Code stringlengths 103 85.9k | Summary listlengths 0 94 |
|---|---|
Please provide a description of the function:def _write_metrics(self, iteration:int, last_metrics:MetricsList, start_idx:int=2)->None:
"Writes training metrics to Tensorboard."
recorder = self.learn.recorder
for i, name in enumerate(recorder.names[start_idx:]):
if last_metrics is Non... | [] |
Please provide a description of the function:def on_batch_end(self, last_loss:Tensor, iteration:int, **kwargs)->None:
"Callback function that writes batch end appropriate data to Tensorboard."
if iteration == 0: return
self._update_batches_if_needed()
if iteration % self.loss_iters == 0:... | [] |
Please provide a description of the function:def on_backward_end(self, iteration:int, **kwargs)->None:
"Callback function that writes backward end appropriate data to Tensorboard."
if iteration == 0: return
self._update_batches_if_needed()
if iteration % self.stats_iters == 0: self._writ... | [] |
Please provide a description of the function:def on_epoch_end(self, last_metrics:MetricsList, iteration:int, **kwargs)->None:
"Callback function that writes epoch end appropriate data to Tensorboard."
self._write_metrics(iteration=iteration, last_metrics=last_metrics) | [] |
Please provide a description of the function:def _write_weight_histograms(self, iteration:int)->None:
"Writes model weight histograms to Tensorboard."
generator, critic = self.learn.gan_trainer.generator, self.learn.gan_trainer.critic
self.hist_writer.write(model=generator, iteration=iteration, ... | [] |
Please provide a description of the function:def _write_gen_model_stats(self, iteration:int)->None:
"Writes gradient statistics for generator to Tensorboard."
generator = self.learn.gan_trainer.generator
self.stats_writer.write(model=generator, iteration=iteration, tbwriter=self.tbwriter, name='... | [] |
Please provide a description of the function:def _write_critic_model_stats(self, iteration:int)->None:
"Writes gradient statistics for critic to Tensorboard."
critic = self.learn.gan_trainer.critic
self.stats_writer.write(model=critic, iteration=iteration, tbwriter=self.tbwriter, name='crit_mode... | [] |
Please provide a description of the function:def _write_model_stats(self, iteration:int)->None:
"Writes gradient statistics to Tensorboard."
# We don't want to write stats when model is not iterated on and hence has zeroed out gradients
gen_mode = self.learn.gan_trainer.gen_mode
if gen_m... | [] |
Please provide a description of the function:def _write_training_loss(self, iteration:int, last_loss:Tensor)->None:
"Writes training loss to Tensorboard."
recorder = self.learn.gan_trainer.recorder
if len(recorder.losses) == 0: return
scalar_value = to_np((recorder.losses[-1:])[0])
... | [] |
Please provide a description of the function:def _write_images(self, iteration:int)->None:
"Writes model generated, original and real images to Tensorboard."
trainer = self.learn.gan_trainer
#TODO: Switching gen_mode temporarily seems a bit hacky here. Certainly not a good side-effect. Is the... | [] |
Please provide a description of the function:def on_batch_end(self, iteration:int, **kwargs)->None:
"Callback function that writes batch end appropriate data to Tensorboard."
super().on_batch_end(iteration=iteration, **kwargs)
if iteration == 0: return
if iteration % self.visual_iters ==... | [] |
Please provide a description of the function:def on_backward_end(self, iteration:int, **kwargs)->None:
"Callback function that writes backward end appropriate data to Tensorboard."
if iteration == 0: return
self._update_batches_if_needed()
#TODO: This could perhaps be implemented as que... | [] |
Please provide a description of the function:def _write_images(self, iteration:int)->None:
"Writes model generated, original and real images to Tensorboard"
self.img_gen_vis.write(learn=self.learn, trn_batch=self.trn_batch, val_batch=self.val_batch, iteration=iteration,
t... | [] |
Please provide a description of the function:def request_write(self, request: TBWriteRequest)->None:
"Queues up an asynchronous write request to Tensorboard."
if self.stop_request.isSet(): return
self.queue.put(request) | [] |
Please provide a description of the function:def _queue_processor(self)->None:
"Processes queued up write requests asynchronously to Tensorboard."
while not self.stop_request.isSet():
while not self.queue.empty():
if self.stop_request.isSet(): return
request =... | [] |
Please provide a description of the function:def get_list_from_model(learn:Learner, ds_type:DatasetType, batch:Tuple)->[]:
"Factory method to convert a batch of model images to a list of ModelImageSet."
image_sets = []
x,y = batch[0],batch[1]
preds = learn.pred_batch(ds_type=ds_type, bat... | [] |
Please provide a description of the function:def _write_histogram(self, param_name:str, values)->None:
"Writes single model histogram to Tensorboard."
tag = self.name + '/weights/' + param_name
self.tbwriter.add_histogram(tag=tag, values=values, global_step=self.iteration) | [] |
Please provide a description of the function:def write(self)->None:
"Writes model histograms to Tensorboard."
for param_name, values in self.params: self._write_histogram(param_name=param_name, values=values) | [] |
Please provide a description of the function:def write(self, model:nn.Module, iteration:int, tbwriter:SummaryWriter, name:str='model')->None:
"Writes model histograms to Tensorboard."
request = HistogramTBRequest(model=model, iteration=iteration, tbwriter=tbwriter, name=name)
asyncTBWriter.reque... | [] |
Please provide a description of the function:def _add_gradient_scalar(self, name:str, scalar_value)->None:
"Writes a single scalar value for a gradient statistic to Tensorboard."
tag = self.name + '/gradients/' + name
self.tbwriter.add_scalar(tag=tag, scalar_value=scalar_value, global_step=self.... | [] |
Please provide a description of the function:def _write_avg_norm(self, norms:[])->None:
"Writes the average norm of the gradients to Tensorboard."
avg_norm = sum(norms)/len(self.gradients)
self._add_gradient_scalar('avg_norm', scalar_value=avg_norm) | [] |
Please provide a description of the function:def _write_median_norm(self, norms:[])->None:
"Writes the median norm of the gradients to Tensorboard."
median_norm = statistics.median(norms)
self._add_gradient_scalar('median_norm', scalar_value=median_norm) | [] |
Please provide a description of the function:def _write_max_norm(self, norms:[])->None:
"Writes the maximum norm of the gradients to Tensorboard."
max_norm = max(norms)
self._add_gradient_scalar('max_norm', scalar_value=max_norm) | [] |
Please provide a description of the function:def _write_min_norm(self, norms:[])->None:
"Writes the minimum norm of the gradients to Tensorboard."
min_norm = min(norms)
self._add_gradient_scalar('min_norm', scalar_value=min_norm) | [] |
Please provide a description of the function:def _write_num_zeros(self)->None:
"Writes the number of zeroes in the gradients to Tensorboard."
gradient_nps = [to_np(x.data) for x in self.gradients]
num_zeros = sum((np.asarray(x) == 0.0).sum() for x in gradient_nps)
self._add_gradient_scal... | [] |
Please provide a description of the function:def _write_avg_gradient(self)->None:
"Writes the average of the gradients to Tensorboard."
avg_gradient = sum(x.data.mean() for x in self.gradients)/len(self.gradients)
self._add_gradient_scalar('avg_gradient', scalar_value=avg_gradient) | [] |
Please provide a description of the function:def _write_median_gradient(self)->None:
"Writes the median of the gradients to Tensorboard."
median_gradient = statistics.median(x.data.median() for x in self.gradients)
self._add_gradient_scalar('median_gradient', scalar_value=median_gradient) | [] |
Please provide a description of the function:def _write_max_gradient(self)->None:
"Writes the maximum of the gradients to Tensorboard."
max_gradient = max(x.data.max() for x in self.gradients)
self._add_gradient_scalar('max_gradient', scalar_value=max_gradient) | [] |
Please provide a description of the function:def _write_min_gradient(self)->None:
"Writes the minimum of the gradients to Tensorboard."
min_gradient = min(x.data.min() for x in self.gradients)
self._add_gradient_scalar('min_gradient', scalar_value=min_gradient) | [] |
Please provide a description of the function:def write(self)->None:
"Writes model gradient statistics to Tensorboard."
if len(self.gradients) == 0: return
norms = [x.data.norm() for x in self.gradients]
self._write_avg_norm(norms=norms)
self._write_median_norm(norms=norms)
... | [] |
Please provide a description of the function:def _write_images(self, name:str, images:[Tensor])->None:
"Writes list of images as tensors to Tensorboard."
tag = self.ds_type.name + ' ' + name
self.tbwriter.add_image(tag=tag, img_tensor=vutils.make_grid(images, normalize=True), global_step=self.it... | [] |
Please provide a description of the function:def _get_image_tensors(self)->([Tensor], [Tensor], [Tensor]):
"Gets list of image tensors from lists of Image objects, as a tuple of original, generated and real(target) images."
orig_images, gen_images, real_images = [], [], []
for image_set in self.... | [] |
Please provide a description of the function:def write(self)->None:
"Writes original, generated and real(target) images to Tensorboard."
orig_images, gen_images, real_images = self._get_image_tensors()
self._write_images(name='orig images', images=orig_images)
self._write_images(name='ge... | [] |
Please provide a description of the function:def write(self, learn:Learner, trn_batch:Tuple, val_batch:Tuple, iteration:int, tbwriter:SummaryWriter)->None:
"Writes training and validation batch images to Tensorboard."
self._write_for_dstype(learn=learn, batch=val_batch, iteration=iteration, tbwriter=tbw... | [] |
Please provide a description of the function:def _write_for_dstype(self, learn:Learner, batch:Tuple, iteration:int, tbwriter:SummaryWriter, ds_type:DatasetType)->None:
"Writes batch images of specified DatasetType to Tensorboard."
request = ImageTBRequest(learn=learn, batch=batch, iteration=iteration, t... | [] |
Please provide a description of the function:def write(self)->None:
"Writes single model graph to Tensorboard."
self.tbwriter.add_graph(model=self.model, input_to_model=self.input_to_model) | [] |
Please provide a description of the function:def write(self, model:nn.Module, tbwriter:SummaryWriter, input_to_model:torch.Tensor)->None:
"Writes model graph to Tensorboard."
request = GraphTBRequest(model=model, tbwriter=tbwriter, input_to_model=input_to_model)
asyncTBWriter.request_write(reque... | [] |
Please provide a description of the function:def fix_batchnorm(swa_model, train_dl):
bn_modules = []
swa_model.apply(lambda module: collect_bn_modules(module, bn_modules))
if not bn_modules: return
swa_model.train()
for module in bn_modules:
module.running_mean = torch.zeros_like... | [
"\n During training, batch norm layers keep track of a running mean and\n variance of the previous layer's activations. Because the parameters\n of the SWA model are computed as the average of other models' parameters,\n the SWA model never sees the training data itself, and therefore has no\n opport... |
Please provide a description of the function:def repackage_var(h):
if IS_TORCH_04: return h.detach() if type(h) == torch.Tensor else tuple(repackage_var(v) for v in h)
else: return Variable(h.data) if type(h) == Variable else tuple(repackage_var(v) for v in h) | [
"Wraps h in new Variables, to detach them from their history."
] |
Please provide a description of the function:def get_language_model(n_tok, emb_sz, n_hid, n_layers, pad_token,
dropout=0.4, dropouth=0.3, dropouti=0.5, dropoute=0.1, wdrop=0.5, tie_weights=True, qrnn=False, bias=False):
rnn_enc = RNN_Encoder(n_tok, emb_sz, n_hid=n_hid, n_layers=n_layers, pad_t... | [
"Returns a SequentialRNN model.\n\n A RNN_Encoder layer is instantiated using the parameters provided.\n\n This is followed by the creation of a LinearDecoder layer.\n\n Also by default (i.e. tie_weights = True), the embedding matrix used in the RNN_Encoder\n is used to instantiate the weights for the ... |
Please provide a description of the function:def forward(self, input):
sl,bs = input.size()
if bs!=self.bs:
self.bs=bs
self.reset()
with set_grad_enabled(self.training):
emb = self.encoder_with_dropout(input, dropout=self.dropoute if self.training els... | [
" Invoked during the forward propagation of the RNN_Encoder module.\n Args:\n input (Tensor): input of shape (sentence length x batch_size)\n\n Returns:\n raw_outputs (tuple(list (Tensor), list(Tensor)): list of tensors evaluated from each RNN layer without using\n dro... |
Please provide a description of the function:def replace_rep(t:str) -> str:
"Replace repetitions at the character level in `t`."
def _replace_rep(m:Collection[str]) -> str:
c,cc = m.groups()
return f' {TK_REP} {len(cc)+1} {c} '
re_rep = re.compile(r'(\S)(\1{3,})')
return re_rep.sub(_repl... | [] |
Please provide a description of the function:def replace_wrep(t:str) -> str:
"Replace word repetitions in `t`."
def _replace_wrep(m:Collection[str]) -> str:
c,cc = m.groups()
return f' {TK_WREP} {len(cc.split())+1} {c} '
re_wrep = re.compile(r'(\b\w+\W+)(\1{3,})')
return re_wrep.sub(_rep... | [] |
Please provide a description of the function:def fix_html(x:str) -> str:
"List of replacements from html strings in `x`."
re1 = re.compile(r' +')
x = x.replace('#39;', "'").replace('amp;', '&').replace('#146;', "'").replace(
'nbsp;', ' ').replace('#36;', '$').replace('\\n', "\n").replace('quot;', "... | [] |
Please provide a description of the function:def replace_all_caps(x:Collection[str]) -> Collection[str]:
"Replace tokens in ALL CAPS in `x` by their lower version and add `TK_UP` before."
res = []
for t in x:
if t.isupper() and len(t) > 1: res.append(TK_UP); res.append(t.lower())
else: res.a... | [] |
Please provide a description of the function:def deal_caps(x:Collection[str]) -> Collection[str]:
"Replace all Capitalized tokens in `x` by their lower version and add `TK_MAJ` before."
res = []
for t in x:
if t == '': continue
if t[0].isupper() and len(t) > 1 and t[1:].islower(): res.append... | [] |
Please provide a description of the function:def process_text(self, t:str, tok:BaseTokenizer) -> List[str]:
"Process one text `t` with tokenizer `tok`."
for rule in self.pre_rules: t = rule(t)
toks = tok.tokenizer(t)
for rule in self.post_rules: toks = rule(toks)
return toks | [] |
Please provide a description of the function:def _process_all_1(self, texts:Collection[str]) -> List[List[str]]:
"Process a list of `texts` in one process."
tok = self.tok_func(self.lang)
if self.special_cases: tok.add_special_cases(self.special_cases)
return [self.process_text(str(t), t... | [] |
Please provide a description of the function:def process_all(self, texts:Collection[str]) -> List[List[str]]:
"Process a list of `texts`."
if self.n_cpus <= 1: return self._process_all_1(texts)
with ProcessPoolExecutor(self.n_cpus) as e:
return sum(e.map(self._process_all_1, partitio... | [] |
Please provide a description of the function:def numericalize(self, t:Collection[str]) -> List[int]:
"Convert a list of tokens `t` to their ids."
return [self.stoi[w] for w in t] | [] |
Please provide a description of the function:def textify(self, nums:Collection[int], sep=' ') -> List[str]:
"Convert a list of `nums` to their tokens."
return sep.join([self.itos[i] for i in nums]) if sep is not None else [self.itos[i] for i in nums] | [] |
Please provide a description of the function:def create(cls, tokens:Tokens, max_vocab:int, min_freq:int) -> 'Vocab':
"Create a vocabulary from a set of `tokens`."
freq = Counter(p for o in tokens for p in o)
itos = [o for o,c in freq.most_common(max_vocab) if c >= min_freq]
for o in reve... | [] |
Please provide a description of the function:def load(cls, path):
"Load the `Vocab` contained in `path`"
itos = pickle.load(open(path, 'rb'))
return cls(itos) | [] |
Please provide a description of the function:def plot_loss(self, n_skip=10, n_skip_end=5):
'''
plots loss function as function of iterations.
When used in Jupyternotebook, plot will be displayed in notebook. Else, plot will be displayed in console and both plot and loss are saved in save_path. ... | [] |
Please provide a description of the function:def plot_lr(self):
'''Plots learning rate in jupyter notebook or console, depending on the enviroment of the learner.'''
if not in_ipynb():
plt.switch_backend('agg')
if self.record_mom:
fig, axs = plt.subplots(1,2,figsize=(12,4... | [] |
Please provide a description of the function:def plot(self, n_skip=10, n_skip_end=5):
'''
Plots the loss function with respect to learning rate, in log scale.
'''
plt.ylabel("validation loss")
plt.xlabel("learning rate (log scale)")
plt.plot(self.lrs[n_skip:-(n_skip_end+... | [] |
Please provide a description of the function:def plot_lr(self, show_text=True, show_moms=True):
phase_limits = [0]
for nb_batch, phase in zip(self.nb_batches, self.phases):
phase_limits.append(phase_limits[-1] + nb_batch * phase.epochs)
if not in_ipynb():
plt.swi... | [
"Plots the lr rate/momentum schedule"
] |
Please provide a description of the function:def main(
gpu:Param("GPU to run on", str)=None,
woof: Param("Use imagewoof (otherwise imagenette)", int)=0,
lr: Param("Learning rate", float)=1e-3,
size: Param("Size (px: 128,192,224)", int)=128,
alpha: Param("Alpha", float)=0.99,
... | [] |
Please provide a description of the function:def on_batch_end(self, last_loss, epoch, num_batch, **kwargs:Any)->None:
"Test if `last_loss` is NaN and interrupts training."
if self.stop: return True #to skip validation after stopping during training
if torch.isnan(last_loss):
print (f... | [] |
Please provide a description of the function:def on_train_begin(self, **kwargs:Any)->None:
"Initializes the best value."
self.best = float('inf') if self.operator == np.less else -float('inf') | [] |
Please provide a description of the function:def get_monitor_value(self):
"Pick the monitored value."
if self.monitor=='trn_loss' and len(self.learn.recorder.losses) == 0: return None
elif len(self.learn.recorder.val_losses) == 0: return None
values = {'train_loss':self.learn.recorder.lo... | [] |
Please provide a description of the function:def on_epoch_end(self, epoch:int, **kwargs:Any)->None:
"Compare the value monitored to its best score and maybe save the model."
if self.every=="epoch": self.learn.save(f'{self.name}_{epoch}')
else: #every="improvement"
current = self.get_... | [] |
Please provide a description of the function:def on_train_end(self, **kwargs):
"Load the best model."
if self.every=="improvement" and (self.learn.path/f'{self.learn.model_dir}/{self.name}.pth').is_file():
self.learn.load(f'{self.name}', purge=False) | [] |
Please provide a description of the function:def on_train_begin(self, **kwargs:Any)->None:
"Initialize inner arguments."
self.wait, self.opt = 0, self.learn.opt
super().on_train_begin(**kwargs) | [] |
Please provide a description of the function:def on_epoch_end(self, epoch, **kwargs:Any)->None:
"Compare the value monitored to its best and maybe reduce lr."
current = self.get_monitor_value()
if current is None: return
if self.operator(current - self.min_delta, self.best): self.best,se... | [] |
Please provide a description of the function:def convert_nb(fname, dest_path='.'):
"Convert a notebook `fname` to html file in `dest_path`."
from .gen_notebooks import remove_undoc_cells, remove_code_cell_jupyter_widget_state_elem
nb = read_nb(fname)
nb['cells'] = remove_undoc_cells(nb['cells'])
nb[... | [] |
Please provide a description of the function:def convert_all(folder, dest_path='.', force_all=False):
"Convert modified notebooks in `folder` to html pages in `dest_path`."
path = Path(folder)
changed_cnt = 0
for fname in path.glob("*.ipynb"):
# only rebuild modified files
fname_out = P... | [] |
Please provide a description of the function:def pad_collate(samples:BatchSamples, pad_idx:int=1, pad_first:bool=True, backwards:bool=False) -> Tuple[LongTensor, LongTensor]:
"Function that collect samples and adds padding. Flips token order if needed"
samples = to_data(samples)
max_len = max([len(s[0]) for... | [] |
Please provide a description of the function:def open_text(fn:PathOrStr, enc='utf-8'):
"Read the text in `fn`."
with open(fn,'r', encoding = enc) as f: return ''.join(f.readlines()) | [] |
Please provide a description of the function:def allocate_buffers(self):
"Create the ragged array that will be filled when we ask for items."
if self.ite_len is None: len(self)
self.idx = LanguageModelPreLoader.CircularIndex(len(self.dataset.x.items), not self.backwards)
self.batch = n... | [] |
Please provide a description of the function:def fill_row(self, forward, items, idx, row, ro, ri, overlap,lengths):
"Fill the row with tokens from the ragged array. --OBS-- overlap != 1 has not been implemented"
ibuf = n = 0
ro -= 1
while ibuf < row.size:
ro += 1
... | [] |
Please provide a description of the function:def from_ids(cls, path:PathOrStr, vocab:Vocab, train_ids:Collection[Collection[int]], valid_ids:Collection[Collection[int]],
test_ids:Collection[Collection[int]]=None, train_lbls:Collection[Union[int,float]]=None,
valid_lbls:Collection[Union... | [] |
Please provide a description of the function:def load(cls, path:PathOrStr, cache_name:PathOrStr='tmp', processor:PreProcessor=None, **kwargs):
"Load a `TextDataBunch` from `path/cache_name`. `kwargs` are passed to the dataloader creation."
warn(, DeprecationWarning)
cache_path = Path(path)/cache... | [
"This method is deprecated and only kept to load data serialized in v1.0.43 or earlier.\n Use `load_data` for data saved with v1.0.44 or later."
] |
Please provide a description of the function:def from_tokens(cls, path:PathOrStr, trn_tok:Collection[Collection[str]], trn_lbls:Collection[Union[int,float]],
val_tok:Collection[Collection[str]], val_lbls:Collection[Union[int,float]], vocab:Vocab=None,
tst_tok:Collection[Collection[str]... | [] |
Please provide a description of the function:def from_df(cls, path:PathOrStr, train_df:DataFrame, valid_df:DataFrame, test_df:Optional[DataFrame]=None,
tokenizer:Tokenizer=None, vocab:Vocab=None, classes:Collection[str]=None, text_cols:IntsOrStrs=1,
label_cols:IntsOrStrs=0, label_delim:s... | [] |
Please provide a description of the function:def from_csv(cls, path:PathOrStr, csv_name, valid_pct:float=0.2, test:Optional[str]=None,
tokenizer:Tokenizer=None, vocab:Vocab=None, classes:Collection[str]=None, delimiter:str=None, header='infer',
text_cols:IntsOrStrs=1, label_cols:IntsOr... | [] |
Please provide a description of the function:def from_folder(cls, path:PathOrStr, train:str='train', valid:str='valid', test:Optional[str]=None,
classes:Collection[Any]=None, tokenizer:Tokenizer=None, vocab:Vocab=None, chunksize:int=10000, max_vocab:int=60000,
min_freq:int=2, mar... | [] |
Please provide a description of the function:def create(cls, train_ds, valid_ds, test_ds=None, path:PathOrStr='.', no_check:bool=False, bs=64, val_bs:int=None,
num_workers:int=0, device:torch.device=None, collate_fn:Callable=data_collate,
dl_tfms:Optional[Collection[Callable]]=None, bptt:i... | [] |
Please provide a description of the function:def create(cls, train_ds, valid_ds, test_ds=None, path:PathOrStr='.', bs:int=32, val_bs:int=None, pad_idx=1,
pad_first=True, device:torch.device=None, no_check:bool=False, backwards:bool=False, **dl_kwargs) -> DataBunch:
"Function that transform the `d... | [] |
Please provide a description of the function:def label_for_lm(self, **kwargs):
"A special labelling method for language models."
self.__class__ = LMTextList
kwargs['label_cls'] = LMLabelList
return self.label_const(0, **kwargs) | [] |
Please provide a description of the function:def from_folder(cls, path:PathOrStr='.', extensions:Collection[str]=text_extensions, vocab:Vocab=None,
processor:PreProcessor=None, **kwargs)->'TextList':
"Get the list of files in `path` that have a text suffix. `recurse` determines if we search ... | [] |
Please provide a description of the function:def show_xys(self, xs, ys, max_len:int=70)->None:
"Show the `xs` (inputs) and `ys` (targets). `max_len` is the maximum number of tokens displayed."
from IPython.display import display, HTML
names = ['idx','text'] if self._is_lm else ['text','target']
... | [] |
Please provide a description of the function:def inceptionv4(pretrained=True):
r
model = InceptionV4()
if pretrained:
model.load_state_dict(model_zoo.load_url(model_urls['imagenet']))
return model | [
"InceptionV4 model architecture from the\n `\"Inception-v4, Inception-ResNet...\" <https://arxiv.org/abs/1602.07261>`_ paper.\n\n Args:\n pretrained (bool): If True, returns a model pre-trained on ImageNet\n "
] |
Please provide a description of the function:def predict_array(self, arr):
precompute = self.precompute
self.precompute = False
pred = super().predict_array(arr)
self.precompute = precompute
return pred | [
"\n This over-ride is necessary because otherwise the learner method accesses the wrong model when it is called\n with precompute set to true\n\n Args:\n arr: a numpy array to be used as input to the model for prediction purposes\n Returns:\n a numpy array containin... |
Please provide a description of the function:def main( gpu:Param("GPU to run on", str)=None ):
gpu = setup_distrib(gpu)
n_gpus = num_distrib()
path = url2path(URLs.CIFAR)
ds_tfms = ([*rand_pad(4, 32), flip_lr(p=0.5)], [])
workers = min(16, num_cpus()//n_gpus)
data = ImageDataBunch.from_fold... | [
"Distrubuted training of CIFAR-10.\n Fastest speed is if you run as follows:\n python -m fastai.launch train_cifar.py"
] |
Please provide a description of the function:def initialize_options(self):
self.dep_groups = ''
self.dep_quote = False
self.dep_conda = False | [
"Set default values for options."
] |
Please provide a description of the function:def run(self):
wanted_groups = self.parse()
deps = []
invalid_groups = []
for grp in wanted_groups:
if grp in dep_groups: deps.extend(dep_groups[grp])
else: invalid_groups.append(grp)
... | [
"Run command."
] |
Please provide a description of the function:def get_sfs_idxs(sfs, last=True):
if last:
feature_szs = [sfs_feats.features.size()[-1] for sfs_feats in sfs]
sfs_idxs = list(np.where(np.array(feature_szs[:-1]) != np.array(feature_szs[1:]))[0])
if feature_szs[0] != feature_szs[1]: sfs_idxs ... | [
"\n Return the saved feature indexes that will be concatenated\n Inputs:\n sfs (list): saved features by hook function, in other words intermediate activations\n last (bool): whether to concatenate only last different activation, or all from the encoder model\n "
] |
Please provide a description of the function:def hook_output (module:nn.Module, detach:bool=True, grad:bool=False)->Hook:
"Return a `Hook` that stores activations of `module` in `self.stored`"
return Hook(module, _hook_inner, detach=detach, is_forward=not grad) | [] |
Please provide a description of the function:def hook_outputs(modules:Collection[nn.Module], detach:bool=True, grad:bool=False)->Hooks:
"Return `Hooks` that store activations of all `modules` in `self.stored`"
return Hooks(modules, _hook_inner, detach=detach, is_forward=not grad) | [] |
Please provide a description of the function:def dummy_batch(m: nn.Module, size:tuple=(64,64))->Tensor:
"Create a dummy batch to go through `m` with `size`."
ch_in = in_channels(m)
return one_param(m).new(1, ch_in, *size).requires_grad_(False).uniform_(-1.,1.) | [] |
Please provide a description of the function:def dummy_eval(m:nn.Module, size:tuple=(64,64)):
"Pass a `dummy_batch` in evaluation mode in `m` with `size`."
return m.eval()(dummy_batch(m, size)) | [] |
Please provide a description of the function:def model_sizes(m:nn.Module, size:tuple=(64,64))->Tuple[Sizes,Tensor,Hooks]:
"Pass a dummy input through the model `m` to get the various sizes of activations."
with hook_outputs(m) as hooks:
x = dummy_eval(m, size)
return [o.stored.shape for o in hoo... | [] |
Please provide a description of the function:def num_features_model(m:nn.Module)->int:
"Return the number of output features for `model`."
sz = 64
while True:
try: return model_sizes(m, size=(sz,sz))[-1][1]
except Exception as e:
sz *= 2
if sz > 2048: raise | [] |
Please provide a description of the function:def params_size(m: Union[nn.Module,Learner], size: tuple = (3, 64, 64))->Tuple[Sizes, Tensor, Hooks]:
"Pass a dummy input through the model to get the various sizes. Returns (res,x,hooks) if `full`"
if isinstance(m, Learner):
if m.data.is_empty:
r... | [] |
Please provide a description of the function:def model_summary(m:Learner, n:int=70):
"Print a summary of `m` using a output text width of `n` chars"
info = layers_info(m)
header = ["Layer (type)", "Output Shape", "Param #", "Trainable"]
res = "=" * n + "\n"
res += f"{header[0]:<20} {header[1]:<20} {... | [] |
Please provide a description of the function:def hook_fn(self, module:nn.Module, input:Tensors, output:Tensors):
"Applies `hook_func` to `module`, `input`, `output`."
if self.detach:
input = (o.detach() for o in input ) if is_listy(input ) else input.detach()
output = (o.detach(... | [] |
Please provide a description of the function:def remove(self):
"Remove the hook from the model."
if not self.removed:
self.hook.remove()
self.removed=True | [] |
Please provide a description of the function:def on_train_begin(self, **kwargs):
"Register the `Hooks` on `self.modules`."
if not self.modules:
self.modules = [m for m in flatten_model(self.learn.model)
if hasattr(m, 'weight')]
self.hooks = Hooks(self.modu... | [] |
Please provide a description of the function:def hook(self, m:nn.Module, i:Tensors, o:Tensors)->Tuple[Rank0Tensor,Rank0Tensor]:
"Take the mean and std of `o`."
return o.mean().item(),o.std().item() | [] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.