-
EfficientNet
https://ai.googleblog.com/2019/05/efficientnet-improving-accuracy-and.html
https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet
-
I noticed a new network (fe3f6...) in http://zero.sjeng.org/networks/
Can you provide some information about win rate over the previous best network?
Number of games it was trained on? How long it…
-
Thank you for your excellent work on the paper.
I encountered a CUDA out-of-memory error with a batch size of 2 on an A100 GPU. I noticed in Table 3 that the GPU memory allocated is significantly lar…
-
as a benchmark for test new weights, such as minigo convert
-
Running the command for the mnist dataset
./run.py train --encoder-layers 1000-500-250-250-250-10 --decoder-spec sig --denoising-cost-x 1000,10,0.1,0.1,0.1,0.1,0.1 --labeled-samples 100 --unlabeled-s…
-
Hi all,
there's 500k games now without a network promoting. This means the training window is "full". I increased it to 750k games, and did a last learning rate lowering (0.00001 @ bs=96). If this …
-
What we can learn from #1113, IMHO, is that a good net still buys us a lot of strength even with the same data. However, a good net is expensive in terms of compute power, and hence I was thinking if…
-
So I've been botting for a while and it seems like my bot usually stops the game being made then restarting.
Sometimes the bot will be successful in making the game but then 10+ more waiting in lin…
-
Now that there is pandoc functionality for the site through the `mkdocs-bibtex` package [see issue #1], the current citations will need updating.
As with standard pandoc citations, this can be don…
-
Some powerful contributors recently started helping this project. While we thank and kudo them for their resources, some changes are needed for the training pipeline, as the AlphaZero pipeline require…