Writes single scalar value to Tensorboard.

Writes training loss to Tensorboard.

Writes model weight histograms to Tensorboard.

Add `val` to statistic

Computes accuracy with `targs` when `input` is bs * n_classes.

Add test set containing `items` with an arbitrary `label`.

Create the ragged array that will be filled when we ask for items.

Look at params (annotated with `Param`) in func and return an `ArgumentParser`

Apply `f` to children of `m`.

Function applied to `df` if it's the train set.

Same as `range_of` but returns an array.

Given `arrs` is [a,b,...] and `mask`index - return[(a[mask],a[~mask]),(b[mask],b[~mask]),...].

Computes the area under the receiver operator characteristic (ROC) curve using the trapezoid method. Restricted binary classification tasks.

convert Bs to MBs and round down

Check if current batch contains already deleted images.

Grab a batch of data and call reduction function `func` per channel

Convert bounding box points from (width,height,center) to (height,width,top,left).

If `module` is batchnorm don't use half precision.

Calculation of the statistics used by the FID. Params: -- files : List of image files paths -- model : Instance of inception m

Calculates the FID of two paths

Numpy implementation of the Frechet Distance. The Frechet distance between two multivariate Gaussians X_1 ~ N(mu_1, C_1) and X_2 ~ N(mu_2, C_2

Decorator to create a simple CLI from `func` using `anno_parser`

Make channel the first axis of `x` and flatten remaining axes

Install required dependencies for DeOldify, with version checks.

Check if a package is installed with an optional version.

Return the children of `m` and its direct parameters not registered in modules.

Yield successive `n`-sized chunks from `l`.

Clear the widget's images preview pane.

Build convnet style learner.

Computes the similarity function between each embedding of `t1` and `t2` matrices.

Create `nn.ConvTranspose2d` layer.

Convert the model `wgts` to go with a new vocabulary.

Create a `TextDataBunch` in `path` from the `datasets` for language modelling. Passes `**dl_kwargs` on to `DataLoader()`

Create custom convnet architecture

Model head that takes `nf` features, runs through `lin_ftrs`, and about `nc` classes.

Create a list of images, filenames and labels but first removing files that are not supposed to be displayed.

Create the documentation notebook for module `mod_name` in path `dest_path`

Critic to train a `GAN`.

Return feature names of date/time cycles as produced by `cyclic_dt_features`.

Convert `batch` items to tensor data.

Pick an embedding size for `n` depending on `classes` if not given in `sz_dict`.

Free the Learner internals, leaving just an empty shell that consumes no memory

Display a few preview images in the notebook

Download a file from a URL to a specific path, showing progress.

Download `url` to destination `fname`.

Search for `n_images` images on Google, matching `search_term` and `size` requirements, download them into `path`/`search_term` and verify th

Download all models if they don't exist locally.

Create a dummy batch to go through `m` with `size`.

Build log-stepped array from `start` to `stop` in `n` steps.

Exp RMSE between `pred` and `targ`.

Explained variance between `pred` and `targ`.

Build pretrained Inception model for FID computation The Inception model for FID computation uses a different set of weights and has a slight

Fill the row with tokens from the ragged array. --OBS-- overlap != 1 has not been implemented

Attempt to resolve keywords such as Learner.lr_find. `match_last` starts matching from last component.

Searches `fastai/tests` folder for any test functions related to `elt`

Searches in `fastai/tests` directory for module tests

Find all functions in `test_file` related to `elt`

Retrieve first layer in a module `m`.

Fit the `model` on `data` and learn using `loss_func` and `opt`.

Merge and format the docstring definition with `arg_comments` and `alt_doc_string`.

Formats function param to `param1:Type=val`. Font weights: param1=bold, val=bold+italic

Format stats before printing.

Create an `ItemList` in `path` from the inputs in the `cols` of `df`.

Create a `TextDataBunch` from DataFrames. `kwargs` are passed to the dataloader creation.

Create an `ItemList` in `path` from the filenames that have a suffix in `extensions`. `recurse` determines if we search subfolders.

Get the list of files in `path` that have an image suffix. `recurse` determines if we search subfolders.

Create a `TextDataBunch` from ids, labels and a `vocab`. `kwargs` are passed to the dataloader creation.

Gets preds, y_true, losses to construct base class from a learner

Create from list of `fnames` in `path` with `label_func`.

Find any lines where `fn_name` is invoked and return the parent test function

Define loss functions for a GAN from `loss_gen` and `loss_crit`.

Generate classes from `items` by taking the sorted unique values.

Calculates the activations of the pool_3 layer for all images. Params: -- files : List of image files paths -- model : Instan

Gets activations at the layer specified by `hook`, applies `pool` of dim `pool_dim` and concatenates

Class definition.

Either max_vocab_sz or one quarter of the number of unique words in `texts`

Default notbook with the minimum metadata.