glides/llama-eap
glides/llama-eap is an 8 billion parameter language model fine-tuned on the Edgar Allan Poe dataset. Developed by nroggendorff, it utilizes Supervised Fine-Tuning (SFT) and Teacher Reinforced Learning (TRL) techniques. This model is specifically designed for generating text in the distinctive style of Edgar Allan Poe, making it suitable for creative writing and stylistic imitation tasks. It supports English language generation with a context length of 32768 tokens.
Loading preview...
Edgar Allan Poe LLM (llama-eap)
This model, developed by nroggendorff, is an 8 billion parameter language model specifically fine-tuned to generate text in the style of Edgar Allan Poe. It leverages a specialized dataset (EAP dataset) and advanced training methodologies to achieve its unique stylistic capabilities.
Key Capabilities
- Stylistic Generation: Excels at producing text that mimics the distinctive literary style of Edgar Allan Poe.
- Advanced Fine-tuning: Utilizes both Supervised Fine-Tuning (SFT) and Teacher Reinforced Learning (TRL) for enhanced performance and stylistic adherence.
- English Language Support: Primarily designed for English text generation.
- Efficient Deployment: Can be loaded and used with quantization (e.g., 4-bit) via the Hugging Face Transformers library, making it accessible for various applications.
Good For
- Creative Writing: Ideal for generating poems, stories, or prose in the gothic and melancholic style of Poe.
- Literary Exploration: Useful for researchers or enthusiasts studying stylistic imitation in large language models.
- Themed Content Creation: Suitable for projects requiring text with a specific, classic literary tone.