-
Due to this line of code,
https://github.com/HazyResearch/sampler/blob/master/src/app/gibbs/single_thread_sampler.cpp#L45.
non-evidence values of variables stay fixed during the sampling phase. This…
-
I’m currently implementing a Dirichlet process model in PyMC 3 and got the stick-breaking process to
work using the new theano decorator `as_op`:
``` python
@theano.compile.ops.as_op(itypes=[t.dv…
-
`logging` messages do not show up in IPython notebook.
At minimum, a brief "Monitoring the sampler" section of the documentation could describe the problem.
Alternatively, provide a traceplot of the…
-
There doesn't seem to be a way of advancing the state of RNGs. Without that, I don't see any way of sampling more than a single random number from a distribution for a given seed.
-
I have not scrutinised this properly, but the way random state is used in the `BernoulliRBM` looks dubious. If `random_state` is set with an integer, `check_random_state` will return a new random gene…
-
It seems to be the case that plotting large models has a somewhat odd behavior. For example, plotting the results of [this example](http://mambajl.readthedocs.org/en/latest/examples/seeds.html) outpu…
ghost updated
9 years ago
-
PaPi does generate SNP profiles as a report now, but there are some assumptions in the code to make sure the impact of noise is minimal.
Although it works pretty awesome for most cases, everything sh…
meren updated
9 years ago
-
We have several reports of high variance in predictions and learned weights between multiple runs (observed by Toshiba, 341 team, and myself).
We want to fully understand what is causing this and how…
-
John,
I've been having some problems debugging memory allocation with BIDMach (BayesNet.scala specifically). I'm using the Java Runtime class which should be a reliable measure of memory allocation. …
-
It should probably take an `n_iter` as an argument.