Seth Troisi
Seth Troisi
This is work against https://github.com/tensorflow/minigo/blob/master/oneoffs/training_curve.py The pickle file is used in the [Colab Figure 3 Investigation](https://colab.research.google.com/drive/1kfzXQh2oG4Fne47bcD-tyTOz8mdvDlFl#scrollTo=nyCr1p6XMaBL)
Makes ```echo "200000004934000028207" | $PM1 -x0 2 1000000000-1000000010 100000000000-100000002000; checkcode $? 8``` and ```echo 154618728587 | $PP1 -x0 3 -go 36 4294957296-4294967295 1; checkcode $? 8``` instant
python3
Re-opening the discussion from https://github.com/mattheww/gomill/issues/1, [Python 2 is now dreadfully close to EOL](https://pythonclock.org/). 1. Would you be supportive of merging a port of the remaining code into sgfmill? 1. Do...
I was trying to understand which of the many `is_X_prp` to use and I benchmarked all the prp functions in [colab](https://colab.research.google.com/drive/1i5Fr0Jt9IZlVGDAreRn94g4tZmhZ4YqK?authuser=0#scrollTo=yzoqALVPxbP8). I'm not sure if this is helpful for other...
This builds on @brilee's https://github.com/jtauber/gtp/pull/7 by updating and adding tests.
# Seth ideas * Virtual Batching ideas (re eval effects of virtual loss https://github.com/tensorflow/minigo/issues/427) * Only add X from batch of Y to the tree, put the rest in NNCache...
Some comments from offline discussion between Andrew & Seth * Launch things via scripts & flag files * record CLI & flag files into a directories * model table *...
Building on https://github.com/tensorflow/minigo/pull/896 in conjunction with https://github.com/tensorflow/minigo/issues/898
We have a number of metrics that measure a bunch of things * metrics about the model weights * l2_cost * model accuracy on constant data * value_cost on pro...