wassname
If you can't possibly remember what something is called, what it is, what it does, or why it does it, it instantly becomes a 'wassname'
I'm just a guy who likes to machine learnPerth, Australia
Pinned Repositories
makehuman-js
A library to build 3D human characters in the browser
attentive-neural-processes
implementing "recurrent attentive neural processes" to forecast power usage (w. LSTM baseline, MCDropout)
keywordshitter2
A website to find long-tail keywords using search suggestions
LoRA_are_lie_detectors
Experiment to see if low rank adapters can work as interventions for lie detection on LLM's
rl-portfolio-management
Attempting to replicate "A Deep Reinforcement Learning Framework for the Financial Portfolio Management Problem" https://arxiv.org/abs/1706.10059 (and an openai gym environment)
rl_2d_walker.js
Teaching a humanoid to walk(ish), then displaying in your browser (using tensorflow.js and reinforcement learning)
viz_torch_optim
Videos of deep learning optimizers moving on 3D problem-landscapes
world-models-sonic-pytorch
Attempt at reinforcement learning with curiosity for Sonic the Hedgehog games. Number 149 on OpenAI retro contest leaderboard, but more work needed
wassname's Repositories
wassname/attentive-neural-processes
implementing "recurrent attentive neural processes" to forecast power usage (w. LSTM baseline, MCDropout)
wassname/seq2seq-time
Bechmarking seq2seq models on a range of multivariate regression datasets
wassname/awesome-rlhf
Lists of datasets, training, and evals for RLHF and similar
wassname/prob_jsonformer
A Bulletproof Way to Generate Structured JSON from Language Models
wassname/quiet-star
Code for Quiet-STaR
wassname/cookiecutter-data-science
A logical, reasonably standardized, but flexible project structure for doing and sharing data science work.
wassname/detect_bs_text
Can we measure how good a text is by how much an LLM learns from it?
wassname/discovering_latent_knowledge
wassname/LoRA_are_lie_detectors
Experiment to see if low rank adapters can work as interventions for lie detection on LLM's
wassname/iris_bigvae
experiment: IRIS but with pretrained LLM
wassname/Alpaca-LoRA-Serve
Alpaca-LoRA as Chatbot service
wassname/alpaca_convert
wassname/awesome-interpretability
Awesome tools for interpreting, manipulating the internals of of deep neural networks.
wassname/baukit
wassname/copilot-gpt4-service
Convert Github Copilot to ChatGPT
wassname/Craftax
(Crafter + NetHack) in JAX
wassname/DebateTree
A langchain app to visualise a debate using Tree-of-Thought reasoning
wassname/epub2audio_xtts
wassname/geopandas
Python tools for geographic data
wassname/jaxtyping
Type annotations and runtime checking for shape and dtype of JAX/NumPy/PyTorch/etc. arrays. https://docs.kidger.site/jaxtyping/
wassname/leaked-system-prompts
Collection of leaked system prompts
wassname/peft
🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.
wassname/Ranger21
Ranger deep learning optimizer rewrite to use newest components
wassname/REM
Implementation of our paper "Improving Token-Based World Models with Parallel Observation Prediction"
wassname/representation-engineering
Representation Engineering: A Top-Down Approach to AI Transparency
wassname/side-by-side
Visual comparison of different translations of itemized texts; e.g. poems, bibles, etc.
wassname/stampy-chat
Conversational chatbot to answer questions about AI Safety & Alignment based on information retrieved from the Alignment Research Dataset
wassname/TTS
🐸💬 - a deep learning toolkit for Text-to-Speech, battle-tested in research and production
wassname/twm
Transformer-based World Models
wassname/word_level_diff_writing_assistant
Proof reading with an AI is easier with tracked changes