Dan Luu on X: ""actually the seed is also a hyper-parameter" https://t.co/KJ9YxL92Gz" / X
To view keyboard shortcuts, press question mark View keyboard shortcuts Post See new posts Conversation Dan Luu @danluu Follow "actually the seed is also a hyper-parameter" ALT 4:11 PM · Aug 15, 2019 9 287 985 36 Post your reply Reply Gary Basin @garybasin · Aug 15, 2019 Lol @Meaningness 2 2 Kyle J Strand @BatmanAoD · Aug 15, 2019 Galaxy brain hot take: all software (i.e. the binary itself) is just a parameter to a Universal Turing Machine 1 29 Ian Soboroff | ian@idf.social @ian_soboroff · Aug 16, 2019 Also, sometimes the same seed does not yield the same random number stream over library versions, language versions, or hardware changes. 1 1 3 Erdem Ertan @Siluuet · Aug 16, 2019 Welcome to the world of NNs 2 Tony Laing @TILaing · Aug 16, 2019 WTF? Sohrab Salimian @SohrabSalimian · Aug 17, 2019 In the recent icml conference a paper discussed how the initialization of a network was critical in getting it to converge. I think this has to do with the input data distribution, these net
Dan Luu @danluu "actually the seed is also a hyper-parameter" 11:11 PM · Aug 15, 2019 7 0 7 205 0 2 0 5 902 0 9 0 2 34 0 3 4 Read 7 replies
Explore this link on the map →related reading
- Hyperparameter optimization - Wikipediaen.wikipedia.org
- GitHub - brexhq/prompt-engineering: Tips and tricks for working with Large Language Models like OpenAI's GPT-4. · GitHubgithub.com
- GitHub - google-research/tuning_playbook: A playbook for systematically maximizing the performance of deep learning models. · GitHubgithub.com
- GitHub - jacobhilton/deep_learning_curriculum: Language model alignment-focused deep learning curriculum · GitHubgithub.com
- The Practitioner’s Guide to the Maximal Update Parameterization - Cerebrascerebras.ai
- The Practitioner's Guide to the Maximal Update Parameterization | EleutherAI Blogblog.eleuther.ai
- jason (@jxnlco) / Xtwitter.com
- GitHub - google-research/tuning_playbook: A playbook for systematically maximizing the performance of deep learning models. · GitHubgithub.com
- Neuronpedianeuronpedia.org
- AlgZoo: uninterpreted models with fewer than 1,500 parameters — LessWronglesswrong.com
- GitHub - openai/parameter-golf: Train the smallest LM you can that fits in 16MB. Best model wins! · GitHubgithub.com
- Greg Yang | Professional pagethegregyang.com