4 ms·
Does anybody know how long it would take to train an alphazero go version using one gpu? In [1] they claim that it took 13 hours until the model was able to bea
by FiberBundle 6y ago
Does anybody know how long it would take to train an alphazero go version using one gpu? In [1] they claim that it took 13 hours until the model was able to beat the original alphago version, but they don't state what hardware they used.
[1] https://deepmind.com/blog/article/alphazero-shedding-new-light-grand-games-chess-shogi-and-go https://deepmind.com/blog/article/alphazero-shedding-new-lig...
- arijun 6y agoI can’t find it now but iirc there was a blog post on HN about a month ago that estimated their training costs at $25 million, using many TPU pods.
- cgreerrun 6y agoHere was the guestimation: https://www.yuzeh.com/data/agz-cost.html https://www.yuzeh.com/data/agz-cost.html
- newswasboring 6y agoFrom an offline chat with the original author, The ELF OpenGo paper[1], which is an open implementation of AlphaGo Zero developed by Facebook AI: "First, we train a superhuman model for ELF OpenGo. Af-ter running our AlphaZero-style training software on 2,000GPUs for 9 days, our 20-block model has achieved super-human performance that is arguably comparable to the 20-block models described in Silver et al. (2017) and Silveret al. (2018)." [1]: https://arxiv.org/pdf/1902.04522.pdf https://arxiv.org/pdf/1902.04522.pdf
- jonath_laurent 6y agoI agree with the quoted numbers. As I mentioned in another comment, you have to keep in mind that AlphaZero is an extremely sample-inefficient learning technique, even for simple problems. However, it has two major strengths: 1) it is pretty generic and 2) it can leverage huge amounts of computing power.
- klipt 6y agoWhat would be an example of a more sample efficient algorithm?
- wrsh07 6y agoThat was with at least one or more tpu pods, iirc https://cloud.google.com/tpu/docs/system-architecture https://cloud.google.com/tpu/docs/system-architecture