Torch Core

Basic pytorch functions used in the fastai library

from PIL import Image

source

setup_cuda

 setup_cuda (benchmark=True)

Sets the main cuda device and sets cudnn.benchmark to benchmark

Arrays and show

source

subplots

 subplots (nrows:int=1, ncols:int=1, figsize:tuple=None, imsize:int=3,
           suptitle:str=None, sharex=False, sharey=False, squeeze=True,
           width_ratios=None, height_ratios=None, subplot_kw=None,
           gridspec_kw=None, **kwargs)

Returns a figure and set of subplots to display images of imsize inches

	Type	Default	Details
nrows	int	1	Number of rows in returned axes grid
ncols	int	1	Number of columns in returned axes grid
figsize	tuple	None	Width, height in inches of the returned figure
imsize	int	3	Size (in inches) of images that will be displayed in the returned figure
suptitle	str	None	Title to be set to returned figure
sharex	bool	False
sharey	bool	False
squeeze	bool	True
width_ratios	NoneType	None
height_ratios	NoneType	None
subplot_kw	NoneType	None
gridspec_kw	NoneType	None
kwargs
Returns	(plt.Figure, plt.Axes)		Returns both fig and ax as a tuple

This is used in get_grid. suptitle, sharex, sharey, squeeze, subplot_kw and gridspec_kw are all passed down to plt.subplots.

source

show_image

 show_image (im, ax=None, figsize=None, title=None, ctx=None, cmap=None,
             norm=None, aspect=None, interpolation=None, alpha=None,
             vmin=None, vmax=None, origin=None, extent=None,
             interpolation_stage=None, filternorm=True, filterrad=4.0,
             resample=None, url=None, data=None, **kwargs)

Show a PIL or PyTorch image on ax.

show_image can show PIL images…

im = Image.open(TEST_IMAGE_BW)
ax = show_image(im, cmap="Greys")

…and color images with standard CHW dim order…

im2 = np.array(Image.open(TEST_IMAGE))
ax = show_image(im2, figsize=(2,2))

…and color images with HWC dim order…

im3 = torch.as_tensor(im2).permute(2,0,1)
ax = show_image(im3, figsize=(2,2))

source

show_titled_image

 show_titled_image (o, ax=None, figsize=None, title=None, ctx=None,
                    cmap=None, norm=None, aspect=None, interpolation=None,
                    alpha=None, vmin=None, vmax=None, origin=None,
                    extent=None, interpolation_stage=None,
                    filternorm=True, filterrad=4.0, resample=None,
                    url=None, data=None, **kwargs)

Call show_image destructuring o to (img,title)

show_titled_image((im3,'A puppy'), figsize=(2,2))

Show all images ims as subplots with rows using titles. suptitle provides a way to create a figure title for all images. If you use suptitle, constrained_layout is used unless you set constrained_layout to False.

source

show_images

 show_images (ims, nrows=1, ncols=None, titles=None, figsize:tuple=None,
              imsize:int=3, suptitle:str=None, sharex=False, sharey=False,
              squeeze=True, width_ratios=None, height_ratios=None,
              subplot_kw=None, gridspec_kw=None)

Show all images ims as subplots with rows using titles.

	Type	Default	Details
ims
nrows	int	1	Number of rows in returned axes grid
ncols	int	1	Number of columns in returned axes grid
titles	NoneType	None
figsize	tuple	None	Width, height in inches of the returned figure
imsize	int	3	Size (in inches) of images that will be displayed in the returned figure
suptitle	str	None	Title to be set to returned figure
sharex	bool	False
sharey	bool	False
squeeze	bool	True
width_ratios	NoneType	None
height_ratios	NoneType	None
subplot_kw	NoneType	None
gridspec_kw	NoneType	None
Returns	(plt.Figure, plt.Axes)		Returns both fig and ax as a tuple

show_images((im,im3),titles=('number','puppy'),suptitle='Number Puppy',  imsize=3)

ArrayImage, ArrayImageBW and ArrayMask are subclasses of ndarray that know how to show themselves.

source

ArrayBase

An ndarray that can modify casting behavior

source

ArrayImageBase

Base class for arrays representing images

source

ArrayImage

An array representing an image

source

ArrayImageBW

An array representing an image

source

ArrayMask

An array representing an image mask

im = Image.open(TEST_IMAGE)

im_t = cast(im, ArrayImage)
test_eq(type(im_t), ArrayImage)

ax = im_t.show(figsize=(2,2))

test_fig_exists(ax)

Basics

source

tensor

 tensor (x, *rest, dtype=None, device=None, requires_grad=False,
         pin_memory=False)

Like torch.as_tensor, but handle lists too, and can pass multiple vector elements directly.

test_eq(tensor(torch.tensor([1,2,3])), torch.tensor([1,2,3]))
test_eq(tensor(array([1,2,3])), torch.tensor([1,2,3]))
test_eq(tensor(1,2,3), torch.tensor([1,2,3]))
test_eq_type(tensor(1.0), torch.tensor(1.0))

set_seed is useful for reproducibility between runs. It is important to remember that certain classes such as Dataloaders have internal random number generators that is not effected by this function, so this must be run before such objects are created in order to guarantee reproducibility.

source

set_seed

 set_seed (s, reproducible=False)

Set random seed for random, torch, and numpy (where available)

Here is an example of how set_seed can be used to reset the state of random number generators.

set_seed(2*33)
a1 = np.random.random()
a2 = torch.rand(())
a3 = random.random()
set_seed(2*33)
b1 = np.random.random()
b2 = torch.rand(())
b3 = random.random()
print('a\'s: {0:3.3f} {1:3.3f} {2:3.3f}'.format(a1,a2,a3))
print('b\'s: {0:3.3f} {1:3.3f} {2:3.3f}'.format(b1,b2,a3))

a's: 0.154 0.498 0.071
b's: 0.154 0.498 0.071

test_eq(a1,b1)
test_eq(a2,b2)
test_eq(a3,b3)

get_random_states and set_random_states are useful for storing a state so you can go back to it later.

source

get_random_states

 get_random_states ()

Gets states for random, torch, and numpy random number generators

source

set_random_states

 set_random_states (random_state, numpy_state, torch_state,
                    torch_cuda_state, torch_deterministic,
                    torch_benchmark)

Set states for random, torch, and numpy random number generators

Below notice that the old values and rewinded values are the same because we were able to return to the previous state.

old_states = get_random_states()
olds = (random.random(),np.random.random(),torch.rand(()))
news = (random.random(),np.random.random(),torch.rand(()))
set_random_states(**old_states)
rewinds = (random.random(),np.random.random(),torch.rand(()))

print('olds:    {0:3.3f} {1:3.3f} {2:3.3f}'.format(*olds))
print('news:    {0:3.3f} {1:3.3f} {2:3.3f}'.format(*news))
print('rewinds: {0:3.3f} {1:3.3f} {2:3.3f}'.format(*rewinds))

olds:    0.435 0.134 0.023
news:    0.246 0.363 0.227
rewinds: 0.435 0.134 0.023

test_ne(olds,news)
test_eq(olds,rewinds)

In no_random we combine the ideas of rewinding state with get_random_states and set_random_states with the ability to set_seed and create a context manager that can allow us to control randomness in a portion of our code.

Note: Similar to torch.random.fork_rng, but also with numpy and random

source

no_random

 no_random (seed=42, reproducible=True)

Stores and retrieves state of random number generators. Sets random seed for random, torch, and numpy.

Here are some examples on how we can use no_random to control the randomness within a block of code.

states=get_random_states()
olds = (random.random(),np.random.random(),torch.rand(()))
set_random_states(**states) #rewinding above random calls

with no_random():
    new1 = (random.random(),np.random.random(),torch.rand(()))
with no_random():
    new2 = (random.random(),np.random.random(),torch.rand(()))
with no_random(seed=100):
    seeded1 = (random.random(),np.random.random(),torch.rand(()))
with no_random(seed=100):
    seeded2 = (random.random(),np.random.random(),torch.rand(()))
        
rewinds = (random.random(),np.random.random(),torch.rand(()))

print('olds:    {0:3.3f} {1:3.3f} {2:3.3f}'.format(*olds))
print('new1:    {0:3.3f} {1:3.3f} {2:3.3f}'.format(*new1))
print('new2:    {0:3.3f} {1:3.3f} {2:3.3f}'.format(*new2))
print('seeded1: {0:3.3f} {1:3.3f} {2:3.3f}'.format(*seeded1))
print('seeded2: {0:3.3f} {1:3.3f} {2:3.3f}'.format(*seeded2))
print('rewinds: {0:3.3f} {1:3.3f} {2:3.3f}'.format(*rewinds))

olds:    0.246 0.363 0.227
new1:    0.639 0.375 0.882
new2:    0.639 0.375 0.882
seeded1: 0.146 0.543 0.112
seeded2: 0.146 0.543 0.112
rewinds: 0.246 0.363 0.227

Notice that olds, and rewinds are alos both equal to each other. From this we can see that everything in the with blocks did not update the state outside of the block. Inside of the block, the state is reset for any particular seed, so for the same seed you should get the same random number generator results.

Note: It is important to remember that classes like Dataloader have internal random number generators, and no_random will have no effect on those random number generators.

test_ne(olds,new1)
test_eq(new1,new2)
test_ne(new1,seeded1)
test_eq(seeded1,seeded2)
test_eq(olds,rewinds)

source

unsqueeze

 unsqueeze (x, dim=-1, n=1)

Same as torch.unsqueeze but can add n dims

t = tensor([1])
t2 = unsqueeze(t, n=2)
test_eq(t2,t[:,None,None])

source

unsqueeze_

 unsqueeze_ (x, dim=-1, n=1)

Same as torch.unsqueeze_ but can add n dims

t = tensor([1])
unsqueeze_(t, n=2)
test_eq(t, tensor([1]).view(1,1,1))

source

apply

 apply (func, x, *args, **kwargs)

Apply func recursively to x, passing on args

source

maybe_gather

 maybe_gather (x, axis=0)

Gather copies of x on axis (if training is distributed)

source

to_detach

 to_detach (b, cpu=True, gather=True)

Recursively detach lists of tensors in b; put them on the CPU if cpu=True.

gather only applies during distributed training and the result tensor will be the one gathered across processes if gather=True (as a result, the batch size will be multiplied by the number of processes).

source

to_half

 to_half (b)

Recursively map floating point tensors in b to FP16.

source

to_float

 to_float (b)

Recursively map floating point tensors in b to float.

source

default_device

 default_device (use_cuda=-1)

Return or set default device; use_cuda: None - CUDA if available; True - error if not available; False - CPU

source

default_device

 default_device (use=-1)

Return or set default device; use_cuda: -1 - CUDA/mps if available; True - error if not available; False - CPU

if torch.cuda.is_available():
    _td = torch.device(torch.cuda.current_device())
    test_eq(default_device(-1), _td)
    test_eq(default_device(True), _td)
else:
    test_eq(default_device(False), torch.device('cpu'))
default_device(-1);

source

to_device

 to_device (b, device=None, non_blocking=False)

Recursively put b on device.

t = to_device((3,(tensor(3),tensor(2))))
t1,(t2,t3) = t

if torch.cuda.is_available():
    test_eq_type(t,(3,(tensor(3).cuda(),tensor(2).cuda())))
    test_eq(t2.type(), "torch.cuda.LongTensor")
    test_eq(t3.type(), "torch.cuda.LongTensor")

source

to_cpu

 to_cpu (b)

Recursively map tensors in b to the cpu.

t3 = to_cpu(t3)
test_eq(t3.type(), "torch.LongTensor")
test_eq(t3, 2)

source

to_np

 to_np (x)

Convert a tensor to a numpy array.

t3 = to_np(t3)
test_eq(type(t3), np.ndarray)
test_eq(t3, 2)

source

to_concat

 to_concat (xs, dim=0)

Concat the element in xs (recursively if they are tuples/lists of tensors)

test_eq(to_concat([tensor([1,2]), tensor([3,4])]), tensor([1,2,3,4]))
test_eq(to_concat([tensor([[1,2]]), tensor([[3,4]])], dim=1), tensor([[1,2,3,4]]))
test_eq_type(to_concat([(tensor([1,2]), tensor([3,4])), (tensor([3,4]), tensor([5,6]))]), (tensor([1,2,3,4]), tensor([3,4,5,6])))
test_eq_type(to_concat([[tensor([1,2]), tensor([3,4])], [tensor([3,4]), tensor([5,6])]]), [tensor([1,2,3,4]), tensor([3,4,5,6])])
test_eq_type(to_concat([(tensor([1,2]),), (tensor([3,4]),)]), (tensor([1,2,3,4]),))

test_eq(to_concat([tensor([[1,2]]), tensor([[3,4], [5,6]])], dim=1), [tensor([1]),tensor([3, 5]),tensor([4, 6])])

test_eq(type(to_concat([dict(foo=tensor([1,2]), bar=tensor(3,4))])), dict)

Tensor subtypes

source

Tensor.set_meta

 Tensor.set_meta (x, as_copy=False)

Set all metadata in __dict__

source

Tensor.as_subclass

 Tensor.as_subclass (typ)

Cast to typ and include __dict__ and meta

Tensor.set_meta and Tensor.as_subclass work together to maintain __dict__ after casting.

class _T(Tensor): pass
t = tensor(1.).requires_grad_()
t.img_size = 1
t2 = t.as_subclass(_T)
test_eq(t.img_size, t2.img_size)
test_eq(t2.img_size, 1)
assert(t2.requires_grad_)

source

TensorBase

 TensorBase (x, **kwargs)

A Tensor which support subclass pickling, and maintains metadata when casting or after methods

TensorBase hooks into __torch_function__ to ensure metadata is not lost. To see all functions being called, set debug.

a = TensorBase(1)
TensorBase.debug=True
1/(a+1)

TensorBase(0.5000)

TensorBase and its subclasses also allow for passing through metadata size as img_size…

from torch.utils.data._utils.collate import default_collate

a = TensorBase(1,img_size=(128,128))
test_eq(a.img_size,(128,128))
b = cast(a,TensorBase)
test_eq(b.img_size,(128,128))
test_eq(torch.stack([a,b],0).img_size,(128,128))

test_eq(default_collate([a,b]).img_size,(128,128))

class _TImage(TensorBase): pass
class _TImage2(_TImage): pass
t1 = _TImage([1.])
t2 = _TImage2([1.])
t2+t1

_TImage2([2.])

class _T(TensorBase): pass

t = _T(range(5))
test_eq(t[0], 0)
test_eq_type(t+1, _T(range(1,6)))
test_eq(repr(t), '_T([0, 1, 2, 3, 4])')
test_eq_type(t[_T([False,False,True,True,True])], _T([2,3,4]))
test_eq_type(t[_T([2,3,4])], _T([2,3,4]))
test_eq(type(pickle.loads(pickle.dumps(t))), _T)
test_eq_type(t.new_ones(1), _T([1]))
test_eq_type(t.new_tensor([1,2]), _T([1,2]))

t = tensor([1,2,3])
m = TensorBase([False,True,True])
test_eq(t[m], tensor([2,3]))
t = tensor([[1,2,3],[1,2,3]])
m = cast(tensor([[False,True,True],
                 [False,True,True]]), TensorBase)
test_eq(t[m], tensor([2,3,2,3]))

t = tensor([[1,2,3],[1,2,3]])
t.img_size = 1
t2 = cast(t, TensorBase)
test_eq(t2.img_size, t.img_size)
x = retain_type(tensor([4,5,6]), t2)
test_eq(x.img_size, t.img_size)
t3 = TensorBase([[1,2,3],[1,2,3]], img_size=1)
test_eq(t3.img_size, t.img_size)
t4 = t2+1
t4.img_size = 2
test_eq(t2.img_size, 1)
test_eq(t4.img_size, 2)
# this will fail with `Tensor` but works with `TensorBase`
test_eq(pickle.loads(pickle.dumps(t2)).img_size, t2.img_size)

source

TensorImageBase

 TensorImageBase (x, **kwargs)

A Tensor which support subclass pickling, and maintains metadata when casting or after methods

source

TensorImage

 TensorImage (x, **kwargs)

A Tensor which support subclass pickling, and maintains metadata when casting or after methods

source

TensorImageBW

 TensorImageBW (x, **kwargs)

A Tensor which support subclass pickling, and maintains metadata when casting or after methods

source

TensorMask

 TensorMask (x, **kwargs)

A Tensor which support subclass pickling, and maintains metadata when casting or after methods

im = Image.open(TEST_IMAGE)
im_t = cast(array(im), TensorImage)
test_eq(type(im_t), TensorImage)

im_t2 = cast(tensor(1), TensorMask)
test_eq(type(im_t2), TensorMask)
test_eq(im_t2, tensor(1))
ax = im_t.show(figsize=(2,2))
_ =(im_t == im_t2)

test_fig_exists(ax)

Operations between TensorMask and TensorImageBase objects return the type of the TensorImageBase object:

a = TensorMask([1,2])
test_eq_type(TensorImage(1)+a, TensorImage([2,3]))
test_eq_type(1-a, TensorMask([0,-1]))

source

TensorFlowField

 TensorFlowField (x, **kwargs)

A Tensor which support subclass pickling, and maintains metadata when casting or after methods

t1 = TensorImage([1.]).view(1,1,1,1)
t2 = TensorFlowField([1.,1.]).view(1,1,1,2)
test_eq_type(F.grid_sample(t1, t2), TensorImage([[[[0.25]]]]))

source

TensorCategory

 TensorCategory (x, **kwargs)

A Tensor which support subclass pickling, and maintains metadata when casting or after methods

tc = TensorCategory([1,2,3])
mask_t = TensorMask([0,2,4,5])
im_t = TensorImage([0,2,4,5])
test_eq(mask_t[tc], tensor([2,4,5]))
test_eq(im_t[tc], tensor([2,4,5]))

source

TensorMultiCategory

 TensorMultiCategory (x, **kwargs)

A Tensor which support subclass pickling, and maintains metadata when casting or after methods

source

TitledTensorScalar

 TitledTensorScalar (x, **kwargs)

A tensor containing a scalar that has a show method

source

L.cat

 L.cat (dim=0)

Same as torch.cat

source

L.stack

 L.stack (dim=0)

Same as torch.stack

source

L.tensored

 L.tensored ()

mapped(tensor)

source

L.tensored

 L.tensored ()

mapped(tensor)

There are shortcuts for torch.stack and torch.cat if your L contains tensors or something convertible. You can manually convert with tensored.

t = L(([1,2],[3,4]))
test_eq(t.tensored(), [tensor(1,2),tensor(3,4)])

source

L.stack

 L.stack (dim=0)

Same as torch.stack

test_eq(t.stack(), tensor([[1,2],[3,4]]))

source

L.cat

 L.cat (dim=0)

Same as torch.cat

test_eq(t.cat(), tensor([1,2,3,4]))

Chunks

source

concat

 concat (*ls)

Concatenate tensors, arrays, lists, or tuples

a,b,c = [1],[1,2],[1,1,2]
test_eq(concat(a,b), c)
test_eq_type(concat(tuple (a),tuple (b)), tuple (c))
test_eq_type(concat(array (a),array (b)), array (c))
test_eq_type(concat(tensor(a),tensor(b)), tensor(c))
test_eq_type(concat(TensorBase(a),TensorBase(b)), TensorBase(c))
test_eq_type(concat([1,1],1), [1,1,1])
test_eq_type(concat(1,1,1), L(1,1,1))
test_eq_type(concat(L(1,2),1), L(1,2,1))

source

Chunks

 Chunks (chunks, lens=None)

Slice and int indexing into a list of lists

docs = L(list(string.ascii_lowercase[a:b]) for a,b in ((0,3),(3,7),(7,8),(8,16),(16,24),(24,26)))

b = Chunks(docs)
test_eq([b[ o] for o in range(0,5)], ['a','b','c','d','e'])
test_eq([b[-o] for o in range(1,6)], ['z','y','x','w','v'])
test_eq(b[6:13], 'g,h,i,j,k,l,m'.split(','))
test_eq(b[20:77], 'u,v,w,x,y,z'.split(','))
test_eq(b[:5], 'a,b,c,d,e'.split(','))
test_eq(b[:2], 'a,b'.split(','))

t = torch.arange(26)
docs = L(t[a:b] for a,b in ((0,3),(3,7),(7,8),(8,16),(16,24),(24,26)))
b = Chunks(docs)
test_eq([b[ o] for o in range(0,5)], range(0,5))
test_eq([b[-o] for o in range(1,6)], [25,24,23,22,21])
test_eq(b[6:13], torch.arange(6,13))
test_eq(b[20:77], torch.arange(20,26))
test_eq(b[:5], torch.arange(5))
test_eq(b[:2], torch.arange(2))

docs = L(TensorBase(t[a:b]) for a,b in ((0,3),(3,7),(7,8),(8,16),(16,24),(24,26)))
b = Chunks(docs)
test_eq_type(b[:2], TensorBase(range(2)))
test_eq_type(b[:5], TensorBase(range(5)))
test_eq_type(b[9:13], TensorBase(range(9,13)))

Simple types

source

show_title

 show_title (o, ax=None, ctx=None, label=None, color='black', **kwargs)

Set title of ax to o, or print o if ax is None

test_stdout(lambda: show_title("title"), "title")
# ensure that col names are unique when showing to a pandas series
assert show_title("title", ctx=pd.Series(dict(a=1)), label='a').equals(pd.Series(dict(a=1,a_='title')))

source

ShowTitle

 ShowTitle ()

Base class that adds a simple show

source

TitledInt

An int with show

source

TitledStr

An str with show

source

TitledFloat

 TitledFloat (x=0)

A float with show

test_stdout(lambda: TitledStr('s').show(), 's')
test_stdout(lambda: TitledInt(1).show(), '1')

source

TitledTuple

 TitledTuple (x=None, *rest)

A fastuple with show

source

TitledStr.truncate

 TitledStr.truncate (n)

Truncate self to n

Other functions

source

get_empty_df

 get_empty_df (n)

Return n empty rows of a dataframe

source

display_df

 display_df (df)

Display df in a notebook or defaults to print

source

get_first

 get_first (c)

Get the first element of c, even if c is a dataframe

source

one_param

 one_param (m)

First parameter in m

source

item_find

 item_find (x, idx=0)

Recursively takes the idx-th element of x

source

find_device

 find_device (b)

Recursively search the device of b.

t2 = to_device(tensor(0))
dev = default_device()
test_eq(find_device(t2), dev)
test_eq(find_device([t2,t2]), dev)
test_eq(find_device({'a':t2,'b':t2}), dev)
test_eq(find_device({'a':[[t2],[t2]],'b':t2}), dev)

source

find_bs

 find_bs (b)

Recursively search the batch size of b.

x = torch.randn(4,5)
x1 = [1,2,3]
test_eq(find_bs(x1), 3)
test_eq(find_bs(x), 4)
test_eq(find_bs((x,x)), 4)
test_eq(find_bs([x, x]), 4)
test_eq(find_bs({'a':x,'b':x}), 4)
test_eq(find_bs({'a':[[x],[x]],'b':x}), 4)

source

np_func

 np_func (f)

Convert a function taking and returning numpy arrays to one taking and returning tensors

This decorator is particularly useful for using numpy functions as fastai metrics, for instance:

from sklearn.metrics import f1_score

@np_func
def f1(inp,targ): return f1_score(targ, inp)

a1,a2 = array([0,1,1]),array([1,0,1])
t = f1(tensor(a1),tensor(a2))
test_eq(f1_score(a1,a2), t)
assert isinstance(t,Tensor)

source

Module

 Module ()

Same as nn.Module, but no need for subclasses to call super().__init__

class _T(Module):
    def __init__(self): self.f = nn.Linear(1,1)
    def forward(self,x): return self.f(x)

t = _T()
t(tensor([1.]))

tensor([-0.0832], grad_fn=<AddBackward0>)

source

get_model

 get_model (model)

Return the model maybe wrapped inside model.

source

one_hot

 one_hot (x, c)

One-hot encode x with c classes.

test_eq(one_hot([1,4], 5), tensor(0,1,0,0,1).byte())
test_eq(one_hot(torch.tensor([]), 5), tensor(0,0,0,0,0).byte())
test_eq(one_hot(2, 5), tensor(0,0,1,0,0).byte())

source

one_hot_decode

 one_hot_decode (x, vocab=None)

test_eq(one_hot_decode(tensor(0,1,0,0,1)), [1,4])
test_eq(one_hot_decode(tensor(0,0,0,0,0)), [   ])
test_eq(one_hot_decode(tensor(0,0,1,0,0)), [2  ])

source

params

 params (m)

Return all parameters of m

source

trainable_params

 trainable_params (m)

Return all trainable parameters of m

m = nn.Linear(4,5)
test_eq(trainable_params(m), [m.weight, m.bias])
m.weight.requires_grad_(False)
test_eq(trainable_params(m), [m.bias])

source

norm_bias_params

 norm_bias_params (m, with_bias=True)

Return all bias and BatchNorm parameters

for norm_func in [nn.BatchNorm1d, partial(nn.InstanceNorm1d, affine=True)]:
    model = nn.Sequential(nn.Linear(10,20), norm_func(20), nn.Conv1d(3,4, 3))
    test_eq(norm_bias_params(model), [model[0].bias, model[1].weight, model[1].bias, model[2].bias])
    model = nn.ModuleList([nn.Linear(10,20, bias=False), nn.Sequential(norm_func(20), nn.Conv1d(3,4,3))])
    test_eq(norm_bias_params(model), [model[1][0].weight, model[1][0].bias, model[1][1].bias])
    model = nn.ModuleList([nn.Linear(10,20), nn.Sequential(norm_func(20), nn.Conv1d(3,4,3))])
    test_eq(norm_bias_params(model, with_bias=False), [model[1][0].weight, model[1][0].bias])

source

batch_to_samples

 batch_to_samples (b, max_n=10)

‘Transposes’ a batch to (at most max_n) samples

t = tensor([1,2,3])
test_eq(batch_to_samples([t,t+1], max_n=2), ([1,2],[2,3]))
test_eq(batch_to_samples(tensor([1,2,3]), 10), [1, 2, 3])
test_eq(batch_to_samples([tensor([1,2,3]), tensor([4,5,6])], 10), [(1, 4), (2, 5), (3, 6)])
test_eq(batch_to_samples([tensor([1,2,3]), tensor([4,5,6])], 2), [(1, 4), (2, 5)])
test_eq(batch_to_samples([tensor([1,2,3]), [tensor([4,5,6]),tensor([7,8,9])]], 10), 
        [(1, (4, 7)), (2, (5, 8)), (3, (6, 9))])
test_eq(batch_to_samples([tensor([1,2,3]), [tensor([4,5,6]),tensor([7,8,9])]], 2), [(1, (4, 7)), (2, (5, 8))])

t = fastuple(tensor([1,2,3]),TensorBase([2,3,4]))
test_eq_type(batch_to_samples(t)[0][1], TensorBase(2))
test_eq(batch_to_samples(t).map(type), [fastuple]*3)

source

Tensor.interp_1d

 Tensor.interp_1d (x:torch.Tensor, xp, fp)

Same as np.interp

brks = tensor(0,1,2,4,8,64).float()
ys = tensor(range_of(brks)).float()
ys /= ys[-1].item()
pts = tensor(0.2,0.5,0.8,3,5,63)

preds = pts.interp_1d(brks, ys)
test_close(preds.numpy(), np.interp(pts.numpy(), brks.numpy(), ys.numpy()))

plt.scatter(brks,ys)
plt.scatter(pts,preds)
plt.legend(['breaks','preds']);

source

Tensor.pca

 Tensor.pca (x:torch.Tensor, k=2)

Compute PCA of x with k dimensions.

source

logit

 logit (x)

Logit of x, clamped to avoid inf.

source

num_distrib

 num_distrib ()

Return the number of processes in distributed training (if applicable).

source

rank_distrib

 rank_distrib ()

Return the distributed rank of this process (if applicable).

source

distrib_barrier

 distrib_barrier ()

Place a synchronization barrier in distributed training

After calling this, ALL sub-processes in the pytorch process group must arrive here before proceeding.

source

Path.save_array

 Path.save_array (p:pathlib.Path, o, complib='lz4', lvl=3)

Save numpy array to a compressed pytables file, using compression level lvl

Compression lib can be any of: blosclz, lz4, lz4hc, snappy, zlib or zstd.

source

Path.load_array

 Path.load_array (p:pathlib.Path)

Save numpy array to a pytables file

source

base_doc

 base_doc (elt)

Print a base documentation of elt

source

doc

 doc (elt)

Try to use doc form nbdev and fall back to base_doc

source

nested_reorder

 nested_reorder (t, idxs)

Reorder all tensors in t using idxs

x = tensor([0,1,2,3,4,5])
idxs = tensor([2,5,1,0,3,4])
test_eq_type(nested_reorder(([x], x), idxs), ([idxs], idxs))

y = L(0,1,2,3,4,5)
z = L(i.item() for i in idxs)
test_eq_type(nested_reorder((y, x), idxs), (z,idxs))

source

flatten_check

 flatten_check (inp, targ)

Check that inp and targ have the same number of elements and flatten them.

x1,x2 = torch.randn(5,4),torch.randn(20)
x1,x2 = flatten_check(x1,x2)
test_eq(x1.shape, [20])
test_eq(x2.shape, [20])
x1,x2 = torch.randn(5,4),torch.randn(21)
test_fail(lambda: flatten_check(x1,x2))

Image helpers

source

make_cross_image

 make_cross_image (bw=True)

Create a tensor containing a cross image, either bw (True) or color

plt.imshow(make_cross_image(), cmap="Greys");

plt.imshow(make_cross_image(False).permute(1,2,0));

source

show_image_batch

 show_image_batch (b, show=<function show_titled_image>, items=9, cols=3,
                   figsize=None, **kwargs)

Display batch b in a grid of size items with cols width

show_image_batch(([Image.open(TEST_IMAGE_BW),Image.open(TEST_IMAGE)],['bw','color']), items=2)

Model init

source

requires_grad

 requires_grad (m)

Check if the first parameter of m requires grad or not

tst = nn.Linear(4,5)
assert requires_grad(tst)
for p in tst.parameters(): p.requires_grad_(False)
assert not requires_grad(tst)

source

init_default

 init_default (m, func=<function kaiming_normal_>)

Initialize m weights with func and set bias to 0.

tst = nn.Linear(4,5)
tst.weight.data.uniform_(-1,1)
tst.bias.data.uniform_(-1,1)
tst = init_default(tst, func = lambda x: x.data.fill_(1.))
test_eq(tst.weight, torch.ones(5,4))
test_eq(tst.bias, torch.zeros(5))

source

cond_init

 cond_init (m, func)

Apply init_default to m unless it’s a batchnorm module

tst = nn.Linear(4,5)
tst.weight.data.uniform_(-1,1)
tst.bias.data.uniform_(-1,1)
cond_init(tst, func = lambda x: x.data.fill_(1.))
test_eq(tst.weight, torch.ones(5,4))
test_eq(tst.bias, torch.zeros(5))

tst = nn.BatchNorm2d(5)
init = [tst.weight.clone(), tst.bias.clone()]
cond_init(tst, func = lambda x: x.data.fill_(1.))
test_eq(tst.weight, init[0])
test_eq(tst.bias, init[1])

source

apply_leaf

 apply_leaf (m, f)

Apply f to children of m.

tst = nn.Sequential(nn.Linear(4,5), nn.Sequential(nn.Linear(4,5), nn.Linear(4,5)))
apply_leaf(tst, partial(init_default, func=lambda x: x.data.fill_(1.)))
for l in [tst[0], *tst[1]]: test_eq(l.weight, torch.ones(5,4))
for l in [tst[0], *tst[1]]: test_eq(l.bias,   torch.zeros(5))

source

apply_init

 apply_init (m, func=<function kaiming_normal_>)

Initialize all non-batchnorm layers of m with func.

tst = nn.Sequential(nn.Linear(4,5), nn.Sequential(nn.Linear(4,5), nn.BatchNorm1d(5)))
init = [tst[1][1].weight.clone(), tst[1][1].bias.clone()]
apply_init(tst, func=lambda x: x.data.fill_(1.))
for l in [tst[0], tst[1][0]]: test_eq(l.weight, torch.ones(5,4))
for l in [tst[0], tst[1][0]]: test_eq(l.bias,   torch.zeros(5))
test_eq(tst[1][1].weight, init[0])
test_eq(tst[1][1].bias,   init[1])