upload_to_hf_example.py

"""
Example: How to Upload SAE Weights to Hugging Face

This script shows how to upload your trained SAE weights to Hugging Face
so they can be easily loaded in the demo app.
"""

from huggingface_hub import HfApi, create_repo
import os

api = HfApi()

REPO_ID = "your-username/sae-weights"
SAE_CHECKPOINT_PATH = "path/to/your/sae_l16.pt"

try:
    create_repo(REPO_ID, repo_type="model", exist_ok=True)
    print(f"Created repository: {REPO_ID}")
except Exception as e:
    print(f"Repository may already exist: {e}")

api.upload_file(
    path_or_fileobj=SAE_CHECKPOINT_PATH,
    path_in_repo="sae_l16.pt",
    repo_id=REPO_ID,
    repo_type="model"
)

print(f"Uploaded SAE weights to {REPO_ID}")
print(f"Load in demo with:")
print(f"   Repository ID: {REPO_ID}")
print(f"   Filename: sae_l16.pt")

print("\n" + "="*60)
print("Example: Upload Multiple SAE Checkpoints")
print("="*60)

checkpoints = {
    "sae_layer_8.pt": "path/to/sae_l8.pt",
    "sae_layer_16.pt": "path/to/sae_l16.pt",
    "sae_layer_24.pt": "path/to/sae_l24.pt"
}

for filename, local_path in checkpoints.items():
    if os.path.exists(local_path):
        api.upload_file(
            path_or_fileobj=local_path,
            path_in_repo=filename,
            repo_id=REPO_ID,
            repo_type="model"
        )
        print(f"Uploaded {filename}")

print("\n" + "="*60)
print("Example: Upload Dataset to Hugging Face")
print("="*60)

from datasets import Dataset, DatasetDict
import pandas as pd

data = {
    "text": [
        "Example text 1",
        "Example text 2",
        "Example text 3"
    ],
    "label": [0, 1, 0]
}

df = pd.DataFrame(data)
dataset = Dataset.from_pandas(df)

dataset_dict = DatasetDict({
    "train": dataset
})

DATASET_REPO_ID = "your-username/custom-dataset"

dataset_dict.push_to_hub(DATASET_REPO_ID)
print(f"Uploaded dataset to {DATASET_REPO_ID}")

print("\n" + "="*60)
print("All uploads complete!")
print("="*60)
print("\nNow in the demo app:")
print("1. Select 'Hugging Face' as SAE source")
print(f"2. Enter Repository ID: {REPO_ID}")
print("3. Enter Filename: sae_l16.pt")
print("4. Click 'Load from Hugging Face'")
print("\nFor custom datasets:")
print("1. Expand 'Advanced: Custom Dataset'")
print(f"2. Enter Dataset Repository ID: {DATASET_REPO_ID}")
print(f"3. Enter SAE Repository ID: {REPO_ID}")
print("4. Click 'Load Custom Configuration'")