sfaira.train.SummarizeGridsearchEmbedding¶
- class sfaira.train.SummarizeGridsearchEmbedding(source_path: dict, cv: bool, loss_idx: int = 0, mse_idx: int = 1, model_id_len: int = 3)¶
Attributes
alias of
ListReturns keys of cross-validation used in dictionaries in this class.
Methods
best_model_by_partition(partition_select, ...)- param partition_select
best_model_embedding([subset, partition, ...])get_best_model_ids(tab, metric_select, ...)- param tab
get_gradients_by_celltype(model_organ, ...)Compute gradients across latent units with respect to input features for each cell type.
load_gs(gs_ids)Loads all relevant data of a grid search.
load_y(hat_or_true, run_id)plot_active_latent_units(organ, topology_version)Plots latent unit activity measured by empirical variance of the expected latent space.
plot_best([rename_levels, partition_select, ...])- param rename_levels
plot_best_model_by_hyperparam(metric_select)Produces boxplots for all hyperparameters choices by organ.
plot_completions([groupby, height_fig, ...])Plot number of completed grid search points by category.
plot_gradient_cor(model_organ, data_organ, ...)Plot correlation heatmap of gradient vectors accumulated on input features between cell types or models.
plot_gradient_distr(model_organ, data_organ, ...)plot_npc(organ, topology_version[, cvs])Plots the explained variance ration that accumulates explained variation of the latent space’s ordered principal components.
plot_training_history(metric_select, metric_show)Plot train and validation loss during training and learning rate reduction for each organ
save_best_weight(path[, partition, metric, ...])Copies weight file from best hyperparameter setting from grid search directory to zoo directory with cleaned file name.
write_best_hyparam(write_path[, subset, ...])