Skip to content

This repository contains alpaca-formatted datasets for creating LoRA (Low Rank Adaptation) models for review analysis. The datasets have been enriched with sentiment analysis and keyword extraction information, in addition to review data.

ayseguldmrblk/Reviews-Datasets

Repository files navigation

Llama LORA Repository

This repository contains datasets and resources for training and fine-tuning AI models for llama-related text generation. We use TextGenWebUI, Hugging Face's Transformers library, and the LoRA (Low Rank Adaptation) fine-tuning technique.

Datasets

reviews_large.json

  • This dataset contains a large collection of llama-related text data. It's designed to be used for training large AI models, enabling them to generate llama-themed content with a broad vocabulary.

reviews_small.json

  • This dataset is a smaller subset of reviews_large.json. It's suitable for quick experimentation and testing of models. Use this dataset if you want to iterate rapidly during development.

reviews_large_alpaca.json

  • This dataset is an extension of reviews_large.json and includes additional data related to alpacas, which can be useful for creating AI models that generate content about both llamas and alpacas.

reviews_small_alpaca.json

  • Similar to reviews_small.json, this dataset is a compact version of reviews_large_alpaca.json. It's ideal for quick prototyping and experimenting with models that incorporate both llama and alpaca content.

Additional Resources

  • For more information on TextGenWebUI, please visit their official documentation.

  • To explore Hugging Face's Transformers library and discover pre-trained models, check out their GitHub repository.

About

This repository contains alpaca-formatted datasets for creating LoRA (Low Rank Adaptation) models for review analysis. The datasets have been enriched with sentiment analysis and keyword extraction information, in addition to review data.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published