Luckily, years of Deep Learning research have honed my ability to sit in front of a graph and refresh every few seconds.
Alex Nichol
7,098 posts
Alex Nichol
@unixpickle
Code, AI, and 3D printing. Opinions are mostly my own, sometimes a computer's. Husband of @thesamnichol. Co-creator of DALL-E 2. Researcher @openai.
Joined September 2011
- Super excited to release Shap-E, our latest work on 3D generative modeling. Paper: arxiv.org/abs/2305.02463 Code/models: github.com/openai/shap-e
GIF - Happy to share a cool hobby project I've been working on for the past few weeks: predicting attributes of cars from images. Blog post: blog.aqnichol.com/2022/12/31/lar… Github: github.com/unixpickle/car… Gradio demo: huggingface.co/spaces/unixpic…
- GPT-4o is a huge step forward for image generation. Not only is it amazing at rendering text and following captions, it also provides a very natural way to iteratively edit and compose visual concepts. 1/8
- Kinda offended that meta didn't try to recruit me 😂
- There is a new 3D generative model in town: Genie by @LumaLabsAI. This model is surprisingly good and insanely fast! As someone who has worked on 3D generative modeling, here are my observations and guesses about how this system works (without any insider info):
GIF - OpenAI is nothing without its people.
- Super excited to share our work on DALL-E 2 pre-training mitigations!
- Now anybody can tinker with and explore the current SOTA generative image models! We have released model weights for our paper "Diffusion Models Beat GANs on Image Synthesis":
- Agent finally learning to beat the 10th floor! A few more days of training and I'll probably be ready to submit this. Whole thing is rewritten in PyTorch now, no more anyrl-py or TensorFlow.
00:00 - Vacation project: convert trained scikit-learn models into Torch modules.
- I'm still not convinced that large-scale RL is anything other than "smoothed-out search distilled into a small model". 1/5
- My theory for why Python is so popular for deep learning: Python is so slow that the CPU basically looks like a GPU--you need vec ops or special kernels just to get good perf. Python devs were used to this from the start (e.g. numpy) so GPU frameworks felt natural.
