Shortcut for `enumerate(subplots.flatten())`

Call `func` for every combination of `r,c` on a subplot

Return predicted class, label and probabilities for `item`.

Process batch `b` of `TensorImage`.

Process a list of `texts`.

Process one text `t` with tokenizer `tok`.

Purge the `Learner` of all cached attributes to release some GPU memory.

Check whether module==version is available on pypi

Returns CLI command to run specific test function

R2 score (coefficient of determination) between `pred` and `targ`.

Randomized version of `crop_pad`.

Randomized version of `zoom`.

Read the values inside the optimizer for the hyper-parameters.

Read the notebook in `fname`.

Build a dictionary containing the position of the `cells`.

Remove the hooks from the model.

If your convolutional window is greater than 1 and you save previous xs, you must reset at the beginning of each new sequence.

Reset the internal memory.

Size to resize to, to hit `targ_sz` at same aspect ratio, in PIL coords (i.e w*h)

Return an image array from run-length encoded string `mask_rle` with `shape`.

Computes the receiver operator characteristic (ROC) curve by determining the true positive ratio (TPR) and false positive ratio (FPR) for various clas

Root mean squared error between `pred` and `targ`.

Check the underlying data in the training set can be properly loaded.

Save model and optimizer state (if `with_opt`) with `file` to `self.model_dir`. `file` can be file-like (file or buffer)

Save `self.itos` in `path`

Sets the seeds for all pseudo random generators in fastai lib

Set bn layers in eval mode for all recursive children of `m`.

For inference, will briefly replace the dataset with one that only contains `item`.

Set parameters that control how we `grid_sample` the image after transforms are applied.

Print user's setup information

Display video in `url`.

Shows `ys` (target images) on a figure of `figsize`.

Show the `xs` (inputs) and `ys` (targets) on a figure of `figsize`.

Show the `xs` (inputs) and `ys` (targets). `max_len` is the maximum number of tokens displayed.

Show `xs` (inputs), `ys` (targets) and `zs` (predictions). `max_len` is the maximum number of tokens displayed.

Sigmoid function with range `(low, high)`

Sorts `similarities` and return the indexes in pairs ordered by highest similarity.

Split the data by using the names in `valid_names` for validation.

Split the data between `train` and `valid`.

Split `model` according to the layers in `splits`.

Split `model` according to the indexes in `idxs`.

Converts 'fastai.foo.bar' into an function 'object' if such exists

Create `pd.DataFrame` containing `items` from `self.x` and `self.y`.

Recursively map lists of tensors in `b ` to FP16.

Recursively map lists of tensors in `b ` to FP16.

Set the `tfms` and `tfm_y` value to be applied to the inputs and targets.

Truncated normal initialization.

Build Unet learner from `data` and `arch`.

Build Unet learner from `data` and `arch`.

Update the documentation notebook of a given module.

Update the `pos_dict` by moving all positions after `start_key` by `nbr`.

Change `url` to a path.

Use only a sample of `sample_pct`of the full dataset and an optional `seed`.

Check if input value is empty.

Convert a value `x` from 0 to 1 (inclusive) to an RGBA tuple according to `cmap` times transparency `alpha_mult`.

Convert a value `x` from 0 to 1 (inclusive) to an RGBA tuple according to `cmap` times transparency `alpha_mult`.

Check if the images in `path` aren't broken, maybe resize them and copy it in `dest`.

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

创建一个执行队列