TheBloke/koala-13B-HF
TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:Apr 7, 2023License:otherArchitecture:Transformer0.0K Cold
TheBloke/koala-13B-HF is a 13 billion parameter dialogue model developed by Berkeley, based on the Llama architecture. This model is specifically designed for academic research in dialogue systems. It was created by merging Koala delta weights with the original Llama 13B model, offering capabilities for conversational AI research within its 4096-token context window.
Loading preview...
Koala-13B-HF: A Dialogue Model for Academic Research
This model, koala-13B-HF, is a 13 billion parameter dialogue model developed by Berkeley, built upon the Llama 13B architecture. It was created by applying Koala delta weights to the original Llama model, specifically for academic research purposes.
Key Capabilities
- Dialogue Generation: Optimized for generating conversational responses.
- Research Focus: Intended for academic exploration and development in the field of conversational AI.
- Llama-based: Leverages the foundational capabilities of the Llama 13B model.
Good For
- Academic Research: Ideal for researchers studying dialogue systems, conversational agents, and large language model fine-tuning.
- Experimentation: Suitable for experimenting with dialogue model architectures and training methodologies.
Important Notes
- License: The model weights are strictly for academic research only, subject to the LLaMA model license and terms of use for data generated by OpenAI and ShareGPT. Commercial usage is prohibited.
- Context Window: Features a 4096-token context length, suitable for moderate-length dialogues.