William Ermlick's starred repositories
RWKV-LM
RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding.
PokemonRedExperiments
Playing Pokemon Red with Reinforcement Learning
trackerjacker
Like nmap for mapping wifi networks you're not connected to, plus device tracking
FastAPI-template
Feature rich robust FastAPI template.
instructor-embedding
[ACL 2023] One Embedder, Any Task: Instruction-Finetuned Text Embeddings
nn-template
Generic template to bootstrap your PyTorch project.
LetsMapYourNetwork
Lets Map Your Network enables you to visualise your physical network in form of graph with zero manual error
Robinhood-1099-Parser
Parse Robinhood 1099 Tax Document from PDF into CSV
d3js-neo4j-example
Some of D3.js v5 example pages visualize the result from Neo4j
NetworkMap
post-exploitation network mapper
py-freerdp
Basic Python wrapper for FreeRDP
yolov8-pidray
PIDray dataset detection using YoloV8 (compared with YoloV5)
datapipeline-TCP-UDP-neo4j
This project is about building a data pipeline to move network traces from a dataset to a Neo4J data store using Python. Each trace over TCP and UDP as well as their subtraces (ARP and ICMP) need to be modelized and stored in Neo4J. The purpose is to visualize each sender and receiver as Nodes, as well as the type of the protocol used as a relationship.