profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/patrl/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.
Patrick Elliott patrl @mit Cambridge https://patrickdelliott.com Theoretical Linguist

patrl/.doom.d 6

Private module for doom emacs

patrl/getting-high-seminar 5

Materials for 24.979 - "Getting high: scope, projection, and evaluation order"

patrl/cont-playground 2

A menagerie of sketches and speculations using continuations (WiP)

patrl/exhMonad 2

Computing exhaustification in haskell

josiet-j/josiet-j.github.io 1

Josie Thaddeus-John's website

patrl/.dots 1

Declarative dots managed via nix and home-manager.

patrl/24.954 1

MIT pragmatics in linguistic theory course materials

patrl/.emacs.d 0

My personal emacs settings.

patrl/agda-nix 0

Some boilerplate for using agda with nix

startedMaximeHerpin/modular_tree

started time in a day

issue closedabrsvn/pyactr-book

reference in readme need medium paid account registration

reference in readme need medium paid account registration

closed time in a day

psigelo

issue closedabrsvn/pyactr-book

Chp 7, Section 2 Lexical Decision model fails during sampling

Hi, I'm trying to fit the lexical decision model in Chapter 7, section 2, and I get an error during inference:

<ipython-input-23-0006ec3856d7>:33: FutureWarning: In v4.0, pm.sample will return an `arviz.InferenceData` object instead of a `MultiTrace` by default. You can pass return_inferencedata=True or return_inferencedata=False to be safe and silence this warning.
  trace = pm.sample(draws=10000, n_init=200000, cores=6, progressbar=True)
Auto-assigning NUTS sampler...
Initializing NUTS using jitter+adapt_diag...
Multiprocess sampling (6 chains in 6 jobs)
NUTS: [threshold, noise, latency_factor, intercept, decay]
<IPython.core.display.HTML object>

C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\hmc\quadpotential.py:224: RuntimeWarning: divide by zero encountered in true_divide
  np.divide(1, self._stds, out=self._inv_stds)
C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\hmc\quadpotential.py:203: RuntimeWarning: invalid value encountered in multiply
  return np.multiply(self._var, x, out=out)
C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\hmc\quadpotential.py:224: RuntimeWarning: divide by zero encountered in true_divide
  np.divide(1, self._stds, out=self._inv_stds)
C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\hmc\quadpotential.py:203: RuntimeWarning: invalid value encountered in multiply
  return np.multiply(self._var, x, out=out)
C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\hmc\quadpotential.py:224: RuntimeWarning: divide by zero encountered in true_divide
  np.divide(1, self._stds, out=self._inv_stds)
C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\hmc\quadpotential.py:203: RuntimeWarning: invalid value encountered in multiply
  return np.multiply(self._var, x, out=out)
RemoteTraceback: 
"""
Traceback (most recent call last):
  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\parallel_sampling.py", line 137, in run
    self._start_loop()
  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\parallel_sampling.py", line 191, in _start_loop
    point, stats = self._compute_point()
  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\parallel_sampling.py", line 216, in _compute_point
    point, stats = self._step_method.step(self._point)
  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\arraystep.py", line 276, in step
    apoint, stats = self.astep(array)
  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\hmc\base_hmc.py", line 147, in astep
    self.potential.raise_ok(self._logp_dlogp_func._ordering.vmap)
  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\step_methods\hmc\quadpotential.py", line 272, in raise_ok
    raise ValueError("\n".join(errmsg))
ValueError: Mass matrix contains zeros on the diagonal. 
The derivative of RV `noise_interval__`.ravel()[0] is zero.
The derivative of RV `intercept_interval__`.ravel()[0] is zero.
"""


The above exception was the direct cause of the following exception:

ValueError: Mass matrix contains zeros on the diagonal. 
The derivative of RV `noise_interval__`.ravel()[0] is zero.
The derivative of RV `intercept_interval__`.ravel()[0] is zero.


The above exception was the direct cause of the following exception:

Traceback (most recent call last):

  File "<ipython-input-23-0006ec3856d7>", line 33, in <module>
    trace = pm.sample(draws=10000, n_init=200000, cores=6, progressbar=True)

  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\sampling.py", line 559, in sample
    trace = _mp_sample(**sample_args, **parallel_args)

  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\sampling.py", line 1477, in _mp_sample
    for draw in sampler:

  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\parallel_sampling.py", line 479, in __iter__
    draw = ProcessAdapter.recv_draw(self._active)

  File "C:\Users\druss\Anaconda3\envs\mypm3env\lib\site-packages\pymc3\parallel_sampling.py", line 359, in recv_draw
    raise error from old_error

RuntimeError: Chain 1 failed.

The only thing I've changed in the code is changing n_jobs to cores, since I'm using a different version of pymc3. Otherwise, I have left your script as is. Also, the first model (the log frequency model) fit just fine.

Have you seen this before? Any idea what the issue is?

closed time in a day

drussellmrichie

issue commentabrsvn/pyactr-book

Chp 7, Section 2 Lexical Decision model fails during sampling

Closing this since the issue is resolved by using the SMC sampler.

drussellmrichie

comment created time in a day

issue closedabrsvn/pyactr-book

Confusing rule name "reanalyse: subject wh"

Correctly me if I wrong - in pyactr-book/book-code/ch7_grodner_gibson/parser_rules.py the purpose of the rule "reanalyse: subject wh" is to revise the initial prediction of subject gapping to a prediction of object gapping. Wouldn't "reanalyse: object wh" be a better name in that case? It confused me when I first saw that rule name. I thought "reanalyse: subject wh" should still predict subject gapping.

closed time in a day

snoions

issue commentabrsvn/pyactr-book

Confusing rule name "reanalyse: subject wh"

Thank you for this. I guess we were thinking about what we were reanalyzing (the initial prediction of subject gap) rather than what we were reanalyzing towards (a non-subject gap). I can't really say which one would be less confusing, but opening this issue here will help either way. Thanks again.

snoions

comment created time in a day

issue openedabrsvn/pyactr-book

Confusing rule name "reanalyse: subject wh"

Correctly me if I wrong - in pyactr-book/book-code/ch7_grodner_gibson/parser_rules.py the purpose of the rule "reanalyse: subject wh" is to revise the initial prediction of subject gapping to a prediction of object gapping. Wouldn't "reanalyse: object wh" be a better name in that case? It confused me when I first saw that rule name, because I thought "reanalyse: subject wh" should still predict subject gapping.

created time in a day

fork mixmix/hyperdrive

Hyperdrive is a secure, real time distributed file system

fork in 3 days

startedDinopony/spyrort-asl

started time in 4 days

startedVelocityRa/pctation

started time in 4 days

startedLiveSplit/livesplit-core

started time in 5 days

release haskell/haskell-language-server

1.2.0

released time in 5 days

release haskell/haskell-language-server

1.2.0

released time in 6 days

startedLiveSplit/LiveSplitOne

started time in 6 days

release haskell/haskell-language-server

1.2.0

released time in 6 days

fork wenkokke/knausj_talon

Config for talon for Mac, Windows and Linux. Very much in progress.

fork in 8 days

fork wenkokke/homebrew-cask

🍻 A CLI workflow for the administration of macOS applications distributed as binaries

https://brew.sh

fork in 8 days

fork wenkokke/brew

🍺 The missing package manager for macOS (or Linux)

https://brew.sh

fork in 8 days

startedChrisPenner/slick

started time in 8 days

startedorg-roam/org-roam

started time in 8 days

release protozoa-nz/whakapapa-ora

v1.5.1

released time in 8 days

startedejoubaud/dom-selector

started time in 8 days

created repositorymasaeedu/infinite

created time in 10 days

startedzpl-c/zpl

started time in 11 days

created repositoryradicle-dev/radicle-http-api

Radicle HTTP API

created time in 11 days

startedbladyjoker/deutschi

started time in 14 days

startedforbole/cosmos-rust-wallet

started time in 15 days

startedorangeduck/Cello

started time in 15 days

startedjoric/nrfmicro

started time in 16 days