Get the list of inputs in the `col` of `path/csv_name`.

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

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

Evaluate the `output` with the critic then uses `self.loss_funcG` to combine it with `target`.

Return the default embedding sizes suitable for this data or takes the ones in `sz_dict`.

Return list of files in `path` that have a suffix in `extensions`; optionally `recurse`.

Return function link to notebook documentation of `ft`. Private functions link to source code

Return globally assigned variables.

Fetch item or user (based on `is_item`) for all in `arr`. (Set model to `cpu` and no grad.)

Finds all submodules of notebook - sorted by submodules > top level modules > manual imports. This gives notebook imports priority

List the inner functions of a class.

Return `label_cls` or guess one from the first element of `labels`.

Returns source code link and pytest command

Find any lines where `elt` is called and return the parent test function

Tuple of predictions and targets, and optional losses (if `loss_func`) using `dl`, max batches `n_batch`.

Returns github link for given file

Create and return a tmp filename, optionally at a specific path. `os.remove` when done with it.

get used memory (in MBs) for the currently selected gpu id, w/o emptying the cache, and needing the `gpu_handle` arg

ICNR init of `x`, with `scale` and `init` function.

Calculate the intrinsic attention of the input w.r.t to an output `class_id`, or the classification given by the model if `None`. For referenc

Calculate the intrinsic attention of the input w.r.t to an output `class_id`, or the classification given by the model if `None`. For referenc

Label every item with `const`.

Give a label to each filename depending on its folder.

Use the labels in `train_labels` and `valid_labels` to label the data. `label_cls` will overwrite the default.

Search `docstring` for backticks and attempt to link those functions to respective documentation.

Create documentation links for all cells in markdown with backticks.

Load a pretrained model and adapts it to the data vocabulary.

Return `ndarray` of `str` of lines of text from `path`.

Return a Button widget with specified `handler`.

Return model path to `filename`, checking locally first then in the config file.

Create a new copy of `self` with `kwargs` replacing current values.

Set state when computing statistics for Global or Group

Normalize `x` with `mean` and `std`.

Add normalize transform using `stats` (defaults to `DataBunch.batch_stats`)

Create normalize/denormalize func using `mean` and `std`, can specify `do_y` and `device`.

Get number of children modules in `m`.

Handle new batch `xb`,`yb` in `train` or validation.

Handle new epoch.

Epoch is done, process `val_loss`.

Initializes the best value.

Return `Image` object created from image in file `fn`.

For a LabelList `ll_input`, resize each image to `size` using `resize_method` and `padding_mode`.

Parse the docstring into its components. :return: a dictionary of form { "short_description": ...,

Split data in `a` equally among `n_cpus` cores

Convert PIL style `image` array to torch style image tensor.

Draw 1 or shape=`size` random booleans (`True` occuring with probability `p`).

Reconstruct one of the underlying item for its data `t`.

Path relative to the `fastai` parent directory

create new OrderedDict that does not contain `module.`

Re-render Jupyter cell for batch of images.

Print delta used+peaked, and an optional context note, which can also be preset in constructor

Scale the coords in `flow` to -1/1 or the image size depending on `to_unit`.

Set the current `dl` used.

Show documentation for element `elt`. Supported types: class, Callable, and enum.

Split the data depending on the folder (`train` or `valid`) in which the filenames are.

Split the data by result of `func` (which returns `True` for validation set).

Split `kwargs` between those expected by `func` and the others.

Separate the parameters in `layer_groups` between `no_wd_types` and bias (`bias_types`) from the rest.

Build anneal schedule for all of the parameters.

Recursively map lists of tensors in `b ` to the cpu.

Recursively map lists of items in `b ` to their wrapped data.

trim function from PEP-257

Draw 1 or shape=`size` random floats from uniform dist: min=`low`, max=`high`.

Download `url` to `fname` if `dest` doesn't exist, and un-tgz to folder `dest`.

向执行队列中添加一个任务

Takes the average of the input.

Same as `nn.BCEWithLogitsLoss`, but flattens input and target.

Apply `nn.AdaptiveAvgPool2d` to `x` and then flatten the result.

执行队列中的任务