Feature Extraction
Transformers
Safetensors
PyTorch
English
eden
text-enhancement
grammar-correction
text-rewriting
encoder-decoder
transformer
custom_code
Instructions to use Rybib/EDEN with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Rybib/EDEN with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="Rybib/EDEN", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Rybib/EDEN", trust_remote_code=True, dtype="auto") - Notebooks
- Google Colab
- Kaggle
| """Talk to EDEN in the terminal, similar to how Ollama works. | |
| This downloads the published model from the Hugging Face Hub the first time it | |
| runs and caches it. After that it works offline. | |
| Usage: | |
| python3 try_eden.py # open the chat interface | |
| python3 try_eden.py "some rough text" # one-shot: clean the given text | |
| """ | |
| import sys | |
| MODEL_ID = "Rybib/EDEN" | |
| # ANSI styles, used only when writing to a real terminal. | |
| _TTY = sys.stdout.isatty() | |
| BOLD = "\033[1m" if _TTY else "" | |
| DIM = "\033[2m" if _TTY else "" | |
| GREEN = "\033[32m" if _TTY else "" | |
| CYAN = "\033[36m" if _TTY else "" | |
| RESET = "\033[0m" if _TTY else "" | |
| BANNER = f"""{CYAN}{BOLD} | |
| EDEN :: Encoder Decoder Enhancement Network | |
| {RESET}{DIM} Type or paste rough text and press Enter to clean it up. | |
| Commands: /help show help /bye quit (Ctrl+D also quits) | |
| {RESET}""" | |
| HELP = f"""{DIM} | |
| Just type or paste text, then press Enter, and EDEN rewrites it. | |
| Commands: | |
| /help show this help | |
| /bye quit (so do /exit, /quit, and Ctrl+D) | |
| {RESET}""" | |
| def load_model(): | |
| try: | |
| from transformers import AutoModel, AutoTokenizer | |
| except ImportError: | |
| print("Missing dependencies. Run this first:") | |
| print(" pip3 install torch transformers") | |
| sys.exit(1) | |
| print(f"{DIM}Loading {MODEL_ID} (first run downloads about 430 MB) ...{RESET}") | |
| tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True) | |
| model = AutoModel.from_pretrained(MODEL_ID, trust_remote_code=True).eval() | |
| return model, tokenizer | |
| def main() -> None: | |
| model, tokenizer = load_model() | |
| # One-shot mode: clean the text passed as arguments and exit. | |
| args = [a for a in sys.argv[1:] if a.strip()] | |
| if args: | |
| print(model.enhance(tokenizer, " ".join(args))) | |
| return | |
| # Interactive chat mode. | |
| print(BANNER) | |
| while True: | |
| try: | |
| text = input(f"{GREEN}{BOLD}>>> {RESET}").strip() | |
| except (EOFError, KeyboardInterrupt): | |
| print(f"\n{DIM}Goodbye.{RESET}") | |
| return | |
| if not text: | |
| continue | |
| if text.lower() in {"/bye", "/exit", "/quit", "/q"}: | |
| print(f"{DIM}Goodbye.{RESET}") | |
| return | |
| if text.lower() in {"/help", "/h", "/?"}: | |
| print(HELP) | |
| continue | |
| cleaned = model.enhance(tokenizer, text) | |
| print(f"{CYAN}{cleaned}{RESET}\n") | |
| if __name__ == "__main__": | |
| main() | |