thebabellibrarybot Goto Github PK

autoencoder-scribes

autoencoder and dataloaders for running anomaly detection in medieval scribes version of the MNIST dataset

babelanno-test

Webapp for annotating manuscripts with a paleographic approach to HOCR ground truth data-labeling

babelhistoricannotator

An annotator with advanced labeling, export and versioning tools to facilitate a paleographer's approach to ground truth creation for HOCR model training of manuscripts

bablib

background_funcs

bokted_backend

backend for booking with prod and dev branches

bokted_frontend

booking app frontend with prod and dev branches

booked

a booking and book keeping application for tattoo artists

bookedoauth

booked

casestudy

masterworks case-study with real-estate DB

gitbot-docker

dockerized project that scraps your github repos and helps you locate exact files and lines you may be confused about.

gitbot_shell

just the frontend for interaction and viewing purposes

honeyhive

API to make prompt templates and generate prompt completions for gpt3, vincua, and cohere

jackandadam

crud

keras-dreambooth-i

fine-tune stable diffusion to learn class-instance of sks pixel-art-character

kraken-bot-api

kraken, boto, FROM public.ecr.aws/lambda/python:3.8

kraken_mk_text_lines

crops original image to bbox dims found with kraken api. generates images per text line from original image metadata

libcurator

allows researchers to make collections for their research topics and review collection statistics quickly

manuscript-segmentation

Semantic segmentation of elements in a manuscript page. Includes code for training on custom annotations. Segmentation of text, image, noise, and marginilia trained on dataset curated with help from the Morgan library. NN built of deeplabv3 transfer learning using resnet50 pretrained weights.

manuscript_text_segmentation

NN designed to identify and make bounding boxes around section of a page that includes text, of course for manuscripts tho (13-15c BOH). Crops to furthers corners of bounding boxes and applies a mask over the access non text space. **Should be followed up with a similar version used to find noise // badlines and mask over that info aka prepping pages for line segmentation.

morganapi

dockerized api and SQL database to look at records from the morgan libraries archives. data visualization views are linked in this repo.

pack2

artwork packing for storage spaces

prep_fasterrcnn_anno

This is a quick tool for creating image masks used for training a FasterRCNN object detection module. Object detection takes several inputs including: image data, bbox dims, object labels, and image data masks. In order to quickly prepare all this data with one tool, this script will automatically create masks from a simple instance of bbox annotations created with VGG image annotator.

puppets-morgan-library

thebabellibrarybot

hello world

visuals

app for looking at houisng data

wand_dhseg_playground

a playground for making a llm with langchain that will either write the integration itself or acuratly id and req methods to integrate any repo, in this case dhSegment, with wandb for training and logging.

yolo-tensor-14latin-bboxes-api

api req to aws lambda function with sage-maker model trained on Latin 14th c. manuscript microfilms.

yolo_dims_edit_img_api

takes dims found with yolo sagemaker model now stored in mongoDB and generates new image_versions via coords