David Beniaguev

SFHQ-dataset. Synthetic Faces High Quality (SFHQ) Dataset. 425,000 curated 1024x1024 synthetic face images

neuron_as_deep_net. Code behind the work "Single Cortical Neurons as Deep Artificial Neural Networks", published in Neuron 2021

visual_taste_approximator. Visual Taste Approximator (VTA) is a very simple tool that helps anyone create an automatic replica of themselves that can approximate their own personal visual taste

SFHQ-T2I-dataset. Synthetic Faces High Quality - Text2Image (SFHQ-T2I) Dataset. 122,726 curated 1024x1024 synthetic face images

filter_and_fire_neuron. Code behind the work "Multiple Synaptic Contacts Combined with Dendritic Filtering Enhance Spatio-Temporal Pattern Recognition of Single Neurons", bioRxiv 2022

repo_to_text. convert github repos or local folders into a single text document that can be used for LLM training, RAG or copy pasting into context while developing

a_chatgpt_never_forgets. A very simple implementation of a chat-bot that remembers everything you ever talked about and possesses rudimentary theory of mind

tutorial_NEURON_pytorch. code for simulating ball and stick models using NEURON, generating simulation datasets, and training various deep networks on the data

my_kaggle_notebooks. Repo with my most popular kaggle notebooks. I've put a lot of effort into them back in the day, so they are highly curated and well documented.

min_local_mongodb. A lightweight implementation of MongoDB-like storage on a local computer without dependencies. provides an identical interface to pymongo

ImSME-dataset. ImSME dataset - Images of Simple Math Equations. 262,144 wide images containing simple math equations

LMMP-protocol. LMMP - a multimodal protocol for the language model era. implements image2string() and string2image() methods

min_local_s3. A lightweight implementation of Amazon S3-like storage on a local computer with no dependencies. Provides an identical interface to boto3's S3 client