If you want to play with local LLMs on the retreat, please install Ollama and download at least one model before you travel. Pulls are still large enough that, if everyone downloads on the venue Wi-Fi in Barcelona, the network will die.
I only have capacity to test this on macOS. If you are using Linux or Windows, please follow the official Ollama installation instructions at https://ollama.com/download.
Install Ollama from the terminal:
| import pymc3 as pm | |
| import numpy as np | |
| import arviz as az | |
| %config InlineBackend.figure_format = 'retina' | |
| # Data from https://twitter.com/tomstafford/status/1456914037195907079?s=20 | |
| N = np.array([1258, 280]) # total number of caffiene and non-caffiene drinkers | |
| k = np.array([966, 168]) # total number of those who have favourite mugs | |
| def cohens_h(p): |
| % Add tooolbox code to path | |
| addpath('~/git-local/darc-experiments-matlab') | |
| % Run the toolbox setup code | |
| env_setup | |
| % Get participant information with GUI | |
| [expt_options] = getHumanExperimentOptions(); | |
| plotting_style = 'full'; |
| % Add tooolbox code to path | |
| addpath('~/git-local/darc-experiments-matlab') | |
| % Run the toolbox setup code | |
| env_setup | |
| % Get participant information with GUI | |
| [expt_options] = getHumanExperimentOptions(); | |
| plotting_style = 'full'; |
| # We have a dataframe with many columns. | |
| # 15 of these columns correspond to responses to questions about hunger. | |
| # They group into 5 subscales (each with 3 questions). | |
| # I want to calculate these subscales (means) as well as an overall score (mean of all). | |
| # This then needs to applied to the control condition (C) and a fasted condition (F) scores | |
| # Once the means are calculated I no longer need the raw scores, so bonus is to automatically remove these columns. Need to retain all other columns in orginal dataframe. | |
| # The code below works just fine, but seems pretty verbose. Is there a more R-like, concise way to do this? | |
| # It seems like a split-apply-combine, but applied to columns rather than rows? |
| true_mean = 0; | |
| true_sigma = 1; | |
| % likelihood_func = @(x, mean, sigma) normpdf(x, mean, sigma); | |
| % the above function to calcalate in matrix form, for speed | |
| likelihood_func = @(params)... | |
| prod(normpdf(repmat(x,[1 numel(params(1))]),... | |
| repmat(params(1), [1 numel(x)])',... | |
| repmat(params(2),[1 numel(x)])' ), 1); | |
| %% generate data |