{"id":"arxiv:1506.01186","title":"Cyclical Learning Rates for Training Neural Networks","date":"2015-06-03","section":"Data, training, optimization","quality":-0.406646,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/","source_url":"https://arxiv.org/abs/1506.01186","text_available":true,"authors":["Leslie N. Smith"],"outgoing":[],"incoming":[{"source":"arxiv:1708.07120","target":"arxiv:1506.01186","evidence":"id","id":"arxiv:1708.07120","title":"Super-Convergence: Very Fast Training of Neural Networks Using Large Learning Rates","url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/604e53f71d7e9128110137811c408ab3b0c041d945f88d001f5095edbfb1a875/"},{"source":"arxiv:2205.13147","target":"arxiv:1506.01186","evidence":"title","id":"arxiv:2205.13147","title":"Matryoshka Representation Learning","url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/803baa0a8ea7ba1bc63bee0ce80006bf4f403ca0db16ada95dd8198da9a4a899/"}],"text_parts":[{"start":0,"end":15809,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html"},{"start":15809,"end":31433,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-2.html"},{"start":31433,"end":33091,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-3.html"}],"contents":[{"title":"Cyclical Learning Rates for Training Neural Networks","offset":0,"level":1,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"Abstract","offset":56,"level":6,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"1 Introduction","offset":1139,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"2 Related work","offset":3924,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"3 Optimal Learning Rates","offset":7061,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"3.1 Cyclical Learning Rates","offset":7090,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"3.2 How can one estimate a good value for the cycle length?","offset":10699,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"3.3 How can one estimate reasonable minimum and maximum boundary values?","offset":12230,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"4 Experiments","offset":14565,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"4.1 CIFAR-10 and CIFAR-100","offset":14930,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"4.1.1 Caffe’s CIFAR-10 architecture","offset":14962,"level":4,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-1.html#text"},{"title":"4.1.2 ResNets, Stochastic Depth, and DenseNets","offset":19627,"level":4,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-2.html#text"},{"title":"4.2 ImageNet","offset":22867,"level":3,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-2.html#text"},{"title":"4.2.1 AlexNet","offset":23167,"level":4,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-2.html#text"},{"title":"4.2.2 GoogLeNet/Inception Architecture","offset":26864,"level":4,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-2.html#text"},{"title":"5 Conclusions","offset":30392,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-2.html#text"},{"title":"Appendix A Instructions for adding CLR to Caffe","offset":31810,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-3.html#text"},{"title":"Appendix B Instructions for adding CLR to Keras","offset":32028,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-3.html#text"},{"title":"Instructions for reporting errors","offset":32127,"level":2,"url":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/part-3.html#text"}],"text_code_points":33091,"source_markdown":"https://vinnik-dmitry07.github.io/favorite-papers/agents/papers/e05cf550191656c287184521d0ff011092f5ed6baa72e50f3a5bfd25d0f5986a/source.md"}