YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

license: mit
language:
- en

AgGPT-20 Banner

AgGPT-20

AgGPT-20 is a revolutionary modular AI training framework using Scalable Feather Architecture that automatically discovers and trains on all .txt files in training_corpora/, creating specialized mini-models stored as .feather files in the models/ directory. The system is extremely modular - you can train individual models using target_train.py, manually add .feather files to the models folder from other sources, or use the full multi-corpora trainer with train.py, and all models will be automatically discovered and used by the chat interface. Install dependencies with pip install pandas pyarrow tqdm pyyaml, place training data in the proper format (user: [input]\n<pad>\nai: [response]\n<eos>) in training_corpora/ or extra_corpora/, run python train.py for full training or python target_train.py for single-file training, then chat with python chat.py. The architecture enables lightning-fast loading, cross-platform compatibility, and seamless model distribution - just drop any compatible .feather model into the models folder and it's instantly available.

Quick Start

# Install dependencies
pip install pandas pyarrow tqdm pyyaml

# Train on all corpora files
python train.py

# Or train on a single target file
python target_train.py

# Start chatting
python chat.py
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support