TheBloke/PMC_LLAMA-7B-10-Epoch-SuperHOT-8K-fp16
TheBloke/PMC_LLAMA-7B-10-Epoch-SuperHOT-8K-fp16 is a 7 billion parameter LLaMA-based model, developed by Chaoyi Wu and merged with Kaio Ken's SuperHOT 8K LoRA. This fp16 PyTorch model is specifically designed for GPU inference and features an extended context window of 8192 tokens. It is fine-tuned on PMC papers from the S2ORC dataset, making it suitable for tasks requiring long-context understanding, particularly in scientific or medical domains, with an added focus on NSFW content from the SuperHOT merge.
Loading preview...
Model Overview
This model, PMC_LLAMA-7B-10-Epoch-SuperHOT-8K-fp16, is a 7 billion parameter LLaMA-based model. It is a merge of Chaoyi Wu's PMC_LLAMA_7B_10_epoch, which was fine-tuned on PMC papers from the S2ORC dataset, and Kaio Ken's SuperHOT 8K LoRA. The primary differentiator is its significantly extended context window of 8192 tokens, achieved through the SuperHOT merge and specific configuration settings.
Key Capabilities
- Extended Context: Supports an 8192-token context length, enabling processing of longer texts and more complex queries.
- Scientific Domain Focus: The base PMC_LLAMA model was trained for 10 epochs on scientific papers, suggesting proficiency in medical and scientific text understanding.
- NSFW Content Generation: The integration of the SuperHOT LoRA indicates an optimization for generating NSFW-focused content.
- FP16 Format: Provided in fp16 PyTorch format, suitable for GPU inference and further conversions.
Good For
- Long-Context Applications: Ideal for tasks requiring extensive contextual understanding, such as summarizing long documents or complex conversations.
- Scientific/Medical Text Processing: Leveraging its PMC_LLAMA base, it can be applied to tasks involving scientific literature.
- NSFW Content Creation: Specifically tuned for generating content in this domain due to the SuperHOT merge.
- Developers: Suitable for those looking for a base fp16 model for further fine-tuning or custom conversions.