CarperAI / squeakily
A library for squeakily cleaning and filtering language datasets.
☆45Updated last year
Alternatives and similar repositories for squeakily:
Users that are interested in squeakily are comparing it to the libraries listed below
- This repository contains code for cleaning your training data of benchmark data to help combat data snooping.☆25Updated last year
- ☆24Updated last year
- ☆22Updated last year
- Experiments with generating opensource language model assistants☆97Updated last year
- QAmeleon introduces synthetic multilingual QA data using PaLM, a 540B large language model. This dataset was generated by prompt tuning P…☆34Updated last year
- Repository containing the SPIN experiments on the DIBT 10k ranked prompts☆24Updated 10 months ago
- Using short models to classify long texts☆21Updated last year
- ☆46Updated 2 months ago
- a pipeline for using api calls to agnostically convert unstructured data into structured training data☆29Updated 3 months ago
- Demonstration that finetuning RoPE model on larger sequences than the pre-trained model adapts the model context limit☆63Updated last year
- QLoRA with Enhanced Multi GPU Support☆36Updated last year
- Exploring finetuning public checkpoints on filter 8K sequences on Pile☆115Updated last year
- ☆18Updated last year
- ☆32Updated last year
- QLoRA for Masked Language Modeling☆21Updated last year
- PyTorch implementation for MRL☆18Updated 10 months ago
- Genalog is an open source, cross-platform python package allowing generation of synthetic document images with custom degradations and te…☆42Updated 11 months ago
- Comprehensive analysis of difference in performance of QLora, Lora, and Full Finetunes.☆82Updated last year
- ☆48Updated last year
- ☆20Updated last year
- Using open source LLMs to build synthetic datasets for direct preference optimization☆49Updated 10 months ago
- Repo for training MLMs, CLMs, or T5-type models on the OLM pretraining data, but it should work with any hugging face text dataset.☆93Updated last year
- Multi-Domain Expert Learning☆67Updated 11 months ago
- SWIM-IR is a Synthetic Wikipedia-based Multilingual Information Retrieval training set with 28 million query-passage pairs spanning 33 la…☆45Updated last year
- Training and Inference Notebooks for the RedPajama (OpenLlama) models☆18Updated last year
- ☆37Updated last year