
return True
except Exception as e:
print(f"\n❌ Error creating dataset: {e}")
import traceback
traceback.print_exc()
return False
def train_model():
"""Train the model"""
print_banner("STEP 2: TRAIN MODEL")
print("\nStarting training...")
print("This may take several minutes depending on your CPU...\n")
try:
# Run training
result = subprocess.run([
sys.executable, "train.py",
"--train_data", "./dataset",
"--output_dir", "./models/bytedream",
"--device", "cpu"
], check=True)
print("\n✅ Training completed successfully!")
return True
except subprocess.CalledProcessError as e:
print(f"\n❌ Training failed: {e}")
return False
except Exception as e:
print(f"\n❌ Error: {e}")
import traceback
traceback.print_exc()
return False
def upload_to_hf(token, repo_id):
"""Upload to Hugging Face"""
print_banner("STEP 3: UPLOAD TO HUGGING FACE")
# Check if model exists
model_dir = Path("./models/bytedream")
if not model_dir.exists():
print("❌ Model directory not found!")
print("Please train the model first.")
return False
print(f"\nUploading to {repo_id}...")
try:
from bytedream.generator import ByteDreamGenerator
# Load generator
print("\nLoading model...")
generator = ByteDreamGenerator(
model_path="./models/bytedream",
config_path="config.yaml",
device="cpu",
)
# Upload to HF
generator.push_to_hub(
repo_id=repo_id,
token=token,
private=False,
commit_message="Upload Byte Dream model",
)
print("\n✅ Upload successful!")
print(f"\n📦 Your model is available at:")
print(f"https://huggingface.co/{repo_id}")
return True
except Exception as e:
print(f"\n❌ Upload failed: {e}")
import traceback
traceback.print_exc()
return False
def main():
"""Main workflow"""
print_banner("BYTE DREAM - COMPLETE WORKFLOW")
print("This script will:")
print("1. Create test dataset")
print("2. Train the model")
print("3. Upload to Hugging Face")
print()
# Step 1: Create dataset
if not create_dataset():
print("\n❌ Failed to create dataset. Exiting...")
return
# Step 2: Train
print("\nReady to train the model.")
response = input("Continue to training? (y/n): ").strip().lower()
if response != 'y':
print("Training cancelled.")
return
if not train_model():
print("\n❌ Training failed. Exiting...")
return
# Step 3: Upload to HF
print("\nReady to upload to Hugging Face.")
response = input("Continue to upload? (y/n): ").strip().lower()
if response != 'y':
print("Upload cancelled.")
print("\nModel saved to: ./models/bytedream")
print("To upload later: python publish_to_hf.py")
return
# Get HF credentials
token = input("\nEnter your Hugging Face token (hf_...): ").strip()
if not token:
print("❌ Token required!")
return
repo_id = input("Enter repository ID (e.g., Enzo8930302/ByteDream): ").strip()
if not repo_id:
print("❌ Repository ID required!")
return
if not upload_to_hf(token, repo_id):
print("\n❌ Upload failed.")
return
# Success!
print_banner("WORKFLOW COMPLETED SUCCESSFULLY!")
print("✅ Dataset created")
print("✅ Model trained")
print(f"✅ Uploaded to Hugging Face: {repo_id}")
print()
if __name__ == "__main__":
try:
main()
except KeyboardInterrupt:
print("\n\nWorkflow interrupted!")
sys.exit(0)