code stringlengths 82 54.1k | code_codestyle int64 0 699 | style_context stringlengths 111 35.6k | style_context_codestyle int64 0 699 | label int64 0 1 |
|---|---|---|---|---|
from abc import ABC, abstractmethod
from argparse import ArgumentParser
class lowerCamelCase (a_ ):
'''simple docstring'''
@staticmethod
@abstractmethod
def __UpperCAmelCase ( _UpperCamelCase ) -> Union[str, Any]:
raise NotImplementedError()
@abstractmethod
def __UpperCAmelCase ( self ) -> Any:
raise NotImplementedError()
| 406 |
"""simple docstring"""
import socket
def _lowerCAmelCase ( ):
'''simple docstring'''
UpperCAmelCase = socket.socket(socket.AF_INET , socket.SOCK_STREAM )
UpperCAmelCase = socket.gethostname()
UpperCAmelCase = 12312
sock.connect((host, port) )
sock.send(b"""Hello server!""" )
with open("""Received_file""" , """wb""" ) as out_file:
print("""File opened""" )
print("""Receiving data...""" )
while True:
UpperCAmelCase = sock.recv(1024 )
if not data:
break
out_file.write(lowerCAmelCase )
print("""Successfully received the file""" )
sock.close()
print("""Connection closed""" )
if __name__ == "__main__":
main()
| 673 | 0 |
"""simple docstring"""
import fcntl
import os
import socket
import torch
import torch.distributed as dist
def __a ( *A ) -> Tuple:
'''simple docstring'''
with open(A , "r" ) as fh:
fcntl.flock(A , fcntl.LOCK_EX )
try:
print(*A )
finally:
fcntl.flock(A , fcntl.LOCK_UN )
__UpperCAmelCase =int(os.environ["""LOCAL_RANK"""])
torch.cuda.set_device(local_rank)
__UpperCAmelCase =torch.device("""cuda""", local_rank)
__UpperCAmelCase =socket.gethostname()
__UpperCAmelCase =F'''[{hostname}-{local_rank}]'''
try:
# test distributed
dist.init_process_group("""nccl""")
dist.all_reduce(torch.ones(1).to(device), op=dist.ReduceOp.SUM)
dist.barrier()
# test cuda is available and can allocate memory
torch.cuda.is_available()
torch.ones(1).cuda(local_rank)
# global rank
__UpperCAmelCase =dist.get_rank()
__UpperCAmelCase =dist.get_world_size()
printflock(F'''{gpu} is OK (global rank: {rank}/{world_size})''')
dist.barrier()
if rank == 0:
printflock(F'''pt={torch.__version__}, cuda={torch.version.cuda}, nccl={torch.cuda.nccl.version()}''')
except Exception:
printflock(F'''{gpu} is broken''')
raise | 337 |
"""simple docstring"""
import math
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
return math.sqrt(lowerCAmelCase ) * math.sqrt(lowerCAmelCase ) == num
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = 0
UpperCAmelCase = n
while left <= right:
UpperCAmelCase = (left + right) // 2
if mid**2 == n:
return True
elif mid**2 > n:
UpperCAmelCase = mid - 1
else:
UpperCAmelCase = mid + 1
return False
if __name__ == "__main__":
import doctest
doctest.testmod()
| 673 | 0 |
'''simple docstring'''
import unittest
from transformers import AutoTokenizer, NystromformerConfig, is_torch_available
from transformers.testing_utils import require_torch, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
NystromformerForMaskedLM,
NystromformerForMultipleChoice,
NystromformerForQuestionAnswering,
NystromformerForSequenceClassification,
NystromformerForTokenClassification,
NystromformerModel,
)
from transformers.models.nystromformer.modeling_nystromformer import NYSTROMFORMER_PRETRAINED_MODEL_ARCHIVE_LIST
class a__ :
def __init__( self : Union[str, Any] , a : int , a : Tuple=13 , a : Union[str, Any]=7 , a : Optional[Any]=True , a : List[Any]=True , a : int=True , a : int=True , a : List[str]=99 , a : Any=32 , a : Union[str, Any]=5 , a : Optional[Any]=4 , a : Optional[int]=37 , a : Any="gelu" , a : Optional[int]=0.1 , a : Optional[int]=0.1 , a : List[Any]=5_12 , a : Union[str, Any]=16 , a : Dict=2 , a : List[Any]=0.02 , a : List[str]=3 , a : Tuple=4 , a : Tuple=None , ):
"""simple docstring"""
__lowerCamelCase = parent
__lowerCamelCase = batch_size
__lowerCamelCase = seq_length
__lowerCamelCase = is_training
__lowerCamelCase = use_input_mask
__lowerCamelCase = use_token_type_ids
__lowerCamelCase = use_labels
__lowerCamelCase = vocab_size
__lowerCamelCase = hidden_size
__lowerCamelCase = num_hidden_layers
__lowerCamelCase = num_attention_heads
__lowerCamelCase = intermediate_size
__lowerCamelCase = hidden_act
__lowerCamelCase = hidden_dropout_prob
__lowerCamelCase = attention_probs_dropout_prob
__lowerCamelCase = max_position_embeddings
__lowerCamelCase = type_vocab_size
__lowerCamelCase = type_sequence_label_size
__lowerCamelCase = initializer_range
__lowerCamelCase = num_labels
__lowerCamelCase = num_choices
__lowerCamelCase = scope
def SCREAMING_SNAKE_CASE__ ( self : Dict ):
"""simple docstring"""
__lowerCamelCase = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__lowerCamelCase = None
if self.use_input_mask:
__lowerCamelCase = random_attention_mask([self.batch_size, self.seq_length] )
__lowerCamelCase = None
if self.use_token_type_ids:
__lowerCamelCase = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
__lowerCamelCase = None
__lowerCamelCase = None
__lowerCamelCase = None
if self.use_labels:
__lowerCamelCase = ids_tensor([self.batch_size] , self.type_sequence_label_size )
__lowerCamelCase = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
__lowerCamelCase = ids_tensor([self.batch_size] , self.num_choices )
__lowerCamelCase = self.get_config()
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def SCREAMING_SNAKE_CASE__ ( self : List[str] ):
"""simple docstring"""
return NystromformerConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=snake_case__ , initializer_range=self.initializer_range , )
def SCREAMING_SNAKE_CASE__ ( self : Tuple , a : Any , a : Tuple , a : int , a : Any , a : List[Any] , a : Any , a : Optional[Any] ):
"""simple docstring"""
__lowerCamelCase = NystromformerModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__lowerCamelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ )
__lowerCamelCase = model(snake_case__ , token_type_ids=snake_case__ )
__lowerCamelCase = model(snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
def SCREAMING_SNAKE_CASE__ ( self : List[Any] , a : Optional[int] , a : List[str] , a : Tuple , a : str , a : Any , a : int , a : Tuple ):
"""simple docstring"""
__lowerCamelCase = NystromformerForMaskedLM(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__lowerCamelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def SCREAMING_SNAKE_CASE__ ( self : List[str] , a : Dict , a : List[str] , a : str , a : int , a : str , a : List[Any] , a : Optional[Any] ):
"""simple docstring"""
__lowerCamelCase = NystromformerForQuestionAnswering(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__lowerCamelCase = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , start_positions=snake_case__ , end_positions=snake_case__ , )
self.parent.assertEqual(result.start_logits.shape , (self.batch_size, self.seq_length) )
self.parent.assertEqual(result.end_logits.shape , (self.batch_size, self.seq_length) )
def SCREAMING_SNAKE_CASE__ ( self : List[str] , a : Dict , a : Optional[int] , a : int , a : Tuple , a : Dict , a : Optional[int] , a : Any ):
"""simple docstring"""
__lowerCamelCase = self.num_labels
__lowerCamelCase = NystromformerForSequenceClassification(snake_case__ )
model.to(snake_case__ )
model.eval()
__lowerCamelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def SCREAMING_SNAKE_CASE__ ( self : str , a : Optional[int] , a : int , a : Optional[int] , a : Optional[int] , a : Optional[int] , a : str , a : Union[str, Any] ):
"""simple docstring"""
__lowerCamelCase = self.num_labels
__lowerCamelCase = NystromformerForTokenClassification(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__lowerCamelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def SCREAMING_SNAKE_CASE__ ( self : Optional[Any] , a : Optional[Any] , a : List[Any] , a : Dict , a : List[Any] , a : Union[str, Any] , a : List[Any] , a : List[Any] ):
"""simple docstring"""
__lowerCamelCase = self.num_choices
__lowerCamelCase = NystromformerForMultipleChoice(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__lowerCamelCase = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
__lowerCamelCase = token_type_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
__lowerCamelCase = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
__lowerCamelCase = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def SCREAMING_SNAKE_CASE__ ( self : Optional[Any] ):
"""simple docstring"""
__lowerCamelCase = self.prepare_config_and_inputs()
(
(
__lowerCamelCase
) , (
__lowerCamelCase
) , (
__lowerCamelCase
) , (
__lowerCamelCase
) , (
__lowerCamelCase
) , (
__lowerCamelCase
) , (
__lowerCamelCase
) ,
) = config_and_inputs
__lowerCamelCase = {'''input_ids''': input_ids, '''token_type_ids''': token_type_ids, '''attention_mask''': input_mask}
return config, inputs_dict
@require_torch
class a__ ( a_ , a_ , unittest.TestCase ):
lowerCamelCase : Optional[Any] =(
(
NystromformerModel,
NystromformerForMaskedLM,
NystromformerForMultipleChoice,
NystromformerForQuestionAnswering,
NystromformerForSequenceClassification,
NystromformerForTokenClassification,
)
if is_torch_available()
else ()
)
lowerCamelCase : Optional[Any] =(
{
'feature-extraction': NystromformerModel,
'fill-mask': NystromformerForMaskedLM,
'question-answering': NystromformerForQuestionAnswering,
'text-classification': NystromformerForSequenceClassification,
'token-classification': NystromformerForTokenClassification,
'zero-shot': NystromformerForSequenceClassification,
}
if is_torch_available()
else {}
)
lowerCamelCase : int =False
lowerCamelCase : Dict =False
def SCREAMING_SNAKE_CASE__ ( self : List[Any] ):
"""simple docstring"""
__lowerCamelCase = NystromformerModelTester(self )
__lowerCamelCase = ConfigTester(self , config_class=snake_case__ , hidden_size=37 )
def SCREAMING_SNAKE_CASE__ ( self : List[str] ):
"""simple docstring"""
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE__ ( self : Any ):
"""simple docstring"""
__lowerCamelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def SCREAMING_SNAKE_CASE__ ( self : Dict ):
"""simple docstring"""
__lowerCamelCase = self.model_tester.prepare_config_and_inputs()
for type in ["absolute", "relative_key", "relative_key_query"]:
__lowerCamelCase = type
self.model_tester.create_and_check_model(*snake_case__ )
def SCREAMING_SNAKE_CASE__ ( self : Optional[Any] ):
"""simple docstring"""
__lowerCamelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_lm(*snake_case__ )
def SCREAMING_SNAKE_CASE__ ( self : Optional[int] ):
"""simple docstring"""
__lowerCamelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_multiple_choice(*snake_case__ )
def SCREAMING_SNAKE_CASE__ ( self : List[str] ):
"""simple docstring"""
__lowerCamelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_question_answering(*snake_case__ )
def SCREAMING_SNAKE_CASE__ ( self : Optional[Any] ):
"""simple docstring"""
__lowerCamelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_sequence_classification(*snake_case__ )
def SCREAMING_SNAKE_CASE__ ( self : Tuple ):
"""simple docstring"""
__lowerCamelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_token_classification(*snake_case__ )
@slow
def SCREAMING_SNAKE_CASE__ ( self : List[Any] ):
"""simple docstring"""
for model_name in NYSTROMFORMER_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__lowerCamelCase = NystromformerModel.from_pretrained(snake_case__ )
self.assertIsNotNone(snake_case__ )
@require_torch
class a__ ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE__ ( self : List[str] ):
"""simple docstring"""
__lowerCamelCase = NystromformerModel.from_pretrained('''uw-madison/nystromformer-512''' )
__lowerCamelCase = torch.tensor([[0, 1, 2, 3, 4, 5]] )
with torch.no_grad():
__lowerCamelCase = model(snake_case__ )[0]
__lowerCamelCase = torch.Size((1, 6, 7_68) )
self.assertEqual(output.shape , snake_case__ )
__lowerCamelCase = torch.tensor(
[[[-0.45_32, -0.09_36, 0.51_37], [-0.26_76, 0.06_28, 0.61_86], [-0.36_29, -0.17_26, 0.47_16]]] )
self.assertTrue(torch.allclose(output[:, :3, :3] , snake_case__ , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE__ ( self : Union[str, Any] ):
"""simple docstring"""
__lowerCamelCase = '''the [MASK] of Belgium is Brussels'''
__lowerCamelCase = AutoTokenizer.from_pretrained('''uw-madison/nystromformer-512''' )
__lowerCamelCase = NystromformerForMaskedLM.from_pretrained('''uw-madison/nystromformer-512''' )
__lowerCamelCase = tokenizer(snake_case__ , return_tensors='''pt''' )
with torch.no_grad():
__lowerCamelCase = model(encoding.input_ids ).logits
__lowerCamelCase = token_logits[:, 2, :].argmax(-1 )[0]
self.assertEqual(tokenizer.decode(snake_case__ ) , '''capital''' )
| 546 |
"""simple docstring"""
import functools
import gc
import inspect
import torch
from .imports import is_npu_available, is_xpu_available
def _lowerCAmelCase ( *lowerCAmelCase ):
'''simple docstring'''
if not isinstance(lowerCAmelCase , lowerCAmelCase ):
UpperCAmelCase = list(lowerCAmelCase )
for i in range(len(lowerCAmelCase ) ):
UpperCAmelCase = None
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
return objects
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = [
"""CUDA out of memory.""", # CUDA OOM
"""cuDNN error: CUDNN_STATUS_NOT_SUPPORTED.""", # CUDNN SNAFU
"""DefaultCPUAllocator: can't allocate memory""", # CPU OOM
]
if isinstance(lowerCAmelCase , lowerCAmelCase ) and len(exception.args ) == 1:
return any(err in exception.args[0] for err in _statements )
return False
def _lowerCAmelCase ( lowerCAmelCase = None , lowerCAmelCase = 128 ):
'''simple docstring'''
if function is None:
return functools.partial(lowerCAmelCase , starting_batch_size=lowerCAmelCase )
UpperCAmelCase = starting_batch_size
def decorator(*lowerCAmelCase , **lowerCAmelCase ):
nonlocal batch_size
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
UpperCAmelCase = list(inspect.signature(lowerCAmelCase ).parameters.keys() )
# Guard against user error
if len(lowerCAmelCase ) < (len(lowerCAmelCase ) + 1):
UpperCAmelCase = """, """.join([F'''{arg}={value}''' for arg, value in zip(params[1:] , args[1:] )] )
raise TypeError(
F'''Batch size was passed into `{function.__name__}` as the first argument when called.'''
F'''Remove this as the decorator already does so: `{function.__name__}({arg_str})`''' )
while True:
if batch_size == 0:
raise RuntimeError("""No executable batch size found, reached zero.""" )
try:
return function(lowerCAmelCase , *lowerCAmelCase , **lowerCAmelCase )
except Exception as e:
if should_reduce_batch_size(lowerCAmelCase ):
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
batch_size //= 2
else:
raise
return decorator
| 673 | 0 |
import warnings
from .state import AcceleratorState, GradientState
warnings.filterwarnings('ignore', category=UserWarning, module='torch.optim.lr_scheduler')
class A :
def __init__( self : str , lowercase_ : Optional[Any] , lowercase_ : List[str] , lowercase_ : Tuple = True , lowercase_ : Optional[Any] = False ) -> int:
"""simple docstring"""
_lowerCamelCase : List[Any] =scheduler
_lowerCamelCase : Dict =optimizers if isinstance(snake_case__ , (list, tuple) ) else [optimizers]
_lowerCamelCase : List[str] =split_batches
_lowerCamelCase : Tuple =step_with_optimizer
_lowerCamelCase : Any =GradientState()
def lowerCamelCase ( self : Any , *lowercase_ : Optional[int] , **lowercase_ : int ) -> Dict:
"""simple docstring"""
if not self.step_with_optimizer:
# No link between scheduler and optimizer -> just step
self.scheduler.step(*snake_case__ , **snake_case__ )
return
# Otherwise, first make sure the optimizer was stepped.
if not self.gradient_state.sync_gradients:
if self.gradient_state.adjust_scheduler:
self.scheduler._step_count += 1
return
for opt in self.optimizers:
if opt.step_was_skipped:
return
if self.split_batches:
# Split batches -> the training dataloader batch size is not changed so one step per training step
self.scheduler.step(*snake_case__ , **snake_case__ )
else:
# Otherwise the training dataloader batch size was multiplied by `num_processes`, so we need to do
# num_processes steps per training step
_lowerCamelCase : Optional[Any] =AcceleratorState().num_processes
for _ in range(snake_case__ ):
# Special case when using OneCycle and `drop_last` was not used
if hasattr(self.scheduler , 'total_steps' ):
if self.scheduler._step_count <= self.scheduler.total_steps:
self.scheduler.step(*snake_case__ , **snake_case__ )
else:
self.scheduler.step(*snake_case__ , **snake_case__ )
def lowerCamelCase ( self : List[str] ) -> Union[str, Any]:
"""simple docstring"""
return self.scheduler.get_last_lr()
def lowerCamelCase ( self : Union[str, Any] ) -> Optional[int]:
"""simple docstring"""
return self.scheduler.state_dict()
def lowerCamelCase ( self : Tuple , lowercase_ : Optional[int] ) -> List[str]:
"""simple docstring"""
self.scheduler.load_state_dict(snake_case__ )
def lowerCamelCase ( self : Tuple ) -> Tuple:
"""simple docstring"""
return self.scheduler.get_lr()
def lowerCamelCase ( self : Optional[int] , *lowercase_ : int , **lowercase_ : int ) -> Tuple:
"""simple docstring"""
return self.scheduler.print_lr(*snake_case__ , **snake_case__ )
| 464 |
"""simple docstring"""
import math
def _lowerCAmelCase ( lowerCAmelCase = 100 ):
'''simple docstring'''
UpperCAmelCase = sum(i * i for i in range(1 , n + 1 ) )
UpperCAmelCase = int(math.pow(sum(range(1 , n + 1 ) ) , 2 ) )
return square_of_sum - sum_of_squares
if __name__ == "__main__":
print(F'{solution() = }')
| 673 | 0 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
UpperCAmelCase : Dict = {
'''configuration_bigbird_pegasus''': [
'''BIGBIRD_PEGASUS_PRETRAINED_CONFIG_ARCHIVE_MAP''',
'''BigBirdPegasusConfig''',
'''BigBirdPegasusOnnxConfig''',
],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCAmelCase : Union[str, Any] = [
'''BIGBIRD_PEGASUS_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''BigBirdPegasusForCausalLM''',
'''BigBirdPegasusForConditionalGeneration''',
'''BigBirdPegasusForQuestionAnswering''',
'''BigBirdPegasusForSequenceClassification''',
'''BigBirdPegasusModel''',
'''BigBirdPegasusPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_bigbird_pegasus import (
BIGBIRD_PEGASUS_PRETRAINED_CONFIG_ARCHIVE_MAP,
BigBirdPegasusConfig,
BigBirdPegasusOnnxConfig,
)
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_bigbird_pegasus import (
BIGBIRD_PEGASUS_PRETRAINED_MODEL_ARCHIVE_LIST,
BigBirdPegasusForCausalLM,
BigBirdPegasusForConditionalGeneration,
BigBirdPegasusForQuestionAnswering,
BigBirdPegasusForSequenceClassification,
BigBirdPegasusModel,
BigBirdPegasusPreTrainedModel,
)
else:
import sys
UpperCAmelCase : Any = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 567 |
"""simple docstring"""
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = [0] * len(lowerCAmelCase )
UpperCAmelCase = []
UpperCAmelCase = [1] * len(lowerCAmelCase )
for values in graph.values():
for i in values:
indegree[i] += 1
for i in range(len(lowerCAmelCase ) ):
if indegree[i] == 0:
queue.append(lowerCAmelCase )
while queue:
UpperCAmelCase = queue.pop(0 )
for x in graph[vertex]:
indegree[x] -= 1
if long_dist[vertex] + 1 > long_dist[x]:
UpperCAmelCase = long_dist[vertex] + 1
if indegree[x] == 0:
queue.append(lowerCAmelCase )
print(max(lowerCAmelCase ) )
# Adjacency list of Graph
lowerCAmelCase_ : str = {0: [2, 3, 4], 1: [2, 7], 2: [5], 3: [5, 7], 4: [7], 5: [6], 6: [7], 7: []}
longest_distance(graph)
| 673 | 0 |
'''simple docstring'''
def lowerCAmelCase_ ( ):
return [
a * b * (10_00 - a - b)
for a in range(1 , 9_99 )
for b in range(_lowerCamelCase , 9_99 )
if (a * a + b * b == (10_00 - a - b) ** 2)
][0]
if __name__ == "__main__":
print(f"{solution() = }") | 578 |
"""simple docstring"""
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import torch
from ..models.auto import AutoModelForSequenceClassification, AutoTokenizer
from .base import PipelineTool
class UpperCamelCase_ ( a_ ):
_A : Optional[int] = 'facebook/bart-large-mnli'
_A : Union[str, Any] = (
'This is a tool that classifies an English text using provided labels. It takes two inputs: `text`, which '
'should be the text to classify, and `labels`, which should be the list of labels to use for classification. '
'It returns the most likely label in the list of provided `labels` for the input text.'
)
_A : Dict = 'text_classifier'
_A : Union[str, Any] = AutoTokenizer
_A : Tuple = AutoModelForSequenceClassification
_A : Optional[int] = ['text', ['text']]
_A : Dict = ['text']
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
super().setup()
UpperCAmelCase = self.model.config
UpperCAmelCase = -1
for idx, label in config.idalabel.items():
if label.lower().startswith("""entail""" ):
UpperCAmelCase = int(snake_case__ )
if self.entailment_id == -1:
raise ValueError("""Could not determine the entailment ID from the model config, please pass it at init.""" )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = labels
return self.pre_processor(
[text] * len(snake_case__ ) , [f'''This example is {label}''' for label in labels] , return_tensors="""pt""" , padding="""max_length""" , )
def UpperCamelCase_ ( self , snake_case__ ) -> str:
"""simple docstring"""
UpperCAmelCase = outputs.logits
UpperCAmelCase = torch.argmax(logits[:, 2] ).item()
return self._labels[label_id]
| 673 | 0 |
"""simple docstring"""
import importlib
import sys
from argparse import REMAINDER, ArgumentParser
from pathlib import Path
import torch_xla.distributed.xla_multiprocessing as xmp
def __lowercase ( ):
snake_case_ : List[str] = ArgumentParser(
description=(
'''PyTorch TPU distributed training launch helper utility that will spawn up multiple distributed processes'''
) )
# Optional arguments for the launch helper
parser.add_argument('''--num_cores''' , type=_a , default=1 , help='''Number of TPU cores to use (1 or 8).''' )
# positional
parser.add_argument(
'''training_script''' , type=_a , help=(
'''The full path to the single TPU training '''
'''program/script to be launched in parallel, '''
'''followed by all the arguments for the '''
'''training script'''
) , )
# rest from the training program
parser.add_argument('''training_script_args''' , nargs=_a )
return parser.parse_args()
def __lowercase ( ):
snake_case_ : str = parse_args()
# Import training_script as a module.
snake_case_ : int = Path(args.training_script )
sys.path.append(str(script_fpath.parent.resolve() ) )
snake_case_ : Any = script_fpath.stem
snake_case_ : int = importlib.import_module(_a )
# Patch sys.argv
snake_case_ : Tuple = [args.training_script] + args.training_script_args + ['''--tpu_num_cores''', str(args.num_cores )]
xmp.spawn(mod._mp_fn , args=() , nprocs=args.num_cores )
if __name__ == "__main__":
main()
| 123 |
"""simple docstring"""
from dataclasses import dataclass
from enum import Enum
from typing import List, Optional, Union
import numpy as np
import PIL
from PIL import Image
from ...utils import BaseOutput, is_torch_available, is_transformers_available
@dataclass
class UpperCamelCase_ ( a_ ):
_A : Union[List[PIL.Image.Image], np.ndarray]
_A : Optional[List[bool]]
if is_transformers_available() and is_torch_available():
from .pipeline_semantic_stable_diffusion import SemanticStableDiffusionPipeline
| 673 | 0 |
'''simple docstring'''
from __future__ import annotations
def __A ( _SCREAMING_SNAKE_CASE : Tuple , _SCREAMING_SNAKE_CASE : List[str] , _SCREAMING_SNAKE_CASE : Optional[int] , _SCREAMING_SNAKE_CASE : Tuple ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : List[Any] = []
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : int = input_list[low:mid], input_list[mid : high + 1]
while left and right:
result.append((left if left[0] <= right[0] else right).pop(0 ) )
__SCREAMING_SNAKE_CASE : List[Any] = result + left + right
return input_list
def __A ( _SCREAMING_SNAKE_CASE : List[Any] ):
"""simple docstring"""
if len(_SCREAMING_SNAKE_CASE ) <= 1:
return input_list
__SCREAMING_SNAKE_CASE : Tuple = list(_SCREAMING_SNAKE_CASE )
# iteration for two-way merging
__SCREAMING_SNAKE_CASE : List[Any] = 2
while p <= len(_SCREAMING_SNAKE_CASE ):
# getting low, high and middle value for merge-sort of single list
for i in range(0 , len(_SCREAMING_SNAKE_CASE ) , _SCREAMING_SNAKE_CASE ):
__SCREAMING_SNAKE_CASE : Optional[Any] = i
__SCREAMING_SNAKE_CASE : List[str] = i + p - 1
__SCREAMING_SNAKE_CASE : Union[str, Any] = (low + high + 1) // 2
__SCREAMING_SNAKE_CASE : Any = merge(_SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE )
# final merge of last two parts
if p * 2 >= len(_SCREAMING_SNAKE_CASE ):
__SCREAMING_SNAKE_CASE : Tuple = i
__SCREAMING_SNAKE_CASE : Any = merge(_SCREAMING_SNAKE_CASE , 0 , _SCREAMING_SNAKE_CASE , len(_SCREAMING_SNAKE_CASE ) - 1 )
break
p *= 2
return input_list
if __name__ == "__main__":
lowercase = input('''Enter numbers separated by a comma:\n''').strip()
if user_input == "":
lowercase = []
else:
lowercase = [int(item.strip()) for item in user_input.split(''',''')]
print(iter_merge_sort(unsorted))
| 211 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_torch_available,
)
lowerCAmelCase_ : Any = {
'''configuration_encodec''': [
'''ENCODEC_PRETRAINED_CONFIG_ARCHIVE_MAP''',
'''EncodecConfig''',
],
'''feature_extraction_encodec''': ['''EncodecFeatureExtractor'''],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
lowerCAmelCase_ : List[str] = [
'''ENCODEC_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''EncodecModel''',
'''EncodecPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_encodec import (
ENCODEC_PRETRAINED_CONFIG_ARCHIVE_MAP,
EncodecConfig,
)
from .feature_extraction_encodec import EncodecFeatureExtractor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_encodec import (
ENCODEC_PRETRAINED_MODEL_ARCHIVE_LIST,
EncodecModel,
EncodecPreTrainedModel,
)
else:
import sys
lowerCAmelCase_ : str = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 673 | 0 |
'''simple docstring'''
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import logging
from .tokenization_mobilebert import MobileBertTokenizer
__a: List[Any] = logging.get_logger(__name__)
__a: List[str] = {'''vocab_file''': '''vocab.txt''', '''tokenizer_file''': '''tokenizer.json'''}
__a: Union[str, Any] = {
'''vocab_file''': {'''mobilebert-uncased''': '''https://huggingface.co/google/mobilebert-uncased/resolve/main/vocab.txt'''},
'''tokenizer_file''': {
'''mobilebert-uncased''': '''https://huggingface.co/google/mobilebert-uncased/resolve/main/tokenizer.json'''
},
}
__a: Tuple = {'''mobilebert-uncased''': 5_12}
__a: List[Any] = {}
class UpperCAmelCase ( a_ ):
'''simple docstring'''
SCREAMING_SNAKE_CASE = VOCAB_FILES_NAMES
SCREAMING_SNAKE_CASE = PRETRAINED_VOCAB_FILES_MAP
SCREAMING_SNAKE_CASE = PRETRAINED_INIT_CONFIGURATION
SCREAMING_SNAKE_CASE = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
SCREAMING_SNAKE_CASE = MobileBertTokenizer
def __init__( self , __lowerCAmelCase=None , __lowerCAmelCase=None , __lowerCAmelCase=True , __lowerCAmelCase="[UNK]" , __lowerCAmelCase="[SEP]" , __lowerCAmelCase="[PAD]" , __lowerCAmelCase="[CLS]" , __lowerCAmelCase="[MASK]" , __lowerCAmelCase=True , __lowerCAmelCase=None , **__lowerCAmelCase , ) -> Optional[Any]:
super().__init__(
snake_case__ , tokenizer_file=snake_case__ , do_lower_case=snake_case__ , unk_token=snake_case__ , sep_token=snake_case__ , pad_token=snake_case__ , cls_token=snake_case__ , mask_token=snake_case__ , tokenize_chinese_chars=snake_case__ , strip_accents=snake_case__ , **snake_case__ , )
lowercase__ : Dict = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get('''lowercase''' , snake_case__ ) != do_lower_case
or normalizer_state.get('''strip_accents''' , snake_case__ ) != strip_accents
or normalizer_state.get('''handle_chinese_chars''' , snake_case__ ) != tokenize_chinese_chars
):
lowercase__ : int = getattr(snake_case__ , normalizer_state.pop('''type''' ) )
lowercase__ : Dict = do_lower_case
lowercase__ : Optional[int] = strip_accents
lowercase__ : Optional[int] = tokenize_chinese_chars
lowercase__ : Tuple = normalizer_class(**snake_case__ )
lowercase__ : int = do_lower_case
def _lowerCAmelCase( self , __lowerCAmelCase , __lowerCAmelCase=None ) -> Dict:
lowercase__ : int = [self.cls_token_id] + token_ids_a + [self.sep_token_id]
if token_ids_a:
output += token_ids_a + [self.sep_token_id]
return output
def _lowerCAmelCase( self , __lowerCAmelCase , __lowerCAmelCase = None ) -> List[int]:
lowercase__ : Dict = [self.sep_token_id]
lowercase__ : Any = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1]
def _lowerCAmelCase( self , __lowerCAmelCase , __lowerCAmelCase = None ) -> Tuple[str]:
lowercase__ : Any = self._tokenizer.model.save(snake_case__ , name=snake_case__ )
return tuple(snake_case__ )
| 152 |
"""simple docstring"""
from ...utils import (
OptionalDependencyNotAvailable,
is_torch_available,
is_transformers_available,
is_transformers_version,
)
try:
if not (is_transformers_available() and is_torch_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ...utils.dummy_torch_and_transformers_objects import (
ImageTextPipelineOutput,
UniDiffuserPipeline,
)
else:
from .modeling_text_decoder import UniDiffuserTextDecoder
from .modeling_uvit import UniDiffuserModel, UTransformeraDModel
from .pipeline_unidiffuser import ImageTextPipelineOutput, UniDiffuserPipeline
| 673 | 0 |
'''simple docstring'''
import math
import os
import unittest
from transformers import MegatronBertConfig, is_torch_available
from transformers.models.auto import get_values
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
MODEL_FOR_PRETRAINING_MAPPING,
MegatronBertForCausalLM,
MegatronBertForMaskedLM,
MegatronBertForMultipleChoice,
MegatronBertForNextSentencePrediction,
MegatronBertForPreTraining,
MegatronBertForQuestionAnswering,
MegatronBertForSequenceClassification,
MegatronBertForTokenClassification,
MegatronBertModel,
)
class __UpperCamelCase :
def __init__( self , __a , __a=13 , __a=7 , __a=True , __a=True , __a=True , __a=True , __a=99 , __a=64 , __a=32 , __a=5 , __a=4 , __a=37 , __a="gelu" , __a=0.1 , __a=0.1 , __a=512 , __a=16 , __a=2 , __a=0.02 , __a=3 , __a=4 , __a=None , ):
'''simple docstring'''
__a : int = parent
__a : List[Any] = batch_size
__a : Tuple = seq_length
__a : List[str] = is_training
__a : str = use_input_mask
__a : str = use_token_type_ids
__a : Optional[int] = use_labels
__a : Any = vocab_size
__a : Optional[Any] = hidden_size
__a : Union[str, Any] = embedding_size
__a : List[Any] = num_hidden_layers
__a : int = num_attention_heads
__a : Dict = intermediate_size
__a : int = hidden_act
__a : Optional[int] = hidden_dropout_prob
__a : List[str] = attention_probs_dropout_prob
__a : int = max_position_embeddings
__a : int = type_vocab_size
__a : Optional[int] = type_sequence_label_size
__a : List[Any] = initializer_range
__a : Dict = num_labels
__a : Optional[int] = num_choices
__a : Optional[int] = scope
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : Any = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__a : Dict = None
if self.use_input_mask:
__a : Optional[Any] = random_attention_mask([self.batch_size, self.seq_length] )
__a : Tuple = None
if self.use_token_type_ids:
__a : List[str] = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
__a : str = None
__a : Tuple = None
__a : List[Any] = None
if self.use_labels:
__a : List[Any] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
__a : Any = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
__a : Optional[int] = ids_tensor([self.batch_size] , self.num_choices )
__a : str = self.get_config()
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def __UpperCAmelCase ( self ):
'''simple docstring'''
return MegatronBertConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , embedding_size=self.embedding_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=snake_case__ , initializer_range=self.initializer_range , )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : Tuple = MegatronBertModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__a : Optional[Any] = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ )
__a : List[Any] = model(snake_case__ , token_type_ids=snake_case__ )
__a : int = model(snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
self.parent.assertEqual(result.pooler_output.shape , (self.batch_size, self.hidden_size) )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : Union[str, Any] = MegatronBertForMaskedLM(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__a : int = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : List[str] = MegatronBertForCausalLM(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__a : Tuple = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : List[str] = MegatronBertForNextSentencePrediction(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__a : Tuple = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, 2) )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : Optional[int] = MegatronBertForPreTraining(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__a : Optional[int] = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ , next_sentence_label=snake_case__ , )
self.parent.assertEqual(result.prediction_logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
self.parent.assertEqual(result.seq_relationship_logits.shape , (self.batch_size, 2) )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : Optional[Any] = MegatronBertForQuestionAnswering(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__a : List[str] = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , start_positions=snake_case__ , end_positions=snake_case__ , )
self.parent.assertEqual(result.start_logits.shape , (self.batch_size, self.seq_length) )
self.parent.assertEqual(result.end_logits.shape , (self.batch_size, self.seq_length) )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : Optional[Any] = self.num_labels
__a : Tuple = MegatronBertForSequenceClassification(snake_case__ )
model.to(snake_case__ )
model.eval()
__a : List[Any] = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : List[str] = self.num_labels
__a : str = MegatronBertForTokenClassification(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__a : Any = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def __UpperCAmelCase ( self , __a , __a , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : Union[str, Any] = self.num_choices
__a : str = MegatronBertForMultipleChoice(config=snake_case__ )
model.to(snake_case__ )
model.eval()
__a : int = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
__a : str = token_type_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
__a : int = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
__a : Any = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : Union[str, Any] = self.prepare_config_and_inputs()
(
(
__a
) , (
__a
) , (
__a
) , (
__a
) , (
__a
) , (
__a
) , (
__a
) ,
) : Optional[Any] = config_and_inputs
__a : str = {'input_ids': input_ids, 'token_type_ids': token_type_ids, 'attention_mask': input_mask}
return config, inputs_dict
@require_torch
class __UpperCamelCase ( a_ , a_ , unittest.TestCase ):
A_ = (
(
MegatronBertModel,
MegatronBertForMaskedLM,
MegatronBertForCausalLM,
MegatronBertForMultipleChoice,
MegatronBertForNextSentencePrediction,
MegatronBertForPreTraining,
MegatronBertForQuestionAnswering,
MegatronBertForSequenceClassification,
MegatronBertForTokenClassification,
)
if is_torch_available()
else ()
)
A_ = (
{
'feature-extraction': MegatronBertModel,
'fill-mask': MegatronBertForMaskedLM,
'question-answering': MegatronBertForQuestionAnswering,
'text-classification': MegatronBertForSequenceClassification,
'text-generation': MegatronBertForCausalLM,
'token-classification': MegatronBertForTokenClassification,
'zero-shot': MegatronBertForSequenceClassification,
}
if is_torch_available()
else {}
)
A_ = True
# test_resize_embeddings = False
A_ = False
def __UpperCAmelCase ( self , __a , __a , __a=False ):
'''simple docstring'''
__a : int = super()._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
if return_labels:
if model_class in get_values(snake_case__ ):
__a : Tuple = torch.zeros(
(self.model_tester.batch_size, self.model_tester.seq_length) , dtype=torch.long , device=snake_case__ )
__a : int = torch.zeros(
self.model_tester.batch_size , dtype=torch.long , device=snake_case__ )
return inputs_dict
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : int = MegatronBertModelTester(self )
__a : Dict = ConfigTester(self , config_class=snake_case__ , hidden_size=37 )
def __UpperCAmelCase ( self ):
'''simple docstring'''
self.config_tester.run_common_tests()
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : Dict = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_megatron_bert_model(*snake_case__ )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_megatron_bert_for_masked_lm(*snake_case__ )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_megatron_bert_for_multiple_choice(*snake_case__ )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_megatron_bert_for_next_sequence_prediction(*snake_case__ )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_megatron_bert_for_pretraining(*snake_case__ )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_megatron_bert_for_question_answering(*snake_case__ )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_megatron_bert_for_sequence_classification(*snake_case__ )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : Optional[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_megatron_bert_for_token_classification(*snake_case__ )
def lowerCamelCase (_SCREAMING_SNAKE_CASE : int ):
return torch.tensor(
_SCREAMING_SNAKE_CASE , dtype=torch.long , device=_SCREAMING_SNAKE_CASE , )
__lowercase : List[Any] = 1E-4
@require_torch
@require_sentencepiece
@require_tokenizers
class __UpperCamelCase ( unittest.TestCase ):
@slow
@unittest.skip('Model is not available.' )
def __UpperCAmelCase ( self ):
'''simple docstring'''
__a : str = 'nvidia/megatron-bert-uncased-345m'
if "MYDIR" in os.environ:
__a : int = os.path.join(os.environ['MYDIR'] , snake_case__ )
__a : List[Any] = MegatronBertModel.from_pretrained(snake_case__ )
model.to(snake_case__ )
model.half()
__a : List[str] = _long_tensor([[101, 7110, 1005, 1056, 2023, 1_1333, 1_7413, 1029, 102]] )
with torch.no_grad():
__a : Dict = model(snake_case__ )[0]
__a : Dict = torch.Size((1, 9, 1024) )
self.assertEqual(output.shape , snake_case__ )
__a : Dict = [-0.6040, -0.2517, -0.1025, 0.3420, -0.6758, -0.0017, -0.1089, -0.1990, 0.5728]
for ii in range(3 ):
for jj in range(3 ):
__a : Dict = output[0, ii, jj]
__a : Dict = expected[3 * ii + jj]
__a : str = 'ii={} jj={} a={} b={}'.format(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
self.assertTrue(math.isclose(snake_case__ , snake_case__ , rel_tol=snake_case__ , abs_tol=snake_case__ ) , msg=snake_case__ )
| 476 |
"""simple docstring"""
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
UNetaDConditionModel,
VideoToVideoSDPipeline,
)
from diffusers.utils import floats_tensor, is_xformers_available, skip_mps
from diffusers.utils.testing_utils import enable_full_determinism, slow, torch_device
from ..pipeline_params import (
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineTesterMixin
enable_full_determinism()
@skip_mps
class UpperCamelCase_ ( a_ , unittest.TestCase ):
_A : str = VideoToVideoSDPipeline
_A : List[str] = TEXT_GUIDED_IMAGE_VARIATION_PARAMS.union({'video'} ) - {'image', 'width', 'height'}
_A : int = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({'video'} ) - {'image'}
_A : int = PipelineTesterMixin.required_optional_params - {'latents'}
_A : List[str] = False
# No `output_type`.
_A : Any = frozenset(
[
'num_inference_steps',
'generator',
'latents',
'return_dict',
'callback',
'callback_steps',
] )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
torch.manual_seed(0 )
UpperCAmelCase = UNetaDConditionModel(
block_out_channels=(32, 64, 64, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=("""CrossAttnDownBlock3D""", """CrossAttnDownBlock3D""", """CrossAttnDownBlock3D""", """DownBlock3D""") , up_block_types=("""UpBlock3D""", """CrossAttnUpBlock3D""", """CrossAttnUpBlock3D""", """CrossAttnUpBlock3D""") , cross_attention_dim=32 , attention_head_dim=4 , )
UpperCAmelCase = DDIMScheduler(
beta_start=0.00_085 , beta_end=0.012 , beta_schedule="""scaled_linear""" , clip_sample=snake_case__ , set_alpha_to_one=snake_case__ , )
torch.manual_seed(0 )
UpperCAmelCase = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] , up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] , latent_channels=4 , sample_size=1_28 , )
torch.manual_seed(0 )
UpperCAmelCase = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=10_00 , hidden_act="""gelu""" , projection_dim=5_12 , )
UpperCAmelCase = CLIPTextModel(snake_case__ )
UpperCAmelCase = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
UpperCAmelCase = {
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
"""text_encoder""": text_encoder,
"""tokenizer""": tokenizer,
}
return components
def UpperCamelCase_ ( self , snake_case__ , snake_case__=0 ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = floats_tensor((1, 3, 3, 32, 32) , rng=random.Random(snake_case__ ) ).to(snake_case__ )
if str(snake_case__ ).startswith("""mps""" ):
UpperCAmelCase = torch.manual_seed(snake_case__ )
else:
UpperCAmelCase = torch.Generator(device=snake_case__ ).manual_seed(snake_case__ )
UpperCAmelCase = {
"""prompt""": """A painting of a squirrel eating a burger""",
"""video""": video,
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 6.0,
"""output_type""": """pt""",
}
return inputs
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
UpperCAmelCase = """cpu""" # ensure determinism for the device-dependent torch.Generator
UpperCAmelCase = self.get_dummy_components()
UpperCAmelCase = VideoToVideoSDPipeline(**snake_case__ )
UpperCAmelCase = sd_pipe.to(snake_case__ )
sd_pipe.set_progress_bar_config(disable=snake_case__ )
UpperCAmelCase = self.get_dummy_inputs(snake_case__ )
UpperCAmelCase = """np"""
UpperCAmelCase = sd_pipe(**snake_case__ ).frames
UpperCAmelCase = frames[0][-3:, -3:, -1]
assert frames[0].shape == (32, 32, 3)
UpperCAmelCase = np.array([1_06, 1_17, 1_13, 1_74, 1_37, 1_12, 1_48, 1_51, 1_31] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(
torch_device != """cuda""" or not is_xformers_available() , reason="""XFormers attention is only available with CUDA and `xformers` installed""" , )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
self._test_xformers_attention_forwardGenerator_pass(test_mean_pixel_difference=snake_case__ , expected_max_diff=5e-3 )
@unittest.skip(reason="""Batching needs to be properly figured out first for this pipeline.""" )
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
pass
@unittest.skip(reason="""Batching needs to be properly figured out first for this pipeline.""" )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
pass
@unittest.skip(reason="""`num_images_per_prompt` argument is not supported for this pipeline.""" )
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
pass
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return super().test_progress_bar()
@slow
@skip_mps
class UpperCamelCase_ ( unittest.TestCase ):
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = VideoToVideoSDPipeline.from_pretrained("""cerspense/zeroscope_v2_XL""" , torch_dtype=torch.floataa )
pipe.enable_model_cpu_offload()
# 10 frames
UpperCAmelCase = torch.Generator(device="""cpu""" ).manual_seed(0 )
UpperCAmelCase = torch.randn((1, 10, 3, 10_24, 5_76) , generator=snake_case__ )
UpperCAmelCase = video.to("""cuda""" )
UpperCAmelCase = """Spiderman is surfing"""
UpperCAmelCase = pipe(snake_case__ , video=snake_case__ , generator=snake_case__ , num_inference_steps=3 , output_type="""pt""" ).frames
UpperCAmelCase = np.array([-1.0_458_984, -1.1_279_297, -0.9_663_086, -0.91_503_906, -0.75_097_656] )
assert np.abs(video_frames.cpu().numpy()[0, 0, 0, 0, -5:] - expected_array ).sum() < 1e-2
| 673 | 0 |
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
A_ = {
'''configuration_mgp_str''': ['''MGP_STR_PRETRAINED_CONFIG_ARCHIVE_MAP''', '''MgpstrConfig'''],
'''processing_mgp_str''': ['''MgpstrProcessor'''],
'''tokenization_mgp_str''': ['''MgpstrTokenizer'''],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
A_ = [
'''MGP_STR_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''MgpstrModel''',
'''MgpstrPreTrainedModel''',
'''MgpstrForSceneTextRecognition''',
]
if TYPE_CHECKING:
from .configuration_mgp_str import MGP_STR_PRETRAINED_CONFIG_ARCHIVE_MAP, MgpstrConfig
from .processing_mgp_str import MgpstrProcessor
from .tokenization_mgp_str import MgpstrTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mgp_str import (
MGP_STR_PRETRAINED_MODEL_ARCHIVE_LIST,
MgpstrForSceneTextRecognition,
MgpstrModel,
MgpstrPreTrainedModel,
)
else:
import sys
A_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 393 |
"""simple docstring"""
import functools
import operator
from ...configuration_utils import PretrainedConfig
from ...utils import logging
lowerCAmelCase_ : int = logging.get_logger(__name__)
lowerCAmelCase_ : Any = {
'''facebook/wav2vec2-base-960h''': '''https://huggingface.co/facebook/wav2vec2-base-960h/resolve/main/config.json''',
# See all Wav2Vec2 models at https://huggingface.co/models?filter=wav2vec2
}
class UpperCamelCase_ ( a_ ):
_A : int = 'wav2vec2'
def __init__( self , snake_case__=32 , snake_case__=7_68 , snake_case__=12 , snake_case__=12 , snake_case__=30_72 , snake_case__="gelu" , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.0 , snake_case__=0.0 , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.02 , snake_case__=1e-5 , snake_case__="group" , snake_case__="gelu" , snake_case__=(5_12, 5_12, 5_12, 5_12, 5_12, 5_12, 5_12) , snake_case__=(5, 2, 2, 2, 2, 2, 2) , snake_case__=(10, 3, 3, 3, 3, 2, 2) , snake_case__=False , snake_case__=1_28 , snake_case__=16 , snake_case__=False , snake_case__=True , snake_case__=0.05 , snake_case__=10 , snake_case__=2 , snake_case__=0.0 , snake_case__=10 , snake_case__=0 , snake_case__=3_20 , snake_case__=2 , snake_case__=0.1 , snake_case__=1_00 , snake_case__=2_56 , snake_case__=2_56 , snake_case__=0.1 , snake_case__="sum" , snake_case__=False , snake_case__=False , snake_case__=2_56 , snake_case__=(5_12, 5_12, 5_12, 5_12, 15_00) , snake_case__=(5, 3, 3, 1, 1) , snake_case__=(1, 2, 3, 1, 1) , snake_case__=5_12 , snake_case__=0 , snake_case__=1 , snake_case__=2 , snake_case__=False , snake_case__=3 , snake_case__=2 , snake_case__=3 , snake_case__=None , snake_case__=None , **snake_case__ , ) -> Optional[Any]:
"""simple docstring"""
super().__init__(**snake_case__ , pad_token_id=snake_case__ , bos_token_id=snake_case__ , eos_token_id=snake_case__ )
UpperCAmelCase = hidden_size
UpperCAmelCase = feat_extract_norm
UpperCAmelCase = feat_extract_activation
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = conv_bias
UpperCAmelCase = num_conv_pos_embeddings
UpperCAmelCase = num_conv_pos_embedding_groups
UpperCAmelCase = len(self.conv_dim )
UpperCAmelCase = num_hidden_layers
UpperCAmelCase = intermediate_size
UpperCAmelCase = hidden_act
UpperCAmelCase = num_attention_heads
UpperCAmelCase = hidden_dropout
UpperCAmelCase = attention_dropout
UpperCAmelCase = activation_dropout
UpperCAmelCase = feat_proj_dropout
UpperCAmelCase = final_dropout
UpperCAmelCase = layerdrop
UpperCAmelCase = layer_norm_eps
UpperCAmelCase = initializer_range
UpperCAmelCase = vocab_size
UpperCAmelCase = do_stable_layer_norm
UpperCAmelCase = use_weighted_layer_sum
if (
(len(self.conv_stride ) != self.num_feat_extract_layers)
or (len(self.conv_kernel ) != self.num_feat_extract_layers)
or (len(self.conv_dim ) != self.num_feat_extract_layers)
):
raise ValueError(
"""Configuration for convolutional layers is incorrect. It is required that `len(config.conv_dim)` =="""
""" `len(config.conv_stride)` == `len(config.conv_kernel)`, but is `len(config.conv_dim) ="""
f''' {len(self.conv_dim )}`, `len(config.conv_stride) = {len(self.conv_stride )}`,'''
f''' `len(config.conv_kernel) = {len(self.conv_kernel )}`.''' )
# fine-tuning config parameters for SpecAugment: https://arxiv.org/abs/1904.08779
UpperCAmelCase = apply_spec_augment
UpperCAmelCase = mask_time_prob
UpperCAmelCase = mask_time_length
UpperCAmelCase = mask_time_min_masks
UpperCAmelCase = mask_feature_prob
UpperCAmelCase = mask_feature_length
UpperCAmelCase = mask_feature_min_masks
# parameters for pretraining with codevector quantized representations
UpperCAmelCase = num_codevectors_per_group
UpperCAmelCase = num_codevector_groups
UpperCAmelCase = contrastive_logits_temperature
UpperCAmelCase = feat_quantizer_dropout
UpperCAmelCase = num_negatives
UpperCAmelCase = codevector_dim
UpperCAmelCase = proj_codevector_dim
UpperCAmelCase = diversity_loss_weight
# ctc loss
UpperCAmelCase = ctc_loss_reduction
UpperCAmelCase = ctc_zero_infinity
# adapter
UpperCAmelCase = add_adapter
UpperCAmelCase = adapter_kernel_size
UpperCAmelCase = adapter_stride
UpperCAmelCase = num_adapter_layers
UpperCAmelCase = output_hidden_size or hidden_size
UpperCAmelCase = adapter_attn_dim
# SequenceClassification-specific parameter. Feel free to ignore for other classes.
UpperCAmelCase = classifier_proj_size
# XVector-specific parameters. Feel free to ignore for other classes.
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = xvector_output_dim
@property
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return functools.reduce(operator.mul , self.conv_stride , 1 )
| 673 | 0 |
def lowercase__ ( ):
'''simple docstring'''
UpperCAmelCase_ : Any = [31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31]
UpperCAmelCase_ : List[Any] = 6
UpperCAmelCase_ : Optional[int] = 1
UpperCAmelCase_ : Any = 1_901
UpperCAmelCase_ : int = 0
while year < 2_001:
day += 7
if (year % 4 == 0 and year % 100 != 0) or (year % 400 == 0):
if day > days_per_month[month - 1] and month != 2:
month += 1
UpperCAmelCase_ : Dict = day - days_per_month[month - 2]
elif day > 29 and month == 2:
month += 1
UpperCAmelCase_ : Any = day - 29
else:
if day > days_per_month[month - 1]:
month += 1
UpperCAmelCase_ : Optional[int] = day - days_per_month[month - 2]
if month > 12:
year += 1
UpperCAmelCase_ : List[str] = 1
if year < 2_001 and day == 1:
sundays += 1
return sundays
if __name__ == "__main__":
print(solution())
| 406 |
"""simple docstring"""
import dataclasses
import json
import sys
import types
from argparse import ArgumentDefaultsHelpFormatter, ArgumentParser, ArgumentTypeError
from copy import copy
from enum import Enum
from inspect import isclass
from pathlib import Path
from typing import Any, Callable, Dict, Iterable, List, Literal, NewType, Optional, Tuple, Union, get_type_hints
import yaml
lowerCAmelCase_ : Optional[Any] = NewType('''DataClass''', Any)
lowerCAmelCase_ : Any = NewType('''DataClassType''', Any)
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
if isinstance(lowerCAmelCase , lowerCAmelCase ):
return v
if v.lower() in ("yes", "true", "t", "y", "1"):
return True
elif v.lower() in ("no", "false", "f", "n", "0"):
return False
else:
raise ArgumentTypeError(
F'''Truthy value expected: got {v} but expected one of yes/no, true/false, t/f, y/n, 1/0 (case insensitive).''' )
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = {str(lowerCAmelCase ): choice for choice in choices}
return lambda lowerCAmelCase : str_to_choice.get(lowerCAmelCase , lowerCAmelCase )
def _lowerCAmelCase ( *,
lowerCAmelCase = None , lowerCAmelCase = None , lowerCAmelCase = dataclasses.MISSING , lowerCAmelCase = dataclasses.MISSING , lowerCAmelCase = None , **lowerCAmelCase , ):
'''simple docstring'''
if metadata is None:
# Important, don't use as default param in function signature because dict is mutable and shared across function calls
UpperCAmelCase = {}
if aliases is not None:
UpperCAmelCase = aliases
if help is not None:
UpperCAmelCase = help
return dataclasses.field(metadata=lowerCAmelCase , default=lowerCAmelCase , default_factory=lowerCAmelCase , **lowerCAmelCase )
class UpperCamelCase_ ( a_ ):
_A : Iterable[DataClassType]
def __init__( self , snake_case__ , **snake_case__ ) -> List[str]:
"""simple docstring"""
if "formatter_class" not in kwargs:
UpperCAmelCase = ArgumentDefaultsHelpFormatter
super().__init__(**snake_case__ )
if dataclasses.is_dataclass(snake_case__ ):
UpperCAmelCase = [dataclass_types]
UpperCAmelCase = list(snake_case__ )
for dtype in self.dataclass_types:
self._add_dataclass_arguments(snake_case__ )
@staticmethod
def UpperCamelCase_ ( snake_case__ , snake_case__ ) -> str:
"""simple docstring"""
UpperCAmelCase = f'''--{field.name}'''
UpperCAmelCase = field.metadata.copy()
# field.metadata is not used at all by Data Classes,
# it is provided as a third-party extension mechanism.
if isinstance(field.type , snake_case__ ):
raise RuntimeError(
"""Unresolved type detected, which should have been done with the help of """
"""`typing.get_type_hints` method by default""" )
UpperCAmelCase = kwargs.pop("""aliases""" , [] )
if isinstance(snake_case__ , snake_case__ ):
UpperCAmelCase = [aliases]
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
if origin_type is Union or (hasattr(snake_case__ , """UnionType""" ) and isinstance(snake_case__ , types.UnionType )):
if str not in field.type.__args__ and (
len(field.type.__args__ ) != 2 or type(snake_case__ ) not in field.type.__args__
):
raise ValueError(
"""Only `Union[X, NoneType]` (i.e., `Optional[X]`) is allowed for `Union` because"""
""" the argument parser only supports one type per argument."""
f''' Problem encountered in field \'{field.name}\'.''' )
if type(snake_case__ ) not in field.type.__args__:
# filter `str` in Union
UpperCAmelCase = field.type.__args__[0] if field.type.__args__[1] == str else field.type.__args__[1]
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
elif bool not in field.type.__args__:
# filter `NoneType` in Union (except for `Union[bool, NoneType]`)
UpperCAmelCase = (
field.type.__args__[0] if isinstance(snake_case__ , field.type.__args__[1] ) else field.type.__args__[1]
)
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
# A variable to store kwargs for a boolean field, if needed
# so that we can init a `no_*` complement argument (see below)
UpperCAmelCase = {}
if origin_type is Literal or (isinstance(field.type , snake_case__ ) and issubclass(field.type , snake_case__ )):
if origin_type is Literal:
UpperCAmelCase = field.type.__args__
else:
UpperCAmelCase = [x.value for x in field.type]
UpperCAmelCase = make_choice_type_function(kwargs["""choices"""] )
if field.default is not dataclasses.MISSING:
UpperCAmelCase = field.default
else:
UpperCAmelCase = True
elif field.type is bool or field.type == Optional[bool]:
# Copy the currect kwargs to use to instantiate a `no_*` complement argument below.
# We do not initialize it here because the `no_*` alternative must be instantiated after the real argument
UpperCAmelCase = copy(snake_case__ )
# Hack because type=bool in argparse does not behave as we want.
UpperCAmelCase = string_to_bool
if field.type is bool or (field.default is not None and field.default is not dataclasses.MISSING):
# Default value is False if we have no default when of type bool.
UpperCAmelCase = False if field.default is dataclasses.MISSING else field.default
# This is the value that will get picked if we don't include --field_name in any way
UpperCAmelCase = default
# This tells argparse we accept 0 or 1 value after --field_name
UpperCAmelCase = """?"""
# This is the value that will get picked if we do --field_name (without value)
UpperCAmelCase = True
elif isclass(snake_case__ ) and issubclass(snake_case__ , snake_case__ ):
UpperCAmelCase = field.type.__args__[0]
UpperCAmelCase = """+"""
if field.default_factory is not dataclasses.MISSING:
UpperCAmelCase = field.default_factory()
elif field.default is dataclasses.MISSING:
UpperCAmelCase = True
else:
UpperCAmelCase = field.type
if field.default is not dataclasses.MISSING:
UpperCAmelCase = field.default
elif field.default_factory is not dataclasses.MISSING:
UpperCAmelCase = field.default_factory()
else:
UpperCAmelCase = True
parser.add_argument(snake_case__ , *snake_case__ , **snake_case__ )
# Add a complement `no_*` argument for a boolean field AFTER the initial field has already been added.
# Order is important for arguments with the same destination!
# We use a copy of earlier kwargs because the original kwargs have changed a lot before reaching down
# here and we do not need those changes/additional keys.
if field.default is True and (field.type is bool or field.type == Optional[bool]):
UpperCAmelCase = False
parser.add_argument(f'''--no_{field.name}''' , action="""store_false""" , dest=field.name , **snake_case__ )
def UpperCamelCase_ ( self , snake_case__ ) -> Any:
"""simple docstring"""
if hasattr(snake_case__ , """_argument_group_name""" ):
UpperCAmelCase = self.add_argument_group(dtype._argument_group_name )
else:
UpperCAmelCase = self
try:
UpperCAmelCase = get_type_hints(snake_case__ )
except NameError:
raise RuntimeError(
f'''Type resolution failed for {dtype}. Try declaring the class in global scope or '''
"""removing line of `from __future__ import annotations` which opts in Postponed """
"""Evaluation of Annotations (PEP 563)""" )
except TypeError as ex:
# Remove this block when we drop Python 3.9 support
if sys.version_info[:2] < (3, 10) and "unsupported operand type(s) for |" in str(snake_case__ ):
UpperCAmelCase = """.""".join(map(snake_case__ , sys.version_info[:3] ) )
raise RuntimeError(
f'''Type resolution failed for {dtype} on Python {python_version}. Try removing '''
"""line of `from __future__ import annotations` which opts in union types as """
"""`X | Y` (PEP 604) via Postponed Evaluation of Annotations (PEP 563). To """
"""support Python versions that lower than 3.10, you need to use """
"""`typing.Union[X, Y]` instead of `X | Y` and `typing.Optional[X]` instead of """
"""`X | None`.""" ) from ex
raise
for field in dataclasses.fields(snake_case__ ):
if not field.init:
continue
UpperCAmelCase = type_hints[field.name]
self._parse_dataclass_field(snake_case__ , snake_case__ )
def UpperCamelCase_ ( self , snake_case__=None , snake_case__=False , snake_case__=True , snake_case__=None , snake_case__=None , ) -> Tuple[DataClass, ...]:
"""simple docstring"""
if args_file_flag or args_filename or (look_for_args_file and len(sys.argv )):
UpperCAmelCase = []
if args_filename:
args_files.append(Path(snake_case__ ) )
elif look_for_args_file and len(sys.argv ):
args_files.append(Path(sys.argv[0] ).with_suffix(""".args""" ) )
# args files specified via command line flag should overwrite default args files so we add them last
if args_file_flag:
# Create special parser just to extract the args_file_flag values
UpperCAmelCase = ArgumentParser()
args_file_parser.add_argument(snake_case__ , type=snake_case__ , action="""append""" )
# Use only remaining args for further parsing (remove the args_file_flag)
UpperCAmelCase , UpperCAmelCase = args_file_parser.parse_known_args(args=snake_case__ )
UpperCAmelCase = vars(snake_case__ ).get(args_file_flag.lstrip("""-""" ) , snake_case__ )
if cmd_args_file_paths:
args_files.extend([Path(snake_case__ ) for p in cmd_args_file_paths] )
UpperCAmelCase = []
for args_file in args_files:
if args_file.exists():
file_args += args_file.read_text().split()
# in case of duplicate arguments the last one has precedence
# args specified via the command line should overwrite args from files, so we add them last
UpperCAmelCase = file_args + args if args is not None else file_args + sys.argv[1:]
UpperCAmelCase , UpperCAmelCase = self.parse_known_args(args=snake_case__ )
UpperCAmelCase = []
for dtype in self.dataclass_types:
UpperCAmelCase = {f.name for f in dataclasses.fields(snake_case__ ) if f.init}
UpperCAmelCase = {k: v for k, v in vars(snake_case__ ).items() if k in keys}
for k in keys:
delattr(snake_case__ , snake_case__ )
UpperCAmelCase = dtype(**snake_case__ )
outputs.append(snake_case__ )
if len(namespace.__dict__ ) > 0:
# additional namespace.
outputs.append(snake_case__ )
if return_remaining_strings:
return (*outputs, remaining_args)
else:
if remaining_args:
raise ValueError(f'''Some specified arguments are not used by the HfArgumentParser: {remaining_args}''' )
return (*outputs,)
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
UpperCAmelCase = set(args.keys() )
UpperCAmelCase = []
for dtype in self.dataclass_types:
UpperCAmelCase = {f.name for f in dataclasses.fields(snake_case__ ) if f.init}
UpperCAmelCase = {k: v for k, v in args.items() if k in keys}
unused_keys.difference_update(inputs.keys() )
UpperCAmelCase = dtype(**snake_case__ )
outputs.append(snake_case__ )
if not allow_extra_keys and unused_keys:
raise ValueError(f'''Some keys are not used by the HfArgumentParser: {sorted(snake_case__ )}''' )
return tuple(snake_case__ )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
with open(Path(snake_case__ ) , encoding="""utf-8""" ) as open_json_file:
UpperCAmelCase = json.loads(open_json_file.read() )
UpperCAmelCase = self.parse_dict(snake_case__ , allow_extra_keys=snake_case__ )
return tuple(snake_case__ )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
UpperCAmelCase = self.parse_dict(yaml.safe_load(Path(snake_case__ ).read_text() ) , allow_extra_keys=snake_case__ )
return tuple(snake_case__ )
| 673 | 0 |
"""simple docstring"""
import os
from pickle import UnpicklingError
from typing import Dict, Tuple
import jax
import jax.numpy as jnp
import numpy as np
from flax.serialization import from_bytes
from flax.traverse_util import flatten_dict, unflatten_dict
import transformers
from .utils import logging
__UpperCAmelCase =logging.get_logger(__name__)
def __a ( A , A , A , A=False ) -> Dict:
'''simple docstring'''
try:
import torch # noqa: F401
except ImportError:
logger.error(
"Loading a PyTorch model in Flax, requires both PyTorch and Flax to be installed. Please see"
" https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation"
" instructions." )
raise
if not is_sharded:
A__ = os.path.abspath(A )
logger.info(f"""Loading PyTorch weights from {pt_path}""" )
A__ = torch.load(A , map_location="cpu" )
logger.info(f"""PyTorch checkpoint contains {sum(t.numel() for t in pt_state_dict.values() ):,} parameters.""" )
A__ = convert_pytorch_state_dict_to_flax(A , A )
else:
# model is sharded and pytorch_checkpoint_path already contains the list of .pt shard files
A__ = convert_pytorch_sharded_state_dict_to_flax(A , A )
return flax_state_dict
def __a ( A , A , A , A , ) -> List[Any]:
'''simple docstring'''
def is_key_or_prefix_key_in_dict(A ) -> bool:
return len(set(A ) & {key, (model_prefix,) + key} ) > 0
# layer norm
A__ = pt_tuple_key[:-1] + ("scale",)
if pt_tuple_key[-1] in ["weight", "gamma"] and is_key_or_prefix_key_in_dict(A ):
return renamed_pt_tuple_key, pt_tensor
# batch norm layer mean
A__ = pt_tuple_key[:-1] + ("mean",)
if pt_tuple_key[-1] == "running_mean" and not is_key_or_prefix_key_in_dict(A ):
return renamed_pt_tuple_key, pt_tensor
# batch norm layer var
A__ = pt_tuple_key[:-1] + ("var",)
if pt_tuple_key[-1] == "running_var" and not is_key_or_prefix_key_in_dict(A ):
return renamed_pt_tuple_key, pt_tensor
# embedding
A__ = pt_tuple_key[:-1] + ("embedding",)
if pt_tuple_key[-1] == "weight" and is_key_or_prefix_key_in_dict(A ):
return renamed_pt_tuple_key, pt_tensor
# conv layer
A__ = pt_tuple_key[:-1] + ("kernel",)
if pt_tuple_key[-1] == "weight" and pt_tensor.ndim == 4 and not is_key_or_prefix_key_in_dict(A ):
A__ = pt_tensor.transpose(2 , 3 , 1 , 0 )
return renamed_pt_tuple_key, pt_tensor
# linear layer
A__ = pt_tuple_key[:-1] + ("kernel",)
if pt_tuple_key[-1] == "weight" and not is_key_or_prefix_key_in_dict(A ):
A__ = pt_tensor.T
return renamed_pt_tuple_key, pt_tensor
# old PyTorch layer norm weight
A__ = pt_tuple_key[:-1] + ("weight",)
if pt_tuple_key[-1] == "gamma":
return renamed_pt_tuple_key, pt_tensor
# old PyTorch layer norm bias
A__ = pt_tuple_key[:-1] + ("bias",)
if pt_tuple_key[-1] == "beta":
return renamed_pt_tuple_key, pt_tensor
# New `weight_norm` from https://github.com/huggingface/transformers/pull/24030
A__ = None
if pt_tuple_key[-3::2] == ("parametrizations", "original0"):
A__ = pt_tuple_key[-2] + "_g"
elif pt_tuple_key[-3::2] == ("parametrizations", "original1"):
A__ = pt_tuple_key[-2] + "_v"
if name is not None:
A__ = pt_tuple_key[:-3] + (name,)
return renamed_pt_tuple_key, pt_tensor
return pt_tuple_key, pt_tensor
def __a ( A , A ) -> Any:
'''simple docstring'''
A__ = {k: v.numpy() for k, v in pt_state_dict.items()}
A__ = flax_model.base_model_prefix
# use params dict if the model contains batch norm layers
if "params" in flax_model.params:
A__ = flax_model.params["params"]
else:
A__ = flax_model.params
A__ = flatten_dict(A )
# add batch_stats keys,values to dict
if "batch_stats" in flax_model.params:
A__ = flatten_dict(flax_model.params["batch_stats"] )
random_flax_state_dict.update(A )
A__ = {}
A__ = (model_prefix not in flax_model_params) and (
model_prefix in {k.split("." )[0] for k in pt_state_dict.keys()}
)
A__ = (model_prefix in flax_model_params) and (
model_prefix not in {k.split("." )[0] for k in pt_state_dict.keys()}
)
# Need to change some parameters name to match Flax names
for pt_key, pt_tensor in pt_state_dict.items():
A__ = tuple(pt_key.split("." ) )
# remove base model prefix if necessary
A__ = pt_tuple_key[0] == model_prefix
if load_model_with_head_into_base_model and has_base_model_prefix:
A__ = pt_tuple_key[1:]
# Correctly rename weight parameters
A__ , A__ = rename_key_and_reshape_tensor(
A , A , A , A )
# add model prefix if necessary
A__ = (model_prefix,) + flax_key in random_flax_state_dict
if load_base_model_into_model_with_head and require_base_model_prefix:
A__ = (model_prefix,) + flax_key
if flax_key in random_flax_state_dict:
if flax_tensor.shape != random_flax_state_dict[flax_key].shape:
raise ValueError(
f"""PyTorch checkpoint seems to be incorrect. Weight {pt_key} was expected to be of shape """
f"""{random_flax_state_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
# add batch stats if the model contains batchnorm layers
if "batch_stats" in flax_model.params:
if "mean" in flax_key[-1] or "var" in flax_key[-1]:
A__ = jnp.asarray(A )
continue
# remove num_batches_tracked key
if "num_batches_tracked" in flax_key[-1]:
flax_state_dict.pop(A , A )
continue
# also add unexpected weight so that warning is thrown
A__ = jnp.asarray(A )
else:
# also add unexpected weight so that warning is thrown
A__ = jnp.asarray(A )
return unflatten_dict(A )
def __a ( A , A ) -> Optional[Any]:
'''simple docstring'''
import torch
# Load the index
A__ = {}
for shard_file in shard_filenames:
# load using msgpack utils
A__ = torch.load(A )
A__ = {k: v.numpy() for k, v in pt_state_dict.items()}
A__ = flax_model.base_model_prefix
# use params dict if the model contains batch norm layers and then add batch_stats keys,values to dict
if "batch_stats" in flax_model.params:
A__ = flax_model.params["params"]
A__ = flatten_dict(A )
random_flax_state_dict.update(flatten_dict(flax_model.params["batch_stats"] ) )
else:
A__ = flax_model.params
A__ = flatten_dict(A )
A__ = (model_prefix not in flax_model_params) and (
model_prefix in {k.split("." )[0] for k in pt_state_dict.keys()}
)
A__ = (model_prefix in flax_model_params) and (
model_prefix not in {k.split("." )[0] for k in pt_state_dict.keys()}
)
# Need to change some parameters name to match Flax names
for pt_key, pt_tensor in pt_state_dict.items():
A__ = tuple(pt_key.split("." ) )
# remove base model prefix if necessary
A__ = pt_tuple_key[0] == model_prefix
if load_model_with_head_into_base_model and has_base_model_prefix:
A__ = pt_tuple_key[1:]
# Correctly rename weight parameters
A__ , A__ = rename_key_and_reshape_tensor(
A , A , A , A )
# add model prefix if necessary
A__ = (model_prefix,) + flax_key in random_flax_state_dict
if load_base_model_into_model_with_head and require_base_model_prefix:
A__ = (model_prefix,) + flax_key
if flax_key in random_flax_state_dict:
if flax_tensor.shape != random_flax_state_dict[flax_key].shape:
raise ValueError(
f"""PyTorch checkpoint seems to be incorrect. Weight {pt_key} was expected to be of shape """
f"""{random_flax_state_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
# add batch stats if the model contains batchnorm layers
if "batch_stats" in flax_model.params:
if "mean" in flax_key[-1]:
A__ = jnp.asarray(A )
continue
if "var" in flax_key[-1]:
A__ = jnp.asarray(A )
continue
# remove num_batches_tracked key
if "num_batches_tracked" in flax_key[-1]:
flax_state_dict.pop(A , A )
continue
# also add unexpected weight so that warning is thrown
A__ = jnp.asarray(A )
else:
# also add unexpected weight so that warning is thrown
A__ = jnp.asarray(A )
return unflatten_dict(A )
def __a ( A , A ) -> Optional[int]:
'''simple docstring'''
A__ = os.path.abspath(A )
logger.info(f"""Loading Flax weights from {flax_checkpoint_path}""" )
# import correct flax class
A__ = getattr(A , "Flax" + model.__class__.__name__ )
# load flax weight dict
with open(A , "rb" ) as state_f:
try:
A__ = from_bytes(A , state_f.read() )
except UnpicklingError:
raise EnvironmentError(f"""Unable to convert {flax_checkpoint_path} to Flax deserializable object. """ )
return load_flax_weights_in_pytorch_model(A , A )
def __a ( A , A ) -> Union[str, Any]:
'''simple docstring'''
try:
import torch # noqa: F401
except ImportError:
logger.error(
"Loading a Flax weights in PyTorch, requires both PyTorch and Flax to be installed. Please see"
" https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation"
" instructions." )
raise
# check if we have bf16 weights
A__ = flatten_dict(jax.tree_util.tree_map(lambda A : x.dtype == jnp.bfloataa , A ) ).values()
if any(A ):
# convert all weights to fp32 if the are bf16 since torch.from_numpy can-not handle bf16
# and bf16 is not fully supported in PT yet.
logger.warning(
"Found ``bfloat16`` weights in Flax model. Casting all ``bfloat16`` weights to ``float32`` "
"before loading those in PyTorch model." )
A__ = jax.tree_util.tree_map(
lambda A : params.astype(np.floataa ) if params.dtype == jnp.bfloataa else params , A )
A__ = flatten_dict(A )
A__ = pt_model.state_dict()
A__ = (pt_model.base_model_prefix in flax_state) and (
pt_model.base_model_prefix not in {k.split("." )[0] for k in pt_model_dict.keys()}
)
A__ = (pt_model.base_model_prefix not in flax_state) and (
pt_model.base_model_prefix in {k.split("." )[0] for k in pt_model_dict.keys()}
)
# keep track of unexpected & missing keys
A__ = []
A__ = set(pt_model_dict.keys() )
for flax_key_tuple, flax_tensor in flax_state_dict.items():
A__ = flax_key_tuple[0] == pt_model.base_model_prefix
A__ = ".".join((pt_model.base_model_prefix,) + flax_key_tuple ) in pt_model_dict
# adapt flax_key to prepare for loading from/to base model only
if load_model_with_head_into_base_model and has_base_model_prefix:
A__ = flax_key_tuple[1:]
elif load_base_model_into_model_with_head and require_base_model_prefix:
A__ = (pt_model.base_model_prefix,) + flax_key_tuple
# rename flax weights to PyTorch format
if flax_key_tuple[-1] == "kernel" and flax_tensor.ndim == 4 and ".".join(A ) not in pt_model_dict:
# conv layer
A__ = flax_key_tuple[:-1] + ("weight",)
A__ = jnp.transpose(A , (3, 2, 0, 1) )
elif flax_key_tuple[-1] == "kernel" and ".".join(A ) not in pt_model_dict:
# linear layer
A__ = flax_key_tuple[:-1] + ("weight",)
A__ = flax_tensor.T
elif flax_key_tuple[-1] in ["scale", "embedding"]:
A__ = flax_key_tuple[:-1] + ("weight",)
# adding batch stats from flax batch norm to pt
elif "mean" in flax_key_tuple[-1]:
A__ = flax_key_tuple[:-1] + ("running_mean",)
elif "var" in flax_key_tuple[-1]:
A__ = flax_key_tuple[:-1] + ("running_var",)
if "batch_stats" in flax_state:
A__ = ".".join(flax_key_tuple[1:] ) # Remove the params/batch_stats header
else:
A__ = ".".join(A )
# We also need to look at `pt_model_dict` and see if there are keys requiring further transformation.
A__ = {}
# New `weight_norm` from https://github.com/huggingface/transformers/pull/24030
for key in pt_model_dict:
A__ = key.split("." )
A__ = None
if key_components[-3::2] == ["parametrizations", "original0"]:
A__ = key_components[-2] + "_g"
elif key_components[-3::2] == ["parametrizations", "original1"]:
A__ = key_components[-2] + "_v"
if name is not None:
A__ = key_components[:-3] + [name]
A__ = ".".join(A )
A__ = key
if flax_key in special_pt_names:
A__ = special_pt_names[flax_key]
if flax_key in pt_model_dict:
if flax_tensor.shape != pt_model_dict[flax_key].shape:
raise ValueError(
f"""Flax checkpoint seems to be incorrect. Weight {flax_key_tuple} was expected """
f"""to be of shape {pt_model_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
else:
# add weight to pytorch dict
A__ = np.asarray(A ) if not isinstance(A , np.ndarray ) else flax_tensor
A__ = torch.from_numpy(A )
# remove from missing keys
missing_keys.remove(A )
else:
# weight is not expected by PyTorch model
unexpected_keys.append(A )
pt_model.load_state_dict(A )
# re-transform missing_keys to list
A__ = list(A )
if len(A ) > 0:
logger.warning(
"Some weights of the Flax model were not used when initializing the PyTorch model"
f""" {pt_model.__class__.__name__}: {unexpected_keys}\n- This IS expected if you are initializing"""
f""" {pt_model.__class__.__name__} from a Flax model trained on another task or with another architecture"""
" (e.g. initializing a BertForSequenceClassification model from a FlaxBertForPreTraining model).\n- This"
f""" IS NOT expected if you are initializing {pt_model.__class__.__name__} from a Flax model that you expect"""
" to be exactly identical (e.g. initializing a BertForSequenceClassification model from a"
" FlaxBertForSequenceClassification model)." )
else:
logger.warning(f"""All Flax model weights were used when initializing {pt_model.__class__.__name__}.\n""" )
if len(A ) > 0:
logger.warning(
f"""Some weights of {pt_model.__class__.__name__} were not initialized from the Flax model and are newly"""
f""" initialized: {missing_keys}\nYou should probably TRAIN this model on a down-stream task to be able to"""
" use it for predictions and inference." )
else:
logger.warning(
f"""All the weights of {pt_model.__class__.__name__} were initialized from the Flax model.\n"""
"If your task is similar to the task the model of the checkpoint was trained on, "
f"""you can already use {pt_model.__class__.__name__} for predictions without further training.""" )
return pt_model | 337 |
"""simple docstring"""
import unittest
import torch
from diffusers import DDIMScheduler, DDPMScheduler, UNetaDModel
from diffusers.training_utils import set_seed
from diffusers.utils.testing_utils import slow
lowerCAmelCase_ : List[str] = False
class UpperCamelCase_ ( unittest.TestCase ):
def UpperCamelCase_ ( self , snake_case__=32 ) -> Optional[Any]:
"""simple docstring"""
set_seed(0 )
UpperCAmelCase = UNetaDModel(sample_size=snake_case__ , in_channels=3 , out_channels=3 )
UpperCAmelCase = torch.optim.SGD(model.parameters() , lr=0.0_001 )
return model, optimizer
@slow
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = """cpu""" # ensure full determinism without setting the CUBLAS_WORKSPACE_CONFIG env variable
UpperCAmelCase = DDPMScheduler(
num_train_timesteps=10_00 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule="""linear""" , clip_sample=snake_case__ , )
UpperCAmelCase = DDIMScheduler(
num_train_timesteps=10_00 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule="""linear""" , clip_sample=snake_case__ , )
assert ddpm_scheduler.config.num_train_timesteps == ddim_scheduler.config.num_train_timesteps
# shared batches for DDPM and DDIM
set_seed(0 )
UpperCAmelCase = [torch.randn((4, 3, 32, 32) ).clip(-1 , 1 ).to(snake_case__ ) for _ in range(4 )]
UpperCAmelCase = [torch.randn((4, 3, 32, 32) ).to(snake_case__ ) for _ in range(4 )]
UpperCAmelCase = [torch.randint(0 , 10_00 , (4,) ).long().to(snake_case__ ) for _ in range(4 )]
# train with a DDPM scheduler
UpperCAmelCase , UpperCAmelCase = self.get_model_optimizer(resolution=32 )
model.train().to(snake_case__ )
for i in range(4 ):
optimizer.zero_grad()
UpperCAmelCase = ddpm_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i] )
UpperCAmelCase = model(snake_case__ , timesteps[i] ).sample
UpperCAmelCase = torch.nn.functional.mse_loss(snake_case__ , noise[i] )
loss.backward()
optimizer.step()
del model, optimizer
# recreate the model and optimizer, and retry with DDIM
UpperCAmelCase , UpperCAmelCase = self.get_model_optimizer(resolution=32 )
model.train().to(snake_case__ )
for i in range(4 ):
optimizer.zero_grad()
UpperCAmelCase = ddim_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i] )
UpperCAmelCase = model(snake_case__ , timesteps[i] ).sample
UpperCAmelCase = torch.nn.functional.mse_loss(snake_case__ , noise[i] )
loss.backward()
optimizer.step()
del model, optimizer
self.assertTrue(torch.allclose(snake_case__ , snake_case__ , atol=1e-5 ) )
self.assertTrue(torch.allclose(snake_case__ , snake_case__ , atol=1e-5 ) )
| 673 | 0 |
'''simple docstring'''
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
UNetaDConditionModel,
VideoToVideoSDPipeline,
)
from diffusers.utils import floats_tensor, is_xformers_available, skip_mps
from diffusers.utils.testing_utils import enable_full_determinism, slow, torch_device
from ..pipeline_params import (
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineTesterMixin
enable_full_determinism()
@skip_mps
class a__ ( a_ , unittest.TestCase ):
lowerCamelCase : str =VideoToVideoSDPipeline
lowerCamelCase : List[str] =TEXT_GUIDED_IMAGE_VARIATION_PARAMS.union({"video"} ) - {'image', 'width', 'height'}
lowerCamelCase : int =TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({"video"} ) - {'image'}
lowerCamelCase : int =PipelineTesterMixin.required_optional_params - {'latents'}
lowerCamelCase : List[str] =False
# No `output_type`.
lowerCamelCase : Any =frozenset(
[
"num_inference_steps",
"generator",
"latents",
"return_dict",
"callback",
"callback_steps",
] )
def SCREAMING_SNAKE_CASE__ ( self : Optional[int] ):
"""simple docstring"""
torch.manual_seed(0 )
__lowerCamelCase = UNetaDConditionModel(
block_out_channels=(32, 64, 64, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=('''CrossAttnDownBlock3D''', '''CrossAttnDownBlock3D''', '''CrossAttnDownBlock3D''', '''DownBlock3D''') , up_block_types=('''UpBlock3D''', '''CrossAttnUpBlock3D''', '''CrossAttnUpBlock3D''', '''CrossAttnUpBlock3D''') , cross_attention_dim=32 , attention_head_dim=4 , )
__lowerCamelCase = DDIMScheduler(
beta_start=0.0_00_85 , beta_end=0.0_12 , beta_schedule='''scaled_linear''' , clip_sample=snake_case__ , set_alpha_to_one=snake_case__ , )
torch.manual_seed(0 )
__lowerCamelCase = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=['''DownEncoderBlock2D''', '''DownEncoderBlock2D'''] , up_block_types=['''UpDecoderBlock2D''', '''UpDecoderBlock2D'''] , latent_channels=4 , sample_size=1_28 , )
torch.manual_seed(0 )
__lowerCamelCase = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-0_5 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=10_00 , hidden_act='''gelu''' , projection_dim=5_12 , )
__lowerCamelCase = CLIPTextModel(snake_case__ )
__lowerCamelCase = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''' )
__lowerCamelCase = {
'''unet''': unet,
'''scheduler''': scheduler,
'''vae''': vae,
'''text_encoder''': text_encoder,
'''tokenizer''': tokenizer,
}
return components
def SCREAMING_SNAKE_CASE__ ( self : Tuple , a : Union[str, Any] , a : Tuple=0 ):
"""simple docstring"""
__lowerCamelCase = floats_tensor((1, 3, 3, 32, 32) , rng=random.Random(snake_case__ ) ).to(snake_case__ )
if str(snake_case__ ).startswith('''mps''' ):
__lowerCamelCase = torch.manual_seed(snake_case__ )
else:
__lowerCamelCase = torch.Generator(device=snake_case__ ).manual_seed(snake_case__ )
__lowerCamelCase = {
'''prompt''': '''A painting of a squirrel eating a burger''',
'''video''': video,
'''generator''': generator,
'''num_inference_steps''': 2,
'''guidance_scale''': 6.0,
'''output_type''': '''pt''',
}
return inputs
def SCREAMING_SNAKE_CASE__ ( self : str ):
"""simple docstring"""
__lowerCamelCase = '''cpu''' # ensure determinism for the device-dependent torch.Generator
__lowerCamelCase = self.get_dummy_components()
__lowerCamelCase = VideoToVideoSDPipeline(**snake_case__ )
__lowerCamelCase = sd_pipe.to(snake_case__ )
sd_pipe.set_progress_bar_config(disable=snake_case__ )
__lowerCamelCase = self.get_dummy_inputs(snake_case__ )
__lowerCamelCase = '''np'''
__lowerCamelCase = sd_pipe(**snake_case__ ).frames
__lowerCamelCase = frames[0][-3:, -3:, -1]
assert frames[0].shape == (32, 32, 3)
__lowerCamelCase = np.array([1_06, 1_17, 1_13, 1_74, 1_37, 1_12, 1_48, 1_51, 1_31] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(
torch_device != '''cuda''' or not is_xformers_available() , reason='''XFormers attention is only available with CUDA and `xformers` installed''' , )
def SCREAMING_SNAKE_CASE__ ( self : List[str] ):
"""simple docstring"""
self._test_xformers_attention_forwardGenerator_pass(test_mean_pixel_difference=snake_case__ , expected_max_diff=5e-3 )
@unittest.skip(reason='''Batching needs to be properly figured out first for this pipeline.''' )
def SCREAMING_SNAKE_CASE__ ( self : int ):
"""simple docstring"""
pass
@unittest.skip(reason='''Batching needs to be properly figured out first for this pipeline.''' )
def SCREAMING_SNAKE_CASE__ ( self : Optional[int] ):
"""simple docstring"""
pass
@unittest.skip(reason='''`num_images_per_prompt` argument is not supported for this pipeline.''' )
def SCREAMING_SNAKE_CASE__ ( self : str ):
"""simple docstring"""
pass
def SCREAMING_SNAKE_CASE__ ( self : int ):
"""simple docstring"""
return super().test_progress_bar()
@slow
@skip_mps
class a__ ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE__ ( self : List[Any] ):
"""simple docstring"""
__lowerCamelCase = VideoToVideoSDPipeline.from_pretrained('''cerspense/zeroscope_v2_XL''' , torch_dtype=torch.floataa )
pipe.enable_model_cpu_offload()
# 10 frames
__lowerCamelCase = torch.Generator(device='''cpu''' ).manual_seed(0 )
__lowerCamelCase = torch.randn((1, 10, 3, 10_24, 5_76) , generator=snake_case__ )
__lowerCamelCase = video.to('''cuda''' )
__lowerCamelCase = '''Spiderman is surfing'''
__lowerCamelCase = pipe(snake_case__ , video=snake_case__ , generator=snake_case__ , num_inference_steps=3 , output_type='''pt''' ).frames
__lowerCamelCase = np.array([-1.0_45_89_84, -1.1_27_92_97, -0.9_66_30_86, -0.91_50_39_06, -0.75_09_76_56] )
assert np.abs(video_frames.cpu().numpy()[0, 0, 0, 0, -5:] - expected_array ).sum() < 1e-2
| 546 |
"""simple docstring"""
import numpy as np
import torch
from torch.utils.data import DataLoader
from accelerate.utils.dataclasses import DistributedType
class UpperCamelCase_ :
def __init__( self , snake_case__=2 , snake_case__=3 , snake_case__=64 , snake_case__=None ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = np.random.default_rng(snake_case__ )
UpperCAmelCase = length
UpperCAmelCase = rng.normal(size=(length,) ).astype(np.floataa )
UpperCAmelCase = a * self.x + b + rng.normal(scale=0.1 , size=(length,) ).astype(np.floataa )
def __len__( self ) -> int:
"""simple docstring"""
return self.length
def __getitem__( self , snake_case__ ) -> Tuple:
"""simple docstring"""
return {"x": self.x[i], "y": self.y[i]}
class UpperCamelCase_ ( torch.nn.Module ):
def __init__( self , snake_case__=0 , snake_case__=0 , snake_case__=False ) -> List[str]:
"""simple docstring"""
super().__init__()
UpperCAmelCase = torch.nn.Parameter(torch.tensor([2, 3] ).float() )
UpperCAmelCase = torch.nn.Parameter(torch.tensor([2, 3] ).float() )
UpperCAmelCase = True
def UpperCamelCase_ ( self , snake_case__=None ) -> List[Any]:
"""simple docstring"""
if self.first_batch:
print(f'''Model dtype: {self.a.dtype}, {self.b.dtype}. Input dtype: {x.dtype}''' )
UpperCAmelCase = False
return x * self.a[0] + self.b[0]
class UpperCamelCase_ ( torch.nn.Module ):
def __init__( self , snake_case__=0 , snake_case__=0 , snake_case__=False ) -> List[Any]:
"""simple docstring"""
super().__init__()
UpperCAmelCase = torch.nn.Parameter(torch.tensor(snake_case__ ).float() )
UpperCAmelCase = torch.nn.Parameter(torch.tensor(snake_case__ ).float() )
UpperCAmelCase = True
def UpperCamelCase_ ( self , snake_case__=None ) -> Optional[Any]:
"""simple docstring"""
if self.first_batch:
print(f'''Model dtype: {self.a.dtype}, {self.b.dtype}. Input dtype: {x.dtype}''' )
UpperCAmelCase = False
return x * self.a + self.b
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase = 16 ):
'''simple docstring'''
from datasets import load_dataset
from transformers import AutoTokenizer
UpperCAmelCase = AutoTokenizer.from_pretrained("""bert-base-cased""" )
UpperCAmelCase = {"""train""": """tests/test_samples/MRPC/train.csv""", """validation""": """tests/test_samples/MRPC/dev.csv"""}
UpperCAmelCase = load_dataset("""csv""" , data_files=lowerCAmelCase )
UpperCAmelCase = datasets["""train"""].unique("""label""" )
UpperCAmelCase = {v: i for i, v in enumerate(lowerCAmelCase )}
def tokenize_function(lowerCAmelCase ):
# max_length=None => use the model max length (it's actually the default)
UpperCAmelCase = tokenizer(
examples["""sentence1"""] , examples["""sentence2"""] , truncation=lowerCAmelCase , max_length=lowerCAmelCase , padding="""max_length""" )
if "label" in examples:
UpperCAmelCase = [label_to_id[l] for l in examples["""label"""]]
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
UpperCAmelCase = datasets.map(
lowerCAmelCase , batched=lowerCAmelCase , remove_columns=["""sentence1""", """sentence2""", """label"""] , )
def collate_fn(lowerCAmelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
if accelerator.distributed_type == DistributedType.TPU:
return tokenizer.pad(lowerCAmelCase , padding="""max_length""" , max_length=128 , return_tensors="""pt""" )
return tokenizer.pad(lowerCAmelCase , padding="""longest""" , return_tensors="""pt""" )
# Instantiate dataloaders.
UpperCAmelCase = DataLoader(tokenized_datasets["""train"""] , shuffle=lowerCAmelCase , collate_fn=lowerCAmelCase , batch_size=2 )
UpperCAmelCase = DataLoader(tokenized_datasets["""validation"""] , shuffle=lowerCAmelCase , collate_fn=lowerCAmelCase , batch_size=1 )
return train_dataloader, eval_dataloader
| 673 | 0 |
import argparse
import os
from accelerate.test_utils import execute_subprocess_async
def a_ ( SCREAMING_SNAKE_CASE__ : Optional[int]=None ):
'''simple docstring'''
if subparsers is not None:
_lowerCamelCase : Optional[Any] =subparsers.add_parser('test' )
else:
_lowerCamelCase : Any =argparse.ArgumentParser('Accelerate test command' )
parser.add_argument(
'--config_file' , default=SCREAMING_SNAKE_CASE__ , help=(
'The path to use to store the config file. Will default to a file named default_config.yaml in the cache '
'location, which is the content of the environment `HF_HOME` suffixed with \'accelerate\', or if you don\'t have '
'such an environment variable, your cache directory (\'~/.cache\' or the content of `XDG_CACHE_HOME`) suffixed '
'with \'huggingface\'.'
) , )
if subparsers is not None:
parser.set_defaults(func=SCREAMING_SNAKE_CASE__ )
return parser
def a_ ( SCREAMING_SNAKE_CASE__ : Union[str, Any] ):
'''simple docstring'''
_lowerCamelCase : List[Any] =os.path.sep.join(__file__.split(os.path.sep )[:-2] + ['test_utils', 'scripts', 'test_script.py'] )
if args.config_file is None:
_lowerCamelCase : Any =script_name
else:
_lowerCamelCase : Optional[int] =F'''--config_file={args.config_file} {script_name}'''
_lowerCamelCase : List[Any] =['accelerate-launch'] + test_args.split()
_lowerCamelCase : Any =execute_subprocess_async(SCREAMING_SNAKE_CASE__ , env=os.environ.copy() )
if result.returncode == 0:
print('Test is a success! You are ready for your distributed training!' )
def a_ ( ):
'''simple docstring'''
_lowerCamelCase : Union[str, Any] =test_command_parser()
_lowerCamelCase : int =parser.parse_args()
test_command(SCREAMING_SNAKE_CASE__ )
if __name__ == "__main__":
main()
| 464 |
"""simple docstring"""
import flax.linen as nn
import jax
import jax.numpy as jnp
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self , snake_case__ ) -> Tuple:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase = hidden_states.shape
UpperCAmelCase = jax.image.resize(
snake_case__ , shape=(batch, height * 2, width * 2, channels) , method="""nearest""" , )
UpperCAmelCase = self.conv(snake_case__ )
return hidden_states
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self , snake_case__ ) -> Any:
"""simple docstring"""
UpperCAmelCase = self.conv(snake_case__ )
return hidden_states
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : int = None
_A : float = 0.0
_A : bool = None
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = self.in_channels if self.out_channels is None else self.out_channels
UpperCAmelCase = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
UpperCAmelCase = nn.Dense(snake_case__ , dtype=self.dtype )
UpperCAmelCase = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
UpperCAmelCase = nn.Dropout(self.dropout_prob )
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
UpperCAmelCase = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
UpperCAmelCase = None
if use_nin_shortcut:
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(1, 1) , strides=(1, 1) , padding="""VALID""" , dtype=self.dtype , )
def __call__( self , snake_case__ , snake_case__ , snake_case__=True ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = hidden_states
UpperCAmelCase = self.norma(snake_case__ )
UpperCAmelCase = nn.swish(snake_case__ )
UpperCAmelCase = self.conva(snake_case__ )
UpperCAmelCase = self.time_emb_proj(nn.swish(snake_case__ ) )
UpperCAmelCase = jnp.expand_dims(jnp.expand_dims(snake_case__ , 1 ) , 1 )
UpperCAmelCase = hidden_states + temb
UpperCAmelCase = self.norma(snake_case__ )
UpperCAmelCase = nn.swish(snake_case__ )
UpperCAmelCase = self.dropout(snake_case__ , snake_case__ )
UpperCAmelCase = self.conva(snake_case__ )
if self.conv_shortcut is not None:
UpperCAmelCase = self.conv_shortcut(snake_case__ )
return hidden_states + residual
| 673 | 0 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
UpperCAmelCase : List[str] = {
'''configuration_luke''': ['''LUKE_PRETRAINED_CONFIG_ARCHIVE_MAP''', '''LukeConfig'''],
'''tokenization_luke''': ['''LukeTokenizer'''],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCAmelCase : int = [
'''LUKE_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''LukeForEntityClassification''',
'''LukeForEntityPairClassification''',
'''LukeForEntitySpanClassification''',
'''LukeForMultipleChoice''',
'''LukeForQuestionAnswering''',
'''LukeForSequenceClassification''',
'''LukeForTokenClassification''',
'''LukeForMaskedLM''',
'''LukeModel''',
'''LukePreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_luke import LUKE_PRETRAINED_CONFIG_ARCHIVE_MAP, LukeConfig
from .tokenization_luke import LukeTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_luke import (
LUKE_PRETRAINED_MODEL_ARCHIVE_LIST,
LukeForEntityClassification,
LukeForEntityPairClassification,
LukeForEntitySpanClassification,
LukeForMaskedLM,
LukeForMultipleChoice,
LukeForQuestionAnswering,
LukeForSequenceClassification,
LukeForTokenClassification,
LukeModel,
LukePreTrainedModel,
)
else:
import sys
UpperCAmelCase : List[str] = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 567 |
"""simple docstring"""
from __future__ import annotations
import inspect
import unittest
from transformers import ViTConfig
from transformers.testing_utils import require_tf, require_vision, slow
from transformers.utils import cached_property, is_tf_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFViTForImageClassification, TFViTModel
if is_vision_available():
from PIL import Image
from transformers import ViTImageProcessor
class UpperCamelCase_ :
def __init__( self , snake_case__ , snake_case__=13 , snake_case__=30 , snake_case__=2 , snake_case__=3 , snake_case__=True , snake_case__=True , snake_case__=32 , snake_case__=2 , snake_case__=4 , snake_case__=37 , snake_case__="gelu" , snake_case__=0.1 , snake_case__=0.1 , snake_case__=10 , snake_case__=0.02 , snake_case__=3 , snake_case__=None , ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = parent
UpperCAmelCase = batch_size
UpperCAmelCase = image_size
UpperCAmelCase = patch_size
UpperCAmelCase = num_channels
UpperCAmelCase = is_training
UpperCAmelCase = use_labels
UpperCAmelCase = hidden_size
UpperCAmelCase = num_hidden_layers
UpperCAmelCase = num_attention_heads
UpperCAmelCase = intermediate_size
UpperCAmelCase = hidden_act
UpperCAmelCase = hidden_dropout_prob
UpperCAmelCase = attention_probs_dropout_prob
UpperCAmelCase = type_sequence_label_size
UpperCAmelCase = initializer_range
UpperCAmelCase = scope
# in ViT, the seq length equals the number of patches + 1 (we add 1 for the [CLS] token)
UpperCAmelCase = (image_size // patch_size) ** 2
UpperCAmelCase = num_patches + 1
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
UpperCAmelCase = None
if self.use_labels:
UpperCAmelCase = ids_tensor([self.batch_size] , self.type_sequence_label_size )
UpperCAmelCase = self.get_config()
return config, pixel_values, labels
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
return ViTConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , is_decoder=snake_case__ , initializer_range=self.initializer_range , )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ ) -> Dict:
"""simple docstring"""
UpperCAmelCase = TFViTModel(config=snake_case__ )
UpperCAmelCase = model(snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
# Test with an image with different size than the one specified in config.
UpperCAmelCase = self.image_size // 2
UpperCAmelCase = pixel_values[:, :, :image_size, :image_size]
UpperCAmelCase = model(snake_case__ , interpolate_pos_encoding=snake_case__ , training=snake_case__ )
UpperCAmelCase = (image_size // self.patch_size) ** 2 + 1
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, seq_length, self.hidden_size) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = self.type_sequence_label_size
UpperCAmelCase = TFViTForImageClassification(snake_case__ )
UpperCAmelCase = model(snake_case__ , labels=snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# Test with an image with different size than the one specified in config.
UpperCAmelCase = self.image_size // 2
UpperCAmelCase = pixel_values[:, :, :image_size, :image_size]
UpperCAmelCase = model(snake_case__ , interpolate_pos_encoding=snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# test greyscale images
UpperCAmelCase = 1
UpperCAmelCase = TFViTForImageClassification(snake_case__ )
UpperCAmelCase = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
UpperCAmelCase = model(snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
UpperCAmelCase = self.prepare_config_and_inputs()
UpperCAmelCase , UpperCAmelCase , UpperCAmelCase = config_and_inputs
UpperCAmelCase = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_tf
class UpperCamelCase_ ( a_ , a_ , unittest.TestCase ):
_A : Optional[int] = (TFViTModel, TFViTForImageClassification) if is_tf_available() else ()
_A : Optional[Any] = (
{'feature-extraction': TFViTModel, 'image-classification': TFViTForImageClassification}
if is_tf_available()
else {}
)
_A : Optional[int] = False
_A : Any = False
_A : List[str] = False
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = TFViTModelTester(self )
UpperCAmelCase = ConfigTester(self , config_class=snake_case__ , has_text_modality=snake_case__ , hidden_size=37 )
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
self.config_tester.run_common_tests()
@unittest.skip(reason="""ViT does not use inputs_embeds""" )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
pass
@unittest.skip(reason="""ViT does not use inputs_embeds""" )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
pass
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
UpperCAmelCase = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (tf.keras.layers.Layer) )
UpperCAmelCase = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , tf.keras.layers.Layer ) )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
UpperCAmelCase = model_class(snake_case__ )
UpperCAmelCase = inspect.signature(model.call )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
UpperCAmelCase = [*signature.parameters.keys()]
UpperCAmelCase = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] , snake_case__ )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*snake_case__ )
@slow
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = TFViTModel.from_pretrained("""google/vit-base-patch16-224""" )
self.assertIsNotNone(snake_case__ )
def _lowerCAmelCase ( ):
'''simple docstring'''
UpperCAmelCase = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
return image
@require_tf
@require_vision
class UpperCamelCase_ ( unittest.TestCase ):
@cached_property
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return ViTImageProcessor.from_pretrained("""google/vit-base-patch16-224""" ) if is_vision_available() else None
@slow
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = TFViTForImageClassification.from_pretrained("""google/vit-base-patch16-224""" )
UpperCAmelCase = self.default_image_processor
UpperCAmelCase = prepare_img()
UpperCAmelCase = image_processor(images=snake_case__ , return_tensors="""tf""" )
# forward pass
UpperCAmelCase = model(**snake_case__ )
# verify the logits
UpperCAmelCase = tf.TensorShape((1, 10_00) )
self.assertEqual(outputs.logits.shape , snake_case__ )
UpperCAmelCase = tf.constant([-0.2_744, 0.8_215, -0.0_836] )
tf.debugging.assert_near(outputs.logits[0, :3] , snake_case__ , atol=1e-4 )
| 673 | 0 |
'''simple docstring'''
import unittest
import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import MobileNetVaImageProcessor
class _UpperCamelCase ( unittest.TestCase ):
'''simple docstring'''
def __init__( self : int , lowerCAmelCase__ : List[str] , lowerCAmelCase__ : List[str]=7 , lowerCAmelCase__ : List[str]=3 , lowerCAmelCase__ : Dict=1_8 , lowerCAmelCase__ : Dict=3_0 , lowerCAmelCase__ : Optional[Any]=4_0_0 , lowerCAmelCase__ : Optional[Any]=True , lowerCAmelCase__ : str=None , lowerCAmelCase__ : Union[str, Any]=True , lowerCAmelCase__ : Dict=None , ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : Tuple = size if size is not None else {"""shortest_edge""": 2_0}
__SCREAMING_SNAKE_CASE : List[str] = crop_size if crop_size is not None else {"""height""": 1_8, """width""": 1_8}
__SCREAMING_SNAKE_CASE : Any = parent
__SCREAMING_SNAKE_CASE : Dict = batch_size
__SCREAMING_SNAKE_CASE : List[Any] = num_channels
__SCREAMING_SNAKE_CASE : Optional[int] = image_size
__SCREAMING_SNAKE_CASE : List[Any] = min_resolution
__SCREAMING_SNAKE_CASE : List[Any] = max_resolution
__SCREAMING_SNAKE_CASE : List[Any] = do_resize
__SCREAMING_SNAKE_CASE : Tuple = size
__SCREAMING_SNAKE_CASE : Optional[Any] = do_center_crop
__SCREAMING_SNAKE_CASE : Dict = crop_size
def UpperCamelCase__ ( self : Any ):
"""simple docstring"""
return {
"do_resize": self.do_resize,
"size": self.size,
"do_center_crop": self.do_center_crop,
"crop_size": self.crop_size,
}
@require_torch
@require_vision
class _UpperCamelCase ( a_ , unittest.TestCase ):
'''simple docstring'''
_A : Tuple = MobileNetVaImageProcessor if is_vision_available() else None
def UpperCamelCase__ ( self : Optional[Any] ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : int = MobileNetVaImageProcessingTester(self )
@property
def UpperCamelCase__ ( self : List[str] ):
"""simple docstring"""
return self.image_processor_tester.prepare_image_processor_dict()
def UpperCamelCase__ ( self : Union[str, Any] ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : Any = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(snake_case__ , """do_resize""" ) )
self.assertTrue(hasattr(snake_case__ , """size""" ) )
self.assertTrue(hasattr(snake_case__ , """do_center_crop""" ) )
self.assertTrue(hasattr(snake_case__ , """crop_size""" ) )
def UpperCamelCase__ ( self : str ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : List[str] = self.image_processing_class.from_dict(self.image_processor_dict )
self.assertEqual(image_processor.size , {"""shortest_edge""": 2_0} )
self.assertEqual(image_processor.crop_size , {"""height""": 1_8, """width""": 1_8} )
__SCREAMING_SNAKE_CASE : int = self.image_processing_class.from_dict(self.image_processor_dict , size=4_2 , crop_size=8_4 )
self.assertEqual(image_processor.size , {"""shortest_edge""": 4_2} )
self.assertEqual(image_processor.crop_size , {"""height""": 8_4, """width""": 8_4} )
def UpperCamelCase__ ( self : str ):
"""simple docstring"""
pass
def UpperCamelCase__ ( self : Dict ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : Optional[Any] = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
__SCREAMING_SNAKE_CASE : Optional[int] = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , Image.Image )
# Test not batched input
__SCREAMING_SNAKE_CASE : Dict = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
self.assertEqual(
encoded_images.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["""height"""],
self.image_processor_tester.crop_size["""width"""],
) , )
# Test batched
__SCREAMING_SNAKE_CASE : str = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["""height"""],
self.image_processor_tester.crop_size["""width"""],
) , )
def UpperCamelCase__ ( self : Dict ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : Dict = self.image_processing_class(**self.image_processor_dict )
# create random numpy tensors
__SCREAMING_SNAKE_CASE : Optional[int] = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ , numpify=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , np.ndarray )
# Test not batched input
__SCREAMING_SNAKE_CASE : Union[str, Any] = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
self.assertEqual(
encoded_images.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["""height"""],
self.image_processor_tester.crop_size["""width"""],
) , )
# Test batched
__SCREAMING_SNAKE_CASE : int = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["""height"""],
self.image_processor_tester.crop_size["""width"""],
) , )
def UpperCamelCase__ ( self : int ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : List[Any] = self.image_processing_class(**self.image_processor_dict )
# create random PyTorch tensors
__SCREAMING_SNAKE_CASE : Optional[int] = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ , torchify=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , torch.Tensor )
# Test not batched input
__SCREAMING_SNAKE_CASE : Optional[int] = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
self.assertEqual(
encoded_images.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["""height"""],
self.image_processor_tester.crop_size["""width"""],
) , )
# Test batched
__SCREAMING_SNAKE_CASE : int = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["""height"""],
self.image_processor_tester.crop_size["""width"""],
) , ) | 578 |
"""simple docstring"""
import unittest
from transformers import AutoTokenizer, NystromformerConfig, is_torch_available
from transformers.testing_utils import require_torch, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
NystromformerForMaskedLM,
NystromformerForMultipleChoice,
NystromformerForQuestionAnswering,
NystromformerForSequenceClassification,
NystromformerForTokenClassification,
NystromformerModel,
)
from transformers.models.nystromformer.modeling_nystromformer import NYSTROMFORMER_PRETRAINED_MODEL_ARCHIVE_LIST
class UpperCamelCase_ :
def __init__( self , snake_case__ , snake_case__=13 , snake_case__=7 , snake_case__=True , snake_case__=True , snake_case__=True , snake_case__=True , snake_case__=99 , snake_case__=32 , snake_case__=5 , snake_case__=4 , snake_case__=37 , snake_case__="gelu" , snake_case__=0.1 , snake_case__=0.1 , snake_case__=5_12 , snake_case__=16 , snake_case__=2 , snake_case__=0.02 , snake_case__=3 , snake_case__=4 , snake_case__=None , ) -> int:
"""simple docstring"""
UpperCAmelCase = parent
UpperCAmelCase = batch_size
UpperCAmelCase = seq_length
UpperCAmelCase = is_training
UpperCAmelCase = use_input_mask
UpperCAmelCase = use_token_type_ids
UpperCAmelCase = use_labels
UpperCAmelCase = vocab_size
UpperCAmelCase = hidden_size
UpperCAmelCase = num_hidden_layers
UpperCAmelCase = num_attention_heads
UpperCAmelCase = intermediate_size
UpperCAmelCase = hidden_act
UpperCAmelCase = hidden_dropout_prob
UpperCAmelCase = attention_probs_dropout_prob
UpperCAmelCase = max_position_embeddings
UpperCAmelCase = type_vocab_size
UpperCAmelCase = type_sequence_label_size
UpperCAmelCase = initializer_range
UpperCAmelCase = num_labels
UpperCAmelCase = num_choices
UpperCAmelCase = scope
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
UpperCAmelCase = None
if self.use_input_mask:
UpperCAmelCase = random_attention_mask([self.batch_size, self.seq_length] )
UpperCAmelCase = None
if self.use_token_type_ids:
UpperCAmelCase = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
UpperCAmelCase = None
UpperCAmelCase = None
UpperCAmelCase = None
if self.use_labels:
UpperCAmelCase = ids_tensor([self.batch_size] , self.type_sequence_label_size )
UpperCAmelCase = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
UpperCAmelCase = ids_tensor([self.batch_size] , self.num_choices )
UpperCAmelCase = self.get_config()
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
return NystromformerConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=snake_case__ , initializer_range=self.initializer_range , )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = NystromformerModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ )
UpperCAmelCase = model(snake_case__ , token_type_ids=snake_case__ )
UpperCAmelCase = model(snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> int:
"""simple docstring"""
UpperCAmelCase = NystromformerForMaskedLM(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> Dict:
"""simple docstring"""
UpperCAmelCase = NystromformerForQuestionAnswering(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , start_positions=snake_case__ , end_positions=snake_case__ , )
self.parent.assertEqual(result.start_logits.shape , (self.batch_size, self.seq_length) )
self.parent.assertEqual(result.end_logits.shape , (self.batch_size, self.seq_length) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = self.num_labels
UpperCAmelCase = NystromformerForSequenceClassification(snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> int:
"""simple docstring"""
UpperCAmelCase = self.num_labels
UpperCAmelCase = NystromformerForTokenClassification(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> Dict:
"""simple docstring"""
UpperCAmelCase = self.num_choices
UpperCAmelCase = NystromformerForMultipleChoice(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
UpperCAmelCase = token_type_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
UpperCAmelCase = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
UpperCAmelCase = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = self.prepare_config_and_inputs()
(
(
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) ,
) = config_and_inputs
UpperCAmelCase = {"""input_ids""": input_ids, """token_type_ids""": token_type_ids, """attention_mask""": input_mask}
return config, inputs_dict
@require_torch
class UpperCamelCase_ ( a_ , a_ , unittest.TestCase ):
_A : Optional[Any] = (
(
NystromformerModel,
NystromformerForMaskedLM,
NystromformerForMultipleChoice,
NystromformerForQuestionAnswering,
NystromformerForSequenceClassification,
NystromformerForTokenClassification,
)
if is_torch_available()
else ()
)
_A : Optional[Any] = (
{
'feature-extraction': NystromformerModel,
'fill-mask': NystromformerForMaskedLM,
'question-answering': NystromformerForQuestionAnswering,
'text-classification': NystromformerForSequenceClassification,
'token-classification': NystromformerForTokenClassification,
'zero-shot': NystromformerForSequenceClassification,
}
if is_torch_available()
else {}
)
_A : int = False
_A : Dict = False
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = NystromformerModelTester(self )
UpperCAmelCase = ConfigTester(self , config_class=snake_case__ , hidden_size=37 )
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
self.config_tester.run_common_tests()
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
for type in ["absolute", "relative_key", "relative_key_query"]:
UpperCAmelCase = type
self.model_tester.create_and_check_model(*snake_case__ )
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_lm(*snake_case__ )
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_multiple_choice(*snake_case__ )
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_question_answering(*snake_case__ )
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_sequence_classification(*snake_case__ )
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_token_classification(*snake_case__ )
@slow
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
for model_name in NYSTROMFORMER_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
UpperCAmelCase = NystromformerModel.from_pretrained(snake_case__ )
self.assertIsNotNone(snake_case__ )
@require_torch
class UpperCamelCase_ ( unittest.TestCase ):
@slow
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = NystromformerModel.from_pretrained("""uw-madison/nystromformer-512""" )
UpperCAmelCase = torch.tensor([[0, 1, 2, 3, 4, 5]] )
with torch.no_grad():
UpperCAmelCase = model(snake_case__ )[0]
UpperCAmelCase = torch.Size((1, 6, 7_68) )
self.assertEqual(output.shape , snake_case__ )
UpperCAmelCase = torch.tensor(
[[[-0.4_532, -0.0_936, 0.5_137], [-0.2_676, 0.0_628, 0.6_186], [-0.3_629, -0.1_726, 0.4_716]]] )
self.assertTrue(torch.allclose(output[:, :3, :3] , snake_case__ , atol=1e-4 ) )
@slow
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = """the [MASK] of Belgium is Brussels"""
UpperCAmelCase = AutoTokenizer.from_pretrained("""uw-madison/nystromformer-512""" )
UpperCAmelCase = NystromformerForMaskedLM.from_pretrained("""uw-madison/nystromformer-512""" )
UpperCAmelCase = tokenizer(snake_case__ , return_tensors="""pt""" )
with torch.no_grad():
UpperCAmelCase = model(encoding.input_ids ).logits
UpperCAmelCase = token_logits[:, 2, :].argmax(-1 )[0]
self.assertEqual(tokenizer.decode(snake_case__ ) , """capital""" )
| 673 | 0 |
"""simple docstring"""
import math
def __lowercase ( _a ):
if 1 < number < 4:
# 2 and 3 are primes
return True
elif number < 2 or number % 2 == 0 or number % 3 == 0:
# Negatives, 0, 1, all even numbers, all multiples of 3 are not primes
return False
# All primes number are in format of 6k +/- 1
for i in range(5 , int(math.sqrt(_a ) + 1 ) , 6 ):
if number % i == 0 or number % (i + 2) == 0:
return False
return True
def __lowercase ( _a = 10_001 ):
try:
snake_case_ : str = int(_a )
except (TypeError, ValueError):
raise TypeError('''Parameter nth must be int or castable to int.''' ) from None
if nth <= 0:
raise ValueError('''Parameter nth must be greater than or equal to one.''' )
snake_case_ : Any = []
snake_case_ : Union[str, Any] = 2
while len(_a ) < nth:
if is_prime(_a ):
primes.append(_a )
num += 1
else:
num += 1
return primes[len(_a ) - 1]
if __name__ == "__main__":
print(f'{solution() = }')
| 123 |
"""simple docstring"""
import os
from argparse import ArgumentParser, Namespace
from ..data import SingleSentenceClassificationProcessor as Processor
from ..pipelines import TextClassificationPipeline
from ..utils import is_tf_available, is_torch_available, logging
from . import BaseTransformersCLICommand
if not is_tf_available() and not is_torch_available():
raise RuntimeError('''At least one of PyTorch or TensorFlow 2.0+ should be installed to use CLI training''')
# TF training parameters
lowerCAmelCase_ : Optional[int] = False
lowerCAmelCase_ : Optional[int] = False
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
return TrainCommand(lowerCAmelCase )
class UpperCamelCase_ ( a_ ):
@staticmethod
def UpperCamelCase_ ( snake_case__ ) -> int:
"""simple docstring"""
UpperCAmelCase = parser.add_parser("""train""" , help="""CLI tool to train a model on a task.""" )
train_parser.add_argument(
"""--train_data""" , type=snake_case__ , required=snake_case__ , help="""path to train (and optionally evaluation) dataset as a csv with tab separated labels and sentences.""" , )
train_parser.add_argument(
"""--column_label""" , type=snake_case__ , default=0 , help="""Column of the dataset csv file with example labels.""" )
train_parser.add_argument(
"""--column_text""" , type=snake_case__ , default=1 , help="""Column of the dataset csv file with example texts.""" )
train_parser.add_argument(
"""--column_id""" , type=snake_case__ , default=2 , help="""Column of the dataset csv file with example ids.""" )
train_parser.add_argument(
"""--skip_first_row""" , action="""store_true""" , help="""Skip the first row of the csv file (headers).""" )
train_parser.add_argument("""--validation_data""" , type=snake_case__ , default="""""" , help="""path to validation dataset.""" )
train_parser.add_argument(
"""--validation_split""" , type=snake_case__ , default=0.1 , help="""if validation dataset is not provided, fraction of train dataset to use as validation dataset.""" , )
train_parser.add_argument("""--output""" , type=snake_case__ , default="""./""" , help="""path to saved the trained model.""" )
train_parser.add_argument(
"""--task""" , type=snake_case__ , default="""text_classification""" , help="""Task to train the model on.""" )
train_parser.add_argument(
"""--model""" , type=snake_case__ , default="""bert-base-uncased""" , help="""Model's name or path to stored model.""" )
train_parser.add_argument("""--train_batch_size""" , type=snake_case__ , default=32 , help="""Batch size for training.""" )
train_parser.add_argument("""--valid_batch_size""" , type=snake_case__ , default=64 , help="""Batch size for validation.""" )
train_parser.add_argument("""--learning_rate""" , type=snake_case__ , default=3e-5 , help="""Learning rate.""" )
train_parser.add_argument("""--adam_epsilon""" , type=snake_case__ , default=1e-08 , help="""Epsilon for Adam optimizer.""" )
train_parser.set_defaults(func=snake_case__ )
def __init__( self , snake_case__ ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = logging.get_logger("""transformers-cli/training""" )
UpperCAmelCase = """tf""" if is_tf_available() else """torch"""
os.makedirs(args.output , exist_ok=snake_case__ )
UpperCAmelCase = args.output
UpperCAmelCase = args.column_label
UpperCAmelCase = args.column_text
UpperCAmelCase = args.column_id
self.logger.info(f'''Loading {args.task} pipeline for {args.model}''' )
if args.task == "text_classification":
UpperCAmelCase = TextClassificationPipeline.from_pretrained(args.model )
elif args.task == "token_classification":
raise NotImplementedError
elif args.task == "question_answering":
raise NotImplementedError
self.logger.info(f'''Loading dataset from {args.train_data}''' )
UpperCAmelCase = Processor.create_from_csv(
args.train_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase = None
if args.validation_data:
self.logger.info(f'''Loading validation dataset from {args.validation_data}''' )
UpperCAmelCase = Processor.create_from_csv(
args.validation_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase = args.validation_split
UpperCAmelCase = args.train_batch_size
UpperCAmelCase = args.valid_batch_size
UpperCAmelCase = args.learning_rate
UpperCAmelCase = args.adam_epsilon
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
if self.framework == "tf":
return self.run_tf()
return self.run_torch()
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
raise NotImplementedError
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
self.pipeline.fit(
self.train_dataset , validation_data=self.valid_dataset , validation_split=self.validation_split , learning_rate=self.learning_rate , adam_epsilon=self.adam_epsilon , train_batch_size=self.train_batch_size , valid_batch_size=self.valid_batch_size , )
# Save trained pipeline
self.pipeline.save_pretrained(self.output )
| 673 | 0 |
'''simple docstring'''
from __future__ import annotations
import inspect
import unittest
from transformers import ViTConfig
from transformers.testing_utils import require_tf, require_vision, slow
from transformers.utils import cached_property, is_tf_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFViTForImageClassification, TFViTModel
if is_vision_available():
from PIL import Image
from transformers import ViTImageProcessor
class __lowerCamelCase :
'''simple docstring'''
def __init__( self , a__ , a__=13 , a__=30 , a__=2 , a__=3 , a__=True , a__=True , a__=32 , a__=2 , a__=4 , a__=37 , a__="gelu" , a__=0.1 , a__=0.1 , a__=10 , a__=0.02 , a__=3 , a__=None , ):
__SCREAMING_SNAKE_CASE : int = parent
__SCREAMING_SNAKE_CASE : Tuple = batch_size
__SCREAMING_SNAKE_CASE : str = image_size
__SCREAMING_SNAKE_CASE : Tuple = patch_size
__SCREAMING_SNAKE_CASE : str = num_channels
__SCREAMING_SNAKE_CASE : Union[str, Any] = is_training
__SCREAMING_SNAKE_CASE : Tuple = use_labels
__SCREAMING_SNAKE_CASE : int = hidden_size
__SCREAMING_SNAKE_CASE : Any = num_hidden_layers
__SCREAMING_SNAKE_CASE : Dict = num_attention_heads
__SCREAMING_SNAKE_CASE : List[Any] = intermediate_size
__SCREAMING_SNAKE_CASE : Optional[Any] = hidden_act
__SCREAMING_SNAKE_CASE : List[Any] = hidden_dropout_prob
__SCREAMING_SNAKE_CASE : List[Any] = attention_probs_dropout_prob
__SCREAMING_SNAKE_CASE : Union[str, Any] = type_sequence_label_size
__SCREAMING_SNAKE_CASE : List[Any] = initializer_range
__SCREAMING_SNAKE_CASE : Optional[Any] = scope
# in ViT, the seq length equals the number of patches + 1 (we add 1 for the [CLS] token)
__SCREAMING_SNAKE_CASE : Any = (image_size // patch_size) ** 2
__SCREAMING_SNAKE_CASE : Optional[int] = num_patches + 1
def a_ ( self ):
__SCREAMING_SNAKE_CASE : Optional[Any] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
__SCREAMING_SNAKE_CASE : List[str] = None
if self.use_labels:
__SCREAMING_SNAKE_CASE : Any = ids_tensor([self.batch_size] , self.type_sequence_label_size )
__SCREAMING_SNAKE_CASE : Dict = self.get_config()
return config, pixel_values, labels
def a_ ( self ):
return ViTConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , is_decoder=snake_case__ , initializer_range=self.initializer_range , )
def a_ ( self , a__ , a__ , a__ ):
__SCREAMING_SNAKE_CASE : str = TFViTModel(config=snake_case__ )
__SCREAMING_SNAKE_CASE : int = model(snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
# Test with an image with different size than the one specified in config.
__SCREAMING_SNAKE_CASE : Optional[int] = self.image_size // 2
__SCREAMING_SNAKE_CASE : Optional[int] = pixel_values[:, :, :image_size, :image_size]
__SCREAMING_SNAKE_CASE : Any = model(snake_case__ , interpolate_pos_encoding=snake_case__ , training=snake_case__ )
__SCREAMING_SNAKE_CASE : Union[str, Any] = (image_size // self.patch_size) ** 2 + 1
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, seq_length, self.hidden_size) )
def a_ ( self , a__ , a__ , a__ ):
__SCREAMING_SNAKE_CASE : List[str] = self.type_sequence_label_size
__SCREAMING_SNAKE_CASE : Dict = TFViTForImageClassification(snake_case__ )
__SCREAMING_SNAKE_CASE : Any = model(snake_case__ , labels=snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# Test with an image with different size than the one specified in config.
__SCREAMING_SNAKE_CASE : Optional[int] = self.image_size // 2
__SCREAMING_SNAKE_CASE : Dict = pixel_values[:, :, :image_size, :image_size]
__SCREAMING_SNAKE_CASE : Tuple = model(snake_case__ , interpolate_pos_encoding=snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# test greyscale images
__SCREAMING_SNAKE_CASE : Optional[int] = 1
__SCREAMING_SNAKE_CASE : Union[str, Any] = TFViTForImageClassification(snake_case__ )
__SCREAMING_SNAKE_CASE : Any = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
__SCREAMING_SNAKE_CASE : str = model(snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def a_ ( self ):
__SCREAMING_SNAKE_CASE : Optional[int] = self.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[int] = config_and_inputs
__SCREAMING_SNAKE_CASE : Dict = {"pixel_values": pixel_values}
return config, inputs_dict
@require_tf
class __lowerCamelCase ( a_ , a_ , unittest.TestCase ):
'''simple docstring'''
snake_case__ : Optional[int] = (TFViTModel, TFViTForImageClassification) if is_tf_available() else ()
snake_case__ : Optional[Any] = (
{'feature-extraction': TFViTModel, 'image-classification': TFViTForImageClassification}
if is_tf_available()
else {}
)
snake_case__ : Optional[int] = False
snake_case__ : Any = False
snake_case__ : List[str] = False
def a_ ( self ):
__SCREAMING_SNAKE_CASE : Optional[int] = TFViTModelTester(self )
__SCREAMING_SNAKE_CASE : List[str] = ConfigTester(self , config_class=snake_case__ , has_text_modality=snake_case__ , hidden_size=37 )
def a_ ( self ):
self.config_tester.run_common_tests()
@unittest.skip(reason="ViT does not use inputs_embeds" )
def a_ ( self ):
pass
@unittest.skip(reason="ViT does not use inputs_embeds" )
def a_ ( self ):
pass
def a_ ( self ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Any = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (tf.keras.layers.Layer) )
__SCREAMING_SNAKE_CASE : Optional[Any] = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , tf.keras.layers.Layer ) )
def a_ ( self ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Union[str, Any] = model_class(snake_case__ )
__SCREAMING_SNAKE_CASE : Tuple = inspect.signature(model.call )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
__SCREAMING_SNAKE_CASE : str = [*signature.parameters.keys()]
__SCREAMING_SNAKE_CASE : Optional[Any] = ["pixel_values"]
self.assertListEqual(arg_names[:1] , snake_case__ )
def a_ ( self ):
__SCREAMING_SNAKE_CASE : List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def a_ ( self ):
__SCREAMING_SNAKE_CASE : List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*snake_case__ )
@slow
def a_ ( self ):
__SCREAMING_SNAKE_CASE : Any = TFViTModel.from_pretrained("google/vit-base-patch16-224" )
self.assertIsNotNone(snake_case__ )
def __A ( ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : Tuple = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png" )
return image
@require_tf
@require_vision
class __lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
@cached_property
def a_ ( self ):
return ViTImageProcessor.from_pretrained("google/vit-base-patch16-224" ) if is_vision_available() else None
@slow
def a_ ( self ):
__SCREAMING_SNAKE_CASE : str = TFViTForImageClassification.from_pretrained("google/vit-base-patch16-224" )
__SCREAMING_SNAKE_CASE : List[Any] = self.default_image_processor
__SCREAMING_SNAKE_CASE : List[Any] = prepare_img()
__SCREAMING_SNAKE_CASE : Dict = image_processor(images=snake_case__ , return_tensors="tf" )
# forward pass
__SCREAMING_SNAKE_CASE : Tuple = model(**snake_case__ )
# verify the logits
__SCREAMING_SNAKE_CASE : List[str] = tf.TensorShape((1, 1000) )
self.assertEqual(outputs.logits.shape , snake_case__ )
__SCREAMING_SNAKE_CASE : Optional[Any] = tf.constant([-0.2744, 0.8215, -0.0836] )
tf.debugging.assert_near(outputs.logits[0, :3] , snake_case__ , atol=1e-4 )
| 211 |
"""simple docstring"""
import sys
from typing import Tuple
import numpy as np
import torch
from PIL import Image
from torch import nn
from transformers.image_utils import PILImageResampling
from utils import img_tensorize
class UpperCamelCase_ :
def __init__( self , snake_case__ , snake_case__=sys.maxsize ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = """bilinear"""
UpperCAmelCase = max_size
UpperCAmelCase = short_edge_length
def __call__( self , snake_case__ ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = []
for img in imgs:
UpperCAmelCase , UpperCAmelCase = img.shape[:2]
# later: provide list and randomly choose index for resize
UpperCAmelCase = np.random.randint(self.short_edge_length[0] , self.short_edge_length[1] + 1 )
if size == 0:
return img
UpperCAmelCase = size * 1.0 / min(snake_case__ , snake_case__ )
if h < w:
UpperCAmelCase , UpperCAmelCase = size, scale * w
else:
UpperCAmelCase , UpperCAmelCase = scale * h, size
if max(snake_case__ , snake_case__ ) > self.max_size:
UpperCAmelCase = self.max_size * 1.0 / max(snake_case__ , snake_case__ )
UpperCAmelCase = newh * scale
UpperCAmelCase = neww * scale
UpperCAmelCase = int(neww + 0.5 )
UpperCAmelCase = int(newh + 0.5 )
if img.dtype == np.uinta:
UpperCAmelCase = Image.fromarray(snake_case__ )
UpperCAmelCase = pil_image.resize((neww, newh) , PILImageResampling.BILINEAR )
UpperCAmelCase = np.asarray(snake_case__ )
else:
UpperCAmelCase = img.permute(2 , 0 , 1 ).unsqueeze(0 ) # 3, 0, 1) # hw(c) -> nchw
UpperCAmelCase = nn.functional.interpolate(
snake_case__ , (newh, neww) , mode=self.interp_method , align_corners=snake_case__ ).squeeze(0 )
img_augs.append(snake_case__ )
return img_augs
class UpperCamelCase_ :
def __init__( self , snake_case__ ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = ResizeShortestEdge([cfg.INPUT.MIN_SIZE_TEST, cfg.INPUT.MIN_SIZE_TEST] , cfg.INPUT.MAX_SIZE_TEST )
UpperCAmelCase = cfg.INPUT.FORMAT
UpperCAmelCase = cfg.SIZE_DIVISIBILITY
UpperCAmelCase = cfg.PAD_VALUE
UpperCAmelCase = cfg.INPUT.MAX_SIZE_TEST
UpperCAmelCase = cfg.MODEL.DEVICE
UpperCAmelCase = torch.tensor(cfg.MODEL.PIXEL_STD ).to(self.device ).view(len(cfg.MODEL.PIXEL_STD ) , 1 , 1 )
UpperCAmelCase = torch.tensor(cfg.MODEL.PIXEL_MEAN ).to(self.device ).view(len(cfg.MODEL.PIXEL_STD ) , 1 , 1 )
UpperCAmelCase = lambda snake_case__ : (x - self.pixel_mean) / self.pixel_std
def UpperCamelCase_ ( self , snake_case__ ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = tuple(max(snake_case__ ) for s in zip(*[img.shape for img in images] ) )
UpperCAmelCase = [im.shape[-2:] for im in images]
UpperCAmelCase = [
nn.functional.pad(
snake_case__ , [0, max_size[-1] - size[1], 0, max_size[-2] - size[0]] , value=self.pad_value , )
for size, im in zip(snake_case__ , snake_case__ )
]
return torch.stack(snake_case__ ), torch.tensor(snake_case__ )
def __call__( self , snake_case__ , snake_case__=False ) -> Optional[Any]:
"""simple docstring"""
with torch.no_grad():
if not isinstance(snake_case__ , snake_case__ ):
UpperCAmelCase = [images]
if single_image:
assert len(snake_case__ ) == 1
for i in range(len(snake_case__ ) ):
if isinstance(images[i] , torch.Tensor ):
images.insert(snake_case__ , images.pop(snake_case__ ).to(self.device ).float() )
elif not isinstance(images[i] , torch.Tensor ):
images.insert(
snake_case__ , torch.as_tensor(img_tensorize(images.pop(snake_case__ ) , input_format=self.input_format ) )
.to(self.device )
.float() , )
# resize smallest edge
UpperCAmelCase = torch.tensor([im.shape[:2] for im in images] )
UpperCAmelCase = self.aug(snake_case__ )
# transpose images and convert to torch tensors
# images = [torch.as_tensor(i.astype("float32")).permute(2, 0, 1).to(self.device) for i in images]
# now normalize before pad to avoid useless arithmetic
UpperCAmelCase = [self.normalizer(snake_case__ ) for x in images]
# now pad them to do the following operations
UpperCAmelCase , UpperCAmelCase = self.pad(snake_case__ )
# Normalize
if self.size_divisibility > 0:
raise NotImplementedError()
# pad
UpperCAmelCase = torch.true_divide(snake_case__ , snake_case__ )
if single_image:
return images[0], sizes[0], scales_yx[0]
else:
return images, sizes, scales_yx
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase ):
'''simple docstring'''
boxes[:, 0::2] *= scale_yx[:, 1]
boxes[:, 1::2] *= scale_yx[:, 0]
return boxes
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase ):
'''simple docstring'''
assert torch.isfinite(lowerCAmelCase ).all(), "Box tensor contains infinite or NaN!"
UpperCAmelCase , UpperCAmelCase = box_size
tensor[:, 0].clamp_(min=0 , max=lowerCAmelCase )
tensor[:, 1].clamp_(min=0 , max=lowerCAmelCase )
tensor[:, 2].clamp_(min=0 , max=lowerCAmelCase )
tensor[:, 3].clamp_(min=0 , max=lowerCAmelCase )
| 673 | 0 |
'''simple docstring'''
from pickle import UnpicklingError
import jax
import jax.numpy as jnp
import numpy as np
from flax.serialization import from_bytes
from flax.traverse_util import flatten_dict
from ..utils import logging
__a: int = logging.get_logger(__name__)
def __UpperCamelCase ( UpperCAmelCase , UpperCAmelCase ):
try:
with open(UpperCAmelCase , '''rb''' ) as flax_state_f:
lowercase__ : List[str] = from_bytes(UpperCAmelCase , flax_state_f.read() )
except UnpicklingError as e:
try:
with open(UpperCAmelCase ) as f:
if f.read().startswith('''version''' ):
raise OSError(
'''You seem to have cloned a repository without having git-lfs installed. Please'''
''' install git-lfs and run `git lfs install` followed by `git lfs pull` in the'''
''' folder you cloned.''' )
else:
raise ValueError from e
except (UnicodeDecodeError, ValueError):
raise EnvironmentError(F"""Unable to convert {model_file} to Flax deserializable object. """ )
return load_flax_weights_in_pytorch_model(UpperCAmelCase , UpperCAmelCase )
def __UpperCamelCase ( UpperCAmelCase , UpperCAmelCase ):
try:
import torch # noqa: F401
except ImportError:
logger.error(
'''Loading Flax weights in PyTorch requires both PyTorch and Flax to be installed. Please see'''
''' https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation'''
''' instructions.''' )
raise
# check if we have bf16 weights
lowercase__ : str = flatten_dict(jax.tree_util.tree_map(lambda UpperCAmelCase : x.dtype == jnp.bfloataa , UpperCAmelCase ) ).values()
if any(UpperCAmelCase ):
# convert all weights to fp32 if they are bf16 since torch.from_numpy can-not handle bf16
# and bf16 is not fully supported in PT yet.
logger.warning(
'''Found ``bfloat16`` weights in Flax model. Casting all ``bfloat16`` weights to ``float32`` '''
'''before loading those in PyTorch model.''' )
lowercase__ : Union[str, Any] = jax.tree_util.tree_map(
lambda UpperCAmelCase : params.astype(np.floataa ) if params.dtype == jnp.bfloataa else params , UpperCAmelCase )
lowercase__ : Optional[Any] = ''''''
lowercase__ : int = flatten_dict(UpperCAmelCase , sep='''.''' )
lowercase__ : Tuple = pt_model.state_dict()
# keep track of unexpected & missing keys
lowercase__ : Optional[int] = []
lowercase__ : int = set(pt_model_dict.keys() )
for flax_key_tuple, flax_tensor in flax_state_dict.items():
lowercase__ : List[Any] = flax_key_tuple.split('''.''' )
if flax_key_tuple_array[-1] == "kernel" and flax_tensor.ndim == 4:
lowercase__ : Union[str, Any] = flax_key_tuple_array[:-1] + ['''weight''']
lowercase__ : Any = jnp.transpose(UpperCAmelCase , (3, 2, 0, 1) )
elif flax_key_tuple_array[-1] == "kernel":
lowercase__ : int = flax_key_tuple_array[:-1] + ['''weight''']
lowercase__ : Tuple = flax_tensor.T
elif flax_key_tuple_array[-1] == "scale":
lowercase__ : List[str] = flax_key_tuple_array[:-1] + ['''weight''']
if "time_embedding" not in flax_key_tuple_array:
for i, flax_key_tuple_string in enumerate(UpperCAmelCase ):
lowercase__ : Any = (
flax_key_tuple_string.replace('''_0''' , '''.0''' )
.replace('''_1''' , '''.1''' )
.replace('''_2''' , '''.2''' )
.replace('''_3''' , '''.3''' )
.replace('''_4''' , '''.4''' )
.replace('''_5''' , '''.5''' )
.replace('''_6''' , '''.6''' )
.replace('''_7''' , '''.7''' )
.replace('''_8''' , '''.8''' )
.replace('''_9''' , '''.9''' )
)
lowercase__ : Any = '''.'''.join(UpperCAmelCase )
if flax_key in pt_model_dict:
if flax_tensor.shape != pt_model_dict[flax_key].shape:
raise ValueError(
F"""Flax checkpoint seems to be incorrect. Weight {flax_key_tuple} was expected """
F"""to be of shape {pt_model_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
else:
# add weight to pytorch dict
lowercase__ : List[Any] = np.asarray(UpperCAmelCase ) if not isinstance(UpperCAmelCase , np.ndarray ) else flax_tensor
lowercase__ : Optional[int] = torch.from_numpy(UpperCAmelCase )
# remove from missing keys
missing_keys.remove(UpperCAmelCase )
else:
# weight is not expected by PyTorch model
unexpected_keys.append(UpperCAmelCase )
pt_model.load_state_dict(UpperCAmelCase )
# re-transform missing_keys to list
lowercase__ : Union[str, Any] = list(UpperCAmelCase )
if len(UpperCAmelCase ) > 0:
logger.warning(
'''Some weights of the Flax model were not used when initializing the PyTorch model'''
F""" {pt_model.__class__.__name__}: {unexpected_keys}\n- This IS expected if you are initializing"""
F""" {pt_model.__class__.__name__} from a Flax model trained on another task or with another architecture"""
''' (e.g. initializing a BertForSequenceClassification model from a FlaxBertForPreTraining model).\n- This'''
F""" IS NOT expected if you are initializing {pt_model.__class__.__name__} from a Flax model that you expect"""
''' to be exactly identical (e.g. initializing a BertForSequenceClassification model from a'''
''' FlaxBertForSequenceClassification model).''' )
if len(UpperCAmelCase ) > 0:
logger.warning(
F"""Some weights of {pt_model.__class__.__name__} were not initialized from the Flax model and are newly"""
F""" initialized: {missing_keys}\nYou should probably TRAIN this model on a down-stream task to be able to"""
''' use it for predictions and inference.''' )
return pt_model
| 152 |
"""simple docstring"""
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import DeiTConfig, DeiTForImageClassificationWithTeacher, DeiTImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
lowerCAmelCase_ : List[str] = logging.get_logger(__name__)
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase=False ):
'''simple docstring'''
UpperCAmelCase = []
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F'''blocks.{i}.norm1.weight''', F'''deit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((F'''blocks.{i}.norm1.bias''', F'''deit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((F'''blocks.{i}.attn.proj.weight''', F'''deit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.attn.proj.bias''', F'''deit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((F'''blocks.{i}.norm2.weight''', F'''deit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((F'''blocks.{i}.norm2.bias''', F'''deit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.weight''', F'''deit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.bias''', F'''deit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.weight''', F'''deit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.bias''', F'''deit.encoder.layer.{i}.output.dense.bias''') )
# projection layer + position embeddings
rename_keys.extend(
[
("""cls_token""", """deit.embeddings.cls_token"""),
("""dist_token""", """deit.embeddings.distillation_token"""),
("""patch_embed.proj.weight""", """deit.embeddings.patch_embeddings.projection.weight"""),
("""patch_embed.proj.bias""", """deit.embeddings.patch_embeddings.projection.bias"""),
("""pos_embed""", """deit.embeddings.position_embeddings"""),
] )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
("""norm.weight""", """layernorm.weight"""),
("""norm.bias""", """layernorm.bias"""),
("""pre_logits.fc.weight""", """pooler.dense.weight"""),
("""pre_logits.fc.bias""", """pooler.dense.bias"""),
] )
# if just the base model, we should remove "deit" from all keys that start with "deit"
UpperCAmelCase = [(pair[0], pair[1][4:]) if pair[1].startswith("""deit""" ) else pair for pair in rename_keys]
else:
# layernorm + classification heads
rename_keys.extend(
[
("""norm.weight""", """deit.layernorm.weight"""),
("""norm.bias""", """deit.layernorm.bias"""),
("""head.weight""", """cls_classifier.weight"""),
("""head.bias""", """cls_classifier.bias"""),
("""head_dist.weight""", """distillation_classifier.weight"""),
("""head_dist.bias""", """distillation_classifier.bias"""),
] )
return rename_keys
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase , lowerCAmelCase=False ):
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
UpperCAmelCase = """"""
else:
UpperCAmelCase = """deit."""
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
UpperCAmelCase = state_dict.pop(F'''blocks.{i}.attn.qkv.weight''' )
UpperCAmelCase = state_dict.pop(F'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
UpperCAmelCase = in_proj_weight[
: config.hidden_size, :
]
UpperCAmelCase = in_proj_bias[: config.hidden_size]
UpperCAmelCase = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
UpperCAmelCase = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
UpperCAmelCase = in_proj_weight[
-config.hidden_size :, :
]
UpperCAmelCase = in_proj_bias[-config.hidden_size :]
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase , lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = dct.pop(lowerCAmelCase )
UpperCAmelCase = val
def _lowerCAmelCase ( ):
'''simple docstring'''
UpperCAmelCase = """http://images.cocodataset.org/val2017/000000039769.jpg"""
UpperCAmelCase = Image.open(requests.get(lowerCAmelCase , stream=lowerCAmelCase ).raw )
return im
@torch.no_grad()
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = DeiTConfig()
# all deit models have fine-tuned heads
UpperCAmelCase = False
# dataset (fine-tuned on ImageNet 2012), patch_size and image_size
UpperCAmelCase = 1000
UpperCAmelCase = """huggingface/label-files"""
UpperCAmelCase = """imagenet-1k-id2label.json"""
UpperCAmelCase = json.load(open(hf_hub_download(lowerCAmelCase , lowerCAmelCase , repo_type="""dataset""" ) , """r""" ) )
UpperCAmelCase = {int(lowerCAmelCase ): v for k, v in idalabel.items()}
UpperCAmelCase = idalabel
UpperCAmelCase = {v: k for k, v in idalabel.items()}
UpperCAmelCase = int(deit_name[-6:-4] )
UpperCAmelCase = int(deit_name[-3:] )
# size of the architecture
if deit_name[9:].startswith("""tiny""" ):
UpperCAmelCase = 192
UpperCAmelCase = 768
UpperCAmelCase = 12
UpperCAmelCase = 3
elif deit_name[9:].startswith("""small""" ):
UpperCAmelCase = 384
UpperCAmelCase = 1536
UpperCAmelCase = 12
UpperCAmelCase = 6
if deit_name[9:].startswith("""base""" ):
pass
elif deit_name[4:].startswith("""large""" ):
UpperCAmelCase = 1024
UpperCAmelCase = 4096
UpperCAmelCase = 24
UpperCAmelCase = 16
# load original model from timm
UpperCAmelCase = timm.create_model(lowerCAmelCase , pretrained=lowerCAmelCase )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
UpperCAmelCase = timm_model.state_dict()
UpperCAmelCase = create_rename_keys(lowerCAmelCase , lowerCAmelCase )
for src, dest in rename_keys:
rename_key(lowerCAmelCase , lowerCAmelCase , lowerCAmelCase )
read_in_q_k_v(lowerCAmelCase , lowerCAmelCase , lowerCAmelCase )
# load HuggingFace model
UpperCAmelCase = DeiTForImageClassificationWithTeacher(lowerCAmelCase ).eval()
model.load_state_dict(lowerCAmelCase )
# Check outputs on an image, prepared by DeiTImageProcessor
UpperCAmelCase = int(
(256 / 224) * config.image_size ) # to maintain same ratio w.r.t. 224 images, see https://github.com/facebookresearch/deit/blob/ab5715372db8c6cad5740714b2216d55aeae052e/datasets.py#L103
UpperCAmelCase = DeiTImageProcessor(size=lowerCAmelCase , crop_size=config.image_size )
UpperCAmelCase = image_processor(images=prepare_img() , return_tensors="""pt""" )
UpperCAmelCase = encoding["""pixel_values"""]
UpperCAmelCase = model(lowerCAmelCase )
UpperCAmelCase = timm_model(lowerCAmelCase )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(lowerCAmelCase , outputs.logits , atol=1e-3 )
Path(lowerCAmelCase ).mkdir(exist_ok=lowerCAmelCase )
print(F'''Saving model {deit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(lowerCAmelCase )
print(F'''Saving image processor to {pytorch_dump_folder_path}''' )
image_processor.save_pretrained(lowerCAmelCase )
if __name__ == "__main__":
lowerCAmelCase_ : Dict = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--deit_name''',
default='''vit_deit_base_distilled_patch16_224''',
type=str,
help='''Name of the DeiT timm model you\'d like to convert.''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model directory.'''
)
lowerCAmelCase_ : str = parser.parse_args()
convert_deit_checkpoint(args.deit_name, args.pytorch_dump_folder_path)
| 673 | 0 |
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
if is_sentencepiece_available():
from ..ta.tokenization_ta import TaTokenizer
else:
from ...utils.dummy_sentencepiece_objects import TaTokenizer
__lowercase : Optional[int] = TaTokenizer
if is_tokenizers_available():
from ..ta.tokenization_ta_fast import TaTokenizerFast
else:
from ...utils.dummy_tokenizers_objects import TaTokenizerFast
__lowercase : int = TaTokenizerFast
__lowercase : Optional[int] = {'''configuration_mt5''': ['''MT5Config''', '''MT5OnnxConfig''']}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowercase : Tuple = [
'''MT5EncoderModel''',
'''MT5ForConditionalGeneration''',
'''MT5ForQuestionAnswering''',
'''MT5Model''',
'''MT5PreTrainedModel''',
'''MT5Stack''',
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowercase : Union[str, Any] = ['''TFMT5EncoderModel''', '''TFMT5ForConditionalGeneration''', '''TFMT5Model''']
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowercase : List[Any] = ['''FlaxMT5EncoderModel''', '''FlaxMT5ForConditionalGeneration''', '''FlaxMT5Model''']
if TYPE_CHECKING:
from .configuration_mta import MTaConfig, MTaOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mta import (
MTaEncoderModel,
MTaForConditionalGeneration,
MTaForQuestionAnswering,
MTaModel,
MTaPreTrainedModel,
MTaStack,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mta import TFMTaEncoderModel, TFMTaForConditionalGeneration, TFMTaModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_mta import FlaxMTaEncoderModel, FlaxMTaForConditionalGeneration, FlaxMTaModel
else:
import sys
__lowercase : Any = _LazyModule(
__name__,
globals()['__file__'],
_import_structure,
extra_objects={'MT5Tokenizer': MTaTokenizer, 'MT5TokenizerFast': MTaTokenizerFast},
module_spec=__spec__,
)
| 476 |
"""simple docstring"""
import unittest
from typing import Dict, List, Optional, Union
import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import BridgeTowerImageProcessor
class UpperCamelCase_ ( unittest.TestCase ):
def __init__( self , snake_case__ , snake_case__ = True , snake_case__ = None , snake_case__ = 32 , snake_case__ = True , snake_case__ = 1 / 2_55 , snake_case__ = True , snake_case__ = True , snake_case__ = [0.48_145_466, 0.4_578_275, 0.40_821_073] , snake_case__ = [0.26_862_954, 0.26_130_258, 0.27_577_711] , snake_case__ = True , snake_case__=7 , snake_case__=30 , snake_case__=4_00 , snake_case__=3 , ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = parent
UpperCAmelCase = do_resize
UpperCAmelCase = size if size is not None else {"""shortest_edge""": 2_88}
UpperCAmelCase = size_divisor
UpperCAmelCase = do_rescale
UpperCAmelCase = rescale_factor
UpperCAmelCase = do_normalize
UpperCAmelCase = do_center_crop
UpperCAmelCase = image_mean
UpperCAmelCase = image_std
UpperCAmelCase = do_pad
UpperCAmelCase = batch_size
UpperCAmelCase = num_channels
UpperCAmelCase = min_resolution
UpperCAmelCase = max_resolution
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
"do_normalize": self.do_normalize,
"do_resize": self.do_resize,
"size": self.size,
"size_divisor": self.size_divisor,
}
def UpperCamelCase_ ( self , snake_case__ , snake_case__=False ) -> int:
"""simple docstring"""
if not batched:
UpperCAmelCase = self.size["""shortest_edge"""]
UpperCAmelCase = image_inputs[0]
if isinstance(snake_case__ , Image.Image ):
UpperCAmelCase , UpperCAmelCase = image.size
else:
UpperCAmelCase , UpperCAmelCase = image.shape[1], image.shape[2]
UpperCAmelCase = size / min(snake_case__ , snake_case__ )
if h < w:
UpperCAmelCase , UpperCAmelCase = size, scale * w
else:
UpperCAmelCase , UpperCAmelCase = scale * h, size
UpperCAmelCase = int((13_33 / 8_00) * size )
if max(snake_case__ , snake_case__ ) > max_size:
UpperCAmelCase = max_size / max(snake_case__ , snake_case__ )
UpperCAmelCase = newh * scale
UpperCAmelCase = neww * scale
UpperCAmelCase , UpperCAmelCase = int(newh + 0.5 ), int(neww + 0.5 )
UpperCAmelCase , UpperCAmelCase = (
newh // self.size_divisor * self.size_divisor,
neww // self.size_divisor * self.size_divisor,
)
else:
UpperCAmelCase = []
for image in image_inputs:
UpperCAmelCase , UpperCAmelCase = self.get_expected_values([image] )
expected_values.append((expected_height, expected_width) )
UpperCAmelCase = max(snake_case__ , key=lambda snake_case__ : item[0] )[0]
UpperCAmelCase = max(snake_case__ , key=lambda snake_case__ : item[1] )[1]
return expected_height, expected_width
@require_torch
@require_vision
class UpperCamelCase_ ( a_ , unittest.TestCase ):
_A : List[Any] = BridgeTowerImageProcessor if is_vision_available() else None
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = BridgeTowerImageProcessingTester(self )
@property
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
return self.image_processor_tester.prepare_image_processor_dict()
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
UpperCAmelCase = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(snake_case__ , """image_mean""" ) )
self.assertTrue(hasattr(snake_case__ , """image_std""" ) )
self.assertTrue(hasattr(snake_case__ , """do_normalize""" ) )
self.assertTrue(hasattr(snake_case__ , """do_resize""" ) )
self.assertTrue(hasattr(snake_case__ , """size""" ) )
self.assertTrue(hasattr(snake_case__ , """size_divisor""" ) )
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
pass
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
UpperCAmelCase = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
UpperCAmelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , Image.Image )
# Test not batched input
UpperCAmelCase = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ , batched=snake_case__ )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = self.image_processing_class(**self.image_processor_dict )
# create random numpy tensors
UpperCAmelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ , numpify=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , np.ndarray )
# Test not batched input
UpperCAmelCase = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ , batched=snake_case__ )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = self.image_processing_class(**self.image_processor_dict )
# create random PyTorch tensors
UpperCAmelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ , torchify=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , torch.Tensor )
# Test not batched input
UpperCAmelCase = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ , batched=snake_case__ )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
| 673 | 0 |
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import DeiTImageProcessor, ViTConfig, ViTForImageClassification, ViTImageProcessor, ViTModel
from transformers.utils import logging
logging.set_verbosity_info()
A_ = logging.get_logger(__name__)
def UpperCAmelCase ( UpperCAmelCase ,UpperCAmelCase=False )-> Union[str, Any]:
'''simple docstring'''
SCREAMING_SNAKE_CASE_ = []
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((f'''blocks.{i}.norm1.weight''', f'''vit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((f'''blocks.{i}.norm1.bias''', f'''vit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((f'''blocks.{i}.attn.proj.weight''', f'''vit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((f'''blocks.{i}.attn.proj.bias''', f'''vit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((f'''blocks.{i}.norm2.weight''', f'''vit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((f'''blocks.{i}.norm2.bias''', f'''vit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((f'''blocks.{i}.mlp.fc1.weight''', f'''vit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((f'''blocks.{i}.mlp.fc1.bias''', f'''vit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((f'''blocks.{i}.mlp.fc2.weight''', f'''vit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((f'''blocks.{i}.mlp.fc2.bias''', f'''vit.encoder.layer.{i}.output.dense.bias''') )
# projection layer + position embeddings
rename_keys.extend(
[
('''cls_token''', '''vit.embeddings.cls_token'''),
('''patch_embed.proj.weight''', '''vit.embeddings.patch_embeddings.projection.weight'''),
('''patch_embed.proj.bias''', '''vit.embeddings.patch_embeddings.projection.bias'''),
('''pos_embed''', '''vit.embeddings.position_embeddings'''),
] )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
('''norm.weight''', '''layernorm.weight'''),
('''norm.bias''', '''layernorm.bias'''),
('''pre_logits.fc.weight''', '''pooler.dense.weight'''),
('''pre_logits.fc.bias''', '''pooler.dense.bias'''),
] )
# if just the base model, we should remove "vit" from all keys that start with "vit"
SCREAMING_SNAKE_CASE_ = [(pair[0], pair[1][4:]) if pair[1].startswith('''vit''' ) else pair for pair in rename_keys]
else:
# layernorm + classification head
rename_keys.extend(
[
('''norm.weight''', '''vit.layernorm.weight'''),
('''norm.bias''', '''vit.layernorm.bias'''),
('''head.weight''', '''classifier.weight'''),
('''head.bias''', '''classifier.bias'''),
] )
return rename_keys
def UpperCAmelCase ( UpperCAmelCase ,UpperCAmelCase ,UpperCAmelCase=False )-> str:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
SCREAMING_SNAKE_CASE_ = ''''''
else:
SCREAMING_SNAKE_CASE_ = '''vit.'''
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
SCREAMING_SNAKE_CASE_ = state_dict.pop(f'''blocks.{i}.attn.qkv.weight''' )
SCREAMING_SNAKE_CASE_ = state_dict.pop(f'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
SCREAMING_SNAKE_CASE_ = in_proj_weight[
: config.hidden_size, :
]
SCREAMING_SNAKE_CASE_ = in_proj_bias[: config.hidden_size]
SCREAMING_SNAKE_CASE_ = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
SCREAMING_SNAKE_CASE_ = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
SCREAMING_SNAKE_CASE_ = in_proj_weight[
-config.hidden_size :, :
]
SCREAMING_SNAKE_CASE_ = in_proj_bias[-config.hidden_size :]
def UpperCAmelCase ( UpperCAmelCase )-> Optional[int]:
'''simple docstring'''
SCREAMING_SNAKE_CASE_ = ['''head.weight''', '''head.bias''']
for k in ignore_keys:
state_dict.pop(UpperCAmelCase ,UpperCAmelCase )
def UpperCAmelCase ( UpperCAmelCase ,UpperCAmelCase ,UpperCAmelCase )-> Any:
'''simple docstring'''
SCREAMING_SNAKE_CASE_ = dct.pop(UpperCAmelCase )
SCREAMING_SNAKE_CASE_ = val
def UpperCAmelCase ( )-> Tuple:
'''simple docstring'''
SCREAMING_SNAKE_CASE_ = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
SCREAMING_SNAKE_CASE_ = Image.open(requests.get(UpperCAmelCase ,stream=UpperCAmelCase ).raw )
return im
@torch.no_grad()
def UpperCAmelCase ( UpperCAmelCase ,UpperCAmelCase )-> Any:
'''simple docstring'''
SCREAMING_SNAKE_CASE_ = ViTConfig()
SCREAMING_SNAKE_CASE_ = False
# dataset (ImageNet-21k only or also fine-tuned on ImageNet 2012), patch_size and image_size
if vit_name[-5:] == "in21k":
SCREAMING_SNAKE_CASE_ = True
SCREAMING_SNAKE_CASE_ = int(vit_name[-12:-10] )
SCREAMING_SNAKE_CASE_ = int(vit_name[-9:-6] )
else:
SCREAMING_SNAKE_CASE_ = 1000
SCREAMING_SNAKE_CASE_ = '''huggingface/label-files'''
SCREAMING_SNAKE_CASE_ = '''imagenet-1k-id2label.json'''
SCREAMING_SNAKE_CASE_ = json.load(open(hf_hub_download(UpperCAmelCase ,UpperCAmelCase ,repo_type='''dataset''' ) ,'''r''' ) )
SCREAMING_SNAKE_CASE_ = {int(UpperCAmelCase ): v for k, v in idalabel.items()}
SCREAMING_SNAKE_CASE_ = idalabel
SCREAMING_SNAKE_CASE_ = {v: k for k, v in idalabel.items()}
SCREAMING_SNAKE_CASE_ = int(vit_name[-6:-4] )
SCREAMING_SNAKE_CASE_ = int(vit_name[-3:] )
# size of the architecture
if "deit" in vit_name:
if vit_name[9:].startswith('''tiny''' ):
SCREAMING_SNAKE_CASE_ = 192
SCREAMING_SNAKE_CASE_ = 768
SCREAMING_SNAKE_CASE_ = 12
SCREAMING_SNAKE_CASE_ = 3
elif vit_name[9:].startswith('''small''' ):
SCREAMING_SNAKE_CASE_ = 384
SCREAMING_SNAKE_CASE_ = 1536
SCREAMING_SNAKE_CASE_ = 12
SCREAMING_SNAKE_CASE_ = 6
else:
pass
else:
if vit_name[4:].startswith('''small''' ):
SCREAMING_SNAKE_CASE_ = 768
SCREAMING_SNAKE_CASE_ = 2304
SCREAMING_SNAKE_CASE_ = 8
SCREAMING_SNAKE_CASE_ = 8
elif vit_name[4:].startswith('''base''' ):
pass
elif vit_name[4:].startswith('''large''' ):
SCREAMING_SNAKE_CASE_ = 1024
SCREAMING_SNAKE_CASE_ = 4096
SCREAMING_SNAKE_CASE_ = 24
SCREAMING_SNAKE_CASE_ = 16
elif vit_name[4:].startswith('''huge''' ):
SCREAMING_SNAKE_CASE_ = 1280
SCREAMING_SNAKE_CASE_ = 5120
SCREAMING_SNAKE_CASE_ = 32
SCREAMING_SNAKE_CASE_ = 16
# load original model from timm
SCREAMING_SNAKE_CASE_ = timm.create_model(UpperCAmelCase ,pretrained=UpperCAmelCase )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
SCREAMING_SNAKE_CASE_ = timm_model.state_dict()
if base_model:
remove_classification_head_(UpperCAmelCase )
SCREAMING_SNAKE_CASE_ = create_rename_keys(UpperCAmelCase ,UpperCAmelCase )
for src, dest in rename_keys:
rename_key(UpperCAmelCase ,UpperCAmelCase ,UpperCAmelCase )
read_in_q_k_v(UpperCAmelCase ,UpperCAmelCase ,UpperCAmelCase )
# load HuggingFace model
if vit_name[-5:] == "in21k":
SCREAMING_SNAKE_CASE_ = ViTModel(UpperCAmelCase ).eval()
else:
SCREAMING_SNAKE_CASE_ = ViTForImageClassification(UpperCAmelCase ).eval()
model.load_state_dict(UpperCAmelCase )
# Check outputs on an image, prepared by ViTImageProcessor/DeiTImageProcessor
if "deit" in vit_name:
SCREAMING_SNAKE_CASE_ = DeiTImageProcessor(size=config.image_size )
else:
SCREAMING_SNAKE_CASE_ = ViTImageProcessor(size=config.image_size )
SCREAMING_SNAKE_CASE_ = image_processor(images=prepare_img() ,return_tensors='''pt''' )
SCREAMING_SNAKE_CASE_ = encoding['''pixel_values''']
SCREAMING_SNAKE_CASE_ = model(UpperCAmelCase )
if base_model:
SCREAMING_SNAKE_CASE_ = timm_model.forward_features(UpperCAmelCase )
assert timm_pooled_output.shape == outputs.pooler_output.shape
assert torch.allclose(UpperCAmelCase ,outputs.pooler_output ,atol=1E-3 )
else:
SCREAMING_SNAKE_CASE_ = timm_model(UpperCAmelCase )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(UpperCAmelCase ,outputs.logits ,atol=1E-3 )
Path(UpperCAmelCase ).mkdir(exist_ok=UpperCAmelCase )
print(f'''Saving model {vit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(UpperCAmelCase )
print(f'''Saving image processor to {pytorch_dump_folder_path}''' )
image_processor.save_pretrained(UpperCAmelCase )
if __name__ == "__main__":
A_ = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--vit_name",
default="vit_base_patch16_224",
type=str,
help="Name of the ViT timm model you\'d like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
A_ = parser.parse_args()
convert_vit_checkpoint(args.vit_name, args.pytorch_dump_folder_path)
| 393 |
"""simple docstring"""
import pickle
import shutil
import tempfile
import unittest
from transformers import SPIECE_UNDERLINE, XLMRobertaTokenizer, XLMRobertaTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
lowerCAmelCase_ : Any = get_tests_dir('''fixtures/test_sentencepiece.model''')
@require_sentencepiece
@require_tokenizers
class UpperCamelCase_ ( a_ , unittest.TestCase ):
_A : List[str] = XLMRobertaTokenizer
_A : List[str] = XLMRobertaTokenizerFast
_A : Optional[Any] = True
_A : List[str] = True
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
super().setUp()
# We have a SentencePiece fixture for testing
UpperCAmelCase = XLMRobertaTokenizer(snake_case__ , keep_accents=snake_case__ )
tokenizer.save_pretrained(self.tmpdirname )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
UpperCAmelCase = """<pad>"""
UpperCAmelCase = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(snake_case__ ) , snake_case__ )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(snake_case__ ) , snake_case__ )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , """<s>""" )
self.assertEqual(vocab_keys[1] , """<pad>""" )
self.assertEqual(vocab_keys[-1] , """<mask>""" )
self.assertEqual(len(snake_case__ ) , 10_02 )
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
self.assertEqual(self.get_tokenizer().vocab_size , 10_02 )
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = XLMRobertaTokenizer(snake_case__ , keep_accents=snake_case__ )
UpperCAmelCase = tokenizer.tokenize("""This is a test""" )
self.assertListEqual(snake_case__ , ["""▁This""", """▁is""", """▁a""", """▁t""", """est"""] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(snake_case__ ) , [value + tokenizer.fairseq_offset for value in [2_85, 46, 10, 1_70, 3_82]] , )
UpperCAmelCase = tokenizer.tokenize("""I was born in 92000, and this is falsé.""" )
self.assertListEqual(
snake_case__ , [
SPIECE_UNDERLINE + """I""",
SPIECE_UNDERLINE + """was""",
SPIECE_UNDERLINE + """b""",
"""or""",
"""n""",
SPIECE_UNDERLINE + """in""",
SPIECE_UNDERLINE + """""",
"""9""",
"""2""",
"""0""",
"""0""",
"""0""",
""",""",
SPIECE_UNDERLINE + """and""",
SPIECE_UNDERLINE + """this""",
SPIECE_UNDERLINE + """is""",
SPIECE_UNDERLINE + """f""",
"""al""",
"""s""",
"""é""",
""".""",
] , )
UpperCAmelCase = tokenizer.convert_tokens_to_ids(snake_case__ )
self.assertListEqual(
snake_case__ , [
value + tokenizer.fairseq_offset
for value in [8, 21, 84, 55, 24, 19, 7, 2, 6_02, 3_47, 3_47, 3_47, 3, 12, 66, 46, 72, 80, 6, 2, 4]
# ^ unk: 2 + 1 = 3 unk: 2 + 1 = 3 ^
] , )
UpperCAmelCase = tokenizer.convert_ids_to_tokens(snake_case__ )
self.assertListEqual(
snake_case__ , [
SPIECE_UNDERLINE + """I""",
SPIECE_UNDERLINE + """was""",
SPIECE_UNDERLINE + """b""",
"""or""",
"""n""",
SPIECE_UNDERLINE + """in""",
SPIECE_UNDERLINE + """""",
"""<unk>""",
"""2""",
"""0""",
"""0""",
"""0""",
""",""",
SPIECE_UNDERLINE + """and""",
SPIECE_UNDERLINE + """this""",
SPIECE_UNDERLINE + """is""",
SPIECE_UNDERLINE + """f""",
"""al""",
"""s""",
"""<unk>""",
""".""",
] , )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
if not self.test_slow_tokenizer:
# as we don't have a slow version, we can't compare the outputs between slow and fast versions
return
UpperCAmelCase = (self.rust_tokenizer_class, """hf-internal-testing/tiny-xlm-roberta""", {})
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(f'''{tokenizer.__class__.__name__} ({pretrained_name})''' ):
UpperCAmelCase = self.rust_tokenizer_class.from_pretrained(snake_case__ , **snake_case__ )
UpperCAmelCase = self.tokenizer_class.from_pretrained(snake_case__ , **snake_case__ )
UpperCAmelCase = tempfile.mkdtemp()
UpperCAmelCase = tokenizer_r.save_pretrained(snake_case__ )
UpperCAmelCase = tokenizer_p.save_pretrained(snake_case__ )
# Checks it save with the same files + the tokenizer.json file for the fast one
self.assertTrue(any("""tokenizer.json""" in f for f in tokenizer_r_files ) )
UpperCAmelCase = tuple(f for f in tokenizer_r_files if """tokenizer.json""" not in f )
self.assertSequenceEqual(snake_case__ , snake_case__ )
# Checks everything loads correctly in the same way
UpperCAmelCase = tokenizer_r.from_pretrained(snake_case__ )
UpperCAmelCase = tokenizer_p.from_pretrained(snake_case__ )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case__ , snake_case__ ) )
# self.assertEqual(getattr(tokenizer_rp, key), getattr(tokenizer_pp, key))
# self.assertEqual(getattr(tokenizer_rp, key + "_id"), getattr(tokenizer_pp, key + "_id"))
shutil.rmtree(snake_case__ )
# Save tokenizer rust, legacy_format=True
UpperCAmelCase = tempfile.mkdtemp()
UpperCAmelCase = tokenizer_r.save_pretrained(snake_case__ , legacy_format=snake_case__ )
UpperCAmelCase = tokenizer_p.save_pretrained(snake_case__ )
# Checks it save with the same files
self.assertSequenceEqual(snake_case__ , snake_case__ )
# Checks everything loads correctly in the same way
UpperCAmelCase = tokenizer_r.from_pretrained(snake_case__ )
UpperCAmelCase = tokenizer_p.from_pretrained(snake_case__ )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case__ , snake_case__ ) )
shutil.rmtree(snake_case__ )
# Save tokenizer rust, legacy_format=False
UpperCAmelCase = tempfile.mkdtemp()
UpperCAmelCase = tokenizer_r.save_pretrained(snake_case__ , legacy_format=snake_case__ )
UpperCAmelCase = tokenizer_p.save_pretrained(snake_case__ )
# Checks it saved the tokenizer.json file
self.assertTrue(any("""tokenizer.json""" in f for f in tokenizer_r_files ) )
# Checks everything loads correctly in the same way
UpperCAmelCase = tokenizer_r.from_pretrained(snake_case__ )
UpperCAmelCase = tokenizer_p.from_pretrained(snake_case__ )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case__ , snake_case__ ) )
shutil.rmtree(snake_case__ )
@cached_property
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
return XLMRobertaTokenizer.from_pretrained("""xlm-roberta-base""" )
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
with tempfile.NamedTemporaryFile() as f:
shutil.copyfile(snake_case__ , f.name )
UpperCAmelCase = XLMRobertaTokenizer(f.name , keep_accents=snake_case__ )
UpperCAmelCase = pickle.dumps(snake_case__ )
pickle.loads(snake_case__ )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
if not self.test_rust_tokenizer:
return
UpperCAmelCase = self.get_tokenizer()
UpperCAmelCase = self.get_rust_tokenizer()
UpperCAmelCase = """I was born in 92000, and this is falsé."""
UpperCAmelCase = tokenizer.tokenize(snake_case__ )
UpperCAmelCase = rust_tokenizer.tokenize(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
UpperCAmelCase = tokenizer.encode(snake_case__ , add_special_tokens=snake_case__ )
UpperCAmelCase = rust_tokenizer.encode(snake_case__ , add_special_tokens=snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
UpperCAmelCase = self.get_rust_tokenizer()
UpperCAmelCase = tokenizer.encode(snake_case__ )
UpperCAmelCase = rust_tokenizer.encode(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
@slow
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = """Hello World!"""
UpperCAmelCase = [0, 3_53_78, 66_61, 38, 2]
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base') # xlmr.large has same tokenizer
# xlmr.eval()
# xlmr.encode(symbols)
self.assertListEqual(snake_case__ , self.big_tokenizer.encode(snake_case__ ) )
@slow
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = (
"""This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) \" [ ] ! : - . Also we will"""
""" add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth"""
)
UpperCAmelCase = [
0,
32_93,
83,
10,
45_52,
49_89,
79_86,
6_78,
10,
59_15,
1_11,
17_94_59,
12_48_50,
4,
60_44,
2_37,
12,
6,
5,
6,
4,
67_80,
7_05,
15,
13_88,
44,
3_78,
1_01_14,
7_11,
1_52,
20,
6,
5,
2_23_76,
6_42,
12_21,
1_51_90,
3_41_53,
4_50,
56_08,
9_59,
11_19,
5_77_02,
1_36,
1_86,
47,
10_98,
2_93_67,
47,
# 4426, # What fairseq tokenizes from "<unk>": "_<"
# 3678, # What fairseq tokenizes from "<unk>": "unk"
# 2740, # What fairseq tokenizes from "<unk>": ">"
3, # What we tokenize from "<unk>": "<unk>"
6, # Residue from the tokenization: an extra sentencepiece underline
4,
60_44,
2_37,
62_84,
5_09_01,
5_28,
31,
90,
34,
9_27,
2,
]
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base') # xlmr.large has same tokenizer
# xlmr.eval()
# xlmr.encode(symbols)
self.assertListEqual(snake_case__ , self.big_tokenizer.encode(snake_case__ ) )
@slow
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
UpperCAmelCase = {"""input_ids""": [[0, 1_10_62, 8_27_72, 7, 15, 8_27_72, 5_38, 5_15_29, 2_37, 1_71_98, 12_90, 2_06, 9, 21_51_75, 13_14, 1_36, 1_71_98, 12_90, 2_06, 9, 5_63_59, 42, 12_20_09, 9, 1_64_66, 16, 8_73_44, 45_37, 9, 47_17, 7_83_81, 6, 15_99_58, 7, 15, 2_44_80, 6_18, 4, 5_27, 2_26_93, 54_28, 4, 27_77, 2_44_80, 98_74, 4, 4_35_23, 5_94, 4, 8_03, 1_83_92, 3_31_89, 18, 4, 4_35_23, 2_44_47, 1_23_99, 1_00, 2_49_55, 8_36_58, 96_26, 14_40_57, 15, 8_39, 2_23_35, 16, 1_36, 2_49_55, 8_36_58, 8_34_79, 15, 3_91_02, 7_24, 16, 6_78, 6_45, 27_89, 13_28, 45_89, 42, 12_20_09, 11_57_74, 23, 8_05, 13_28, 4_68_76, 7, 1_36, 5_38_94, 19_40, 4_22_27, 4_11_59, 1_77_21, 8_23, 4_25, 4, 2_75_12, 9_87_22, 2_06, 1_36, 55_31, 49_70, 9_19, 1_73_36, 5, 2], [0, 2_00_80, 6_18, 83, 8_27_75, 47, 4_79, 9, 15_17, 73, 5_38_94, 3_33, 8_05_81, 11_01_17, 1_88_11, 52_56, 12_95, 51, 15_25_26, 2_97, 79_86, 3_90, 12_44_16, 5_38, 3_54_31, 2_14, 98, 1_50_44, 2_57_37, 1_36, 71_08, 4_37_01, 23, 7_56, 13_53_55, 7, 5, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [0, 5_81, 6_37_73, 11_94_55, 6, 14_77_97, 8_82_03, 7, 6_45, 70, 21, 32_85, 1_02_69, 5, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]], """attention_mask""": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=snake_case__ , model_name="""xlm-roberta-base""" , revision="""d9d8a8ea5eb94b1c6654ae9249df7793cd2933d3""" , )
| 673 | 0 |
__UpperCAmelCase = '''Tobias Carryer'''
from time import time
class lowerCamelCase :
'''simple docstring'''
def __init__( self , _UpperCamelCase , _UpperCamelCase , _UpperCamelCase , _UpperCamelCase=int(time() ) ) -> List[Any]: # noqa: B008
UpperCAmelCase_ : Optional[Any] = multiplier
UpperCAmelCase_ : Tuple = increment
UpperCAmelCase_ : int = modulo
UpperCAmelCase_ : Dict = seed
def __UpperCAmelCase ( self ) -> List[Any]:
UpperCAmelCase_ : Any = (self.multiplier * self.seed + self.increment) % self.modulo
return self.seed
if __name__ == "__main__":
# Show the LCG in action.
__UpperCAmelCase = LinearCongruentialGenerator(1664525, 1013904223, 2 << 31)
while True:
print(lcg.next_number())
| 406 |
"""simple docstring"""
import socket
def _lowerCAmelCase ( ):
'''simple docstring'''
UpperCAmelCase = socket.socket(socket.AF_INET , socket.SOCK_STREAM )
UpperCAmelCase = socket.gethostname()
UpperCAmelCase = 12312
sock.connect((host, port) )
sock.send(b"""Hello server!""" )
with open("""Received_file""" , """wb""" ) as out_file:
print("""File opened""" )
print("""Receiving data...""" )
while True:
UpperCAmelCase = sock.recv(1024 )
if not data:
break
out_file.write(lowerCAmelCase )
print("""Successfully received the file""" )
sock.close()
print("""Connection closed""" )
if __name__ == "__main__":
main()
| 673 | 0 |
"""simple docstring"""
def __a ( A , A ) -> Dict:
'''simple docstring'''
A__ = 0
A__ = len(A ) - 1
while left <= right:
# avoid divided by 0 during interpolation
if sorted_collection[left] == sorted_collection[right]:
if sorted_collection[left] == item:
return left
else:
return None
A__ = left + ((item - sorted_collection[left]) * (right - left)) // (
sorted_collection[right] - sorted_collection[left]
)
# out of range check
if point < 0 or point >= len(A ):
return None
A__ = sorted_collection[point]
if current_item == item:
return point
else:
if point < left:
A__ = left
A__ = point
elif point > right:
A__ = right
A__ = point
else:
if item < current_item:
A__ = point - 1
else:
A__ = point + 1
return None
def __a ( A , A , A , A ) -> Optional[int]:
'''simple docstring'''
if sorted_collection[left] == sorted_collection[right]:
if sorted_collection[left] == item:
return left
else:
return None
A__ = left + ((item - sorted_collection[left]) * (right - left)) // (
sorted_collection[right] - sorted_collection[left]
)
# out of range check
if point < 0 or point >= len(A ):
return None
if sorted_collection[point] == item:
return point
elif point < left:
return interpolation_search_by_recursion(A , A , A , A )
elif point > right:
return interpolation_search_by_recursion(A , A , A , A )
else:
if sorted_collection[point] > item:
return interpolation_search_by_recursion(
A , A , A , point - 1 )
else:
return interpolation_search_by_recursion(
A , A , point + 1 , A )
def __a ( A ) -> Tuple:
'''simple docstring'''
if collection != sorted(A ):
raise ValueError("Collection must be ascending sorted" )
return True
if __name__ == "__main__":
import sys
__UpperCAmelCase =0
if debug == 1:
__UpperCAmelCase =[10, 30, 40, 45, 50, 66, 77, 93]
try:
__assert_sorted(collection)
except ValueError:
sys.exit("""Sequence must be ascending sorted to apply interpolation search""")
__UpperCAmelCase =67
__UpperCAmelCase =interpolation_search(collection, target)
if result is not None:
print(F'''{target} found at positions: {result}''')
else:
print("""Not found""") | 337 |
"""simple docstring"""
import math
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
return math.sqrt(lowerCAmelCase ) * math.sqrt(lowerCAmelCase ) == num
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = 0
UpperCAmelCase = n
while left <= right:
UpperCAmelCase = (left + right) // 2
if mid**2 == n:
return True
elif mid**2 > n:
UpperCAmelCase = mid - 1
else:
UpperCAmelCase = mid + 1
return False
if __name__ == "__main__":
import doctest
doctest.testmod()
| 673 | 0 |
'''simple docstring'''
from typing import List, Optional, Tuple
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import logging
from .tokenization_herbert import HerbertTokenizer
__UpperCAmelCase =logging.get_logger(__name__)
__UpperCAmelCase ={'''vocab_file''': '''vocab.json''', '''merges_file''': '''merges.txt''', '''tokenizer_file''': '''tokenizer.json'''}
__UpperCAmelCase ={
'''vocab_file''': {
'''allegro/herbert-base-cased''': '''https://huggingface.co/allegro/herbert-base-cased/resolve/main/vocab.json'''
},
'''merges_file''': {
'''allegro/herbert-base-cased''': '''https://huggingface.co/allegro/herbert-base-cased/resolve/main/merges.txt'''
},
}
__UpperCAmelCase ={'''allegro/herbert-base-cased''': 5_1_4}
__UpperCAmelCase ={}
class a__ ( a_ ):
lowerCamelCase : Dict =VOCAB_FILES_NAMES
lowerCamelCase : Any =PRETRAINED_VOCAB_FILES_MAP
lowerCamelCase : List[str] =PRETRAINED_INIT_CONFIGURATION
lowerCamelCase : Dict =PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
lowerCamelCase : Tuple =HerbertTokenizer
def __init__( self : Tuple , a : Any=None , a : int=None , a : Union[str, Any]=None , a : Union[str, Any]="<s>" , a : Optional[int]="<unk>" , a : List[Any]="<pad>" , a : Optional[Any]="<mask>" , a : str="</s>" , **a : Optional[Any] , ):
"""simple docstring"""
super().__init__(
snake_case__ , snake_case__ , tokenizer_file=snake_case__ , cls_token=snake_case__ , unk_token=snake_case__ , pad_token=snake_case__ , mask_token=snake_case__ , sep_token=snake_case__ , **snake_case__ , )
def SCREAMING_SNAKE_CASE__ ( self : Dict , a : int , a : Union[str, Any] = None ):
"""simple docstring"""
__lowerCamelCase = [self.cls_token_id]
__lowerCamelCase = [self.sep_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE__ ( self : Tuple , a : str , a : List[Any] = None , a : Any = False ):
"""simple docstring"""
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=snake_case__ , token_ids_a=snake_case__ , already_has_special_tokens=snake_case__ )
if token_ids_a is None:
return [1] + ([0] * len(snake_case__ )) + [1]
return [1] + ([0] * len(snake_case__ )) + [1] + ([0] * len(snake_case__ )) + [1]
def SCREAMING_SNAKE_CASE__ ( self : Union[str, Any] , a : Tuple , a : str = None ):
"""simple docstring"""
__lowerCamelCase = [self.sep_token_id]
__lowerCamelCase = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1]
def SCREAMING_SNAKE_CASE__ ( self : Tuple , a : Any , a : Tuple = None ):
"""simple docstring"""
__lowerCamelCase = self._tokenizer.model.save(snake_case__ , name=snake_case__ )
return tuple(snake_case__ )
| 546 |
"""simple docstring"""
import functools
import gc
import inspect
import torch
from .imports import is_npu_available, is_xpu_available
def _lowerCAmelCase ( *lowerCAmelCase ):
'''simple docstring'''
if not isinstance(lowerCAmelCase , lowerCAmelCase ):
UpperCAmelCase = list(lowerCAmelCase )
for i in range(len(lowerCAmelCase ) ):
UpperCAmelCase = None
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
return objects
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = [
"""CUDA out of memory.""", # CUDA OOM
"""cuDNN error: CUDNN_STATUS_NOT_SUPPORTED.""", # CUDNN SNAFU
"""DefaultCPUAllocator: can't allocate memory""", # CPU OOM
]
if isinstance(lowerCAmelCase , lowerCAmelCase ) and len(exception.args ) == 1:
return any(err in exception.args[0] for err in _statements )
return False
def _lowerCAmelCase ( lowerCAmelCase = None , lowerCAmelCase = 128 ):
'''simple docstring'''
if function is None:
return functools.partial(lowerCAmelCase , starting_batch_size=lowerCAmelCase )
UpperCAmelCase = starting_batch_size
def decorator(*lowerCAmelCase , **lowerCAmelCase ):
nonlocal batch_size
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
UpperCAmelCase = list(inspect.signature(lowerCAmelCase ).parameters.keys() )
# Guard against user error
if len(lowerCAmelCase ) < (len(lowerCAmelCase ) + 1):
UpperCAmelCase = """, """.join([F'''{arg}={value}''' for arg, value in zip(params[1:] , args[1:] )] )
raise TypeError(
F'''Batch size was passed into `{function.__name__}` as the first argument when called.'''
F'''Remove this as the decorator already does so: `{function.__name__}({arg_str})`''' )
while True:
if batch_size == 0:
raise RuntimeError("""No executable batch size found, reached zero.""" )
try:
return function(lowerCAmelCase , *lowerCAmelCase , **lowerCAmelCase )
except Exception as e:
if should_reduce_batch_size(lowerCAmelCase ):
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
batch_size //= 2
else:
raise
return decorator
| 673 | 0 |
from __future__ import annotations
import os
from typing import Any
import requests
lowerCamelCase = '''https://api.github.com'''
# https://docs.github.com/en/free-pro-team@latest/rest/reference/users#get-the-authenticated-user
lowerCamelCase = BASE_URL + '''/user'''
# https://github.com/settings/tokens
lowerCamelCase = os.environ.get('USER_TOKEN', '')
def a_ ( SCREAMING_SNAKE_CASE__ : str ):
'''simple docstring'''
_lowerCamelCase : List[Any] ={
'Authorization': F'''token {auth_token}''',
'Accept': 'application/vnd.github.v3+json',
}
return requests.get(SCREAMING_SNAKE_CASE__ , headers=SCREAMING_SNAKE_CASE__ ).json()
if __name__ == "__main__": # pragma: no cover
if USER_TOKEN:
for key, value in fetch_github_info(USER_TOKEN).items():
print(F"""{key}: {value}""")
else:
raise ValueError('\'USER_TOKEN\' field cannot be empty.')
| 464 |
"""simple docstring"""
import math
def _lowerCAmelCase ( lowerCAmelCase = 100 ):
'''simple docstring'''
UpperCAmelCase = sum(i * i for i in range(1 , n + 1 ) )
UpperCAmelCase = int(math.pow(sum(range(1 , n + 1 ) ) , 2 ) )
return square_of_sum - sum_of_squares
if __name__ == "__main__":
print(F'{solution() = }')
| 673 | 0 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
UpperCAmelCase : str = {
'''configuration_poolformer''': [
'''POOLFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP''',
'''PoolFormerConfig''',
'''PoolFormerOnnxConfig''',
]
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCAmelCase : Any = ['''PoolFormerFeatureExtractor''']
UpperCAmelCase : Dict = ['''PoolFormerImageProcessor''']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCAmelCase : Optional[int] = [
'''POOLFORMER_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''PoolFormerForImageClassification''',
'''PoolFormerModel''',
'''PoolFormerPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_poolformer import (
POOLFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP,
PoolFormerConfig,
PoolFormerOnnxConfig,
)
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .feature_extraction_poolformer import PoolFormerFeatureExtractor
from .image_processing_poolformer import PoolFormerImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_poolformer import (
POOLFORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
PoolFormerForImageClassification,
PoolFormerModel,
PoolFormerPreTrainedModel,
)
else:
import sys
UpperCAmelCase : Tuple = _LazyModule(__name__, globals()["__file__"], _import_structure)
| 567 |
"""simple docstring"""
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = [0] * len(lowerCAmelCase )
UpperCAmelCase = []
UpperCAmelCase = [1] * len(lowerCAmelCase )
for values in graph.values():
for i in values:
indegree[i] += 1
for i in range(len(lowerCAmelCase ) ):
if indegree[i] == 0:
queue.append(lowerCAmelCase )
while queue:
UpperCAmelCase = queue.pop(0 )
for x in graph[vertex]:
indegree[x] -= 1
if long_dist[vertex] + 1 > long_dist[x]:
UpperCAmelCase = long_dist[vertex] + 1
if indegree[x] == 0:
queue.append(lowerCAmelCase )
print(max(lowerCAmelCase ) )
# Adjacency list of Graph
lowerCAmelCase_ : str = {0: [2, 3, 4], 1: [2, 7], 2: [5], 3: [5, 7], 4: [7], 5: [6], 6: [7], 7: []}
longest_distance(graph)
| 673 | 0 |
'''simple docstring'''
import math
def lowerCAmelCase_ ( _lowerCamelCase: Dict ):
return math.sqrt(_lowerCamelCase ) * math.sqrt(_lowerCamelCase ) == num
def lowerCAmelCase_ ( _lowerCamelCase: Optional[Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = 0
__SCREAMING_SNAKE_CASE : Optional[Any] = n
while left <= right:
__SCREAMING_SNAKE_CASE : str = (left + right) // 2
if mid**2 == n:
return True
elif mid**2 > n:
__SCREAMING_SNAKE_CASE : Tuple = mid - 1
else:
__SCREAMING_SNAKE_CASE : Optional[int] = mid + 1
return False
if __name__ == "__main__":
import doctest
doctest.testmod() | 578 |
"""simple docstring"""
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import torch
from ..models.auto import AutoModelForSequenceClassification, AutoTokenizer
from .base import PipelineTool
class UpperCamelCase_ ( a_ ):
_A : Optional[int] = 'facebook/bart-large-mnli'
_A : Union[str, Any] = (
'This is a tool that classifies an English text using provided labels. It takes two inputs: `text`, which '
'should be the text to classify, and `labels`, which should be the list of labels to use for classification. '
'It returns the most likely label in the list of provided `labels` for the input text.'
)
_A : Dict = 'text_classifier'
_A : Union[str, Any] = AutoTokenizer
_A : Tuple = AutoModelForSequenceClassification
_A : Optional[int] = ['text', ['text']]
_A : Dict = ['text']
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
super().setup()
UpperCAmelCase = self.model.config
UpperCAmelCase = -1
for idx, label in config.idalabel.items():
if label.lower().startswith("""entail""" ):
UpperCAmelCase = int(snake_case__ )
if self.entailment_id == -1:
raise ValueError("""Could not determine the entailment ID from the model config, please pass it at init.""" )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = labels
return self.pre_processor(
[text] * len(snake_case__ ) , [f'''This example is {label}''' for label in labels] , return_tensors="""pt""" , padding="""max_length""" , )
def UpperCamelCase_ ( self , snake_case__ ) -> str:
"""simple docstring"""
UpperCAmelCase = outputs.logits
UpperCAmelCase = torch.argmax(logits[:, 2] ).item()
return self._labels[label_id]
| 673 | 0 |
"""simple docstring"""
from ...configuration_utils import PretrainedConfig
from ...utils import logging
lowercase__ : Optional[Any] = logging.get_logger(__name__)
lowercase__ : Dict = {
'''naver-clova-ix/donut-base''': '''https://huggingface.co/naver-clova-ix/donut-base/resolve/main/config.json''',
# See all Donut models at https://huggingface.co/models?filter=donut-swin
}
class _UpperCAmelCase ( a_):
_lowerCAmelCase : Union[str, Any] = 'donut-swin'
_lowerCAmelCase : str = {
'num_attention_heads': 'num_heads',
'num_hidden_layers': 'num_layers',
}
def __init__( self : int , lowercase_ : List[Any]=224 , lowercase_ : Union[str, Any]=4 , lowercase_ : Union[str, Any]=3 , lowercase_ : Optional[Any]=96 , lowercase_ : Any=[2, 2, 6, 2] , lowercase_ : Any=[3, 6, 12, 24] , lowercase_ : List[Any]=7 , lowercase_ : Optional[Any]=4.0 , lowercase_ : str=True , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.0 , lowercase_ : List[Any]=0.1 , lowercase_ : Union[str, Any]="gelu" , lowercase_ : Tuple=False , lowercase_ : Any=0.02 , lowercase_ : int=1E-5 , **lowercase_ : Union[str, Any] , ):
super().__init__(**snake_case__ )
snake_case_ : List[Any] = image_size
snake_case_ : int = patch_size
snake_case_ : Optional[Any] = num_channels
snake_case_ : Dict = embed_dim
snake_case_ : List[Any] = depths
snake_case_ : List[str] = len(snake_case__ )
snake_case_ : int = num_heads
snake_case_ : str = window_size
snake_case_ : Tuple = mlp_ratio
snake_case_ : Tuple = qkv_bias
snake_case_ : Optional[Any] = hidden_dropout_prob
snake_case_ : Optional[int] = attention_probs_dropout_prob
snake_case_ : Optional[Any] = drop_path_rate
snake_case_ : int = hidden_act
snake_case_ : str = use_absolute_embeddings
snake_case_ : Any = layer_norm_eps
snake_case_ : Tuple = initializer_range
# we set the hidden_size attribute in order to make Swin work with VisionEncoderDecoderModel
# this indicates the channel dimension after the last stage of the model
snake_case_ : int = int(embed_dim * 2 ** (len(snake_case__ ) - 1) )
| 123 |
"""simple docstring"""
from dataclasses import dataclass
from enum import Enum
from typing import List, Optional, Union
import numpy as np
import PIL
from PIL import Image
from ...utils import BaseOutput, is_torch_available, is_transformers_available
@dataclass
class UpperCamelCase_ ( a_ ):
_A : Union[List[PIL.Image.Image], np.ndarray]
_A : Optional[List[bool]]
if is_transformers_available() and is_torch_available():
from .pipeline_semantic_stable_diffusion import SemanticStableDiffusionPipeline
| 673 | 0 |
'''simple docstring'''
import math
from typing import Union
import torch
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import randn_tensor
from .scheduling_utils import SchedulerMixin
class __lowerCamelCase ( a_ , a_ ):
'''simple docstring'''
snake_case__ : int = 1
@register_to_config
def __init__( self , a__=2000 , a__=0.1 , a__=20 , a__=1e-3 ):
__SCREAMING_SNAKE_CASE : List[Any] = None
__SCREAMING_SNAKE_CASE : Any = None
__SCREAMING_SNAKE_CASE : Optional[Any] = None
def a_ ( self , a__ , a__ = None ):
__SCREAMING_SNAKE_CASE : Optional[Any] = torch.linspace(1 , self.config.sampling_eps , snake_case__ , device=snake_case__ )
def a_ ( self , a__ , a__ , a__ , a__=None ):
if self.timesteps is None:
raise ValueError(
"`self.timesteps` is not set, you need to run 'set_timesteps' after creating the scheduler" )
# TODO(Patrick) better comments + non-PyTorch
# postprocess model score
__SCREAMING_SNAKE_CASE : List[str] = (
-0.25 * t**2 * (self.config.beta_max - self.config.beta_min) - 0.5 * t * self.config.beta_min
)
__SCREAMING_SNAKE_CASE : Optional[Any] = torch.sqrt(1.0 - torch.exp(2.0 * log_mean_coeff ) )
__SCREAMING_SNAKE_CASE : List[str] = std.flatten()
while len(std.shape ) < len(score.shape ):
__SCREAMING_SNAKE_CASE : str = std.unsqueeze(-1 )
__SCREAMING_SNAKE_CASE : Optional[Any] = -score / std
# compute
__SCREAMING_SNAKE_CASE : Union[str, Any] = -1.0 / len(self.timesteps )
__SCREAMING_SNAKE_CASE : List[str] = self.config.beta_min + t * (self.config.beta_max - self.config.beta_min)
__SCREAMING_SNAKE_CASE : List[Any] = beta_t.flatten()
while len(beta_t.shape ) < len(x.shape ):
__SCREAMING_SNAKE_CASE : int = beta_t.unsqueeze(-1 )
__SCREAMING_SNAKE_CASE : List[Any] = -0.5 * beta_t * x
__SCREAMING_SNAKE_CASE : Any = torch.sqrt(snake_case__ )
__SCREAMING_SNAKE_CASE : List[Any] = drift - diffusion**2 * score
__SCREAMING_SNAKE_CASE : Tuple = x + drift * dt
# add noise
__SCREAMING_SNAKE_CASE : Any = randn_tensor(x.shape , layout=x.layout , generator=snake_case__ , device=x.device , dtype=x.dtype )
__SCREAMING_SNAKE_CASE : Any = x_mean + diffusion * math.sqrt(-dt ) * noise
return x, x_mean
def __len__( self ):
return self.config.num_train_timesteps
| 211 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_torch_available,
)
lowerCAmelCase_ : Any = {
'''configuration_encodec''': [
'''ENCODEC_PRETRAINED_CONFIG_ARCHIVE_MAP''',
'''EncodecConfig''',
],
'''feature_extraction_encodec''': ['''EncodecFeatureExtractor'''],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
lowerCAmelCase_ : List[str] = [
'''ENCODEC_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''EncodecModel''',
'''EncodecPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_encodec import (
ENCODEC_PRETRAINED_CONFIG_ARCHIVE_MAP,
EncodecConfig,
)
from .feature_extraction_encodec import EncodecFeatureExtractor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_encodec import (
ENCODEC_PRETRAINED_MODEL_ARCHIVE_LIST,
EncodecModel,
EncodecPreTrainedModel,
)
else:
import sys
lowerCAmelCase_ : str = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 673 | 0 |
'''simple docstring'''
from typing import Any, Dict, List, Optional, Tuple, Union
import torch
from torch import nn
from torch.utils.data import DistributedSampler, RandomSampler
from transformers import PreTrainedModel, Trainer, logging
from transformers.integrations import is_fairscale_available
from transformers.models.fsmt.configuration_fsmt import FSMTConfig
from transformers.optimization import (
Adafactor,
AdamW,
get_constant_schedule,
get_constant_schedule_with_warmup,
get_cosine_schedule_with_warmup,
get_cosine_with_hard_restarts_schedule_with_warmup,
get_linear_schedule_with_warmup,
get_polynomial_decay_schedule_with_warmup,
)
from transformers.trainer_pt_utils import get_tpu_sampler
from transformers.training_args import ParallelMode
from transformers.utils import is_torch_tpu_available
if is_fairscale_available():
from fairscale.optim import OSS
__a: Union[str, Any] = logging.get_logger(__name__)
__a: int = {
'''linear''': get_linear_schedule_with_warmup,
'''cosine''': get_cosine_schedule_with_warmup,
'''cosine_w_restarts''': get_cosine_with_hard_restarts_schedule_with_warmup,
'''polynomial''': get_polynomial_decay_schedule_with_warmup,
'''constant''': get_constant_schedule,
'''constant_w_warmup''': get_constant_schedule_with_warmup,
}
class UpperCAmelCase ( a_ ):
'''simple docstring'''
def __init__( self , __lowerCAmelCase=None , __lowerCAmelCase=None , *__lowerCAmelCase , **__lowerCAmelCase ) -> Tuple:
super().__init__(*snake_case__ , **snake_case__ )
if config is None:
assert isinstance(self.model , snake_case__ ), (
"If no `config` is passed the model to be trained has to be of type `PreTrainedModel`, but is"
F""" {self.model.__class__}"""
)
lowercase__ : List[Any] = self.model.config
else:
lowercase__ : Optional[int] = config
lowercase__ : Dict = data_args
lowercase__ : List[str] = self.config.tgt_vocab_size if isinstance(self.config , snake_case__ ) else self.config.vocab_size
if self.args.label_smoothing != 0 or (self.data_args is not None and self.data_args.ignore_pad_token_for_loss):
assert self.config.pad_token_id is not None, (
"Make sure that `config.pad_token_id` is correcly defined when ignoring `pad_token` for loss"
" calculation or doing label smoothing."
)
if self.config.pad_token_id is None and self.config.eos_token_id is not None:
logger.warning(
F"""The `config.pad_token_id` is `None`. Using `config.eos_token_id` = {self.config.eos_token_id} for"""
''' padding..''' )
if self.args.label_smoothing == 0:
lowercase__ : Any = torch.nn.CrossEntropyLoss(ignore_index=self.config.pad_token_id )
else:
# dynamically import label_smoothed_nll_loss
from utils import label_smoothed_nll_loss
lowercase__ : Tuple = label_smoothed_nll_loss
def _lowerCAmelCase( self , __lowerCAmelCase ) -> Union[str, Any]:
if self.optimizer is None:
lowercase__ : int = ['''bias''', '''LayerNorm.weight''']
lowercase__ : Optional[Any] = [
{
'''params''': [p for n, p in self.model.named_parameters() if not any(nd in n for nd in no_decay )],
'''weight_decay''': self.args.weight_decay,
},
{
'''params''': [p for n, p in self.model.named_parameters() if any(nd in n for nd in no_decay )],
'''weight_decay''': 0.0,
},
]
lowercase__ : Any = Adafactor if self.args.adafactor else AdamW
if self.args.adafactor:
lowercase__ : List[str] = Adafactor
lowercase__ : int = {'''scale_parameter''': False, '''relative_step''': False}
else:
lowercase__ : int = AdamW
lowercase__ : List[Any] = {
'''betas''': (self.args.adam_betaa, self.args.adam_betaa),
'''eps''': self.args.adam_epsilon,
}
lowercase__ : Union[str, Any] = self.args.learning_rate
if self.sharded_ddp:
lowercase__ : List[Any] = OSS(
params=snake_case__ , optim=snake_case__ , **snake_case__ , )
else:
lowercase__ : Tuple = optimizer_cls(snake_case__ , **snake_case__ )
if self.lr_scheduler is None:
lowercase__ : Optional[Any] = self._get_lr_scheduler(snake_case__ )
else: # ignoring --lr_scheduler
logger.warning('''scheduler is passed to `Seq2SeqTrainer`, `--lr_scheduler` arg is ignored.''' )
def _lowerCAmelCase( self , __lowerCAmelCase ) -> str:
lowercase__ : Optional[Any] = arg_to_scheduler[self.args.lr_scheduler]
if self.args.lr_scheduler == "constant":
lowercase__ : Optional[Any] = schedule_func(self.optimizer )
elif self.args.lr_scheduler == "constant_w_warmup":
lowercase__ : str = schedule_func(self.optimizer , num_warmup_steps=self.args.warmup_steps )
else:
lowercase__ : Optional[int] = schedule_func(
self.optimizer , num_warmup_steps=self.args.warmup_steps , num_training_steps=snake_case__ )
return scheduler
def _lowerCAmelCase( self ) -> Optional[torch.utils.data.Sampler]:
if isinstance(self.train_dataset , torch.utils.data.IterableDataset ):
return None
elif is_torch_tpu_available():
return get_tpu_sampler(self.train_dataset )
else:
if self.args.sortish_sampler:
self.train_dataset.make_sortish_sampler(
self.args.per_device_train_batch_size , distributed=(self.args.parallel_mode == ParallelMode.DISTRIBUTED) , )
return (
RandomSampler(self.train_dataset )
if self.args.local_rank == -1
else DistributedSampler(self.train_dataset )
)
def _lowerCAmelCase( self , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) -> Union[str, Any]:
if self.args.label_smoothing == 0:
if self.data_args is not None and self.data_args.ignore_pad_token_for_loss:
# force training to ignore pad token
lowercase__ : Optional[Any] = model(**snake_case__ , use_cache=snake_case__ )[0]
lowercase__ : str = self.loss_fn(logits.view(-1 , logits.shape[-1] ) , labels.view(-1 ) )
else:
# compute usual loss via models
lowercase__ , lowercase__ : List[Any] = model(**snake_case__ , labels=snake_case__ , use_cache=snake_case__ )[:2]
else:
# compute label smoothed loss
lowercase__ : Tuple = model(**snake_case__ , use_cache=snake_case__ )[0]
lowercase__ : Union[str, Any] = torch.nn.functional.log_softmax(snake_case__ , dim=-1 )
lowercase__ , lowercase__ : Tuple = self.loss_fn(snake_case__ , snake_case__ , self.args.label_smoothing , ignore_index=self.config.pad_token_id )
return loss, logits
def _lowerCAmelCase( self , __lowerCAmelCase , __lowerCAmelCase ) -> Optional[int]:
lowercase__ : List[str] = inputs.pop('''labels''' )
lowercase__ , lowercase__ : int = self._compute_loss(snake_case__ , snake_case__ , snake_case__ )
return loss
def _lowerCAmelCase( self , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase = None , ) -> Tuple[Optional[float], Optional[torch.Tensor], Optional[torch.Tensor]]:
lowercase__ : int = self._prepare_inputs(snake_case__ )
lowercase__ : Union[str, Any] = {
'''max_length''': self.data_args.val_max_target_length
if self.data_args is not None
else self.config.max_length,
'''num_beams''': self.data_args.eval_beams if self.data_args is not None else self.config.num_beams,
}
if self.args.predict_with_generate and not self.args.prediction_loss_only:
lowercase__ : Tuple = self.model.generate(
inputs['''input_ids'''] , attention_mask=inputs['''attention_mask'''] , **snake_case__ , )
# in case the batch is shorter than max length, the output should be padded
if generated_tokens.shape[-1] < gen_kwargs["max_length"]:
lowercase__ : List[Any] = self._pad_tensors_to_max_len(snake_case__ , gen_kwargs['''max_length'''] )
lowercase__ : Any = inputs.pop('''labels''' )
with torch.no_grad():
# compute loss on predict data
lowercase__ , lowercase__ : List[str] = self._compute_loss(snake_case__ , snake_case__ , snake_case__ )
lowercase__ : Any = loss.mean().detach()
if self.args.prediction_loss_only:
return (loss, None, None)
lowercase__ : str = generated_tokens if self.args.predict_with_generate else logits
if labels.shape[-1] < gen_kwargs["max_length"]:
lowercase__ : Tuple = self._pad_tensors_to_max_len(snake_case__ , gen_kwargs['''max_length'''] )
return (loss, logits, labels)
def _lowerCAmelCase( self , __lowerCAmelCase , __lowerCAmelCase ) -> Tuple:
lowercase__ : List[str] = self.config.pad_token_id if self.config.pad_token_id is not None else self.config.eos_token_id
if pad_token_id is None:
raise ValueError(
'''Make sure that either `config.pad_token_id` or `config.eos_token_id` is defined if tensor has to be'''
F""" padded to `max_length`={max_length}""" )
lowercase__ : Tuple = pad_token_id * torch.ones(
(tensor.shape[0], max_length) , dtype=tensor.dtype , device=tensor.device )
lowercase__ : Optional[Any] = tensor
return padded_tensor
| 152 |
"""simple docstring"""
from ...utils import (
OptionalDependencyNotAvailable,
is_torch_available,
is_transformers_available,
is_transformers_version,
)
try:
if not (is_transformers_available() and is_torch_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ...utils.dummy_torch_and_transformers_objects import (
ImageTextPipelineOutput,
UniDiffuserPipeline,
)
else:
from .modeling_text_decoder import UniDiffuserTextDecoder
from .modeling_uvit import UniDiffuserModel, UTransformeraDModel
from .pipeline_unidiffuser import ImageTextPipelineOutput, UniDiffuserPipeline
| 673 | 0 |
'''simple docstring'''
import argparse
from transformers import CLIPImageProcessor, CLIPVisionModelWithProjection
from diffusers import UnCLIPImageVariationPipeline, UnCLIPPipeline
if __name__ == "__main__":
__lowercase : List[str] = argparse.ArgumentParser()
parser.add_argument('--dump_path', default=None, type=str, required=True, help='Path to the output model.')
parser.add_argument(
'--txt2img_unclip',
default='kakaobrain/karlo-v1-alpha',
type=str,
required=False,
help='The pretrained txt2img unclip.',
)
__lowercase : Optional[Any] = parser.parse_args()
__lowercase : Optional[Any] = UnCLIPPipeline.from_pretrained(args.txtaimg_unclip)
__lowercase : List[str] = CLIPImageProcessor()
__lowercase : str = CLIPVisionModelWithProjection.from_pretrained('openai/clip-vit-large-patch14')
__lowercase : Optional[Any] = UnCLIPImageVariationPipeline(
decoder=txtaimg.decoder,
text_encoder=txtaimg.text_encoder,
tokenizer=txtaimg.tokenizer,
text_proj=txtaimg.text_proj,
feature_extractor=feature_extractor,
image_encoder=image_encoder,
super_res_first=txtaimg.super_res_first,
super_res_last=txtaimg.super_res_last,
decoder_scheduler=txtaimg.decoder_scheduler,
super_res_scheduler=txtaimg.super_res_scheduler,
)
imgaimg.save_pretrained(args.dump_path)
| 476 |
"""simple docstring"""
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
UNetaDConditionModel,
VideoToVideoSDPipeline,
)
from diffusers.utils import floats_tensor, is_xformers_available, skip_mps
from diffusers.utils.testing_utils import enable_full_determinism, slow, torch_device
from ..pipeline_params import (
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineTesterMixin
enable_full_determinism()
@skip_mps
class UpperCamelCase_ ( a_ , unittest.TestCase ):
_A : str = VideoToVideoSDPipeline
_A : List[str] = TEXT_GUIDED_IMAGE_VARIATION_PARAMS.union({'video'} ) - {'image', 'width', 'height'}
_A : int = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({'video'} ) - {'image'}
_A : int = PipelineTesterMixin.required_optional_params - {'latents'}
_A : List[str] = False
# No `output_type`.
_A : Any = frozenset(
[
'num_inference_steps',
'generator',
'latents',
'return_dict',
'callback',
'callback_steps',
] )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
torch.manual_seed(0 )
UpperCAmelCase = UNetaDConditionModel(
block_out_channels=(32, 64, 64, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=("""CrossAttnDownBlock3D""", """CrossAttnDownBlock3D""", """CrossAttnDownBlock3D""", """DownBlock3D""") , up_block_types=("""UpBlock3D""", """CrossAttnUpBlock3D""", """CrossAttnUpBlock3D""", """CrossAttnUpBlock3D""") , cross_attention_dim=32 , attention_head_dim=4 , )
UpperCAmelCase = DDIMScheduler(
beta_start=0.00_085 , beta_end=0.012 , beta_schedule="""scaled_linear""" , clip_sample=snake_case__ , set_alpha_to_one=snake_case__ , )
torch.manual_seed(0 )
UpperCAmelCase = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] , up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] , latent_channels=4 , sample_size=1_28 , )
torch.manual_seed(0 )
UpperCAmelCase = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=10_00 , hidden_act="""gelu""" , projection_dim=5_12 , )
UpperCAmelCase = CLIPTextModel(snake_case__ )
UpperCAmelCase = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
UpperCAmelCase = {
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
"""text_encoder""": text_encoder,
"""tokenizer""": tokenizer,
}
return components
def UpperCamelCase_ ( self , snake_case__ , snake_case__=0 ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = floats_tensor((1, 3, 3, 32, 32) , rng=random.Random(snake_case__ ) ).to(snake_case__ )
if str(snake_case__ ).startswith("""mps""" ):
UpperCAmelCase = torch.manual_seed(snake_case__ )
else:
UpperCAmelCase = torch.Generator(device=snake_case__ ).manual_seed(snake_case__ )
UpperCAmelCase = {
"""prompt""": """A painting of a squirrel eating a burger""",
"""video""": video,
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 6.0,
"""output_type""": """pt""",
}
return inputs
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
UpperCAmelCase = """cpu""" # ensure determinism for the device-dependent torch.Generator
UpperCAmelCase = self.get_dummy_components()
UpperCAmelCase = VideoToVideoSDPipeline(**snake_case__ )
UpperCAmelCase = sd_pipe.to(snake_case__ )
sd_pipe.set_progress_bar_config(disable=snake_case__ )
UpperCAmelCase = self.get_dummy_inputs(snake_case__ )
UpperCAmelCase = """np"""
UpperCAmelCase = sd_pipe(**snake_case__ ).frames
UpperCAmelCase = frames[0][-3:, -3:, -1]
assert frames[0].shape == (32, 32, 3)
UpperCAmelCase = np.array([1_06, 1_17, 1_13, 1_74, 1_37, 1_12, 1_48, 1_51, 1_31] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(
torch_device != """cuda""" or not is_xformers_available() , reason="""XFormers attention is only available with CUDA and `xformers` installed""" , )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
self._test_xformers_attention_forwardGenerator_pass(test_mean_pixel_difference=snake_case__ , expected_max_diff=5e-3 )
@unittest.skip(reason="""Batching needs to be properly figured out first for this pipeline.""" )
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
pass
@unittest.skip(reason="""Batching needs to be properly figured out first for this pipeline.""" )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
pass
@unittest.skip(reason="""`num_images_per_prompt` argument is not supported for this pipeline.""" )
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
pass
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return super().test_progress_bar()
@slow
@skip_mps
class UpperCamelCase_ ( unittest.TestCase ):
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = VideoToVideoSDPipeline.from_pretrained("""cerspense/zeroscope_v2_XL""" , torch_dtype=torch.floataa )
pipe.enable_model_cpu_offload()
# 10 frames
UpperCAmelCase = torch.Generator(device="""cpu""" ).manual_seed(0 )
UpperCAmelCase = torch.randn((1, 10, 3, 10_24, 5_76) , generator=snake_case__ )
UpperCAmelCase = video.to("""cuda""" )
UpperCAmelCase = """Spiderman is surfing"""
UpperCAmelCase = pipe(snake_case__ , video=snake_case__ , generator=snake_case__ , num_inference_steps=3 , output_type="""pt""" ).frames
UpperCAmelCase = np.array([-1.0_458_984, -1.1_279_297, -0.9_663_086, -0.91_503_906, -0.75_097_656] )
assert np.abs(video_frames.cpu().numpy()[0, 0, 0, 0, -5:] - expected_array ).sum() < 1e-2
| 673 | 0 |
from __future__ import annotations
from typing import Any
def UpperCAmelCase ( UpperCAmelCase )-> Union[str, Any]:
'''simple docstring'''
create_state_space_tree(UpperCAmelCase ,[] ,0 )
def UpperCAmelCase ( UpperCAmelCase ,UpperCAmelCase ,UpperCAmelCase )-> Optional[int]:
'''simple docstring'''
if index == len(UpperCAmelCase ):
print(UpperCAmelCase )
return
create_state_space_tree(UpperCAmelCase ,UpperCAmelCase ,index + 1 )
current_subsequence.append(sequence[index] )
create_state_space_tree(UpperCAmelCase ,UpperCAmelCase ,index + 1 )
current_subsequence.pop()
if __name__ == "__main__":
A_ = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["A", "B", "C"])
generate_all_subsequences(seq)
| 393 |
"""simple docstring"""
import functools
import operator
from ...configuration_utils import PretrainedConfig
from ...utils import logging
lowerCAmelCase_ : int = logging.get_logger(__name__)
lowerCAmelCase_ : Any = {
'''facebook/wav2vec2-base-960h''': '''https://huggingface.co/facebook/wav2vec2-base-960h/resolve/main/config.json''',
# See all Wav2Vec2 models at https://huggingface.co/models?filter=wav2vec2
}
class UpperCamelCase_ ( a_ ):
_A : int = 'wav2vec2'
def __init__( self , snake_case__=32 , snake_case__=7_68 , snake_case__=12 , snake_case__=12 , snake_case__=30_72 , snake_case__="gelu" , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.0 , snake_case__=0.0 , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.02 , snake_case__=1e-5 , snake_case__="group" , snake_case__="gelu" , snake_case__=(5_12, 5_12, 5_12, 5_12, 5_12, 5_12, 5_12) , snake_case__=(5, 2, 2, 2, 2, 2, 2) , snake_case__=(10, 3, 3, 3, 3, 2, 2) , snake_case__=False , snake_case__=1_28 , snake_case__=16 , snake_case__=False , snake_case__=True , snake_case__=0.05 , snake_case__=10 , snake_case__=2 , snake_case__=0.0 , snake_case__=10 , snake_case__=0 , snake_case__=3_20 , snake_case__=2 , snake_case__=0.1 , snake_case__=1_00 , snake_case__=2_56 , snake_case__=2_56 , snake_case__=0.1 , snake_case__="sum" , snake_case__=False , snake_case__=False , snake_case__=2_56 , snake_case__=(5_12, 5_12, 5_12, 5_12, 15_00) , snake_case__=(5, 3, 3, 1, 1) , snake_case__=(1, 2, 3, 1, 1) , snake_case__=5_12 , snake_case__=0 , snake_case__=1 , snake_case__=2 , snake_case__=False , snake_case__=3 , snake_case__=2 , snake_case__=3 , snake_case__=None , snake_case__=None , **snake_case__ , ) -> Optional[Any]:
"""simple docstring"""
super().__init__(**snake_case__ , pad_token_id=snake_case__ , bos_token_id=snake_case__ , eos_token_id=snake_case__ )
UpperCAmelCase = hidden_size
UpperCAmelCase = feat_extract_norm
UpperCAmelCase = feat_extract_activation
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = conv_bias
UpperCAmelCase = num_conv_pos_embeddings
UpperCAmelCase = num_conv_pos_embedding_groups
UpperCAmelCase = len(self.conv_dim )
UpperCAmelCase = num_hidden_layers
UpperCAmelCase = intermediate_size
UpperCAmelCase = hidden_act
UpperCAmelCase = num_attention_heads
UpperCAmelCase = hidden_dropout
UpperCAmelCase = attention_dropout
UpperCAmelCase = activation_dropout
UpperCAmelCase = feat_proj_dropout
UpperCAmelCase = final_dropout
UpperCAmelCase = layerdrop
UpperCAmelCase = layer_norm_eps
UpperCAmelCase = initializer_range
UpperCAmelCase = vocab_size
UpperCAmelCase = do_stable_layer_norm
UpperCAmelCase = use_weighted_layer_sum
if (
(len(self.conv_stride ) != self.num_feat_extract_layers)
or (len(self.conv_kernel ) != self.num_feat_extract_layers)
or (len(self.conv_dim ) != self.num_feat_extract_layers)
):
raise ValueError(
"""Configuration for convolutional layers is incorrect. It is required that `len(config.conv_dim)` =="""
""" `len(config.conv_stride)` == `len(config.conv_kernel)`, but is `len(config.conv_dim) ="""
f''' {len(self.conv_dim )}`, `len(config.conv_stride) = {len(self.conv_stride )}`,'''
f''' `len(config.conv_kernel) = {len(self.conv_kernel )}`.''' )
# fine-tuning config parameters for SpecAugment: https://arxiv.org/abs/1904.08779
UpperCAmelCase = apply_spec_augment
UpperCAmelCase = mask_time_prob
UpperCAmelCase = mask_time_length
UpperCAmelCase = mask_time_min_masks
UpperCAmelCase = mask_feature_prob
UpperCAmelCase = mask_feature_length
UpperCAmelCase = mask_feature_min_masks
# parameters for pretraining with codevector quantized representations
UpperCAmelCase = num_codevectors_per_group
UpperCAmelCase = num_codevector_groups
UpperCAmelCase = contrastive_logits_temperature
UpperCAmelCase = feat_quantizer_dropout
UpperCAmelCase = num_negatives
UpperCAmelCase = codevector_dim
UpperCAmelCase = proj_codevector_dim
UpperCAmelCase = diversity_loss_weight
# ctc loss
UpperCAmelCase = ctc_loss_reduction
UpperCAmelCase = ctc_zero_infinity
# adapter
UpperCAmelCase = add_adapter
UpperCAmelCase = adapter_kernel_size
UpperCAmelCase = adapter_stride
UpperCAmelCase = num_adapter_layers
UpperCAmelCase = output_hidden_size or hidden_size
UpperCAmelCase = adapter_attn_dim
# SequenceClassification-specific parameter. Feel free to ignore for other classes.
UpperCAmelCase = classifier_proj_size
# XVector-specific parameters. Feel free to ignore for other classes.
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = xvector_output_dim
@property
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return functools.reduce(operator.mul , self.conv_stride , 1 )
| 673 | 0 |
def lowercase__ ( __snake_case : Optional[int] ):
'''simple docstring'''
if n == 1 or not isinstance(__snake_case , __snake_case ):
return 0
elif n == 2:
return 1
else:
UpperCAmelCase_ : Tuple = [0, 1]
for i in range(2 , n + 1 ):
sequence.append(sequence[i - 1] + sequence[i - 2] )
return sequence[n]
def lowercase__ ( __snake_case : Dict ):
'''simple docstring'''
UpperCAmelCase_ : int = 0
UpperCAmelCase_ : Dict = 2
while digits < n:
index += 1
UpperCAmelCase_ : Tuple = len(str(fibonacci(__snake_case ) ) )
return index
def lowercase__ ( __snake_case : Union[str, Any] = 1_000 ):
'''simple docstring'''
return fibonacci_digits_index(__snake_case )
if __name__ == "__main__":
print(solution(int(str(input()).strip())))
| 406 |
"""simple docstring"""
import dataclasses
import json
import sys
import types
from argparse import ArgumentDefaultsHelpFormatter, ArgumentParser, ArgumentTypeError
from copy import copy
from enum import Enum
from inspect import isclass
from pathlib import Path
from typing import Any, Callable, Dict, Iterable, List, Literal, NewType, Optional, Tuple, Union, get_type_hints
import yaml
lowerCAmelCase_ : Optional[Any] = NewType('''DataClass''', Any)
lowerCAmelCase_ : Any = NewType('''DataClassType''', Any)
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
if isinstance(lowerCAmelCase , lowerCAmelCase ):
return v
if v.lower() in ("yes", "true", "t", "y", "1"):
return True
elif v.lower() in ("no", "false", "f", "n", "0"):
return False
else:
raise ArgumentTypeError(
F'''Truthy value expected: got {v} but expected one of yes/no, true/false, t/f, y/n, 1/0 (case insensitive).''' )
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = {str(lowerCAmelCase ): choice for choice in choices}
return lambda lowerCAmelCase : str_to_choice.get(lowerCAmelCase , lowerCAmelCase )
def _lowerCAmelCase ( *,
lowerCAmelCase = None , lowerCAmelCase = None , lowerCAmelCase = dataclasses.MISSING , lowerCAmelCase = dataclasses.MISSING , lowerCAmelCase = None , **lowerCAmelCase , ):
'''simple docstring'''
if metadata is None:
# Important, don't use as default param in function signature because dict is mutable and shared across function calls
UpperCAmelCase = {}
if aliases is not None:
UpperCAmelCase = aliases
if help is not None:
UpperCAmelCase = help
return dataclasses.field(metadata=lowerCAmelCase , default=lowerCAmelCase , default_factory=lowerCAmelCase , **lowerCAmelCase )
class UpperCamelCase_ ( a_ ):
_A : Iterable[DataClassType]
def __init__( self , snake_case__ , **snake_case__ ) -> List[str]:
"""simple docstring"""
if "formatter_class" not in kwargs:
UpperCAmelCase = ArgumentDefaultsHelpFormatter
super().__init__(**snake_case__ )
if dataclasses.is_dataclass(snake_case__ ):
UpperCAmelCase = [dataclass_types]
UpperCAmelCase = list(snake_case__ )
for dtype in self.dataclass_types:
self._add_dataclass_arguments(snake_case__ )
@staticmethod
def UpperCamelCase_ ( snake_case__ , snake_case__ ) -> str:
"""simple docstring"""
UpperCAmelCase = f'''--{field.name}'''
UpperCAmelCase = field.metadata.copy()
# field.metadata is not used at all by Data Classes,
# it is provided as a third-party extension mechanism.
if isinstance(field.type , snake_case__ ):
raise RuntimeError(
"""Unresolved type detected, which should have been done with the help of """
"""`typing.get_type_hints` method by default""" )
UpperCAmelCase = kwargs.pop("""aliases""" , [] )
if isinstance(snake_case__ , snake_case__ ):
UpperCAmelCase = [aliases]
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
if origin_type is Union or (hasattr(snake_case__ , """UnionType""" ) and isinstance(snake_case__ , types.UnionType )):
if str not in field.type.__args__ and (
len(field.type.__args__ ) != 2 or type(snake_case__ ) not in field.type.__args__
):
raise ValueError(
"""Only `Union[X, NoneType]` (i.e., `Optional[X]`) is allowed for `Union` because"""
""" the argument parser only supports one type per argument."""
f''' Problem encountered in field \'{field.name}\'.''' )
if type(snake_case__ ) not in field.type.__args__:
# filter `str` in Union
UpperCAmelCase = field.type.__args__[0] if field.type.__args__[1] == str else field.type.__args__[1]
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
elif bool not in field.type.__args__:
# filter `NoneType` in Union (except for `Union[bool, NoneType]`)
UpperCAmelCase = (
field.type.__args__[0] if isinstance(snake_case__ , field.type.__args__[1] ) else field.type.__args__[1]
)
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
# A variable to store kwargs for a boolean field, if needed
# so that we can init a `no_*` complement argument (see below)
UpperCAmelCase = {}
if origin_type is Literal or (isinstance(field.type , snake_case__ ) and issubclass(field.type , snake_case__ )):
if origin_type is Literal:
UpperCAmelCase = field.type.__args__
else:
UpperCAmelCase = [x.value for x in field.type]
UpperCAmelCase = make_choice_type_function(kwargs["""choices"""] )
if field.default is not dataclasses.MISSING:
UpperCAmelCase = field.default
else:
UpperCAmelCase = True
elif field.type is bool or field.type == Optional[bool]:
# Copy the currect kwargs to use to instantiate a `no_*` complement argument below.
# We do not initialize it here because the `no_*` alternative must be instantiated after the real argument
UpperCAmelCase = copy(snake_case__ )
# Hack because type=bool in argparse does not behave as we want.
UpperCAmelCase = string_to_bool
if field.type is bool or (field.default is not None and field.default is not dataclasses.MISSING):
# Default value is False if we have no default when of type bool.
UpperCAmelCase = False if field.default is dataclasses.MISSING else field.default
# This is the value that will get picked if we don't include --field_name in any way
UpperCAmelCase = default
# This tells argparse we accept 0 or 1 value after --field_name
UpperCAmelCase = """?"""
# This is the value that will get picked if we do --field_name (without value)
UpperCAmelCase = True
elif isclass(snake_case__ ) and issubclass(snake_case__ , snake_case__ ):
UpperCAmelCase = field.type.__args__[0]
UpperCAmelCase = """+"""
if field.default_factory is not dataclasses.MISSING:
UpperCAmelCase = field.default_factory()
elif field.default is dataclasses.MISSING:
UpperCAmelCase = True
else:
UpperCAmelCase = field.type
if field.default is not dataclasses.MISSING:
UpperCAmelCase = field.default
elif field.default_factory is not dataclasses.MISSING:
UpperCAmelCase = field.default_factory()
else:
UpperCAmelCase = True
parser.add_argument(snake_case__ , *snake_case__ , **snake_case__ )
# Add a complement `no_*` argument for a boolean field AFTER the initial field has already been added.
# Order is important for arguments with the same destination!
# We use a copy of earlier kwargs because the original kwargs have changed a lot before reaching down
# here and we do not need those changes/additional keys.
if field.default is True and (field.type is bool or field.type == Optional[bool]):
UpperCAmelCase = False
parser.add_argument(f'''--no_{field.name}''' , action="""store_false""" , dest=field.name , **snake_case__ )
def UpperCamelCase_ ( self , snake_case__ ) -> Any:
"""simple docstring"""
if hasattr(snake_case__ , """_argument_group_name""" ):
UpperCAmelCase = self.add_argument_group(dtype._argument_group_name )
else:
UpperCAmelCase = self
try:
UpperCAmelCase = get_type_hints(snake_case__ )
except NameError:
raise RuntimeError(
f'''Type resolution failed for {dtype}. Try declaring the class in global scope or '''
"""removing line of `from __future__ import annotations` which opts in Postponed """
"""Evaluation of Annotations (PEP 563)""" )
except TypeError as ex:
# Remove this block when we drop Python 3.9 support
if sys.version_info[:2] < (3, 10) and "unsupported operand type(s) for |" in str(snake_case__ ):
UpperCAmelCase = """.""".join(map(snake_case__ , sys.version_info[:3] ) )
raise RuntimeError(
f'''Type resolution failed for {dtype} on Python {python_version}. Try removing '''
"""line of `from __future__ import annotations` which opts in union types as """
"""`X | Y` (PEP 604) via Postponed Evaluation of Annotations (PEP 563). To """
"""support Python versions that lower than 3.10, you need to use """
"""`typing.Union[X, Y]` instead of `X | Y` and `typing.Optional[X]` instead of """
"""`X | None`.""" ) from ex
raise
for field in dataclasses.fields(snake_case__ ):
if not field.init:
continue
UpperCAmelCase = type_hints[field.name]
self._parse_dataclass_field(snake_case__ , snake_case__ )
def UpperCamelCase_ ( self , snake_case__=None , snake_case__=False , snake_case__=True , snake_case__=None , snake_case__=None , ) -> Tuple[DataClass, ...]:
"""simple docstring"""
if args_file_flag or args_filename or (look_for_args_file and len(sys.argv )):
UpperCAmelCase = []
if args_filename:
args_files.append(Path(snake_case__ ) )
elif look_for_args_file and len(sys.argv ):
args_files.append(Path(sys.argv[0] ).with_suffix(""".args""" ) )
# args files specified via command line flag should overwrite default args files so we add them last
if args_file_flag:
# Create special parser just to extract the args_file_flag values
UpperCAmelCase = ArgumentParser()
args_file_parser.add_argument(snake_case__ , type=snake_case__ , action="""append""" )
# Use only remaining args for further parsing (remove the args_file_flag)
UpperCAmelCase , UpperCAmelCase = args_file_parser.parse_known_args(args=snake_case__ )
UpperCAmelCase = vars(snake_case__ ).get(args_file_flag.lstrip("""-""" ) , snake_case__ )
if cmd_args_file_paths:
args_files.extend([Path(snake_case__ ) for p in cmd_args_file_paths] )
UpperCAmelCase = []
for args_file in args_files:
if args_file.exists():
file_args += args_file.read_text().split()
# in case of duplicate arguments the last one has precedence
# args specified via the command line should overwrite args from files, so we add them last
UpperCAmelCase = file_args + args if args is not None else file_args + sys.argv[1:]
UpperCAmelCase , UpperCAmelCase = self.parse_known_args(args=snake_case__ )
UpperCAmelCase = []
for dtype in self.dataclass_types:
UpperCAmelCase = {f.name for f in dataclasses.fields(snake_case__ ) if f.init}
UpperCAmelCase = {k: v for k, v in vars(snake_case__ ).items() if k in keys}
for k in keys:
delattr(snake_case__ , snake_case__ )
UpperCAmelCase = dtype(**snake_case__ )
outputs.append(snake_case__ )
if len(namespace.__dict__ ) > 0:
# additional namespace.
outputs.append(snake_case__ )
if return_remaining_strings:
return (*outputs, remaining_args)
else:
if remaining_args:
raise ValueError(f'''Some specified arguments are not used by the HfArgumentParser: {remaining_args}''' )
return (*outputs,)
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
UpperCAmelCase = set(args.keys() )
UpperCAmelCase = []
for dtype in self.dataclass_types:
UpperCAmelCase = {f.name for f in dataclasses.fields(snake_case__ ) if f.init}
UpperCAmelCase = {k: v for k, v in args.items() if k in keys}
unused_keys.difference_update(inputs.keys() )
UpperCAmelCase = dtype(**snake_case__ )
outputs.append(snake_case__ )
if not allow_extra_keys and unused_keys:
raise ValueError(f'''Some keys are not used by the HfArgumentParser: {sorted(snake_case__ )}''' )
return tuple(snake_case__ )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
with open(Path(snake_case__ ) , encoding="""utf-8""" ) as open_json_file:
UpperCAmelCase = json.loads(open_json_file.read() )
UpperCAmelCase = self.parse_dict(snake_case__ , allow_extra_keys=snake_case__ )
return tuple(snake_case__ )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
UpperCAmelCase = self.parse_dict(yaml.safe_load(Path(snake_case__ ).read_text() ) , allow_extra_keys=snake_case__ )
return tuple(snake_case__ )
| 673 | 0 |
"""simple docstring"""
import gc
import tempfile
import unittest
import numpy as np
import torch
from diffusers import VersatileDiffusionPipeline
from diffusers.utils.testing_utils import load_image, nightly, require_torch_gpu, torch_device
__UpperCAmelCase =False
class lowerCAmelCase__ ( unittest.TestCase ):
pass
@nightly
@require_torch_gpu
class lowerCAmelCase__ ( unittest.TestCase ):
def lowercase_ ( self ):
'''simple docstring'''
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def lowercase_ ( self ):
'''simple docstring'''
A__ = VersatileDiffusionPipeline.from_pretrained("shi-labs/versatile-diffusion" , torch_dtype=torch.floataa )
pipe.to(snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
A__ = load_image(
"https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/versatile_diffusion/benz.jpg" )
A__ = torch.manual_seed(0 )
A__ = pipe.dual_guided(
prompt="first prompt" , image=snake_case__ , text_to_image_strength=0.75 , generator=snake_case__ , guidance_scale=7.5 , num_inference_steps=2 , output_type="numpy" , ).images
with tempfile.TemporaryDirectory() as tmpdirname:
pipe.save_pretrained(snake_case__ )
A__ = VersatileDiffusionPipeline.from_pretrained(snake_case__ , torch_dtype=torch.floataa )
pipe.to(snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
A__ = generator.manual_seed(0 )
A__ = pipe.dual_guided(
prompt="first prompt" , image=snake_case__ , text_to_image_strength=0.75 , generator=snake_case__ , guidance_scale=7.5 , num_inference_steps=2 , output_type="numpy" , ).images
assert np.abs(image - new_image ).sum() < 1e-5, "Models don't have the same forward pass"
def lowercase_ ( self ):
'''simple docstring'''
A__ = VersatileDiffusionPipeline.from_pretrained("shi-labs/versatile-diffusion" , torch_dtype=torch.floataa )
pipe.to(snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
A__ = "cyberpunk 2077"
A__ = load_image(
"https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/versatile_diffusion/benz.jpg" )
A__ = torch.manual_seed(0 )
A__ = pipe.dual_guided(
prompt=snake_case__ , image=snake_case__ , text_to_image_strength=0.75 , generator=snake_case__ , guidance_scale=7.5 , num_inference_steps=50 , output_type="numpy" , ).images
A__ = image[0, 2_53:2_56, 2_53:2_56, -1]
assert image.shape == (1, 5_12, 5_12, 3)
A__ = np.array([0.1448, 0.1619, 0.1741, 0.1086, 0.1147, 0.1128, 0.1199, 0.1165, 0.1001] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-1
A__ = "A painting of a squirrel eating a burger "
A__ = torch.manual_seed(0 )
A__ = pipe.text_to_image(
prompt=snake_case__ , generator=snake_case__ , guidance_scale=7.5 , num_inference_steps=50 , output_type="numpy" ).images
A__ = image[0, 2_53:2_56, 2_53:2_56, -1]
assert image.shape == (1, 5_12, 5_12, 3)
A__ = np.array([0.3367, 0.3169, 0.2656, 0.3870, 0.4790, 0.3796, 0.4009, 0.4878, 0.4778] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-1
A__ = pipe.image_variation(snake_case__ , generator=snake_case__ , output_type="numpy" ).images
A__ = image[0, 2_53:2_56, 2_53:2_56, -1]
assert image.shape == (1, 5_12, 5_12, 3)
A__ = np.array([0.3076, 0.3123, 0.3284, 0.3782, 0.3770, 0.3894, 0.4297, 0.4331, 0.4456] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-1 | 337 |
"""simple docstring"""
import unittest
import torch
from diffusers import DDIMScheduler, DDPMScheduler, UNetaDModel
from diffusers.training_utils import set_seed
from diffusers.utils.testing_utils import slow
lowerCAmelCase_ : List[str] = False
class UpperCamelCase_ ( unittest.TestCase ):
def UpperCamelCase_ ( self , snake_case__=32 ) -> Optional[Any]:
"""simple docstring"""
set_seed(0 )
UpperCAmelCase = UNetaDModel(sample_size=snake_case__ , in_channels=3 , out_channels=3 )
UpperCAmelCase = torch.optim.SGD(model.parameters() , lr=0.0_001 )
return model, optimizer
@slow
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = """cpu""" # ensure full determinism without setting the CUBLAS_WORKSPACE_CONFIG env variable
UpperCAmelCase = DDPMScheduler(
num_train_timesteps=10_00 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule="""linear""" , clip_sample=snake_case__ , )
UpperCAmelCase = DDIMScheduler(
num_train_timesteps=10_00 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule="""linear""" , clip_sample=snake_case__ , )
assert ddpm_scheduler.config.num_train_timesteps == ddim_scheduler.config.num_train_timesteps
# shared batches for DDPM and DDIM
set_seed(0 )
UpperCAmelCase = [torch.randn((4, 3, 32, 32) ).clip(-1 , 1 ).to(snake_case__ ) for _ in range(4 )]
UpperCAmelCase = [torch.randn((4, 3, 32, 32) ).to(snake_case__ ) for _ in range(4 )]
UpperCAmelCase = [torch.randint(0 , 10_00 , (4,) ).long().to(snake_case__ ) for _ in range(4 )]
# train with a DDPM scheduler
UpperCAmelCase , UpperCAmelCase = self.get_model_optimizer(resolution=32 )
model.train().to(snake_case__ )
for i in range(4 ):
optimizer.zero_grad()
UpperCAmelCase = ddpm_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i] )
UpperCAmelCase = model(snake_case__ , timesteps[i] ).sample
UpperCAmelCase = torch.nn.functional.mse_loss(snake_case__ , noise[i] )
loss.backward()
optimizer.step()
del model, optimizer
# recreate the model and optimizer, and retry with DDIM
UpperCAmelCase , UpperCAmelCase = self.get_model_optimizer(resolution=32 )
model.train().to(snake_case__ )
for i in range(4 ):
optimizer.zero_grad()
UpperCAmelCase = ddim_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i] )
UpperCAmelCase = model(snake_case__ , timesteps[i] ).sample
UpperCAmelCase = torch.nn.functional.mse_loss(snake_case__ , noise[i] )
loss.backward()
optimizer.step()
del model, optimizer
self.assertTrue(torch.allclose(snake_case__ , snake_case__ , atol=1e-5 ) )
self.assertTrue(torch.allclose(snake_case__ , snake_case__ , atol=1e-5 ) )
| 673 | 0 |
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__UpperCAmelCase =logging.get_logger(__name__)
__UpperCAmelCase ={
'''sayakpaul/vit-msn-base''': '''https://huggingface.co/sayakpaul/vit-msn-base/resolve/main/config.json''',
# See all ViT MSN models at https://huggingface.co/models?filter=vit_msn
}
class a__ ( a_ ):
lowerCamelCase : str ='vit_msn'
def __init__( self : List[str] , a : Dict=7_68 , a : List[Any]=12 , a : Dict=12 , a : Dict=30_72 , a : Dict="gelu" , a : Optional[int]=0.0 , a : Union[str, Any]=0.0 , a : Any=0.02 , a : Dict=1e-0_6 , a : Optional[Any]=2_24 , a : Optional[Any]=16 , a : Dict=3 , a : Dict=True , **a : List[Any] , ):
"""simple docstring"""
super().__init__(**snake_case__ )
__lowerCamelCase = hidden_size
__lowerCamelCase = num_hidden_layers
__lowerCamelCase = num_attention_heads
__lowerCamelCase = intermediate_size
__lowerCamelCase = hidden_act
__lowerCamelCase = hidden_dropout_prob
__lowerCamelCase = attention_probs_dropout_prob
__lowerCamelCase = initializer_range
__lowerCamelCase = layer_norm_eps
__lowerCamelCase = image_size
__lowerCamelCase = patch_size
__lowerCamelCase = num_channels
__lowerCamelCase = qkv_bias
| 546 |
"""simple docstring"""
import numpy as np
import torch
from torch.utils.data import DataLoader
from accelerate.utils.dataclasses import DistributedType
class UpperCamelCase_ :
def __init__( self , snake_case__=2 , snake_case__=3 , snake_case__=64 , snake_case__=None ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = np.random.default_rng(snake_case__ )
UpperCAmelCase = length
UpperCAmelCase = rng.normal(size=(length,) ).astype(np.floataa )
UpperCAmelCase = a * self.x + b + rng.normal(scale=0.1 , size=(length,) ).astype(np.floataa )
def __len__( self ) -> int:
"""simple docstring"""
return self.length
def __getitem__( self , snake_case__ ) -> Tuple:
"""simple docstring"""
return {"x": self.x[i], "y": self.y[i]}
class UpperCamelCase_ ( torch.nn.Module ):
def __init__( self , snake_case__=0 , snake_case__=0 , snake_case__=False ) -> List[str]:
"""simple docstring"""
super().__init__()
UpperCAmelCase = torch.nn.Parameter(torch.tensor([2, 3] ).float() )
UpperCAmelCase = torch.nn.Parameter(torch.tensor([2, 3] ).float() )
UpperCAmelCase = True
def UpperCamelCase_ ( self , snake_case__=None ) -> List[Any]:
"""simple docstring"""
if self.first_batch:
print(f'''Model dtype: {self.a.dtype}, {self.b.dtype}. Input dtype: {x.dtype}''' )
UpperCAmelCase = False
return x * self.a[0] + self.b[0]
class UpperCamelCase_ ( torch.nn.Module ):
def __init__( self , snake_case__=0 , snake_case__=0 , snake_case__=False ) -> List[Any]:
"""simple docstring"""
super().__init__()
UpperCAmelCase = torch.nn.Parameter(torch.tensor(snake_case__ ).float() )
UpperCAmelCase = torch.nn.Parameter(torch.tensor(snake_case__ ).float() )
UpperCAmelCase = True
def UpperCamelCase_ ( self , snake_case__=None ) -> Optional[Any]:
"""simple docstring"""
if self.first_batch:
print(f'''Model dtype: {self.a.dtype}, {self.b.dtype}. Input dtype: {x.dtype}''' )
UpperCAmelCase = False
return x * self.a + self.b
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase = 16 ):
'''simple docstring'''
from datasets import load_dataset
from transformers import AutoTokenizer
UpperCAmelCase = AutoTokenizer.from_pretrained("""bert-base-cased""" )
UpperCAmelCase = {"""train""": """tests/test_samples/MRPC/train.csv""", """validation""": """tests/test_samples/MRPC/dev.csv"""}
UpperCAmelCase = load_dataset("""csv""" , data_files=lowerCAmelCase )
UpperCAmelCase = datasets["""train"""].unique("""label""" )
UpperCAmelCase = {v: i for i, v in enumerate(lowerCAmelCase )}
def tokenize_function(lowerCAmelCase ):
# max_length=None => use the model max length (it's actually the default)
UpperCAmelCase = tokenizer(
examples["""sentence1"""] , examples["""sentence2"""] , truncation=lowerCAmelCase , max_length=lowerCAmelCase , padding="""max_length""" )
if "label" in examples:
UpperCAmelCase = [label_to_id[l] for l in examples["""label"""]]
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
UpperCAmelCase = datasets.map(
lowerCAmelCase , batched=lowerCAmelCase , remove_columns=["""sentence1""", """sentence2""", """label"""] , )
def collate_fn(lowerCAmelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
if accelerator.distributed_type == DistributedType.TPU:
return tokenizer.pad(lowerCAmelCase , padding="""max_length""" , max_length=128 , return_tensors="""pt""" )
return tokenizer.pad(lowerCAmelCase , padding="""longest""" , return_tensors="""pt""" )
# Instantiate dataloaders.
UpperCAmelCase = DataLoader(tokenized_datasets["""train"""] , shuffle=lowerCAmelCase , collate_fn=lowerCAmelCase , batch_size=2 )
UpperCAmelCase = DataLoader(tokenized_datasets["""validation"""] , shuffle=lowerCAmelCase , collate_fn=lowerCAmelCase , batch_size=1 )
return train_dataloader, eval_dataloader
| 673 | 0 |
import logging
import os
import sys
from dataclasses import dataclass, field
from typing import Optional
import torch
from datasets import load_dataset
from torchvision.transforms import Compose, Lambda, Normalize, RandomHorizontalFlip, RandomResizedCrop, ToTensor
from torchvision.transforms.functional import InterpolationMode
import transformers
from transformers import (
HfArgumentParser,
Trainer,
TrainingArguments,
ViTImageProcessor,
ViTMAEConfig,
ViTMAEForPreTraining,
)
from transformers.trainer_utils import get_last_checkpoint
from transformers.utils import check_min_version, send_example_telemetry
from transformers.utils.versions import require_version
lowerCamelCase = logging.getLogger(__name__)
# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
check_min_version('4.31.0')
require_version('datasets>=1.8.0', 'To fix: pip install -r examples/pytorch/image-pretraining/requirements.txt')
@dataclass
class A :
UpperCamelCase__ : Optional[str] =field(
default='cifar10' , metadata={'help': 'Name of a dataset from the datasets package'} )
UpperCamelCase__ : Optional[str] =field(
default=a_ , metadata={'help': 'The configuration name of the dataset to use (via the datasets library).'} )
UpperCamelCase__ : Optional[str] =field(
default=a_ , metadata={'help': 'The column name of the images in the files.'} )
UpperCamelCase__ : Optional[str] =field(default=a_ , metadata={'help': 'A folder containing the training data.'} )
UpperCamelCase__ : Optional[str] =field(default=a_ , metadata={'help': 'A folder containing the validation data.'} )
UpperCamelCase__ : Optional[float] =field(
default=0.15 , metadata={'help': 'Percent to split off of train for validation.'} )
UpperCamelCase__ : Optional[int] =field(
default=a_ , metadata={
'help': (
'For debugging purposes or quicker training, truncate the number of training examples to this '
'value if set.'
)
} , )
UpperCamelCase__ : Optional[int] =field(
default=a_ , metadata={
'help': (
'For debugging purposes or quicker training, truncate the number of evaluation examples to this '
'value if set.'
)
} , )
def lowerCamelCase ( self : List[str] ) -> List[Any]:
"""simple docstring"""
_lowerCamelCase : List[str] ={}
if self.train_dir is not None:
_lowerCamelCase : List[str] =self.train_dir
if self.validation_dir is not None:
_lowerCamelCase : Tuple =self.validation_dir
_lowerCamelCase : List[str] =data_files if data_files else None
@dataclass
class A :
UpperCamelCase__ : str =field(
default=a_ , metadata={
'help': (
'The model checkpoint for weights initialization.Don\'t set if you want to train a model from scratch.'
)
} , )
UpperCamelCase__ : Optional[str] =field(
default=a_ , metadata={'help': 'Pretrained config name or path if not the same as model_name_or_path'} )
UpperCamelCase__ : Optional[str] =field(
default=a_ , metadata={
'help': (
'Override some existing default config settings when a model is trained from scratch. Example: '
'n_embd=10,resid_pdrop=0.2,scale_attn_weights=false,summary_type=cls_index'
)
} , )
UpperCamelCase__ : Optional[str] =field(
default=a_ , metadata={'help': 'Where do you want to store the pretrained models downloaded from s3'} )
UpperCamelCase__ : str =field(
default='main' , metadata={'help': 'The specific model version to use (can be a branch name, tag name or commit id).'} , )
UpperCamelCase__ : str =field(default=a_ , metadata={'help': 'Name or path of preprocessor config.'} )
UpperCamelCase__ : bool =field(
default=a_ , metadata={
'help': (
'Will use the token generated when running `huggingface-cli login` (necessary to use this script '
'with private models).'
)
} , )
UpperCamelCase__ : float =field(
default=0.75 , metadata={'help': 'The ratio of the number of masked tokens in the input sequence.'} )
UpperCamelCase__ : bool =field(
default=a_ , metadata={'help': 'Whether or not to train with normalized pixel values as target.'} )
@dataclass
class A ( a_ ):
UpperCamelCase__ : float =field(
default=1E-3 , metadata={'help': 'Base learning rate: absolute_lr = base_lr * total_batch_size / 256.'} )
def a_ ( SCREAMING_SNAKE_CASE__ : str ):
'''simple docstring'''
_lowerCamelCase : Tuple =torch.stack([example['pixel_values'] for example in examples] )
return {"pixel_values": pixel_values}
def a_ ( ):
'''simple docstring'''
_lowerCamelCase : Any =HfArgumentParser((ModelArguments, DataTrainingArguments, CustomTrainingArguments) )
if len(sys.argv ) == 2 and sys.argv[1].endswith('.json' ):
# If we pass only one argument to the script and it's the path to a json file,
# let's parse it to get our arguments.
_lowerCamelCase , _lowerCamelCase , _lowerCamelCase : Any =parser.parse_json_file(json_file=os.path.abspath(sys.argv[1] ) )
else:
_lowerCamelCase , _lowerCamelCase , _lowerCamelCase : List[str] =parser.parse_args_into_dataclasses()
# Sending telemetry. Tracking the example usage helps us better allocate resources to maintain them. The
# information sent is the one passed as arguments along with your Python/PyTorch versions.
send_example_telemetry('run_mae' , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
# Setup logging
logging.basicConfig(
format='%(asctime)s - %(levelname)s - %(name)s - %(message)s' , datefmt='%m/%d/%Y %H:%M:%S' , handlers=[logging.StreamHandler(sys.stdout )] , )
if training_args.should_log:
# The default of training_args.log_level is passive, so we set log level at info here to have that default.
transformers.utils.logging.set_verbosity_info()
_lowerCamelCase : Optional[int] =training_args.get_process_log_level()
logger.setLevel(SCREAMING_SNAKE_CASE__ )
transformers.utils.logging.set_verbosity(SCREAMING_SNAKE_CASE__ )
transformers.utils.logging.enable_default_handler()
transformers.utils.logging.enable_explicit_format()
# Log on each process the small summary:
logger.warning(
F'''Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}'''
+ F'''distributed training: {bool(training_args.local_rank != -1 )}, 16-bits training: {training_args.fpaa}''' )
logger.info(F'''Training/evaluation parameters {training_args}''' )
# Detecting last checkpoint.
_lowerCamelCase : List[Any] =None
if os.path.isdir(training_args.output_dir ) and training_args.do_train and not training_args.overwrite_output_dir:
_lowerCamelCase : Union[str, Any] =get_last_checkpoint(training_args.output_dir )
if last_checkpoint is None and len(os.listdir(training_args.output_dir ) ) > 0:
raise ValueError(
F'''Output directory ({training_args.output_dir}) already exists and is not empty. '''
'Use --overwrite_output_dir to overcome.' )
elif last_checkpoint is not None and training_args.resume_from_checkpoint is None:
logger.info(
F'''Checkpoint detected, resuming training at {last_checkpoint}. To avoid this behavior, change '''
'the `--output_dir` or add `--overwrite_output_dir` to train from scratch.' )
# Initialize our dataset.
_lowerCamelCase : int =load_dataset(
data_args.dataset_name , data_args.dataset_config_name , data_files=data_args.data_files , cache_dir=model_args.cache_dir , use_auth_token=True if model_args.use_auth_token else None , )
# If we don't have a validation split, split off a percentage of train as validation.
_lowerCamelCase : str =None if 'validation' in ds.keys() else data_args.train_val_split
if isinstance(data_args.train_val_split , SCREAMING_SNAKE_CASE__ ) and data_args.train_val_split > 0.0:
_lowerCamelCase : Union[str, Any] =ds['train'].train_test_split(data_args.train_val_split )
_lowerCamelCase : Dict =split['train']
_lowerCamelCase : List[str] =split['test']
# Load pretrained model and image processor
#
# Distributed training:
# The .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
_lowerCamelCase : Optional[int] ={
'cache_dir': model_args.cache_dir,
'revision': model_args.model_revision,
'use_auth_token': True if model_args.use_auth_token else None,
}
if model_args.config_name:
_lowerCamelCase : Tuple =ViTMAEConfig.from_pretrained(model_args.config_name , **SCREAMING_SNAKE_CASE__ )
elif model_args.model_name_or_path:
_lowerCamelCase : List[Any] =ViTMAEConfig.from_pretrained(model_args.model_name_or_path , **SCREAMING_SNAKE_CASE__ )
else:
_lowerCamelCase : Union[str, Any] =ViTMAEConfig()
logger.warning('You are instantiating a new config instance from scratch.' )
if model_args.config_overrides is not None:
logger.info(F'''Overriding config: {model_args.config_overrides}''' )
config.update_from_string(model_args.config_overrides )
logger.info(F'''New config: {config}''' )
# adapt config
config.update(
{
'mask_ratio': model_args.mask_ratio,
'norm_pix_loss': model_args.norm_pix_loss,
} )
# create image processor
if model_args.image_processor_name:
_lowerCamelCase : List[Any] =ViTImageProcessor.from_pretrained(model_args.image_processor_name , **SCREAMING_SNAKE_CASE__ )
elif model_args.model_name_or_path:
_lowerCamelCase : Optional[int] =ViTImageProcessor.from_pretrained(model_args.model_name_or_path , **SCREAMING_SNAKE_CASE__ )
else:
_lowerCamelCase : int =ViTImageProcessor()
# create model
if model_args.model_name_or_path:
_lowerCamelCase : List[str] =ViTMAEForPreTraining.from_pretrained(
model_args.model_name_or_path , from_tf=bool('.ckpt' in model_args.model_name_or_path ) , config=SCREAMING_SNAKE_CASE__ , cache_dir=model_args.cache_dir , revision=model_args.model_revision , use_auth_token=True if model_args.use_auth_token else None , )
else:
logger.info('Training new model from scratch' )
_lowerCamelCase : Any =ViTMAEForPreTraining(SCREAMING_SNAKE_CASE__ )
if training_args.do_train:
_lowerCamelCase : List[str] =ds['train'].column_names
else:
_lowerCamelCase : Optional[Any] =ds['validation'].column_names
if data_args.image_column_name is not None:
_lowerCamelCase : int =data_args.image_column_name
elif "image" in column_names:
_lowerCamelCase : Union[str, Any] ='image'
elif "img" in column_names:
_lowerCamelCase : int ='img'
else:
_lowerCamelCase : List[str] =column_names[0]
# transformations as done in original MAE paper
# source: https://github.com/facebookresearch/mae/blob/main/main_pretrain.py
if "shortest_edge" in image_processor.size:
_lowerCamelCase : Optional[int] =image_processor.size['shortest_edge']
else:
_lowerCamelCase : List[str] =(image_processor.size['height'], image_processor.size['width'])
_lowerCamelCase : Optional[int] =Compose(
[
Lambda(lambda SCREAMING_SNAKE_CASE__ : img.convert('RGB' ) if img.mode != "RGB" else img ),
RandomResizedCrop(SCREAMING_SNAKE_CASE__ , scale=(0.2, 1.0) , interpolation=InterpolationMode.BICUBIC ),
RandomHorizontalFlip(),
ToTensor(),
Normalize(mean=image_processor.image_mean , std=image_processor.image_std ),
] )
def preprocess_images(SCREAMING_SNAKE_CASE__ : int ):
_lowerCamelCase : str =[transforms(SCREAMING_SNAKE_CASE__ ) for image in examples[image_column_name]]
return examples
if training_args.do_train:
if "train" not in ds:
raise ValueError('--do_train requires a train dataset' )
if data_args.max_train_samples is not None:
_lowerCamelCase : Dict =ds['train'].shuffle(seed=training_args.seed ).select(range(data_args.max_train_samples ) )
# Set the training transforms
ds["train"].set_transform(SCREAMING_SNAKE_CASE__ )
if training_args.do_eval:
if "validation" not in ds:
raise ValueError('--do_eval requires a validation dataset' )
if data_args.max_eval_samples is not None:
_lowerCamelCase : List[Any] =(
ds['validation'].shuffle(seed=training_args.seed ).select(range(data_args.max_eval_samples ) )
)
# Set the validation transforms
ds["validation"].set_transform(SCREAMING_SNAKE_CASE__ )
# Compute absolute learning rate
_lowerCamelCase : Any =(
training_args.train_batch_size * training_args.gradient_accumulation_steps * training_args.world_size
)
if training_args.base_learning_rate is not None:
_lowerCamelCase : List[str] =training_args.base_learning_rate * total_train_batch_size / 256
# Initialize our trainer
_lowerCamelCase : List[Any] =Trainer(
model=SCREAMING_SNAKE_CASE__ , args=SCREAMING_SNAKE_CASE__ , train_dataset=ds['train'] if training_args.do_train else None , eval_dataset=ds['validation'] if training_args.do_eval else None , tokenizer=SCREAMING_SNAKE_CASE__ , data_collator=SCREAMING_SNAKE_CASE__ , )
# Training
if training_args.do_train:
_lowerCamelCase : Tuple =None
if training_args.resume_from_checkpoint is not None:
_lowerCamelCase : Optional[Any] =training_args.resume_from_checkpoint
elif last_checkpoint is not None:
_lowerCamelCase : Any =last_checkpoint
_lowerCamelCase : List[str] =trainer.train(resume_from_checkpoint=SCREAMING_SNAKE_CASE__ )
trainer.save_model()
trainer.log_metrics('train' , train_result.metrics )
trainer.save_metrics('train' , train_result.metrics )
trainer.save_state()
# Evaluation
if training_args.do_eval:
_lowerCamelCase : Dict =trainer.evaluate()
trainer.log_metrics('eval' , SCREAMING_SNAKE_CASE__ )
trainer.save_metrics('eval' , SCREAMING_SNAKE_CASE__ )
# Write model card and (optionally) push to hub
_lowerCamelCase : List[Any] ={
'tasks': 'masked-auto-encoding',
'dataset': data_args.dataset_name,
'tags': ['masked-auto-encoding'],
}
if training_args.push_to_hub:
trainer.push_to_hub(**SCREAMING_SNAKE_CASE__ )
else:
trainer.create_model_card(**SCREAMING_SNAKE_CASE__ )
def a_ ( SCREAMING_SNAKE_CASE__ : Dict ):
'''simple docstring'''
main()
if __name__ == "__main__":
main()
| 464 |
"""simple docstring"""
import flax.linen as nn
import jax
import jax.numpy as jnp
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self , snake_case__ ) -> Tuple:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase = hidden_states.shape
UpperCAmelCase = jax.image.resize(
snake_case__ , shape=(batch, height * 2, width * 2, channels) , method="""nearest""" , )
UpperCAmelCase = self.conv(snake_case__ )
return hidden_states
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self , snake_case__ ) -> Any:
"""simple docstring"""
UpperCAmelCase = self.conv(snake_case__ )
return hidden_states
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : int = None
_A : float = 0.0
_A : bool = None
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = self.in_channels if self.out_channels is None else self.out_channels
UpperCAmelCase = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
UpperCAmelCase = nn.Dense(snake_case__ , dtype=self.dtype )
UpperCAmelCase = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
UpperCAmelCase = nn.Dropout(self.dropout_prob )
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
UpperCAmelCase = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
UpperCAmelCase = None
if use_nin_shortcut:
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(1, 1) , strides=(1, 1) , padding="""VALID""" , dtype=self.dtype , )
def __call__( self , snake_case__ , snake_case__ , snake_case__=True ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = hidden_states
UpperCAmelCase = self.norma(snake_case__ )
UpperCAmelCase = nn.swish(snake_case__ )
UpperCAmelCase = self.conva(snake_case__ )
UpperCAmelCase = self.time_emb_proj(nn.swish(snake_case__ ) )
UpperCAmelCase = jnp.expand_dims(jnp.expand_dims(snake_case__ , 1 ) , 1 )
UpperCAmelCase = hidden_states + temb
UpperCAmelCase = self.norma(snake_case__ )
UpperCAmelCase = nn.swish(snake_case__ )
UpperCAmelCase = self.dropout(snake_case__ , snake_case__ )
UpperCAmelCase = self.conva(snake_case__ )
if self.conv_shortcut is not None:
UpperCAmelCase = self.conv_shortcut(snake_case__ )
return hidden_states + residual
| 673 | 0 |
"""simple docstring"""
from __future__ import annotations
import copy
import inspect
import json
import math
import os
import tempfile
import unittest
from importlib import import_module
import numpy as np
from transformers import ViTMAEConfig
from transformers.file_utils import cached_property, is_tf_available, is_vision_available
from transformers.testing_utils import require_tf, require_vision, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFViTMAEForPreTraining, TFViTMAEModel
if is_vision_available():
from PIL import Image
from transformers import ViTImageProcessor
class SCREAMING_SNAKE_CASE__ :
def __init__( self : Dict , lowerCAmelCase_ : Dict , lowerCAmelCase_ : Optional[Any]=1_3 , lowerCAmelCase_ : int=3_0 , lowerCAmelCase_ : List[Any]=2 , lowerCAmelCase_ : Union[str, Any]=3 , lowerCAmelCase_ : Optional[Any]=True , lowerCAmelCase_ : List[Any]=True , lowerCAmelCase_ : Dict=3_2 , lowerCAmelCase_ : Optional[int]=2 , lowerCAmelCase_ : Any=4 , lowerCAmelCase_ : Tuple=3_7 , lowerCAmelCase_ : Optional[Any]="gelu" , lowerCAmelCase_ : List[str]=0.1 , lowerCAmelCase_ : Tuple=0.1 , lowerCAmelCase_ : Optional[Any]=1_0 , lowerCAmelCase_ : str=0.02 , lowerCAmelCase_ : List[str]=3 , lowerCAmelCase_ : Tuple=0.6 , lowerCAmelCase_ : List[str]=None , ):
"""simple docstring"""
lowercase_ = parent
lowercase_ = batch_size
lowercase_ = image_size
lowercase_ = patch_size
lowercase_ = num_channels
lowercase_ = is_training
lowercase_ = use_labels
lowercase_ = hidden_size
lowercase_ = num_hidden_layers
lowercase_ = num_attention_heads
lowercase_ = intermediate_size
lowercase_ = hidden_act
lowercase_ = hidden_dropout_prob
lowercase_ = attention_probs_dropout_prob
lowercase_ = type_sequence_label_size
lowercase_ = initializer_range
lowercase_ = mask_ratio
lowercase_ = scope
# in ViTMAE, the expected sequence length = (num_patches + 1) * (1 - config.mask_ratio), rounded above
# (we add 1 for the [CLS] token)
lowercase_ = (image_size // patch_size) ** 2
lowercase_ = int(math.ceil((1 - mask_ratio) * (num_patches + 1)))
def _UpperCAmelCase ( self : Union[str, Any]):
"""simple docstring"""
lowercase_ = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size])
lowercase_ = None
if self.use_labels:
lowercase_ = ids_tensor([self.batch_size] , self.type_sequence_label_size)
lowercase_ = self.get_config()
return config, pixel_values, labels
def _UpperCAmelCase ( self : str):
"""simple docstring"""
return ViTMAEConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , decoder_hidden_size=self.hidden_size , decoder_num_hidden_layers=self.num_hidden_layers , decoder_num_attention_heads=self.num_attention_heads , decoder_intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , is_decoder=snake_case__ , initializer_range=self.initializer_range , mask_ratio=self.mask_ratio , )
def _UpperCAmelCase ( self : Dict , lowerCAmelCase_ : Dict , lowerCAmelCase_ : List[Any] , lowerCAmelCase_ : Tuple):
"""simple docstring"""
lowercase_ = TFViTMAEModel(config=snake_case__)
lowercase_ = model(snake_case__ , training=snake_case__)
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size))
def _UpperCAmelCase ( self : Union[str, Any] , lowerCAmelCase_ : Optional[int] , lowerCAmelCase_ : str , lowerCAmelCase_ : Union[str, Any]):
"""simple docstring"""
lowercase_ = TFViTMAEForPreTraining(snake_case__)
lowercase_ = model(snake_case__ , training=snake_case__)
# expected sequence length = num_patches
lowercase_ = (self.image_size // self.patch_size) ** 2
lowercase_ = self.patch_size**2 * self.num_channels
self.parent.assertEqual(result.logits.shape , (self.batch_size, num_patches, expected_num_channels))
# test greyscale images
lowercase_ = 1
lowercase_ = TFViTMAEForPreTraining(snake_case__)
lowercase_ = floats_tensor([self.batch_size, 1, self.image_size, self.image_size])
lowercase_ = model(snake_case__ , training=snake_case__)
lowercase_ = self.patch_size**2
self.parent.assertEqual(result.logits.shape , (self.batch_size, num_patches, expected_num_channels))
def _UpperCAmelCase ( self : List[Any]):
"""simple docstring"""
lowercase_ = self.prepare_config_and_inputs()
((lowercase_) , (lowercase_) , (lowercase_)) = config_and_inputs
lowercase_ = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_tf
class SCREAMING_SNAKE_CASE__ ( a_ , a_ , unittest.TestCase ):
lowercase__ = (TFViTMAEModel, TFViTMAEForPreTraining) if is_tf_available() else ()
lowercase__ = {'feature-extraction': TFViTMAEModel} if is_tf_available() else {}
lowercase__ = False
lowercase__ = False
lowercase__ = False
lowercase__ = False
def _UpperCAmelCase ( self : Tuple):
"""simple docstring"""
lowercase_ = TFViTMAEModelTester(self)
lowercase_ = ConfigTester(self , config_class=snake_case__ , has_text_modality=snake_case__ , hidden_size=3_7)
def _UpperCAmelCase ( self : Optional[int]):
"""simple docstring"""
self.config_tester.run_common_tests()
@unittest.skip(reason="""ViTMAE does not use inputs_embeds""")
def _UpperCAmelCase ( self : Optional[int]):
"""simple docstring"""
pass
def _UpperCAmelCase ( self : Optional[Any]):
"""simple docstring"""
lowercase_ , lowercase_ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowercase_ = model_class(snake_case__)
self.assertIsInstance(model.get_input_embeddings() , (tf.keras.layers.Layer))
lowercase_ = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , tf.keras.layers.Layer))
def _UpperCAmelCase ( self : str):
"""simple docstring"""
lowercase_ , lowercase_ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowercase_ = model_class(snake_case__)
lowercase_ = inspect.signature(model.call)
# signature.parameters is an OrderedDict => so arg_names order is deterministic
lowercase_ = [*signature.parameters.keys()]
lowercase_ = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] , snake_case__)
def _UpperCAmelCase ( self : int):
"""simple docstring"""
lowercase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__)
def _UpperCAmelCase ( self : Union[str, Any]):
"""simple docstring"""
lowercase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_pretraining(*snake_case__)
def _UpperCAmelCase ( self : List[Any]):
"""simple docstring"""
np.random.seed(2)
lowercase_ , lowercase_ = self.model_tester.prepare_config_and_inputs_for_common()
lowercase_ = int((config.image_size // config.patch_size) ** 2)
lowercase_ = np.random.uniform(size=(self.model_tester.batch_size, num_patches))
for model_class in self.all_model_classes:
lowercase_ = model_class(snake_case__)
lowercase_ = self._prepare_for_class(snake_case__ , snake_case__)
lowercase_ = model(snake_case__ , noise=snake_case__)
lowercase_ = copy.deepcopy(self._prepare_for_class(snake_case__ , snake_case__))
lowercase_ = model(**snake_case__ , noise=snake_case__)
lowercase_ = outputs_dict[0].numpy()
lowercase_ = outputs_keywords[0].numpy()
self.assertLess(np.sum(np.abs(output_dict - output_keywords)) , 1E-6)
def _UpperCAmelCase ( self : Union[str, Any]):
"""simple docstring"""
np.random.seed(2)
lowercase_ , lowercase_ = self.model_tester.prepare_config_and_inputs_for_common()
lowercase_ = int((config.image_size // config.patch_size) ** 2)
lowercase_ = np.random.uniform(size=(self.model_tester.batch_size, num_patches))
def prepare_numpy_arrays(lowerCAmelCase_ : List[str]):
lowercase_ = {}
for k, v in inputs_dict.items():
if tf.is_tensor(snake_case__):
lowercase_ = v.numpy()
else:
lowercase_ = np.array(snake_case__)
return inputs_np_dict
for model_class in self.all_model_classes:
lowercase_ = model_class(snake_case__)
lowercase_ = self._prepare_for_class(snake_case__ , snake_case__)
lowercase_ = prepare_numpy_arrays(snake_case__)
lowercase_ = model(snake_case__ , noise=snake_case__)
lowercase_ = model(**snake_case__ , noise=snake_case__)
self.assert_outputs_same(snake_case__ , snake_case__)
def _UpperCAmelCase ( self : Tuple , lowerCAmelCase_ : List[Any] , lowerCAmelCase_ : Dict , lowerCAmelCase_ : List[Any]):
"""simple docstring"""
np.random.seed(2)
lowercase_ = int((tf_model.config.image_size // tf_model.config.patch_size) ** 2)
lowercase_ = np.random.uniform(size=(self.model_tester.batch_size, num_patches))
lowercase_ = tf.constant(snake_case__)
# Add `noise` argument.
# PT inputs will be prepared in `super().check_pt_tf_models()` with this added `noise` argument
lowercase_ = tf_noise
super().check_pt_tf_models(snake_case__ , snake_case__ , snake_case__)
def _UpperCAmelCase ( self : Optional[int]):
"""simple docstring"""
np.random.seed(2)
lowercase_ , lowercase_ = self.model_tester.prepare_config_and_inputs_for_common()
lowercase_ = {
module_member
for model_class in self.all_model_classes
for module in (import_module(model_class.__module__),)
for module_member_name in dir(snake_case__)
if module_member_name.endswith("""MainLayer""")
# This condition is required, since `modeling_tf_clip.py` has 3 classes whose names end with `MainLayer`.
and module_member_name[: -len("""MainLayer""")] == model_class.__name__[: -len("""Model""")]
for module_member in (getattr(snake_case__ , snake_case__),)
if isinstance(snake_case__ , snake_case__)
and tf.keras.layers.Layer in module_member.__bases__
and getattr(snake_case__ , """_keras_serializable""" , snake_case__)
}
lowercase_ = int((config.image_size // config.patch_size) ** 2)
lowercase_ = np.random.uniform(size=(self.model_tester.batch_size, num_patches))
lowercase_ = tf.convert_to_tensor(snake_case__)
inputs_dict.update({"""noise""": noise})
for main_layer_class in tf_main_layer_classes:
lowercase_ = main_layer_class(snake_case__)
lowercase_ = {
name: tf.keras.Input(tensor.shape[1:] , dtype=tensor.dtype) for name, tensor in inputs_dict.items()
}
lowercase_ = tf.keras.Model(snake_case__ , outputs=main_layer(snake_case__))
lowercase_ = model(snake_case__)
with tempfile.TemporaryDirectory() as tmpdirname:
lowercase_ = os.path.join(snake_case__ , """keras_model.h5""")
model.save(snake_case__)
lowercase_ = tf.keras.models.load_model(
snake_case__ , custom_objects={main_layer_class.__name__: main_layer_class})
assert isinstance(snake_case__ , tf.keras.Model)
lowercase_ = model(snake_case__)
self.assert_outputs_same(snake_case__ , snake_case__)
@slow
def _UpperCAmelCase ( self : Any):
"""simple docstring"""
np.random.seed(2)
lowercase_ , lowercase_ = self.model_tester.prepare_config_and_inputs_for_common()
lowercase_ = int((config.image_size // config.patch_size) ** 2)
lowercase_ = np.random.uniform(size=(self.model_tester.batch_size, num_patches))
for model_class in self.all_model_classes:
lowercase_ = model_class(snake_case__)
lowercase_ = self._prepare_for_class(snake_case__ , snake_case__)
lowercase_ = model(snake_case__ , noise=snake_case__)
if model_class.__name__ == "TFViTMAEModel":
lowercase_ = outputs.last_hidden_state.numpy()
lowercase_ = 0
else:
lowercase_ = outputs.logits.numpy()
lowercase_ = 0
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(snake_case__ , saved_model=snake_case__)
lowercase_ = model_class.from_pretrained(snake_case__)
lowercase_ = model(snake_case__ , noise=snake_case__)
if model_class.__name__ == "TFViTMAEModel":
lowercase_ = after_outputs["""last_hidden_state"""].numpy()
lowercase_ = 0
else:
lowercase_ = after_outputs["""logits"""].numpy()
lowercase_ = 0
lowercase_ = np.amax(np.abs(out_a - out_a))
self.assertLessEqual(snake_case__ , 1E-5)
def _UpperCAmelCase ( self : Optional[Any]):
"""simple docstring"""
np.random.seed(2)
lowercase_ , lowercase_ = self.model_tester.prepare_config_and_inputs_for_common()
lowercase_ = int((config.image_size // config.patch_size) ** 2)
lowercase_ = np.random.uniform(size=(self.model_tester.batch_size, num_patches))
for model_class in self.all_model_classes:
lowercase_ = model_class(snake_case__)
lowercase_ = self._prepare_for_class(snake_case__ , snake_case__)
lowercase_ = model(snake_case__ , noise=snake_case__)
lowercase_ = model.get_config()
# make sure that returned config is jsonifiable, which is required by keras
json.dumps(snake_case__)
lowercase_ = model_class.from_config(model.get_config())
# make sure it also accepts a normal config
lowercase_ = model_class.from_config(model.config)
lowercase_ = new_model(snake_case__) # Build model
new_model.set_weights(model.get_weights())
lowercase_ = new_model(snake_case__ , noise=snake_case__)
self.assert_outputs_same(snake_case__ , snake_case__)
@unittest.skip(
reason="""ViTMAE returns a random mask + ids_restore in each forward pass. See test_save_load
to get deterministic results.""")
def _UpperCAmelCase ( self : int):
"""simple docstring"""
pass
@unittest.skip(reason="""ViTMAE returns a random mask + ids_restore in each forward pass. See test_save_load""")
def _UpperCAmelCase ( self : List[str]):
"""simple docstring"""
pass
@slow
def _UpperCAmelCase ( self : List[Any]):
"""simple docstring"""
lowercase_ = TFViTMAEModel.from_pretrained("""google/vit-base-patch16-224""")
self.assertIsNotNone(snake_case__)
def _SCREAMING_SNAKE_CASE () -> str:
'''simple docstring'''
lowercase_ = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
return image
@require_tf
@require_vision
class SCREAMING_SNAKE_CASE__ ( unittest.TestCase ):
@cached_property
def _UpperCAmelCase ( self : List[Any]):
"""simple docstring"""
return ViTImageProcessor.from_pretrained("""facebook/vit-mae-base""") if is_vision_available() else None
@slow
def _UpperCAmelCase ( self : Optional[Any]):
"""simple docstring"""
np.random.seed(2)
lowercase_ = TFViTMAEForPreTraining.from_pretrained("""facebook/vit-mae-base""")
lowercase_ = self.default_image_processor
lowercase_ = prepare_img()
lowercase_ = image_processor(images=snake_case__ , return_tensors="""tf""")
# prepare a noise vector that will be also used for testing the TF model
# (this way we can ensure that the PT and TF models operate on the same inputs)
lowercase_ = ViTMAEConfig()
lowercase_ = int((vit_mae_config.image_size // vit_mae_config.patch_size) ** 2)
lowercase_ = np.random.uniform(size=(1, num_patches))
# forward pass
lowercase_ = model(**snake_case__ , noise=snake_case__)
# verify the logits
lowercase_ = tf.convert_to_tensor([1, 1_9_6, 7_6_8])
self.assertEqual(outputs.logits.shape , snake_case__)
lowercase_ = tf.convert_to_tensor(
[[-0.0_548, -1.7_023, -0.9_325], [0.3_721, -0.5_670, -0.2_233], [0.8_235, -1.3_878, -0.3_524]])
tf.debugging.assert_near(outputs.logits[0, :3, :3] , snake_case__ , atol=1E-4)
| 567 |
"""simple docstring"""
from __future__ import annotations
import inspect
import unittest
from transformers import ViTConfig
from transformers.testing_utils import require_tf, require_vision, slow
from transformers.utils import cached_property, is_tf_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFViTForImageClassification, TFViTModel
if is_vision_available():
from PIL import Image
from transformers import ViTImageProcessor
class UpperCamelCase_ :
def __init__( self , snake_case__ , snake_case__=13 , snake_case__=30 , snake_case__=2 , snake_case__=3 , snake_case__=True , snake_case__=True , snake_case__=32 , snake_case__=2 , snake_case__=4 , snake_case__=37 , snake_case__="gelu" , snake_case__=0.1 , snake_case__=0.1 , snake_case__=10 , snake_case__=0.02 , snake_case__=3 , snake_case__=None , ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = parent
UpperCAmelCase = batch_size
UpperCAmelCase = image_size
UpperCAmelCase = patch_size
UpperCAmelCase = num_channels
UpperCAmelCase = is_training
UpperCAmelCase = use_labels
UpperCAmelCase = hidden_size
UpperCAmelCase = num_hidden_layers
UpperCAmelCase = num_attention_heads
UpperCAmelCase = intermediate_size
UpperCAmelCase = hidden_act
UpperCAmelCase = hidden_dropout_prob
UpperCAmelCase = attention_probs_dropout_prob
UpperCAmelCase = type_sequence_label_size
UpperCAmelCase = initializer_range
UpperCAmelCase = scope
# in ViT, the seq length equals the number of patches + 1 (we add 1 for the [CLS] token)
UpperCAmelCase = (image_size // patch_size) ** 2
UpperCAmelCase = num_patches + 1
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
UpperCAmelCase = None
if self.use_labels:
UpperCAmelCase = ids_tensor([self.batch_size] , self.type_sequence_label_size )
UpperCAmelCase = self.get_config()
return config, pixel_values, labels
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
return ViTConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , is_decoder=snake_case__ , initializer_range=self.initializer_range , )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ ) -> Dict:
"""simple docstring"""
UpperCAmelCase = TFViTModel(config=snake_case__ )
UpperCAmelCase = model(snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
# Test with an image with different size than the one specified in config.
UpperCAmelCase = self.image_size // 2
UpperCAmelCase = pixel_values[:, :, :image_size, :image_size]
UpperCAmelCase = model(snake_case__ , interpolate_pos_encoding=snake_case__ , training=snake_case__ )
UpperCAmelCase = (image_size // self.patch_size) ** 2 + 1
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, seq_length, self.hidden_size) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = self.type_sequence_label_size
UpperCAmelCase = TFViTForImageClassification(snake_case__ )
UpperCAmelCase = model(snake_case__ , labels=snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# Test with an image with different size than the one specified in config.
UpperCAmelCase = self.image_size // 2
UpperCAmelCase = pixel_values[:, :, :image_size, :image_size]
UpperCAmelCase = model(snake_case__ , interpolate_pos_encoding=snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# test greyscale images
UpperCAmelCase = 1
UpperCAmelCase = TFViTForImageClassification(snake_case__ )
UpperCAmelCase = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
UpperCAmelCase = model(snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
UpperCAmelCase = self.prepare_config_and_inputs()
UpperCAmelCase , UpperCAmelCase , UpperCAmelCase = config_and_inputs
UpperCAmelCase = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_tf
class UpperCamelCase_ ( a_ , a_ , unittest.TestCase ):
_A : Optional[int] = (TFViTModel, TFViTForImageClassification) if is_tf_available() else ()
_A : Optional[Any] = (
{'feature-extraction': TFViTModel, 'image-classification': TFViTForImageClassification}
if is_tf_available()
else {}
)
_A : Optional[int] = False
_A : Any = False
_A : List[str] = False
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = TFViTModelTester(self )
UpperCAmelCase = ConfigTester(self , config_class=snake_case__ , has_text_modality=snake_case__ , hidden_size=37 )
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
self.config_tester.run_common_tests()
@unittest.skip(reason="""ViT does not use inputs_embeds""" )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
pass
@unittest.skip(reason="""ViT does not use inputs_embeds""" )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
pass
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
UpperCAmelCase = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (tf.keras.layers.Layer) )
UpperCAmelCase = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , tf.keras.layers.Layer ) )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
UpperCAmelCase = model_class(snake_case__ )
UpperCAmelCase = inspect.signature(model.call )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
UpperCAmelCase = [*signature.parameters.keys()]
UpperCAmelCase = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] , snake_case__ )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*snake_case__ )
@slow
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = TFViTModel.from_pretrained("""google/vit-base-patch16-224""" )
self.assertIsNotNone(snake_case__ )
def _lowerCAmelCase ( ):
'''simple docstring'''
UpperCAmelCase = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
return image
@require_tf
@require_vision
class UpperCamelCase_ ( unittest.TestCase ):
@cached_property
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return ViTImageProcessor.from_pretrained("""google/vit-base-patch16-224""" ) if is_vision_available() else None
@slow
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = TFViTForImageClassification.from_pretrained("""google/vit-base-patch16-224""" )
UpperCAmelCase = self.default_image_processor
UpperCAmelCase = prepare_img()
UpperCAmelCase = image_processor(images=snake_case__ , return_tensors="""tf""" )
# forward pass
UpperCAmelCase = model(**snake_case__ )
# verify the logits
UpperCAmelCase = tf.TensorShape((1, 10_00) )
self.assertEqual(outputs.logits.shape , snake_case__ )
UpperCAmelCase = tf.constant([-0.2_744, 0.8_215, -0.0_836] )
tf.debugging.assert_near(outputs.logits[0, :3] , snake_case__ , atol=1e-4 )
| 673 | 0 |
'''simple docstring'''
import numpy as np
from nltk.translate import meteor_score
import datasets
from datasets.config import importlib_metadata, version
UpperCamelCase__ : int = version.parse(importlib_metadata.version('''nltk'''))
if NLTK_VERSION >= version.Version('''3.6.4'''):
from nltk import word_tokenize
UpperCamelCase__ : Optional[Any] = '''\
@inproceedings{banarjee2005,
title = {{METEOR}: An Automatic Metric for {MT} Evaluation with Improved Correlation with Human Judgments},
author = {Banerjee, Satanjeev and Lavie, Alon},
booktitle = {Proceedings of the {ACL} Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and/or Summarization},
month = jun,
year = {2005},
address = {Ann Arbor, Michigan},
publisher = {Association for Computational Linguistics},
url = {https://www.aclweb.org/anthology/W05-0909},
pages = {65--72},
}
'''
UpperCamelCase__ : Union[str, Any] = '''\
METEOR, an automatic metric for machine translation evaluation
that is based on a generalized concept of unigram matching between the
machine-produced translation and human-produced reference translations.
Unigrams can be matched based on their surface forms, stemmed forms,
and meanings; furthermore, METEOR can be easily extended to include more
advanced matching strategies. Once all generalized unigram matches
between the two strings have been found, METEOR computes a score for
this matching using a combination of unigram-precision, unigram-recall, and
a measure of fragmentation that is designed to directly capture how
well-ordered the matched words in the machine translation are in relation
to the reference.
METEOR gets an R correlation value of 0.347 with human evaluation on the Arabic
data and 0.331 on the Chinese data. This is shown to be an improvement on
using simply unigram-precision, unigram-recall and their harmonic F1
combination.
'''
UpperCamelCase__ : Tuple = '''
Computes METEOR score of translated segments against one or more references.
Args:
predictions: list of predictions to score. Each prediction
should be a string with tokens separated by spaces.
references: list of reference for each prediction. Each
reference should be a string with tokens separated by spaces.
alpha: Parameter for controlling relative weights of precision and recall. default: 0.9
beta: Parameter for controlling shape of penalty as a function of fragmentation. default: 3
gamma: Relative weight assigned to fragmentation penalty. default: 0.5
Returns:
\'meteor\': meteor score.
Examples:
>>> meteor = datasets.load_metric(\'meteor\')
>>> predictions = ["It is a guide to action which ensures that the military always obeys the commands of the party"]
>>> references = ["It is a guide to action that ensures that the military will forever heed Party commands"]
>>> results = meteor.compute(predictions=predictions, references=references)
>>> print(round(results["meteor"], 4))
0.6944
'''
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class _UpperCamelCase ( datasets.Metric ):
'''simple docstring'''
def UpperCamelCase__ ( self : str ):
"""simple docstring"""
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features(
{
"""predictions""": datasets.Value("""string""" , id="""sequence""" ),
"""references""": datasets.Value("""string""" , id="""sequence""" ),
} ) , codebase_urls=["""https://github.com/nltk/nltk/blob/develop/nltk/translate/meteor_score.py"""] , reference_urls=[
"""https://www.nltk.org/api/nltk.translate.html#module-nltk.translate.meteor_score""",
"""https://en.wikipedia.org/wiki/METEOR""",
] , )
def UpperCamelCase__ ( self : List[str] , lowerCAmelCase__ : List[str] ):
"""simple docstring"""
import nltk
nltk.download("""wordnet""" )
if NLTK_VERSION >= version.Version("""3.6.5""" ):
nltk.download("""punkt""" )
if NLTK_VERSION >= version.Version("""3.6.6""" ):
nltk.download("""omw-1.4""" )
def UpperCamelCase__ ( self : str , lowerCAmelCase__ : Tuple , lowerCAmelCase__ : List[str] , lowerCAmelCase__ : Dict=0.9 , lowerCAmelCase__ : int=3 , lowerCAmelCase__ : Union[str, Any]=0.5 ):
"""simple docstring"""
if NLTK_VERSION >= version.Version("""3.6.5""" ):
__SCREAMING_SNAKE_CASE : List[Any] = [
meteor_score.single_meteor_score(
word_tokenize(snake_case__ ) , word_tokenize(snake_case__ ) , alpha=snake_case__ , beta=snake_case__ , gamma=snake_case__ )
for ref, pred in zip(snake_case__ , snake_case__ )
]
else:
__SCREAMING_SNAKE_CASE : Optional[int] = [
meteor_score.single_meteor_score(snake_case__ , snake_case__ , alpha=snake_case__ , beta=snake_case__ , gamma=snake_case__ )
for ref, pred in zip(snake_case__ , snake_case__ )
]
return {"meteor": np.mean(snake_case__ )} | 578 |
"""simple docstring"""
import unittest
from transformers import AutoTokenizer, NystromformerConfig, is_torch_available
from transformers.testing_utils import require_torch, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
NystromformerForMaskedLM,
NystromformerForMultipleChoice,
NystromformerForQuestionAnswering,
NystromformerForSequenceClassification,
NystromformerForTokenClassification,
NystromformerModel,
)
from transformers.models.nystromformer.modeling_nystromformer import NYSTROMFORMER_PRETRAINED_MODEL_ARCHIVE_LIST
class UpperCamelCase_ :
def __init__( self , snake_case__ , snake_case__=13 , snake_case__=7 , snake_case__=True , snake_case__=True , snake_case__=True , snake_case__=True , snake_case__=99 , snake_case__=32 , snake_case__=5 , snake_case__=4 , snake_case__=37 , snake_case__="gelu" , snake_case__=0.1 , snake_case__=0.1 , snake_case__=5_12 , snake_case__=16 , snake_case__=2 , snake_case__=0.02 , snake_case__=3 , snake_case__=4 , snake_case__=None , ) -> int:
"""simple docstring"""
UpperCAmelCase = parent
UpperCAmelCase = batch_size
UpperCAmelCase = seq_length
UpperCAmelCase = is_training
UpperCAmelCase = use_input_mask
UpperCAmelCase = use_token_type_ids
UpperCAmelCase = use_labels
UpperCAmelCase = vocab_size
UpperCAmelCase = hidden_size
UpperCAmelCase = num_hidden_layers
UpperCAmelCase = num_attention_heads
UpperCAmelCase = intermediate_size
UpperCAmelCase = hidden_act
UpperCAmelCase = hidden_dropout_prob
UpperCAmelCase = attention_probs_dropout_prob
UpperCAmelCase = max_position_embeddings
UpperCAmelCase = type_vocab_size
UpperCAmelCase = type_sequence_label_size
UpperCAmelCase = initializer_range
UpperCAmelCase = num_labels
UpperCAmelCase = num_choices
UpperCAmelCase = scope
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
UpperCAmelCase = None
if self.use_input_mask:
UpperCAmelCase = random_attention_mask([self.batch_size, self.seq_length] )
UpperCAmelCase = None
if self.use_token_type_ids:
UpperCAmelCase = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
UpperCAmelCase = None
UpperCAmelCase = None
UpperCAmelCase = None
if self.use_labels:
UpperCAmelCase = ids_tensor([self.batch_size] , self.type_sequence_label_size )
UpperCAmelCase = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
UpperCAmelCase = ids_tensor([self.batch_size] , self.num_choices )
UpperCAmelCase = self.get_config()
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
return NystromformerConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=snake_case__ , initializer_range=self.initializer_range , )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = NystromformerModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ )
UpperCAmelCase = model(snake_case__ , token_type_ids=snake_case__ )
UpperCAmelCase = model(snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> int:
"""simple docstring"""
UpperCAmelCase = NystromformerForMaskedLM(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> Dict:
"""simple docstring"""
UpperCAmelCase = NystromformerForQuestionAnswering(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , start_positions=snake_case__ , end_positions=snake_case__ , )
self.parent.assertEqual(result.start_logits.shape , (self.batch_size, self.seq_length) )
self.parent.assertEqual(result.end_logits.shape , (self.batch_size, self.seq_length) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = self.num_labels
UpperCAmelCase = NystromformerForSequenceClassification(snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> int:
"""simple docstring"""
UpperCAmelCase = self.num_labels
UpperCAmelCase = NystromformerForTokenClassification(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = model(snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ ) -> Dict:
"""simple docstring"""
UpperCAmelCase = self.num_choices
UpperCAmelCase = NystromformerForMultipleChoice(config=snake_case__ )
model.to(snake_case__ )
model.eval()
UpperCAmelCase = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
UpperCAmelCase = token_type_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
UpperCAmelCase = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
UpperCAmelCase = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = self.prepare_config_and_inputs()
(
(
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) , (
UpperCAmelCase
) ,
) = config_and_inputs
UpperCAmelCase = {"""input_ids""": input_ids, """token_type_ids""": token_type_ids, """attention_mask""": input_mask}
return config, inputs_dict
@require_torch
class UpperCamelCase_ ( a_ , a_ , unittest.TestCase ):
_A : Optional[Any] = (
(
NystromformerModel,
NystromformerForMaskedLM,
NystromformerForMultipleChoice,
NystromformerForQuestionAnswering,
NystromformerForSequenceClassification,
NystromformerForTokenClassification,
)
if is_torch_available()
else ()
)
_A : Optional[Any] = (
{
'feature-extraction': NystromformerModel,
'fill-mask': NystromformerForMaskedLM,
'question-answering': NystromformerForQuestionAnswering,
'text-classification': NystromformerForSequenceClassification,
'token-classification': NystromformerForTokenClassification,
'zero-shot': NystromformerForSequenceClassification,
}
if is_torch_available()
else {}
)
_A : int = False
_A : Dict = False
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = NystromformerModelTester(self )
UpperCAmelCase = ConfigTester(self , config_class=snake_case__ , hidden_size=37 )
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
self.config_tester.run_common_tests()
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
for type in ["absolute", "relative_key", "relative_key_query"]:
UpperCAmelCase = type
self.model_tester.create_and_check_model(*snake_case__ )
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_lm(*snake_case__ )
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_multiple_choice(*snake_case__ )
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_question_answering(*snake_case__ )
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_sequence_classification(*snake_case__ )
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_token_classification(*snake_case__ )
@slow
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
for model_name in NYSTROMFORMER_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
UpperCAmelCase = NystromformerModel.from_pretrained(snake_case__ )
self.assertIsNotNone(snake_case__ )
@require_torch
class UpperCamelCase_ ( unittest.TestCase ):
@slow
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = NystromformerModel.from_pretrained("""uw-madison/nystromformer-512""" )
UpperCAmelCase = torch.tensor([[0, 1, 2, 3, 4, 5]] )
with torch.no_grad():
UpperCAmelCase = model(snake_case__ )[0]
UpperCAmelCase = torch.Size((1, 6, 7_68) )
self.assertEqual(output.shape , snake_case__ )
UpperCAmelCase = torch.tensor(
[[[-0.4_532, -0.0_936, 0.5_137], [-0.2_676, 0.0_628, 0.6_186], [-0.3_629, -0.1_726, 0.4_716]]] )
self.assertTrue(torch.allclose(output[:, :3, :3] , snake_case__ , atol=1e-4 ) )
@slow
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = """the [MASK] of Belgium is Brussels"""
UpperCAmelCase = AutoTokenizer.from_pretrained("""uw-madison/nystromformer-512""" )
UpperCAmelCase = NystromformerForMaskedLM.from_pretrained("""uw-madison/nystromformer-512""" )
UpperCAmelCase = tokenizer(snake_case__ , return_tensors="""pt""" )
with torch.no_grad():
UpperCAmelCase = model(encoding.input_ids ).logits
UpperCAmelCase = token_logits[:, 2, :].argmax(-1 )[0]
self.assertEqual(tokenizer.decode(snake_case__ ) , """capital""" )
| 673 | 0 |
"""simple docstring"""
from typing import Dict, List, Optional, Tuple, Union
import torch
from ...models import AutoencoderKL, TransformeraDModel
from ...schedulers import KarrasDiffusionSchedulers
from ...utils import randn_tensor
from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
class _UpperCAmelCase ( a_):
def __init__( self : Optional[int] , lowercase_ : Tuple , lowercase_ : List[Any] , lowercase_ : int , lowercase_ : Union[str, Any] = None , ):
super().__init__()
self.register_modules(transformer=snake_case__ , vae=snake_case__ , scheduler=snake_case__ )
# create a imagenet -> id dictionary for easier use
snake_case_ : Tuple = {}
if idalabel is not None:
for key, value in idalabel.items():
for label in value.split(''',''' ):
snake_case_ : List[Any] = int(snake_case__ )
snake_case_ : Dict = dict(sorted(self.labels.items() ) )
def _snake_case ( self : Tuple , lowercase_ : List[str] ):
if not isinstance(snake_case__ , snake_case__ ):
snake_case_ : Dict = list(snake_case__ )
for l in label:
if l not in self.labels:
raise ValueError(
f"{l} does not exist. Please make sure to select one of the following labels: \n {self.labels}." )
return [self.labels[l] for l in label]
@torch.no_grad()
def __call__( self : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[Any] = 4.0 , lowercase_ : Any = None , lowercase_ : Tuple = 50 , lowercase_ : List[str] = "pil" , lowercase_ : Dict = True , ):
snake_case_ : int = len(snake_case__ )
snake_case_ : int = self.transformer.config.sample_size
snake_case_ : Union[str, Any] = self.transformer.config.in_channels
snake_case_ : Dict = randn_tensor(
shape=(batch_size, latent_channels, latent_size, latent_size) , generator=snake_case__ , device=self.device , dtype=self.transformer.dtype , )
snake_case_ : Any = torch.cat([latents] * 2 ) if guidance_scale > 1 else latents
snake_case_ : Any = torch.tensor(snake_case__ , device=self.device ).reshape(-1 )
snake_case_ : int = torch.tensor([1000] * batch_size , device=self.device )
snake_case_ : Optional[Any] = torch.cat([class_labels, class_null] , 0 ) if guidance_scale > 1 else class_labels
# set step values
self.scheduler.set_timesteps(snake_case__ )
for t in self.progress_bar(self.scheduler.timesteps ):
if guidance_scale > 1:
snake_case_ : List[Any] = latent_model_input[: len(snake_case__ ) // 2]
snake_case_ : Optional[Any] = torch.cat([half, half] , dim=0 )
snake_case_ : str = self.scheduler.scale_model_input(snake_case__ , snake_case__ )
snake_case_ : Tuple = t
if not torch.is_tensor(snake_case__ ):
# TODO: this requires sync between CPU and GPU. So try to pass timesteps as tensors if you can
# This would be a good case for the `match` statement (Python 3.10+)
snake_case_ : Optional[int] = latent_model_input.device.type == '''mps'''
if isinstance(snake_case__ , snake_case__ ):
snake_case_ : str = torch.floataa if is_mps else torch.floataa
else:
snake_case_ : List[str] = torch.intaa if is_mps else torch.intaa
snake_case_ : Optional[Any] = torch.tensor([timesteps] , dtype=snake_case__ , device=latent_model_input.device )
elif len(timesteps.shape ) == 0:
snake_case_ : str = timesteps[None].to(latent_model_input.device )
# broadcast to batch dimension in a way that's compatible with ONNX/Core ML
snake_case_ : Optional[int] = timesteps.expand(latent_model_input.shape[0] )
# predict noise model_output
snake_case_ : str = self.transformer(
snake_case__ , timestep=snake_case__ , class_labels=snake_case__ ).sample
# perform guidance
if guidance_scale > 1:
snake_case_, snake_case_ : Any = noise_pred[:, :latent_channels], noise_pred[:, latent_channels:]
snake_case_, snake_case_ : List[Any] = torch.split(snake_case__ , len(snake_case__ ) // 2 , dim=0 )
snake_case_ : Tuple = uncond_eps + guidance_scale * (cond_eps - uncond_eps)
snake_case_ : str = torch.cat([half_eps, half_eps] , dim=0 )
snake_case_ : Any = torch.cat([eps, rest] , dim=1 )
# learned sigma
if self.transformer.config.out_channels // 2 == latent_channels:
snake_case_, snake_case_ : List[str] = torch.split(snake_case__ , snake_case__ , dim=1 )
else:
snake_case_ : List[Any] = noise_pred
# compute previous image: x_t -> x_t-1
snake_case_ : List[str] = self.scheduler.step(snake_case__ , snake_case__ , snake_case__ ).prev_sample
if guidance_scale > 1:
snake_case_, snake_case_ : Tuple = latent_model_input.chunk(2 , dim=0 )
else:
snake_case_ : Optional[Any] = latent_model_input
snake_case_ : Tuple = 1 / self.vae.config.scaling_factor * latents
snake_case_ : List[str] = self.vae.decode(snake_case__ ).sample
snake_case_ : Tuple = (samples / 2 + 0.5).clamp(0 , 1 )
# we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16
snake_case_ : Tuple = samples.cpu().permute(0 , 2 , 3 , 1 ).float().numpy()
if output_type == "pil":
snake_case_ : Any = self.numpy_to_pil(snake_case__ )
if not return_dict:
return (samples,)
return ImagePipelineOutput(images=snake_case__ )
| 123 |
"""simple docstring"""
import os
from argparse import ArgumentParser, Namespace
from ..data import SingleSentenceClassificationProcessor as Processor
from ..pipelines import TextClassificationPipeline
from ..utils import is_tf_available, is_torch_available, logging
from . import BaseTransformersCLICommand
if not is_tf_available() and not is_torch_available():
raise RuntimeError('''At least one of PyTorch or TensorFlow 2.0+ should be installed to use CLI training''')
# TF training parameters
lowerCAmelCase_ : Optional[int] = False
lowerCAmelCase_ : Optional[int] = False
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
return TrainCommand(lowerCAmelCase )
class UpperCamelCase_ ( a_ ):
@staticmethod
def UpperCamelCase_ ( snake_case__ ) -> int:
"""simple docstring"""
UpperCAmelCase = parser.add_parser("""train""" , help="""CLI tool to train a model on a task.""" )
train_parser.add_argument(
"""--train_data""" , type=snake_case__ , required=snake_case__ , help="""path to train (and optionally evaluation) dataset as a csv with tab separated labels and sentences.""" , )
train_parser.add_argument(
"""--column_label""" , type=snake_case__ , default=0 , help="""Column of the dataset csv file with example labels.""" )
train_parser.add_argument(
"""--column_text""" , type=snake_case__ , default=1 , help="""Column of the dataset csv file with example texts.""" )
train_parser.add_argument(
"""--column_id""" , type=snake_case__ , default=2 , help="""Column of the dataset csv file with example ids.""" )
train_parser.add_argument(
"""--skip_first_row""" , action="""store_true""" , help="""Skip the first row of the csv file (headers).""" )
train_parser.add_argument("""--validation_data""" , type=snake_case__ , default="""""" , help="""path to validation dataset.""" )
train_parser.add_argument(
"""--validation_split""" , type=snake_case__ , default=0.1 , help="""if validation dataset is not provided, fraction of train dataset to use as validation dataset.""" , )
train_parser.add_argument("""--output""" , type=snake_case__ , default="""./""" , help="""path to saved the trained model.""" )
train_parser.add_argument(
"""--task""" , type=snake_case__ , default="""text_classification""" , help="""Task to train the model on.""" )
train_parser.add_argument(
"""--model""" , type=snake_case__ , default="""bert-base-uncased""" , help="""Model's name or path to stored model.""" )
train_parser.add_argument("""--train_batch_size""" , type=snake_case__ , default=32 , help="""Batch size for training.""" )
train_parser.add_argument("""--valid_batch_size""" , type=snake_case__ , default=64 , help="""Batch size for validation.""" )
train_parser.add_argument("""--learning_rate""" , type=snake_case__ , default=3e-5 , help="""Learning rate.""" )
train_parser.add_argument("""--adam_epsilon""" , type=snake_case__ , default=1e-08 , help="""Epsilon for Adam optimizer.""" )
train_parser.set_defaults(func=snake_case__ )
def __init__( self , snake_case__ ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = logging.get_logger("""transformers-cli/training""" )
UpperCAmelCase = """tf""" if is_tf_available() else """torch"""
os.makedirs(args.output , exist_ok=snake_case__ )
UpperCAmelCase = args.output
UpperCAmelCase = args.column_label
UpperCAmelCase = args.column_text
UpperCAmelCase = args.column_id
self.logger.info(f'''Loading {args.task} pipeline for {args.model}''' )
if args.task == "text_classification":
UpperCAmelCase = TextClassificationPipeline.from_pretrained(args.model )
elif args.task == "token_classification":
raise NotImplementedError
elif args.task == "question_answering":
raise NotImplementedError
self.logger.info(f'''Loading dataset from {args.train_data}''' )
UpperCAmelCase = Processor.create_from_csv(
args.train_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase = None
if args.validation_data:
self.logger.info(f'''Loading validation dataset from {args.validation_data}''' )
UpperCAmelCase = Processor.create_from_csv(
args.validation_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase = args.validation_split
UpperCAmelCase = args.train_batch_size
UpperCAmelCase = args.valid_batch_size
UpperCAmelCase = args.learning_rate
UpperCAmelCase = args.adam_epsilon
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
if self.framework == "tf":
return self.run_tf()
return self.run_torch()
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
raise NotImplementedError
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
self.pipeline.fit(
self.train_dataset , validation_data=self.valid_dataset , validation_split=self.validation_split , learning_rate=self.learning_rate , adam_epsilon=self.adam_epsilon , train_batch_size=self.train_batch_size , valid_batch_size=self.valid_batch_size , )
# Save trained pipeline
self.pipeline.save_pretrained(self.output )
| 673 | 0 |
'''simple docstring'''
import unittest
from transformers import load_tool
from .test_tools_common import ToolTesterMixin
class __lowerCamelCase ( unittest.TestCase , a_ ):
'''simple docstring'''
def a_ ( self ):
__SCREAMING_SNAKE_CASE : str = load_tool("text-classification" )
self.tool.setup()
__SCREAMING_SNAKE_CASE : str = load_tool("text-classification" , remote=snake_case__ )
def a_ ( self ):
__SCREAMING_SNAKE_CASE : int = self.tool("That's quite cool" , ["positive", "negative"] )
self.assertEqual(snake_case__ , "positive" )
def a_ ( self ):
__SCREAMING_SNAKE_CASE : Dict = self.remote_tool("That's quite cool" , ["positive", "negative"] )
self.assertEqual(snake_case__ , "positive" )
def a_ ( self ):
__SCREAMING_SNAKE_CASE : Any = self.tool(text="That's quite cool" , labels=["positive", "negative"] )
self.assertEqual(snake_case__ , "positive" )
def a_ ( self ):
__SCREAMING_SNAKE_CASE : Any = self.remote_tool(text="That's quite cool" , labels=["positive", "negative"] )
self.assertEqual(snake_case__ , "positive" )
| 211 |
"""simple docstring"""
import sys
from typing import Tuple
import numpy as np
import torch
from PIL import Image
from torch import nn
from transformers.image_utils import PILImageResampling
from utils import img_tensorize
class UpperCamelCase_ :
def __init__( self , snake_case__ , snake_case__=sys.maxsize ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = """bilinear"""
UpperCAmelCase = max_size
UpperCAmelCase = short_edge_length
def __call__( self , snake_case__ ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = []
for img in imgs:
UpperCAmelCase , UpperCAmelCase = img.shape[:2]
# later: provide list and randomly choose index for resize
UpperCAmelCase = np.random.randint(self.short_edge_length[0] , self.short_edge_length[1] + 1 )
if size == 0:
return img
UpperCAmelCase = size * 1.0 / min(snake_case__ , snake_case__ )
if h < w:
UpperCAmelCase , UpperCAmelCase = size, scale * w
else:
UpperCAmelCase , UpperCAmelCase = scale * h, size
if max(snake_case__ , snake_case__ ) > self.max_size:
UpperCAmelCase = self.max_size * 1.0 / max(snake_case__ , snake_case__ )
UpperCAmelCase = newh * scale
UpperCAmelCase = neww * scale
UpperCAmelCase = int(neww + 0.5 )
UpperCAmelCase = int(newh + 0.5 )
if img.dtype == np.uinta:
UpperCAmelCase = Image.fromarray(snake_case__ )
UpperCAmelCase = pil_image.resize((neww, newh) , PILImageResampling.BILINEAR )
UpperCAmelCase = np.asarray(snake_case__ )
else:
UpperCAmelCase = img.permute(2 , 0 , 1 ).unsqueeze(0 ) # 3, 0, 1) # hw(c) -> nchw
UpperCAmelCase = nn.functional.interpolate(
snake_case__ , (newh, neww) , mode=self.interp_method , align_corners=snake_case__ ).squeeze(0 )
img_augs.append(snake_case__ )
return img_augs
class UpperCamelCase_ :
def __init__( self , snake_case__ ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = ResizeShortestEdge([cfg.INPUT.MIN_SIZE_TEST, cfg.INPUT.MIN_SIZE_TEST] , cfg.INPUT.MAX_SIZE_TEST )
UpperCAmelCase = cfg.INPUT.FORMAT
UpperCAmelCase = cfg.SIZE_DIVISIBILITY
UpperCAmelCase = cfg.PAD_VALUE
UpperCAmelCase = cfg.INPUT.MAX_SIZE_TEST
UpperCAmelCase = cfg.MODEL.DEVICE
UpperCAmelCase = torch.tensor(cfg.MODEL.PIXEL_STD ).to(self.device ).view(len(cfg.MODEL.PIXEL_STD ) , 1 , 1 )
UpperCAmelCase = torch.tensor(cfg.MODEL.PIXEL_MEAN ).to(self.device ).view(len(cfg.MODEL.PIXEL_STD ) , 1 , 1 )
UpperCAmelCase = lambda snake_case__ : (x - self.pixel_mean) / self.pixel_std
def UpperCamelCase_ ( self , snake_case__ ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase = tuple(max(snake_case__ ) for s in zip(*[img.shape for img in images] ) )
UpperCAmelCase = [im.shape[-2:] for im in images]
UpperCAmelCase = [
nn.functional.pad(
snake_case__ , [0, max_size[-1] - size[1], 0, max_size[-2] - size[0]] , value=self.pad_value , )
for size, im in zip(snake_case__ , snake_case__ )
]
return torch.stack(snake_case__ ), torch.tensor(snake_case__ )
def __call__( self , snake_case__ , snake_case__=False ) -> Optional[Any]:
"""simple docstring"""
with torch.no_grad():
if not isinstance(snake_case__ , snake_case__ ):
UpperCAmelCase = [images]
if single_image:
assert len(snake_case__ ) == 1
for i in range(len(snake_case__ ) ):
if isinstance(images[i] , torch.Tensor ):
images.insert(snake_case__ , images.pop(snake_case__ ).to(self.device ).float() )
elif not isinstance(images[i] , torch.Tensor ):
images.insert(
snake_case__ , torch.as_tensor(img_tensorize(images.pop(snake_case__ ) , input_format=self.input_format ) )
.to(self.device )
.float() , )
# resize smallest edge
UpperCAmelCase = torch.tensor([im.shape[:2] for im in images] )
UpperCAmelCase = self.aug(snake_case__ )
# transpose images and convert to torch tensors
# images = [torch.as_tensor(i.astype("float32")).permute(2, 0, 1).to(self.device) for i in images]
# now normalize before pad to avoid useless arithmetic
UpperCAmelCase = [self.normalizer(snake_case__ ) for x in images]
# now pad them to do the following operations
UpperCAmelCase , UpperCAmelCase = self.pad(snake_case__ )
# Normalize
if self.size_divisibility > 0:
raise NotImplementedError()
# pad
UpperCAmelCase = torch.true_divide(snake_case__ , snake_case__ )
if single_image:
return images[0], sizes[0], scales_yx[0]
else:
return images, sizes, scales_yx
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase ):
'''simple docstring'''
boxes[:, 0::2] *= scale_yx[:, 1]
boxes[:, 1::2] *= scale_yx[:, 0]
return boxes
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase ):
'''simple docstring'''
assert torch.isfinite(lowerCAmelCase ).all(), "Box tensor contains infinite or NaN!"
UpperCAmelCase , UpperCAmelCase = box_size
tensor[:, 0].clamp_(min=0 , max=lowerCAmelCase )
tensor[:, 1].clamp_(min=0 , max=lowerCAmelCase )
tensor[:, 2].clamp_(min=0 , max=lowerCAmelCase )
tensor[:, 3].clamp_(min=0 , max=lowerCAmelCase )
| 673 | 0 |
'''simple docstring'''
import unittest
import numpy as np
from diffusers import LMSDiscreteScheduler, OnnxStableDiffusionInpaintPipeline
from diffusers.utils.testing_utils import (
is_onnx_available,
load_image,
nightly,
require_onnxruntime,
require_torch_gpu,
)
from ..test_pipelines_onnx_common import OnnxPipelineTesterMixin
if is_onnx_available():
import onnxruntime as ort
class UpperCAmelCase ( a_ , unittest.TestCase ):
'''simple docstring'''
pass
@nightly
@require_onnxruntime
@require_torch_gpu
class UpperCAmelCase ( unittest.TestCase ):
'''simple docstring'''
@property
def _lowerCAmelCase( self ) -> Dict:
return (
"CUDAExecutionProvider",
{
"gpu_mem_limit": "15000000000", # 15GB
"arena_extend_strategy": "kSameAsRequested",
},
)
@property
def _lowerCAmelCase( self ) -> Tuple:
lowercase__ : Optional[int] = ort.SessionOptions()
lowercase__ : List[Any] = False
return options
def _lowerCAmelCase( self ) -> str:
lowercase__ : Optional[Any] = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/in_paint/overture-creations-5sI6fQgYIuo.png''' )
lowercase__ : List[str] = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/in_paint/overture-creations-5sI6fQgYIuo_mask.png''' )
lowercase__ : List[Any] = OnnxStableDiffusionInpaintPipeline.from_pretrained(
'''runwayml/stable-diffusion-inpainting''' , revision='''onnx''' , safety_checker=snake_case__ , feature_extractor=snake_case__ , provider=self.gpu_provider , sess_options=self.gpu_options , )
pipe.set_progress_bar_config(disable=snake_case__ )
lowercase__ : List[Any] = '''A red cat sitting on a park bench'''
lowercase__ : Dict = np.random.RandomState(0 )
lowercase__ : Optional[Any] = pipe(
prompt=snake_case__ , image=snake_case__ , mask_image=snake_case__ , guidance_scale=7.5 , num_inference_steps=10 , generator=snake_case__ , output_type='''np''' , )
lowercase__ : List[Any] = output.images
lowercase__ : Optional[int] = images[0, 255:258, 255:258, -1]
assert images.shape == (1, 512, 512, 3)
lowercase__ : List[str] = np.array([0.2_5_1_4, 0.3_0_0_7, 0.3_5_1_7, 0.1_7_9_0, 0.2_3_8_2, 0.3_1_6_7, 0.1_9_4_4, 0.2_2_7_3, 0.2_4_6_4] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-3
def _lowerCAmelCase( self ) -> Dict:
lowercase__ : Dict = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/in_paint/overture-creations-5sI6fQgYIuo.png''' )
lowercase__ : str = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/in_paint/overture-creations-5sI6fQgYIuo_mask.png''' )
lowercase__ : int = LMSDiscreteScheduler.from_pretrained(
'''runwayml/stable-diffusion-inpainting''' , subfolder='''scheduler''' , revision='''onnx''' )
lowercase__ : str = OnnxStableDiffusionInpaintPipeline.from_pretrained(
'''runwayml/stable-diffusion-inpainting''' , revision='''onnx''' , scheduler=snake_case__ , safety_checker=snake_case__ , feature_extractor=snake_case__ , provider=self.gpu_provider , sess_options=self.gpu_options , )
pipe.set_progress_bar_config(disable=snake_case__ )
lowercase__ : Optional[int] = '''A red cat sitting on a park bench'''
lowercase__ : Dict = np.random.RandomState(0 )
lowercase__ : Optional[int] = pipe(
prompt=snake_case__ , image=snake_case__ , mask_image=snake_case__ , guidance_scale=7.5 , num_inference_steps=20 , generator=snake_case__ , output_type='''np''' , )
lowercase__ : str = output.images
lowercase__ : Optional[int] = images[0, 255:258, 255:258, -1]
assert images.shape == (1, 512, 512, 3)
lowercase__ : Optional[Any] = np.array([0.0_0_8_6, 0.0_0_7_7, 0.0_0_8_3, 0.0_0_9_3, 0.0_1_0_7, 0.0_1_3_9, 0.0_0_9_4, 0.0_0_9_7, 0.0_1_2_5] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-3
| 152 |
"""simple docstring"""
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import DeiTConfig, DeiTForImageClassificationWithTeacher, DeiTImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
lowerCAmelCase_ : List[str] = logging.get_logger(__name__)
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase=False ):
'''simple docstring'''
UpperCAmelCase = []
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F'''blocks.{i}.norm1.weight''', F'''deit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((F'''blocks.{i}.norm1.bias''', F'''deit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((F'''blocks.{i}.attn.proj.weight''', F'''deit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.attn.proj.bias''', F'''deit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((F'''blocks.{i}.norm2.weight''', F'''deit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((F'''blocks.{i}.norm2.bias''', F'''deit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.weight''', F'''deit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.bias''', F'''deit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.weight''', F'''deit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.bias''', F'''deit.encoder.layer.{i}.output.dense.bias''') )
# projection layer + position embeddings
rename_keys.extend(
[
("""cls_token""", """deit.embeddings.cls_token"""),
("""dist_token""", """deit.embeddings.distillation_token"""),
("""patch_embed.proj.weight""", """deit.embeddings.patch_embeddings.projection.weight"""),
("""patch_embed.proj.bias""", """deit.embeddings.patch_embeddings.projection.bias"""),
("""pos_embed""", """deit.embeddings.position_embeddings"""),
] )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
("""norm.weight""", """layernorm.weight"""),
("""norm.bias""", """layernorm.bias"""),
("""pre_logits.fc.weight""", """pooler.dense.weight"""),
("""pre_logits.fc.bias""", """pooler.dense.bias"""),
] )
# if just the base model, we should remove "deit" from all keys that start with "deit"
UpperCAmelCase = [(pair[0], pair[1][4:]) if pair[1].startswith("""deit""" ) else pair for pair in rename_keys]
else:
# layernorm + classification heads
rename_keys.extend(
[
("""norm.weight""", """deit.layernorm.weight"""),
("""norm.bias""", """deit.layernorm.bias"""),
("""head.weight""", """cls_classifier.weight"""),
("""head.bias""", """cls_classifier.bias"""),
("""head_dist.weight""", """distillation_classifier.weight"""),
("""head_dist.bias""", """distillation_classifier.bias"""),
] )
return rename_keys
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase , lowerCAmelCase=False ):
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
UpperCAmelCase = """"""
else:
UpperCAmelCase = """deit."""
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
UpperCAmelCase = state_dict.pop(F'''blocks.{i}.attn.qkv.weight''' )
UpperCAmelCase = state_dict.pop(F'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
UpperCAmelCase = in_proj_weight[
: config.hidden_size, :
]
UpperCAmelCase = in_proj_bias[: config.hidden_size]
UpperCAmelCase = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
UpperCAmelCase = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
UpperCAmelCase = in_proj_weight[
-config.hidden_size :, :
]
UpperCAmelCase = in_proj_bias[-config.hidden_size :]
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase , lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = dct.pop(lowerCAmelCase )
UpperCAmelCase = val
def _lowerCAmelCase ( ):
'''simple docstring'''
UpperCAmelCase = """http://images.cocodataset.org/val2017/000000039769.jpg"""
UpperCAmelCase = Image.open(requests.get(lowerCAmelCase , stream=lowerCAmelCase ).raw )
return im
@torch.no_grad()
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = DeiTConfig()
# all deit models have fine-tuned heads
UpperCAmelCase = False
# dataset (fine-tuned on ImageNet 2012), patch_size and image_size
UpperCAmelCase = 1000
UpperCAmelCase = """huggingface/label-files"""
UpperCAmelCase = """imagenet-1k-id2label.json"""
UpperCAmelCase = json.load(open(hf_hub_download(lowerCAmelCase , lowerCAmelCase , repo_type="""dataset""" ) , """r""" ) )
UpperCAmelCase = {int(lowerCAmelCase ): v for k, v in idalabel.items()}
UpperCAmelCase = idalabel
UpperCAmelCase = {v: k for k, v in idalabel.items()}
UpperCAmelCase = int(deit_name[-6:-4] )
UpperCAmelCase = int(deit_name[-3:] )
# size of the architecture
if deit_name[9:].startswith("""tiny""" ):
UpperCAmelCase = 192
UpperCAmelCase = 768
UpperCAmelCase = 12
UpperCAmelCase = 3
elif deit_name[9:].startswith("""small""" ):
UpperCAmelCase = 384
UpperCAmelCase = 1536
UpperCAmelCase = 12
UpperCAmelCase = 6
if deit_name[9:].startswith("""base""" ):
pass
elif deit_name[4:].startswith("""large""" ):
UpperCAmelCase = 1024
UpperCAmelCase = 4096
UpperCAmelCase = 24
UpperCAmelCase = 16
# load original model from timm
UpperCAmelCase = timm.create_model(lowerCAmelCase , pretrained=lowerCAmelCase )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
UpperCAmelCase = timm_model.state_dict()
UpperCAmelCase = create_rename_keys(lowerCAmelCase , lowerCAmelCase )
for src, dest in rename_keys:
rename_key(lowerCAmelCase , lowerCAmelCase , lowerCAmelCase )
read_in_q_k_v(lowerCAmelCase , lowerCAmelCase , lowerCAmelCase )
# load HuggingFace model
UpperCAmelCase = DeiTForImageClassificationWithTeacher(lowerCAmelCase ).eval()
model.load_state_dict(lowerCAmelCase )
# Check outputs on an image, prepared by DeiTImageProcessor
UpperCAmelCase = int(
(256 / 224) * config.image_size ) # to maintain same ratio w.r.t. 224 images, see https://github.com/facebookresearch/deit/blob/ab5715372db8c6cad5740714b2216d55aeae052e/datasets.py#L103
UpperCAmelCase = DeiTImageProcessor(size=lowerCAmelCase , crop_size=config.image_size )
UpperCAmelCase = image_processor(images=prepare_img() , return_tensors="""pt""" )
UpperCAmelCase = encoding["""pixel_values"""]
UpperCAmelCase = model(lowerCAmelCase )
UpperCAmelCase = timm_model(lowerCAmelCase )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(lowerCAmelCase , outputs.logits , atol=1e-3 )
Path(lowerCAmelCase ).mkdir(exist_ok=lowerCAmelCase )
print(F'''Saving model {deit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(lowerCAmelCase )
print(F'''Saving image processor to {pytorch_dump_folder_path}''' )
image_processor.save_pretrained(lowerCAmelCase )
if __name__ == "__main__":
lowerCAmelCase_ : Dict = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--deit_name''',
default='''vit_deit_base_distilled_patch16_224''',
type=str,
help='''Name of the DeiT timm model you\'d like to convert.''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model directory.'''
)
lowerCAmelCase_ : str = parser.parse_args()
convert_deit_checkpoint(args.deit_name, args.pytorch_dump_folder_path)
| 673 | 0 |
'''simple docstring'''
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import DeiTConfig, DeiTForImageClassificationWithTeacher, DeiTImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
__lowercase : List[str] = logging.get_logger(__name__)
def lowerCamelCase (_SCREAMING_SNAKE_CASE : str , _SCREAMING_SNAKE_CASE : List[str]=False ):
__a : Tuple = []
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F"""blocks.{i}.norm1.weight""", F"""deit.encoder.layer.{i}.layernorm_before.weight""") )
rename_keys.append((F"""blocks.{i}.norm1.bias""", F"""deit.encoder.layer.{i}.layernorm_before.bias""") )
rename_keys.append((F"""blocks.{i}.attn.proj.weight""", F"""deit.encoder.layer.{i}.attention.output.dense.weight""") )
rename_keys.append((F"""blocks.{i}.attn.proj.bias""", F"""deit.encoder.layer.{i}.attention.output.dense.bias""") )
rename_keys.append((F"""blocks.{i}.norm2.weight""", F"""deit.encoder.layer.{i}.layernorm_after.weight""") )
rename_keys.append((F"""blocks.{i}.norm2.bias""", F"""deit.encoder.layer.{i}.layernorm_after.bias""") )
rename_keys.append((F"""blocks.{i}.mlp.fc1.weight""", F"""deit.encoder.layer.{i}.intermediate.dense.weight""") )
rename_keys.append((F"""blocks.{i}.mlp.fc1.bias""", F"""deit.encoder.layer.{i}.intermediate.dense.bias""") )
rename_keys.append((F"""blocks.{i}.mlp.fc2.weight""", F"""deit.encoder.layer.{i}.output.dense.weight""") )
rename_keys.append((F"""blocks.{i}.mlp.fc2.bias""", F"""deit.encoder.layer.{i}.output.dense.bias""") )
# projection layer + position embeddings
rename_keys.extend(
[
('cls_token', 'deit.embeddings.cls_token'),
('dist_token', 'deit.embeddings.distillation_token'),
('patch_embed.proj.weight', 'deit.embeddings.patch_embeddings.projection.weight'),
('patch_embed.proj.bias', 'deit.embeddings.patch_embeddings.projection.bias'),
('pos_embed', 'deit.embeddings.position_embeddings'),
] )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
('norm.weight', 'layernorm.weight'),
('norm.bias', 'layernorm.bias'),
('pre_logits.fc.weight', 'pooler.dense.weight'),
('pre_logits.fc.bias', 'pooler.dense.bias'),
] )
# if just the base model, we should remove "deit" from all keys that start with "deit"
__a : Any = [(pair[0], pair[1][4:]) if pair[1].startswith('deit' ) else pair for pair in rename_keys]
else:
# layernorm + classification heads
rename_keys.extend(
[
('norm.weight', 'deit.layernorm.weight'),
('norm.bias', 'deit.layernorm.bias'),
('head.weight', 'cls_classifier.weight'),
('head.bias', 'cls_classifier.bias'),
('head_dist.weight', 'distillation_classifier.weight'),
('head_dist.bias', 'distillation_classifier.bias'),
] )
return rename_keys
def lowerCamelCase (_SCREAMING_SNAKE_CASE : Any , _SCREAMING_SNAKE_CASE : Any , _SCREAMING_SNAKE_CASE : str=False ):
for i in range(config.num_hidden_layers ):
if base_model:
__a : Union[str, Any] = ''
else:
__a : List[Any] = 'deit.'
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
__a : Tuple = state_dict.pop(F"""blocks.{i}.attn.qkv.weight""" )
__a : List[str] = state_dict.pop(F"""blocks.{i}.attn.qkv.bias""" )
# next, add query, keys and values (in that order) to the state dict
__a : Union[str, Any] = in_proj_weight[
: config.hidden_size, :
]
__a : List[str] = in_proj_bias[: config.hidden_size]
__a : Dict = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
__a : str = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
__a : List[str] = in_proj_weight[
-config.hidden_size :, :
]
__a : Optional[Any] = in_proj_bias[-config.hidden_size :]
def lowerCamelCase (_SCREAMING_SNAKE_CASE : int , _SCREAMING_SNAKE_CASE : Optional[int] , _SCREAMING_SNAKE_CASE : Optional[int] ):
__a : Optional[int] = dct.pop(_SCREAMING_SNAKE_CASE )
__a : Optional[Any] = val
def lowerCamelCase ():
__a : Union[str, Any] = 'http://images.cocodataset.org/val2017/000000039769.jpg'
__a : Optional[int] = Image.open(requests.get(_SCREAMING_SNAKE_CASE , stream=_SCREAMING_SNAKE_CASE ).raw )
return im
@torch.no_grad()
def lowerCamelCase (_SCREAMING_SNAKE_CASE : int , _SCREAMING_SNAKE_CASE : int ):
__a : Optional[int] = DeiTConfig()
# all deit models have fine-tuned heads
__a : int = False
# dataset (fine-tuned on ImageNet 2012), patch_size and image_size
__a : str = 1_000
__a : Tuple = 'huggingface/label-files'
__a : Optional[int] = 'imagenet-1k-id2label.json'
__a : Dict = json.load(open(hf_hub_download(_SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE , repo_type='dataset' ) , 'r' ) )
__a : Dict = {int(_SCREAMING_SNAKE_CASE ): v for k, v in idalabel.items()}
__a : Optional[Any] = idalabel
__a : Any = {v: k for k, v in idalabel.items()}
__a : Optional[int] = int(deit_name[-6:-4] )
__a : Tuple = int(deit_name[-3:] )
# size of the architecture
if deit_name[9:].startswith('tiny' ):
__a : str = 192
__a : str = 768
__a : Union[str, Any] = 12
__a : Tuple = 3
elif deit_name[9:].startswith('small' ):
__a : Dict = 384
__a : Optional[int] = 1_536
__a : Tuple = 12
__a : Tuple = 6
if deit_name[9:].startswith('base' ):
pass
elif deit_name[4:].startswith('large' ):
__a : List[str] = 1_024
__a : str = 4_096
__a : Optional[Any] = 24
__a : List[str] = 16
# load original model from timm
__a : Any = timm.create_model(_SCREAMING_SNAKE_CASE , pretrained=_SCREAMING_SNAKE_CASE )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
__a : Optional[int] = timm_model.state_dict()
__a : List[str] = create_rename_keys(_SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE )
for src, dest in rename_keys:
rename_key(_SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE )
read_in_q_k_v(_SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE , _SCREAMING_SNAKE_CASE )
# load HuggingFace model
__a : int = DeiTForImageClassificationWithTeacher(_SCREAMING_SNAKE_CASE ).eval()
model.load_state_dict(_SCREAMING_SNAKE_CASE )
# Check outputs on an image, prepared by DeiTImageProcessor
__a : List[Any] = int(
(256 / 224) * config.image_size ) # to maintain same ratio w.r.t. 224 images, see https://github.com/facebookresearch/deit/blob/ab5715372db8c6cad5740714b2216d55aeae052e/datasets.py#L103
__a : List[str] = DeiTImageProcessor(size=_SCREAMING_SNAKE_CASE , crop_size=config.image_size )
__a : Optional[int] = image_processor(images=prepare_img() , return_tensors='pt' )
__a : Dict = encoding['pixel_values']
__a : Optional[int] = model(_SCREAMING_SNAKE_CASE )
__a : int = timm_model(_SCREAMING_SNAKE_CASE )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(_SCREAMING_SNAKE_CASE , outputs.logits , atol=1e-3 )
Path(_SCREAMING_SNAKE_CASE ).mkdir(exist_ok=_SCREAMING_SNAKE_CASE )
print(F"""Saving model {deit_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(_SCREAMING_SNAKE_CASE )
print(F"""Saving image processor to {pytorch_dump_folder_path}""" )
image_processor.save_pretrained(_SCREAMING_SNAKE_CASE )
if __name__ == "__main__":
__lowercase : Dict = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--deit_name',
default='vit_deit_base_distilled_patch16_224',
type=str,
help='Name of the DeiT timm model you\'d like to convert.',
)
parser.add_argument(
'--pytorch_dump_folder_path', default=None, type=str, help='Path to the output PyTorch model directory.'
)
__lowercase : str = parser.parse_args()
convert_deit_checkpoint(args.deit_name, args.pytorch_dump_folder_path)
| 476 |
"""simple docstring"""
import unittest
from typing import Dict, List, Optional, Union
import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import BridgeTowerImageProcessor
class UpperCamelCase_ ( unittest.TestCase ):
def __init__( self , snake_case__ , snake_case__ = True , snake_case__ = None , snake_case__ = 32 , snake_case__ = True , snake_case__ = 1 / 2_55 , snake_case__ = True , snake_case__ = True , snake_case__ = [0.48_145_466, 0.4_578_275, 0.40_821_073] , snake_case__ = [0.26_862_954, 0.26_130_258, 0.27_577_711] , snake_case__ = True , snake_case__=7 , snake_case__=30 , snake_case__=4_00 , snake_case__=3 , ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = parent
UpperCAmelCase = do_resize
UpperCAmelCase = size if size is not None else {"""shortest_edge""": 2_88}
UpperCAmelCase = size_divisor
UpperCAmelCase = do_rescale
UpperCAmelCase = rescale_factor
UpperCAmelCase = do_normalize
UpperCAmelCase = do_center_crop
UpperCAmelCase = image_mean
UpperCAmelCase = image_std
UpperCAmelCase = do_pad
UpperCAmelCase = batch_size
UpperCAmelCase = num_channels
UpperCAmelCase = min_resolution
UpperCAmelCase = max_resolution
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
"do_normalize": self.do_normalize,
"do_resize": self.do_resize,
"size": self.size,
"size_divisor": self.size_divisor,
}
def UpperCamelCase_ ( self , snake_case__ , snake_case__=False ) -> int:
"""simple docstring"""
if not batched:
UpperCAmelCase = self.size["""shortest_edge"""]
UpperCAmelCase = image_inputs[0]
if isinstance(snake_case__ , Image.Image ):
UpperCAmelCase , UpperCAmelCase = image.size
else:
UpperCAmelCase , UpperCAmelCase = image.shape[1], image.shape[2]
UpperCAmelCase = size / min(snake_case__ , snake_case__ )
if h < w:
UpperCAmelCase , UpperCAmelCase = size, scale * w
else:
UpperCAmelCase , UpperCAmelCase = scale * h, size
UpperCAmelCase = int((13_33 / 8_00) * size )
if max(snake_case__ , snake_case__ ) > max_size:
UpperCAmelCase = max_size / max(snake_case__ , snake_case__ )
UpperCAmelCase = newh * scale
UpperCAmelCase = neww * scale
UpperCAmelCase , UpperCAmelCase = int(newh + 0.5 ), int(neww + 0.5 )
UpperCAmelCase , UpperCAmelCase = (
newh // self.size_divisor * self.size_divisor,
neww // self.size_divisor * self.size_divisor,
)
else:
UpperCAmelCase = []
for image in image_inputs:
UpperCAmelCase , UpperCAmelCase = self.get_expected_values([image] )
expected_values.append((expected_height, expected_width) )
UpperCAmelCase = max(snake_case__ , key=lambda snake_case__ : item[0] )[0]
UpperCAmelCase = max(snake_case__ , key=lambda snake_case__ : item[1] )[1]
return expected_height, expected_width
@require_torch
@require_vision
class UpperCamelCase_ ( a_ , unittest.TestCase ):
_A : List[Any] = BridgeTowerImageProcessor if is_vision_available() else None
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = BridgeTowerImageProcessingTester(self )
@property
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
return self.image_processor_tester.prepare_image_processor_dict()
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
UpperCAmelCase = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(snake_case__ , """image_mean""" ) )
self.assertTrue(hasattr(snake_case__ , """image_std""" ) )
self.assertTrue(hasattr(snake_case__ , """do_normalize""" ) )
self.assertTrue(hasattr(snake_case__ , """do_resize""" ) )
self.assertTrue(hasattr(snake_case__ , """size""" ) )
self.assertTrue(hasattr(snake_case__ , """size_divisor""" ) )
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
pass
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
UpperCAmelCase = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
UpperCAmelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , Image.Image )
# Test not batched input
UpperCAmelCase = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ , batched=snake_case__ )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = self.image_processing_class(**self.image_processor_dict )
# create random numpy tensors
UpperCAmelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ , numpify=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , np.ndarray )
# Test not batched input
UpperCAmelCase = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ , batched=snake_case__ )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = self.image_processing_class(**self.image_processor_dict )
# create random PyTorch tensors
UpperCAmelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=snake_case__ , torchify=snake_case__ )
for image in image_inputs:
self.assertIsInstance(snake_case__ , torch.Tensor )
# Test not batched input
UpperCAmelCase = image_processing(image_inputs[0] , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase = image_processing(snake_case__ , return_tensors="""pt""" ).pixel_values
UpperCAmelCase , UpperCAmelCase = self.image_processor_tester.get_expected_values(snake_case__ , batched=snake_case__ )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
| 673 | 0 |
import argparse
import os
from accelerate.utils import ComputeEnvironment
from .cluster import get_cluster_input
from .config_args import cache_dir, default_config_file, default_yaml_config_file, load_config_from_file # noqa: F401
from .config_utils import _ask_field, _ask_options, _convert_compute_environment # noqa: F401
from .sagemaker import get_sagemaker_input
A_ = '''Launches a series of prompts to create and save a `default_config.yaml` configuration file for your training system. Should always be ran first on your machine'''
def UpperCAmelCase ( )-> List[str]:
'''simple docstring'''
SCREAMING_SNAKE_CASE_ = _ask_options(
'''In which compute environment are you running?''' ,['''This machine''', '''AWS (Amazon SageMaker)'''] ,_convert_compute_environment ,)
if compute_environment == ComputeEnvironment.AMAZON_SAGEMAKER:
SCREAMING_SNAKE_CASE_ = get_sagemaker_input()
else:
SCREAMING_SNAKE_CASE_ = get_cluster_input()
return config
def UpperCAmelCase ( UpperCAmelCase=None )-> Optional[Any]:
'''simple docstring'''
if subparsers is not None:
SCREAMING_SNAKE_CASE_ = subparsers.add_parser('''config''' ,description=UpperCAmelCase )
else:
SCREAMING_SNAKE_CASE_ = argparse.ArgumentParser('''Accelerate config command''' ,description=UpperCAmelCase )
parser.add_argument(
'''--config_file''' ,default=UpperCAmelCase ,help=(
'''The path to use to store the config file. Will default to a file named default_config.yaml in the cache '''
'''location, which is the content of the environment `HF_HOME` suffixed with \'accelerate\', or if you don\'t have '''
'''such an environment variable, your cache directory (\'~/.cache\' or the content of `XDG_CACHE_HOME`) suffixed '''
'''with \'huggingface\'.'''
) ,)
if subparsers is not None:
parser.set_defaults(func=UpperCAmelCase )
return parser
def UpperCAmelCase ( UpperCAmelCase )-> Tuple:
'''simple docstring'''
SCREAMING_SNAKE_CASE_ = get_user_input()
if args.config_file is not None:
SCREAMING_SNAKE_CASE_ = args.config_file
else:
if not os.path.isdir(UpperCAmelCase ):
os.makedirs(UpperCAmelCase )
SCREAMING_SNAKE_CASE_ = default_yaml_config_file
if config_file.endswith('''.json''' ):
config.to_json_file(UpperCAmelCase )
else:
config.to_yaml_file(UpperCAmelCase )
print(f'''accelerate configuration saved at {config_file}''' )
def UpperCAmelCase ( )-> List[Any]:
'''simple docstring'''
SCREAMING_SNAKE_CASE_ = config_command_parser()
SCREAMING_SNAKE_CASE_ = parser.parse_args()
config_command(UpperCAmelCase )
if __name__ == "__main__":
main()
| 393 |
"""simple docstring"""
import pickle
import shutil
import tempfile
import unittest
from transformers import SPIECE_UNDERLINE, XLMRobertaTokenizer, XLMRobertaTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
lowerCAmelCase_ : Any = get_tests_dir('''fixtures/test_sentencepiece.model''')
@require_sentencepiece
@require_tokenizers
class UpperCamelCase_ ( a_ , unittest.TestCase ):
_A : List[str] = XLMRobertaTokenizer
_A : List[str] = XLMRobertaTokenizerFast
_A : Optional[Any] = True
_A : List[str] = True
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
super().setUp()
# We have a SentencePiece fixture for testing
UpperCAmelCase = XLMRobertaTokenizer(snake_case__ , keep_accents=snake_case__ )
tokenizer.save_pretrained(self.tmpdirname )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
UpperCAmelCase = """<pad>"""
UpperCAmelCase = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(snake_case__ ) , snake_case__ )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(snake_case__ ) , snake_case__ )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , """<s>""" )
self.assertEqual(vocab_keys[1] , """<pad>""" )
self.assertEqual(vocab_keys[-1] , """<mask>""" )
self.assertEqual(len(snake_case__ ) , 10_02 )
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
self.assertEqual(self.get_tokenizer().vocab_size , 10_02 )
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = XLMRobertaTokenizer(snake_case__ , keep_accents=snake_case__ )
UpperCAmelCase = tokenizer.tokenize("""This is a test""" )
self.assertListEqual(snake_case__ , ["""▁This""", """▁is""", """▁a""", """▁t""", """est"""] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(snake_case__ ) , [value + tokenizer.fairseq_offset for value in [2_85, 46, 10, 1_70, 3_82]] , )
UpperCAmelCase = tokenizer.tokenize("""I was born in 92000, and this is falsé.""" )
self.assertListEqual(
snake_case__ , [
SPIECE_UNDERLINE + """I""",
SPIECE_UNDERLINE + """was""",
SPIECE_UNDERLINE + """b""",
"""or""",
"""n""",
SPIECE_UNDERLINE + """in""",
SPIECE_UNDERLINE + """""",
"""9""",
"""2""",
"""0""",
"""0""",
"""0""",
""",""",
SPIECE_UNDERLINE + """and""",
SPIECE_UNDERLINE + """this""",
SPIECE_UNDERLINE + """is""",
SPIECE_UNDERLINE + """f""",
"""al""",
"""s""",
"""é""",
""".""",
] , )
UpperCAmelCase = tokenizer.convert_tokens_to_ids(snake_case__ )
self.assertListEqual(
snake_case__ , [
value + tokenizer.fairseq_offset
for value in [8, 21, 84, 55, 24, 19, 7, 2, 6_02, 3_47, 3_47, 3_47, 3, 12, 66, 46, 72, 80, 6, 2, 4]
# ^ unk: 2 + 1 = 3 unk: 2 + 1 = 3 ^
] , )
UpperCAmelCase = tokenizer.convert_ids_to_tokens(snake_case__ )
self.assertListEqual(
snake_case__ , [
SPIECE_UNDERLINE + """I""",
SPIECE_UNDERLINE + """was""",
SPIECE_UNDERLINE + """b""",
"""or""",
"""n""",
SPIECE_UNDERLINE + """in""",
SPIECE_UNDERLINE + """""",
"""<unk>""",
"""2""",
"""0""",
"""0""",
"""0""",
""",""",
SPIECE_UNDERLINE + """and""",
SPIECE_UNDERLINE + """this""",
SPIECE_UNDERLINE + """is""",
SPIECE_UNDERLINE + """f""",
"""al""",
"""s""",
"""<unk>""",
""".""",
] , )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
if not self.test_slow_tokenizer:
# as we don't have a slow version, we can't compare the outputs between slow and fast versions
return
UpperCAmelCase = (self.rust_tokenizer_class, """hf-internal-testing/tiny-xlm-roberta""", {})
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(f'''{tokenizer.__class__.__name__} ({pretrained_name})''' ):
UpperCAmelCase = self.rust_tokenizer_class.from_pretrained(snake_case__ , **snake_case__ )
UpperCAmelCase = self.tokenizer_class.from_pretrained(snake_case__ , **snake_case__ )
UpperCAmelCase = tempfile.mkdtemp()
UpperCAmelCase = tokenizer_r.save_pretrained(snake_case__ )
UpperCAmelCase = tokenizer_p.save_pretrained(snake_case__ )
# Checks it save with the same files + the tokenizer.json file for the fast one
self.assertTrue(any("""tokenizer.json""" in f for f in tokenizer_r_files ) )
UpperCAmelCase = tuple(f for f in tokenizer_r_files if """tokenizer.json""" not in f )
self.assertSequenceEqual(snake_case__ , snake_case__ )
# Checks everything loads correctly in the same way
UpperCAmelCase = tokenizer_r.from_pretrained(snake_case__ )
UpperCAmelCase = tokenizer_p.from_pretrained(snake_case__ )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case__ , snake_case__ ) )
# self.assertEqual(getattr(tokenizer_rp, key), getattr(tokenizer_pp, key))
# self.assertEqual(getattr(tokenizer_rp, key + "_id"), getattr(tokenizer_pp, key + "_id"))
shutil.rmtree(snake_case__ )
# Save tokenizer rust, legacy_format=True
UpperCAmelCase = tempfile.mkdtemp()
UpperCAmelCase = tokenizer_r.save_pretrained(snake_case__ , legacy_format=snake_case__ )
UpperCAmelCase = tokenizer_p.save_pretrained(snake_case__ )
# Checks it save with the same files
self.assertSequenceEqual(snake_case__ , snake_case__ )
# Checks everything loads correctly in the same way
UpperCAmelCase = tokenizer_r.from_pretrained(snake_case__ )
UpperCAmelCase = tokenizer_p.from_pretrained(snake_case__ )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case__ , snake_case__ ) )
shutil.rmtree(snake_case__ )
# Save tokenizer rust, legacy_format=False
UpperCAmelCase = tempfile.mkdtemp()
UpperCAmelCase = tokenizer_r.save_pretrained(snake_case__ , legacy_format=snake_case__ )
UpperCAmelCase = tokenizer_p.save_pretrained(snake_case__ )
# Checks it saved the tokenizer.json file
self.assertTrue(any("""tokenizer.json""" in f for f in tokenizer_r_files ) )
# Checks everything loads correctly in the same way
UpperCAmelCase = tokenizer_r.from_pretrained(snake_case__ )
UpperCAmelCase = tokenizer_p.from_pretrained(snake_case__ )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case__ , snake_case__ ) )
shutil.rmtree(snake_case__ )
@cached_property
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
return XLMRobertaTokenizer.from_pretrained("""xlm-roberta-base""" )
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
with tempfile.NamedTemporaryFile() as f:
shutil.copyfile(snake_case__ , f.name )
UpperCAmelCase = XLMRobertaTokenizer(f.name , keep_accents=snake_case__ )
UpperCAmelCase = pickle.dumps(snake_case__ )
pickle.loads(snake_case__ )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
if not self.test_rust_tokenizer:
return
UpperCAmelCase = self.get_tokenizer()
UpperCAmelCase = self.get_rust_tokenizer()
UpperCAmelCase = """I was born in 92000, and this is falsé."""
UpperCAmelCase = tokenizer.tokenize(snake_case__ )
UpperCAmelCase = rust_tokenizer.tokenize(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
UpperCAmelCase = tokenizer.encode(snake_case__ , add_special_tokens=snake_case__ )
UpperCAmelCase = rust_tokenizer.encode(snake_case__ , add_special_tokens=snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
UpperCAmelCase = self.get_rust_tokenizer()
UpperCAmelCase = tokenizer.encode(snake_case__ )
UpperCAmelCase = rust_tokenizer.encode(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
@slow
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = """Hello World!"""
UpperCAmelCase = [0, 3_53_78, 66_61, 38, 2]
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base') # xlmr.large has same tokenizer
# xlmr.eval()
# xlmr.encode(symbols)
self.assertListEqual(snake_case__ , self.big_tokenizer.encode(snake_case__ ) )
@slow
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = (
"""This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) \" [ ] ! : - . Also we will"""
""" add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth"""
)
UpperCAmelCase = [
0,
32_93,
83,
10,
45_52,
49_89,
79_86,
6_78,
10,
59_15,
1_11,
17_94_59,
12_48_50,
4,
60_44,
2_37,
12,
6,
5,
6,
4,
67_80,
7_05,
15,
13_88,
44,
3_78,
1_01_14,
7_11,
1_52,
20,
6,
5,
2_23_76,
6_42,
12_21,
1_51_90,
3_41_53,
4_50,
56_08,
9_59,
11_19,
5_77_02,
1_36,
1_86,
47,
10_98,
2_93_67,
47,
# 4426, # What fairseq tokenizes from "<unk>": "_<"
# 3678, # What fairseq tokenizes from "<unk>": "unk"
# 2740, # What fairseq tokenizes from "<unk>": ">"
3, # What we tokenize from "<unk>": "<unk>"
6, # Residue from the tokenization: an extra sentencepiece underline
4,
60_44,
2_37,
62_84,
5_09_01,
5_28,
31,
90,
34,
9_27,
2,
]
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base') # xlmr.large has same tokenizer
# xlmr.eval()
# xlmr.encode(symbols)
self.assertListEqual(snake_case__ , self.big_tokenizer.encode(snake_case__ ) )
@slow
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
UpperCAmelCase = {"""input_ids""": [[0, 1_10_62, 8_27_72, 7, 15, 8_27_72, 5_38, 5_15_29, 2_37, 1_71_98, 12_90, 2_06, 9, 21_51_75, 13_14, 1_36, 1_71_98, 12_90, 2_06, 9, 5_63_59, 42, 12_20_09, 9, 1_64_66, 16, 8_73_44, 45_37, 9, 47_17, 7_83_81, 6, 15_99_58, 7, 15, 2_44_80, 6_18, 4, 5_27, 2_26_93, 54_28, 4, 27_77, 2_44_80, 98_74, 4, 4_35_23, 5_94, 4, 8_03, 1_83_92, 3_31_89, 18, 4, 4_35_23, 2_44_47, 1_23_99, 1_00, 2_49_55, 8_36_58, 96_26, 14_40_57, 15, 8_39, 2_23_35, 16, 1_36, 2_49_55, 8_36_58, 8_34_79, 15, 3_91_02, 7_24, 16, 6_78, 6_45, 27_89, 13_28, 45_89, 42, 12_20_09, 11_57_74, 23, 8_05, 13_28, 4_68_76, 7, 1_36, 5_38_94, 19_40, 4_22_27, 4_11_59, 1_77_21, 8_23, 4_25, 4, 2_75_12, 9_87_22, 2_06, 1_36, 55_31, 49_70, 9_19, 1_73_36, 5, 2], [0, 2_00_80, 6_18, 83, 8_27_75, 47, 4_79, 9, 15_17, 73, 5_38_94, 3_33, 8_05_81, 11_01_17, 1_88_11, 52_56, 12_95, 51, 15_25_26, 2_97, 79_86, 3_90, 12_44_16, 5_38, 3_54_31, 2_14, 98, 1_50_44, 2_57_37, 1_36, 71_08, 4_37_01, 23, 7_56, 13_53_55, 7, 5, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [0, 5_81, 6_37_73, 11_94_55, 6, 14_77_97, 8_82_03, 7, 6_45, 70, 21, 32_85, 1_02_69, 5, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]], """attention_mask""": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=snake_case__ , model_name="""xlm-roberta-base""" , revision="""d9d8a8ea5eb94b1c6654ae9249df7793cd2933d3""" , )
| 673 | 0 |
import math
from enum import Enum
from typing import Optional, Union
from torch.optim import Optimizer
from torch.optim.lr_scheduler import LambdaLR
from .utils import logging
__UpperCAmelCase = logging.get_logger(__name__)
class lowerCamelCase (a_ ):
'''simple docstring'''
_snake_case : List[str] = 'linear'
_snake_case : Union[str, Any] = 'cosine'
_snake_case : Dict = 'cosine_with_restarts'
_snake_case : List[Any] = 'polynomial'
_snake_case : int = 'constant'
_snake_case : Optional[int] = 'constant_with_warmup'
_snake_case : str = 'piecewise_constant'
def lowercase__ ( __snake_case : str , __snake_case : Tuple = -1 ):
'''simple docstring'''
return LambdaLR(__snake_case , lambda __snake_case : 1 , last_epoch=__snake_case )
def lowercase__ ( __snake_case : List[str] , __snake_case : Optional[int] , __snake_case : Optional[Any] = -1 ):
'''simple docstring'''
def lr_lambda(__snake_case : List[Any] ):
if current_step < num_warmup_steps:
return float(__snake_case ) / float(max(1.0 , __snake_case ) )
return 1.0
return LambdaLR(__snake_case , __snake_case , last_epoch=__snake_case )
def lowercase__ ( __snake_case : Any , __snake_case : Any , __snake_case : int = -1 ):
'''simple docstring'''
UpperCAmelCase_ : List[str] = {}
UpperCAmelCase_ : int = step_rules.split(',' )
for rule_str in rule_list[:-1]:
UpperCAmelCase_ , UpperCAmelCase_ : str = rule_str.split(':' )
UpperCAmelCase_ : Any = int(__snake_case )
UpperCAmelCase_ : List[str] = float(__snake_case )
UpperCAmelCase_ : Dict = value
UpperCAmelCase_ : str = float(rule_list[-1] )
def create_rules_function(__snake_case : Tuple , __snake_case : Dict ):
def rule_func(__snake_case : Dict ) -> float:
UpperCAmelCase_ : Union[str, Any] = sorted(rules_dict.keys() )
for i, sorted_step in enumerate(__snake_case ):
if steps < sorted_step:
return rules_dict[sorted_steps[i]]
return last_lr_multiple
return rule_func
UpperCAmelCase_ : Dict = create_rules_function(__snake_case , __snake_case )
return LambdaLR(__snake_case , __snake_case , last_epoch=__snake_case )
def lowercase__ ( __snake_case : Union[str, Any] , __snake_case : List[Any] , __snake_case : int , __snake_case : int=-1 ):
'''simple docstring'''
def lr_lambda(__snake_case : Optional[int] ):
if current_step < num_warmup_steps:
return float(__snake_case ) / float(max(1 , __snake_case ) )
return max(
0.0 , float(num_training_steps - current_step ) / float(max(1 , num_training_steps - num_warmup_steps ) ) )
return LambdaLR(__snake_case , __snake_case , __snake_case )
def lowercase__ ( __snake_case : List[str] , __snake_case : Tuple , __snake_case : Optional[Any] , __snake_case : List[str] = 0.5 , __snake_case : List[str] = -1 ):
'''simple docstring'''
def lr_lambda(__snake_case : Optional[int] ):
if current_step < num_warmup_steps:
return float(__snake_case ) / float(max(1 , __snake_case ) )
UpperCAmelCase_ : Tuple = float(current_step - num_warmup_steps ) / float(max(1 , num_training_steps - num_warmup_steps ) )
return max(0.0 , 0.5 * (1.0 + math.cos(math.pi * float(__snake_case ) * 2.0 * progress )) )
return LambdaLR(__snake_case , __snake_case , __snake_case )
def lowercase__ ( __snake_case : Optional[Any] , __snake_case : List[str] , __snake_case : Dict , __snake_case : int = 1 , __snake_case : List[str] = -1 ):
'''simple docstring'''
def lr_lambda(__snake_case : str ):
if current_step < num_warmup_steps:
return float(__snake_case ) / float(max(1 , __snake_case ) )
UpperCAmelCase_ : Any = float(current_step - num_warmup_steps ) / float(max(1 , num_training_steps - num_warmup_steps ) )
if progress >= 1.0:
return 0.0
return max(0.0 , 0.5 * (1.0 + math.cos(math.pi * ((float(__snake_case ) * progress) % 1.0) )) )
return LambdaLR(__snake_case , __snake_case , __snake_case )
def lowercase__ ( __snake_case : List[str] , __snake_case : List[Any] , __snake_case : Any , __snake_case : Optional[Any]=1E-7 , __snake_case : List[str]=1.0 , __snake_case : Any=-1 ):
'''simple docstring'''
UpperCAmelCase_ : Optional[int] = optimizer.defaults['lr']
if not (lr_init > lr_end):
raise ValueError(F"lr_end ({lr_end}) must be be smaller than initial lr ({lr_init})" )
def lr_lambda(__snake_case : str ):
if current_step < num_warmup_steps:
return float(__snake_case ) / float(max(1 , __snake_case ) )
elif current_step > num_training_steps:
return lr_end / lr_init # as LambdaLR multiplies by lr_init
else:
UpperCAmelCase_ : Union[str, Any] = lr_init - lr_end
UpperCAmelCase_ : List[Any] = num_training_steps - num_warmup_steps
UpperCAmelCase_ : Tuple = 1 - (current_step - num_warmup_steps) / decay_steps
UpperCAmelCase_ : Tuple = lr_range * pct_remaining**power + lr_end
return decay / lr_init # as LambdaLR multiplies by lr_init
return LambdaLR(__snake_case , __snake_case , __snake_case )
__UpperCAmelCase = {
SchedulerType.LINEAR: get_linear_schedule_with_warmup,
SchedulerType.COSINE: get_cosine_schedule_with_warmup,
SchedulerType.COSINE_WITH_RESTARTS: get_cosine_with_hard_restarts_schedule_with_warmup,
SchedulerType.POLYNOMIAL: get_polynomial_decay_schedule_with_warmup,
SchedulerType.CONSTANT: get_constant_schedule,
SchedulerType.CONSTANT_WITH_WARMUP: get_constant_schedule_with_warmup,
SchedulerType.PIECEWISE_CONSTANT: get_piecewise_constant_schedule,
}
def lowercase__ ( __snake_case : List[Any] , __snake_case : Union[str, Any] , __snake_case : Optional[Any] = None , __snake_case : Dict = None , __snake_case : Dict = None , __snake_case : Tuple = 1 , __snake_case : int = 1.0 , __snake_case : Any = -1 , ):
'''simple docstring'''
UpperCAmelCase_ : str = SchedulerType(__snake_case )
UpperCAmelCase_ : Optional[Any] = TYPE_TO_SCHEDULER_FUNCTION[name]
if name == SchedulerType.CONSTANT:
return schedule_func(__snake_case , last_epoch=__snake_case )
if name == SchedulerType.PIECEWISE_CONSTANT:
return schedule_func(__snake_case , step_rules=__snake_case , last_epoch=__snake_case )
# All other schedulers require `num_warmup_steps`
if num_warmup_steps is None:
raise ValueError(F"{name} requires `num_warmup_steps`, please provide that argument." )
if name == SchedulerType.CONSTANT_WITH_WARMUP:
return schedule_func(__snake_case , num_warmup_steps=__snake_case , last_epoch=__snake_case )
# All other schedulers require `num_training_steps`
if num_training_steps is None:
raise ValueError(F"{name} requires `num_training_steps`, please provide that argument." )
if name == SchedulerType.COSINE_WITH_RESTARTS:
return schedule_func(
__snake_case , num_warmup_steps=__snake_case , num_training_steps=__snake_case , num_cycles=__snake_case , last_epoch=__snake_case , )
if name == SchedulerType.POLYNOMIAL:
return schedule_func(
__snake_case , num_warmup_steps=__snake_case , num_training_steps=__snake_case , power=__snake_case , last_epoch=__snake_case , )
return schedule_func(
__snake_case , num_warmup_steps=__snake_case , num_training_steps=__snake_case , last_epoch=__snake_case )
| 406 |
"""simple docstring"""
import socket
def _lowerCAmelCase ( ):
'''simple docstring'''
UpperCAmelCase = socket.socket(socket.AF_INET , socket.SOCK_STREAM )
UpperCAmelCase = socket.gethostname()
UpperCAmelCase = 12312
sock.connect((host, port) )
sock.send(b"""Hello server!""" )
with open("""Received_file""" , """wb""" ) as out_file:
print("""File opened""" )
print("""Receiving data...""" )
while True:
UpperCAmelCase = sock.recv(1024 )
if not data:
break
out_file.write(lowerCAmelCase )
print("""Successfully received the file""" )
sock.close()
print("""Connection closed""" )
if __name__ == "__main__":
main()
| 673 | 0 |
"""simple docstring"""
def __a ( A ) -> str:
'''simple docstring'''
A__ = [0] * len(A )
A__ = []
A__ = [1] * len(A )
for values in graph.values():
for i in values:
indegree[i] += 1
for i in range(len(A ) ):
if indegree[i] == 0:
queue.append(A )
while queue:
A__ = queue.pop(0 )
for x in graph[vertex]:
indegree[x] -= 1
if long_dist[vertex] + 1 > long_dist[x]:
A__ = long_dist[vertex] + 1
if indegree[x] == 0:
queue.append(A )
print(max(A ) )
# Adjacency list of Graph
__UpperCAmelCase ={0: [2, 3, 4], 1: [2, 7], 2: [5], 3: [5, 7], 4: [7], 5: [6], 6: [7], 7: []}
longest_distance(graph) | 337 |
"""simple docstring"""
import math
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
return math.sqrt(lowerCAmelCase ) * math.sqrt(lowerCAmelCase ) == num
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = 0
UpperCAmelCase = n
while left <= right:
UpperCAmelCase = (left + right) // 2
if mid**2 == n:
return True
elif mid**2 > n:
UpperCAmelCase = mid - 1
else:
UpperCAmelCase = mid + 1
return False
if __name__ == "__main__":
import doctest
doctest.testmod()
| 673 | 0 |
'''simple docstring'''
from ...utils import (
OptionalDependencyNotAvailable,
is_torch_available,
is_transformers_available,
is_transformers_version,
)
try:
if not (is_transformers_available() and is_torch_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ...utils.dummy_torch_and_transformers_objects import (
ImageTextPipelineOutput,
UniDiffuserPipeline,
)
else:
from .modeling_text_decoder import UniDiffuserTextDecoder
from .modeling_uvit import UniDiffuserModel, UTransformeraDModel
from .pipeline_unidiffuser import ImageTextPipelineOutput, UniDiffuserPipeline
| 546 |
"""simple docstring"""
import functools
import gc
import inspect
import torch
from .imports import is_npu_available, is_xpu_available
def _lowerCAmelCase ( *lowerCAmelCase ):
'''simple docstring'''
if not isinstance(lowerCAmelCase , lowerCAmelCase ):
UpperCAmelCase = list(lowerCAmelCase )
for i in range(len(lowerCAmelCase ) ):
UpperCAmelCase = None
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
return objects
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = [
"""CUDA out of memory.""", # CUDA OOM
"""cuDNN error: CUDNN_STATUS_NOT_SUPPORTED.""", # CUDNN SNAFU
"""DefaultCPUAllocator: can't allocate memory""", # CPU OOM
]
if isinstance(lowerCAmelCase , lowerCAmelCase ) and len(exception.args ) == 1:
return any(err in exception.args[0] for err in _statements )
return False
def _lowerCAmelCase ( lowerCAmelCase = None , lowerCAmelCase = 128 ):
'''simple docstring'''
if function is None:
return functools.partial(lowerCAmelCase , starting_batch_size=lowerCAmelCase )
UpperCAmelCase = starting_batch_size
def decorator(*lowerCAmelCase , **lowerCAmelCase ):
nonlocal batch_size
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
UpperCAmelCase = list(inspect.signature(lowerCAmelCase ).parameters.keys() )
# Guard against user error
if len(lowerCAmelCase ) < (len(lowerCAmelCase ) + 1):
UpperCAmelCase = """, """.join([F'''{arg}={value}''' for arg, value in zip(params[1:] , args[1:] )] )
raise TypeError(
F'''Batch size was passed into `{function.__name__}` as the first argument when called.'''
F'''Remove this as the decorator already does so: `{function.__name__}({arg_str})`''' )
while True:
if batch_size == 0:
raise RuntimeError("""No executable batch size found, reached zero.""" )
try:
return function(lowerCAmelCase , *lowerCAmelCase , **lowerCAmelCase )
except Exception as e:
if should_reduce_batch_size(lowerCAmelCase ):
gc.collect()
if is_xpu_available():
torch.xpu.empty_cache()
elif is_npu_available():
torch.npu.empty_cache()
else:
torch.cuda.empty_cache()
batch_size //= 2
else:
raise
return decorator
| 673 | 0 |
import numpy as np
def a_ ( SCREAMING_SNAKE_CASE__ : Tuple , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : Optional[int] = 1e-12 , SCREAMING_SNAKE_CASE__ : Any = 100 , ):
'''simple docstring'''
assert np.shape(SCREAMING_SNAKE_CASE__ )[0] == np.shape(SCREAMING_SNAKE_CASE__ )[1]
# Ensure proper dimensionality.
assert np.shape(SCREAMING_SNAKE_CASE__ )[0] == np.shape(SCREAMING_SNAKE_CASE__ )[0]
# Ensure inputs are either both complex or both real
assert np.iscomplexobj(SCREAMING_SNAKE_CASE__ ) == np.iscomplexobj(SCREAMING_SNAKE_CASE__ )
_lowerCamelCase : Union[str, Any] =np.iscomplexobj(SCREAMING_SNAKE_CASE__ )
if is_complex:
# Ensure complex input_matrix is Hermitian
assert np.array_equal(SCREAMING_SNAKE_CASE__ , input_matrix.conj().T )
# Set convergence to False. Will define convergence when we exceed max_iterations
# or when we have small changes from one iteration to next.
_lowerCamelCase : Optional[int] =False
_lowerCamelCase : Optional[Any] =0
_lowerCamelCase : int =0
_lowerCamelCase : Tuple =1e12
while not convergence:
# Multiple matrix by the vector.
_lowerCamelCase : Optional[int] =np.dot(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
# Normalize the resulting output vector.
_lowerCamelCase : str =w / np.linalg.norm(SCREAMING_SNAKE_CASE__ )
# Find rayleigh quotient
# (faster than usual b/c we know vector is normalized already)
_lowerCamelCase : Tuple =vector.conj().T if is_complex else vector.T
_lowerCamelCase : Optional[Any] =np.dot(SCREAMING_SNAKE_CASE__ , np.dot(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) )
# Check convergence.
_lowerCamelCase : Optional[Any] =np.abs(lambda_ - lambda_previous ) / lambda_
iterations += 1
if error <= error_tol or iterations >= max_iterations:
_lowerCamelCase : Any =True
_lowerCamelCase : List[str] =lambda_
if is_complex:
_lowerCamelCase : int =np.real(lambda_ )
return lambda_, vector
def a_ ( ):
'''simple docstring'''
_lowerCamelCase : Optional[int] =np.array([[41, 4, 20], [4, 26, 30], [20, 30, 50]] )
_lowerCamelCase : Optional[Any] =np.array([41, 4, 20] )
_lowerCamelCase : str =real_input_matrix.astype(np.complexaaa )
_lowerCamelCase : Optional[Any] =np.triu(1J * complex_input_matrix , 1 )
complex_input_matrix += imag_matrix
complex_input_matrix += -1 * imag_matrix.T
_lowerCamelCase : Any =np.array([41, 4, 20] ).astype(np.complexaaa )
for problem_type in ["real", "complex"]:
if problem_type == "real":
_lowerCamelCase : str =real_input_matrix
_lowerCamelCase : int =real_vector
elif problem_type == "complex":
_lowerCamelCase : Dict =complex_input_matrix
_lowerCamelCase : Optional[Any] =complex_vector
# Our implementation.
_lowerCamelCase , _lowerCamelCase : Tuple =power_iteration(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
# Numpy implementation.
# Get eigenvalues and eigenvectors using built-in numpy
# eigh (eigh used for symmetric or hermetian matrices).
_lowerCamelCase , _lowerCamelCase : Optional[int] =np.linalg.eigh(SCREAMING_SNAKE_CASE__ )
# Last eigenvalue is the maximum one.
_lowerCamelCase : Union[str, Any] =eigen_values[-1]
# Last column in this matrix is eigenvector corresponding to largest eigenvalue.
_lowerCamelCase : Tuple =eigen_vectors[:, -1]
# Check our implementation and numpy gives close answers.
assert np.abs(eigen_value - eigen_value_max ) <= 1e-6
# Take absolute values element wise of each eigenvector.
# as they are only unique to a minus sign.
assert np.linalg.norm(np.abs(SCREAMING_SNAKE_CASE__ ) - np.abs(SCREAMING_SNAKE_CASE__ ) ) <= 1e-6
if __name__ == "__main__":
import doctest
doctest.testmod()
test_power_iteration()
| 464 |
"""simple docstring"""
import math
def _lowerCAmelCase ( lowerCAmelCase = 100 ):
'''simple docstring'''
UpperCAmelCase = sum(i * i for i in range(1 , n + 1 ) )
UpperCAmelCase = int(math.pow(sum(range(1 , n + 1 ) ) , 2 ) )
return square_of_sum - sum_of_squares
if __name__ == "__main__":
print(F'{solution() = }')
| 673 | 0 |
"""simple docstring"""
import numpy as np
def _SCREAMING_SNAKE_CASE (__lowerCAmelCase ) -> Tuple:
'''simple docstring'''
return 1 / (1 + np.exp(-vector ))
def _SCREAMING_SNAKE_CASE (__lowerCAmelCase ) -> int:
'''simple docstring'''
return vector * sigmoid(__lowerCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 567 |
"""simple docstring"""
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = [0] * len(lowerCAmelCase )
UpperCAmelCase = []
UpperCAmelCase = [1] * len(lowerCAmelCase )
for values in graph.values():
for i in values:
indegree[i] += 1
for i in range(len(lowerCAmelCase ) ):
if indegree[i] == 0:
queue.append(lowerCAmelCase )
while queue:
UpperCAmelCase = queue.pop(0 )
for x in graph[vertex]:
indegree[x] -= 1
if long_dist[vertex] + 1 > long_dist[x]:
UpperCAmelCase = long_dist[vertex] + 1
if indegree[x] == 0:
queue.append(lowerCAmelCase )
print(max(lowerCAmelCase ) )
# Adjacency list of Graph
lowerCAmelCase_ : str = {0: [2, 3, 4], 1: [2, 7], 2: [5], 3: [5, 7], 4: [7], 5: [6], 6: [7], 7: []}
longest_distance(graph)
| 673 | 0 |
'''simple docstring'''
from typing import Optional
from urllib.parse import quote
import huggingface_hub as hfh
from packaging import version
def lowerCAmelCase_ ( _lowerCamelCase: List[str] , _lowerCamelCase: List[str] , _lowerCamelCase: str = None ):
if version.parse(hfh.__version__ ).release < version.parse("""0.11.0""" ).release:
# old versions of hfh don't url-encode the file path
__SCREAMING_SNAKE_CASE : Optional[int] = quote(_lowerCamelCase )
return hfh.hf_hub_url(_lowerCamelCase , _lowerCamelCase , repo_type="""dataset""" , revision=_lowerCamelCase ) | 578 |
"""simple docstring"""
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import torch
from ..models.auto import AutoModelForSequenceClassification, AutoTokenizer
from .base import PipelineTool
class UpperCamelCase_ ( a_ ):
_A : Optional[int] = 'facebook/bart-large-mnli'
_A : Union[str, Any] = (
'This is a tool that classifies an English text using provided labels. It takes two inputs: `text`, which '
'should be the text to classify, and `labels`, which should be the list of labels to use for classification. '
'It returns the most likely label in the list of provided `labels` for the input text.'
)
_A : Dict = 'text_classifier'
_A : Union[str, Any] = AutoTokenizer
_A : Tuple = AutoModelForSequenceClassification
_A : Optional[int] = ['text', ['text']]
_A : Dict = ['text']
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
super().setup()
UpperCAmelCase = self.model.config
UpperCAmelCase = -1
for idx, label in config.idalabel.items():
if label.lower().startswith("""entail""" ):
UpperCAmelCase = int(snake_case__ )
if self.entailment_id == -1:
raise ValueError("""Could not determine the entailment ID from the model config, please pass it at init.""" )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = labels
return self.pre_processor(
[text] * len(snake_case__ ) , [f'''This example is {label}''' for label in labels] , return_tensors="""pt""" , padding="""max_length""" , )
def UpperCamelCase_ ( self , snake_case__ ) -> str:
"""simple docstring"""
UpperCAmelCase = outputs.logits
UpperCAmelCase = torch.argmax(logits[:, 2] ).item()
return self._labels[label_id]
| 673 | 0 |
"""simple docstring"""
def __lowercase ( _a ):
snake_case_ : Any = int(_a )
if n_element < 1:
snake_case_ : Optional[int] = ValueError('''a should be a positive number''' )
raise my_error
snake_case_ : Dict = [1]
snake_case_, snake_case_, snake_case_ : Tuple = (0, 0, 0)
snake_case_ : List[Any] = 1
while index < n_element:
while hamming_list[i] * 2 <= hamming_list[-1]:
i += 1
while hamming_list[j] * 3 <= hamming_list[-1]:
j += 1
while hamming_list[k] * 5 <= hamming_list[-1]:
k += 1
hamming_list.append(
min(hamming_list[i] * 2 , hamming_list[j] * 3 , hamming_list[k] * 5 ) )
index += 1
return hamming_list
if __name__ == "__main__":
lowercase__ : Dict = input('''Enter the last number (nth term) of the Hamming Number Series: ''')
print('''Formula of Hamming Number Series => 2^i * 3^j * 5^k''')
lowercase__ : str = hamming(int(n))
print('''-----------------------------------------------------''')
print(f'The list with nth numbers is: {hamming_numbers}')
print('''-----------------------------------------------------''')
| 123 |
"""simple docstring"""
from dataclasses import dataclass
from enum import Enum
from typing import List, Optional, Union
import numpy as np
import PIL
from PIL import Image
from ...utils import BaseOutput, is_torch_available, is_transformers_available
@dataclass
class UpperCamelCase_ ( a_ ):
_A : Union[List[PIL.Image.Image], np.ndarray]
_A : Optional[List[bool]]
if is_transformers_available() and is_torch_available():
from .pipeline_semantic_stable_diffusion import SemanticStableDiffusionPipeline
| 673 | 0 |
'''simple docstring'''
from typing import List, Optional, Union
import numpy as np
import PIL
import torch
from PIL import Image
from ...models import UNetaDConditionModel, VQModel
from ...pipelines import DiffusionPipeline
from ...pipelines.pipeline_utils import ImagePipelineOutput
from ...schedulers import DDPMScheduler
from ...utils import (
is_accelerate_available,
is_accelerate_version,
logging,
randn_tensor,
replace_example_docstring,
)
lowercase = logging.get_logger(__name__) # pylint: disable=invalid-name
lowercase = '''
Examples:
```py
>>> from diffusers import KandinskyV22Img2ImgPipeline, KandinskyV22PriorPipeline
>>> from diffusers.utils import load_image
>>> import torch
>>> pipe_prior = KandinskyV22PriorPipeline.from_pretrained(
... "kandinsky-community/kandinsky-2-2-prior", torch_dtype=torch.float16
... )
>>> pipe_prior.to("cuda")
>>> prompt = "A red cartoon frog, 4k"
>>> image_emb, zero_image_emb = pipe_prior(prompt, return_dict=False)
>>> pipe = KandinskyV22Img2ImgPipeline.from_pretrained(
... "kandinsky-community/kandinsky-2-2-decoder", torch_dtype=torch.float16
... )
>>> pipe.to("cuda")
>>> init_image = load_image(
... "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main"
... "/kandinsky/frog.png"
... )
>>> image = pipe(
... image=init_image,
... image_embeds=image_emb,
... negative_image_embeds=zero_image_emb,
... height=768,
... width=768,
... num_inference_steps=100,
... strength=0.2,
... ).images
>>> image[0].save("red_frog.png")
```
'''
def __A ( _SCREAMING_SNAKE_CASE : Any , _SCREAMING_SNAKE_CASE : List[str] , _SCREAMING_SNAKE_CASE : List[str]=8 ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : Optional[Any] = height // scale_factor**2
if height % scale_factor**2 != 0:
new_height += 1
__SCREAMING_SNAKE_CASE : Optional[int] = width // scale_factor**2
if width % scale_factor**2 != 0:
new_width += 1
return new_height * scale_factor, new_width * scale_factor
def __A ( _SCREAMING_SNAKE_CASE : str , _SCREAMING_SNAKE_CASE : str=5_1_2 , _SCREAMING_SNAKE_CASE : List[Any]=5_1_2 ):
"""simple docstring"""
__SCREAMING_SNAKE_CASE : List[Any] = pil_image.resize((w, h) , resample=Image.BICUBIC , reducing_gap=1 )
__SCREAMING_SNAKE_CASE : Tuple = np.array(pil_image.convert("RGB" ) )
__SCREAMING_SNAKE_CASE : Union[str, Any] = arr.astype(np.floataa ) / 1_2_7.5 - 1
__SCREAMING_SNAKE_CASE : int = np.transpose(_SCREAMING_SNAKE_CASE , [2, 0, 1] )
__SCREAMING_SNAKE_CASE : List[str] = torch.from_numpy(_SCREAMING_SNAKE_CASE ).unsqueeze(0 )
return image
class __lowerCamelCase ( a_ ):
'''simple docstring'''
def __init__( self , a__ , a__ , a__ , ):
super().__init__()
self.register_modules(
unet=snake_case__ , scheduler=snake_case__ , movq=snake_case__ , )
__SCREAMING_SNAKE_CASE : List[Any] = 2 ** (len(self.movq.config.block_out_channels ) - 1)
def a_ ( self , a__ , a__ , a__ ):
__SCREAMING_SNAKE_CASE : Optional[Any] = min(int(num_inference_steps * strength ) , snake_case__ )
__SCREAMING_SNAKE_CASE : Tuple = max(num_inference_steps - init_timestep , 0 )
__SCREAMING_SNAKE_CASE : List[str] = self.scheduler.timesteps[t_start:]
return timesteps, num_inference_steps - t_start
def a_ ( self , a__ , a__ , a__ , a__ , a__ , a__ , a__=None ):
if not isinstance(snake_case__ , (torch.Tensor, PIL.Image.Image, list) ):
raise ValueError(
f'`image` has to be of type `torch.Tensor`, `PIL.Image.Image` or list but is {type(snake_case__ )}' )
__SCREAMING_SNAKE_CASE : Optional[Any] = image.to(device=snake_case__ , dtype=snake_case__ )
__SCREAMING_SNAKE_CASE : str = batch_size * num_images_per_prompt
if image.shape[1] == 4:
__SCREAMING_SNAKE_CASE : Tuple = image
else:
if isinstance(snake_case__ , snake_case__ ) and len(snake_case__ ) != batch_size:
raise ValueError(
f'You have passed a list of generators of length {len(snake_case__ )}, but requested an effective batch'
f' size of {batch_size}. Make sure the batch size matches the length of the generators.' )
elif isinstance(snake_case__ , snake_case__ ):
__SCREAMING_SNAKE_CASE : Tuple = [
self.movq.encode(image[i : i + 1] ).latent_dist.sample(generator[i] ) for i in range(snake_case__ )
]
__SCREAMING_SNAKE_CASE : Optional[int] = torch.cat(snake_case__ , dim=0 )
else:
__SCREAMING_SNAKE_CASE : Optional[Any] = self.movq.encode(snake_case__ ).latent_dist.sample(snake_case__ )
__SCREAMING_SNAKE_CASE : List[Any] = self.movq.config.scaling_factor * init_latents
__SCREAMING_SNAKE_CASE : Tuple = torch.cat([init_latents] , dim=0 )
__SCREAMING_SNAKE_CASE : Any = init_latents.shape
__SCREAMING_SNAKE_CASE : Any = randn_tensor(snake_case__ , generator=snake_case__ , device=snake_case__ , dtype=snake_case__ )
# get latents
__SCREAMING_SNAKE_CASE : Any = self.scheduler.add_noise(snake_case__ , snake_case__ , snake_case__ )
__SCREAMING_SNAKE_CASE : Optional[Any] = init_latents
return latents
def a_ ( self , a__=0 ):
if is_accelerate_available():
from accelerate import cpu_offload
else:
raise ImportError("Please install accelerate via `pip install accelerate`" )
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.device(f'cuda:{gpu_id}' )
__SCREAMING_SNAKE_CASE : str = [
self.unet,
self.movq,
]
for cpu_offloaded_model in models:
if cpu_offloaded_model is not None:
cpu_offload(snake_case__ , snake_case__ )
def a_ ( self , a__=0 ):
if is_accelerate_available() and is_accelerate_version(">=" , "0.17.0.dev0" ):
from accelerate import cpu_offload_with_hook
else:
raise ImportError("`enable_model_cpu_offload` requires `accelerate v0.17.0` or higher." )
__SCREAMING_SNAKE_CASE : Dict = torch.device(f'cuda:{gpu_id}' )
if self.device.type != "cpu":
self.to("cpu" , silence_dtype_warnings=snake_case__ )
torch.cuda.empty_cache() # otherwise we don't see the memory savings (but they probably exist)
__SCREAMING_SNAKE_CASE : Tuple = None
for cpu_offloaded_model in [self.unet, self.movq]:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = cpu_offload_with_hook(snake_case__ , snake_case__ , prev_module_hook=snake_case__ )
# We'll offload the last model manually.
__SCREAMING_SNAKE_CASE : Optional[Any] = hook
@property
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._execution_device
def a_ ( self ):
if not hasattr(self.unet , "_hf_hook" ):
return self.device
for module in self.unet.modules():
if (
hasattr(snake_case__ , "_hf_hook" )
and hasattr(module._hf_hook , "execution_device" )
and module._hf_hook.execution_device is not None
):
return torch.device(module._hf_hook.execution_device )
return self.device
@torch.no_grad()
@replace_example_docstring(snake_case__ )
def __call__( self , a__ , a__ , a__ , a__ = 512 , a__ = 512 , a__ = 100 , a__ = 4.0 , a__ = 0.3 , a__ = 1 , a__ = None , a__ = "pil" , a__ = True , ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = self._execution_device
__SCREAMING_SNAKE_CASE : Dict = guidance_scale > 1.0
if isinstance(snake_case__ , snake_case__ ):
__SCREAMING_SNAKE_CASE : Optional[Any] = torch.cat(snake_case__ , dim=0 )
__SCREAMING_SNAKE_CASE : Optional[int] = image_embeds.shape[0]
if isinstance(snake_case__ , snake_case__ ):
__SCREAMING_SNAKE_CASE : List[str] = torch.cat(snake_case__ , dim=0 )
if do_classifier_free_guidance:
__SCREAMING_SNAKE_CASE : Optional[int] = image_embeds.repeat_interleave(snake_case__ , dim=0 )
__SCREAMING_SNAKE_CASE : List[str] = negative_image_embeds.repeat_interleave(snake_case__ , dim=0 )
__SCREAMING_SNAKE_CASE : Any = torch.cat([negative_image_embeds, image_embeds] , dim=0 ).to(dtype=self.unet.dtype , device=snake_case__ )
if not isinstance(snake_case__ , snake_case__ ):
__SCREAMING_SNAKE_CASE : Optional[int] = [image]
if not all(isinstance(snake_case__ , (PIL.Image.Image, torch.Tensor) ) for i in image ):
raise ValueError(
f'Input is in incorrect format: {[type(snake_case__ ) for i in image]}. Currently, we only support PIL image and pytorch tensor' )
__SCREAMING_SNAKE_CASE : Any = torch.cat([prepare_image(snake_case__ , snake_case__ , snake_case__ ) for i in image] , dim=0 )
__SCREAMING_SNAKE_CASE : Any = image.to(dtype=image_embeds.dtype , device=snake_case__ )
__SCREAMING_SNAKE_CASE : Dict = self.movq.encode(snake_case__ )["latents"]
__SCREAMING_SNAKE_CASE : Optional[int] = latents.repeat_interleave(snake_case__ , dim=0 )
self.scheduler.set_timesteps(snake_case__ , device=snake_case__ )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = self.get_timesteps(snake_case__ , snake_case__ , snake_case__ )
__SCREAMING_SNAKE_CASE : List[str] = timesteps[:1].repeat(batch_size * num_images_per_prompt )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = downscale_height_and_width(snake_case__ , snake_case__ , self.movq_scale_factor )
__SCREAMING_SNAKE_CASE : Any = self.prepare_latents(
snake_case__ , snake_case__ , snake_case__ , snake_case__ , image_embeds.dtype , snake_case__ , snake_case__ )
for i, t in enumerate(self.progress_bar(snake_case__ ) ):
# expand the latents if we are doing classifier free guidance
__SCREAMING_SNAKE_CASE : int = torch.cat([latents] * 2 ) if do_classifier_free_guidance else latents
__SCREAMING_SNAKE_CASE : str = {"image_embeds": image_embeds}
__SCREAMING_SNAKE_CASE : Dict = self.unet(
sample=snake_case__ , timestep=snake_case__ , encoder_hidden_states=snake_case__ , added_cond_kwargs=snake_case__ , return_dict=snake_case__ , )[0]
if do_classifier_free_guidance:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = noise_pred.split(latents.shape[1] , dim=1 )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[int] = noise_pred.chunk(2 )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[int] = variance_pred.chunk(2 )
__SCREAMING_SNAKE_CASE : Optional[int] = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.cat([noise_pred, variance_pred_text] , dim=1 )
if not (
hasattr(self.scheduler.config , "variance_type" )
and self.scheduler.config.variance_type in ["learned", "learned_range"]
):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : str = noise_pred.split(latents.shape[1] , dim=1 )
# compute the previous noisy sample x_t -> x_t-1
__SCREAMING_SNAKE_CASE : List[Any] = self.scheduler.step(
snake_case__ , snake_case__ , snake_case__ , generator=snake_case__ , )[0]
# post-processing
__SCREAMING_SNAKE_CASE : int = self.movq.decode(snake_case__ , force_not_quantize=snake_case__ )["sample"]
if output_type not in ["pt", "np", "pil"]:
raise ValueError(f'Only the output types `pt`, `pil` and `np` are supported not output_type={output_type}' )
if output_type in ["np", "pil"]:
__SCREAMING_SNAKE_CASE : Any = image * 0.5 + 0.5
__SCREAMING_SNAKE_CASE : Union[str, Any] = image.clamp(0 , 1 )
__SCREAMING_SNAKE_CASE : List[str] = image.cpu().permute(0 , 2 , 3 , 1 ).float().numpy()
if output_type == "pil":
__SCREAMING_SNAKE_CASE : List[Any] = self.numpy_to_pil(snake_case__ )
if not return_dict:
return (image,)
return ImagePipelineOutput(images=snake_case__ )
| 211 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_torch_available,
)
lowerCAmelCase_ : Any = {
'''configuration_encodec''': [
'''ENCODEC_PRETRAINED_CONFIG_ARCHIVE_MAP''',
'''EncodecConfig''',
],
'''feature_extraction_encodec''': ['''EncodecFeatureExtractor'''],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
lowerCAmelCase_ : List[str] = [
'''ENCODEC_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''EncodecModel''',
'''EncodecPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_encodec import (
ENCODEC_PRETRAINED_CONFIG_ARCHIVE_MAP,
EncodecConfig,
)
from .feature_extraction_encodec import EncodecFeatureExtractor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_encodec import (
ENCODEC_PRETRAINED_MODEL_ARCHIVE_LIST,
EncodecModel,
EncodecPreTrainedModel,
)
else:
import sys
lowerCAmelCase_ : str = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 673 | 0 |
'''simple docstring'''
import sys
from collections.abc import Mapping
from typing import TYPE_CHECKING, Dict, Optional
import numpy as np
import pyarrow as pa
from .. import config
from ..utils.logging import get_logger
from ..utils.py_utils import map_nested
from .formatting import TensorFormatter
if TYPE_CHECKING:
import jax
import jaxlib
__a: List[str] = get_logger()
__a: Optional[dict] = None
class UpperCAmelCase ( TensorFormatter[Mapping, "jax.Array", Mapping] ):
'''simple docstring'''
def __init__( self , __lowerCAmelCase=None , __lowerCAmelCase=None , **__lowerCAmelCase ) -> int:
super().__init__(features=snake_case__ )
import jax
from jaxlib.xla_client import Device
if isinstance(snake_case__ , snake_case__ ):
raise ValueError(
F"""Expected {device} to be a `str` not {type(snake_case__ )}, as `jaxlib.xla_extension.Device` """
'''is not serializable neither with `pickle` nor with `dill`. Instead you can surround '''
'''the device with `str()` to get its string identifier that will be internally mapped '''
'''to the actual `jaxlib.xla_extension.Device`.''' )
lowercase__ : Optional[int] = device if isinstance(snake_case__ , snake_case__ ) else str(jax.devices()[0] )
# using global variable since `jaxlib.xla_extension.Device` is not serializable neither
# with `pickle` nor with `dill`, so we need to use a global variable instead
global DEVICE_MAPPING
if DEVICE_MAPPING is None:
lowercase__ : str = self._map_devices_to_str()
if self.device not in list(DEVICE_MAPPING.keys() ):
logger.warning(
F"""Device with string identifier {self.device} not listed among the available """
F"""devices: {list(DEVICE_MAPPING.keys() )}, so falling back to the default """
F"""device: {str(jax.devices()[0] )}.""" )
lowercase__ : Union[str, Any] = str(jax.devices()[0] )
lowercase__ : Dict = jnp_array_kwargs
@staticmethod
def _lowerCAmelCase( ) -> Dict[str, "jaxlib.xla_extension.Device"]:
import jax
return {str(snake_case__ ): device for device in jax.devices()}
def _lowerCAmelCase( self , __lowerCAmelCase ) -> Tuple:
import jax
import jax.numpy as jnp
if isinstance(snake_case__ , snake_case__ ) and column:
if all(
isinstance(snake_case__ , jax.Array ) and x.shape == column[0].shape and x.dtype == column[0].dtype for x in column ):
return jnp.stack(snake_case__ , axis=0 )
return column
def _lowerCAmelCase( self , __lowerCAmelCase ) -> Dict:
import jax
import jax.numpy as jnp
if isinstance(snake_case__ , (str, bytes, type(snake_case__ )) ):
return value
elif isinstance(snake_case__ , (np.character, np.ndarray) ) and np.issubdtype(value.dtype , np.character ):
return value.tolist()
lowercase__ : Dict = {}
if isinstance(snake_case__ , (np.number, np.ndarray) ) and np.issubdtype(value.dtype , np.integer ):
# the default int precision depends on the jax config
# see https://jax.readthedocs.io/en/latest/notebooks/Common_Gotchas_in_JAX.html#double-64bit-precision
if jax.config.jax_enable_xaa:
lowercase__ : List[str] = {'''dtype''': jnp.intaa}
else:
lowercase__ : Optional[int] = {'''dtype''': jnp.intaa}
elif isinstance(snake_case__ , (np.number, np.ndarray) ) and np.issubdtype(value.dtype , np.floating ):
lowercase__ : List[Any] = {'''dtype''': jnp.floataa}
elif config.PIL_AVAILABLE and "PIL" in sys.modules:
import PIL.Image
if isinstance(snake_case__ , PIL.Image.Image ):
lowercase__ : Tuple = np.asarray(snake_case__ )
# using global variable since `jaxlib.xla_extension.Device` is not serializable neither
# with `pickle` nor with `dill`, so we need to use a global variable instead
global DEVICE_MAPPING
if DEVICE_MAPPING is None:
lowercase__ : Optional[Any] = self._map_devices_to_str()
with jax.default_device(DEVICE_MAPPING[self.device] ):
# calling jnp.array on a np.ndarray does copy the data
# see https://github.com/google/jax/issues/4486
return jnp.array(snake_case__ , **{**default_dtype, **self.jnp_array_kwargs} )
def _lowerCAmelCase( self , __lowerCAmelCase ) -> Any:
import jax
# support for torch, tf, jax etc.
if config.TORCH_AVAILABLE and "torch" in sys.modules:
import torch
if isinstance(snake_case__ , torch.Tensor ):
return self._tensorize(data_struct.detach().cpu().numpy()[()] )
if hasattr(snake_case__ , '''__array__''' ) and not isinstance(snake_case__ , jax.Array ):
lowercase__ : Any = data_struct.__array__()
# support for nested types like struct of list of struct
if isinstance(snake_case__ , np.ndarray ):
if data_struct.dtype == object: # jax arrays cannot be instantied from an array of objects
return self._consolidate([self.recursive_tensorize(snake_case__ ) for substruct in data_struct] )
elif isinstance(snake_case__ , (list, tuple) ):
return self._consolidate([self.recursive_tensorize(snake_case__ ) for substruct in data_struct] )
return self._tensorize(snake_case__ )
def _lowerCAmelCase( self , __lowerCAmelCase ) -> int:
return map_nested(self._recursive_tensorize , snake_case__ , map_list=snake_case__ )
def _lowerCAmelCase( self , __lowerCAmelCase ) -> Mapping:
lowercase__ : List[Any] = self.numpy_arrow_extractor().extract_row(snake_case__ )
lowercase__ : Optional[int] = self.python_features_decoder.decode_row(snake_case__ )
return self.recursive_tensorize(snake_case__ )
def _lowerCAmelCase( self , __lowerCAmelCase ) -> "jax.Array":
lowercase__ : Any = self.numpy_arrow_extractor().extract_column(snake_case__ )
lowercase__ : Any = self.python_features_decoder.decode_column(snake_case__ , pa_table.column_names[0] )
lowercase__ : Any = self.recursive_tensorize(snake_case__ )
lowercase__ : Optional[int] = self._consolidate(snake_case__ )
return column
def _lowerCAmelCase( self , __lowerCAmelCase ) -> Mapping:
lowercase__ : Dict = self.numpy_arrow_extractor().extract_batch(snake_case__ )
lowercase__ : str = self.python_features_decoder.decode_batch(snake_case__ )
lowercase__ : List[Any] = self.recursive_tensorize(snake_case__ )
for column_name in batch:
lowercase__ : Any = self._consolidate(batch[column_name] )
return batch
| 152 |
"""simple docstring"""
from ...utils import (
OptionalDependencyNotAvailable,
is_torch_available,
is_transformers_available,
is_transformers_version,
)
try:
if not (is_transformers_available() and is_torch_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ...utils.dummy_torch_and_transformers_objects import (
ImageTextPipelineOutput,
UniDiffuserPipeline,
)
else:
from .modeling_text_decoder import UniDiffuserTextDecoder
from .modeling_uvit import UniDiffuserModel, UTransformeraDModel
from .pipeline_unidiffuser import ImageTextPipelineOutput, UniDiffuserPipeline
| 673 | 0 |
'''simple docstring'''
from __future__ import annotations
import time
__lowercase : Dict = list[tuple[int, int]]
__lowercase : List[str] = [
[0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 0, 0, 0], # 0 are free path whereas 1's are obstacles
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 0, 0, 0],
[1, 0, 1, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 1, 0, 0],
]
__lowercase : Union[str, Any] = [[-1, 0], [0, -1], [1, 0], [0, 1]] # up, left, down, right
class __UpperCamelCase :
def __init__( self , __a , __a , __a , __a , __a ):
'''simple docstring'''
__a : str = pos_x
__a : Dict = pos_y
__a : str = (pos_y, pos_x)
__a : Optional[int] = goal_x
__a : Union[str, Any] = goal_y
__a : Union[str, Any] = parent
class __UpperCamelCase :
def __init__( self , __a , __a ):
'''simple docstring'''
__a : Any = Node(start[1] , start[0] , goal[1] , goal[0] , snake_case__ )
__a : Optional[Any] = Node(goal[1] , goal[0] , goal[1] , goal[0] , snake_case__ )
__a : List[str] = [self.start]
__a : Any = False
def __UpperCAmelCase ( self ):
'''simple docstring'''
while self.node_queue:
__a : str = self.node_queue.pop(0 )
if current_node.pos == self.target.pos:
__a : Optional[int] = True
return self.retrace_path(snake_case__ )
__a : Tuple = self.get_successors(snake_case__ )
for node in successors:
self.node_queue.append(snake_case__ )
if not self.reached:
return [self.start.pos]
return None
def __UpperCAmelCase ( self , __a ):
'''simple docstring'''
__a : int = []
for action in delta:
__a : Any = parent.pos_x + action[1]
__a : Any = parent.pos_y + action[0]
if not (0 <= pos_x <= len(grid[0] ) - 1 and 0 <= pos_y <= len(snake_case__ ) - 1):
continue
if grid[pos_y][pos_x] != 0:
continue
successors.append(
Node(snake_case__ , snake_case__ , self.target.pos_y , self.target.pos_x , snake_case__ ) )
return successors
def __UpperCAmelCase ( self , __a ):
'''simple docstring'''
__a : Union[str, Any] = node
__a : Tuple = []
while current_node is not None:
path.append((current_node.pos_y, current_node.pos_x) )
__a : Optional[int] = current_node.parent
path.reverse()
return path
class __UpperCamelCase :
def __init__( self , __a , __a ):
'''simple docstring'''
__a : int = BreadthFirstSearch(snake_case__ , snake_case__ )
__a : Optional[int] = BreadthFirstSearch(snake_case__ , snake_case__ )
__a : Tuple = False
def __UpperCAmelCase ( self ):
'''simple docstring'''
while self.fwd_bfs.node_queue or self.bwd_bfs.node_queue:
__a : Dict = self.fwd_bfs.node_queue.pop(0 )
__a : Optional[Any] = self.bwd_bfs.node_queue.pop(0 )
if current_bwd_node.pos == current_fwd_node.pos:
__a : Optional[Any] = True
return self.retrace_bidirectional_path(
snake_case__ , snake_case__ )
__a : Any = current_bwd_node
__a : Union[str, Any] = current_fwd_node
__a : Dict = {
self.fwd_bfs: self.fwd_bfs.get_successors(snake_case__ ),
self.bwd_bfs: self.bwd_bfs.get_successors(snake_case__ ),
}
for bfs in [self.fwd_bfs, self.bwd_bfs]:
for node in successors[bfs]:
bfs.node_queue.append(snake_case__ )
if not self.reached:
return [self.fwd_bfs.start.pos]
return None
def __UpperCAmelCase ( self , __a , __a ):
'''simple docstring'''
__a : List[str] = self.fwd_bfs.retrace_path(snake_case__ )
__a : str = self.bwd_bfs.retrace_path(snake_case__ )
bwd_path.pop()
bwd_path.reverse()
__a : Tuple = fwd_path + bwd_path
return path
if __name__ == "__main__":
# all coordinates are given in format [y,x]
import doctest
doctest.testmod()
__lowercase : str = (0, 0)
__lowercase : Dict = (len(grid) - 1, len(grid[0]) - 1)
for elem in grid:
print(elem)
__lowercase : Optional[int] = time.time()
__lowercase : str = BreadthFirstSearch(init, goal)
__lowercase : Dict = bfs.search()
__lowercase : List[Any] = time.time() - start_bfs_time
print('Unidirectional BFS computation time : ', bfs_time)
__lowercase : Dict = time.time()
__lowercase : List[Any] = BidirectionalBreadthFirstSearch(init, goal)
__lowercase : Optional[Any] = bd_bfs.search()
__lowercase : Any = time.time() - start_bd_bfs_time
print('Bidirectional BFS computation time : ', bd_bfs_time)
| 476 |
"""simple docstring"""
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
UNetaDConditionModel,
VideoToVideoSDPipeline,
)
from diffusers.utils import floats_tensor, is_xformers_available, skip_mps
from diffusers.utils.testing_utils import enable_full_determinism, slow, torch_device
from ..pipeline_params import (
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineTesterMixin
enable_full_determinism()
@skip_mps
class UpperCamelCase_ ( a_ , unittest.TestCase ):
_A : str = VideoToVideoSDPipeline
_A : List[str] = TEXT_GUIDED_IMAGE_VARIATION_PARAMS.union({'video'} ) - {'image', 'width', 'height'}
_A : int = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({'video'} ) - {'image'}
_A : int = PipelineTesterMixin.required_optional_params - {'latents'}
_A : List[str] = False
# No `output_type`.
_A : Any = frozenset(
[
'num_inference_steps',
'generator',
'latents',
'return_dict',
'callback',
'callback_steps',
] )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
torch.manual_seed(0 )
UpperCAmelCase = UNetaDConditionModel(
block_out_channels=(32, 64, 64, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=("""CrossAttnDownBlock3D""", """CrossAttnDownBlock3D""", """CrossAttnDownBlock3D""", """DownBlock3D""") , up_block_types=("""UpBlock3D""", """CrossAttnUpBlock3D""", """CrossAttnUpBlock3D""", """CrossAttnUpBlock3D""") , cross_attention_dim=32 , attention_head_dim=4 , )
UpperCAmelCase = DDIMScheduler(
beta_start=0.00_085 , beta_end=0.012 , beta_schedule="""scaled_linear""" , clip_sample=snake_case__ , set_alpha_to_one=snake_case__ , )
torch.manual_seed(0 )
UpperCAmelCase = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] , up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] , latent_channels=4 , sample_size=1_28 , )
torch.manual_seed(0 )
UpperCAmelCase = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=10_00 , hidden_act="""gelu""" , projection_dim=5_12 , )
UpperCAmelCase = CLIPTextModel(snake_case__ )
UpperCAmelCase = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
UpperCAmelCase = {
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
"""text_encoder""": text_encoder,
"""tokenizer""": tokenizer,
}
return components
def UpperCamelCase_ ( self , snake_case__ , snake_case__=0 ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = floats_tensor((1, 3, 3, 32, 32) , rng=random.Random(snake_case__ ) ).to(snake_case__ )
if str(snake_case__ ).startswith("""mps""" ):
UpperCAmelCase = torch.manual_seed(snake_case__ )
else:
UpperCAmelCase = torch.Generator(device=snake_case__ ).manual_seed(snake_case__ )
UpperCAmelCase = {
"""prompt""": """A painting of a squirrel eating a burger""",
"""video""": video,
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 6.0,
"""output_type""": """pt""",
}
return inputs
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
UpperCAmelCase = """cpu""" # ensure determinism for the device-dependent torch.Generator
UpperCAmelCase = self.get_dummy_components()
UpperCAmelCase = VideoToVideoSDPipeline(**snake_case__ )
UpperCAmelCase = sd_pipe.to(snake_case__ )
sd_pipe.set_progress_bar_config(disable=snake_case__ )
UpperCAmelCase = self.get_dummy_inputs(snake_case__ )
UpperCAmelCase = """np"""
UpperCAmelCase = sd_pipe(**snake_case__ ).frames
UpperCAmelCase = frames[0][-3:, -3:, -1]
assert frames[0].shape == (32, 32, 3)
UpperCAmelCase = np.array([1_06, 1_17, 1_13, 1_74, 1_37, 1_12, 1_48, 1_51, 1_31] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(
torch_device != """cuda""" or not is_xformers_available() , reason="""XFormers attention is only available with CUDA and `xformers` installed""" , )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
self._test_xformers_attention_forwardGenerator_pass(test_mean_pixel_difference=snake_case__ , expected_max_diff=5e-3 )
@unittest.skip(reason="""Batching needs to be properly figured out first for this pipeline.""" )
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
pass
@unittest.skip(reason="""Batching needs to be properly figured out first for this pipeline.""" )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
pass
@unittest.skip(reason="""`num_images_per_prompt` argument is not supported for this pipeline.""" )
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
pass
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return super().test_progress_bar()
@slow
@skip_mps
class UpperCamelCase_ ( unittest.TestCase ):
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = VideoToVideoSDPipeline.from_pretrained("""cerspense/zeroscope_v2_XL""" , torch_dtype=torch.floataa )
pipe.enable_model_cpu_offload()
# 10 frames
UpperCAmelCase = torch.Generator(device="""cpu""" ).manual_seed(0 )
UpperCAmelCase = torch.randn((1, 10, 3, 10_24, 5_76) , generator=snake_case__ )
UpperCAmelCase = video.to("""cuda""" )
UpperCAmelCase = """Spiderman is surfing"""
UpperCAmelCase = pipe(snake_case__ , video=snake_case__ , generator=snake_case__ , num_inference_steps=3 , output_type="""pt""" ).frames
UpperCAmelCase = np.array([-1.0_458_984, -1.1_279_297, -0.9_663_086, -0.91_503_906, -0.75_097_656] )
assert np.abs(video_frames.cpu().numpy()[0, 0, 0, 0, -5:] - expected_array ).sum() < 1e-2
| 673 | 0 |
import logging
import os
from typing import List, Tuple
import numpy as np
import psutil
import torch
import torch.distributed as dist
from transformers import RagRetriever
A_ = logging.getLogger(__name__)
class snake_case ( a_ ):
'''simple docstring'''
def __init__( self : Any , lowerCAmelCase_ : Tuple , lowerCAmelCase_ : Dict , lowerCAmelCase_ : Optional[Any] , lowerCAmelCase_ : Union[str, Any]=None ) -> Union[str, Any]:
"""simple docstring"""
super().__init__(
snake_case__ , question_encoder_tokenizer=snake_case__ , generator_tokenizer=snake_case__ , index=snake_case__ , init_retrieval=snake_case__ , )
SCREAMING_SNAKE_CASE_ = None
def _lowercase ( self : Optional[int] , lowerCAmelCase_ : Optional[int] ) -> Tuple:
"""simple docstring"""
logger.info('''initializing retrieval''' )
# initializing a separate process group for retrieval as the default
# nccl backend doesn't support gather/scatter operations while gloo
# is too slow to replace nccl for the core gpu communication
if dist.is_initialized():
logger.info('''dist initialized''' )
# needs to be set manually
SCREAMING_SNAKE_CASE_ = self._infer_socket_ifname()
# avoid clash with the NCCL port
SCREAMING_SNAKE_CASE_ = str(distributed_port + 1 )
SCREAMING_SNAKE_CASE_ = dist.new_group(ranks=snake_case__ , backend='''gloo''' )
# initialize retriever only on the main worker
if not dist.is_initialized() or self._is_main():
logger.info('''dist not initialized / main''' )
self.index.init_index()
# all processes wait untill the retriever is initialized by the main process
if dist.is_initialized():
torch.distributed.barrier(group=self.process_group )
def _lowercase ( self : int ) -> Optional[int]:
"""simple docstring"""
return dist.get_rank(group=self.process_group ) == 0
def _lowercase ( self : Optional[Any] , lowerCAmelCase_ : str , lowerCAmelCase_ : int , lowerCAmelCase_ : Optional[Any]=torch.floataa ) -> Union[str, Any]:
"""simple docstring"""
SCREAMING_SNAKE_CASE_ = torch.empty(snake_case__ , dtype=snake_case__ )
dist.scatter(snake_case__ , src=0 , scatter_list=snake_case__ , group=self.process_group )
return target_tensor
def _lowercase ( self : Union[str, Any] ) -> List[Any]:
"""simple docstring"""
SCREAMING_SNAKE_CASE_ = psutil.net_if_addrs()
# a hacky way to deal with varying network interface names
SCREAMING_SNAKE_CASE_ = next((addr for addr in addrs if addr.startswith('''e''' )) , snake_case__ )
return ifname
def _lowercase ( self : Optional[Any] , lowerCAmelCase_ : Union[str, Any] , lowerCAmelCase_ : Tuple ) -> Tuple[np.ndarray, List[dict]]:
"""simple docstring"""
if not dist.is_initialized():
SCREAMING_SNAKE_CASE_ ,SCREAMING_SNAKE_CASE_ = self._main_retrieve(snake_case__ , snake_case__ )
return retrieved_doc_embeds, doc_ids, self.index.get_doc_dicts(snake_case__ )
# distributed training
SCREAMING_SNAKE_CASE_ = dist.get_world_size(group=self.process_group )
# gather logic
SCREAMING_SNAKE_CASE_ = None
if self._is_main():
SCREAMING_SNAKE_CASE_ = [torch.empty(question_hidden_states.shape , dtype=torch.floataa ) for _ in range(snake_case__ )]
dist.gather(torch.tensor(snake_case__ ) , dst=0 , gather_list=snake_case__ , group=self.process_group )
# scatter logic
SCREAMING_SNAKE_CASE_ = question_hidden_states.shape[0]
SCREAMING_SNAKE_CASE_ = []
SCREAMING_SNAKE_CASE_ = []
if self._is_main():
assert len(snake_case__ ) == world_size
SCREAMING_SNAKE_CASE_ ,SCREAMING_SNAKE_CASE_ = self._main_retrieve(torch.cat(snake_case__ ).numpy() , snake_case__ )
SCREAMING_SNAKE_CASE_ ,SCREAMING_SNAKE_CASE_ = torch.tensor(snake_case__ ), torch.tensor(snake_case__ )
SCREAMING_SNAKE_CASE_ = self._chunk_tensor(snake_case__ , snake_case__ )
SCREAMING_SNAKE_CASE_ = self._chunk_tensor(snake_case__ , snake_case__ )
SCREAMING_SNAKE_CASE_ = self._scattered(snake_case__ , [n_queries, n_docs] , target_type=torch.intaa )
SCREAMING_SNAKE_CASE_ = self._scattered(snake_case__ , [n_queries, n_docs, question_hidden_states.shape[1]] )
return retrieved_doc_embeds.numpy(), doc_ids.numpy(), self.index.get_doc_dicts(snake_case__ )
| 393 |
"""simple docstring"""
import functools
import operator
from ...configuration_utils import PretrainedConfig
from ...utils import logging
lowerCAmelCase_ : int = logging.get_logger(__name__)
lowerCAmelCase_ : Any = {
'''facebook/wav2vec2-base-960h''': '''https://huggingface.co/facebook/wav2vec2-base-960h/resolve/main/config.json''',
# See all Wav2Vec2 models at https://huggingface.co/models?filter=wav2vec2
}
class UpperCamelCase_ ( a_ ):
_A : int = 'wav2vec2'
def __init__( self , snake_case__=32 , snake_case__=7_68 , snake_case__=12 , snake_case__=12 , snake_case__=30_72 , snake_case__="gelu" , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.0 , snake_case__=0.0 , snake_case__=0.1 , snake_case__=0.1 , snake_case__=0.02 , snake_case__=1e-5 , snake_case__="group" , snake_case__="gelu" , snake_case__=(5_12, 5_12, 5_12, 5_12, 5_12, 5_12, 5_12) , snake_case__=(5, 2, 2, 2, 2, 2, 2) , snake_case__=(10, 3, 3, 3, 3, 2, 2) , snake_case__=False , snake_case__=1_28 , snake_case__=16 , snake_case__=False , snake_case__=True , snake_case__=0.05 , snake_case__=10 , snake_case__=2 , snake_case__=0.0 , snake_case__=10 , snake_case__=0 , snake_case__=3_20 , snake_case__=2 , snake_case__=0.1 , snake_case__=1_00 , snake_case__=2_56 , snake_case__=2_56 , snake_case__=0.1 , snake_case__="sum" , snake_case__=False , snake_case__=False , snake_case__=2_56 , snake_case__=(5_12, 5_12, 5_12, 5_12, 15_00) , snake_case__=(5, 3, 3, 1, 1) , snake_case__=(1, 2, 3, 1, 1) , snake_case__=5_12 , snake_case__=0 , snake_case__=1 , snake_case__=2 , snake_case__=False , snake_case__=3 , snake_case__=2 , snake_case__=3 , snake_case__=None , snake_case__=None , **snake_case__ , ) -> Optional[Any]:
"""simple docstring"""
super().__init__(**snake_case__ , pad_token_id=snake_case__ , bos_token_id=snake_case__ , eos_token_id=snake_case__ )
UpperCAmelCase = hidden_size
UpperCAmelCase = feat_extract_norm
UpperCAmelCase = feat_extract_activation
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = conv_bias
UpperCAmelCase = num_conv_pos_embeddings
UpperCAmelCase = num_conv_pos_embedding_groups
UpperCAmelCase = len(self.conv_dim )
UpperCAmelCase = num_hidden_layers
UpperCAmelCase = intermediate_size
UpperCAmelCase = hidden_act
UpperCAmelCase = num_attention_heads
UpperCAmelCase = hidden_dropout
UpperCAmelCase = attention_dropout
UpperCAmelCase = activation_dropout
UpperCAmelCase = feat_proj_dropout
UpperCAmelCase = final_dropout
UpperCAmelCase = layerdrop
UpperCAmelCase = layer_norm_eps
UpperCAmelCase = initializer_range
UpperCAmelCase = vocab_size
UpperCAmelCase = do_stable_layer_norm
UpperCAmelCase = use_weighted_layer_sum
if (
(len(self.conv_stride ) != self.num_feat_extract_layers)
or (len(self.conv_kernel ) != self.num_feat_extract_layers)
or (len(self.conv_dim ) != self.num_feat_extract_layers)
):
raise ValueError(
"""Configuration for convolutional layers is incorrect. It is required that `len(config.conv_dim)` =="""
""" `len(config.conv_stride)` == `len(config.conv_kernel)`, but is `len(config.conv_dim) ="""
f''' {len(self.conv_dim )}`, `len(config.conv_stride) = {len(self.conv_stride )}`,'''
f''' `len(config.conv_kernel) = {len(self.conv_kernel )}`.''' )
# fine-tuning config parameters for SpecAugment: https://arxiv.org/abs/1904.08779
UpperCAmelCase = apply_spec_augment
UpperCAmelCase = mask_time_prob
UpperCAmelCase = mask_time_length
UpperCAmelCase = mask_time_min_masks
UpperCAmelCase = mask_feature_prob
UpperCAmelCase = mask_feature_length
UpperCAmelCase = mask_feature_min_masks
# parameters for pretraining with codevector quantized representations
UpperCAmelCase = num_codevectors_per_group
UpperCAmelCase = num_codevector_groups
UpperCAmelCase = contrastive_logits_temperature
UpperCAmelCase = feat_quantizer_dropout
UpperCAmelCase = num_negatives
UpperCAmelCase = codevector_dim
UpperCAmelCase = proj_codevector_dim
UpperCAmelCase = diversity_loss_weight
# ctc loss
UpperCAmelCase = ctc_loss_reduction
UpperCAmelCase = ctc_zero_infinity
# adapter
UpperCAmelCase = add_adapter
UpperCAmelCase = adapter_kernel_size
UpperCAmelCase = adapter_stride
UpperCAmelCase = num_adapter_layers
UpperCAmelCase = output_hidden_size or hidden_size
UpperCAmelCase = adapter_attn_dim
# SequenceClassification-specific parameter. Feel free to ignore for other classes.
UpperCAmelCase = classifier_proj_size
# XVector-specific parameters. Feel free to ignore for other classes.
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = list(snake_case__ )
UpperCAmelCase = xvector_output_dim
@property
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return functools.reduce(operator.mul , self.conv_stride , 1 )
| 673 | 0 |
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
__UpperCAmelCase = logging.get_logger(__name__)
__UpperCAmelCase = {
'''roberta-base''': '''https://huggingface.co/roberta-base/resolve/main/config.json''',
'''roberta-large''': '''https://huggingface.co/roberta-large/resolve/main/config.json''',
'''roberta-large-mnli''': '''https://huggingface.co/roberta-large-mnli/resolve/main/config.json''',
'''distilroberta-base''': '''https://huggingface.co/distilroberta-base/resolve/main/config.json''',
'''roberta-base-openai-detector''': '''https://huggingface.co/roberta-base-openai-detector/resolve/main/config.json''',
'''roberta-large-openai-detector''': '''https://huggingface.co/roberta-large-openai-detector/resolve/main/config.json''',
}
class lowerCamelCase (a_ ):
'''simple docstring'''
_snake_case : Optional[int] = 'roberta'
def __init__( self , _UpperCamelCase=5_0_2_6_5 , _UpperCamelCase=7_6_8 , _UpperCamelCase=1_2 , _UpperCamelCase=1_2 , _UpperCamelCase=3_0_7_2 , _UpperCamelCase="gelu" , _UpperCamelCase=0.1 , _UpperCamelCase=0.1 , _UpperCamelCase=5_1_2 , _UpperCamelCase=2 , _UpperCamelCase=0.02 , _UpperCamelCase=1E-12 , _UpperCamelCase=1 , _UpperCamelCase=0 , _UpperCamelCase=2 , _UpperCamelCase="absolute" , _UpperCamelCase=True , _UpperCamelCase=None , **_UpperCamelCase , ) -> List[str]:
super().__init__(pad_token_id=snake_case__ , bos_token_id=snake_case__ , eos_token_id=snake_case__ , **snake_case__ )
UpperCAmelCase_ : str = vocab_size
UpperCAmelCase_ : Any = hidden_size
UpperCAmelCase_ : Optional[Any] = num_hidden_layers
UpperCAmelCase_ : List[str] = num_attention_heads
UpperCAmelCase_ : Tuple = hidden_act
UpperCAmelCase_ : int = intermediate_size
UpperCAmelCase_ : Tuple = hidden_dropout_prob
UpperCAmelCase_ : Any = attention_probs_dropout_prob
UpperCAmelCase_ : List[Any] = max_position_embeddings
UpperCAmelCase_ : int = type_vocab_size
UpperCAmelCase_ : Tuple = initializer_range
UpperCAmelCase_ : Optional[int] = layer_norm_eps
UpperCAmelCase_ : List[Any] = position_embedding_type
UpperCAmelCase_ : List[Any] = use_cache
UpperCAmelCase_ : Union[str, Any] = classifier_dropout
class lowerCamelCase (a_ ):
'''simple docstring'''
@property
def __UpperCAmelCase ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
UpperCAmelCase_ : Optional[Any] = {0: 'batch', 1: 'choice', 2: 'sequence'}
else:
UpperCAmelCase_ : Tuple = {0: 'batch', 1: 'sequence'}
return OrderedDict(
[
('input_ids', dynamic_axis),
('attention_mask', dynamic_axis),
] )
| 406 |
"""simple docstring"""
import dataclasses
import json
import sys
import types
from argparse import ArgumentDefaultsHelpFormatter, ArgumentParser, ArgumentTypeError
from copy import copy
from enum import Enum
from inspect import isclass
from pathlib import Path
from typing import Any, Callable, Dict, Iterable, List, Literal, NewType, Optional, Tuple, Union, get_type_hints
import yaml
lowerCAmelCase_ : Optional[Any] = NewType('''DataClass''', Any)
lowerCAmelCase_ : Any = NewType('''DataClassType''', Any)
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
if isinstance(lowerCAmelCase , lowerCAmelCase ):
return v
if v.lower() in ("yes", "true", "t", "y", "1"):
return True
elif v.lower() in ("no", "false", "f", "n", "0"):
return False
else:
raise ArgumentTypeError(
F'''Truthy value expected: got {v} but expected one of yes/no, true/false, t/f, y/n, 1/0 (case insensitive).''' )
def _lowerCAmelCase ( lowerCAmelCase ):
'''simple docstring'''
UpperCAmelCase = {str(lowerCAmelCase ): choice for choice in choices}
return lambda lowerCAmelCase : str_to_choice.get(lowerCAmelCase , lowerCAmelCase )
def _lowerCAmelCase ( *,
lowerCAmelCase = None , lowerCAmelCase = None , lowerCAmelCase = dataclasses.MISSING , lowerCAmelCase = dataclasses.MISSING , lowerCAmelCase = None , **lowerCAmelCase , ):
'''simple docstring'''
if metadata is None:
# Important, don't use as default param in function signature because dict is mutable and shared across function calls
UpperCAmelCase = {}
if aliases is not None:
UpperCAmelCase = aliases
if help is not None:
UpperCAmelCase = help
return dataclasses.field(metadata=lowerCAmelCase , default=lowerCAmelCase , default_factory=lowerCAmelCase , **lowerCAmelCase )
class UpperCamelCase_ ( a_ ):
_A : Iterable[DataClassType]
def __init__( self , snake_case__ , **snake_case__ ) -> List[str]:
"""simple docstring"""
if "formatter_class" not in kwargs:
UpperCAmelCase = ArgumentDefaultsHelpFormatter
super().__init__(**snake_case__ )
if dataclasses.is_dataclass(snake_case__ ):
UpperCAmelCase = [dataclass_types]
UpperCAmelCase = list(snake_case__ )
for dtype in self.dataclass_types:
self._add_dataclass_arguments(snake_case__ )
@staticmethod
def UpperCamelCase_ ( snake_case__ , snake_case__ ) -> str:
"""simple docstring"""
UpperCAmelCase = f'''--{field.name}'''
UpperCAmelCase = field.metadata.copy()
# field.metadata is not used at all by Data Classes,
# it is provided as a third-party extension mechanism.
if isinstance(field.type , snake_case__ ):
raise RuntimeError(
"""Unresolved type detected, which should have been done with the help of """
"""`typing.get_type_hints` method by default""" )
UpperCAmelCase = kwargs.pop("""aliases""" , [] )
if isinstance(snake_case__ , snake_case__ ):
UpperCAmelCase = [aliases]
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
if origin_type is Union or (hasattr(snake_case__ , """UnionType""" ) and isinstance(snake_case__ , types.UnionType )):
if str not in field.type.__args__ and (
len(field.type.__args__ ) != 2 or type(snake_case__ ) not in field.type.__args__
):
raise ValueError(
"""Only `Union[X, NoneType]` (i.e., `Optional[X]`) is allowed for `Union` because"""
""" the argument parser only supports one type per argument."""
f''' Problem encountered in field \'{field.name}\'.''' )
if type(snake_case__ ) not in field.type.__args__:
# filter `str` in Union
UpperCAmelCase = field.type.__args__[0] if field.type.__args__[1] == str else field.type.__args__[1]
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
elif bool not in field.type.__args__:
# filter `NoneType` in Union (except for `Union[bool, NoneType]`)
UpperCAmelCase = (
field.type.__args__[0] if isinstance(snake_case__ , field.type.__args__[1] ) else field.type.__args__[1]
)
UpperCAmelCase = getattr(field.type , """__origin__""" , field.type )
# A variable to store kwargs for a boolean field, if needed
# so that we can init a `no_*` complement argument (see below)
UpperCAmelCase = {}
if origin_type is Literal or (isinstance(field.type , snake_case__ ) and issubclass(field.type , snake_case__ )):
if origin_type is Literal:
UpperCAmelCase = field.type.__args__
else:
UpperCAmelCase = [x.value for x in field.type]
UpperCAmelCase = make_choice_type_function(kwargs["""choices"""] )
if field.default is not dataclasses.MISSING:
UpperCAmelCase = field.default
else:
UpperCAmelCase = True
elif field.type is bool or field.type == Optional[bool]:
# Copy the currect kwargs to use to instantiate a `no_*` complement argument below.
# We do not initialize it here because the `no_*` alternative must be instantiated after the real argument
UpperCAmelCase = copy(snake_case__ )
# Hack because type=bool in argparse does not behave as we want.
UpperCAmelCase = string_to_bool
if field.type is bool or (field.default is not None and field.default is not dataclasses.MISSING):
# Default value is False if we have no default when of type bool.
UpperCAmelCase = False if field.default is dataclasses.MISSING else field.default
# This is the value that will get picked if we don't include --field_name in any way
UpperCAmelCase = default
# This tells argparse we accept 0 or 1 value after --field_name
UpperCAmelCase = """?"""
# This is the value that will get picked if we do --field_name (without value)
UpperCAmelCase = True
elif isclass(snake_case__ ) and issubclass(snake_case__ , snake_case__ ):
UpperCAmelCase = field.type.__args__[0]
UpperCAmelCase = """+"""
if field.default_factory is not dataclasses.MISSING:
UpperCAmelCase = field.default_factory()
elif field.default is dataclasses.MISSING:
UpperCAmelCase = True
else:
UpperCAmelCase = field.type
if field.default is not dataclasses.MISSING:
UpperCAmelCase = field.default
elif field.default_factory is not dataclasses.MISSING:
UpperCAmelCase = field.default_factory()
else:
UpperCAmelCase = True
parser.add_argument(snake_case__ , *snake_case__ , **snake_case__ )
# Add a complement `no_*` argument for a boolean field AFTER the initial field has already been added.
# Order is important for arguments with the same destination!
# We use a copy of earlier kwargs because the original kwargs have changed a lot before reaching down
# here and we do not need those changes/additional keys.
if field.default is True and (field.type is bool or field.type == Optional[bool]):
UpperCAmelCase = False
parser.add_argument(f'''--no_{field.name}''' , action="""store_false""" , dest=field.name , **snake_case__ )
def UpperCamelCase_ ( self , snake_case__ ) -> Any:
"""simple docstring"""
if hasattr(snake_case__ , """_argument_group_name""" ):
UpperCAmelCase = self.add_argument_group(dtype._argument_group_name )
else:
UpperCAmelCase = self
try:
UpperCAmelCase = get_type_hints(snake_case__ )
except NameError:
raise RuntimeError(
f'''Type resolution failed for {dtype}. Try declaring the class in global scope or '''
"""removing line of `from __future__ import annotations` which opts in Postponed """
"""Evaluation of Annotations (PEP 563)""" )
except TypeError as ex:
# Remove this block when we drop Python 3.9 support
if sys.version_info[:2] < (3, 10) and "unsupported operand type(s) for |" in str(snake_case__ ):
UpperCAmelCase = """.""".join(map(snake_case__ , sys.version_info[:3] ) )
raise RuntimeError(
f'''Type resolution failed for {dtype} on Python {python_version}. Try removing '''
"""line of `from __future__ import annotations` which opts in union types as """
"""`X | Y` (PEP 604) via Postponed Evaluation of Annotations (PEP 563). To """
"""support Python versions that lower than 3.10, you need to use """
"""`typing.Union[X, Y]` instead of `X | Y` and `typing.Optional[X]` instead of """
"""`X | None`.""" ) from ex
raise
for field in dataclasses.fields(snake_case__ ):
if not field.init:
continue
UpperCAmelCase = type_hints[field.name]
self._parse_dataclass_field(snake_case__ , snake_case__ )
def UpperCamelCase_ ( self , snake_case__=None , snake_case__=False , snake_case__=True , snake_case__=None , snake_case__=None , ) -> Tuple[DataClass, ...]:
"""simple docstring"""
if args_file_flag or args_filename or (look_for_args_file and len(sys.argv )):
UpperCAmelCase = []
if args_filename:
args_files.append(Path(snake_case__ ) )
elif look_for_args_file and len(sys.argv ):
args_files.append(Path(sys.argv[0] ).with_suffix(""".args""" ) )
# args files specified via command line flag should overwrite default args files so we add them last
if args_file_flag:
# Create special parser just to extract the args_file_flag values
UpperCAmelCase = ArgumentParser()
args_file_parser.add_argument(snake_case__ , type=snake_case__ , action="""append""" )
# Use only remaining args for further parsing (remove the args_file_flag)
UpperCAmelCase , UpperCAmelCase = args_file_parser.parse_known_args(args=snake_case__ )
UpperCAmelCase = vars(snake_case__ ).get(args_file_flag.lstrip("""-""" ) , snake_case__ )
if cmd_args_file_paths:
args_files.extend([Path(snake_case__ ) for p in cmd_args_file_paths] )
UpperCAmelCase = []
for args_file in args_files:
if args_file.exists():
file_args += args_file.read_text().split()
# in case of duplicate arguments the last one has precedence
# args specified via the command line should overwrite args from files, so we add them last
UpperCAmelCase = file_args + args if args is not None else file_args + sys.argv[1:]
UpperCAmelCase , UpperCAmelCase = self.parse_known_args(args=snake_case__ )
UpperCAmelCase = []
for dtype in self.dataclass_types:
UpperCAmelCase = {f.name for f in dataclasses.fields(snake_case__ ) if f.init}
UpperCAmelCase = {k: v for k, v in vars(snake_case__ ).items() if k in keys}
for k in keys:
delattr(snake_case__ , snake_case__ )
UpperCAmelCase = dtype(**snake_case__ )
outputs.append(snake_case__ )
if len(namespace.__dict__ ) > 0:
# additional namespace.
outputs.append(snake_case__ )
if return_remaining_strings:
return (*outputs, remaining_args)
else:
if remaining_args:
raise ValueError(f'''Some specified arguments are not used by the HfArgumentParser: {remaining_args}''' )
return (*outputs,)
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
UpperCAmelCase = set(args.keys() )
UpperCAmelCase = []
for dtype in self.dataclass_types:
UpperCAmelCase = {f.name for f in dataclasses.fields(snake_case__ ) if f.init}
UpperCAmelCase = {k: v for k, v in args.items() if k in keys}
unused_keys.difference_update(inputs.keys() )
UpperCAmelCase = dtype(**snake_case__ )
outputs.append(snake_case__ )
if not allow_extra_keys and unused_keys:
raise ValueError(f'''Some keys are not used by the HfArgumentParser: {sorted(snake_case__ )}''' )
return tuple(snake_case__ )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
with open(Path(snake_case__ ) , encoding="""utf-8""" ) as open_json_file:
UpperCAmelCase = json.loads(open_json_file.read() )
UpperCAmelCase = self.parse_dict(snake_case__ , allow_extra_keys=snake_case__ )
return tuple(snake_case__ )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ = False ) -> Tuple[DataClass, ...]:
"""simple docstring"""
UpperCAmelCase = self.parse_dict(yaml.safe_load(Path(snake_case__ ).read_text() ) , allow_extra_keys=snake_case__ )
return tuple(snake_case__ )
| 673 | 0 |
"""simple docstring"""
import numpy
class lowerCAmelCase__ :
def __init__( self , UpperCamelCase__ , UpperCamelCase__ ):
'''simple docstring'''
A__ = input_array
# Random initial weights are assigned where first argument is the
# number of nodes in previous layer and second argument is the
# number of nodes in the next layer.
# Random initial weights are assigned.
# self.input_array.shape[1] is used to represent number of nodes in input layer.
# First hidden layer consists of 4 nodes.
A__ = numpy.random.rand(
self.input_array.shape[1] , 4 )
# Random initial values for the first hidden layer.
# First hidden layer has 4 nodes.
# Second hidden layer has 3 nodes.
A__ = numpy.random.rand(
4 , 3 )
# Random initial values for the second hidden layer.
# Second hidden layer has 3 nodes.
# Output layer has 1 node.
A__ = numpy.random.rand(3 , 1 )
# Real output values provided.
A__ = output_array
# Predicted output values by the neural network.
# Predicted_output array initially consists of zeroes.
A__ = numpy.zeros(output_array.shape )
def lowercase_ ( self ):
'''simple docstring'''
A__ = sigmoid(
numpy.dot(self.input_array , self.input_layer_and_first_hidden_layer_weights ) )
# layer_between_first_hidden_layer_and_second_hidden_layer is the layer
# connecting the first hidden set of nodes with the second hidden set of nodes.
A__ = sigmoid(
numpy.dot(
self.layer_between_input_and_first_hidden_layer , self.first_hidden_layer_and_second_hidden_layer_weights , ) )
# layer_between_second_hidden_layer_and_output is the layer connecting
# second hidden layer with the output node.
A__ = sigmoid(
numpy.dot(
self.layer_between_first_hidden_layer_and_second_hidden_layer , self.second_hidden_layer_and_output_layer_weights , ) )
return self.layer_between_second_hidden_layer_and_output
def lowercase_ ( self ):
'''simple docstring'''
A__ = numpy.dot(
self.layer_between_first_hidden_layer_and_second_hidden_layer.T , 2
* (self.output_array - self.predicted_output)
* sigmoid_derivative(self.predicted_output ) , )
A__ = numpy.dot(
self.layer_between_input_and_first_hidden_layer.T , numpy.dot(
2
* (self.output_array - self.predicted_output)
* sigmoid_derivative(self.predicted_output ) , self.second_hidden_layer_and_output_layer_weights.T , )
* sigmoid_derivative(
self.layer_between_first_hidden_layer_and_second_hidden_layer ) , )
A__ = numpy.dot(
self.input_array.T , numpy.dot(
numpy.dot(
2
* (self.output_array - self.predicted_output)
* sigmoid_derivative(self.predicted_output ) , self.second_hidden_layer_and_output_layer_weights.T , )
* sigmoid_derivative(
self.layer_between_first_hidden_layer_and_second_hidden_layer ) , self.first_hidden_layer_and_second_hidden_layer_weights.T , )
* sigmoid_derivative(self.layer_between_input_and_first_hidden_layer ) , )
self.input_layer_and_first_hidden_layer_weights += (
updated_input_layer_and_first_hidden_layer_weights
)
self.first_hidden_layer_and_second_hidden_layer_weights += (
updated_first_hidden_layer_and_second_hidden_layer_weights
)
self.second_hidden_layer_and_output_layer_weights += (
updated_second_hidden_layer_and_output_layer_weights
)
def lowercase_ ( self , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ):
'''simple docstring'''
for iteration in range(1 , iterations + 1 ):
A__ = self.feedforward()
self.back_propagation()
if give_loss:
A__ = numpy.mean(numpy.square(output - self.feedforward() ) )
print(f"""Iteration {iteration} Loss: {loss}""" )
def lowercase_ ( self , UpperCamelCase__ ):
'''simple docstring'''
A__ = input_arr
A__ = sigmoid(
numpy.dot(self.array , self.input_layer_and_first_hidden_layer_weights ) )
A__ = sigmoid(
numpy.dot(
self.layer_between_input_and_first_hidden_layer , self.first_hidden_layer_and_second_hidden_layer_weights , ) )
A__ = sigmoid(
numpy.dot(
self.layer_between_first_hidden_layer_and_second_hidden_layer , self.second_hidden_layer_and_output_layer_weights , ) )
return int(self.layer_between_second_hidden_layer_and_output > 0.6 )
def __a ( A ) -> Dict:
'''simple docstring'''
return 1 / (1 + numpy.exp(-value ))
def __a ( A ) -> List[str]:
'''simple docstring'''
return (value) * (1 - (value))
def __a ( ) -> int:
'''simple docstring'''
A__ = numpy.array(
(
[0, 0, 0],
[0, 0, 1],
[0, 1, 0],
[0, 1, 1],
[1, 0, 0],
[1, 0, 1],
[1, 1, 0],
[1, 1, 1],
) , dtype=numpy.floataa , )
# True output values for the given input values.
A__ = numpy.array(([0], [1], [1], [0], [1], [0], [0], [1]) , dtype=numpy.floataa )
# Calling neural network class.
A__ = TwoHiddenLayerNeuralNetwork(
input_array=A , output_array=A )
# Calling training function.
# Set give_loss to True if you want to see loss in every iteration.
neural_network.train(output=A , iterations=10 , give_loss=A )
return neural_network.predict(numpy.array(([1, 1, 1]) , dtype=numpy.floataa ) )
if __name__ == "__main__":
example() | 337 |
"""simple docstring"""
import unittest
import torch
from diffusers import DDIMScheduler, DDPMScheduler, UNetaDModel
from diffusers.training_utils import set_seed
from diffusers.utils.testing_utils import slow
lowerCAmelCase_ : List[str] = False
class UpperCamelCase_ ( unittest.TestCase ):
def UpperCamelCase_ ( self , snake_case__=32 ) -> Optional[Any]:
"""simple docstring"""
set_seed(0 )
UpperCAmelCase = UNetaDModel(sample_size=snake_case__ , in_channels=3 , out_channels=3 )
UpperCAmelCase = torch.optim.SGD(model.parameters() , lr=0.0_001 )
return model, optimizer
@slow
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = """cpu""" # ensure full determinism without setting the CUBLAS_WORKSPACE_CONFIG env variable
UpperCAmelCase = DDPMScheduler(
num_train_timesteps=10_00 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule="""linear""" , clip_sample=snake_case__ , )
UpperCAmelCase = DDIMScheduler(
num_train_timesteps=10_00 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule="""linear""" , clip_sample=snake_case__ , )
assert ddpm_scheduler.config.num_train_timesteps == ddim_scheduler.config.num_train_timesteps
# shared batches for DDPM and DDIM
set_seed(0 )
UpperCAmelCase = [torch.randn((4, 3, 32, 32) ).clip(-1 , 1 ).to(snake_case__ ) for _ in range(4 )]
UpperCAmelCase = [torch.randn((4, 3, 32, 32) ).to(snake_case__ ) for _ in range(4 )]
UpperCAmelCase = [torch.randint(0 , 10_00 , (4,) ).long().to(snake_case__ ) for _ in range(4 )]
# train with a DDPM scheduler
UpperCAmelCase , UpperCAmelCase = self.get_model_optimizer(resolution=32 )
model.train().to(snake_case__ )
for i in range(4 ):
optimizer.zero_grad()
UpperCAmelCase = ddpm_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i] )
UpperCAmelCase = model(snake_case__ , timesteps[i] ).sample
UpperCAmelCase = torch.nn.functional.mse_loss(snake_case__ , noise[i] )
loss.backward()
optimizer.step()
del model, optimizer
# recreate the model and optimizer, and retry with DDIM
UpperCAmelCase , UpperCAmelCase = self.get_model_optimizer(resolution=32 )
model.train().to(snake_case__ )
for i in range(4 ):
optimizer.zero_grad()
UpperCAmelCase = ddim_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i] )
UpperCAmelCase = model(snake_case__ , timesteps[i] ).sample
UpperCAmelCase = torch.nn.functional.mse_loss(snake_case__ , noise[i] )
loss.backward()
optimizer.step()
del model, optimizer
self.assertTrue(torch.allclose(snake_case__ , snake_case__ , atol=1e-5 ) )
self.assertTrue(torch.allclose(snake_case__ , snake_case__ , atol=1e-5 ) )
| 673 | 0 |
'''simple docstring'''
import re
from filelock import FileLock
try:
import nltk
__UpperCAmelCase =True
except (ImportError, ModuleNotFoundError):
__UpperCAmelCase =False
if NLTK_AVAILABLE:
with FileLock(".lock") as lock:
nltk.download("punkt", quiet=True)
def __lowerCAmelCase ( UpperCamelCase__ ) -> Union[str, Any]:
re.sub('''<n>''' , '''''' , UpperCamelCase__ ) # remove pegasus newline char
assert NLTK_AVAILABLE, "nltk must be installed to separate newlines between sentences. (pip install nltk)"
return "\n".join(nltk.sent_tokenize(UpperCamelCase__ ) )
| 546 |
"""simple docstring"""
import numpy as np
import torch
from torch.utils.data import DataLoader
from accelerate.utils.dataclasses import DistributedType
class UpperCamelCase_ :
def __init__( self , snake_case__=2 , snake_case__=3 , snake_case__=64 , snake_case__=None ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = np.random.default_rng(snake_case__ )
UpperCAmelCase = length
UpperCAmelCase = rng.normal(size=(length,) ).astype(np.floataa )
UpperCAmelCase = a * self.x + b + rng.normal(scale=0.1 , size=(length,) ).astype(np.floataa )
def __len__( self ) -> int:
"""simple docstring"""
return self.length
def __getitem__( self , snake_case__ ) -> Tuple:
"""simple docstring"""
return {"x": self.x[i], "y": self.y[i]}
class UpperCamelCase_ ( torch.nn.Module ):
def __init__( self , snake_case__=0 , snake_case__=0 , snake_case__=False ) -> List[str]:
"""simple docstring"""
super().__init__()
UpperCAmelCase = torch.nn.Parameter(torch.tensor([2, 3] ).float() )
UpperCAmelCase = torch.nn.Parameter(torch.tensor([2, 3] ).float() )
UpperCAmelCase = True
def UpperCamelCase_ ( self , snake_case__=None ) -> List[Any]:
"""simple docstring"""
if self.first_batch:
print(f'''Model dtype: {self.a.dtype}, {self.b.dtype}. Input dtype: {x.dtype}''' )
UpperCAmelCase = False
return x * self.a[0] + self.b[0]
class UpperCamelCase_ ( torch.nn.Module ):
def __init__( self , snake_case__=0 , snake_case__=0 , snake_case__=False ) -> List[Any]:
"""simple docstring"""
super().__init__()
UpperCAmelCase = torch.nn.Parameter(torch.tensor(snake_case__ ).float() )
UpperCAmelCase = torch.nn.Parameter(torch.tensor(snake_case__ ).float() )
UpperCAmelCase = True
def UpperCamelCase_ ( self , snake_case__=None ) -> Optional[Any]:
"""simple docstring"""
if self.first_batch:
print(f'''Model dtype: {self.a.dtype}, {self.b.dtype}. Input dtype: {x.dtype}''' )
UpperCAmelCase = False
return x * self.a + self.b
def _lowerCAmelCase ( lowerCAmelCase , lowerCAmelCase = 16 ):
'''simple docstring'''
from datasets import load_dataset
from transformers import AutoTokenizer
UpperCAmelCase = AutoTokenizer.from_pretrained("""bert-base-cased""" )
UpperCAmelCase = {"""train""": """tests/test_samples/MRPC/train.csv""", """validation""": """tests/test_samples/MRPC/dev.csv"""}
UpperCAmelCase = load_dataset("""csv""" , data_files=lowerCAmelCase )
UpperCAmelCase = datasets["""train"""].unique("""label""" )
UpperCAmelCase = {v: i for i, v in enumerate(lowerCAmelCase )}
def tokenize_function(lowerCAmelCase ):
# max_length=None => use the model max length (it's actually the default)
UpperCAmelCase = tokenizer(
examples["""sentence1"""] , examples["""sentence2"""] , truncation=lowerCAmelCase , max_length=lowerCAmelCase , padding="""max_length""" )
if "label" in examples:
UpperCAmelCase = [label_to_id[l] for l in examples["""label"""]]
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
UpperCAmelCase = datasets.map(
lowerCAmelCase , batched=lowerCAmelCase , remove_columns=["""sentence1""", """sentence2""", """label"""] , )
def collate_fn(lowerCAmelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
if accelerator.distributed_type == DistributedType.TPU:
return tokenizer.pad(lowerCAmelCase , padding="""max_length""" , max_length=128 , return_tensors="""pt""" )
return tokenizer.pad(lowerCAmelCase , padding="""longest""" , return_tensors="""pt""" )
# Instantiate dataloaders.
UpperCAmelCase = DataLoader(tokenized_datasets["""train"""] , shuffle=lowerCAmelCase , collate_fn=lowerCAmelCase , batch_size=2 )
UpperCAmelCase = DataLoader(tokenized_datasets["""validation"""] , shuffle=lowerCAmelCase , collate_fn=lowerCAmelCase , batch_size=1 )
return train_dataloader, eval_dataloader
| 673 | 0 |
import os
from tempfile import TemporaryDirectory
from unittest import TestCase
import pytest
from absl.testing import parameterized
from datasets import config
from datasets.arrow_reader import HF_GCP_BASE_URL
from datasets.builder import DatasetBuilder
from datasets.dataset_dict import IterableDatasetDict
from datasets.iterable_dataset import IterableDataset
from datasets.load import dataset_module_factory, import_main_class
from datasets.utils.file_utils import cached_path
lowerCamelCase = [
{'''dataset''': '''wikipedia''', '''config_name''': '''20220301.de'''},
{'''dataset''': '''wikipedia''', '''config_name''': '''20220301.en'''},
{'''dataset''': '''wikipedia''', '''config_name''': '''20220301.fr'''},
{'''dataset''': '''wikipedia''', '''config_name''': '''20220301.frr'''},
{'''dataset''': '''wikipedia''', '''config_name''': '''20220301.it'''},
{'''dataset''': '''wikipedia''', '''config_name''': '''20220301.simple'''},
{'''dataset''': '''snli''', '''config_name''': '''plain_text'''},
{'''dataset''': '''eli5''', '''config_name''': '''LFQA_reddit'''},
{'''dataset''': '''wiki40b''', '''config_name''': '''en'''},
{'''dataset''': '''wiki_dpr''', '''config_name''': '''psgs_w100.nq.compressed'''},
{'''dataset''': '''wiki_dpr''', '''config_name''': '''psgs_w100.nq.no_index'''},
{'''dataset''': '''wiki_dpr''', '''config_name''': '''psgs_w100.multiset.no_index'''},
{'''dataset''': '''natural_questions''', '''config_name''': '''default'''},
]
def a_ ( SCREAMING_SNAKE_CASE__ : Optional[Any]=True ):
'''simple docstring'''
if with_config:
return [
{
"testcase_name": d["dataset"] + "/" + d["config_name"],
"dataset": d["dataset"],
"config_name": d["config_name"],
}
for d in DATASETS_ON_HF_GCP
]
else:
return [
{"testcase_name": dataset, "dataset": dataset} for dataset in {d["dataset"] for d in DATASETS_ON_HF_GCP}
]
@parameterized.named_parameters(list_datasets_on_hf_gcp_parameters(with_config=a_ ) )
class A ( a_ ):
UpperCamelCase__ : Tuple =None
UpperCamelCase__ : List[Any] =None
def lowerCamelCase ( self : List[Any] , lowercase_ : str , lowercase_ : Optional[int] ) -> Union[str, Any]:
"""simple docstring"""
with TemporaryDirectory() as tmp_dir:
_lowerCamelCase : Optional[Any] =dataset_module_factory(snake_case__ , cache_dir=snake_case__ )
_lowerCamelCase : Tuple =import_main_class(dataset_module.module_path , dataset=snake_case__ )
_lowerCamelCase : int =builder_cls(
cache_dir=snake_case__ , config_name=snake_case__ , hash=dataset_module.hash , )
_lowerCamelCase : Optional[int] ='/'.join(
[
HF_GCP_BASE_URL,
builder_instance._relative_data_dir(with_hash=snake_case__ ).replace(os.sep , '/' ),
config.DATASET_INFO_FILENAME,
] )
_lowerCamelCase : Optional[Any] =cached_path(snake_case__ , cache_dir=snake_case__ )
self.assertTrue(os.path.exists(snake_case__ ) )
@pytest.mark.integration
def a_ ( SCREAMING_SNAKE_CASE__ : List[str] ):
'''simple docstring'''
_lowerCamelCase : Optional[int] =tmp_path_factory.mktemp('test_hf_gcp' ) / 'test_wikipedia_simple'
_lowerCamelCase : Tuple =dataset_module_factory('wikipedia' , cache_dir=SCREAMING_SNAKE_CASE__ )
_lowerCamelCase : List[str] =import_main_class(dataset_module.module_path )
_lowerCamelCase : Optional[Any] =builder_cls(
cache_dir=SCREAMING_SNAKE_CASE__ , config_name='20220301.frr' , hash=dataset_module.hash , )
# use the HF cloud storage, not the original download_and_prepare that uses apache-beam
_lowerCamelCase : Tuple =None
builder_instance.download_and_prepare()
_lowerCamelCase : List[str] =builder_instance.as_dataset()
assert ds
@pytest.mark.integration
def a_ ( SCREAMING_SNAKE_CASE__ : Tuple ):
'''simple docstring'''
_lowerCamelCase : Any =dataset_module_factory('wikipedia' , cache_dir=SCREAMING_SNAKE_CASE__ )
_lowerCamelCase : Optional[int] =import_main_class(dataset_module.module_path , dataset=SCREAMING_SNAKE_CASE__ )
_lowerCamelCase : Union[str, Any] =builder_cls(
cache_dir=SCREAMING_SNAKE_CASE__ , config_name='20220301.frr' , hash=dataset_module.hash , )
_lowerCamelCase : str =builder_instance.as_streaming_dataset()
assert ds
assert isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
assert "train" in ds
assert isinstance(ds['train'] , SCREAMING_SNAKE_CASE__ )
assert next(iter(ds['train'] ) )
| 464 |
"""simple docstring"""
import flax.linen as nn
import jax
import jax.numpy as jnp
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self , snake_case__ ) -> Tuple:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase = hidden_states.shape
UpperCAmelCase = jax.image.resize(
snake_case__ , shape=(batch, height * 2, width * 2, channels) , method="""nearest""" , )
UpperCAmelCase = self.conv(snake_case__ )
return hidden_states
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self , snake_case__ ) -> Any:
"""simple docstring"""
UpperCAmelCase = self.conv(snake_case__ )
return hidden_states
class UpperCamelCase_ ( nn.Module ):
_A : int
_A : int = None
_A : float = 0.0
_A : bool = None
_A : jnp.dtype = jnp.floataa
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = self.in_channels if self.out_channels is None else self.out_channels
UpperCAmelCase = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
UpperCAmelCase = nn.Dense(snake_case__ , dtype=self.dtype )
UpperCAmelCase = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
UpperCAmelCase = nn.Dropout(self.dropout_prob )
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
UpperCAmelCase = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
UpperCAmelCase = None
if use_nin_shortcut:
UpperCAmelCase = nn.Conv(
snake_case__ , kernel_size=(1, 1) , strides=(1, 1) , padding="""VALID""" , dtype=self.dtype , )
def __call__( self , snake_case__ , snake_case__ , snake_case__=True ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = hidden_states
UpperCAmelCase = self.norma(snake_case__ )
UpperCAmelCase = nn.swish(snake_case__ )
UpperCAmelCase = self.conva(snake_case__ )
UpperCAmelCase = self.time_emb_proj(nn.swish(snake_case__ ) )
UpperCAmelCase = jnp.expand_dims(jnp.expand_dims(snake_case__ , 1 ) , 1 )
UpperCAmelCase = hidden_states + temb
UpperCAmelCase = self.norma(snake_case__ )
UpperCAmelCase = nn.swish(snake_case__ )
UpperCAmelCase = self.dropout(snake_case__ , snake_case__ )
UpperCAmelCase = self.conva(snake_case__ )
if self.conv_shortcut is not None:
UpperCAmelCase = self.conv_shortcut(snake_case__ )
return hidden_states + residual
| 673 | 0 |
"""simple docstring"""
import unittest
import torch
from diffusers import DDIMScheduler, DDPMScheduler, UNetaDModel
from diffusers.training_utils import set_seed
from diffusers.utils.testing_utils import slow
UpperCAmelCase : List[str] = False
class SCREAMING_SNAKE_CASE__ ( unittest.TestCase ):
def _UpperCAmelCase ( self : Optional[Any] , lowerCAmelCase_ : List[str]=3_2):
"""simple docstring"""
set_seed(0)
lowercase_ = UNetaDModel(sample_size=snake_case__ , in_channels=3 , out_channels=3)
lowercase_ = torch.optim.SGD(model.parameters() , lr=0.0_001)
return model, optimizer
@slow
def _UpperCAmelCase ( self : Dict):
"""simple docstring"""
lowercase_ = """cpu""" # ensure full determinism without setting the CUBLAS_WORKSPACE_CONFIG env variable
lowercase_ = DDPMScheduler(
num_train_timesteps=1_0_0_0 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule="""linear""" , clip_sample=snake_case__ , )
lowercase_ = DDIMScheduler(
num_train_timesteps=1_0_0_0 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule="""linear""" , clip_sample=snake_case__ , )
assert ddpm_scheduler.config.num_train_timesteps == ddim_scheduler.config.num_train_timesteps
# shared batches for DDPM and DDIM
set_seed(0)
lowercase_ = [torch.randn((4, 3, 3_2, 3_2)).clip(-1 , 1).to(snake_case__) for _ in range(4)]
lowercase_ = [torch.randn((4, 3, 3_2, 3_2)).to(snake_case__) for _ in range(4)]
lowercase_ = [torch.randint(0 , 1_0_0_0 , (4,)).long().to(snake_case__) for _ in range(4)]
# train with a DDPM scheduler
lowercase_ , lowercase_ = self.get_model_optimizer(resolution=3_2)
model.train().to(snake_case__)
for i in range(4):
optimizer.zero_grad()
lowercase_ = ddpm_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i])
lowercase_ = model(snake_case__ , timesteps[i]).sample
lowercase_ = torch.nn.functional.mse_loss(snake_case__ , noise[i])
loss.backward()
optimizer.step()
del model, optimizer
# recreate the model and optimizer, and retry with DDIM
lowercase_ , lowercase_ = self.get_model_optimizer(resolution=3_2)
model.train().to(snake_case__)
for i in range(4):
optimizer.zero_grad()
lowercase_ = ddim_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i])
lowercase_ = model(snake_case__ , timesteps[i]).sample
lowercase_ = torch.nn.functional.mse_loss(snake_case__ , noise[i])
loss.backward()
optimizer.step()
del model, optimizer
self.assertTrue(torch.allclose(snake_case__ , snake_case__ , atol=1E-5))
self.assertTrue(torch.allclose(snake_case__ , snake_case__ , atol=1E-5))
| 567 |
"""simple docstring"""
from __future__ import annotations
import inspect
import unittest
from transformers import ViTConfig
from transformers.testing_utils import require_tf, require_vision, slow
from transformers.utils import cached_property, is_tf_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFViTForImageClassification, TFViTModel
if is_vision_available():
from PIL import Image
from transformers import ViTImageProcessor
class UpperCamelCase_ :
def __init__( self , snake_case__ , snake_case__=13 , snake_case__=30 , snake_case__=2 , snake_case__=3 , snake_case__=True , snake_case__=True , snake_case__=32 , snake_case__=2 , snake_case__=4 , snake_case__=37 , snake_case__="gelu" , snake_case__=0.1 , snake_case__=0.1 , snake_case__=10 , snake_case__=0.02 , snake_case__=3 , snake_case__=None , ) -> Optional[Any]:
"""simple docstring"""
UpperCAmelCase = parent
UpperCAmelCase = batch_size
UpperCAmelCase = image_size
UpperCAmelCase = patch_size
UpperCAmelCase = num_channels
UpperCAmelCase = is_training
UpperCAmelCase = use_labels
UpperCAmelCase = hidden_size
UpperCAmelCase = num_hidden_layers
UpperCAmelCase = num_attention_heads
UpperCAmelCase = intermediate_size
UpperCAmelCase = hidden_act
UpperCAmelCase = hidden_dropout_prob
UpperCAmelCase = attention_probs_dropout_prob
UpperCAmelCase = type_sequence_label_size
UpperCAmelCase = initializer_range
UpperCAmelCase = scope
# in ViT, the seq length equals the number of patches + 1 (we add 1 for the [CLS] token)
UpperCAmelCase = (image_size // patch_size) ** 2
UpperCAmelCase = num_patches + 1
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
UpperCAmelCase = None
if self.use_labels:
UpperCAmelCase = ids_tensor([self.batch_size] , self.type_sequence_label_size )
UpperCAmelCase = self.get_config()
return config, pixel_values, labels
def UpperCamelCase_ ( self ) -> Optional[int]:
"""simple docstring"""
return ViTConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , is_decoder=snake_case__ , initializer_range=self.initializer_range , )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ ) -> Dict:
"""simple docstring"""
UpperCAmelCase = TFViTModel(config=snake_case__ )
UpperCAmelCase = model(snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
# Test with an image with different size than the one specified in config.
UpperCAmelCase = self.image_size // 2
UpperCAmelCase = pixel_values[:, :, :image_size, :image_size]
UpperCAmelCase = model(snake_case__ , interpolate_pos_encoding=snake_case__ , training=snake_case__ )
UpperCAmelCase = (image_size // self.patch_size) ** 2 + 1
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, seq_length, self.hidden_size) )
def UpperCamelCase_ ( self , snake_case__ , snake_case__ , snake_case__ ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = self.type_sequence_label_size
UpperCAmelCase = TFViTForImageClassification(snake_case__ )
UpperCAmelCase = model(snake_case__ , labels=snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# Test with an image with different size than the one specified in config.
UpperCAmelCase = self.image_size // 2
UpperCAmelCase = pixel_values[:, :, :image_size, :image_size]
UpperCAmelCase = model(snake_case__ , interpolate_pos_encoding=snake_case__ , training=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# test greyscale images
UpperCAmelCase = 1
UpperCAmelCase = TFViTForImageClassification(snake_case__ )
UpperCAmelCase = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
UpperCAmelCase = model(snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def UpperCamelCase_ ( self ) -> int:
"""simple docstring"""
UpperCAmelCase = self.prepare_config_and_inputs()
UpperCAmelCase , UpperCAmelCase , UpperCAmelCase = config_and_inputs
UpperCAmelCase = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_tf
class UpperCamelCase_ ( a_ , a_ , unittest.TestCase ):
_A : Optional[int] = (TFViTModel, TFViTForImageClassification) if is_tf_available() else ()
_A : Optional[Any] = (
{'feature-extraction': TFViTModel, 'image-classification': TFViTForImageClassification}
if is_tf_available()
else {}
)
_A : Optional[int] = False
_A : Any = False
_A : List[str] = False
def UpperCamelCase_ ( self ) -> Dict:
"""simple docstring"""
UpperCAmelCase = TFViTModelTester(self )
UpperCAmelCase = ConfigTester(self , config_class=snake_case__ , has_text_modality=snake_case__ , hidden_size=37 )
def UpperCamelCase_ ( self ) -> Tuple:
"""simple docstring"""
self.config_tester.run_common_tests()
@unittest.skip(reason="""ViT does not use inputs_embeds""" )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
pass
@unittest.skip(reason="""ViT does not use inputs_embeds""" )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
pass
def UpperCamelCase_ ( self ) -> Union[str, Any]:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
UpperCAmelCase = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (tf.keras.layers.Layer) )
UpperCAmelCase = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , tf.keras.layers.Layer ) )
def UpperCamelCase_ ( self ) -> List[str]:
"""simple docstring"""
UpperCAmelCase , UpperCAmelCase = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
UpperCAmelCase = model_class(snake_case__ )
UpperCAmelCase = inspect.signature(model.call )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
UpperCAmelCase = [*signature.parameters.keys()]
UpperCAmelCase = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] , snake_case__ )
def UpperCamelCase_ ( self ) -> Any:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*snake_case__ )
@slow
def UpperCamelCase_ ( self ) -> str:
"""simple docstring"""
UpperCAmelCase = TFViTModel.from_pretrained("""google/vit-base-patch16-224""" )
self.assertIsNotNone(snake_case__ )
def _lowerCAmelCase ( ):
'''simple docstring'''
UpperCAmelCase = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
return image
@require_tf
@require_vision
class UpperCamelCase_ ( unittest.TestCase ):
@cached_property
def UpperCamelCase_ ( self ) -> Optional[Any]:
"""simple docstring"""
return ViTImageProcessor.from_pretrained("""google/vit-base-patch16-224""" ) if is_vision_available() else None
@slow
def UpperCamelCase_ ( self ) -> List[Any]:
"""simple docstring"""
UpperCAmelCase = TFViTForImageClassification.from_pretrained("""google/vit-base-patch16-224""" )
UpperCAmelCase = self.default_image_processor
UpperCAmelCase = prepare_img()
UpperCAmelCase = image_processor(images=snake_case__ , return_tensors="""tf""" )
# forward pass
UpperCAmelCase = model(**snake_case__ )
# verify the logits
UpperCAmelCase = tf.TensorShape((1, 10_00) )
self.assertEqual(outputs.logits.shape , snake_case__ )
UpperCAmelCase = tf.constant([-0.2_744, 0.8_215, -0.0_836] )
tf.debugging.assert_near(outputs.logits[0, :3] , snake_case__ , atol=1e-4 )
| 673 | 0 |
"""simple docstring"""
import argparse
import torch
from transformers import (
SpeechTaConfig,
SpeechTaFeatureExtractor,
SpeechTaForSpeechToSpeech,
SpeechTaForSpeechToText,
SpeechTaForTextToSpeech,
SpeechTaProcessor,
SpeechTaTokenizer,
logging,
)
from transformers.tokenization_utils import AddedToken
logging.set_verbosity_info()
_lowerCamelCase = logging.get_logger('''transformers.models.speecht5''')
_lowerCamelCase = {
'''speech_encoder_prenet.layer_norm''': '''speecht5.encoder.prenet.feature_projection.layer_norm''',
'''speech_encoder_prenet.post_extract_proj''': '''speecht5.encoder.prenet.feature_projection.projection''',
'''speech_encoder_prenet.pos_conv.0''': '''speecht5.encoder.prenet.pos_conv_embed.conv''',
'''speech_encoder_prenet.mask_emb''': '''speecht5.encoder.prenet.masked_spec_embed''',
}
_lowerCamelCase = {
'''text_encoder_prenet.encoder_prenet.0''': '''speecht5.encoder.prenet.embed_tokens''',
'''text_encoder_prenet.encoder_prenet.1.alpha''': '''speecht5.encoder.prenet.encode_positions.alpha''',
}
_lowerCamelCase = {
'''speech_decoder_prenet.decoder_prenet.0.0.prenet.0.0''': '''speecht5.decoder.prenet.layers.0''',
'''speech_decoder_prenet.decoder_prenet.0.0.prenet.1.0''': '''speecht5.decoder.prenet.layers.1''',
'''speech_decoder_prenet.decoder_prenet.0.1''': '''speecht5.decoder.prenet.final_layer''',
'''speech_decoder_prenet.decoder_prenet.1.alpha''': '''speecht5.decoder.prenet.encode_positions.alpha''',
'''speech_decoder_prenet.spkembs_layer.0''': '''speecht5.decoder.prenet.speaker_embeds_layer''',
}
_lowerCamelCase = {
'''speech_decoder_postnet.feat_out''': '''speech_decoder_postnet.feat_out''',
'''speech_decoder_postnet.prob_out''': '''speech_decoder_postnet.prob_out''',
'''speech_decoder_postnet.postnet.postnet.0.0''': '''speech_decoder_postnet.layers.0.conv''',
'''speech_decoder_postnet.postnet.postnet.0.1''': '''speech_decoder_postnet.layers.0.batch_norm''',
'''speech_decoder_postnet.postnet.postnet.1.0''': '''speech_decoder_postnet.layers.1.conv''',
'''speech_decoder_postnet.postnet.postnet.1.1''': '''speech_decoder_postnet.layers.1.batch_norm''',
'''speech_decoder_postnet.postnet.postnet.2.0''': '''speech_decoder_postnet.layers.2.conv''',
'''speech_decoder_postnet.postnet.postnet.2.1''': '''speech_decoder_postnet.layers.2.batch_norm''',
'''speech_decoder_postnet.postnet.postnet.3.0''': '''speech_decoder_postnet.layers.3.conv''',
'''speech_decoder_postnet.postnet.postnet.3.1''': '''speech_decoder_postnet.layers.3.batch_norm''',
'''speech_decoder_postnet.postnet.postnet.4.0''': '''speech_decoder_postnet.layers.4.conv''',
'''speech_decoder_postnet.postnet.postnet.4.1''': '''speech_decoder_postnet.layers.4.batch_norm''',
}
_lowerCamelCase = {
'''text_decoder_prenet.embed_tokens''': '''speecht5.decoder.prenet.embed_tokens''',
}
_lowerCamelCase = {
'''text_decoder_postnet.output_projection''': '''text_decoder_postnet.lm_head''',
}
_lowerCamelCase = {
'''encoder.layers.*.self_attn.k_proj''': '''speecht5.encoder.wrapped_encoder.layers.*.attention.k_proj''',
'''encoder.layers.*.self_attn.v_proj''': '''speecht5.encoder.wrapped_encoder.layers.*.attention.v_proj''',
'''encoder.layers.*.self_attn.q_proj''': '''speecht5.encoder.wrapped_encoder.layers.*.attention.q_proj''',
'''encoder.layers.*.self_attn.out_proj''': '''speecht5.encoder.wrapped_encoder.layers.*.attention.out_proj''',
'''encoder.layers.*.self_attn_layer_norm''': '''speecht5.encoder.wrapped_encoder.layers.*.layer_norm''',
'''encoder.layers.*.fc1''': '''speecht5.encoder.wrapped_encoder.layers.*.feed_forward.intermediate_dense''',
'''encoder.layers.*.fc2''': '''speecht5.encoder.wrapped_encoder.layers.*.feed_forward.output_dense''',
'''encoder.layers.*.final_layer_norm''': '''speecht5.encoder.wrapped_encoder.layers.*.final_layer_norm''',
'''encoder.layer_norm''': '''speecht5.encoder.wrapped_encoder.layer_norm''',
'''encoder.pos_emb.pe_k''': '''speecht5.encoder.wrapped_encoder.embed_positions.pe_k''',
}
_lowerCamelCase = {
'''decoder.layers.*.self_attn.k_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn.k_proj''',
'''decoder.layers.*.self_attn.v_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn.v_proj''',
'''decoder.layers.*.self_attn.q_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn.q_proj''',
'''decoder.layers.*.self_attn.out_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn.out_proj''',
'''decoder.layers.*.self_attn_layer_norm''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn_layer_norm''',
'''decoder.layers.*.encoder_attn.k_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn.k_proj''',
'''decoder.layers.*.encoder_attn.v_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn.v_proj''',
'''decoder.layers.*.encoder_attn.q_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn.q_proj''',
'''decoder.layers.*.encoder_attn.out_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn.out_proj''',
'''decoder.layers.*.encoder_attn_layer_norm''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn_layer_norm''',
'''decoder.layers.*.fc1''': '''speecht5.decoder.wrapped_decoder.layers.*.feed_forward.intermediate_dense''',
'''decoder.layers.*.fc2''': '''speecht5.decoder.wrapped_decoder.layers.*.feed_forward.output_dense''',
'''decoder.layers.*.final_layer_norm''': '''speecht5.decoder.wrapped_decoder.layers.*.final_layer_norm''',
}
_lowerCamelCase = {
**MAPPING_SPEECH_ENCODER_PRENET,
**MAPPING_ENCODER,
**MAPPING_DECODER,
**MAPPING_TEXT_DECODER_PRENET,
**MAPPING_TEXT_DECODER_POSTNET,
}
_lowerCamelCase = {
**MAPPING_TEXT_ENCODER_PRENET,
**MAPPING_ENCODER,
**MAPPING_DECODER,
**MAPPING_SPEECH_DECODER_PRENET,
**MAPPING_SPEECH_DECODER_POSTNET,
}
_lowerCamelCase = {
**MAPPING_SPEECH_ENCODER_PRENET,
**MAPPING_ENCODER,
**MAPPING_DECODER,
**MAPPING_SPEECH_DECODER_PRENET,
**MAPPING_SPEECH_DECODER_POSTNET,
}
_lowerCamelCase = []
_lowerCamelCase = [
'''encoder.version''',
'''encoder.layers.*.norm_k.weight''',
'''encoder.layers.*.norm_k.bias''',
'''decoder.version''',
'''decoder.layers.*.norm_k.weight''',
'''decoder.layers.*.norm_k.bias''',
'''decoder.pos_emb.pe_k''',
'''speech_encoder_prenet.embed_positions._float_tensor''',
'''text_decoder_prenet.embed_positions._float_tensor''',
]
_lowerCamelCase = IGNORE_KEYS + [
'''encoder.proj''',
'''text_encoder_prenet.*''',
'''speech_decoder_prenet.*''',
'''speech_decoder_postnet.*''',
]
_lowerCamelCase = IGNORE_KEYS + [
'''encoder.proj''',
'''speech_encoder_prenet.*''',
'''text_decoder_prenet.*''',
'''text_decoder_postnet.*''',
]
_lowerCamelCase = IGNORE_KEYS + [
'''encoder.proj''',
'''text_encoder_prenet.*''',
'''text_decoder_prenet.*''',
'''text_decoder_postnet.*''',
]
def lowerCAmelCase_ ( lowercase_ : Any , lowercase_ : str , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple ):
'''simple docstring'''
for attribute in key.split('''.''' ):
__SCREAMING_SNAKE_CASE : Dict = getattr(lowercase_ , lowercase_ )
if weight_type is not None:
__SCREAMING_SNAKE_CASE : str = getattr(lowercase_ , lowercase_ ).shape
else:
__SCREAMING_SNAKE_CASE : List[str] = hf_pointer.shape
if hf_shape != value.shape:
raise ValueError(
F'''Shape of hf {key + '.' + weight_type if weight_type is not None else ''} is {hf_shape}, but should be'''
F''' {value.shape} for {full_name}''' )
if weight_type == "weight":
__SCREAMING_SNAKE_CASE : str = value
elif weight_type == "weight_g":
__SCREAMING_SNAKE_CASE : int = value
elif weight_type == "weight_v":
__SCREAMING_SNAKE_CASE : Dict = value
elif weight_type == "bias":
__SCREAMING_SNAKE_CASE : List[Any] = value
elif weight_type == "running_mean":
__SCREAMING_SNAKE_CASE : List[Any] = value
elif weight_type == "running_var":
__SCREAMING_SNAKE_CASE : int = value
elif weight_type == "num_batches_tracked":
__SCREAMING_SNAKE_CASE : Optional[int] = value
else:
__SCREAMING_SNAKE_CASE : List[Any] = value
logger.info(F'''{key + ('.' + weight_type if weight_type is not None else '')} was initialized from {full_name}.''' )
def lowerCAmelCase_ ( lowercase_ : Optional[int] , lowercase_ : int ):
'''simple docstring'''
for key in ignore_keys:
if key.endswith('''.*''' ):
if name.startswith(key[:-1] ):
return True
elif ".*." in key:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : int = key.split('''.*.''' )
if prefix in name and suffix in name:
return True
elif key in name:
return True
return False
def lowerCAmelCase_ ( lowercase_ : Any , lowercase_ : Dict , lowercase_ : Dict ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = []
if task == "s2t":
__SCREAMING_SNAKE_CASE : int = hf_model.speechta.encoder.prenet.feature_encoder
__SCREAMING_SNAKE_CASE : List[str] = MAPPING_S2T
__SCREAMING_SNAKE_CASE : Optional[Any] = IGNORE_KEYS_S2T
elif task == "t2s":
__SCREAMING_SNAKE_CASE : List[str] = None
__SCREAMING_SNAKE_CASE : Optional[int] = MAPPING_T2S
__SCREAMING_SNAKE_CASE : Dict = IGNORE_KEYS_T2S
elif task == "s2s":
__SCREAMING_SNAKE_CASE : Dict = hf_model.speechta.encoder.prenet.feature_encoder
__SCREAMING_SNAKE_CASE : Any = MAPPING_S2S
__SCREAMING_SNAKE_CASE : List[Any] = IGNORE_KEYS_S2S
else:
raise ValueError(F'''Unsupported task: {task}''' )
for name, value in fairseq_dict.items():
if should_ignore(lowercase_ , lowercase_ ):
logger.info(F'''{name} was ignored''' )
continue
__SCREAMING_SNAKE_CASE : Optional[Any] = False
if "conv_layers" in name:
load_conv_layer(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , hf_model.config.feat_extract_norm == '''group''' , )
__SCREAMING_SNAKE_CASE : Tuple = True
else:
for key, mapped_key in MAPPING.items():
# mapped_key = "speecht5." + mapped_key if mapped_key not in TOP_LEVEL_KEYS else mapped_key
if "*" in key:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = key.split('''.*.''' )
if prefix in name and suffix in name:
__SCREAMING_SNAKE_CASE : str = suffix
# if key in name or key.split("w2v_model.")[-1] == name.split(".")[0]:
if key in name:
__SCREAMING_SNAKE_CASE : Any = True
if "*" in mapped_key:
__SCREAMING_SNAKE_CASE : Tuple = name.split(lowercase_ )[0].split('''.''' )[-2]
__SCREAMING_SNAKE_CASE : int = mapped_key.replace('''*''' , lowercase_ )
if "weight_g" in name:
__SCREAMING_SNAKE_CASE : int = '''weight_g'''
elif "weight_v" in name:
__SCREAMING_SNAKE_CASE : str = '''weight_v'''
elif "bias" in name:
__SCREAMING_SNAKE_CASE : Any = '''bias'''
elif "weight" in name:
__SCREAMING_SNAKE_CASE : Optional[Any] = '''weight'''
elif "running_mean" in name:
__SCREAMING_SNAKE_CASE : int = '''running_mean'''
elif "running_var" in name:
__SCREAMING_SNAKE_CASE : List[Any] = '''running_var'''
elif "num_batches_tracked" in name:
__SCREAMING_SNAKE_CASE : int = '''num_batches_tracked'''
else:
__SCREAMING_SNAKE_CASE : List[str] = None
set_recursively(lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
continue
if not is_used:
unused_weights.append(lowercase_ )
logger.warning(F'''Unused weights: {unused_weights}''' )
def lowerCAmelCase_ ( lowercase_ : int , lowercase_ : List[str] , lowercase_ : Dict , lowercase_ : Union[str, Any] , lowercase_ : List[str] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[Any] = full_name.split('''conv_layers.''' )[-1]
__SCREAMING_SNAKE_CASE : int = name.split('''.''' )
__SCREAMING_SNAKE_CASE : Any = int(items[0] )
__SCREAMING_SNAKE_CASE : Optional[Any] = int(items[1] )
if type_id == 0:
if "bias" in name:
if value.shape != feature_extractor.conv_layers[layer_id].conv.bias.data.shape:
raise ValueError(
F'''{full_name} has size {value.shape}, but'''
F''' {feature_extractor.conv_layers[layer_id].conv.bias.data.shape} was found.''' )
__SCREAMING_SNAKE_CASE : Any = value
logger.info(F'''Feat extract conv layer {layer_id} was initialized from {full_name}.''' )
elif "weight" in name:
if value.shape != feature_extractor.conv_layers[layer_id].conv.weight.data.shape:
raise ValueError(
F'''{full_name} has size {value.shape}, but'''
F''' {feature_extractor.conv_layers[layer_id].conv.weight.data.shape} was found.''' )
__SCREAMING_SNAKE_CASE : Dict = value
logger.info(F'''Feat extract conv layer {layer_id} was initialized from {full_name}.''' )
elif (type_id == 2 and not use_group_norm) or (type_id == 2 and layer_id == 0 and use_group_norm):
if "bias" in name:
if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape:
raise ValueError(
F'''{full_name} has size {value.shape}, but'''
F''' {feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape} was found.''' )
__SCREAMING_SNAKE_CASE : List[str] = value
logger.info(F'''Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.''' )
elif "weight" in name:
if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape:
raise ValueError(
F'''{full_name} has size {value.shape}, but'''
F''' {feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape} was found.''' )
__SCREAMING_SNAKE_CASE : str = value
logger.info(F'''Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.''' )
else:
unused_weights.append(lowercase_ )
@torch.no_grad()
def lowerCAmelCase_ ( lowercase_ : Union[str, Any] , lowercase_ : Any , lowercase_ : Any , lowercase_ : Union[str, Any]=None , lowercase_ : Tuple=None , lowercase_ : str=None , ):
'''simple docstring'''
if config_path is not None:
__SCREAMING_SNAKE_CASE : Dict = SpeechTaConfig.from_pretrained(lowercase_ )
else:
__SCREAMING_SNAKE_CASE : List[str] = SpeechTaConfig()
if task == "s2t":
__SCREAMING_SNAKE_CASE : Optional[int] = config.max_text_positions
__SCREAMING_SNAKE_CASE : int = SpeechTaForSpeechToText(lowercase_ )
elif task == "t2s":
__SCREAMING_SNAKE_CASE : Optional[int] = 1876
__SCREAMING_SNAKE_CASE : Optional[int] = 600
__SCREAMING_SNAKE_CASE : List[str] = config.max_speech_positions
__SCREAMING_SNAKE_CASE : int = SpeechTaForTextToSpeech(lowercase_ )
elif task == "s2s":
__SCREAMING_SNAKE_CASE : Optional[Any] = 1876
__SCREAMING_SNAKE_CASE : Any = config.max_speech_positions
__SCREAMING_SNAKE_CASE : Optional[int] = SpeechTaForSpeechToSpeech(lowercase_ )
else:
raise ValueError(F'''Unknown task name: {task}''' )
if vocab_path:
__SCREAMING_SNAKE_CASE : Any = SpeechTaTokenizer(lowercase_ , model_max_length=config.max_text_positions )
# Mask token behaves like a normal word, i.e. include the space before it
__SCREAMING_SNAKE_CASE : List[Any] = AddedToken('''<mask>''' , lstrip=lowercase_ , rstrip=lowercase_ )
__SCREAMING_SNAKE_CASE : int = mask_token
tokenizer.add_special_tokens({'''mask_token''': mask_token} )
tokenizer.add_tokens(['''<ctc_blank>'''] )
__SCREAMING_SNAKE_CASE : Optional[Any] = SpeechTaFeatureExtractor()
__SCREAMING_SNAKE_CASE : Dict = SpeechTaProcessor(tokenizer=lowercase_ , feature_extractor=lowercase_ )
processor.save_pretrained(lowercase_ )
__SCREAMING_SNAKE_CASE : List[Any] = torch.load(lowercase_ )
recursively_load_weights(fairseq_checkpoint['''model'''] , lowercase_ , lowercase_ )
model.save_pretrained(lowercase_ )
if repo_id:
print('''Pushing to the hub...''' )
processor.push_to_hub(lowercase_ )
model.push_to_hub(lowercase_ )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
parser.add_argument(
'''--task''',
default='''s2t''',
type=str,
help='''Type of the SpeechT5 model you\'d like to convert. Should be one of \'s2t\', \'t2s\', \'s2s\'.''',
)
parser.add_argument('''--checkpoint_path''', required=True, default=None, type=str, help='''Path to fairseq checkpoint''')
parser.add_argument('''--vocab_path''', default=None, type=str, help='''Path to SentencePiece model''')
parser.add_argument('''--config_path''', default=None, type=str, help='''Path to hf config.json of model to convert''')
parser.add_argument(
'''--pytorch_dump_folder_path''', required=True, default=None, type=str, help='''Path to the output PyTorch model.'''
)
parser.add_argument(
'''--push_to_hub''', default=None, type=str, help='''Where to upload the converted model on the 🤗 hub.'''
)
_lowerCamelCase = parser.parse_args()
convert_speechta_checkpoint(
args.task,
args.checkpoint_path,
args.pytorch_dump_folder_path,
args.config_path,
args.vocab_path,
args.push_to_hub,
)
| 674 |
"""simple docstring"""
from collections.abc import Callable
import numpy as np
def lowerCAmelCase_ ( lowercase_ : Callable , lowercase_ : float , lowercase_ : float , lowercase_ : float , lowercase_ : float ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : int = int(np.ceil((x_end - xa) / step_size ) )
__SCREAMING_SNAKE_CASE : Dict = np.zeros((n + 1,) )
__SCREAMING_SNAKE_CASE : List[Any] = ya
__SCREAMING_SNAKE_CASE : Dict = xa
for k in range(lowercase_ ):
__SCREAMING_SNAKE_CASE : str = y[k] + step_size * ode_func(lowercase_ , y[k] )
__SCREAMING_SNAKE_CASE : int = y[k] + (
(step_size / 2) * (ode_func(lowercase_ , y[k] ) + ode_func(x + step_size , lowercase_ ))
)
x += step_size
return y
if __name__ == "__main__":
import doctest
doctest.testmod()
| 674 | 1 |
"""simple docstring"""
import numpy as np
def lowerCAmelCase_ ( lowercase_ : np.ndarray , lowercase_ : np.ndarray , lowercase_ : float = 1E-1_2 , lowercase_ : int = 100 , ):
'''simple docstring'''
assert np.shape(lowercase_ )[0] == np.shape(lowercase_ )[1]
# Ensure proper dimensionality.
assert np.shape(lowercase_ )[0] == np.shape(lowercase_ )[0]
# Ensure inputs are either both complex or both real
assert np.iscomplexobj(lowercase_ ) == np.iscomplexobj(lowercase_ )
__SCREAMING_SNAKE_CASE : Any = np.iscomplexobj(lowercase_ )
if is_complex:
# Ensure complex input_matrix is Hermitian
assert np.array_equal(lowercase_ , input_matrix.conj().T )
# Set convergence to False. Will define convergence when we exceed max_iterations
# or when we have small changes from one iteration to next.
__SCREAMING_SNAKE_CASE : Optional[Any] = False
__SCREAMING_SNAKE_CASE : Union[str, Any] = 0
__SCREAMING_SNAKE_CASE : Any = 0
__SCREAMING_SNAKE_CASE : Any = 1E1_2
while not convergence:
# Multiple matrix by the vector.
__SCREAMING_SNAKE_CASE : int = np.dot(lowercase_ , lowercase_ )
# Normalize the resulting output vector.
__SCREAMING_SNAKE_CASE : int = w / np.linalg.norm(lowercase_ )
# Find rayleigh quotient
# (faster than usual b/c we know vector is normalized already)
__SCREAMING_SNAKE_CASE : Optional[Any] = vector.conj().T if is_complex else vector.T
__SCREAMING_SNAKE_CASE : Optional[int] = np.dot(lowercase_ , np.dot(lowercase_ , lowercase_ ) )
# Check convergence.
__SCREAMING_SNAKE_CASE : int = np.abs(lambda_ - lambda_previous ) / lambda_
iterations += 1
if error <= error_tol or iterations >= max_iterations:
__SCREAMING_SNAKE_CASE : Optional[int] = True
__SCREAMING_SNAKE_CASE : str = lambda_
if is_complex:
__SCREAMING_SNAKE_CASE : Optional[int] = np.real(lambda_ )
return lambda_, vector
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.array([[41, 4, 20], [4, 26, 30], [20, 30, 50]] )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.array([41, 4, 20] )
__SCREAMING_SNAKE_CASE : Dict = real_input_matrix.astype(np.complexaaa )
__SCREAMING_SNAKE_CASE : List[str] = np.triu(1J * complex_input_matrix , 1 )
complex_input_matrix += imag_matrix
complex_input_matrix += -1 * imag_matrix.T
__SCREAMING_SNAKE_CASE : int = np.array([41, 4, 20] ).astype(np.complexaaa )
for problem_type in ["real", "complex"]:
if problem_type == "real":
__SCREAMING_SNAKE_CASE : int = real_input_matrix
__SCREAMING_SNAKE_CASE : str = real_vector
elif problem_type == "complex":
__SCREAMING_SNAKE_CASE : Dict = complex_input_matrix
__SCREAMING_SNAKE_CASE : Optional[int] = complex_vector
# Our implementation.
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : str = power_iteration(lowercase_ , lowercase_ )
# Numpy implementation.
# Get eigenvalues and eigenvectors using built-in numpy
# eigh (eigh used for symmetric or hermetian matrices).
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : int = np.linalg.eigh(lowercase_ )
# Last eigenvalue is the maximum one.
__SCREAMING_SNAKE_CASE : int = eigen_values[-1]
# Last column in this matrix is eigenvector corresponding to largest eigenvalue.
__SCREAMING_SNAKE_CASE : Dict = eigen_vectors[:, -1]
# Check our implementation and numpy gives close answers.
assert np.abs(eigen_value - eigen_value_max ) <= 1E-6
# Take absolute values element wise of each eigenvector.
# as they are only unique to a minus sign.
assert np.linalg.norm(np.abs(lowercase_ ) - np.abs(lowercase_ ) ) <= 1E-6
if __name__ == "__main__":
import doctest
doctest.testmod()
test_power_iteration()
| 674 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...file_utils import _LazyModule, is_torch_available
from ...utils import OptionalDependencyNotAvailable
_lowerCamelCase = {
'''configuration_gpt_neox_japanese''': ['''GPT_NEOX_JAPANESE_PRETRAINED_CONFIG_ARCHIVE_MAP''', '''GPTNeoXJapaneseConfig'''],
'''tokenization_gpt_neox_japanese''': ['''GPTNeoXJapaneseTokenizer'''],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCamelCase = [
'''GPT_NEOX_JAPANESE_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''GPTNeoXJapaneseForCausalLM''',
'''GPTNeoXJapaneseLayer''',
'''GPTNeoXJapaneseModel''',
'''GPTNeoXJapanesePreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_gpt_neox_japanese import GPT_NEOX_JAPANESE_PRETRAINED_CONFIG_ARCHIVE_MAP, GPTNeoXJapaneseConfig
from .tokenization_gpt_neox_japanese import GPTNeoXJapaneseTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_gpt_neox_japanese import (
GPT_NEOX_JAPANESE_PRETRAINED_MODEL_ARCHIVE_LIST,
GPTNeoXJapaneseForCausalLM,
GPTNeoXJapaneseLayer,
GPTNeoXJapaneseModel,
GPTNeoXJapanesePreTrainedModel,
)
else:
import sys
_lowerCamelCase = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 674 | 1 |
"""simple docstring"""
import os
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, BatchEncoding, PreTrainedTokenizer
from ...utils import logging
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = '''▁'''
_lowerCamelCase = {'''vocab_file''': '''sentencepiece.bpe.model'''}
_lowerCamelCase = {
'''vocab_file''': {
'''facebook/mbart-large-50-one-to-many-mmt''': (
'''https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/sentencepiece.bpe.model'''
),
}
}
_lowerCamelCase = {
'''facebook/mbart-large-50-one-to-many-mmt''': 10_24,
}
# fmt: off
_lowerCamelCase = ['''ar_AR''', '''cs_CZ''', '''de_DE''', '''en_XX''', '''es_XX''', '''et_EE''', '''fi_FI''', '''fr_XX''', '''gu_IN''', '''hi_IN''', '''it_IT''', '''ja_XX''', '''kk_KZ''', '''ko_KR''', '''lt_LT''', '''lv_LV''', '''my_MM''', '''ne_NP''', '''nl_XX''', '''ro_RO''', '''ru_RU''', '''si_LK''', '''tr_TR''', '''vi_VN''', '''zh_CN''', '''af_ZA''', '''az_AZ''', '''bn_IN''', '''fa_IR''', '''he_IL''', '''hr_HR''', '''id_ID''', '''ka_GE''', '''km_KH''', '''mk_MK''', '''ml_IN''', '''mn_MN''', '''mr_IN''', '''pl_PL''', '''ps_AF''', '''pt_XX''', '''sv_SE''', '''sw_KE''', '''ta_IN''', '''te_IN''', '''th_TH''', '''tl_XX''', '''uk_UA''', '''ur_PK''', '''xh_ZA''', '''gl_ES''', '''sl_SI''']
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = VOCAB_FILES_NAMES
lowerCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
lowerCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
lowerCamelCase__ = ['''input_ids''', '''attention_mask''']
lowerCamelCase__ = []
lowerCamelCase__ = []
def __init__( self :Tuple , _lowerCamelCase :Dict , _lowerCamelCase :Tuple=None , _lowerCamelCase :Optional[int]=None , _lowerCamelCase :int="</s>" , _lowerCamelCase :Optional[Any]="</s>" , _lowerCamelCase :Union[str, Any]="<s>" , _lowerCamelCase :int="<unk>" , _lowerCamelCase :List[str]="<pad>" , _lowerCamelCase :int="<mask>" , _lowerCamelCase :Optional[Dict[str, Any]] = None , **_lowerCamelCase :List[Any] , ):
# Mask token behave like a normal word, i.e. include the space before it
__SCREAMING_SNAKE_CASE : Any = AddedToken(_lowerCamelCase , lstrip=_lowerCamelCase , rstrip=_lowerCamelCase ) if isinstance(_lowerCamelCase , _lowerCamelCase ) else mask_token
__SCREAMING_SNAKE_CASE : List[Any] = {} if sp_model_kwargs is None else sp_model_kwargs
__SCREAMING_SNAKE_CASE : Dict = kwargs.get('''additional_special_tokens''' , [] )
kwargs["additional_special_tokens"] += [
code for code in FAIRSEQ_LANGUAGE_CODES if code not in kwargs["additional_special_tokens"]
]
super().__init__(
src_lang=_lowerCamelCase , tgt_lang=_lowerCamelCase , eos_token=_lowerCamelCase , unk_token=_lowerCamelCase , sep_token=_lowerCamelCase , cls_token=_lowerCamelCase , pad_token=_lowerCamelCase , mask_token=_lowerCamelCase , sp_model_kwargs=self.sp_model_kwargs , **_lowerCamelCase , )
__SCREAMING_SNAKE_CASE : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(str(_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : int = vocab_file
# Original fairseq vocab and spm vocab must be "aligned":
# Vocab | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9
# -------- | ------- | ------- | ------ | ------- | --- | --- | --- | ----- | ----- | ----
# fairseq | '<s>' | '<pad>' | '</s>' | '<unk>' | ',' | '.' | '▁' | 's' | '▁de' | '-'
# spm | '<unk>' | '<s>' | '</s>' | ',' | '.' | '▁' | 's' | '▁de' | '-' | '▁a'
# Mimic fairseq token-to-id alignment for the first 4 token
__SCREAMING_SNAKE_CASE : List[Any] = {'''<s>''': 0, '''<pad>''': 1, '''</s>''': 2, '''<unk>''': 3}
# The first "real" token "," has position 4 in the original fairseq vocab and position 3 in the spm vocab
__SCREAMING_SNAKE_CASE : Union[str, Any] = 1
__SCREAMING_SNAKE_CASE : Optional[int] = len(self.sp_model )
__SCREAMING_SNAKE_CASE : List[str] = {
code: self.sp_model_size + i + self.fairseq_offset for i, code in enumerate(_lowerCamelCase )
}
__SCREAMING_SNAKE_CASE : Tuple = {v: k for k, v in self.lang_code_to_id.items()}
__SCREAMING_SNAKE_CASE : Union[str, Any] = len(self.sp_model ) + len(self.lang_code_to_id ) + self.fairseq_offset
self.fairseq_tokens_to_ids.update(self.lang_code_to_id )
__SCREAMING_SNAKE_CASE : Optional[int] = {v: k for k, v in self.fairseq_tokens_to_ids.items()}
__SCREAMING_SNAKE_CASE : List[str] = src_lang if src_lang is not None else '''en_XX'''
__SCREAMING_SNAKE_CASE : Any = self.lang_code_to_id[self._src_lang]
__SCREAMING_SNAKE_CASE : Tuple = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
return len(self.sp_model ) + len(self.lang_code_to_id ) + self.fairseq_offset + 1 # Plus 1 for the mask token
@property
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
return self._src_lang
@src_lang.setter
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] , _lowerCamelCase :str ):
__SCREAMING_SNAKE_CASE : List[str] = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def __getstate__( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Any = self.__dict__.copy()
__SCREAMING_SNAKE_CASE : Dict = None
return state
def __setstate__( self :Tuple , _lowerCamelCase :Dict ):
__SCREAMING_SNAKE_CASE : int = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs''' ):
__SCREAMING_SNAKE_CASE : List[Any] = {}
__SCREAMING_SNAKE_CASE : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = {self.convert_ids_to_tokens(_lowerCamelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :str ):
return self.sp_model.encode(_lowerCamelCase , out_type=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :str ):
if token in self.fairseq_tokens_to_ids:
return self.fairseq_tokens_to_ids[token]
__SCREAMING_SNAKE_CASE : int = self.sp_model.PieceToId(_lowerCamelCase )
# Need to return unknown token if the SP model returned 0
return spm_id + self.fairseq_offset if spm_id else self.unk_token_id
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :int ):
if index in self.fairseq_ids_to_tokens:
return self.fairseq_ids_to_tokens[index]
return self.sp_model.IdToPiece(index - self.fairseq_offset )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : Tuple = []
__SCREAMING_SNAKE_CASE : Any = ''''''
__SCREAMING_SNAKE_CASE : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_lowerCamelCase ) + token
__SCREAMING_SNAKE_CASE : Union[str, Any] = True
__SCREAMING_SNAKE_CASE : Optional[int] = []
else:
current_sub_tokens.append(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = False
out_string += self.sp_model.decode(_lowerCamelCase )
return out_string.strip()
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :str , _lowerCamelCase :Optional[str] = None ):
if not os.path.isdir(_lowerCamelCase ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__SCREAMING_SNAKE_CASE : Tuple = os.path.join(
_lowerCamelCase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_lowerCamelCase ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _lowerCamelCase )
elif not os.path.isfile(self.vocab_file ):
with open(_lowerCamelCase , '''wb''' ) as fi:
__SCREAMING_SNAKE_CASE : Optional[int] = self.sp_model.serialized_model_proto()
fi.write(_lowerCamelCase )
return (out_vocab_file,)
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :List[int] , _lowerCamelCase :Optional[List[int]] = None , _lowerCamelCase :bool = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_lowerCamelCase , token_ids_a=_lowerCamelCase , already_has_special_tokens=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = [1] * len(self.prefix_tokens )
__SCREAMING_SNAKE_CASE : Any = [1] * len(self.suffix_tokens )
if token_ids_a is None:
return prefix_ones + ([0] * len(_lowerCamelCase )) + suffix_ones
return prefix_ones + ([0] * len(_lowerCamelCase )) + ([0] * len(_lowerCamelCase )) + suffix_ones
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :List[int] , _lowerCamelCase :Optional[List[int]] = None ):
if token_ids_a is None:
return self.prefix_tokens + token_ids_a + self.suffix_tokens
# We don't expect to process pairs, but leave the pair logic for API consistency
return self.prefix_tokens + token_ids_a + token_ids_a + self.suffix_tokens
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :str , _lowerCamelCase :Optional[str] , _lowerCamelCase :Optional[str] , **_lowerCamelCase :List[Any] ):
if src_lang is None or tgt_lang is None:
raise ValueError('''Translation requires a `src_lang` and a `tgt_lang` for this model''' )
__SCREAMING_SNAKE_CASE : Optional[int] = src_lang
__SCREAMING_SNAKE_CASE : Union[str, Any] = self(_lowerCamelCase , add_special_tokens=_lowerCamelCase , return_tensors=_lowerCamelCase , **_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = self.convert_tokens_to_ids(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = tgt_lang_id
return inputs
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] , _lowerCamelCase :List[str] , _lowerCamelCase :str = "en_XX" , _lowerCamelCase :Optional[List[str]] = None , _lowerCamelCase :str = "ro_RO" , **_lowerCamelCase :Optional[int] , ):
__SCREAMING_SNAKE_CASE : Optional[Any] = src_lang
__SCREAMING_SNAKE_CASE : int = tgt_lang
return super().prepare_seqaseq_batch(_lowerCamelCase , _lowerCamelCase , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
return self.set_src_lang_special_tokens(self.src_lang )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :str ):
__SCREAMING_SNAKE_CASE : Dict = self.lang_code_to_id[src_lang]
__SCREAMING_SNAKE_CASE : List[Any] = [self.cur_lang_code_id]
__SCREAMING_SNAKE_CASE : List[Any] = [self.eos_token_id]
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] , _lowerCamelCase :str ):
__SCREAMING_SNAKE_CASE : str = self.lang_code_to_id[tgt_lang]
__SCREAMING_SNAKE_CASE : Dict = [self.cur_lang_code_id]
__SCREAMING_SNAKE_CASE : List[str] = [self.eos_token_id]
| 674 |
"""simple docstring"""
from __future__ import annotations
from typing import Any
class snake_case :
def __init__( self :Optional[Any] , _lowerCamelCase :int ):
__SCREAMING_SNAKE_CASE : int = num_of_nodes
__SCREAMING_SNAKE_CASE : list[list[int]] = []
__SCREAMING_SNAKE_CASE : dict[int, int] = {}
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :int , _lowerCamelCase :int , _lowerCamelCase :int ):
self.m_edges.append([u_node, v_node, weight] )
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :int ):
if self.m_component[u_node] == u_node:
return u_node
return self.find_component(self.m_component[u_node] )
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :int ):
if self.m_component[u_node] != u_node:
for k in self.m_component:
__SCREAMING_SNAKE_CASE : Optional[Any] = self.find_component(_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :list[int] , _lowerCamelCase :int , _lowerCamelCase :int ):
if component_size[u_node] <= component_size[v_node]:
__SCREAMING_SNAKE_CASE : List[Any] = v_node
component_size[v_node] += component_size[u_node]
self.set_component(_lowerCamelCase )
elif component_size[u_node] >= component_size[v_node]:
__SCREAMING_SNAKE_CASE : Dict = self.find_component(_lowerCamelCase )
component_size[u_node] += component_size[v_node]
self.set_component(_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Optional[int] = []
__SCREAMING_SNAKE_CASE : str = 0
__SCREAMING_SNAKE_CASE : list[Any] = [-1] * self.m_num_of_nodes
# A list of components (initialized to all of the nodes)
for node in range(self.m_num_of_nodes ):
self.m_component.update({node: node} )
component_size.append(1 )
__SCREAMING_SNAKE_CASE : str = self.m_num_of_nodes
while num_of_components > 1:
for edge in self.m_edges:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = edge
__SCREAMING_SNAKE_CASE : Optional[Any] = self.m_component[u]
__SCREAMING_SNAKE_CASE : int = self.m_component[v]
if u_component != v_component:
for component in (u_component, v_component):
if (
minimum_weight_edge[component] == -1
or minimum_weight_edge[component][2] > w
):
__SCREAMING_SNAKE_CASE : Optional[Any] = [u, v, w]
for edge in minimum_weight_edge:
if isinstance(_lowerCamelCase , _lowerCamelCase ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : int = edge
__SCREAMING_SNAKE_CASE : Tuple = self.m_component[u]
__SCREAMING_SNAKE_CASE : int = self.m_component[v]
if u_component != v_component:
mst_weight += w
self.union(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
print(f'''Added edge [{u} - {v}]\nAdded weight: {w}\n''' )
num_of_components -= 1
__SCREAMING_SNAKE_CASE : Optional[Any] = [-1] * self.m_num_of_nodes
print(f'''The total weight of the minimal spanning tree is: {mst_weight}''' )
def lowerCAmelCase_ ( ):
'''simple docstring'''
if __name__ == "__main__":
import doctest
doctest.testmod()
| 674 | 1 |
"""simple docstring"""
import argparse
import pickle
import numpy as np
import torch
from torch import nn
from transformers import ReformerConfig, ReformerModelWithLMHead
from transformers.utils import logging
logging.set_verbosity_info()
def lowerCAmelCase_ ( lowercase_ : int , lowercase_ : Any , lowercase_ : int=None ):
'''simple docstring'''
assert torch_layer.weight.shape == weight.shape, F'''{torch_layer} layer.weight does not match'''
__SCREAMING_SNAKE_CASE : str = nn.Parameter(lowercase_ )
if bias is not None:
assert torch_layer.bias.shape == bias.shape, F'''{torch_layer} layer.bias does not match'''
__SCREAMING_SNAKE_CASE : Tuple = nn.Parameter(lowercase_ )
def lowerCAmelCase_ ( lowercase_ : Tuple , lowercase_ : int , lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = np.asarray(weights[0] )
__SCREAMING_SNAKE_CASE : Optional[int] = np.asarray(weights[1] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(weights[2] )
set_param(
torch_layer.self_attention.query_key , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.value , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.output.dense , torch.tensor(lowercase_ ).view(-1 , lowercase_ ).contiguous().transpose(0 , 1 ) , )
def lowerCAmelCase_ ( lowercase_ : Dict , lowercase_ : List[str] , lowercase_ : List[str] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(weights[0] )
__SCREAMING_SNAKE_CASE : Any = np.asarray(weights[1] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(weights[2] )
__SCREAMING_SNAKE_CASE : Tuple = np.asarray(weights[3] )
set_param(
torch_layer.self_attention.query , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.key , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.value , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.output.dense , torch.tensor(lowercase_ ).view(-1 , lowercase_ ).contiguous().transpose(0 , 1 ) , )
def lowerCAmelCase_ ( lowercase_ : int , lowercase_ : List[str] , lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = weights[0][0][0]
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(layer_norm_a[0] )
__SCREAMING_SNAKE_CASE : List[Any] = np.asarray(layer_norm_a[1] )
set_param(
torch_block.attention.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# lsh weights + output
__SCREAMING_SNAKE_CASE : Tuple = weights[0][1]
if len(lowercase_ ) < 4:
set_layer_weights_in_torch_lsh(lowercase_ , torch_block.attention , lowercase_ )
else:
set_layer_weights_in_torch_local(lowercase_ , torch_block.attention , lowercase_ )
# intermediate weighs
__SCREAMING_SNAKE_CASE : Any = weights[2][0][1][2]
# Chunked Feed Forward
if len(lowercase_ ) == 4:
__SCREAMING_SNAKE_CASE : List[str] = intermediate_weights[2]
# layernorm 2
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(intermediate_weights[0][0] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(intermediate_weights[0][1] )
set_param(
torch_block.feed_forward.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# intermediate dense
__SCREAMING_SNAKE_CASE : int = np.asarray(intermediate_weights[1][0] )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(intermediate_weights[1][1] )
set_param(
torch_block.feed_forward.dense.dense , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
# intermediate out
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(intermediate_weights[4][0] )
__SCREAMING_SNAKE_CASE : Any = np.asarray(intermediate_weights[4][1] )
set_param(
torch_block.feed_forward.output.dense , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
def lowerCAmelCase_ ( lowercase_ : str , lowercase_ : Optional[Any] , lowercase_ : Optional[Any] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = torch_model.reformer
# word embeds
__SCREAMING_SNAKE_CASE : int = np.asarray(weights[1] )
set_param(
torch_model_reformer.embeddings.word_embeddings , torch.tensor(lowercase_ ) , )
if isinstance(weights[3] , lowercase_ ):
__SCREAMING_SNAKE_CASE : int = torch_model_reformer.embeddings.position_embeddings
for emb_idx in range(len(position_embeddings.weights ) ):
__SCREAMING_SNAKE_CASE : Dict = np.asarray(weights[3][emb_idx][0] )
assert (
position_embeddings.weights[emb_idx].shape == emb_weights.shape
), F'''{position_embeddings[emb_idx]} emb does not match'''
__SCREAMING_SNAKE_CASE : str = nn.Parameter(torch.tensor(lowercase_ ) )
__SCREAMING_SNAKE_CASE : List[Any] = weights[5]
assert len(torch_model_reformer.encoder.layers ) * 4 == len(
lowercase_ ), "HF and trax model do not have the same number of layers"
for layer_idx, layer in enumerate(torch_model_reformer.encoder.layers ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = trax_layer_weights[4 * layer_idx : 4 * (layer_idx + 1)]
set_block_weights_in_torch(lowercase_ , lowercase_ , lowercase_ )
# output layer norm
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(weights[7][0] )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(weights[7][1] )
set_param(
torch_model_reformer.encoder.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# output embeddings
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(weights[9][0] )
__SCREAMING_SNAKE_CASE : List[Any] = np.asarray(weights[9][1] )
set_param(
torch_model.lm_head.decoder , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
def lowerCAmelCase_ ( lowercase_ : List[str] , lowercase_ : Any , lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = ReformerConfig.from_json_file(lowercase_ )
print(F'''Building PyTorch model from configuration: {config}''' )
__SCREAMING_SNAKE_CASE : List[str] = ReformerModelWithLMHead(lowercase_ )
with open(lowercase_ , '''rb''' ) as f:
__SCREAMING_SNAKE_CASE : int = pickle.load(lowercase_ )['''weights''']
set_model_weights_in_torch(lowercase_ , lowercase_ , config.hidden_size )
# Save pytorch-model
print(F'''Save PyTorch model to {pytorch_dump_path}''' )
torch.save(model.state_dict() , lowercase_ )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--trax_model_pkl_path''', default=None, type=str, required=True, help='''Path to the TensorFlow checkpoint path.'''
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
required=True,
help=(
'''The config json file corresponding to the pre-trained Reformer model. \n'''
'''This specifies the model architecture.'''
),
)
parser.add_argument(
'''--pytorch_dump_path''', default=None, type=str, required=True, help='''Path to the output PyTorch model.'''
)
_lowerCamelCase = parser.parse_args()
convert_trax_checkpoint_to_pytorch(args.trax_model_pkl_path, args.config_file, args.pytorch_dump_path)
| 674 |
"""simple docstring"""
import argparse
import pickle
import numpy as np
import torch
from torch import nn
from transformers import ReformerConfig, ReformerModelWithLMHead
from transformers.utils import logging
logging.set_verbosity_info()
def lowerCAmelCase_ ( lowercase_ : int , lowercase_ : Any , lowercase_ : int=None ):
'''simple docstring'''
assert torch_layer.weight.shape == weight.shape, F'''{torch_layer} layer.weight does not match'''
__SCREAMING_SNAKE_CASE : str = nn.Parameter(lowercase_ )
if bias is not None:
assert torch_layer.bias.shape == bias.shape, F'''{torch_layer} layer.bias does not match'''
__SCREAMING_SNAKE_CASE : Tuple = nn.Parameter(lowercase_ )
def lowerCAmelCase_ ( lowercase_ : Tuple , lowercase_ : int , lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = np.asarray(weights[0] )
__SCREAMING_SNAKE_CASE : Optional[int] = np.asarray(weights[1] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(weights[2] )
set_param(
torch_layer.self_attention.query_key , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.value , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.output.dense , torch.tensor(lowercase_ ).view(-1 , lowercase_ ).contiguous().transpose(0 , 1 ) , )
def lowerCAmelCase_ ( lowercase_ : Dict , lowercase_ : List[str] , lowercase_ : List[str] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(weights[0] )
__SCREAMING_SNAKE_CASE : Any = np.asarray(weights[1] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(weights[2] )
__SCREAMING_SNAKE_CASE : Tuple = np.asarray(weights[3] )
set_param(
torch_layer.self_attention.query , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.key , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.value , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.output.dense , torch.tensor(lowercase_ ).view(-1 , lowercase_ ).contiguous().transpose(0 , 1 ) , )
def lowerCAmelCase_ ( lowercase_ : int , lowercase_ : List[str] , lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = weights[0][0][0]
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(layer_norm_a[0] )
__SCREAMING_SNAKE_CASE : List[Any] = np.asarray(layer_norm_a[1] )
set_param(
torch_block.attention.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# lsh weights + output
__SCREAMING_SNAKE_CASE : Tuple = weights[0][1]
if len(lowercase_ ) < 4:
set_layer_weights_in_torch_lsh(lowercase_ , torch_block.attention , lowercase_ )
else:
set_layer_weights_in_torch_local(lowercase_ , torch_block.attention , lowercase_ )
# intermediate weighs
__SCREAMING_SNAKE_CASE : Any = weights[2][0][1][2]
# Chunked Feed Forward
if len(lowercase_ ) == 4:
__SCREAMING_SNAKE_CASE : List[str] = intermediate_weights[2]
# layernorm 2
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(intermediate_weights[0][0] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(intermediate_weights[0][1] )
set_param(
torch_block.feed_forward.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# intermediate dense
__SCREAMING_SNAKE_CASE : int = np.asarray(intermediate_weights[1][0] )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(intermediate_weights[1][1] )
set_param(
torch_block.feed_forward.dense.dense , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
# intermediate out
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(intermediate_weights[4][0] )
__SCREAMING_SNAKE_CASE : Any = np.asarray(intermediate_weights[4][1] )
set_param(
torch_block.feed_forward.output.dense , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
def lowerCAmelCase_ ( lowercase_ : str , lowercase_ : Optional[Any] , lowercase_ : Optional[Any] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = torch_model.reformer
# word embeds
__SCREAMING_SNAKE_CASE : int = np.asarray(weights[1] )
set_param(
torch_model_reformer.embeddings.word_embeddings , torch.tensor(lowercase_ ) , )
if isinstance(weights[3] , lowercase_ ):
__SCREAMING_SNAKE_CASE : int = torch_model_reformer.embeddings.position_embeddings
for emb_idx in range(len(position_embeddings.weights ) ):
__SCREAMING_SNAKE_CASE : Dict = np.asarray(weights[3][emb_idx][0] )
assert (
position_embeddings.weights[emb_idx].shape == emb_weights.shape
), F'''{position_embeddings[emb_idx]} emb does not match'''
__SCREAMING_SNAKE_CASE : str = nn.Parameter(torch.tensor(lowercase_ ) )
__SCREAMING_SNAKE_CASE : List[Any] = weights[5]
assert len(torch_model_reformer.encoder.layers ) * 4 == len(
lowercase_ ), "HF and trax model do not have the same number of layers"
for layer_idx, layer in enumerate(torch_model_reformer.encoder.layers ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = trax_layer_weights[4 * layer_idx : 4 * (layer_idx + 1)]
set_block_weights_in_torch(lowercase_ , lowercase_ , lowercase_ )
# output layer norm
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(weights[7][0] )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(weights[7][1] )
set_param(
torch_model_reformer.encoder.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# output embeddings
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(weights[9][0] )
__SCREAMING_SNAKE_CASE : List[Any] = np.asarray(weights[9][1] )
set_param(
torch_model.lm_head.decoder , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
def lowerCAmelCase_ ( lowercase_ : List[str] , lowercase_ : Any , lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = ReformerConfig.from_json_file(lowercase_ )
print(F'''Building PyTorch model from configuration: {config}''' )
__SCREAMING_SNAKE_CASE : List[str] = ReformerModelWithLMHead(lowercase_ )
with open(lowercase_ , '''rb''' ) as f:
__SCREAMING_SNAKE_CASE : int = pickle.load(lowercase_ )['''weights''']
set_model_weights_in_torch(lowercase_ , lowercase_ , config.hidden_size )
# Save pytorch-model
print(F'''Save PyTorch model to {pytorch_dump_path}''' )
torch.save(model.state_dict() , lowercase_ )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--trax_model_pkl_path''', default=None, type=str, required=True, help='''Path to the TensorFlow checkpoint path.'''
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
required=True,
help=(
'''The config json file corresponding to the pre-trained Reformer model. \n'''
'''This specifies the model architecture.'''
),
)
parser.add_argument(
'''--pytorch_dump_path''', default=None, type=str, required=True, help='''Path to the output PyTorch model.'''
)
_lowerCamelCase = parser.parse_args()
convert_trax_checkpoint_to_pytorch(args.trax_model_pkl_path, args.config_file, args.pytorch_dump_path)
| 674 | 1 |
"""simple docstring"""
import os
from argparse import ArgumentParser
from typing import List
import torch.utils.data
from datasets import Dataset, IterableDataset
from datasets.distributed import split_dataset_by_node
_lowerCamelCase = 4
_lowerCamelCase = 3
class snake_case ( __UpperCAmelCase ):
pass
def lowerCAmelCase_ ( lowercase_ : List[str] ):
'''simple docstring'''
for shard in shards:
for i in range(lowercase_ ):
yield {"i": i, "shard": shard}
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[Any] = int(os.environ['''RANK'''] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = int(os.environ['''WORLD_SIZE'''] )
__SCREAMING_SNAKE_CASE : List[str] = ArgumentParser()
parser.add_argument('''--streaming''' , type=lowercase_ )
parser.add_argument('''--local_rank''' , type=lowercase_ )
parser.add_argument('''--num_workers''' , type=lowercase_ , default=0 )
__SCREAMING_SNAKE_CASE : Tuple = parser.parse_args()
__SCREAMING_SNAKE_CASE : int = args.streaming
__SCREAMING_SNAKE_CASE : Any = args.num_workers
__SCREAMING_SNAKE_CASE : Union[str, Any] = {'''shards''': [F'''shard_{shard_idx}''' for shard_idx in range(lowercase_ )]}
__SCREAMING_SNAKE_CASE : Optional[int] = IterableDataset.from_generator(lowercase_ , gen_kwargs=lowercase_ )
if not streaming:
__SCREAMING_SNAKE_CASE : List[str] = Dataset.from_list(list(lowercase_ ) )
__SCREAMING_SNAKE_CASE : List[Any] = split_dataset_by_node(lowercase_ , rank=lowercase_ , world_size=lowercase_ )
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.utils.data.DataLoader(lowercase_ , num_workers=lowercase_ )
__SCREAMING_SNAKE_CASE : Any = NUM_SHARDS * NUM_ITEMS_PER_SHARD
__SCREAMING_SNAKE_CASE : int = full_size // world_size
expected_local_size += int(rank < (full_size % world_size) )
__SCREAMING_SNAKE_CASE : Dict = sum(1 for _ in dataloader )
if local_size != expected_local_size:
raise FailedTestError(F'''local_size {local_size} != expected_local_size {expected_local_size}''' )
if __name__ == "__main__":
main()
| 674 |
"""simple docstring"""
from typing import Callable, Optional, Union
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = {
'''microsoft/xprophetnet-large-wiki100-cased''': (
'''https://huggingface.co/microsoft/xprophetnet-large-wiki100-cased/resolve/main/config.json'''
),
}
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = '''xlm-prophetnet'''
lowerCamelCase__ = ['''past_key_values''']
lowerCamelCase__ = {
'''num_attention_heads''': '''num_encoder_attention_heads''',
}
def __init__( self :List[str] , _lowerCamelCase :Optional[float] = 0.1 , _lowerCamelCase :Optional[Union[str, Callable]] = "gelu" , _lowerCamelCase :Optional[int] = 3_0_5_2_2 , _lowerCamelCase :Optional[int] = 1_0_2_4 , _lowerCamelCase :Optional[int] = 4_0_9_6 , _lowerCamelCase :Optional[int] = 1_2 , _lowerCamelCase :Optional[int] = 1_6 , _lowerCamelCase :Optional[int] = 4_0_9_6 , _lowerCamelCase :Optional[int] = 1_2 , _lowerCamelCase :Optional[int] = 1_6 , _lowerCamelCase :Optional[float] = 0.1 , _lowerCamelCase :Optional[float] = 0.1 , _lowerCamelCase :Optional[int] = 5_1_2 , _lowerCamelCase :Optional[float] = 0.0_2 , _lowerCamelCase :Optional[bool] = True , _lowerCamelCase :Optional[bool] = True , _lowerCamelCase :Optional[int] = 0 , _lowerCamelCase :Optional[int] = 2 , _lowerCamelCase :Optional[int] = 3_2 , _lowerCamelCase :Optional[int] = 1_2_8 , _lowerCamelCase :Optional[bool] = False , _lowerCamelCase :Optional[float] = 0.0 , _lowerCamelCase :Optional[bool] = True , _lowerCamelCase :Optional[int] = 0 , _lowerCamelCase :Optional[int] = 1 , _lowerCamelCase :Optional[int] = 2 , **_lowerCamelCase :int , ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = vocab_size
__SCREAMING_SNAKE_CASE : Optional[int] = hidden_size
__SCREAMING_SNAKE_CASE : List[Any] = encoder_ffn_dim
__SCREAMING_SNAKE_CASE : str = num_encoder_layers
__SCREAMING_SNAKE_CASE : Optional[Any] = num_encoder_attention_heads
__SCREAMING_SNAKE_CASE : str = decoder_ffn_dim
__SCREAMING_SNAKE_CASE : List[Any] = num_decoder_layers
__SCREAMING_SNAKE_CASE : List[str] = num_decoder_attention_heads
__SCREAMING_SNAKE_CASE : Dict = max_position_embeddings
__SCREAMING_SNAKE_CASE : Any = init_std # Normal(0, this parameter)
__SCREAMING_SNAKE_CASE : Any = activation_function
# parameters for xlmprophetnet
__SCREAMING_SNAKE_CASE : List[Any] = ngram
__SCREAMING_SNAKE_CASE : int = num_buckets
__SCREAMING_SNAKE_CASE : List[str] = relative_max_distance
__SCREAMING_SNAKE_CASE : str = disable_ngram_loss
__SCREAMING_SNAKE_CASE : Optional[int] = eps
# 3 Types of Dropout
__SCREAMING_SNAKE_CASE : int = attention_dropout
__SCREAMING_SNAKE_CASE : Optional[Any] = activation_dropout
__SCREAMING_SNAKE_CASE : Dict = dropout
__SCREAMING_SNAKE_CASE : Any = use_cache
super().__init__(
pad_token_id=_lowerCamelCase , bos_token_id=_lowerCamelCase , eos_token_id=_lowerCamelCase , is_encoder_decoder=_lowerCamelCase , add_cross_attention=_lowerCamelCase , decoder_start_token_id=_lowerCamelCase , **_lowerCamelCase , )
@property
def SCREAMING_SNAKE_CASE_ ( self :int ):
return self.num_encoder_layers + self.num_decoder_layers
@num_hidden_layers.setter
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :List[Any] ):
raise NotImplementedError(
'''This model does not support the setting of `num_hidden_layers`. Please set `num_encoder_layers` and'''
''' `num_decoder_layers`.''' )
| 674 | 1 |
"""simple docstring"""
_lowerCamelCase = [4, 1, 7, 4, 2, 6, 4, 1, 5, 3, 7, 5]
_lowerCamelCase = [3, 7, 7, 4, 2, 6, 4, 1, 5, 3, 7, 5]
_lowerCamelCase = {
0: '''Sunday''',
1: '''Monday''',
2: '''Tuesday''',
3: '''Wednesday''',
4: '''Thursday''',
5: '''Friday''',
6: '''Saturday''',
}
def lowerCAmelCase_ ( lowercase_ : int , lowercase_ : int , lowercase_ : int ):
'''simple docstring'''
assert len(str(lowercase_ ) ) > 2, "year should be in YYYY format"
assert 1 <= month <= 12, "month should be between 1 to 12"
assert 1 <= day <= 31, "day should be between 1 to 31"
# Doomsday algorithm:
__SCREAMING_SNAKE_CASE : List[str] = year // 100
__SCREAMING_SNAKE_CASE : str = (5 * (century % 4) + 2) % 7
__SCREAMING_SNAKE_CASE : Dict = year % 100
__SCREAMING_SNAKE_CASE : List[Any] = centurian % 12
__SCREAMING_SNAKE_CASE : Union[str, Any] = (
(centurian // 12) + centurian_m + (centurian_m // 4) + century_anchor
) % 7
__SCREAMING_SNAKE_CASE : List[Any] = (
DOOMSDAY_NOT_LEAP[month - 1]
if (year % 4 != 0) or (centurian == 0 and (year % 400) == 0)
else DOOMSDAY_LEAP[month - 1]
)
__SCREAMING_SNAKE_CASE : Union[str, Any] = (dooms_day + day - day_anchor) % 7
return WEEK_DAY_NAMES[week_day]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 674 |
"""simple docstring"""
import ast
import os
import re
import shutil
import tempfile
import unittest
from unittest import mock
import torch
from accelerate.test_utils.examples import compare_against_test
from accelerate.test_utils.testing import TempDirTestCase, require_trackers, run_command, slow
from accelerate.utils import write_basic_config
# DataLoaders built from `test_samples/MRPC` for quick testing
# Should mock `{script_name}.get_dataloaders` via:
# @mock.patch("{script_name}.get_dataloaders", mocked_dataloaders)
_lowerCamelCase = [
'''cross_validation.py''',
'''gradient_accumulation.py''',
'''local_sgd.py''',
'''multi_process_metrics.py''',
'''memory.py''',
'''automatic_gradient_accumulation.py''',
'''fsdp_with_peak_mem_tracking.py''',
'''deepspeed_with_config_support.py''',
'''megatron_lm_gpt_pretraining.py''',
]
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :str , _lowerCamelCase :bool , _lowerCamelCase :str = None , _lowerCamelCase :list = None ):
__SCREAMING_SNAKE_CASE : List[str] = None
__SCREAMING_SNAKE_CASE : Optional[Any] = os.path.abspath(os.path.join('''examples''' , '''by_feature''' ) )
__SCREAMING_SNAKE_CASE : Union[str, Any] = os.path.abspath('''examples''' )
for item in os.listdir(_lowerCamelCase ):
if item not in EXCLUDE_EXAMPLES:
__SCREAMING_SNAKE_CASE : List[Any] = os.path.join(_lowerCamelCase , _lowerCamelCase )
if os.path.isfile(_lowerCamelCase ) and ".py" in item_path:
with self.subTest(
tested_script=_lowerCamelCase , feature_script=_lowerCamelCase , tested_section='''main()''' if parser_only else '''training_function()''' , ):
__SCREAMING_SNAKE_CASE : Tuple = compare_against_test(
os.path.join(_lowerCamelCase , _lowerCamelCase ) , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = '''\n'''.join(_lowerCamelCase )
if special_strings is not None:
for string in special_strings:
__SCREAMING_SNAKE_CASE : List[Any] = diff.replace(_lowerCamelCase , '''''' )
self.assertEqual(_lowerCamelCase , '''''' )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
self.one_complete_example('''complete_nlp_example.py''' , _lowerCamelCase )
self.one_complete_example('''complete_nlp_example.py''' , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = os.path.abspath(os.path.join('''examples''' , '''cv_example.py''' ) )
__SCREAMING_SNAKE_CASE : Optional[int] = [
''' ''' * 1_6 + '''{\n\n''',
''' ''' * 2_0 + '''"accuracy": eval_metric["accuracy"],\n\n''',
''' ''' * 2_0 + '''"f1": eval_metric["f1"],\n\n''',
''' ''' * 2_0 + '''"train_loss": total_loss.item() / len(train_dataloader),\n\n''',
''' ''' * 2_0 + '''"epoch": epoch,\n\n''',
''' ''' * 1_6 + '''},\n\n''',
''' ''' * 1_6 + '''step=epoch,\n''',
''' ''' * 1_2,
''' ''' * 8 + '''for step, batch in enumerate(active_dataloader):\n''',
]
self.one_complete_example('''complete_cv_example.py''' , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
self.one_complete_example('''complete_cv_example.py''' , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
@mock.patch.dict(os.environ , {'''TESTING_MOCKED_DATALOADERS''': '''1'''} )
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = False
@classmethod
def SCREAMING_SNAKE_CASE_ ( cls :Dict ):
super().setUpClass()
__SCREAMING_SNAKE_CASE : Dict = tempfile.mkdtemp()
__SCREAMING_SNAKE_CASE : str = os.path.join(cls._tmpdir , '''default_config.yml''' )
write_basic_config(save_location=cls.configPath )
__SCREAMING_SNAKE_CASE : List[Any] = ['''accelerate''', '''launch''', '''--config_file''', cls.configPath]
@classmethod
def SCREAMING_SNAKE_CASE_ ( cls :Dict ):
super().tearDownClass()
shutil.rmtree(cls._tmpdir )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : str = f'''
examples/by_feature/checkpointing.py
--checkpointing_steps epoch
--output_dir {self.tmpdir}
'''.split()
run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(self.tmpdir , '''epoch_0''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Optional[Any] = f'''
examples/by_feature/checkpointing.py
--checkpointing_steps 1
--output_dir {self.tmpdir}
'''.split()
__SCREAMING_SNAKE_CASE : Optional[int] = run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(self.tmpdir , '''step_2''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Optional[int] = f'''
examples/by_feature/checkpointing.py
--resume_from_checkpoint {os.path.join(self.tmpdir , 'epoch_0' )}
'''.split()
__SCREAMING_SNAKE_CASE : Any = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
self.assertNotIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Optional[int] = f'''
examples/by_feature/checkpointing.py
--resume_from_checkpoint {os.path.join(self.tmpdir , 'step_2' )}
'''.split()
__SCREAMING_SNAKE_CASE : List[str] = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
if torch.cuda.is_available():
__SCREAMING_SNAKE_CASE : List[Any] = torch.cuda.device_count()
else:
__SCREAMING_SNAKE_CASE : Optional[int] = 1
if num_processes > 1:
self.assertNotIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
else:
self.assertIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = '''
examples/by_feature/cross_validation.py
--num_folds 2
'''.split()
with mock.patch.dict(os.environ , {'''TESTING_MOCKED_DATALOADERS''': '''0'''} ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = re.findall('''({.+})''' , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = [r for r in results if '''accuracy''' in r][-1]
__SCREAMING_SNAKE_CASE : Tuple = ast.literal_eval(_lowerCamelCase )
self.assertGreaterEqual(results['''accuracy'''] , 0.7_5 )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = ['''examples/by_feature/multi_process_metrics.py''']
run_command(self._launch_args + testargs )
@require_trackers
@mock.patch.dict(os.environ , {'''WANDB_MODE''': '''offline'''} )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
with tempfile.TemporaryDirectory() as tmpdir:
__SCREAMING_SNAKE_CASE : int = f'''
examples/by_feature/tracking.py
--with_tracking
--project_dir {tmpdir}
'''.split()
run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(_lowerCamelCase , '''tracking''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = ['''examples/by_feature/gradient_accumulation.py''']
run_command(self._launch_args + testargs )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[Any] = ['''examples/by_feature/local_sgd.py''']
run_command(self._launch_args + testargs )
| 674 | 1 |
"""simple docstring"""
import copy
import os
import cva
import numpy as np
from matplotlib import pyplot as plt
class snake_case :
def __init__( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Tuple = ''''''
__SCREAMING_SNAKE_CASE : Tuple = ''''''
__SCREAMING_SNAKE_CASE : Union[str, Any] = []
__SCREAMING_SNAKE_CASE : Any = 0
__SCREAMING_SNAKE_CASE : Any = 2_5_6
__SCREAMING_SNAKE_CASE : Union[str, Any] = 0
__SCREAMING_SNAKE_CASE : int = 0
__SCREAMING_SNAKE_CASE : int = 0
__SCREAMING_SNAKE_CASE : Dict = 0
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :List[Any] ):
__SCREAMING_SNAKE_CASE : str = cva.imread(_lowerCamelCase , 0 )
__SCREAMING_SNAKE_CASE : List[str] = copy.deepcopy(self.img )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = plt.hist(self.img.ravel() , 2_5_6 , [0, 2_5_6] , label='''x''' )
__SCREAMING_SNAKE_CASE : Optional[int] = np.sum(_lowerCamelCase )
for i in range(len(_lowerCamelCase ) ):
__SCREAMING_SNAKE_CASE : List[str] = x[i] / self.k
self.sk += prk
__SCREAMING_SNAKE_CASE : str = (self.L - 1) * self.sk
if self.rem != 0:
__SCREAMING_SNAKE_CASE : Optional[int] = int(last % last )
__SCREAMING_SNAKE_CASE : int = int(last + 1 if self.rem >= 0.5 else last )
self.last_list.append(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = int(np.ma.count(self.img ) / self.img[1].size )
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.img[1].size
for i in range(self.number_of_cols ):
for j in range(self.number_of_rows ):
__SCREAMING_SNAKE_CASE : Optional[int] = self.img[j][i]
if num != self.last_list[num]:
__SCREAMING_SNAKE_CASE : Dict = self.last_list[num]
cva.imwrite('''output_data/output.jpg''' , self.img )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
plt.hist(self.img.ravel() , 2_5_6 , [0, 2_5_6] )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
cva.imshow('''Output-Image''' , self.img )
cva.imshow('''Input-Image''' , self.original_image )
cva.waitKey(5_0_0_0 )
cva.destroyAllWindows()
if __name__ == "__main__":
_lowerCamelCase = os.path.join(os.path.basename(__file__), '''image_data/input.jpg''')
_lowerCamelCase = ConstantStretch()
stretcher.stretch(file_path)
stretcher.plot_histogram()
stretcher.show_image()
| 674 |
"""simple docstring"""
import argparse
import logging
import os
import time
import timeit
import datasets
import numpy as np
import pycuda.autoinit # noqa: F401
import pycuda.driver as cuda
import tensorrt as trt
import torch
from absl import logging as absl_logging
from accelerate import Accelerator
from datasets import load_dataset, load_metric
from torch.utils.data import DataLoader
from utils_qa import postprocess_qa_predictions
import transformers
from transformers import AutoTokenizer, EvalPrediction, default_data_collator, set_seed
from transformers.trainer_pt_utils import nested_concat, nested_truncate
_lowerCamelCase = trt.Logger(trt.Logger.WARNING)
_lowerCamelCase = absl_logging.get_absl_logger()
absl_logger.setLevel(logging.WARNING)
_lowerCamelCase = logging.getLogger(__name__)
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--onnx_model_path''',
default=None,
type=str,
required=True,
help='''Path to ONNX model: ''',
)
parser.add_argument(
'''--output_dir''',
default=None,
type=str,
required=True,
help='''The output directory where the model checkpoints and predictions will be written.''',
)
# Other parameters
parser.add_argument(
'''--tokenizer_name''',
default='''''',
type=str,
required=True,
help='''Pretrained tokenizer name or path if not the same as model_name''',
)
parser.add_argument(
'''--version_2_with_negative''',
action='''store_true''',
help='''If true, the SQuAD examples contain some that do not have an answer.''',
)
parser.add_argument(
'''--null_score_diff_threshold''',
type=float,
default=0.0,
help='''If null_score - best_non_null is greater than the threshold predict null.''',
)
parser.add_argument(
'''--max_seq_length''',
default=3_84,
type=int,
help=(
'''The maximum total input sequence length after WordPiece tokenization. Sequences '''
'''longer than this will be truncated, and sequences shorter than this will be padded.'''
),
)
parser.add_argument(
'''--doc_stride''',
default=1_28,
type=int,
help='''When splitting up a long document into chunks, how much stride to take between chunks.''',
)
parser.add_argument('''--per_device_eval_batch_size''', default=8, type=int, help='''Batch size per GPU/CPU for evaluation.''')
parser.add_argument(
'''--n_best_size''',
default=20,
type=int,
help='''The total number of n-best predictions to generate in the nbest_predictions.json output file.''',
)
parser.add_argument(
'''--max_answer_length''',
default=30,
type=int,
help=(
'''The maximum length of an answer that can be generated. This is needed because the start '''
'''and end predictions are not conditioned on one another.'''
),
)
parser.add_argument('''--seed''', type=int, default=42, help='''random seed for initialization''')
parser.add_argument(
'''--dataset_name''',
type=str,
default=None,
required=True,
help='''The name of the dataset to use (via the datasets library).''',
)
parser.add_argument(
'''--dataset_config_name''',
type=str,
default=None,
help='''The configuration name of the dataset to use (via the datasets library).''',
)
parser.add_argument(
'''--preprocessing_num_workers''', type=int, default=4, help='''A csv or a json file containing the training data.'''
)
parser.add_argument('''--overwrite_cache''', action='''store_true''', help='''Overwrite the cached training and evaluation sets''')
parser.add_argument(
'''--fp16''',
action='''store_true''',
help='''Whether to use 16-bit (mixed) precision instead of 32-bit''',
)
parser.add_argument(
'''--int8''',
action='''store_true''',
help='''Whether to use INT8''',
)
_lowerCamelCase = parser.parse_args()
if args.tokenizer_name:
_lowerCamelCase = AutoTokenizer.from_pretrained(args.tokenizer_name, use_fast=True)
else:
raise ValueError(
'''You are instantiating a new tokenizer from scratch. This is not supported by this script.'''
'''You can do it from another script, save it, and load it from here, using --tokenizer_name.'''
)
logger.info('''Training/evaluation parameters %s''', args)
_lowerCamelCase = args.per_device_eval_batch_size
_lowerCamelCase = (args.eval_batch_size, args.max_seq_length)
# TRT Engine properties
_lowerCamelCase = True
_lowerCamelCase = '''temp_engine/bert-fp32.engine'''
if args.fpaa:
_lowerCamelCase = '''temp_engine/bert-fp16.engine'''
if args.inta:
_lowerCamelCase = '''temp_engine/bert-int8.engine'''
# import ONNX file
if not os.path.exists('''temp_engine'''):
os.makedirs('''temp_engine''')
_lowerCamelCase = 1 << (int)(trt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH)
with trt.Builder(TRT_LOGGER) as builder, builder.create_network(EXPLICIT_BATCH) as network, trt.OnnxParser(
network, TRT_LOGGER
) as parser:
with open(args.onnx_model_path, '''rb''') as model:
if not parser.parse(model.read()):
for error in range(parser.num_errors):
print(parser.get_error(error))
# Query input names and shapes from parsed TensorRT network
_lowerCamelCase = [network.get_input(i) for i in range(network.num_inputs)]
_lowerCamelCase = [_input.name for _input in network_inputs] # ex: ["actual_input1"]
with builder.create_builder_config() as config:
_lowerCamelCase = 1 << 50
if STRICT_TYPES:
config.set_flag(trt.BuilderFlag.STRICT_TYPES)
if args.fpaa:
config.set_flag(trt.BuilderFlag.FPaa)
if args.inta:
config.set_flag(trt.BuilderFlag.INTa)
_lowerCamelCase = builder.create_optimization_profile()
config.add_optimization_profile(profile)
for i in range(len(input_names)):
profile.set_shape(input_names[i], INPUT_SHAPE, INPUT_SHAPE, INPUT_SHAPE)
_lowerCamelCase = builder.build_engine(network, config)
# serialize_engine and store in file (can be directly loaded and deserialized):
with open(engine_name, '''wb''') as f:
f.write(engine.serialize())
def lowerCAmelCase_ ( lowercase_ : List[Any] , lowercase_ : Tuple , lowercase_ : List[Any] , lowercase_ : List[Any] , lowercase_ : str , lowercase_ : Tuple , lowercase_ : Any , lowercase_ : Optional[int] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = np.asarray(inputs['''input_ids'''] , dtype=np.intaa )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(inputs['''attention_mask'''] , dtype=np.intaa )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(inputs['''token_type_ids'''] , dtype=np.intaa )
# Copy inputs
cuda.memcpy_htod_async(d_inputs[0] , input_ids.ravel() , lowercase_ )
cuda.memcpy_htod_async(d_inputs[1] , attention_mask.ravel() , lowercase_ )
cuda.memcpy_htod_async(d_inputs[2] , token_type_ids.ravel() , lowercase_ )
# start time
__SCREAMING_SNAKE_CASE : Tuple = time.time()
# Run inference
context.execute_async(
bindings=[int(lowercase_ ) for d_inp in d_inputs] + [int(lowercase_ ), int(lowercase_ )] , stream_handle=stream.handle )
# Transfer predictions back from GPU
cuda.memcpy_dtoh_async(lowercase_ , lowercase_ , lowercase_ )
cuda.memcpy_dtoh_async(lowercase_ , lowercase_ , lowercase_ )
# Synchronize the stream and take time
stream.synchronize()
# end time
__SCREAMING_SNAKE_CASE : List[str] = time.time()
__SCREAMING_SNAKE_CASE : int = end_time - start_time
__SCREAMING_SNAKE_CASE : int = (h_outputa, h_outputa)
# print(outputs)
return outputs, infer_time
# Initialize the accelerator. We will let the accelerator handle device placement for us in this example.
_lowerCamelCase = Accelerator()
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format='''%(asctime)s - %(levelname)s - %(name)s - %(message)s''',
datefmt='''%m/%d/%Y %H:%M:%S''',
level=logging.INFO,
)
# Setup logging, we only want one process per machine to log things on the screen.
# accelerator.is_local_main_process is only True for one process per machine.
logger.setLevel(logging.INFO if accelerator.is_local_main_process else logging.ERROR)
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_info()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
# If passed along, set the training seed now.
if args.seed is not None:
set_seed(args.seed)
# Get the datasets: you can either provide your own CSV/JSON/TXT training and evaluation files (see below)
# or just provide the name of one of the public datasets available on the hub at https://huggingface.co/datasets/
# (the dataset will be downloaded automatically from the datasets Hub).
#
# For CSV/JSON files, this script will use the column called 'text' or the first column if no column called
# 'text' is found. You can easily tweak this behavior (see below).
if args.dataset_name is not None:
# Downloading and loading a dataset from the hub.
_lowerCamelCase = load_dataset(args.dataset_name, args.dataset_config_name)
else:
raise ValueError('''Evaluation requires a dataset name''')
# See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
# https://huggingface.co/docs/datasets/loading_datasets.html.
# Preprocessing the datasets.
# Preprocessing is slighlty different for training and evaluation.
_lowerCamelCase = raw_datasets['''validation'''].column_names
_lowerCamelCase = '''question''' if '''question''' in column_names else column_names[0]
_lowerCamelCase = '''context''' if '''context''' in column_names else column_names[1]
_lowerCamelCase = '''answers''' if '''answers''' in column_names else column_names[2]
# Padding side determines if we do (question|context) or (context|question).
_lowerCamelCase = tokenizer.padding_side == '''right'''
if args.max_seq_length > tokenizer.model_max_length:
logger.warning(
f'The max_seq_length passed ({args.max_seq_length}) is larger than the maximum length for the'
f'model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}.'
)
_lowerCamelCase = min(args.max_seq_length, tokenizer.model_max_length)
def lowerCAmelCase_ ( lowercase_ : Tuple ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = [q.lstrip() for q in examples[question_column_name]]
# Tokenize our examples with truncation and maybe padding, but keep the overflows using a stride. This results
# in one example possible giving several features when a context is long, each of those features having a
# context that overlaps a bit the context of the previous feature.
__SCREAMING_SNAKE_CASE : Optional[Any] = tokenizer(
examples[question_column_name if pad_on_right else context_column_name] , examples[context_column_name if pad_on_right else question_column_name] , truncation='''only_second''' if pad_on_right else '''only_first''' , max_length=lowercase_ , stride=args.doc_stride , return_overflowing_tokens=lowercase_ , return_offsets_mapping=lowercase_ , padding='''max_length''' , )
# Since one example might give us several features if it has a long context, we need a map from a feature to
# its corresponding example. This key gives us just that.
__SCREAMING_SNAKE_CASE : Optional[int] = tokenized_examples.pop('''overflow_to_sample_mapping''' )
# For evaluation, we will need to convert our predictions to substrings of the context, so we keep the
# corresponding example_id and we will store the offset mappings.
__SCREAMING_SNAKE_CASE : Any = []
for i in range(len(tokenized_examples['''input_ids'''] ) ):
# Grab the sequence corresponding to that example (to know what is the context and what is the question).
__SCREAMING_SNAKE_CASE : int = tokenized_examples.sequence_ids(lowercase_ )
__SCREAMING_SNAKE_CASE : str = 1 if pad_on_right else 0
# One example can give several spans, this is the index of the example containing this span of text.
__SCREAMING_SNAKE_CASE : str = sample_mapping[i]
tokenized_examples["example_id"].append(examples['''id'''][sample_index] )
# Set to None the offset_mapping that are not part of the context so it's easy to determine if a token
# position is part of the context or not.
__SCREAMING_SNAKE_CASE : List[str] = [
(o if sequence_ids[k] == context_index else None)
for k, o in enumerate(tokenized_examples['''offset_mapping'''][i] )
]
return tokenized_examples
_lowerCamelCase = raw_datasets['''validation''']
# Validation Feature Creation
_lowerCamelCase = eval_examples.map(
prepare_validation_features,
batched=True,
num_proc=args.preprocessing_num_workers,
remove_columns=column_names,
load_from_cache_file=not args.overwrite_cache,
desc='''Running tokenizer on validation dataset''',
)
_lowerCamelCase = default_data_collator
_lowerCamelCase = eval_dataset.remove_columns(['''example_id''', '''offset_mapping'''])
_lowerCamelCase = DataLoader(
eval_dataset_for_model, collate_fn=data_collator, batch_size=args.per_device_eval_batch_size
)
def lowerCAmelCase_ ( lowercase_ : Dict , lowercase_ : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : List[Any]="eval" ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = postprocess_qa_predictions(
examples=lowercase_ , features=lowercase_ , predictions=lowercase_ , version_2_with_negative=args.version_2_with_negative , n_best_size=args.n_best_size , max_answer_length=args.max_answer_length , null_score_diff_threshold=args.null_score_diff_threshold , output_dir=args.output_dir , prefix=lowercase_ , )
# Format the result to the format the metric expects.
if args.version_2_with_negative:
__SCREAMING_SNAKE_CASE : Union[str, Any] = [
{'''id''': k, '''prediction_text''': v, '''no_answer_probability''': 0.0} for k, v in predictions.items()
]
else:
__SCREAMING_SNAKE_CASE : int = [{'''id''': k, '''prediction_text''': v} for k, v in predictions.items()]
__SCREAMING_SNAKE_CASE : Any = [{'''id''': ex['''id'''], '''answers''': ex[answer_column_name]} for ex in examples]
return EvalPrediction(predictions=lowercase_ , label_ids=lowercase_ )
_lowerCamelCase = load_metric('''squad_v2''' if args.version_2_with_negative else '''squad''')
# Evaluation!
logger.info('''Loading ONNX model %s for evaluation''', args.onnx_model_path)
with open(engine_name, '''rb''') as f, trt.Runtime(TRT_LOGGER) as runtime, runtime.deserialize_cuda_engine(
f.read()
) as engine, engine.create_execution_context() as context:
# setup for TRT inferrence
for i in range(len(input_names)):
context.set_binding_shape(i, INPUT_SHAPE)
assert context.all_binding_shapes_specified
def lowerCAmelCase_ ( lowercase_ : Any ):
'''simple docstring'''
return trt.volume(engine.get_binding_shape(lowercase_ ) ) * engine.get_binding_dtype(lowercase_ ).itemsize
# Allocate device memory for inputs and outputs.
_lowerCamelCase = [cuda.mem_alloc(binding_nbytes(binding)) for binding in engine if engine.binding_is_input(binding)]
# Allocate output buffer
_lowerCamelCase = cuda.pagelocked_empty(tuple(context.get_binding_shape(3)), dtype=np.floataa)
_lowerCamelCase = cuda.pagelocked_empty(tuple(context.get_binding_shape(4)), dtype=np.floataa)
_lowerCamelCase = cuda.mem_alloc(h_outputa.nbytes)
_lowerCamelCase = cuda.mem_alloc(h_outputa.nbytes)
# Create a stream in which to copy inputs/outputs and run inference.
_lowerCamelCase = cuda.Stream()
# Evaluation
logger.info('''***** Running Evaluation *****''')
logger.info(f' Num examples = {len(eval_dataset)}')
logger.info(f' Batch size = {args.per_device_eval_batch_size}')
_lowerCamelCase = 0.0
_lowerCamelCase = 0
_lowerCamelCase = timeit.default_timer()
_lowerCamelCase = None
for step, batch in enumerate(eval_dataloader):
_lowerCamelCase , _lowerCamelCase = model_infer(batch, context, d_inputs, h_outputa, h_outputa, d_outputa, d_outputa, stream)
total_time += infer_time
niter += 1
_lowerCamelCase , _lowerCamelCase = outputs
_lowerCamelCase = torch.tensor(start_logits)
_lowerCamelCase = torch.tensor(end_logits)
# necessary to pad predictions and labels for being gathered
_lowerCamelCase = accelerator.pad_across_processes(start_logits, dim=1, pad_index=-1_00)
_lowerCamelCase = accelerator.pad_across_processes(end_logits, dim=1, pad_index=-1_00)
_lowerCamelCase = (accelerator.gather(start_logits).cpu().numpy(), accelerator.gather(end_logits).cpu().numpy())
_lowerCamelCase = logits if all_preds is None else nested_concat(all_preds, logits, padding_index=-1_00)
if all_preds is not None:
_lowerCamelCase = nested_truncate(all_preds, len(eval_dataset))
_lowerCamelCase = timeit.default_timer() - start_time
logger.info(''' Evaluation done in total %f secs (%f sec per example)''', evalTime, evalTime / len(eval_dataset))
# Inference time from TRT
logger.info('''Average Inference Time = {:.3f} ms'''.format(total_time * 10_00 / niter))
logger.info('''Total Inference Time = {:.3f} ms'''.format(total_time * 10_00))
logger.info('''Total Number of Inference = %d''', niter)
_lowerCamelCase = post_processing_function(eval_examples, eval_dataset, all_preds)
_lowerCamelCase = metric.compute(predictions=prediction.predictions, references=prediction.label_ids)
logger.info(f'Evaluation metrics: {eval_metric}')
| 674 | 1 |
"""simple docstring"""
import inspect
import unittest
from transformers import ConvNextVaConfig
from transformers.models.auto import get_values
from transformers.models.auto.modeling_auto import MODEL_FOR_BACKBONE_MAPPING_NAMES, MODEL_MAPPING_NAMES
from transformers.testing_utils import require_torch, require_vision, slow, torch_device
from transformers.utils import cached_property, is_torch_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import ConvNextVaBackbone, ConvNextVaForImageClassification, ConvNextVaModel
from transformers.models.convnextva.modeling_convnextva import CONVNEXTV2_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class snake_case :
def __init__( self :Dict , _lowerCamelCase :List[Any] , _lowerCamelCase :Union[str, Any]=1_3 , _lowerCamelCase :List[Any]=3_2 , _lowerCamelCase :Tuple=3 , _lowerCamelCase :int=4 , _lowerCamelCase :List[Any]=[1_0, 2_0, 3_0, 4_0] , _lowerCamelCase :Optional[int]=[2, 2, 3, 2] , _lowerCamelCase :Dict=True , _lowerCamelCase :Tuple=True , _lowerCamelCase :int=3_7 , _lowerCamelCase :Optional[Any]="gelu" , _lowerCamelCase :Any=1_0 , _lowerCamelCase :Optional[Any]=0.0_2 , _lowerCamelCase :str=["stage2", "stage3", "stage4"] , _lowerCamelCase :List[Any]=[2, 3, 4] , _lowerCamelCase :int=None , ):
__SCREAMING_SNAKE_CASE : Optional[int] = parent
__SCREAMING_SNAKE_CASE : str = batch_size
__SCREAMING_SNAKE_CASE : int = image_size
__SCREAMING_SNAKE_CASE : Optional[Any] = num_channels
__SCREAMING_SNAKE_CASE : Tuple = num_stages
__SCREAMING_SNAKE_CASE : Union[str, Any] = hidden_sizes
__SCREAMING_SNAKE_CASE : Any = depths
__SCREAMING_SNAKE_CASE : int = is_training
__SCREAMING_SNAKE_CASE : str = use_labels
__SCREAMING_SNAKE_CASE : int = intermediate_size
__SCREAMING_SNAKE_CASE : Dict = hidden_act
__SCREAMING_SNAKE_CASE : Any = num_labels
__SCREAMING_SNAKE_CASE : Tuple = initializer_range
__SCREAMING_SNAKE_CASE : Any = out_features
__SCREAMING_SNAKE_CASE : int = out_indices
__SCREAMING_SNAKE_CASE : Dict = scope
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : Any = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
__SCREAMING_SNAKE_CASE : Optional[int] = None
if self.use_labels:
__SCREAMING_SNAKE_CASE : str = ids_tensor([self.batch_size] , self.num_labels )
__SCREAMING_SNAKE_CASE : str = self.get_config()
return config, pixel_values, labels
def SCREAMING_SNAKE_CASE_ ( self :int ):
return ConvNextVaConfig(
num_channels=self.num_channels , hidden_sizes=self.hidden_sizes , depths=self.depths , num_stages=self.num_stages , hidden_act=self.hidden_act , is_decoder=_lowerCamelCase , initializer_range=self.initializer_range , out_features=self.out_features , out_indices=self.out_indices , num_labels=self.num_labels , )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :int , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : Any = ConvNextVaModel(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : Any = model(_lowerCamelCase )
# expected last hidden states: B, C, H // 32, W // 32
self.parent.assertEqual(
result.last_hidden_state.shape , (self.batch_size, self.hidden_sizes[-1], self.image_size // 3_2, self.image_size // 3_2) , )
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :str , _lowerCamelCase :List[str] , _lowerCamelCase :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : List[Any] = ConvNextVaForImageClassification(_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : Optional[Any] = model(_lowerCamelCase , labels=_lowerCamelCase )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :Any , _lowerCamelCase :Any , _lowerCamelCase :Any ):
__SCREAMING_SNAKE_CASE : Optional[int] = ConvNextVaBackbone(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : Optional[int] = model(_lowerCamelCase )
# verify hidden states
self.parent.assertEqual(len(result.feature_maps ) , len(config.out_features ) )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [self.batch_size, self.hidden_sizes[1], 4, 4] )
# verify channels
self.parent.assertEqual(len(model.channels ) , len(config.out_features ) )
self.parent.assertListEqual(model.channels , config.hidden_sizes[1:] )
# verify backbone works with out_features=None
__SCREAMING_SNAKE_CASE : str = None
__SCREAMING_SNAKE_CASE : List[Any] = ConvNextVaBackbone(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : Tuple = model(_lowerCamelCase )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , 1 )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [self.batch_size, self.hidden_sizes[-1], 1, 1] )
# verify channels
self.parent.assertEqual(len(model.channels ) , 1 )
self.parent.assertListEqual(model.channels , [config.hidden_sizes[-1]] )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : List[str] = self.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = config_and_inputs
__SCREAMING_SNAKE_CASE : List[str] = {'''pixel_values''': pixel_values}
return config, inputs_dict
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Optional[Any] = self.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = config_and_inputs
__SCREAMING_SNAKE_CASE : Tuple = {'''pixel_values''': pixel_values, '''labels''': labels}
return config, inputs_dict
@require_torch
class snake_case ( __UpperCAmelCase , __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = (
(
ConvNextVaModel,
ConvNextVaForImageClassification,
ConvNextVaBackbone,
)
if is_torch_available()
else ()
)
lowerCamelCase__ = (
{'''feature-extraction''': ConvNextVaModel, '''image-classification''': ConvNextVaForImageClassification}
if is_torch_available()
else {}
)
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Optional[Any] = ConvNextVaModelTester(self )
__SCREAMING_SNAKE_CASE : int = ConfigTester(self , config_class=_lowerCamelCase , has_text_modality=_lowerCamelCase , hidden_size=3_7 )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
return
@unittest.skip(reason='''ConvNextV2 does not use inputs_embeds''' )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
pass
@unittest.skip(reason='''ConvNextV2 does not support input and output embeddings''' )
def SCREAMING_SNAKE_CASE_ ( self :str ):
pass
@unittest.skip(reason='''ConvNextV2 does not use feedforward chunking''' )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
pass
def SCREAMING_SNAKE_CASE_ ( self :Any ):
if not self.model_tester.is_training:
return
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = self.model_tester.prepare_config_and_inputs_with_labels()
__SCREAMING_SNAKE_CASE : List[Any] = True
if model_class.__name__ in [
*get_values(_lowerCamelCase ),
*get_values(_lowerCamelCase ),
]:
continue
__SCREAMING_SNAKE_CASE : List[str] = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.train()
__SCREAMING_SNAKE_CASE : List[str] = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = model(**_lowerCamelCase ).loss
loss.backward()
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
if not self.model_tester.is_training:
return
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : str = self.model_tester.prepare_config_and_inputs_with_labels()
__SCREAMING_SNAKE_CASE : Any = False
__SCREAMING_SNAKE_CASE : Union[str, Any] = True
if (
model_class.__name__
in [*get_values(_lowerCamelCase ), *get_values(_lowerCamelCase )]
or not model_class.supports_gradient_checkpointing
):
continue
__SCREAMING_SNAKE_CASE : str = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.gradient_checkpointing_enable()
model.train()
__SCREAMING_SNAKE_CASE : int = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = model(**_lowerCamelCase ).loss
loss.backward()
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : int = model_class(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
__SCREAMING_SNAKE_CASE : Optional[int] = [*signature.parameters.keys()]
__SCREAMING_SNAKE_CASE : str = ['''pixel_values''']
self.assertListEqual(arg_names[:1] , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Any ):
def check_hidden_states_output(_lowerCamelCase :Tuple , _lowerCamelCase :int , _lowerCamelCase :List[str] ):
__SCREAMING_SNAKE_CASE : Optional[Any] = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[Any] = model(**self._prepare_for_class(_lowerCamelCase , _lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : List[Any] = outputs.encoder_hidden_states if config.is_encoder_decoder else outputs.hidden_states
__SCREAMING_SNAKE_CASE : Tuple = self.model_tester.num_stages
self.assertEqual(len(_lowerCamelCase ) , expected_num_stages + 1 )
# ConvNextV2's feature maps are of shape (batch_size, num_channels, height, width)
self.assertListEqual(
list(hidden_states[0].shape[-2:] ) , [self.model_tester.image_size // 4, self.model_tester.image_size // 4] , )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Optional[Any] = True
check_hidden_states_output(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
__SCREAMING_SNAKE_CASE : Optional[Any] = True
check_hidden_states_output(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*_lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
for model_name in CONVNEXTV2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Optional[int] = ConvNextVaModel.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
return image
@require_torch
@require_vision
class snake_case ( unittest.TestCase ):
@cached_property
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
return AutoImageProcessor.from_pretrained('''facebook/convnextv2-tiny-1k-224''' ) if is_vision_available() else None
@slow
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : str = ConvNextVaForImageClassification.from_pretrained('''facebook/convnextv2-tiny-1k-224''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = self.default_image_processor
__SCREAMING_SNAKE_CASE : Any = prepare_img()
__SCREAMING_SNAKE_CASE : List[str] = preprocessor(images=_lowerCamelCase , return_tensors='''pt''' ).to(_lowerCamelCase )
# forward pass
with torch.no_grad():
__SCREAMING_SNAKE_CASE : Dict = model(**_lowerCamelCase )
# verify the logits
__SCREAMING_SNAKE_CASE : List[str] = torch.Size((1, 1_0_0_0) )
self.assertEqual(outputs.logits.shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = torch.tensor([0.9_9_9_6, 0.1_9_6_6, -0.4_3_8_6] ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(outputs.logits[0, :3] , _lowerCamelCase , atol=1e-4 ) )
| 674 |
"""simple docstring"""
from __future__ import annotations
import unittest
from transformers import is_tf_available, is_torch_available
from transformers.testing_utils import DUMMY_UNKNOWN_IDENTIFIER, SMALL_MODEL_IDENTIFIER, is_pt_tf_cross_test, slow
if is_tf_available():
from transformers import (
AutoConfig,
BertConfig,
GPTaConfig,
TaConfig,
TFAutoModel,
TFAutoModelForCausalLM,
TFAutoModelForMaskedLM,
TFAutoModelForPreTraining,
TFAutoModelForQuestionAnswering,
TFAutoModelForSeqaSeqLM,
TFAutoModelForSequenceClassification,
TFAutoModelWithLMHead,
TFBertForMaskedLM,
TFBertForPreTraining,
TFBertForQuestionAnswering,
TFBertForSequenceClassification,
TFBertModel,
TFGPTaLMHeadModel,
TFRobertaForMaskedLM,
TFTaForConditionalGeneration,
)
from transformers.models.bert.modeling_tf_bert import TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST
from transformers.models.gpta.modeling_tf_gpta import TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST
from transformers.models.ta.modeling_tf_ta import TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST
if is_torch_available():
from transformers import (
AutoModel,
AutoModelForCausalLM,
AutoModelForMaskedLM,
AutoModelForPreTraining,
AutoModelForQuestionAnswering,
AutoModelForSeqaSeqLM,
AutoModelForSequenceClassification,
AutoModelWithLMHead,
BertForMaskedLM,
BertForPreTraining,
BertForQuestionAnswering,
BertForSequenceClassification,
BertModel,
GPTaLMHeadModel,
RobertaForMaskedLM,
TaForConditionalGeneration,
)
@is_pt_tf_cross_test
class snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : List[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = TFAutoModel.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModel.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : List[str] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelForPreTraining.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = AutoModelForPreTraining.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Any ):
for model_name in TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Optional[int] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = TFAutoModelForCausalLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = TFAutoModelForCausalLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelForCausalLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = AutoModelForCausalLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Any = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Union[str, Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = TFAutoModelForMaskedLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[str] = TFAutoModelForMaskedLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelForMaskedLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = AutoModelForMaskedLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :int ):
for model_name in TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : str = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelForSeqaSeqLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : str = TFAutoModelForSeqaSeqLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = AutoModelForSeqaSeqLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = AutoModelForSeqaSeqLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = TFAutoModelForSequenceClassification.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModelForSequenceClassification.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = TFAutoModelForQuestionAnswering.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = AutoModelForQuestionAnswering.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
__SCREAMING_SNAKE_CASE : Union[str, Any] = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
| 674 | 1 |
"""simple docstring"""
import unittest
from huggingface_hub import hf_hub_download
from transformers import MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING, VideoMAEFeatureExtractor
from transformers.pipelines import VideoClassificationPipeline, pipeline
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_decord,
require_tf,
require_torch,
require_torch_or_tf,
require_vision,
)
from .test_pipelines_common import ANY
@is_pipeline_test
@require_torch_or_tf
@require_vision
@require_decord
class snake_case ( unittest.TestCase ):
lowerCamelCase__ = MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :Any , _lowerCamelCase :int , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[str] = hf_hub_download(
repo_id='''nateraw/video-demo''' , filename='''archery.mp4''' , repo_type='''dataset''' )
__SCREAMING_SNAKE_CASE : List[Any] = VideoClassificationPipeline(model=_lowerCamelCase , image_processor=_lowerCamelCase , top_k=2 )
__SCREAMING_SNAKE_CASE : Union[str, Any] = [
example_video_filepath,
'''https://huggingface.co/datasets/nateraw/video-demo/resolve/main/archery.mp4''',
]
return video_classifier, examples
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :Tuple , _lowerCamelCase :Dict ):
for example in examples:
__SCREAMING_SNAKE_CASE : int = video_classifier(_lowerCamelCase )
self.assertEqual(
_lowerCamelCase , [
{'''score''': ANY(_lowerCamelCase ), '''label''': ANY(_lowerCamelCase )},
{'''score''': ANY(_lowerCamelCase ), '''label''': ANY(_lowerCamelCase )},
] , )
@require_torch
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE : Optional[int] = '''hf-internal-testing/tiny-random-VideoMAEForVideoClassification'''
__SCREAMING_SNAKE_CASE : Tuple = VideoMAEFeatureExtractor(
size={'''shortest_edge''': 1_0} , crop_size={'''height''': 1_0, '''width''': 1_0} )
__SCREAMING_SNAKE_CASE : Optional[Any] = pipeline(
'''video-classification''' , model=_lowerCamelCase , feature_extractor=_lowerCamelCase , frame_sampling_rate=4 )
__SCREAMING_SNAKE_CASE : int = hf_hub_download(repo_id='''nateraw/video-demo''' , filename='''archery.mp4''' , repo_type='''dataset''' )
__SCREAMING_SNAKE_CASE : List[Any] = video_classifier(_lowerCamelCase , top_k=2 )
self.assertEqual(
nested_simplify(_lowerCamelCase , decimals=4 ) , [{'''score''': 0.5_1_9_9, '''label''': '''LABEL_0'''}, {'''score''': 0.4_8_0_1, '''label''': '''LABEL_1'''}] , )
__SCREAMING_SNAKE_CASE : List[str] = video_classifier(
[
video_file_path,
video_file_path,
] , top_k=2 , )
self.assertEqual(
nested_simplify(_lowerCamelCase , decimals=4 ) , [
[{'''score''': 0.5_1_9_9, '''label''': '''LABEL_0'''}, {'''score''': 0.4_8_0_1, '''label''': '''LABEL_1'''}],
[{'''score''': 0.5_1_9_9, '''label''': '''LABEL_0'''}, {'''score''': 0.4_8_0_1, '''label''': '''LABEL_1'''}],
] , )
@require_tf
def SCREAMING_SNAKE_CASE_ ( self :str ):
pass
| 674 |
"""simple docstring"""
import os
import re
import warnings
from shutil import copyfile
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
if TYPE_CHECKING:
from ...tokenization_utils_base import TextInput
from ...utils import logging
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCamelCase = {
'''vocab_file''': {
'''t5-small''': '''https://huggingface.co/t5-small/resolve/main/spiece.model''',
'''t5-base''': '''https://huggingface.co/t5-base/resolve/main/spiece.model''',
'''t5-large''': '''https://huggingface.co/t5-large/resolve/main/spiece.model''',
'''t5-3b''': '''https://huggingface.co/t5-3b/resolve/main/spiece.model''',
'''t5-11b''': '''https://huggingface.co/t5-11b/resolve/main/spiece.model''',
}
}
# TODO(PVP) - this should be removed in Transformers v5
_lowerCamelCase = {
'''t5-small''': 5_12,
'''t5-base''': 5_12,
'''t5-large''': 5_12,
'''t5-3b''': 5_12,
'''t5-11b''': 5_12,
}
_lowerCamelCase = '''▁'''
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = VOCAB_FILES_NAMES
lowerCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
lowerCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
lowerCamelCase__ = ['''input_ids''', '''attention_mask''']
def __init__( self :int , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Union[str, Any]="</s>" , _lowerCamelCase :List[Any]="<unk>" , _lowerCamelCase :Union[str, Any]="<pad>" , _lowerCamelCase :int=1_0_0 , _lowerCamelCase :Union[str, Any]=None , _lowerCamelCase :Optional[Dict[str, Any]] = None , _lowerCamelCase :int=True , **_lowerCamelCase :List[Any] , ):
# Add extra_ids to the special token list
if extra_ids > 0 and additional_special_tokens is None:
__SCREAMING_SNAKE_CASE : Union[str, Any] = [f'''<extra_id_{i}>''' for i in range(_lowerCamelCase )]
elif extra_ids > 0 and additional_special_tokens is not None:
# Check that we have the right number of extra_id special tokens
__SCREAMING_SNAKE_CASE : Optional[int] = len(set(filter(lambda _lowerCamelCase : bool('''extra_id''' in str(_lowerCamelCase ) ) , _lowerCamelCase ) ) )
if extra_tokens != extra_ids:
raise ValueError(
f'''Both extra_ids ({extra_ids}) and additional_special_tokens ({additional_special_tokens}) are'''
''' provided to T5Tokenizer. In this case the additional_special_tokens must include the extra_ids'''
''' tokens''' )
if legacy:
logger.warning_once(
f'''You are using the legacy behaviour of the {self.__class__}. This means that tokens that come after special tokens will not be properly handled. We recommend you to'''
''' read the related pull request available at https://github.com/huggingface/transformers/pull/24565''' )
__SCREAMING_SNAKE_CASE : Optional[Any] = legacy
__SCREAMING_SNAKE_CASE : Optional[int] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
eos_token=_lowerCamelCase , unk_token=_lowerCamelCase , pad_token=_lowerCamelCase , extra_ids=_lowerCamelCase , additional_special_tokens=_lowerCamelCase , sp_model_kwargs=self.sp_model_kwargs , legacy=_lowerCamelCase , **_lowerCamelCase , )
__SCREAMING_SNAKE_CASE : Tuple = vocab_file
__SCREAMING_SNAKE_CASE : List[str] = extra_ids
__SCREAMING_SNAKE_CASE : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(_lowerCamelCase )
@staticmethod
def SCREAMING_SNAKE_CASE_ ( _lowerCamelCase :str , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :int ):
if pretrained_model_name_or_path in TaTokenizer.max_model_input_sizes:
__SCREAMING_SNAKE_CASE : Any = TaTokenizer.max_model_input_sizes[pretrained_model_name_or_path]
if init_max_model_length is not None and init_max_model_length != max_model_length:
return init_max_model_length
elif init_max_model_length is None:
warnings.warn(
'''This tokenizer was incorrectly instantiated with a model max length of'''
f''' {deprecated_max_model_length} which will be corrected in Transformers v5.\nFor now, this'''
''' behavior is kept to avoid breaking backwards compatibility when padding/encoding with'''
''' `truncation is True`.\n- Be aware that you SHOULD NOT rely on'''
f''' {pretrained_model_name_or_path} automatically truncating your input to'''
f''' {deprecated_max_model_length} when padding/encoding.\n- If you want to encode/pad to sequences'''
f''' longer than {deprecated_max_model_length} you can either instantiate this tokenizer with'''
''' `model_max_length` or pass `max_length` when encoding/padding.\n- To avoid this warning, please'''
''' instantiate this tokenizer with `model_max_length` set to your preferred value.''' , _lowerCamelCase , )
return max_model_length
@property
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
return self.sp_model.get_piece_size() + self._extra_ids
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : str = {self.convert_ids_to_tokens(_lowerCamelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :List[int] , _lowerCamelCase :Optional[List[int]] = None , _lowerCamelCase :bool = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_lowerCamelCase , token_ids_a=_lowerCamelCase , already_has_special_tokens=_lowerCamelCase )
# normal case: some special tokens
if token_ids_a is None:
return ([0] * len(_lowerCamelCase )) + [1]
return ([0] * len(_lowerCamelCase )) + [1] + ([0] * len(_lowerCamelCase )) + [1]
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
return list(
set(filter(lambda _lowerCamelCase : bool(re.search(r'''<extra_id_\d+>''' , _lowerCamelCase ) ) is not None , self.additional_special_tokens ) ) )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
return [self._convert_token_to_id(_lowerCamelCase ) for token in self.get_sentinel_tokens()]
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :List[int] ):
if len(_lowerCamelCase ) > 0 and token_ids[-1] == self.eos_token_id:
warnings.warn(
f'''This sequence already has {self.eos_token}. In future versions this behavior may lead to duplicated'''
''' eos tokens being added.''' )
return token_ids
else:
return token_ids + [self.eos_token_id]
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :List[int] , _lowerCamelCase :Optional[List[int]] = None ):
__SCREAMING_SNAKE_CASE : List[str] = [self.eos_token_id]
if token_ids_a is None:
return len(token_ids_a + eos ) * [0]
return len(token_ids_a + eos + token_ids_a + eos ) * [0]
def SCREAMING_SNAKE_CASE_ ( self :List[Any] , _lowerCamelCase :List[int] , _lowerCamelCase :Optional[List[int]] = None ):
__SCREAMING_SNAKE_CASE : Optional[Any] = self._add_eos_if_not_present(_lowerCamelCase )
if token_ids_a is None:
return token_ids_a
else:
__SCREAMING_SNAKE_CASE : Union[str, Any] = self._add_eos_if_not_present(_lowerCamelCase )
return token_ids_a + token_ids_a
def __getstate__( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Any = self.__dict__.copy()
__SCREAMING_SNAKE_CASE : List[str] = None
return state
def __setstate__( self :Optional[Any] , _lowerCamelCase :List[str] ):
__SCREAMING_SNAKE_CASE : Tuple = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs''' ):
__SCREAMING_SNAKE_CASE : Optional[int] = {}
__SCREAMING_SNAKE_CASE : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :"TextInput" , **_lowerCamelCase :str ):
# Replace the SPIECE_UNDERLINE with a space to make sure SPIECE_UNDERLINE is only used at
# the beginning of the text
if not self.legacy:
__SCREAMING_SNAKE_CASE : Dict = SPIECE_UNDERLINE + text.replace(_lowerCamelCase , ''' ''' )
return super().tokenize(_lowerCamelCase , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :List[Any] , **_lowerCamelCase :Dict ):
if not self.legacy:
__SCREAMING_SNAKE_CASE : str = text.startswith(_lowerCamelCase )
if is_first:
__SCREAMING_SNAKE_CASE : str = text[1:]
__SCREAMING_SNAKE_CASE : Tuple = self.sp_model.encode(_lowerCamelCase , out_type=_lowerCamelCase )
if not self.legacy and not is_first and not text.startswith(''' ''' ) and tokens[0].startswith(_lowerCamelCase ):
__SCREAMING_SNAKE_CASE : Optional[int] = ([tokens[0][1:]] if len(tokens[0] ) > 1 else []) + tokens[1:]
return tokens
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Optional[Any] ):
if token.startswith('''<extra_id_''' ):
__SCREAMING_SNAKE_CASE : Tuple = re.match(r'''<extra_id_(\d+)>''' , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = int(match.group(1 ) )
return self.vocab_size - num - 1
return self.sp_model.piece_to_id(_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :Optional[int] ):
if index < self.sp_model.get_piece_size():
__SCREAMING_SNAKE_CASE : List[Any] = self.sp_model.IdToPiece(_lowerCamelCase )
else:
__SCREAMING_SNAKE_CASE : Dict = f'''<extra_id_{self.vocab_size - 1 - index}>'''
return token
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :Any ):
__SCREAMING_SNAKE_CASE : str = []
__SCREAMING_SNAKE_CASE : Dict = ''''''
__SCREAMING_SNAKE_CASE : Dict = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_lowerCamelCase ) + token
__SCREAMING_SNAKE_CASE : List[str] = True
__SCREAMING_SNAKE_CASE : str = []
else:
current_sub_tokens.append(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = False
out_string += self.sp_model.decode(_lowerCamelCase )
return out_string.strip()
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :str , _lowerCamelCase :Optional[str] = None ):
if not os.path.isdir(_lowerCamelCase ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__SCREAMING_SNAKE_CASE : List[str] = os.path.join(
_lowerCamelCase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_lowerCamelCase ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _lowerCamelCase )
elif not os.path.isfile(self.vocab_file ):
with open(_lowerCamelCase , '''wb''' ) as fi:
__SCREAMING_SNAKE_CASE : Any = self.sp_model.serialized_model_proto()
fi.write(_lowerCamelCase )
return (out_vocab_file,)
| 674 | 1 |
"""simple docstring"""
import ast
import os
import re
import shutil
import tempfile
import unittest
from unittest import mock
import torch
from accelerate.test_utils.examples import compare_against_test
from accelerate.test_utils.testing import TempDirTestCase, require_trackers, run_command, slow
from accelerate.utils import write_basic_config
# DataLoaders built from `test_samples/MRPC` for quick testing
# Should mock `{script_name}.get_dataloaders` via:
# @mock.patch("{script_name}.get_dataloaders", mocked_dataloaders)
_lowerCamelCase = [
'''cross_validation.py''',
'''gradient_accumulation.py''',
'''local_sgd.py''',
'''multi_process_metrics.py''',
'''memory.py''',
'''automatic_gradient_accumulation.py''',
'''fsdp_with_peak_mem_tracking.py''',
'''deepspeed_with_config_support.py''',
'''megatron_lm_gpt_pretraining.py''',
]
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :str , _lowerCamelCase :bool , _lowerCamelCase :str = None , _lowerCamelCase :list = None ):
__SCREAMING_SNAKE_CASE : List[str] = None
__SCREAMING_SNAKE_CASE : Optional[Any] = os.path.abspath(os.path.join('''examples''' , '''by_feature''' ) )
__SCREAMING_SNAKE_CASE : Union[str, Any] = os.path.abspath('''examples''' )
for item in os.listdir(_lowerCamelCase ):
if item not in EXCLUDE_EXAMPLES:
__SCREAMING_SNAKE_CASE : List[Any] = os.path.join(_lowerCamelCase , _lowerCamelCase )
if os.path.isfile(_lowerCamelCase ) and ".py" in item_path:
with self.subTest(
tested_script=_lowerCamelCase , feature_script=_lowerCamelCase , tested_section='''main()''' if parser_only else '''training_function()''' , ):
__SCREAMING_SNAKE_CASE : Tuple = compare_against_test(
os.path.join(_lowerCamelCase , _lowerCamelCase ) , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = '''\n'''.join(_lowerCamelCase )
if special_strings is not None:
for string in special_strings:
__SCREAMING_SNAKE_CASE : List[Any] = diff.replace(_lowerCamelCase , '''''' )
self.assertEqual(_lowerCamelCase , '''''' )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
self.one_complete_example('''complete_nlp_example.py''' , _lowerCamelCase )
self.one_complete_example('''complete_nlp_example.py''' , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = os.path.abspath(os.path.join('''examples''' , '''cv_example.py''' ) )
__SCREAMING_SNAKE_CASE : Optional[int] = [
''' ''' * 1_6 + '''{\n\n''',
''' ''' * 2_0 + '''"accuracy": eval_metric["accuracy"],\n\n''',
''' ''' * 2_0 + '''"f1": eval_metric["f1"],\n\n''',
''' ''' * 2_0 + '''"train_loss": total_loss.item() / len(train_dataloader),\n\n''',
''' ''' * 2_0 + '''"epoch": epoch,\n\n''',
''' ''' * 1_6 + '''},\n\n''',
''' ''' * 1_6 + '''step=epoch,\n''',
''' ''' * 1_2,
''' ''' * 8 + '''for step, batch in enumerate(active_dataloader):\n''',
]
self.one_complete_example('''complete_cv_example.py''' , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
self.one_complete_example('''complete_cv_example.py''' , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
@mock.patch.dict(os.environ , {'''TESTING_MOCKED_DATALOADERS''': '''1'''} )
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = False
@classmethod
def SCREAMING_SNAKE_CASE_ ( cls :Dict ):
super().setUpClass()
__SCREAMING_SNAKE_CASE : Dict = tempfile.mkdtemp()
__SCREAMING_SNAKE_CASE : str = os.path.join(cls._tmpdir , '''default_config.yml''' )
write_basic_config(save_location=cls.configPath )
__SCREAMING_SNAKE_CASE : List[Any] = ['''accelerate''', '''launch''', '''--config_file''', cls.configPath]
@classmethod
def SCREAMING_SNAKE_CASE_ ( cls :Dict ):
super().tearDownClass()
shutil.rmtree(cls._tmpdir )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : str = f'''
examples/by_feature/checkpointing.py
--checkpointing_steps epoch
--output_dir {self.tmpdir}
'''.split()
run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(self.tmpdir , '''epoch_0''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Optional[Any] = f'''
examples/by_feature/checkpointing.py
--checkpointing_steps 1
--output_dir {self.tmpdir}
'''.split()
__SCREAMING_SNAKE_CASE : Optional[int] = run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(self.tmpdir , '''step_2''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Optional[int] = f'''
examples/by_feature/checkpointing.py
--resume_from_checkpoint {os.path.join(self.tmpdir , 'epoch_0' )}
'''.split()
__SCREAMING_SNAKE_CASE : Any = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
self.assertNotIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Optional[int] = f'''
examples/by_feature/checkpointing.py
--resume_from_checkpoint {os.path.join(self.tmpdir , 'step_2' )}
'''.split()
__SCREAMING_SNAKE_CASE : List[str] = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
if torch.cuda.is_available():
__SCREAMING_SNAKE_CASE : List[Any] = torch.cuda.device_count()
else:
__SCREAMING_SNAKE_CASE : Optional[int] = 1
if num_processes > 1:
self.assertNotIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
else:
self.assertIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = '''
examples/by_feature/cross_validation.py
--num_folds 2
'''.split()
with mock.patch.dict(os.environ , {'''TESTING_MOCKED_DATALOADERS''': '''0'''} ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = re.findall('''({.+})''' , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = [r for r in results if '''accuracy''' in r][-1]
__SCREAMING_SNAKE_CASE : Tuple = ast.literal_eval(_lowerCamelCase )
self.assertGreaterEqual(results['''accuracy'''] , 0.7_5 )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = ['''examples/by_feature/multi_process_metrics.py''']
run_command(self._launch_args + testargs )
@require_trackers
@mock.patch.dict(os.environ , {'''WANDB_MODE''': '''offline'''} )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
with tempfile.TemporaryDirectory() as tmpdir:
__SCREAMING_SNAKE_CASE : int = f'''
examples/by_feature/tracking.py
--with_tracking
--project_dir {tmpdir}
'''.split()
run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(_lowerCamelCase , '''tracking''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = ['''examples/by_feature/gradient_accumulation.py''']
run_command(self._launch_args + testargs )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[Any] = ['''examples/by_feature/local_sgd.py''']
run_command(self._launch_args + testargs )
| 674 |
"""simple docstring"""
import os
import pytest
import yaml
from datasets.features.features import Features, Value
from datasets.info import DatasetInfo, DatasetInfosDict
@pytest.mark.parametrize(
'''files''' , [
['''full:README.md''', '''dataset_infos.json'''],
['''empty:README.md''', '''dataset_infos.json'''],
['''dataset_infos.json'''],
['''full:README.md'''],
] , )
def lowerCAmelCase_ ( lowercase_ : List[str] , lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = tmp_path_factory.mktemp('''dset_infos_dir''' )
if "full:README.md" in files:
with open(dataset_infos_dir / '''README.md''' , '''w''' ) as f:
f.write('''---\ndataset_info:\n dataset_size: 42\n---''' )
if "empty:README.md" in files:
with open(dataset_infos_dir / '''README.md''' , '''w''' ) as f:
f.write('''''' )
# we want to support dataset_infos.json for backward compatibility
if "dataset_infos.json" in files:
with open(dataset_infos_dir / '''dataset_infos.json''' , '''w''' ) as f:
f.write('''{"default": {"dataset_size": 42}}''' )
__SCREAMING_SNAKE_CASE : Optional[Any] = DatasetInfosDict.from_directory(lowercase_ )
assert dataset_infos
assert dataset_infos["default"].dataset_size == 42
@pytest.mark.parametrize(
'''dataset_info''' , [
DatasetInfo(),
DatasetInfo(
description='''foo''' , features=Features({'''a''': Value('''int32''' )} ) , builder_name='''builder''' , config_name='''config''' , version='''1.0.0''' , splits=[{'''name''': '''train'''}] , download_size=42 , ),
] , )
def lowerCAmelCase_ ( lowercase_ : Any , lowercase_ : DatasetInfo ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[str] = str(lowercase_ )
dataset_info.write_to_directory(lowercase_ )
__SCREAMING_SNAKE_CASE : Dict = DatasetInfo.from_directory(lowercase_ )
assert dataset_info == reloaded
assert os.path.exists(os.path.join(lowercase_ , '''dataset_info.json''' ) )
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = DatasetInfo(
description='''foo''' , citation='''bar''' , homepage='''https://foo.bar''' , license='''CC0''' , features=Features({'''a''': Value('''int32''' )} ) , post_processed={} , supervised_keys=() , task_templates=[] , builder_name='''builder''' , config_name='''config''' , version='''1.0.0''' , splits=[{'''name''': '''train''', '''num_examples''': 42}] , download_checksums={} , download_size=1337 , post_processing_size=442 , dataset_size=1234 , size_in_bytes=1337 + 442 + 1234 , )
__SCREAMING_SNAKE_CASE : Optional[int] = dataset_info._to_yaml_dict()
assert sorted(lowercase_ ) == sorted(DatasetInfo._INCLUDED_INFO_IN_YAML )
for key in DatasetInfo._INCLUDED_INFO_IN_YAML:
assert key in dataset_info_yaml_dict
assert isinstance(dataset_info_yaml_dict[key] , (list, dict, int, str) )
__SCREAMING_SNAKE_CASE : int = yaml.safe_dump(lowercase_ )
__SCREAMING_SNAKE_CASE : Union[str, Any] = yaml.safe_load(lowercase_ )
assert dataset_info_yaml_dict == reloaded
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = DatasetInfo()
__SCREAMING_SNAKE_CASE : Union[str, Any] = dataset_info._to_yaml_dict()
assert dataset_info_yaml_dict == {}
@pytest.mark.parametrize(
'''dataset_infos_dict''' , [
DatasetInfosDict(),
DatasetInfosDict({'''default''': DatasetInfo()} ),
DatasetInfosDict({'''my_config_name''': DatasetInfo()} ),
DatasetInfosDict(
{
'''default''': DatasetInfo(
description='''foo''' , features=Features({'''a''': Value('''int32''' )} ) , builder_name='''builder''' , config_name='''config''' , version='''1.0.0''' , splits=[{'''name''': '''train'''}] , download_size=42 , )
} ),
DatasetInfosDict(
{
'''v1''': DatasetInfo(dataset_size=42 ),
'''v2''': DatasetInfo(dataset_size=1337 ),
} ),
] , )
def lowerCAmelCase_ ( lowercase_ : List[Any] , lowercase_ : DatasetInfosDict ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = str(lowercase_ )
dataset_infos_dict.write_to_directory(lowercase_ )
__SCREAMING_SNAKE_CASE : Optional[Any] = DatasetInfosDict.from_directory(lowercase_ )
# the config_name of the dataset_infos_dict take over the attribute
for config_name, dataset_info in dataset_infos_dict.items():
__SCREAMING_SNAKE_CASE : Optional[int] = config_name
# the yaml representation doesn't include fields like description or citation
# so we just test that we can recover what we can from the yaml
__SCREAMING_SNAKE_CASE : Tuple = DatasetInfo._from_yaml_dict(dataset_info._to_yaml_dict() )
assert dataset_infos_dict == reloaded
if dataset_infos_dict:
assert os.path.exists(os.path.join(lowercase_ , '''README.md''' ) )
| 674 | 1 |
"""simple docstring"""
from __future__ import annotations
def lowerCAmelCase_ ( lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = str(lowercase_ )
return len(lowercase_ ) == 9 and set(lowercase_ ) == set('''123456789''' )
def lowerCAmelCase_ ( ):
'''simple docstring'''
for base_num in range(9999 , 4999 , -1 ):
__SCREAMING_SNAKE_CASE : List[str] = 10_0002 * base_num
if is_9_pandigital(lowercase_ ):
return candidate
for base_num in range(333 , 99 , -1 ):
__SCREAMING_SNAKE_CASE : List[Any] = 100_2003 * base_num
if is_9_pandigital(lowercase_ ):
return candidate
return None
if __name__ == "__main__":
print(f'{solution() = }')
| 674 |
"""simple docstring"""
from typing import Optional, Union
import torch
from torch import nn
from ...configuration_utils import ConfigMixin, register_to_config
from ...models.modeling_utils import ModelMixin
class snake_case ( __UpperCAmelCase , __UpperCAmelCase ):
@register_to_config
def __init__( self :List[str] , _lowerCamelCase :int = 7_6_8 , ):
super().__init__()
__SCREAMING_SNAKE_CASE : Union[str, Any] = nn.Parameter(torch.zeros(1 , _lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : List[Any] = nn.Parameter(torch.ones(1 , _lowerCamelCase ) )
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :Optional[Union[str, torch.device]] = None , _lowerCamelCase :Optional[torch.dtype] = None , ):
__SCREAMING_SNAKE_CASE : Optional[int] = nn.Parameter(self.mean.to(_lowerCamelCase ).to(_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Tuple = nn.Parameter(self.std.to(_lowerCamelCase ).to(_lowerCamelCase ) )
return self
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :Tuple ):
__SCREAMING_SNAKE_CASE : Tuple = (embeds - self.mean) * 1.0 / self.std
return embeds
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[Any] = (embeds * self.std) + self.mean
return embeds
| 674 | 1 |
"""simple docstring"""
def lowerCAmelCase_ ( lowercase_ : list , lowercase_ : list ):
'''simple docstring'''
_validate_point(lowercase_ )
_validate_point(lowercase_ )
if len(lowercase_ ) != len(lowercase_ ):
raise ValueError('''Both points must be in the same n-dimensional space''' )
return float(sum(abs(a - b ) for a, b in zip(lowercase_ , lowercase_ ) ) )
def lowerCAmelCase_ ( lowercase_ : list[float] ):
'''simple docstring'''
if point:
if isinstance(lowercase_ , lowercase_ ):
for item in point:
if not isinstance(lowercase_ , (int, float) ):
__SCREAMING_SNAKE_CASE : Dict = (
'''Expected a list of numbers as input, found '''
F'''{type(lowercase_ ).__name__}'''
)
raise TypeError(lowercase_ )
else:
__SCREAMING_SNAKE_CASE : int = F'''Expected a list of numbers as input, found {type(lowercase_ ).__name__}'''
raise TypeError(lowercase_ )
else:
raise ValueError('''Missing an input''' )
def lowerCAmelCase_ ( lowercase_ : list , lowercase_ : list ):
'''simple docstring'''
_validate_point(lowercase_ )
_validate_point(lowercase_ )
if len(lowercase_ ) != len(lowercase_ ):
raise ValueError('''Both points must be in the same n-dimensional space''' )
return float(sum(abs(x - y ) for x, y in zip(lowercase_ , lowercase_ ) ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 674 |
"""simple docstring"""
import argparse
import torch
from transformers import BertConfig, BertForPreTraining, load_tf_weights_in_bert
from transformers.utils import logging
logging.set_verbosity_info()
def lowerCAmelCase_ ( lowercase_ : Dict , lowercase_ : int , lowercase_ : Any ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = BertConfig.from_json_file(lowercase_ )
print(F'''Building PyTorch model from configuration: {config}''' )
__SCREAMING_SNAKE_CASE : Optional[int] = BertForPreTraining(lowercase_ )
# Load weights from tf checkpoint
load_tf_weights_in_bert(lowercase_ , lowercase_ , lowercase_ )
# Save pytorch-model
print(F'''Save PyTorch model to {pytorch_dump_path}''' )
torch.save(model.state_dict() , lowercase_ )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--tf_checkpoint_path''', default=None, type=str, required=True, help='''Path to the TensorFlow checkpoint path.'''
)
parser.add_argument(
'''--bert_config_file''',
default=None,
type=str,
required=True,
help=(
'''The config json file corresponding to the pre-trained BERT model. \n'''
'''This specifies the model architecture.'''
),
)
parser.add_argument(
'''--pytorch_dump_path''', default=None, type=str, required=True, help='''Path to the output PyTorch model.'''
)
_lowerCamelCase = parser.parse_args()
convert_tf_checkpoint_to_pytorch(args.tf_checkpoint_path, args.bert_config_file, args.pytorch_dump_path)
| 674 | 1 |
"""simple docstring"""
# Copyright 2023 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
_lowerCamelCase = {
'''configuration_xmod''': [
'''XMOD_PRETRAINED_CONFIG_ARCHIVE_MAP''',
'''XmodConfig''',
'''XmodOnnxConfig''',
],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCamelCase = [
'''XMOD_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''XmodForCausalLM''',
'''XmodForMaskedLM''',
'''XmodForMultipleChoice''',
'''XmodForQuestionAnswering''',
'''XmodForSequenceClassification''',
'''XmodForTokenClassification''',
'''XmodModel''',
'''XmodPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_xmod import XMOD_PRETRAINED_CONFIG_ARCHIVE_MAP, XmodConfig, XmodOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_xmod import (
XMOD_PRETRAINED_MODEL_ARCHIVE_LIST,
XmodForCausalLM,
XmodForMaskedLM,
XmodForMultipleChoice,
XmodForQuestionAnswering,
XmodForSequenceClassification,
XmodForTokenClassification,
XmodModel,
XmodPreTrainedModel,
)
else:
import sys
_lowerCamelCase = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 674 |
"""simple docstring"""
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import numpy as np
import torch
from ..models.clipseg import CLIPSegForImageSegmentation
from ..utils import is_vision_available, requires_backends
from .base import PipelineTool
if is_vision_available():
from PIL import Image
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = (
'''This is a tool that creates a segmentation mask of an image according to a label. It cannot create an image.'''
'''It takes two arguments named `image` which should be the original image, and `label` which should be a text '''
'''describing the elements what should be identified in the segmentation mask. The tool returns the mask.'''
)
lowerCamelCase__ = '''CIDAS/clipseg-rd64-refined'''
lowerCamelCase__ = '''image_segmenter'''
lowerCamelCase__ = CLIPSegForImageSegmentation
lowerCamelCase__ = ['''image''', '''text''']
lowerCamelCase__ = ['''image''']
def __init__( self :Dict , *_lowerCamelCase :Union[str, Any] , **_lowerCamelCase :Tuple ):
requires_backends(self , ['''vision'''] )
super().__init__(*_lowerCamelCase , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :"Image" , _lowerCamelCase :str ):
return self.pre_processor(text=[label] , images=[image] , padding=_lowerCamelCase , return_tensors='''pt''' )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] , _lowerCamelCase :Optional[int] ):
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[Any] = self.model(**_lowerCamelCase ).logits
return logits
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Tuple ):
__SCREAMING_SNAKE_CASE : Optional[int] = outputs.cpu().detach().numpy()
__SCREAMING_SNAKE_CASE : str = 0
__SCREAMING_SNAKE_CASE : str = 1
return Image.fromarray((array * 2_5_5).astype(np.uinta ) )
| 674 | 1 |
"""simple docstring"""
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = {
'''bert-base-uncased''': '''https://huggingface.co/bert-base-uncased/resolve/main/config.json''',
'''bert-large-uncased''': '''https://huggingface.co/bert-large-uncased/resolve/main/config.json''',
'''bert-base-cased''': '''https://huggingface.co/bert-base-cased/resolve/main/config.json''',
'''bert-large-cased''': '''https://huggingface.co/bert-large-cased/resolve/main/config.json''',
'''bert-base-multilingual-uncased''': '''https://huggingface.co/bert-base-multilingual-uncased/resolve/main/config.json''',
'''bert-base-multilingual-cased''': '''https://huggingface.co/bert-base-multilingual-cased/resolve/main/config.json''',
'''bert-base-chinese''': '''https://huggingface.co/bert-base-chinese/resolve/main/config.json''',
'''bert-base-german-cased''': '''https://huggingface.co/bert-base-german-cased/resolve/main/config.json''',
'''bert-large-uncased-whole-word-masking''': (
'''https://huggingface.co/bert-large-uncased-whole-word-masking/resolve/main/config.json'''
),
'''bert-large-cased-whole-word-masking''': (
'''https://huggingface.co/bert-large-cased-whole-word-masking/resolve/main/config.json'''
),
'''bert-large-uncased-whole-word-masking-finetuned-squad''': (
'''https://huggingface.co/bert-large-uncased-whole-word-masking-finetuned-squad/resolve/main/config.json'''
),
'''bert-large-cased-whole-word-masking-finetuned-squad''': (
'''https://huggingface.co/bert-large-cased-whole-word-masking-finetuned-squad/resolve/main/config.json'''
),
'''bert-base-cased-finetuned-mrpc''': '''https://huggingface.co/bert-base-cased-finetuned-mrpc/resolve/main/config.json''',
'''bert-base-german-dbmdz-cased''': '''https://huggingface.co/bert-base-german-dbmdz-cased/resolve/main/config.json''',
'''bert-base-german-dbmdz-uncased''': '''https://huggingface.co/bert-base-german-dbmdz-uncased/resolve/main/config.json''',
'''cl-tohoku/bert-base-japanese''': '''https://huggingface.co/cl-tohoku/bert-base-japanese/resolve/main/config.json''',
'''cl-tohoku/bert-base-japanese-whole-word-masking''': (
'''https://huggingface.co/cl-tohoku/bert-base-japanese-whole-word-masking/resolve/main/config.json'''
),
'''cl-tohoku/bert-base-japanese-char''': (
'''https://huggingface.co/cl-tohoku/bert-base-japanese-char/resolve/main/config.json'''
),
'''cl-tohoku/bert-base-japanese-char-whole-word-masking''': (
'''https://huggingface.co/cl-tohoku/bert-base-japanese-char-whole-word-masking/resolve/main/config.json'''
),
'''TurkuNLP/bert-base-finnish-cased-v1''': (
'''https://huggingface.co/TurkuNLP/bert-base-finnish-cased-v1/resolve/main/config.json'''
),
'''TurkuNLP/bert-base-finnish-uncased-v1''': (
'''https://huggingface.co/TurkuNLP/bert-base-finnish-uncased-v1/resolve/main/config.json'''
),
'''wietsedv/bert-base-dutch-cased''': '''https://huggingface.co/wietsedv/bert-base-dutch-cased/resolve/main/config.json''',
# See all BERT models at https://huggingface.co/models?filter=bert
}
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = '''bert'''
def __init__( self :Optional[int] , _lowerCamelCase :Optional[Any]=3_0_5_2_2 , _lowerCamelCase :int=7_6_8 , _lowerCamelCase :Union[str, Any]=1_2 , _lowerCamelCase :Any=1_2 , _lowerCamelCase :List[str]=3_0_7_2 , _lowerCamelCase :Optional[int]="gelu" , _lowerCamelCase :List[Any]=0.1 , _lowerCamelCase :Union[str, Any]=0.1 , _lowerCamelCase :List[str]=5_1_2 , _lowerCamelCase :Union[str, Any]=2 , _lowerCamelCase :Tuple=0.0_2 , _lowerCamelCase :Any=1e-12 , _lowerCamelCase :Tuple=0 , _lowerCamelCase :List[str]="absolute" , _lowerCamelCase :Optional[int]=True , _lowerCamelCase :Optional[int]=None , **_lowerCamelCase :Optional[Any] , ):
super().__init__(pad_token_id=_lowerCamelCase , **_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = vocab_size
__SCREAMING_SNAKE_CASE : Union[str, Any] = hidden_size
__SCREAMING_SNAKE_CASE : str = num_hidden_layers
__SCREAMING_SNAKE_CASE : List[str] = num_attention_heads
__SCREAMING_SNAKE_CASE : str = hidden_act
__SCREAMING_SNAKE_CASE : Optional[Any] = intermediate_size
__SCREAMING_SNAKE_CASE : Tuple = hidden_dropout_prob
__SCREAMING_SNAKE_CASE : Tuple = attention_probs_dropout_prob
__SCREAMING_SNAKE_CASE : Union[str, Any] = max_position_embeddings
__SCREAMING_SNAKE_CASE : Optional[int] = type_vocab_size
__SCREAMING_SNAKE_CASE : int = initializer_range
__SCREAMING_SNAKE_CASE : int = layer_norm_eps
__SCREAMING_SNAKE_CASE : Tuple = position_embedding_type
__SCREAMING_SNAKE_CASE : Tuple = use_cache
__SCREAMING_SNAKE_CASE : Tuple = classifier_dropout
class snake_case ( __UpperCAmelCase ):
@property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
if self.task == "multiple-choice":
__SCREAMING_SNAKE_CASE : Dict = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
__SCREAMING_SNAKE_CASE : Tuple = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''input_ids''', dynamic_axis),
('''attention_mask''', dynamic_axis),
('''token_type_ids''', dynamic_axis),
] )
| 674 |
"""simple docstring"""
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
from transformers import BertTokenizerFast
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES, BertTokenizer
from transformers.testing_utils import require_tokenizers, require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import VisionTextDualEncoderProcessor, ViTImageProcessor
@require_tokenizers
@require_vision
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = tempfile.mkdtemp()
# fmt: off
__SCREAMING_SNAKE_CASE : Optional[int] = ['''[UNK]''', '''[CLS]''', '''[SEP]''', '''[PAD]''', '''[MASK]''', '''want''', '''##want''', '''##ed''', '''wa''', '''un''', '''runn''', '''##ing''', ''',''', '''low''', '''lowest''']
# fmt: on
__SCREAMING_SNAKE_CASE : List[Any] = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
__SCREAMING_SNAKE_CASE : Optional[int] = {
'''do_resize''': True,
'''size''': {'''height''': 1_8, '''width''': 1_8},
'''do_normalize''': True,
'''image_mean''': [0.5, 0.5, 0.5],
'''image_std''': [0.5, 0.5, 0.5],
}
__SCREAMING_SNAKE_CASE : List[Any] = os.path.join(self.tmpdirname , _lowerCamelCase )
with open(self.image_processor_file , '''w''' , encoding='''utf-8''' ) as fp:
json.dump(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , **_lowerCamelCase :List[str] ):
return BertTokenizer.from_pretrained(self.tmpdirname , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] , **_lowerCamelCase :Optional[int] ):
return ViTImageProcessor.from_pretrained(self.tmpdirname , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
shutil.rmtree(self.tmpdirname )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = [np.random.randint(2_5_5 , size=(3, 3_0, 4_0_0) , dtype=np.uinta )]
__SCREAMING_SNAKE_CASE : Tuple = [Image.fromarray(np.moveaxis(_lowerCamelCase , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Optional[Any] = self.get_image_processor()
__SCREAMING_SNAKE_CASE : Optional[Any] = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
processor.save_pretrained(self.tmpdirname )
__SCREAMING_SNAKE_CASE : int = VisionTextDualEncoderProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor.image_processor , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : str = VisionTextDualEncoderProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''' )
__SCREAMING_SNAKE_CASE : List[str] = self.get_image_processor(do_normalize=_lowerCamelCase , padding_value=1.0 )
__SCREAMING_SNAKE_CASE : int = VisionTextDualEncoderProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=_lowerCamelCase , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.get_image_processor()
__SCREAMING_SNAKE_CASE : Dict = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Tuple = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = self.prepare_image_inputs()
__SCREAMING_SNAKE_CASE : Optional[int] = image_processor(_lowerCamelCase , return_tensors='''np''' )
__SCREAMING_SNAKE_CASE : Tuple = processor(images=_lowerCamelCase , return_tensors='''np''' )
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1e-2 )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : Any = self.get_image_processor()
__SCREAMING_SNAKE_CASE : Any = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Tuple = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = '''lower newer'''
__SCREAMING_SNAKE_CASE : Optional[int] = processor(text=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = tokenizer(_lowerCamelCase )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : str = self.get_image_processor()
__SCREAMING_SNAKE_CASE : int = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : int = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = '''lower newer'''
__SCREAMING_SNAKE_CASE : int = self.prepare_image_inputs()
__SCREAMING_SNAKE_CASE : int = processor(text=_lowerCamelCase , images=_lowerCamelCase )
self.assertListEqual(list(inputs.keys() ) , ['''input_ids''', '''token_type_ids''', '''attention_mask''', '''pixel_values'''] )
# test if it raises when no input is passed
with self.assertRaises(_lowerCamelCase ):
processor()
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : Dict = self.get_image_processor()
__SCREAMING_SNAKE_CASE : List[Any] = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Union[str, Any] = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
__SCREAMING_SNAKE_CASE : Tuple = processor.batch_decode(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = tokenizer.batch_decode(_lowerCamelCase )
self.assertListEqual(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : str = self.get_image_processor()
__SCREAMING_SNAKE_CASE : List[str] = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Optional[int] = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = '''lower newer'''
__SCREAMING_SNAKE_CASE : Dict = self.prepare_image_inputs()
__SCREAMING_SNAKE_CASE : str = processor(text=_lowerCamelCase , images=_lowerCamelCase )
self.assertListEqual(list(inputs.keys() ) , processor.model_input_names )
| 674 | 1 |
"""simple docstring"""
import collections
import inspect
import unittest
from typing import Dict, List, Tuple
from transformers import MaskFormerSwinConfig
from transformers.testing_utils import require_torch, require_torch_multi_gpu, torch_device
from transformers.utils import is_torch_available
from ...test_backbone_common import BackboneTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import MaskFormerSwinBackbone
from transformers.models.maskformer import MaskFormerSwinModel
class snake_case :
def __init__( self :Dict , _lowerCamelCase :int , _lowerCamelCase :Optional[Any]=1_3 , _lowerCamelCase :Union[str, Any]=3_2 , _lowerCamelCase :Dict=2 , _lowerCamelCase :int=3 , _lowerCamelCase :Optional[int]=1_6 , _lowerCamelCase :str=[1, 2, 1] , _lowerCamelCase :int=[2, 2, 4] , _lowerCamelCase :List[Any]=2 , _lowerCamelCase :List[str]=2.0 , _lowerCamelCase :int=True , _lowerCamelCase :Union[str, Any]=0.0 , _lowerCamelCase :Any=0.0 , _lowerCamelCase :Tuple=0.1 , _lowerCamelCase :Tuple="gelu" , _lowerCamelCase :str=False , _lowerCamelCase :int=True , _lowerCamelCase :Dict=0.0_2 , _lowerCamelCase :Union[str, Any]=1e-5 , _lowerCamelCase :Any=True , _lowerCamelCase :List[str]=None , _lowerCamelCase :List[str]=True , _lowerCamelCase :Dict=1_0 , _lowerCamelCase :Union[str, Any]=8 , _lowerCamelCase :List[Any]=["stage1", "stage2", "stage3"] , _lowerCamelCase :str=[1, 2, 3] , ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = parent
__SCREAMING_SNAKE_CASE : Optional[int] = batch_size
__SCREAMING_SNAKE_CASE : Any = image_size
__SCREAMING_SNAKE_CASE : Optional[int] = patch_size
__SCREAMING_SNAKE_CASE : str = num_channels
__SCREAMING_SNAKE_CASE : List[Any] = embed_dim
__SCREAMING_SNAKE_CASE : Optional[Any] = depths
__SCREAMING_SNAKE_CASE : List[str] = num_heads
__SCREAMING_SNAKE_CASE : Union[str, Any] = window_size
__SCREAMING_SNAKE_CASE : int = mlp_ratio
__SCREAMING_SNAKE_CASE : Any = qkv_bias
__SCREAMING_SNAKE_CASE : Tuple = hidden_dropout_prob
__SCREAMING_SNAKE_CASE : Optional[int] = attention_probs_dropout_prob
__SCREAMING_SNAKE_CASE : Dict = drop_path_rate
__SCREAMING_SNAKE_CASE : Dict = hidden_act
__SCREAMING_SNAKE_CASE : Union[str, Any] = use_absolute_embeddings
__SCREAMING_SNAKE_CASE : Optional[int] = patch_norm
__SCREAMING_SNAKE_CASE : Any = layer_norm_eps
__SCREAMING_SNAKE_CASE : Union[str, Any] = initializer_range
__SCREAMING_SNAKE_CASE : Dict = is_training
__SCREAMING_SNAKE_CASE : Optional[Any] = scope
__SCREAMING_SNAKE_CASE : Union[str, Any] = use_labels
__SCREAMING_SNAKE_CASE : Optional[Any] = type_sequence_label_size
__SCREAMING_SNAKE_CASE : List[Any] = encoder_stride
__SCREAMING_SNAKE_CASE : List[str] = out_features
__SCREAMING_SNAKE_CASE : Optional[int] = out_indices
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : str = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
__SCREAMING_SNAKE_CASE : int = None
if self.use_labels:
__SCREAMING_SNAKE_CASE : Union[str, Any] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
__SCREAMING_SNAKE_CASE : Tuple = self.get_config()
return config, pixel_values, labels
def SCREAMING_SNAKE_CASE_ ( self :int ):
return MaskFormerSwinConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , embed_dim=self.embed_dim , depths=self.depths , num_heads=self.num_heads , window_size=self.window_size , mlp_ratio=self.mlp_ratio , qkv_bias=self.qkv_bias , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , drop_path_rate=self.drop_path_rate , hidden_act=self.hidden_act , use_absolute_embeddings=self.use_absolute_embeddings , path_norm=self.patch_norm , layer_norm_eps=self.layer_norm_eps , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , out_features=self.out_features , out_indices=self.out_indices , )
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :Dict , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :Optional[Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = MaskFormerSwinModel(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : Tuple = model(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
__SCREAMING_SNAKE_CASE : Optional[int] = int(config.embed_dim * 2 ** (len(config.depths ) - 1) )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, expected_seq_len, expected_dim) )
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :List[str] , _lowerCamelCase :Any , _lowerCamelCase :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Any = MaskFormerSwinBackbone(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : Union[str, Any] = model(_lowerCamelCase )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , len(config.out_features ) )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [1_3, 1_6, 1_6, 1_6] )
# verify channels
self.parent.assertEqual(len(model.channels ) , len(config.out_features ) )
self.parent.assertListEqual(model.channels , [1_6, 3_2, 6_4] )
# verify ValueError
with self.parent.assertRaises(_lowerCamelCase ):
__SCREAMING_SNAKE_CASE : Optional[Any] = ['''stem''']
__SCREAMING_SNAKE_CASE : Optional[Any] = MaskFormerSwinBackbone(config=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Dict = self.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : int = config_and_inputs
__SCREAMING_SNAKE_CASE : str = {'''pixel_values''': pixel_values}
return config, inputs_dict
@require_torch
class snake_case ( __UpperCAmelCase , __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
lowerCamelCase__ = {'''feature-extraction''': MaskFormerSwinModel} if is_torch_available() else {}
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Tuple = MaskFormerSwinModelTester(self )
__SCREAMING_SNAKE_CASE : str = ConfigTester(self , config_class=_lowerCamelCase , embed_dim=3_7 )
@require_torch_multi_gpu
@unittest.skip(
reason=(
'''`MaskFormerSwinModel` outputs `hidden_states_spatial_dimensions` which doesn\'t work well with'''
''' `nn.DataParallel`'''
) )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
pass
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
return
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[int] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_backbone(*_lowerCamelCase )
@unittest.skip('''Swin does not use inputs_embeds''' )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
pass
@unittest.skip('''Swin does not support feedforward chunking''' )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
pass
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : int = model_class(_lowerCamelCase )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
__SCREAMING_SNAKE_CASE : List[str] = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(_lowerCamelCase , nn.Linear ) )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Union[str, Any] = model_class(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
__SCREAMING_SNAKE_CASE : Any = [*signature.parameters.keys()]
__SCREAMING_SNAKE_CASE : Any = ['''pixel_values''']
self.assertListEqual(arg_names[:1] , _lowerCamelCase )
@unittest.skip(reason='''MaskFormerSwin is only used as backbone and doesn\'t support output_attentions''' )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
pass
@unittest.skip(reason='''MaskFormerSwin is only used as an internal backbone''' )
def SCREAMING_SNAKE_CASE_ ( self :int ):
pass
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Optional[int] , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Optional[int] , _lowerCamelCase :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
with torch.no_grad():
__SCREAMING_SNAKE_CASE : Union[str, Any] = model(**self._prepare_for_class(_lowerCamelCase , _lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Dict = outputs.hidden_states
__SCREAMING_SNAKE_CASE : int = getattr(
self.model_tester , '''expected_num_hidden_layers''' , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(_lowerCamelCase ) , _lowerCamelCase )
# Swin has a different seq_length
__SCREAMING_SNAKE_CASE : Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
__SCREAMING_SNAKE_CASE : List[Any] = (image_size[1] // patch_size[1]) * (image_size[0] // patch_size[0])
self.assertListEqual(
list(hidden_states[0].shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
__SCREAMING_SNAKE_CASE : Dict = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Union[str, Any] = True
self.check_hidden_states_output(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
__SCREAMING_SNAKE_CASE : Optional[Any] = True
self.check_hidden_states_output(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs_for_common()
__SCREAMING_SNAKE_CASE : int = 3
__SCREAMING_SNAKE_CASE : int = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
__SCREAMING_SNAKE_CASE : List[Any] = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
__SCREAMING_SNAKE_CASE : Any = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
__SCREAMING_SNAKE_CASE : List[Any] = image_size[1] + patch_size[1] - (image_size[1] % patch_size[1])
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Dict = True
self.check_hidden_states_output(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase , (padded_height, padded_width) )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
__SCREAMING_SNAKE_CASE : Any = True
self.check_hidden_states_output(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase , (padded_height, padded_width) )
@unittest.skip(reason='''MaskFormerSwin doesn\'t have pretrained checkpoints''' )
def SCREAMING_SNAKE_CASE_ ( self :int ):
pass
@unittest.skip(reason='''This will be fixed once MaskFormerSwin is replaced by native Swin''' )
def SCREAMING_SNAKE_CASE_ ( self :Any ):
pass
@unittest.skip(reason='''This will be fixed once MaskFormerSwin is replaced by native Swin''' )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
pass
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(_lowerCamelCase :List[Any] ):
__SCREAMING_SNAKE_CASE : Any = 0
return t
def check_equivalence(_lowerCamelCase :Any , _lowerCamelCase :Any , _lowerCamelCase :int , _lowerCamelCase :str={} ):
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[str] = model(**_lowerCamelCase , return_dict=_lowerCamelCase , **_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = model(**_lowerCamelCase , return_dict=_lowerCamelCase , **_lowerCamelCase ).to_tuple()
def recursive_check(_lowerCamelCase :List[str] , _lowerCamelCase :int ):
if isinstance(_lowerCamelCase , (List, Tuple) ):
for tuple_iterable_value, dict_iterable_value in zip(_lowerCamelCase , _lowerCamelCase ):
recursive_check(_lowerCamelCase , _lowerCamelCase )
elif isinstance(_lowerCamelCase , _lowerCamelCase ):
for tuple_iterable_value, dict_iterable_value in zip(
tuple_object.values() , dict_object.values() ):
recursive_check(_lowerCamelCase , _lowerCamelCase )
elif tuple_object is None:
return
else:
self.assertTrue(
torch.allclose(
set_nan_tensor_to_zero(_lowerCamelCase ) , set_nan_tensor_to_zero(_lowerCamelCase ) , atol=1e-5 ) , msg=(
'''Tuple and dict output are not equal. Difference:'''
f''' {torch.max(torch.abs(tuple_object - dict_object ) )}. Tuple has `nan`:'''
f''' {torch.isnan(_lowerCamelCase ).any()} and `inf`: {torch.isinf(_lowerCamelCase )}. Dict has'''
f''' `nan`: {torch.isnan(_lowerCamelCase ).any()} and `inf`: {torch.isinf(_lowerCamelCase )}.'''
) , )
recursive_check(_lowerCamelCase , _lowerCamelCase )
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : List[str] = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : List[Any] = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase )
check_equivalence(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
check_equivalence(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase )
check_equivalence(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase , {'''output_hidden_states''': True} )
__SCREAMING_SNAKE_CASE : Any = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
check_equivalence(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase , {'''output_hidden_states''': True} )
@require_torch
class snake_case ( unittest.TestCase , __UpperCAmelCase ):
lowerCamelCase__ = (MaskFormerSwinBackbone,) if is_torch_available() else ()
lowerCamelCase__ = MaskFormerSwinConfig
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Tuple = MaskFormerSwinModelTester(self )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs_for_common()
__SCREAMING_SNAKE_CASE : List[str] = inputs_dict['''pixel_values'''].shape[0]
for backbone_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : int = backbone_class(_lowerCamelCase )
backbone.to(_lowerCamelCase )
backbone.eval()
__SCREAMING_SNAKE_CASE : Optional[int] = backbone(**_lowerCamelCase )
# Test default outputs and verify feature maps
self.assertIsInstance(outputs.feature_maps , _lowerCamelCase )
self.assertTrue(len(outputs.feature_maps ) == len(backbone.channels ) )
for feature_map, n_channels in zip(outputs.feature_maps , backbone.channels ):
self.assertTrue(feature_map.shape[:2] , (batch_size, n_channels) )
self.assertIsNone(outputs.hidden_states )
self.assertIsNone(outputs.attentions )
# Test output_hidden_states=True
__SCREAMING_SNAKE_CASE : str = backbone(**_lowerCamelCase , output_hidden_states=_lowerCamelCase )
self.assertIsNotNone(outputs.hidden_states )
self.assertTrue(len(outputs.hidden_states ) , len(backbone.stage_names ) )
# We skip the stem layer
for hidden_states, n_channels in zip(outputs.hidden_states[1:] , backbone.channels ):
for hidden_state in hidden_states:
# Hidden states are in the format (batch_size, (height * width), n_channels)
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = hidden_state.shape
self.assertTrue((h_batch_size, h_n_channels) , (batch_size, n_channels) )
# Test output_attentions=True
if self.has_attentions:
__SCREAMING_SNAKE_CASE : Any = backbone(**_lowerCamelCase , output_attentions=_lowerCamelCase )
self.assertIsNotNone(outputs.attentions )
| 674 |
"""simple docstring"""
import inspect
import unittest
import numpy as np
from tests.test_modeling_common import floats_tensor
from transformers import DetrConfig, MaskFormerConfig, SwinConfig, is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device
from transformers.utils import cached_property
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import MaskFormerForInstanceSegmentation, MaskFormerModel
if is_vision_available():
from transformers import MaskFormerImageProcessor
if is_vision_available():
from PIL import Image
class snake_case :
def __init__( self :Optional[int] , _lowerCamelCase :int , _lowerCamelCase :Any=2 , _lowerCamelCase :List[str]=True , _lowerCamelCase :str=False , _lowerCamelCase :Tuple=1_0 , _lowerCamelCase :str=3 , _lowerCamelCase :str=3_2 * 4 , _lowerCamelCase :Dict=3_2 * 6 , _lowerCamelCase :str=4 , _lowerCamelCase :Any=3_2 , ):
__SCREAMING_SNAKE_CASE : List[str] = parent
__SCREAMING_SNAKE_CASE : Tuple = batch_size
__SCREAMING_SNAKE_CASE : Optional[Any] = is_training
__SCREAMING_SNAKE_CASE : Dict = use_auxiliary_loss
__SCREAMING_SNAKE_CASE : List[str] = num_queries
__SCREAMING_SNAKE_CASE : Optional[int] = num_channels
__SCREAMING_SNAKE_CASE : List[Any] = min_size
__SCREAMING_SNAKE_CASE : int = max_size
__SCREAMING_SNAKE_CASE : Any = num_labels
__SCREAMING_SNAKE_CASE : Union[str, Any] = mask_feature_size
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : str = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to(
_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = torch.ones([self.batch_size, self.min_size, self.max_size] , device=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = (
torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] , device=_lowerCamelCase ) > 0.5
).float()
__SCREAMING_SNAKE_CASE : Dict = (torch.rand((self.batch_size, self.num_labels) , device=_lowerCamelCase ) > 0.5).long()
__SCREAMING_SNAKE_CASE : str = self.get_config()
return config, pixel_values, pixel_mask, mask_labels, class_labels
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
return MaskFormerConfig.from_backbone_and_decoder_configs(
backbone_config=SwinConfig(
depths=[1, 1, 1, 1] , ) , decoder_config=DetrConfig(
decoder_ffn_dim=1_2_8 , num_queries=self.num_queries , decoder_attention_heads=2 , d_model=self.mask_feature_size , ) , mask_feature_size=self.mask_feature_size , fpn_feature_size=self.mask_feature_size , num_channels=self.num_channels , num_labels=self.num_labels , )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = self.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE : Any = {'''pixel_values''': pixel_values, '''pixel_mask''': pixel_mask}
return config, inputs_dict
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :int , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = output.encoder_hidden_states
__SCREAMING_SNAKE_CASE : int = output.pixel_decoder_hidden_states
__SCREAMING_SNAKE_CASE : Optional[int] = output.transformer_decoder_hidden_states
self.parent.assertTrue(len(_lowerCamelCase ) , len(config.backbone_config.depths ) )
self.parent.assertTrue(len(_lowerCamelCase ) , len(config.backbone_config.depths ) )
self.parent.assertTrue(len(_lowerCamelCase ) , config.decoder_config.decoder_layers )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Dict , _lowerCamelCase :str , _lowerCamelCase :Any , _lowerCamelCase :Optional[Any]=False ):
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[str] = MaskFormerModel(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : str = model(pixel_values=_lowerCamelCase , pixel_mask=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = model(_lowerCamelCase , output_hidden_states=_lowerCamelCase )
# the correct shape of output.transformer_decoder_hidden_states ensure the correcteness of the
# encoder and pixel decoder
self.parent.assertEqual(
output.transformer_decoder_last_hidden_state.shape , (self.batch_size, self.num_queries, self.mask_feature_size) , )
# let's ensure the other two hidden state exists
self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(output.encoder_last_hidden_state is not None )
if output_hidden_states:
self.check_output_hidden_state(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Optional[int] , _lowerCamelCase :Optional[Any] , _lowerCamelCase :List[str] , _lowerCamelCase :Dict ):
__SCREAMING_SNAKE_CASE : Optional[Any] = MaskFormerForInstanceSegmentation(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
def comm_check_on_output(_lowerCamelCase :Optional[Any] ):
# let's still check that all the required stuff is there
self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.encoder_last_hidden_state is not None )
# okay, now we need to check the logits shape
# due to the encoder compression, masks have a //4 spatial size
self.parent.assertEqual(
result.masks_queries_logits.shape , (self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) , )
# + 1 for null class
self.parent.assertEqual(
result.class_queries_logits.shape , (self.batch_size, self.num_queries, self.num_labels + 1) )
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[str] = model(pixel_values=_lowerCamelCase , pixel_mask=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = model(_lowerCamelCase )
comm_check_on_output(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = model(
pixel_values=_lowerCamelCase , pixel_mask=_lowerCamelCase , mask_labels=_lowerCamelCase , class_labels=_lowerCamelCase )
comm_check_on_output(_lowerCamelCase )
self.parent.assertTrue(result.loss is not None )
self.parent.assertEqual(result.loss.shape , torch.Size([1] ) )
@require_torch
class snake_case ( __UpperCAmelCase , __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = (MaskFormerModel, MaskFormerForInstanceSegmentation) if is_torch_available() else ()
lowerCamelCase__ = (
{'''feature-extraction''': MaskFormerModel, '''image-segmentation''': MaskFormerForInstanceSegmentation}
if is_torch_available()
else {}
)
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Optional[Any] = MaskFormerModelTester(self )
__SCREAMING_SNAKE_CASE : str = ConfigTester(self , config_class=_lowerCamelCase , has_text_modality=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(_lowerCamelCase , **_lowerCamelCase , output_hidden_states=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_maskformer_instance_segmentation_head_model(*_lowerCamelCase )
@unittest.skip(reason='''MaskFormer does not use inputs_embeds''' )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
pass
@unittest.skip(reason='''MaskFormer does not have a get_input_embeddings method''' )
def SCREAMING_SNAKE_CASE_ ( self :int ):
pass
@unittest.skip(reason='''MaskFormer is not a generative model''' )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
pass
@unittest.skip(reason='''MaskFormer does not use token embeddings''' )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
pass
@require_torch_multi_gpu
@unittest.skip(
reason='''MaskFormer has some layers using `add_module` which doesn\'t work well with `nn.DataParallel`''' )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
pass
@unittest.skip('''Will be fixed soon by reducing the size of the model used for common tests.''' )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
pass
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Tuple = model_class(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
__SCREAMING_SNAKE_CASE : List[str] = [*signature.parameters.keys()]
__SCREAMING_SNAKE_CASE : Optional[Any] = ['''pixel_values''']
self.assertListEqual(arg_names[:1] , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
for model_name in ["facebook/maskformer-swin-small-coco"]:
__SCREAMING_SNAKE_CASE : Tuple = MaskFormerModel.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Dict = (self.model_tester.min_size,) * 2
__SCREAMING_SNAKE_CASE : Dict = {
'''pixel_values''': torch.randn((2, 3, *size) , device=_lowerCamelCase ),
'''mask_labels''': torch.randn((2, 1_0, *size) , device=_lowerCamelCase ),
'''class_labels''': torch.zeros(2 , 1_0 , device=_lowerCamelCase ).long(),
}
__SCREAMING_SNAKE_CASE : Dict = MaskFormerForInstanceSegmentation(MaskFormerConfig() ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = model(**_lowerCamelCase )
self.assertTrue(outputs.loss is not None )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(_lowerCamelCase , **_lowerCamelCase , output_hidden_states=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Dict = model_class(_lowerCamelCase ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = model(**_lowerCamelCase , output_attentions=_lowerCamelCase )
self.assertTrue(outputs.attentions is not None )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
if not self.model_tester.is_training:
return
# only MaskFormerForInstanceSegmentation has the loss
__SCREAMING_SNAKE_CASE : List[Any] = self.all_model_classes[1]
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE : int = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.train()
__SCREAMING_SNAKE_CASE : Any = model(_lowerCamelCase , mask_labels=_lowerCamelCase , class_labels=_lowerCamelCase ).loss
loss.backward()
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# only MaskFormerForInstanceSegmentation has the loss
__SCREAMING_SNAKE_CASE : Tuple = self.all_model_classes[1]
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE : Tuple = True
__SCREAMING_SNAKE_CASE : List[str] = True
__SCREAMING_SNAKE_CASE : Any = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.train()
__SCREAMING_SNAKE_CASE : Optional[int] = model(_lowerCamelCase , mask_labels=_lowerCamelCase , class_labels=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = outputs.encoder_hidden_states[0]
encoder_hidden_states.retain_grad()
__SCREAMING_SNAKE_CASE : str = outputs.pixel_decoder_hidden_states[0]
pixel_decoder_hidden_states.retain_grad()
# we requires_grad=True in inputs_embeds (line 2152), the original implementation don't
__SCREAMING_SNAKE_CASE : Optional[int] = outputs.transformer_decoder_hidden_states[0]
transformer_decoder_hidden_states.retain_grad()
__SCREAMING_SNAKE_CASE : int = outputs.attentions[0]
attentions.retain_grad()
outputs.loss.backward(retain_graph=_lowerCamelCase )
self.assertIsNotNone(encoder_hidden_states.grad )
self.assertIsNotNone(pixel_decoder_hidden_states.grad )
self.assertIsNotNone(transformer_decoder_hidden_states.grad )
self.assertIsNotNone(attentions.grad )
_lowerCamelCase = 1e-4
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[Any] = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
return image
@require_vision
@slow
class snake_case ( unittest.TestCase ):
@cached_property
def SCREAMING_SNAKE_CASE_ ( self :str ):
return (
MaskFormerImageProcessor.from_pretrained('''facebook/maskformer-swin-small-coco''' )
if is_vision_available()
else None
)
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = MaskFormerModel.from_pretrained('''facebook/maskformer-swin-small-coco''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = self.default_image_processor
__SCREAMING_SNAKE_CASE : Tuple = prepare_img()
__SCREAMING_SNAKE_CASE : List[Any] = image_processor(_lowerCamelCase , return_tensors='''pt''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = inputs['''pixel_values'''].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 3_2) == 0 and (inputs_shape[-2] % 3_2) == 0 )
# check size
self.assertEqual(_lowerCamelCase , (1, 3, 8_0_0, 1_0_8_8) )
with torch.no_grad():
__SCREAMING_SNAKE_CASE : Tuple = model(**_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.tensor(
[[-0.0_4_8_2, 0.9_2_2_8, 0.4_9_5_1], [-0.2_5_4_7, 0.8_0_1_7, 0.8_5_2_7], [-0.0_0_6_9, 0.3_3_8_5, -0.0_0_8_9]] ).to(_lowerCamelCase )
self.assertTrue(
torch.allclose(
outputs.encoder_last_hidden_state[0, 0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Any = torch.tensor(
[[-0.8_4_2_2, -0.8_4_3_4, -0.9_7_1_8], [-1.0_1_4_4, -0.5_5_6_5, -0.4_1_9_5], [-1.0_0_3_8, -0.4_4_8_4, -0.1_9_6_1]] ).to(_lowerCamelCase )
self.assertTrue(
torch.allclose(
outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Optional[int] = torch.tensor(
[[0.2_8_5_2, -0.0_1_5_9, 0.9_7_3_5], [0.6_2_5_4, 0.1_8_5_8, 0.8_5_2_9], [-0.0_6_8_0, -0.4_1_1_6, 1.8_4_1_3]] ).to(_lowerCamelCase )
self.assertTrue(
torch.allclose(
outputs.transformer_decoder_last_hidden_state[0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE : Tuple = (
MaskFormerForInstanceSegmentation.from_pretrained('''facebook/maskformer-swin-small-coco''' )
.to(_lowerCamelCase )
.eval()
)
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.default_image_processor
__SCREAMING_SNAKE_CASE : Tuple = prepare_img()
__SCREAMING_SNAKE_CASE : str = image_processor(_lowerCamelCase , return_tensors='''pt''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = inputs['''pixel_values'''].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 3_2) == 0 and (inputs_shape[-2] % 3_2) == 0 )
# check size
self.assertEqual(_lowerCamelCase , (1, 3, 8_0_0, 1_0_8_8) )
with torch.no_grad():
__SCREAMING_SNAKE_CASE : str = model(**_lowerCamelCase )
# masks_queries_logits
__SCREAMING_SNAKE_CASE : Any = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape , (1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) , )
__SCREAMING_SNAKE_CASE : List[Any] = [
[-1.3_7_3_7_1_2_4, -1.7_7_2_4_9_3_7, -1.9_3_6_4_2_3_3],
[-1.5_9_7_7_2_8_1, -1.9_8_6_7_9_3_9, -2.1_5_2_3_6_9_5],
[-1.5_7_9_5_3_9_8, -1.9_2_6_9_8_3_2, -2.0_9_3_9_4_2],
]
__SCREAMING_SNAKE_CASE : Any = torch.tensor(_lowerCamelCase ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
# class_queries_logits
__SCREAMING_SNAKE_CASE : List[Any] = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape , (1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor(
[
[1.65_12e00, -5.25_72e00, -3.35_19e00],
[3.61_69e-02, -5.90_25e00, -2.93_13e00],
[1.07_66e-04, -7.76_30e00, -5.12_63e00],
] ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : Tuple = (
MaskFormerForInstanceSegmentation.from_pretrained('''facebook/maskformer-resnet101-coco-stuff''' )
.to(_lowerCamelCase )
.eval()
)
__SCREAMING_SNAKE_CASE : Tuple = self.default_image_processor
__SCREAMING_SNAKE_CASE : str = prepare_img()
__SCREAMING_SNAKE_CASE : List[Any] = image_processor(_lowerCamelCase , return_tensors='''pt''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = inputs['''pixel_values'''].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 3_2) == 0 and (inputs_shape[-2] % 3_2) == 0 )
# check size
self.assertEqual(_lowerCamelCase , (1, 3, 8_0_0, 1_0_8_8) )
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[Any] = model(**_lowerCamelCase )
# masks_queries_logits
__SCREAMING_SNAKE_CASE : Union[str, Any] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape , (1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) , )
__SCREAMING_SNAKE_CASE : List[str] = [[-0.9_0_4_6, -2.6_3_6_6, -4.6_0_6_2], [-3.4_1_7_9, -5.7_8_9_0, -8.8_0_5_7], [-4.9_1_7_9, -7.6_5_6_0, -1_0.7_7_1_1]]
__SCREAMING_SNAKE_CASE : Any = torch.tensor(_lowerCamelCase ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
# class_queries_logits
__SCREAMING_SNAKE_CASE : int = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape , (1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor(
[[4.7_1_8_8, -3.2_5_8_5, -2.8_8_5_7], [6.6_8_7_1, -2.9_1_8_1, -1.2_4_8_7], [7.2_4_4_9, -2.2_7_6_4, -2.1_8_7_4]] ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : int = (
MaskFormerForInstanceSegmentation.from_pretrained('''facebook/maskformer-swin-small-coco''' )
.to(_lowerCamelCase )
.eval()
)
__SCREAMING_SNAKE_CASE : Any = self.default_image_processor
__SCREAMING_SNAKE_CASE : int = image_processor(
[np.zeros((3, 8_0_0, 1_3_3_3) ), np.zeros((3, 8_0_0, 1_3_3_3) )] , segmentation_maps=[np.zeros((3_8_4, 3_8_4) ).astype(np.floataa ), np.zeros((3_8_4, 3_8_4) ).astype(np.floataa )] , return_tensors='''pt''' , )
__SCREAMING_SNAKE_CASE : Dict = inputs['''pixel_values'''].to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = [el.to(_lowerCamelCase ) for el in inputs['''mask_labels''']]
__SCREAMING_SNAKE_CASE : str = [el.to(_lowerCamelCase ) for el in inputs['''class_labels''']]
with torch.no_grad():
__SCREAMING_SNAKE_CASE : int = model(**_lowerCamelCase )
self.assertTrue(outputs.loss is not None )
| 674 | 1 |
"""simple docstring"""
from __future__ import annotations
import unittest
import numpy as np
from transformers import OPTConfig, is_tf_available
from transformers.testing_utils import require_sentencepiece, require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import GPTaTokenizer, TFOPTForCausalLM, TFOPTModel
def lowerCAmelCase_ ( lowercase_ : Union[str, Any] , lowercase_ : Any , lowercase_ : Optional[Any]=None , lowercase_ : str=None ):
'''simple docstring'''
if attention_mask is None:
__SCREAMING_SNAKE_CASE : Optional[int] = tf.cast(tf.math.not_equal(lowercase_ , config.pad_token_id ) , tf.inta )
return {"input_ids": input_ids, "attention_mask": attention_mask}
@require_tf
class snake_case :
lowerCamelCase__ = OPTConfig
lowerCamelCase__ = {}
lowerCamelCase__ = '''gelu'''
def __init__( self :Any , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Optional[Any]=1_3 , _lowerCamelCase :Any=7 , _lowerCamelCase :Any=True , _lowerCamelCase :List[str]=False , _lowerCamelCase :int=9_9 , _lowerCamelCase :Dict=1_6 , _lowerCamelCase :Any=2 , _lowerCamelCase :Any=4 , _lowerCamelCase :Optional[Any]=4 , _lowerCamelCase :Union[str, Any]="gelu" , _lowerCamelCase :List[str]=0.1 , _lowerCamelCase :Dict=0.1 , _lowerCamelCase :List[str]=2_0 , _lowerCamelCase :int=2 , _lowerCamelCase :str=1 , _lowerCamelCase :Optional[int]=0 , _lowerCamelCase :Union[str, Any]=1_6 , _lowerCamelCase :Optional[Any]=1_6 , ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = parent
__SCREAMING_SNAKE_CASE : Optional[Any] = batch_size
__SCREAMING_SNAKE_CASE : Any = seq_length
__SCREAMING_SNAKE_CASE : Any = is_training
__SCREAMING_SNAKE_CASE : Optional[int] = use_labels
__SCREAMING_SNAKE_CASE : List[str] = vocab_size
__SCREAMING_SNAKE_CASE : List[str] = hidden_size
__SCREAMING_SNAKE_CASE : List[Any] = num_hidden_layers
__SCREAMING_SNAKE_CASE : Tuple = num_attention_heads
__SCREAMING_SNAKE_CASE : str = intermediate_size
__SCREAMING_SNAKE_CASE : Optional[Any] = hidden_act
__SCREAMING_SNAKE_CASE : List[str] = hidden_dropout_prob
__SCREAMING_SNAKE_CASE : Optional[int] = attention_probs_dropout_prob
__SCREAMING_SNAKE_CASE : int = max_position_embeddings
__SCREAMING_SNAKE_CASE : Union[str, Any] = eos_token_id
__SCREAMING_SNAKE_CASE : Optional[Any] = pad_token_id
__SCREAMING_SNAKE_CASE : int = bos_token_id
__SCREAMING_SNAKE_CASE : str = embed_dim
__SCREAMING_SNAKE_CASE : List[Any] = word_embed_proj_dim
__SCREAMING_SNAKE_CASE : str = False
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : int = ids_tensor([self.batch_size, self.seq_length - 1] , self.vocab_size )
__SCREAMING_SNAKE_CASE : str = tf.expand_dims(tf.constant([self.eos_token_id] * self.batch_size ) , 1 )
__SCREAMING_SNAKE_CASE : Union[str, Any] = tf.concat([input_ids, eos_tensor] , axis=1 )
__SCREAMING_SNAKE_CASE : List[Any] = self.config_cls(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , ffn_dim=self.intermediate_size , dropout=self.hidden_dropout_prob , attention_dropout=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , eos_token_id=self.eos_token_id , bos_token_id=self.bos_token_id , pad_token_id=self.pad_token_id , embed_dim=self.embed_dim , word_embed_proj_dim=self.word_embed_proj_dim , is_encoder_decoder=_lowerCamelCase , **self.config_updates , )
__SCREAMING_SNAKE_CASE : Optional[int] = prepare_opt_inputs_dict(_lowerCamelCase , _lowerCamelCase )
return config, inputs_dict
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = TFOPTModel(config=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = inputs_dict['''input_ids''']
__SCREAMING_SNAKE_CASE : List[str] = input_ids[:1, :]
__SCREAMING_SNAKE_CASE : Any = inputs_dict['''attention_mask'''][:1, :]
__SCREAMING_SNAKE_CASE : Tuple = 1
# first forward pass
__SCREAMING_SNAKE_CASE : Optional[int] = model(_lowerCamelCase , attention_mask=_lowerCamelCase , use_cache=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
__SCREAMING_SNAKE_CASE : Optional[int] = ids_tensor((self.batch_size, 3) , config.vocab_size )
__SCREAMING_SNAKE_CASE : List[Any] = tf.cast(ids_tensor((self.batch_size, 3) , 2 ) , tf.inta )
# append to next input_ids and
__SCREAMING_SNAKE_CASE : int = tf.concat([input_ids, next_tokens] , axis=-1 )
__SCREAMING_SNAKE_CASE : int = tf.concat([attention_mask, next_attn_mask] , axis=-1 )
__SCREAMING_SNAKE_CASE : Any = model(_lowerCamelCase , attention_mask=_lowerCamelCase )[0]
__SCREAMING_SNAKE_CASE : List[str] = model(_lowerCamelCase , attention_mask=_lowerCamelCase , past_key_values=_lowerCamelCase )[0]
self.parent.assertEqual(next_tokens.shape[1] , output_from_past.shape[1] )
# select random slice
__SCREAMING_SNAKE_CASE : Tuple = int(ids_tensor((1,) , output_from_past.shape[-1] ) )
__SCREAMING_SNAKE_CASE : Optional[int] = output_from_no_past[:, -3:, random_slice_idx]
__SCREAMING_SNAKE_CASE : Union[str, Any] = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(_lowerCamelCase , _lowerCamelCase , rtol=1e-3 )
@require_tf
class snake_case ( __UpperCAmelCase , __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = (TFOPTModel, TFOPTForCausalLM) if is_tf_available() else ()
lowerCamelCase__ = (TFOPTForCausalLM,) if is_tf_available() else ()
lowerCamelCase__ = (
{'''feature-extraction''': TFOPTModel, '''text-generation''': TFOPTForCausalLM} if is_tf_available() else {}
)
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = 10
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Any = TFOPTModelTester(self )
__SCREAMING_SNAKE_CASE : str = ConfigTester(self , config_class=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[int] = self.model_tester.prepare_config_and_inputs_for_common()
def _get_word_embedding_weight(_lowerCamelCase :Any , _lowerCamelCase :Optional[Any] ):
if hasattr(_lowerCamelCase , '''weight''' ):
return embedding_layer.weight
else:
# Here we build the word embeddings weights if not exists.
# And then we retry to get the attribute once built.
model.build()
if hasattr(_lowerCamelCase , '''weight''' ):
return embedding_layer.weight
else:
return None
for model_class in self.all_model_classes:
for size in [config.vocab_size - 1_0, config.vocab_size + 1_0]:
# build the embeddings
__SCREAMING_SNAKE_CASE : str = model_class(config=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = _get_word_embedding_weight(_lowerCamelCase , model.get_input_embeddings() )
__SCREAMING_SNAKE_CASE : Union[str, Any] = _get_word_embedding_weight(_lowerCamelCase , model.get_output_embeddings() )
# reshape the embeddings
model.resize_token_embeddings(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = _get_word_embedding_weight(_lowerCamelCase , model.get_input_embeddings() )
__SCREAMING_SNAKE_CASE : str = _get_word_embedding_weight(_lowerCamelCase , model.get_output_embeddings() )
# check that the resized embeddings size matches the desired size.
__SCREAMING_SNAKE_CASE : Optional[Any] = size if size is not None else config.vocab_size
self.assertEqual(new_input_embeddings.shape[0] , _lowerCamelCase )
# check that weights remain the same after resizing
__SCREAMING_SNAKE_CASE : int = True
for pa, pa in zip(old_input_embeddings.value() , new_input_embeddings.value() ):
if tf.math.reduce_sum(tf.math.abs(pa - pa ) ) > 0:
__SCREAMING_SNAKE_CASE : List[str] = False
self.assertTrue(_lowerCamelCase )
if old_output_embeddings is not None and new_output_embeddings is not None:
self.assertEqual(new_output_embeddings.shape[0] , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = True
for pa, pa in zip(old_output_embeddings.value() , new_output_embeddings.value() ):
if tf.math.reduce_sum(tf.math.abs(pa - pa ) ) > 0:
__SCREAMING_SNAKE_CASE : Tuple = False
self.assertTrue(_lowerCamelCase )
def lowerCAmelCase_ ( lowercase_ : int ):
'''simple docstring'''
return tf.constant(lowercase_ , dtype=tf.intaa )
@require_tf
class snake_case ( unittest.TestCase ):
lowerCamelCase__ = 99
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Dict = tf.ones((4, 1) , dtype=tf.intaa ) * 2
__SCREAMING_SNAKE_CASE : int = tf.concat([ids_tensor((4, 6) , self.vocab_size - 3 ) + 3, eos_column_vector] , axis=1 )
__SCREAMING_SNAKE_CASE : List[Any] = input_ids.shape[0]
__SCREAMING_SNAKE_CASE : Optional[Any] = OPTConfig(
vocab_size=self.vocab_size , hidden_size=2_4 , num_hidden_layers=2 , num_attention_heads=2 , ffn_dim=3_2 , max_position_embeddings=4_8 , eos_token_id=2 , pad_token_id=1 , bos_token_id=0 , )
return config, input_ids, batch_size
@require_sentencepiece
@require_tf
class snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : List[Any] = TFOPTModel.from_pretrained('''facebook/opt-350m''' )
__SCREAMING_SNAKE_CASE : Any = _long_tensor([[0, 3_1_4_1_4, 2_3_2, 3_2_8, 7_4_0, 1_1_4_0, 1_2_6_9_5, 6_9, 4_6_0_7_8, 1_5_8_8, 2]] )
__SCREAMING_SNAKE_CASE : List[str] = tf.not_equal(_lowerCamelCase , model.config.pad_token_id )
with tf.GradientTape():
__SCREAMING_SNAKE_CASE : Any = model(input_ids=_lowerCamelCase , attention_mask=_lowerCamelCase ).last_hidden_state
__SCREAMING_SNAKE_CASE : Dict = (1, 1_1, 5_1_2)
self.assertEqual(output.shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = tf.constant(
[[-0.2_8_7_3, -1.9_2_1_8, -0.3_0_3_3], [-1.2_7_1_0, -0.1_3_3_8, -0.1_9_0_2], [0.4_0_9_5, 0.1_2_1_4, -1.3_1_2_1]] )
self.assertTrue(np.allclose(output[:, :3, :3] , _lowerCamelCase , atol=4e-3 ) )
__SCREAMING_SNAKE_CASE : int = tf.function(_lowerCamelCase , jit_compile=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = xla_generate(_lowerCamelCase , _lowerCamelCase )[0]
self.assertTrue(np.allclose(output[:, :3, :3] , _lowerCamelCase , atol=4e-2 ) )
@require_tf
@slow
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
super().setUp()
__SCREAMING_SNAKE_CASE : Tuple = '''facebook/opt-350m'''
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : str = TFOPTForCausalLM.from_pretrained(self.path_model )
__SCREAMING_SNAKE_CASE : Dict = GPTaTokenizer.from_pretrained(self.path_model )
__SCREAMING_SNAKE_CASE : Union[str, Any] = [
'''Today is a beautiful day and I want to''',
'''In the city of''',
'''Paris is the capital of France and''',
'''Computers and mobile phones have taken''',
]
# verify that prompt without BOS token is identical to Metaseq -> add_special_tokens=False
__SCREAMING_SNAKE_CASE : Tuple = tokenizer(_lowerCamelCase , return_tensors='''tf''' , padding=_lowerCamelCase , add_special_tokens=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = tf.math.reduce_mean(model(inputs.input_ids , attention_mask=inputs.attention_mask )[0] , axis=-1 )
__SCREAMING_SNAKE_CASE : Optional[Any] = tf.constant(
[
[1.3_8_5_1, -1_3.8_9_2_3, -1_0.5_2_2_9, -1_0.7_5_3_3, -0.2_3_0_9, -1_0.2_3_8_4, -0.5_3_6_5, -9.0_9_4_7, -5.1_6_7_0],
[-4.7_0_7_3, -1_0.6_2_7_6, -3.9_4_1_5, -2_1.5_2_4_2, -0.2_8_2_2, -0.2_8_2_2, -0.2_8_2_2, -0.2_8_2_2, -0.2_8_2_2],
[0.6_2_4_7, -3.4_2_2_9, -8.9_1_7_9, -1.4_2_9_7, -1_4.1_6_5_0, 1.4_1_4_6, -9.0_2_1_8, -0.2_7_0_3, -0.2_7_0_3],
[6.4_7_8_3, -1.9_9_1_3, -1_0.7_9_2_6, -2.3_3_3_6, 1.5_0_9_2, -0.9_9_7_4, -6.8_2_1_3, 1.3_4_7_7, 1.3_4_7_7],
] )
self.assertTrue(np.allclose(_lowerCamelCase , _lowerCamelCase , atol=1e-4 ) )
__SCREAMING_SNAKE_CASE : int = tf.function(_lowerCamelCase , jit_compile=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = tf.math.reduce_mean(xla_generate(inputs.input_ids , attention_mask=inputs.attention_mask )[0] , axis=-1 )
self.assertTrue(np.allclose(_lowerCamelCase , _lowerCamelCase , atol=1e-4 ) )
@require_tf
@slow
class snake_case ( unittest.TestCase ):
@property
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
return [
"Today is a beautiful day and I want",
"In the city of",
"Paris is the capital of France and",
"Computers and mobile phones have taken",
]
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = '''facebook/opt-125m'''
__SCREAMING_SNAKE_CASE : int = [
'''Today is a beautiful day and I want to''',
'''In the city of New York, the city''',
'''Paris is the capital of France and the capital''',
'''Computers and mobile phones have taken over the''',
]
__SCREAMING_SNAKE_CASE : str = []
__SCREAMING_SNAKE_CASE : Any = GPTaTokenizer.from_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = TFOPTForCausalLM.from_pretrained(_lowerCamelCase )
for prompt in self.prompts:
__SCREAMING_SNAKE_CASE : int = tokenizer(_lowerCamelCase , return_tensors='''tf''' ).input_ids
__SCREAMING_SNAKE_CASE : int = model.generate(_lowerCamelCase , max_length=1_0 )
__SCREAMING_SNAKE_CASE : Dict = tokenizer.batch_decode(_lowerCamelCase , skip_special_tokens=_lowerCamelCase )
predicted_outputs += generated_string
self.assertListEqual(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Dict = '''facebook/opt-350m'''
__SCREAMING_SNAKE_CASE : str = GPTaTokenizer.from_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = TFOPTForCausalLM.from_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = '''left'''
# use different length sentences to test batching
__SCREAMING_SNAKE_CASE : Union[str, Any] = [
'''Hello, my dog is a little''',
'''Today, I''',
]
__SCREAMING_SNAKE_CASE : str = tokenizer(_lowerCamelCase , return_tensors='''tf''' , padding=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = inputs['''input_ids''']
__SCREAMING_SNAKE_CASE : Union[str, Any] = model.generate(input_ids=_lowerCamelCase , attention_mask=inputs['''attention_mask'''] )
__SCREAMING_SNAKE_CASE : Any = tokenizer(sentences[0] , return_tensors='''tf''' ).input_ids
__SCREAMING_SNAKE_CASE : List[Any] = model.generate(input_ids=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = inputs_non_padded.shape[-1] - tf.math.reduce_sum(
tf.cast(inputs['''attention_mask'''][-1] , tf.intaa ) )
__SCREAMING_SNAKE_CASE : int = tokenizer(sentences[1] , return_tensors='''tf''' ).input_ids
__SCREAMING_SNAKE_CASE : Optional[int] = model.generate(input_ids=_lowerCamelCase , max_length=model.config.max_length - num_paddings )
__SCREAMING_SNAKE_CASE : List[Any] = tokenizer.batch_decode(_lowerCamelCase , skip_special_tokens=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = tokenizer.decode(output_non_padded[0] , skip_special_tokens=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = tokenizer.decode(output_padded[0] , skip_special_tokens=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = [
'''Hello, my dog is a little bit of a dork.\nI\'m a little bit''',
'''Today, I was in the middle of a conversation with a friend about the''',
]
self.assertListEqual(_lowerCamelCase , _lowerCamelCase )
self.assertListEqual(_lowerCamelCase , [non_padded_sentence, padded_sentence] )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : Tuple = '''facebook/opt-350m'''
__SCREAMING_SNAKE_CASE : Dict = [
'''Today is a beautiful day and I want to''',
'''In the city of San Francisco, the city''',
'''Paris is the capital of France and the capital''',
'''Computers and mobile phones have taken over the''',
]
__SCREAMING_SNAKE_CASE : int = []
__SCREAMING_SNAKE_CASE : Any = GPTaTokenizer.from_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = TFOPTForCausalLM.from_pretrained(_lowerCamelCase )
for prompt in self.prompts:
__SCREAMING_SNAKE_CASE : Optional[int] = tokenizer(_lowerCamelCase , return_tensors='''tf''' ).input_ids
__SCREAMING_SNAKE_CASE : List[Any] = model.generate(_lowerCamelCase , max_length=1_0 )
__SCREAMING_SNAKE_CASE : str = tokenizer.batch_decode(_lowerCamelCase , skip_special_tokens=_lowerCamelCase )
predicted_outputs += generated_string
self.assertListEqual(_lowerCamelCase , _lowerCamelCase )
| 674 |
"""simple docstring"""
import contextlib
import importlib
import io
import unittest
import transformers
# Try to import everything from transformers to ensure every object can be loaded.
from transformers import * # noqa F406
from transformers.testing_utils import DUMMY_UNKNOWN_IDENTIFIER, require_flax, require_tf, require_torch
from transformers.utils import ContextManagers, find_labels, is_flax_available, is_tf_available, is_torch_available
if is_torch_available():
from transformers import BertForPreTraining, BertForQuestionAnswering, BertForSequenceClassification
if is_tf_available():
from transformers import TFBertForPreTraining, TFBertForQuestionAnswering, TFBertForSequenceClassification
if is_flax_available():
from transformers import FlaxBertForPreTraining, FlaxBertForQuestionAnswering, FlaxBertForSequenceClassification
_lowerCamelCase = DUMMY_UNKNOWN_IDENTIFIER
# An actual model hosted on huggingface.co
_lowerCamelCase = '''main'''
# Default branch name
_lowerCamelCase = '''f2c752cfc5c0ab6f4bdec59acea69eefbee381c2'''
# One particular commit (not the top of `main`)
_lowerCamelCase = '''aaaaaaa'''
# This commit does not exist, so we should 404.
_lowerCamelCase = '''d9e9f15bc825e4b2c9249e9578f884bbcb5e3684'''
# Sha-1 of config.json on the top of `main`, for checking purposes
_lowerCamelCase = '''4b243c475af8d0a7754e87d7d096c92e5199ec2fe168a2ee7998e3b8e9bcb1d3'''
@contextlib.contextmanager
def lowerCAmelCase_ ( ):
'''simple docstring'''
print('''Welcome!''' )
yield
print('''Bye!''' )
@contextlib.contextmanager
def lowerCAmelCase_ ( ):
'''simple docstring'''
print('''Bonjour!''' )
yield
print('''Au revoir!''' )
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
# If the spec is missing, importlib would not be able to import the module dynamically.
assert transformers.__spec__ is not None
assert importlib.util.find_spec('''transformers''' ) is not None
class snake_case ( unittest.TestCase ):
@unittest.mock.patch('''sys.stdout''' , new_callable=io.StringIO )
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :Optional[int] ):
with ContextManagers([] ):
print('''Transformers are awesome!''' )
# The print statement adds a new line at the end of the output
self.assertEqual(mock_stdout.getvalue() , '''Transformers are awesome!\n''' )
@unittest.mock.patch('''sys.stdout''' , new_callable=io.StringIO )
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :Optional[int] ):
with ContextManagers([context_en()] ):
print('''Transformers are awesome!''' )
# The output should be wrapped with an English welcome and goodbye
self.assertEqual(mock_stdout.getvalue() , '''Welcome!\nTransformers are awesome!\nBye!\n''' )
@unittest.mock.patch('''sys.stdout''' , new_callable=io.StringIO )
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :List[str] ):
with ContextManagers([context_fr(), context_en()] ):
print('''Transformers are awesome!''' )
# The output should be wrapped with an English and French welcome and goodbye
self.assertEqual(mock_stdout.getvalue() , '''Bonjour!\nWelcome!\nTransformers are awesome!\nBye!\nAu revoir!\n''' )
@require_torch
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels'''] )
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels''', '''next_sentence_label'''] )
self.assertEqual(find_labels(_lowerCamelCase ) , ['''start_positions''', '''end_positions'''] )
class snake_case ( __UpperCAmelCase ):
pass
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels'''] )
@require_tf
def SCREAMING_SNAKE_CASE_ ( self :int ):
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels'''] )
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels''', '''next_sentence_label'''] )
self.assertEqual(find_labels(_lowerCamelCase ) , ['''start_positions''', '''end_positions'''] )
class snake_case ( __UpperCAmelCase ):
pass
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels'''] )
@require_flax
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
# Flax models don't have labels
self.assertEqual(find_labels(_lowerCamelCase ) , [] )
self.assertEqual(find_labels(_lowerCamelCase ) , [] )
self.assertEqual(find_labels(_lowerCamelCase ) , [] )
class snake_case ( __UpperCAmelCase ):
pass
self.assertEqual(find_labels(_lowerCamelCase ) , [] )
| 674 | 1 |
"""simple docstring"""
import gc
import random
import unittest
import numpy as np
import torch
from transformers import (
CLIPImageProcessor,
CLIPTextConfig,
CLIPTextModel,
CLIPTokenizer,
CLIPVisionConfig,
CLIPVisionModelWithProjection,
)
from diffusers import AutoencoderKL, DDIMScheduler, DDPMScheduler, StableUnCLIPImgaImgPipeline, UNetaDConditionModel
from diffusers.pipelines.pipeline_utils import DiffusionPipeline
from diffusers.pipelines.stable_diffusion.stable_unclip_image_normalizer import StableUnCLIPImageNormalizer
from diffusers.utils.import_utils import is_xformers_available
from diffusers.utils.testing_utils import (
enable_full_determinism,
floats_tensor,
load_image,
load_numpy,
require_torch_gpu,
skip_mps,
slow,
torch_device,
)
from ..pipeline_params import TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS, TEXT_GUIDED_IMAGE_VARIATION_PARAMS
from ..test_pipelines_common import (
PipelineKarrasSchedulerTesterMixin,
PipelineLatentTesterMixin,
PipelineTesterMixin,
assert_mean_pixel_difference,
)
enable_full_determinism()
class snake_case ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = StableUnCLIPImgaImgPipeline
lowerCamelCase__ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS
lowerCamelCase__ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS
lowerCamelCase__ = frozenset(
[] ) # TO-DO: update image_params once pipeline is refactored with VaeImageProcessor.preprocess
lowerCamelCase__ = frozenset([] )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : List[Any] = 3_2
__SCREAMING_SNAKE_CASE : Any = embedder_hidden_size
# image encoding components
__SCREAMING_SNAKE_CASE : int = CLIPImageProcessor(crop_size=3_2 , size=3_2 )
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : str = CLIPVisionModelWithProjection(
CLIPVisionConfig(
hidden_size=_lowerCamelCase , projection_dim=_lowerCamelCase , num_hidden_layers=5 , num_attention_heads=4 , image_size=3_2 , intermediate_size=3_7 , patch_size=1 , ) )
# regular denoising components
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Tuple = StableUnCLIPImageNormalizer(embedding_dim=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = DDPMScheduler(beta_schedule='''squaredcos_cap_v2''' )
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : List[str] = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''' )
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : List[Any] = CLIPTextModel(
CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=_lowerCamelCase , projection_dim=3_2 , intermediate_size=3_7 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_0_0_0 , ) )
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : int = UNetaDConditionModel(
sample_size=3_2 , in_channels=4 , out_channels=4 , down_block_types=('''CrossAttnDownBlock2D''', '''DownBlock2D''') , up_block_types=('''UpBlock2D''', '''CrossAttnUpBlock2D''') , block_out_channels=(3_2, 6_4) , attention_head_dim=(2, 4) , class_embed_type='''projection''' , projection_class_embeddings_input_dim=embedder_projection_dim * 2 , cross_attention_dim=_lowerCamelCase , layers_per_block=1 , upcast_attention=_lowerCamelCase , use_linear_projection=_lowerCamelCase , )
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : List[Any] = DDIMScheduler(
beta_schedule='''scaled_linear''' , beta_start=0.0_0_0_8_5 , beta_end=0.0_1_2 , prediction_type='''v_prediction''' , set_alpha_to_one=_lowerCamelCase , steps_offset=1 , )
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Any = AutoencoderKL()
__SCREAMING_SNAKE_CASE : Optional[int] = {
# image encoding components
'''feature_extractor''': feature_extractor,
'''image_encoder''': image_encoder.eval(),
# image noising components
'''image_normalizer''': image_normalizer.eval(),
'''image_noising_scheduler''': image_noising_scheduler,
# regular denoising components
'''tokenizer''': tokenizer,
'''text_encoder''': text_encoder.eval(),
'''unet''': unet.eval(),
'''scheduler''': scheduler,
'''vae''': vae.eval(),
}
return components
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :List[str]=0 , _lowerCamelCase :Optional[int]=True ):
if str(_lowerCamelCase ).startswith('''mps''' ):
__SCREAMING_SNAKE_CASE : str = torch.manual_seed(_lowerCamelCase )
else:
__SCREAMING_SNAKE_CASE : List[Any] = torch.Generator(device=_lowerCamelCase ).manual_seed(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = floats_tensor((1, 3, 3_2, 3_2) , rng=random.Random(_lowerCamelCase ) ).to(_lowerCamelCase )
if pil_image:
__SCREAMING_SNAKE_CASE : Dict = input_image * 0.5 + 0.5
__SCREAMING_SNAKE_CASE : Optional[Any] = input_image.clamp(0 , 1 )
__SCREAMING_SNAKE_CASE : List[Any] = input_image.cpu().permute(0 , 2 , 3 , 1 ).float().numpy()
__SCREAMING_SNAKE_CASE : Union[str, Any] = DiffusionPipeline.numpy_to_pil(_lowerCamelCase )[0]
return {
"prompt": "An anime racoon running a marathon",
"image": input_image,
"generator": generator,
"num_inference_steps": 2,
"output_type": "np",
}
@skip_mps
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : int = '''cpu''' # ensure determinism for the device-dependent torch.Generator
__SCREAMING_SNAKE_CASE : str = self.get_dummy_components()
__SCREAMING_SNAKE_CASE : List[str] = StableUnCLIPImgaImgPipeline(**_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = sd_pipe.to(_lowerCamelCase )
sd_pipe.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = self.get_dummy_inputs(_lowerCamelCase )
inputs.update({'''image_embeds''': None} )
__SCREAMING_SNAKE_CASE : str = sd_pipe(**_lowerCamelCase ).images
__SCREAMING_SNAKE_CASE : Tuple = image[0, -3:, -3:, -1]
assert image.shape == (1, 3_2, 3_2, 3)
__SCREAMING_SNAKE_CASE : List[str] = np.array([0.3_8_7_2, 0.7_2_2_4, 0.5_6_0_1, 0.4_7_4_1, 0.6_8_7_2, 0.5_8_1_4, 0.4_6_3_6, 0.3_8_6_7, 0.5_0_7_8] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : Optional[int] = torch_device in ['''cpu''', '''mps''']
self._test_attention_slicing_forward_pass(test_max_difference=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Any = torch_device in ['''cpu''', '''mps''']
self._test_inference_batch_single_identical(test_max_difference=_lowerCamelCase )
@unittest.skipIf(
torch_device != '''cuda''' or not is_xformers_available() , reason='''XFormers attention is only available with CUDA and `xformers` installed''' , )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
self._test_xformers_attention_forwardGenerator_pass(test_max_difference=_lowerCamelCase )
@slow
@require_torch_gpu
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Any ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE : int = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/stable_unclip/turtle.png''' )
__SCREAMING_SNAKE_CASE : List[str] = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/stable_unclip/stable_unclip_2_1_l_img2img_anime_turtle_fp16.npy''' )
__SCREAMING_SNAKE_CASE : Optional[Any] = StableUnCLIPImgaImgPipeline.from_pretrained(
'''fusing/stable-unclip-2-1-l-img2img''' , torch_dtype=torch.floataa )
pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
# stable unclip will oom when integration tests are run on a V100,
# so turn on memory savings
pipe.enable_attention_slicing()
pipe.enable_sequential_cpu_offload()
__SCREAMING_SNAKE_CASE : List[str] = torch.Generator(device='''cpu''' ).manual_seed(0 )
__SCREAMING_SNAKE_CASE : str = pipe(_lowerCamelCase , '''anime turle''' , generator=_lowerCamelCase , output_type='''np''' )
__SCREAMING_SNAKE_CASE : str = output.images[0]
assert image.shape == (7_6_8, 7_6_8, 3)
assert_mean_pixel_difference(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : List[str] = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/stable_unclip/turtle.png''' )
__SCREAMING_SNAKE_CASE : Dict = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/stable_unclip/stable_unclip_2_1_h_img2img_anime_turtle_fp16.npy''' )
__SCREAMING_SNAKE_CASE : List[Any] = StableUnCLIPImgaImgPipeline.from_pretrained(
'''fusing/stable-unclip-2-1-h-img2img''' , torch_dtype=torch.floataa )
pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
# stable unclip will oom when integration tests are run on a V100,
# so turn on memory savings
pipe.enable_attention_slicing()
pipe.enable_sequential_cpu_offload()
__SCREAMING_SNAKE_CASE : Any = torch.Generator(device='''cpu''' ).manual_seed(0 )
__SCREAMING_SNAKE_CASE : Union[str, Any] = pipe(_lowerCamelCase , '''anime turle''' , generator=_lowerCamelCase , output_type='''np''' )
__SCREAMING_SNAKE_CASE : List[Any] = output.images[0]
assert image.shape == (7_6_8, 7_6_8, 3)
assert_mean_pixel_difference(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : Optional[int] = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/stable_unclip/turtle.png''' )
torch.cuda.empty_cache()
torch.cuda.reset_max_memory_allocated()
torch.cuda.reset_peak_memory_stats()
__SCREAMING_SNAKE_CASE : Any = StableUnCLIPImgaImgPipeline.from_pretrained(
'''fusing/stable-unclip-2-1-h-img2img''' , torch_dtype=torch.floataa )
__SCREAMING_SNAKE_CASE : str = pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
pipe.enable_attention_slicing()
pipe.enable_sequential_cpu_offload()
__SCREAMING_SNAKE_CASE : int = pipe(
_lowerCamelCase , '''anime turtle''' , num_inference_steps=2 , output_type='''np''' , )
__SCREAMING_SNAKE_CASE : Tuple = torch.cuda.max_memory_allocated()
# make sure that less than 7 GB is allocated
assert mem_bytes < 7 * 1_0**9
| 674 |
"""simple docstring"""
import json
import pathlib
import unittest
import numpy as np
from transformers.testing_utils import require_torch, require_vision, slow
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import YolosImageProcessor
class snake_case ( unittest.TestCase ):
def __init__( self :List[Any] , _lowerCamelCase :Dict , _lowerCamelCase :Tuple=7 , _lowerCamelCase :Dict=3 , _lowerCamelCase :Optional[Any]=3_0 , _lowerCamelCase :List[str]=4_0_0 , _lowerCamelCase :Union[str, Any]=True , _lowerCamelCase :Union[str, Any]=None , _lowerCamelCase :List[Any]=True , _lowerCamelCase :Any=[0.5, 0.5, 0.5] , _lowerCamelCase :Dict=[0.5, 0.5, 0.5] , _lowerCamelCase :Dict=True , _lowerCamelCase :str=1 / 2_5_5 , _lowerCamelCase :Union[str, Any]=True , ):
# by setting size["longest_edge"] > max_resolution we're effectively not testing this :p
__SCREAMING_SNAKE_CASE : Optional[Any] = size if size is not None else {'''shortest_edge''': 1_8, '''longest_edge''': 1_3_3_3}
__SCREAMING_SNAKE_CASE : List[str] = parent
__SCREAMING_SNAKE_CASE : Dict = batch_size
__SCREAMING_SNAKE_CASE : str = num_channels
__SCREAMING_SNAKE_CASE : Union[str, Any] = min_resolution
__SCREAMING_SNAKE_CASE : Union[str, Any] = max_resolution
__SCREAMING_SNAKE_CASE : Tuple = do_resize
__SCREAMING_SNAKE_CASE : Union[str, Any] = size
__SCREAMING_SNAKE_CASE : int = do_normalize
__SCREAMING_SNAKE_CASE : List[Any] = image_mean
__SCREAMING_SNAKE_CASE : Tuple = image_std
__SCREAMING_SNAKE_CASE : Dict = do_rescale
__SCREAMING_SNAKE_CASE : Optional[int] = rescale_factor
__SCREAMING_SNAKE_CASE : List[Any] = do_pad
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
return {
"do_resize": self.do_resize,
"size": self.size,
"do_normalize": self.do_normalize,
"image_mean": self.image_mean,
"image_std": self.image_std,
"do_rescale": self.do_rescale,
"rescale_factor": self.rescale_factor,
"do_pad": self.do_pad,
}
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :Dict , _lowerCamelCase :List[Any]=False ):
if not batched:
__SCREAMING_SNAKE_CASE : str = image_inputs[0]
if isinstance(_lowerCamelCase , Image.Image ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[int] = image.size
else:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = image.shape[1], image.shape[2]
if w < h:
__SCREAMING_SNAKE_CASE : str = int(self.size['''shortest_edge'''] * h / w )
__SCREAMING_SNAKE_CASE : int = self.size['''shortest_edge''']
elif w > h:
__SCREAMING_SNAKE_CASE : Optional[Any] = self.size['''shortest_edge''']
__SCREAMING_SNAKE_CASE : int = int(self.size['''shortest_edge'''] * w / h )
else:
__SCREAMING_SNAKE_CASE : List[str] = self.size['''shortest_edge''']
__SCREAMING_SNAKE_CASE : List[str] = self.size['''shortest_edge''']
else:
__SCREAMING_SNAKE_CASE : Optional[Any] = []
for image in image_inputs:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = self.get_expected_values([image] )
expected_values.append((expected_height, expected_width) )
__SCREAMING_SNAKE_CASE : Optional[int] = max(_lowerCamelCase , key=lambda _lowerCamelCase : item[0] )[0]
__SCREAMING_SNAKE_CASE : int = max(_lowerCamelCase , key=lambda _lowerCamelCase : item[1] )[1]
return expected_height, expected_width
@require_torch
@require_vision
class snake_case ( __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = YolosImageProcessor if is_vision_available() else None
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[Any] = YolosImageProcessingTester(self )
@property
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
return self.image_processor_tester.prepare_image_processor_dict()
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : Any = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(_lowerCamelCase , '''image_mean''' ) )
self.assertTrue(hasattr(_lowerCamelCase , '''image_std''' ) )
self.assertTrue(hasattr(_lowerCamelCase , '''do_normalize''' ) )
self.assertTrue(hasattr(_lowerCamelCase , '''do_resize''' ) )
self.assertTrue(hasattr(_lowerCamelCase , '''size''' ) )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.image_processing_class.from_dict(self.image_processor_dict )
self.assertEqual(image_processor.size , {'''shortest_edge''': 1_8, '''longest_edge''': 1_3_3_3} )
self.assertEqual(image_processor.do_pad , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = self.image_processing_class.from_dict(
self.image_processor_dict , size=4_2 , max_size=8_4 , pad_and_return_pixel_mask=_lowerCamelCase )
self.assertEqual(image_processor.size , {'''shortest_edge''': 4_2, '''longest_edge''': 8_4} )
self.assertEqual(image_processor.do_pad , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
pass
def SCREAMING_SNAKE_CASE_ ( self :int ):
# Initialize image_processing
__SCREAMING_SNAKE_CASE : Optional[int] = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
__SCREAMING_SNAKE_CASE : List[str] = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowerCamelCase )
for image in image_inputs:
self.assertIsInstance(_lowerCamelCase , Image.Image )
# Test not batched input
__SCREAMING_SNAKE_CASE : str = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.image_processor_tester.get_expected_values(_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.image_processor_tester.get_expected_values(_lowerCamelCase , batched=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = image_processing(_lowerCamelCase , return_tensors='''pt''' ).pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# Initialize image_processing
__SCREAMING_SNAKE_CASE : Optional[Any] = self.image_processing_class(**self.image_processor_dict )
# create random numpy tensors
__SCREAMING_SNAKE_CASE : Optional[int] = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowerCamelCase , numpify=_lowerCamelCase )
for image in image_inputs:
self.assertIsInstance(_lowerCamelCase , np.ndarray )
# Test not batched input
__SCREAMING_SNAKE_CASE : Dict = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[str] = self.image_processor_tester.get_expected_values(_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
__SCREAMING_SNAKE_CASE : List[Any] = image_processing(_lowerCamelCase , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = self.image_processor_tester.get_expected_values(_lowerCamelCase , batched=_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def SCREAMING_SNAKE_CASE_ ( self :Any ):
# Initialize image_processing
__SCREAMING_SNAKE_CASE : str = self.image_processing_class(**self.image_processor_dict )
# create random PyTorch tensors
__SCREAMING_SNAKE_CASE : Any = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowerCamelCase , torchify=_lowerCamelCase )
for image in image_inputs:
self.assertIsInstance(_lowerCamelCase , torch.Tensor )
# Test not batched input
__SCREAMING_SNAKE_CASE : List[str] = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = self.image_processor_tester.get_expected_values(_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
__SCREAMING_SNAKE_CASE : Optional[int] = image_processing(_lowerCamelCase , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = self.image_processor_tester.get_expected_values(_lowerCamelCase , batched=_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
# Initialize image_processings
__SCREAMING_SNAKE_CASE : Dict = self.image_processing_class(**self.image_processor_dict )
__SCREAMING_SNAKE_CASE : Tuple = self.image_processing_class(do_resize=_lowerCamelCase , do_normalize=_lowerCamelCase , do_rescale=_lowerCamelCase )
# create random PyTorch tensors
__SCREAMING_SNAKE_CASE : int = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowerCamelCase , torchify=_lowerCamelCase )
for image in image_inputs:
self.assertIsInstance(_lowerCamelCase , torch.Tensor )
# Test whether the method "pad" and calling the image processor return the same tensors
__SCREAMING_SNAKE_CASE : Union[str, Any] = image_processing_a.pad(_lowerCamelCase , return_tensors='''pt''' )
__SCREAMING_SNAKE_CASE : Union[str, Any] = image_processing_a(_lowerCamelCase , return_tensors='''pt''' )
self.assertTrue(
torch.allclose(encoded_images_with_method['''pixel_values'''] , encoded_images['''pixel_values'''] , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE_ ( self :int ):
# prepare image and target
__SCREAMING_SNAKE_CASE : Tuple = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
with open('''./tests/fixtures/tests_samples/COCO/coco_annotations.txt''' , '''r''' ) as f:
__SCREAMING_SNAKE_CASE : Tuple = json.loads(f.read() )
__SCREAMING_SNAKE_CASE : Union[str, Any] = {'''image_id''': 3_9_7_6_9, '''annotations''': target}
# encode them
__SCREAMING_SNAKE_CASE : List[Any] = YolosImageProcessor.from_pretrained('''hustvl/yolos-small''' )
__SCREAMING_SNAKE_CASE : Optional[Any] = image_processing(images=_lowerCamelCase , annotations=_lowerCamelCase , return_tensors='''pt''' )
# verify pixel values
__SCREAMING_SNAKE_CASE : Optional[int] = torch.Size([1, 3, 8_0_0, 1_0_6_6] )
self.assertEqual(encoding['''pixel_values'''].shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor([0.2_7_9_6, 0.3_1_3_8, 0.3_4_8_1] )
self.assertTrue(torch.allclose(encoding['''pixel_values'''][0, 0, 0, :3] , _lowerCamelCase , atol=1e-4 ) )
# verify area
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.tensor([5_8_8_7.9_6_0_0, 1_1_2_5_0.2_0_6_1, 4_8_9_3_5_3.8_4_3_8, 8_3_7_1_2_2.7_5_0_0, 1_4_7_9_6_7.5_1_5_6, 1_6_5_7_3_2.3_4_3_8] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''area'''] , _lowerCamelCase ) )
# verify boxes
__SCREAMING_SNAKE_CASE : List[str] = torch.Size([6, 4] )
self.assertEqual(encoding['''labels'''][0]['''boxes'''].shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = torch.tensor([0.5_5_0_3, 0.2_7_6_5, 0.0_6_0_4, 0.2_2_1_5] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''boxes'''][0] , _lowerCamelCase , atol=1e-3 ) )
# verify image_id
__SCREAMING_SNAKE_CASE : Optional[Any] = torch.tensor([3_9_7_6_9] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''image_id'''] , _lowerCamelCase ) )
# verify is_crowd
__SCREAMING_SNAKE_CASE : str = torch.tensor([0, 0, 0, 0, 0, 0] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''iscrowd'''] , _lowerCamelCase ) )
# verify class_labels
__SCREAMING_SNAKE_CASE : Dict = torch.tensor([7_5, 7_5, 6_3, 6_5, 1_7, 1_7] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''class_labels'''] , _lowerCamelCase ) )
# verify orig_size
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.tensor([4_8_0, 6_4_0] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''orig_size'''] , _lowerCamelCase ) )
# verify size
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor([8_0_0, 1_0_6_6] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''size'''] , _lowerCamelCase ) )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# prepare image, target and masks_path
__SCREAMING_SNAKE_CASE : List[str] = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
with open('''./tests/fixtures/tests_samples/COCO/coco_panoptic_annotations.txt''' , '''r''' ) as f:
__SCREAMING_SNAKE_CASE : Union[str, Any] = json.loads(f.read() )
__SCREAMING_SNAKE_CASE : Dict = {'''file_name''': '''000000039769.png''', '''image_id''': 3_9_7_6_9, '''segments_info''': target}
__SCREAMING_SNAKE_CASE : Optional[int] = pathlib.Path('''./tests/fixtures/tests_samples/COCO/coco_panoptic''' )
# encode them
__SCREAMING_SNAKE_CASE : Any = YolosImageProcessor(format='''coco_panoptic''' )
__SCREAMING_SNAKE_CASE : Dict = image_processing(images=_lowerCamelCase , annotations=_lowerCamelCase , masks_path=_lowerCamelCase , return_tensors='''pt''' )
# verify pixel values
__SCREAMING_SNAKE_CASE : str = torch.Size([1, 3, 8_0_0, 1_0_6_6] )
self.assertEqual(encoding['''pixel_values'''].shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = torch.tensor([0.2_7_9_6, 0.3_1_3_8, 0.3_4_8_1] )
self.assertTrue(torch.allclose(encoding['''pixel_values'''][0, 0, 0, :3] , _lowerCamelCase , atol=1e-4 ) )
# verify area
__SCREAMING_SNAKE_CASE : Any = torch.tensor([1_4_7_9_7_9.6_8_7_5, 1_6_5_5_2_7.0_4_6_9, 4_8_4_6_3_8.5_9_3_8, 1_1_2_9_2.9_3_7_5, 5_8_7_9.6_5_6_2, 7_6_3_4.1_1_4_7] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''area'''] , _lowerCamelCase ) )
# verify boxes
__SCREAMING_SNAKE_CASE : List[str] = torch.Size([6, 4] )
self.assertEqual(encoding['''labels'''][0]['''boxes'''].shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = torch.tensor([0.2_6_2_5, 0.5_4_3_7, 0.4_6_8_8, 0.8_6_2_5] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''boxes'''][0] , _lowerCamelCase , atol=1e-3 ) )
# verify image_id
__SCREAMING_SNAKE_CASE : Dict = torch.tensor([3_9_7_6_9] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''image_id'''] , _lowerCamelCase ) )
# verify is_crowd
__SCREAMING_SNAKE_CASE : List[Any] = torch.tensor([0, 0, 0, 0, 0, 0] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''iscrowd'''] , _lowerCamelCase ) )
# verify class_labels
__SCREAMING_SNAKE_CASE : Optional[Any] = torch.tensor([1_7, 1_7, 6_3, 7_5, 7_5, 9_3] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''class_labels'''] , _lowerCamelCase ) )
# verify masks
__SCREAMING_SNAKE_CASE : Optional[Any] = 8_2_2_8_7_3
self.assertEqual(encoding['''labels'''][0]['''masks'''].sum().item() , _lowerCamelCase )
# verify orig_size
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor([4_8_0, 6_4_0] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''orig_size'''] , _lowerCamelCase ) )
# verify size
__SCREAMING_SNAKE_CASE : Any = torch.tensor([8_0_0, 1_0_6_6] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''size'''] , _lowerCamelCase ) )
| 674 | 1 |
"""simple docstring"""
import inspect
import unittest
import warnings
from math import ceil, floor
from transformers import LevitConfig
from transformers.file_utils import cached_property, is_torch_available, is_vision_available
from transformers.models.auto import get_values
from transformers.testing_utils import require_torch, require_vision, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING,
MODEL_MAPPING,
LevitForImageClassification,
LevitForImageClassificationWithTeacher,
LevitModel,
)
from transformers.models.levit.modeling_levit import LEVIT_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import LevitImageProcessor
class snake_case ( __UpperCAmelCase ):
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Optional[Any] = self.config_class(**self.inputs_dict )
self.parent.assertTrue(hasattr(_lowerCamelCase , '''hidden_sizes''' ) )
self.parent.assertTrue(hasattr(_lowerCamelCase , '''num_attention_heads''' ) )
class snake_case :
def __init__( self :List[Any] , _lowerCamelCase :str , _lowerCamelCase :Optional[int]=1_3 , _lowerCamelCase :Union[str, Any]=6_4 , _lowerCamelCase :Optional[Any]=3 , _lowerCamelCase :Any=3 , _lowerCamelCase :Tuple=2 , _lowerCamelCase :Tuple=1 , _lowerCamelCase :Union[str, Any]=1_6 , _lowerCamelCase :Any=[1_2_8, 2_5_6, 3_8_4] , _lowerCamelCase :Dict=[4, 6, 8] , _lowerCamelCase :List[str]=[2, 3, 4] , _lowerCamelCase :Union[str, Any]=[1_6, 1_6, 1_6] , _lowerCamelCase :Any=0 , _lowerCamelCase :Union[str, Any]=[2, 2, 2] , _lowerCamelCase :List[str]=[2, 2, 2] , _lowerCamelCase :str=0.0_2 , _lowerCamelCase :Dict=True , _lowerCamelCase :List[str]=True , _lowerCamelCase :int=2 , ):
__SCREAMING_SNAKE_CASE : Dict = parent
__SCREAMING_SNAKE_CASE : List[str] = batch_size
__SCREAMING_SNAKE_CASE : Tuple = image_size
__SCREAMING_SNAKE_CASE : Union[str, Any] = num_channels
__SCREAMING_SNAKE_CASE : Optional[Any] = kernel_size
__SCREAMING_SNAKE_CASE : Optional[int] = stride
__SCREAMING_SNAKE_CASE : List[str] = padding
__SCREAMING_SNAKE_CASE : Tuple = hidden_sizes
__SCREAMING_SNAKE_CASE : List[str] = num_attention_heads
__SCREAMING_SNAKE_CASE : Optional[Any] = depths
__SCREAMING_SNAKE_CASE : Optional[int] = key_dim
__SCREAMING_SNAKE_CASE : Optional[int] = drop_path_rate
__SCREAMING_SNAKE_CASE : int = patch_size
__SCREAMING_SNAKE_CASE : List[str] = attention_ratio
__SCREAMING_SNAKE_CASE : List[str] = mlp_ratio
__SCREAMING_SNAKE_CASE : int = initializer_range
__SCREAMING_SNAKE_CASE : Dict = [
['''Subsample''', key_dim[0], hidden_sizes[0] // key_dim[0], 4, 2, 2],
['''Subsample''', key_dim[0], hidden_sizes[1] // key_dim[0], 4, 2, 2],
]
__SCREAMING_SNAKE_CASE : List[Any] = is_training
__SCREAMING_SNAKE_CASE : Any = use_labels
__SCREAMING_SNAKE_CASE : Any = num_labels
__SCREAMING_SNAKE_CASE : List[str] = initializer_range
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : List[str] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
__SCREAMING_SNAKE_CASE : List[Any] = None
if self.use_labels:
__SCREAMING_SNAKE_CASE : Dict = ids_tensor([self.batch_size] , self.num_labels )
__SCREAMING_SNAKE_CASE : Optional[Any] = self.get_config()
return config, pixel_values, labels
def SCREAMING_SNAKE_CASE_ ( self :str ):
return LevitConfig(
image_size=self.image_size , num_channels=self.num_channels , kernel_size=self.kernel_size , stride=self.stride , padding=self.padding , patch_size=self.patch_size , hidden_sizes=self.hidden_sizes , num_attention_heads=self.num_attention_heads , depths=self.depths , key_dim=self.key_dim , drop_path_rate=self.drop_path_rate , mlp_ratio=self.mlp_ratio , attention_ratio=self.attention_ratio , initializer_range=self.initializer_range , down_ops=self.down_ops , )
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :str , _lowerCamelCase :str , _lowerCamelCase :int ):
__SCREAMING_SNAKE_CASE : Tuple = LevitModel(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : List[str] = model(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = (self.image_size, self.image_size)
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = image_size[0], image_size[1]
for _ in range(4 ):
__SCREAMING_SNAKE_CASE : str = floor(((height + 2 * self.padding - self.kernel_size) / self.stride) + 1 )
__SCREAMING_SNAKE_CASE : Tuple = floor(((width + 2 * self.padding - self.kernel_size) / self.stride) + 1 )
self.parent.assertEqual(
result.last_hidden_state.shape , (self.batch_size, ceil(height / 4 ) * ceil(width / 4 ), self.hidden_sizes[-1]) , )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] , _lowerCamelCase :Optional[int] , _lowerCamelCase :Tuple , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : Optional[Any] = self.num_labels
__SCREAMING_SNAKE_CASE : Dict = LevitForImageClassification(_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : Optional[Any] = model(_lowerCamelCase , labels=_lowerCamelCase )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE : Any = self.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = config_and_inputs
__SCREAMING_SNAKE_CASE : Optional[Any] = {'''pixel_values''': pixel_values}
return config, inputs_dict
@require_torch
class snake_case ( __UpperCAmelCase , __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = (
(LevitModel, LevitForImageClassification, LevitForImageClassificationWithTeacher)
if is_torch_available()
else ()
)
lowerCamelCase__ = (
{
'''feature-extraction''': LevitModel,
'''image-classification''': (LevitForImageClassification, LevitForImageClassificationWithTeacher),
}
if is_torch_available()
else {}
)
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = LevitModelTester(self )
__SCREAMING_SNAKE_CASE : int = ConfigTester(self , config_class=_lowerCamelCase , has_text_modality=_lowerCamelCase , hidden_size=3_7 )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
return
@unittest.skip(reason='''Levit does not use inputs_embeds''' )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
pass
@unittest.skip(reason='''Levit does not support input and output embeddings''' )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
pass
@unittest.skip(reason='''Levit does not output attentions''' )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
pass
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Optional[int] = model_class(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
__SCREAMING_SNAKE_CASE : int = [*signature.parameters.keys()]
__SCREAMING_SNAKE_CASE : Optional[int] = ['''pixel_values''']
self.assertListEqual(arg_names[:1] , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
def check_hidden_states_output(_lowerCamelCase :List[str] , _lowerCamelCase :List[str] , _lowerCamelCase :int ):
__SCREAMING_SNAKE_CASE : int = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
with torch.no_grad():
__SCREAMING_SNAKE_CASE : Any = model(**self._prepare_for_class(_lowerCamelCase , _lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : List[str] = outputs.hidden_states
__SCREAMING_SNAKE_CASE : List[Any] = len(self.model_tester.depths ) + 1
self.assertEqual(len(_lowerCamelCase ) , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = (self.model_tester.image_size, self.model_tester.image_size)
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : str = image_size[0], image_size[1]
for _ in range(4 ):
__SCREAMING_SNAKE_CASE : Any = floor(
(
(height + 2 * self.model_tester.padding - self.model_tester.kernel_size)
/ self.model_tester.stride
)
+ 1 )
__SCREAMING_SNAKE_CASE : int = floor(
(
(width + 2 * self.model_tester.padding - self.model_tester.kernel_size)
/ self.model_tester.stride
)
+ 1 )
# verify the first hidden states (first block)
self.assertListEqual(
list(hidden_states[0].shape[-2:] ) , [
height * width,
self.model_tester.hidden_sizes[0],
] , )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : List[str] = True
check_hidden_states_output(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
__SCREAMING_SNAKE_CASE : Optional[int] = True
check_hidden_states_output(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
@unittest.skip('''Will be fixed soon by reducing the size of the model used for common tests.''' )
def SCREAMING_SNAKE_CASE_ ( self :str ):
pass
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :str , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :Optional[Any]=False ):
__SCREAMING_SNAKE_CASE : int = super()._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
if return_labels:
if model_class.__name__ == "LevitForImageClassificationWithTeacher":
del inputs_dict["labels"]
return inputs_dict
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
if not self.model_tester.is_training:
return
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
__SCREAMING_SNAKE_CASE : Optional[Any] = True
for model_class in self.all_model_classes:
# LevitForImageClassificationWithTeacher supports inference-only
if (
model_class in get_values(_lowerCamelCase )
or model_class.__name__ == "LevitForImageClassificationWithTeacher"
):
continue
__SCREAMING_SNAKE_CASE : Optional[Any] = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.train()
__SCREAMING_SNAKE_CASE : Union[str, Any] = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = model(**_lowerCamelCase ).loss
loss.backward()
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
if not self.model_tester.is_training:
return
__SCREAMING_SNAKE_CASE : List[Any] = False
__SCREAMING_SNAKE_CASE : int = True
for model_class in self.all_model_classes:
if model_class in get_values(_lowerCamelCase ) or not model_class.supports_gradient_checkpointing:
continue
# LevitForImageClassificationWithTeacher supports inference-only
if model_class.__name__ == "LevitForImageClassificationWithTeacher":
continue
__SCREAMING_SNAKE_CASE : Union[str, Any] = model_class(_lowerCamelCase )
model.gradient_checkpointing_enable()
model.to(_lowerCamelCase )
model.train()
__SCREAMING_SNAKE_CASE : Any = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = model(**_lowerCamelCase ).loss
loss.backward()
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
__SCREAMING_SNAKE_CASE : int = [
{'''title''': '''multi_label_classification''', '''num_labels''': 2, '''dtype''': torch.float},
{'''title''': '''single_label_classification''', '''num_labels''': 1, '''dtype''': torch.long},
{'''title''': '''regression''', '''num_labels''': 1, '''dtype''': torch.float},
]
for model_class in self.all_model_classes:
if (
model_class
not in [
*get_values(_lowerCamelCase ),
]
or model_class.__name__ == "LevitForImageClassificationWithTeacher"
):
continue
for problem_type in problem_types:
with self.subTest(msg=f'''Testing {model_class} with {problem_type['title']}''' ):
__SCREAMING_SNAKE_CASE : str = problem_type['''title''']
__SCREAMING_SNAKE_CASE : List[Any] = problem_type['''num_labels''']
__SCREAMING_SNAKE_CASE : Optional[Any] = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.train()
__SCREAMING_SNAKE_CASE : Dict = self._prepare_for_class(_lowerCamelCase , _lowerCamelCase , return_labels=_lowerCamelCase )
if problem_type["num_labels"] > 1:
__SCREAMING_SNAKE_CASE : Optional[int] = inputs['''labels'''].unsqueeze(1 ).repeat(1 , problem_type['''num_labels'''] )
__SCREAMING_SNAKE_CASE : Optional[int] = inputs['''labels'''].to(problem_type['''dtype'''] )
# This tests that we do not trigger the warning form PyTorch "Using a target size that is different
# to the input size. This will likely lead to incorrect results due to broadcasting. Please ensure
# they have the same size." which is a symptom something in wrong for the regression problem.
# See https://github.com/huggingface/transformers/issues/11780
with warnings.catch_warnings(record=_lowerCamelCase ) as warning_list:
__SCREAMING_SNAKE_CASE : Union[str, Any] = model(**_lowerCamelCase ).loss
for w in warning_list:
if "Using a target size that is different to the input size" in str(w.message ):
raise ValueError(
f'''Something is going wrong in the regression problem: intercepted {w.message}''' )
loss.backward()
@slow
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
for model_name in LEVIT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Union[str, Any] = LevitModel.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
return image
@require_torch
@require_vision
class snake_case ( unittest.TestCase ):
@cached_property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
return LevitImageProcessor.from_pretrained(LEVIT_PRETRAINED_MODEL_ARCHIVE_LIST[0] )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : str = LevitForImageClassificationWithTeacher.from_pretrained(LEVIT_PRETRAINED_MODEL_ARCHIVE_LIST[0] ).to(
_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = self.default_image_processor
__SCREAMING_SNAKE_CASE : Any = prepare_img()
__SCREAMING_SNAKE_CASE : Union[str, Any] = image_processor(images=_lowerCamelCase , return_tensors='''pt''' ).to(_lowerCamelCase )
# forward pass
with torch.no_grad():
__SCREAMING_SNAKE_CASE : Union[str, Any] = model(**_lowerCamelCase )
# verify the logits
__SCREAMING_SNAKE_CASE : List[Any] = torch.Size((1, 1_0_0_0) )
self.assertEqual(outputs.logits.shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = torch.tensor([1.0_4_4_8, -0.3_7_4_5, -1.8_3_1_7] ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(outputs.logits[0, :3] , _lowerCamelCase , atol=1e-4 ) )
| 674 |
"""simple docstring"""
from __future__ import annotations
def lowerCAmelCase_ ( lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = str(lowercase_ )
return len(lowercase_ ) == 9 and set(lowercase_ ) == set('''123456789''' )
def lowerCAmelCase_ ( ):
'''simple docstring'''
for base_num in range(9999 , 4999 , -1 ):
__SCREAMING_SNAKE_CASE : List[str] = 10_0002 * base_num
if is_9_pandigital(lowercase_ ):
return candidate
for base_num in range(333 , 99 , -1 ):
__SCREAMING_SNAKE_CASE : List[Any] = 100_2003 * base_num
if is_9_pandigital(lowercase_ ):
return candidate
return None
if __name__ == "__main__":
print(f'{solution() = }')
| 674 | 1 |
"""simple docstring"""
import gc
import unittest
import numpy as np
import torch
from diffusers import (
AudioDiffusionPipeline,
AutoencoderKL,
DDIMScheduler,
DDPMScheduler,
DiffusionPipeline,
Mel,
UNetaDConditionModel,
UNetaDModel,
)
from diffusers.utils import slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
enable_full_determinism()
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :str ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
@property
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : List[str] = UNetaDModel(
sample_size=(3_2, 6_4) , in_channels=1 , out_channels=1 , layers_per_block=2 , block_out_channels=(1_2_8, 1_2_8) , down_block_types=('''AttnDownBlock2D''', '''DownBlock2D''') , up_block_types=('''UpBlock2D''', '''AttnUpBlock2D''') , )
return model
@property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : str = UNetaDConditionModel(
sample_size=(6_4, 3_2) , in_channels=1 , out_channels=1 , layers_per_block=2 , block_out_channels=(1_2_8, 1_2_8) , down_block_types=('''CrossAttnDownBlock2D''', '''DownBlock2D''') , up_block_types=('''UpBlock2D''', '''CrossAttnUpBlock2D''') , cross_attention_dim=1_0 , )
return model
@property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Tuple = AutoencoderKL(
sample_size=(1_2_8, 6_4) , in_channels=1 , out_channels=1 , latent_channels=1 , layers_per_block=2 , block_out_channels=(1_2_8, 1_2_8) , down_block_types=('''DownEncoderBlock2D''', '''DownEncoderBlock2D''') , up_block_types=('''UpDecoderBlock2D''', '''UpDecoderBlock2D''') , )
__SCREAMING_SNAKE_CASE : List[str] = UNetaDModel(
sample_size=(6_4, 3_2) , in_channels=1 , out_channels=1 , layers_per_block=2 , block_out_channels=(1_2_8, 1_2_8) , down_block_types=('''AttnDownBlock2D''', '''DownBlock2D''') , up_block_types=('''UpBlock2D''', '''AttnUpBlock2D''') , )
return vqvae, unet
@slow
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : str = '''cpu''' # ensure determinism for the device-dependent torch.Generator
__SCREAMING_SNAKE_CASE : Tuple = Mel(
x_res=self.dummy_unet.config.sample_size[1] , y_res=self.dummy_unet.config.sample_size[0] , )
__SCREAMING_SNAKE_CASE : List[str] = DDPMScheduler()
__SCREAMING_SNAKE_CASE : Optional[int] = AudioDiffusionPipeline(vqvae=_lowerCamelCase , unet=self.dummy_unet , mel=_lowerCamelCase , scheduler=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = torch.Generator(device=_lowerCamelCase ).manual_seed(4_2 )
__SCREAMING_SNAKE_CASE : Optional[int] = pipe(generator=_lowerCamelCase , steps=4 )
__SCREAMING_SNAKE_CASE : Tuple = output.audios[0]
__SCREAMING_SNAKE_CASE : List[Any] = output.images[0]
__SCREAMING_SNAKE_CASE : Optional[Any] = torch.Generator(device=_lowerCamelCase ).manual_seed(4_2 )
__SCREAMING_SNAKE_CASE : int = pipe(generator=_lowerCamelCase , steps=4 , return_dict=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = output[0][0]
assert audio.shape == (1, (self.dummy_unet.config.sample_size[1] - 1) * mel.hop_length)
assert (
image.height == self.dummy_unet.config.sample_size[0]
and image.width == self.dummy_unet.config.sample_size[1]
)
__SCREAMING_SNAKE_CASE : List[Any] = np.frombuffer(image.tobytes() , dtype='''uint8''' )[:1_0]
__SCREAMING_SNAKE_CASE : int = np.frombuffer(image_from_tuple.tobytes() , dtype='''uint8''' )[:1_0]
__SCREAMING_SNAKE_CASE : str = np.array([6_9, 2_5_5, 2_5_5, 2_5_5, 0, 0, 7_7, 1_8_1, 1_2, 1_2_7] )
assert np.abs(image_slice.flatten() - expected_slice ).max() == 0
assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() == 0
__SCREAMING_SNAKE_CASE : Dict = Mel(
x_res=self.dummy_vqvae_and_unet[0].config.sample_size[1] , y_res=self.dummy_vqvae_and_unet[0].config.sample_size[0] , )
__SCREAMING_SNAKE_CASE : List[Any] = DDIMScheduler()
__SCREAMING_SNAKE_CASE : str = self.dummy_vqvae_and_unet
__SCREAMING_SNAKE_CASE : List[str] = AudioDiffusionPipeline(
vqvae=self.dummy_vqvae_and_unet[0] , unet=dummy_vqvae_and_unet[1] , mel=_lowerCamelCase , scheduler=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
np.random.seed(0 )
__SCREAMING_SNAKE_CASE : int = np.random.uniform(-1 , 1 , ((dummy_vqvae_and_unet[0].config.sample_size[1] - 1) * mel.hop_length,) )
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.Generator(device=_lowerCamelCase ).manual_seed(4_2 )
__SCREAMING_SNAKE_CASE : Tuple = pipe(raw_audio=_lowerCamelCase , generator=_lowerCamelCase , start_step=5 , steps=1_0 )
__SCREAMING_SNAKE_CASE : Any = output.images[0]
assert (
image.height == self.dummy_vqvae_and_unet[0].config.sample_size[0]
and image.width == self.dummy_vqvae_and_unet[0].config.sample_size[1]
)
__SCREAMING_SNAKE_CASE : Any = np.frombuffer(image.tobytes() , dtype='''uint8''' )[:1_0]
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.array([1_2_0, 1_1_7, 1_1_0, 1_0_9, 1_3_8, 1_6_7, 1_3_8, 1_4_8, 1_3_2, 1_2_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() == 0
__SCREAMING_SNAKE_CASE : Optional[Any] = self.dummy_unet_condition
__SCREAMING_SNAKE_CASE : Dict = AudioDiffusionPipeline(
vqvae=self.dummy_vqvae_and_unet[0] , unet=_lowerCamelCase , mel=_lowerCamelCase , scheduler=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
np.random.seed(0 )
__SCREAMING_SNAKE_CASE : Optional[int] = torch.rand((1, 1, 1_0) )
__SCREAMING_SNAKE_CASE : Union[str, Any] = pipe(generator=_lowerCamelCase , encoding=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = output.images[0]
__SCREAMING_SNAKE_CASE : List[Any] = np.frombuffer(image.tobytes() , dtype='''uint8''' )[:1_0]
__SCREAMING_SNAKE_CASE : List[str] = np.array([1_0_7, 1_0_3, 1_2_0, 1_2_7, 1_4_2, 1_2_2, 1_1_3, 1_2_2, 9_7, 1_1_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() == 0
@slow
@require_torch_gpu
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : str = torch_device
__SCREAMING_SNAKE_CASE : Union[str, Any] = DiffusionPipeline.from_pretrained('''teticio/audio-diffusion-ddim-256''' )
__SCREAMING_SNAKE_CASE : int = pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = torch.Generator(device=_lowerCamelCase ).manual_seed(4_2 )
__SCREAMING_SNAKE_CASE : Optional[int] = pipe(generator=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = output.audios[0]
__SCREAMING_SNAKE_CASE : Union[str, Any] = output.images[0]
assert audio.shape == (1, (pipe.unet.config.sample_size[1] - 1) * pipe.mel.hop_length)
assert image.height == pipe.unet.config.sample_size[0] and image.width == pipe.unet.config.sample_size[1]
__SCREAMING_SNAKE_CASE : Dict = np.frombuffer(image.tobytes() , dtype='''uint8''' )[:1_0]
__SCREAMING_SNAKE_CASE : Any = np.array([1_5_1, 1_6_7, 1_5_4, 1_4_4, 1_2_2, 1_3_4, 1_2_1, 1_0_5, 7_0, 2_6] )
assert np.abs(image_slice.flatten() - expected_slice ).max() == 0
| 674 |
"""simple docstring"""
import json
import sys
import tempfile
import unittest
from pathlib import Path
import transformers
from transformers import (
CONFIG_MAPPING,
IMAGE_PROCESSOR_MAPPING,
AutoConfig,
AutoImageProcessor,
CLIPConfig,
CLIPImageProcessor,
)
from transformers.testing_utils import DUMMY_UNKNOWN_IDENTIFIER
sys.path.append(str(Path(__file__).parent.parent.parent.parent / '''utils'''))
from test_module.custom_configuration import CustomConfig # noqa E402
from test_module.custom_image_processing import CustomImageProcessor # noqa E402
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : Tuple = 0
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoImageProcessor.from_pretrained('''openai/clip-vit-base-patch32''' )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :int ):
with tempfile.TemporaryDirectory() as tmpdirname:
__SCREAMING_SNAKE_CASE : Optional[Any] = Path(_lowerCamelCase ) / '''preprocessor_config.json'''
__SCREAMING_SNAKE_CASE : Dict = Path(_lowerCamelCase ) / '''config.json'''
json.dump(
{'''image_processor_type''': '''CLIPImageProcessor''', '''processor_class''': '''CLIPProcessor'''} , open(_lowerCamelCase , '''w''' ) , )
json.dump({'''model_type''': '''clip'''} , open(_lowerCamelCase , '''w''' ) )
__SCREAMING_SNAKE_CASE : Dict = AutoImageProcessor.from_pretrained(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# Ensure we can load the image processor from the feature extractor config
with tempfile.TemporaryDirectory() as tmpdirname:
__SCREAMING_SNAKE_CASE : Tuple = Path(_lowerCamelCase ) / '''preprocessor_config.json'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = Path(_lowerCamelCase ) / '''config.json'''
json.dump(
{'''feature_extractor_type''': '''CLIPFeatureExtractor''', '''processor_class''': '''CLIPProcessor'''} , open(_lowerCamelCase , '''w''' ) , )
json.dump({'''model_type''': '''clip'''} , open(_lowerCamelCase , '''w''' ) )
__SCREAMING_SNAKE_CASE : int = AutoImageProcessor.from_pretrained(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :int ):
with tempfile.TemporaryDirectory() as tmpdirname:
__SCREAMING_SNAKE_CASE : str = CLIPConfig()
# Create a dummy config file with image_proceesor_type
__SCREAMING_SNAKE_CASE : Tuple = Path(_lowerCamelCase ) / '''preprocessor_config.json'''
__SCREAMING_SNAKE_CASE : Any = Path(_lowerCamelCase ) / '''config.json'''
json.dump(
{'''image_processor_type''': '''CLIPImageProcessor''', '''processor_class''': '''CLIPProcessor'''} , open(_lowerCamelCase , '''w''' ) , )
json.dump({'''model_type''': '''clip'''} , open(_lowerCamelCase , '''w''' ) )
# remove image_processor_type to make sure config.json alone is enough to load image processor locally
__SCREAMING_SNAKE_CASE : int = AutoImageProcessor.from_pretrained(_lowerCamelCase ).to_dict()
config_dict.pop('''image_processor_type''' )
__SCREAMING_SNAKE_CASE : Union[str, Any] = CLIPImageProcessor(**_lowerCamelCase )
# save in new folder
model_config.save_pretrained(_lowerCamelCase )
config.save_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoImageProcessor.from_pretrained(_lowerCamelCase )
# make sure private variable is not incorrectly saved
__SCREAMING_SNAKE_CASE : Tuple = json.loads(config.to_json_string() )
self.assertTrue('''_processor_class''' not in dict_as_saved )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
with tempfile.TemporaryDirectory() as tmpdirname:
__SCREAMING_SNAKE_CASE : Dict = Path(_lowerCamelCase ) / '''preprocessor_config.json'''
json.dump(
{'''image_processor_type''': '''CLIPImageProcessor''', '''processor_class''': '''CLIPProcessor'''} , open(_lowerCamelCase , '''w''' ) , )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoImageProcessor.from_pretrained(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
with self.assertRaisesRegex(
_lowerCamelCase , '''clip-base is not a local folder and is not a valid model identifier''' ):
__SCREAMING_SNAKE_CASE : Tuple = AutoImageProcessor.from_pretrained('''clip-base''' )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
with self.assertRaisesRegex(
_lowerCamelCase , r'''aaaaaa is not a valid git identifier \(branch name, tag name or commit id\)''' ):
__SCREAMING_SNAKE_CASE : Optional[int] = AutoImageProcessor.from_pretrained(_lowerCamelCase , revision='''aaaaaa''' )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
with self.assertRaisesRegex(
_lowerCamelCase , '''hf-internal-testing/config-no-model does not appear to have a file named preprocessor_config.json.''' , ):
__SCREAMING_SNAKE_CASE : Optional[int] = AutoImageProcessor.from_pretrained('''hf-internal-testing/config-no-model''' )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
# If remote code is not set, we will time out when asking whether to load the model.
with self.assertRaises(_lowerCamelCase ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = AutoImageProcessor.from_pretrained('''hf-internal-testing/test_dynamic_image_processor''' )
# If remote code is disabled, we can't load this config.
with self.assertRaises(_lowerCamelCase ):
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoImageProcessor.from_pretrained(
'''hf-internal-testing/test_dynamic_image_processor''' , trust_remote_code=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = AutoImageProcessor.from_pretrained(
'''hf-internal-testing/test_dynamic_image_processor''' , trust_remote_code=_lowerCamelCase )
self.assertEqual(image_processor.__class__.__name__ , '''NewImageProcessor''' )
# Test image processor can be reloaded.
with tempfile.TemporaryDirectory() as tmp_dir:
image_processor.save_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoImageProcessor.from_pretrained(_lowerCamelCase , trust_remote_code=_lowerCamelCase )
self.assertEqual(reloaded_image_processor.__class__.__name__ , '''NewImageProcessor''' )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
try:
AutoConfig.register('''custom''' , _lowerCamelCase )
AutoImageProcessor.register(_lowerCamelCase , _lowerCamelCase )
# Trying to register something existing in the Transformers library will raise an error
with self.assertRaises(_lowerCamelCase ):
AutoImageProcessor.register(_lowerCamelCase , _lowerCamelCase )
with tempfile.TemporaryDirectory() as tmpdirname:
__SCREAMING_SNAKE_CASE : Dict = Path(_lowerCamelCase ) / '''preprocessor_config.json'''
__SCREAMING_SNAKE_CASE : Dict = Path(_lowerCamelCase ) / '''config.json'''
json.dump(
{'''feature_extractor_type''': '''CLIPFeatureExtractor''', '''processor_class''': '''CLIPProcessor'''} , open(_lowerCamelCase , '''w''' ) , )
json.dump({'''model_type''': '''clip'''} , open(_lowerCamelCase , '''w''' ) )
__SCREAMING_SNAKE_CASE : Optional[int] = CustomImageProcessor.from_pretrained(_lowerCamelCase )
# Now that the config is registered, it can be used as any other config with the auto-API
with tempfile.TemporaryDirectory() as tmp_dir:
image_processor.save_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoImageProcessor.from_pretrained(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
finally:
if "custom" in CONFIG_MAPPING._extra_content:
del CONFIG_MAPPING._extra_content["custom"]
if CustomConfig in IMAGE_PROCESSOR_MAPPING._extra_content:
del IMAGE_PROCESSOR_MAPPING._extra_content[CustomConfig]
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = True
try:
AutoConfig.register('''custom''' , _lowerCamelCase )
AutoImageProcessor.register(_lowerCamelCase , _lowerCamelCase )
# If remote code is not set, the default is to use local
__SCREAMING_SNAKE_CASE : List[str] = AutoImageProcessor.from_pretrained('''hf-internal-testing/test_dynamic_image_processor''' )
self.assertEqual(image_processor.__class__.__name__ , '''NewImageProcessor''' )
self.assertTrue(image_processor.is_local )
# If remote code is disabled, we load the local one.
__SCREAMING_SNAKE_CASE : str = AutoImageProcessor.from_pretrained(
'''hf-internal-testing/test_dynamic_image_processor''' , trust_remote_code=_lowerCamelCase )
self.assertEqual(image_processor.__class__.__name__ , '''NewImageProcessor''' )
self.assertTrue(image_processor.is_local )
# If remote is enabled, we load from the Hub
__SCREAMING_SNAKE_CASE : Optional[int] = AutoImageProcessor.from_pretrained(
'''hf-internal-testing/test_dynamic_image_processor''' , trust_remote_code=_lowerCamelCase )
self.assertEqual(image_processor.__class__.__name__ , '''NewImageProcessor''' )
self.assertTrue(not hasattr(_lowerCamelCase , '''is_local''' ) )
finally:
if "custom" in CONFIG_MAPPING._extra_content:
del CONFIG_MAPPING._extra_content["custom"]
if CustomConfig in IMAGE_PROCESSOR_MAPPING._extra_content:
del IMAGE_PROCESSOR_MAPPING._extra_content[CustomConfig]
| 674 | 1 |
"""simple docstring"""
import unittest
import numpy as np
import torch
from diffusers import ScoreSdeVePipeline, ScoreSdeVeScheduler, UNetaDModel
from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device
enable_full_determinism()
class snake_case ( unittest.TestCase ):
@property
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Optional[Any] = UNetaDModel(
block_out_channels=(3_2, 6_4) , layers_per_block=2 , sample_size=3_2 , in_channels=3 , out_channels=3 , down_block_types=('''DownBlock2D''', '''AttnDownBlock2D''') , up_block_types=('''AttnUpBlock2D''', '''UpBlock2D''') , )
return model
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : int = self.dummy_uncond_unet
__SCREAMING_SNAKE_CASE : str = ScoreSdeVeScheduler()
__SCREAMING_SNAKE_CASE : str = ScoreSdeVePipeline(unet=_lowerCamelCase , scheduler=_lowerCamelCase )
sde_ve.to(_lowerCamelCase )
sde_ve.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : int = sde_ve(num_inference_steps=2 , output_type='''numpy''' , generator=_lowerCamelCase ).images
__SCREAMING_SNAKE_CASE : int = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : str = sde_ve(num_inference_steps=2 , output_type='''numpy''' , generator=_lowerCamelCase , return_dict=_lowerCamelCase )[
0
]
__SCREAMING_SNAKE_CASE : Dict = image[0, -3:, -3:, -1]
__SCREAMING_SNAKE_CASE : Any = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 3_2, 3_2, 3)
__SCREAMING_SNAKE_CASE : Any = np.array([0.0, 1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1e-2
@slow
@require_torch
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : List[str] = '''google/ncsnpp-church-256'''
__SCREAMING_SNAKE_CASE : Optional[Any] = UNetaDModel.from_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = ScoreSdeVeScheduler.from_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = ScoreSdeVePipeline(unet=_lowerCamelCase , scheduler=_lowerCamelCase )
sde_ve.to(_lowerCamelCase )
sde_ve.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Dict = sde_ve(num_inference_steps=1_0 , output_type='''numpy''' , generator=_lowerCamelCase ).images
__SCREAMING_SNAKE_CASE : List[str] = image[0, -3:, -3:, -1]
assert image.shape == (1, 2_5_6, 2_5_6, 3)
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.array([0.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
| 674 |
"""simple docstring"""
from __future__ import annotations
from typing import Any
class snake_case ( __UpperCAmelCase ):
pass
class snake_case :
def __init__( self :List[Any] , _lowerCamelCase :Any ):
__SCREAMING_SNAKE_CASE : Any = data
__SCREAMING_SNAKE_CASE : Node | None = None
def __iter__( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : List[str] = self
__SCREAMING_SNAKE_CASE : List[str] = []
while node:
if node in visited:
raise ContainsLoopError
visited.append(_lowerCamelCase )
yield node.data
__SCREAMING_SNAKE_CASE : List[str] = node.next_node
@property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
try:
list(self )
return False
except ContainsLoopError:
return True
if __name__ == "__main__":
_lowerCamelCase = Node(1)
_lowerCamelCase = Node(2)
_lowerCamelCase = Node(3)
_lowerCamelCase = Node(4)
print(root_node.has_loop) # False
_lowerCamelCase = root_node.next_node
print(root_node.has_loop) # True
_lowerCamelCase = Node(5)
_lowerCamelCase = Node(6)
_lowerCamelCase = Node(5)
_lowerCamelCase = Node(6)
print(root_node.has_loop) # False
_lowerCamelCase = Node(1)
print(root_node.has_loop) # False
| 674 | 1 |
"""simple docstring"""
import os
def lowerCAmelCase_ ( lowercase_ : str = "matrix.txt" ):
'''simple docstring'''
with open(os.path.join(os.path.dirname(lowercase_ ) , lowercase_ ) ) as in_file:
__SCREAMING_SNAKE_CASE : List[Any] = in_file.read()
__SCREAMING_SNAKE_CASE : Any = [[int(lowercase_ ) for cell in row.split(''',''' )] for row in data.strip().splitlines()]
__SCREAMING_SNAKE_CASE : str = [[0 for cell in row] for row in grid]
__SCREAMING_SNAKE_CASE : List[Any] = len(grid[0] )
__SCREAMING_SNAKE_CASE : Optional[Any] = [[0 for i in range(lowercase_ )] for j in range(lowercase_ )]
__SCREAMING_SNAKE_CASE : Any = grid[0][0]
for i in range(1 , lowercase_ ):
__SCREAMING_SNAKE_CASE : Dict = grid[0][i] + dp[0][i - 1]
for i in range(1 , lowercase_ ):
__SCREAMING_SNAKE_CASE : List[str] = grid[i][0] + dp[i - 1][0]
for i in range(1 , lowercase_ ):
for j in range(1 , lowercase_ ):
__SCREAMING_SNAKE_CASE : Optional[int] = grid[i][j] + min(dp[i - 1][j] , dp[i][j - 1] )
return dp[-1][-1]
if __name__ == "__main__":
print(f'{solution() = }')
| 674 |
"""simple docstring"""
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = {
'''weiweishi/roc-bert-base-zh''': '''https://huggingface.co/weiweishi/roc-bert-base-zh/resolve/main/config.json''',
}
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = '''roc_bert'''
def __init__( self :Union[str, Any] , _lowerCamelCase :Any=3_0_5_2_2 , _lowerCamelCase :str=7_6_8 , _lowerCamelCase :Optional[Any]=1_2 , _lowerCamelCase :List[str]=1_2 , _lowerCamelCase :str=3_0_7_2 , _lowerCamelCase :Tuple="gelu" , _lowerCamelCase :List[Any]=0.1 , _lowerCamelCase :List[str]=0.1 , _lowerCamelCase :Optional[int]=5_1_2 , _lowerCamelCase :Dict=2 , _lowerCamelCase :Any=0.0_2 , _lowerCamelCase :Optional[int]=1e-12 , _lowerCamelCase :str=True , _lowerCamelCase :Any=0 , _lowerCamelCase :List[str]="absolute" , _lowerCamelCase :List[Any]=None , _lowerCamelCase :Any=True , _lowerCamelCase :Union[str, Any]=True , _lowerCamelCase :str=7_6_8 , _lowerCamelCase :Union[str, Any]=9_1_0 , _lowerCamelCase :List[Any]=5_1_2 , _lowerCamelCase :Optional[int]=2_4_8_5_8 , _lowerCamelCase :Union[str, Any]=True , **_lowerCamelCase :str , ):
__SCREAMING_SNAKE_CASE : List[str] = vocab_size
__SCREAMING_SNAKE_CASE : int = max_position_embeddings
__SCREAMING_SNAKE_CASE : List[str] = hidden_size
__SCREAMING_SNAKE_CASE : str = num_hidden_layers
__SCREAMING_SNAKE_CASE : int = num_attention_heads
__SCREAMING_SNAKE_CASE : Any = intermediate_size
__SCREAMING_SNAKE_CASE : Optional[int] = hidden_act
__SCREAMING_SNAKE_CASE : List[Any] = hidden_dropout_prob
__SCREAMING_SNAKE_CASE : Optional[Any] = attention_probs_dropout_prob
__SCREAMING_SNAKE_CASE : Union[str, Any] = initializer_range
__SCREAMING_SNAKE_CASE : Union[str, Any] = type_vocab_size
__SCREAMING_SNAKE_CASE : List[str] = layer_norm_eps
__SCREAMING_SNAKE_CASE : Optional[int] = use_cache
__SCREAMING_SNAKE_CASE : str = enable_pronunciation
__SCREAMING_SNAKE_CASE : List[str] = enable_shape
__SCREAMING_SNAKE_CASE : Tuple = pronunciation_embed_dim
__SCREAMING_SNAKE_CASE : Optional[Any] = pronunciation_vocab_size
__SCREAMING_SNAKE_CASE : str = shape_embed_dim
__SCREAMING_SNAKE_CASE : Union[str, Any] = shape_vocab_size
__SCREAMING_SNAKE_CASE : Tuple = concat_input
__SCREAMING_SNAKE_CASE : Union[str, Any] = position_embedding_type
__SCREAMING_SNAKE_CASE : str = classifier_dropout
super().__init__(pad_token_id=_lowerCamelCase , **_lowerCamelCase )
| 674 | 1 |
"""simple docstring"""
from __future__ import annotations
from math import pi, sqrt
def lowerCAmelCase_ ( lowercase_ : float , lowercase_ : float ):
'''simple docstring'''
if inductance <= 0:
raise ValueError('''Inductance cannot be 0 or negative''' )
elif capacitance <= 0:
raise ValueError('''Capacitance cannot be 0 or negative''' )
else:
return (
"Resonant frequency",
float(1 / (2 * pi * (sqrt(inductance * capacitance ))) ),
)
if __name__ == "__main__":
import doctest
doctest.testmod()
| 674 |
"""simple docstring"""
import itertools
import json
import linecache
import os
import pickle
import re
import socket
import string
from collections import Counter
from logging import getLogger
from pathlib import Path
from typing import Callable, Dict, Iterable, List
import git
import torch
from torch.utils.data import Dataset
from transformers import BartTokenizer, RagTokenizer, TaTokenizer
def lowerCAmelCase_ ( lowercase_ : List[str] , lowercase_ : List[str] , lowercase_ : Dict , lowercase_ : Tuple , lowercase_ : Optional[Any]=True , lowercase_ : Any="pt" ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = {'''add_prefix_space''': True} if isinstance(lowercase_ , lowercase_ ) and not line.startswith(''' ''' ) else {}
__SCREAMING_SNAKE_CASE : Optional[int] = padding_side
return tokenizer(
[line] , max_length=lowercase_ , padding='''max_length''' if pad_to_max_length else None , truncation=lowercase_ , return_tensors=lowercase_ , add_special_tokens=lowercase_ , **lowercase_ , )
def lowerCAmelCase_ ( lowercase_ : Optional[int] , lowercase_ : Tuple , lowercase_ : List[Any]=None , ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = input_ids.ne(lowercase_ ).any(dim=0 )
if attention_mask is None:
return input_ids[:, keep_column_mask]
else:
return (input_ids[:, keep_column_mask], attention_mask[:, keep_column_mask])
class snake_case ( __UpperCAmelCase ):
def __init__( self :Optional[Any] , _lowerCamelCase :Dict , _lowerCamelCase :Any , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Tuple , _lowerCamelCase :Any="train" , _lowerCamelCase :str=None , _lowerCamelCase :Optional[Any]=None , _lowerCamelCase :List[Any]=None , _lowerCamelCase :Tuple="" , ):
super().__init__()
__SCREAMING_SNAKE_CASE : Dict = Path(_lowerCamelCase ).joinpath(type_path + '''.source''' )
__SCREAMING_SNAKE_CASE : Any = Path(_lowerCamelCase ).joinpath(type_path + '''.target''' )
__SCREAMING_SNAKE_CASE : Any = self.get_char_lens(self.src_file )
__SCREAMING_SNAKE_CASE : List[str] = max_source_length
__SCREAMING_SNAKE_CASE : Dict = max_target_length
assert min(self.src_lens ) > 0, f'''found empty line in {self.src_file}'''
__SCREAMING_SNAKE_CASE : Dict = tokenizer
__SCREAMING_SNAKE_CASE : Union[str, Any] = prefix
if n_obs is not None:
__SCREAMING_SNAKE_CASE : Any = self.src_lens[:n_obs]
__SCREAMING_SNAKE_CASE : List[str] = src_lang
__SCREAMING_SNAKE_CASE : str = tgt_lang
def __len__( self :int ):
return len(self.src_lens )
def __getitem__( self :Optional[Any] , _lowerCamelCase :Any ):
__SCREAMING_SNAKE_CASE : Optional[Any] = index + 1 # linecache starts at 1
__SCREAMING_SNAKE_CASE : Any = self.prefix + linecache.getline(str(self.src_file ) , _lowerCamelCase ).rstrip('''\n''' )
__SCREAMING_SNAKE_CASE : Dict = linecache.getline(str(self.tgt_file ) , _lowerCamelCase ).rstrip('''\n''' )
assert source_line, f'''empty source line for index {index}'''
assert tgt_line, f'''empty tgt line for index {index}'''
# Need to add eos token manually for T5
if isinstance(self.tokenizer , _lowerCamelCase ):
source_line += self.tokenizer.eos_token
tgt_line += self.tokenizer.eos_token
# Pad source and target to the right
__SCREAMING_SNAKE_CASE : Dict = (
self.tokenizer.question_encoder if isinstance(self.tokenizer , _lowerCamelCase ) else self.tokenizer
)
__SCREAMING_SNAKE_CASE : Optional[Any] = self.tokenizer.generator if isinstance(self.tokenizer , _lowerCamelCase ) else self.tokenizer
__SCREAMING_SNAKE_CASE : Dict = encode_line(_lowerCamelCase , _lowerCamelCase , self.max_source_length , '''right''' )
__SCREAMING_SNAKE_CASE : Dict = encode_line(_lowerCamelCase , _lowerCamelCase , self.max_target_length , '''right''' )
__SCREAMING_SNAKE_CASE : Any = source_inputs['''input_ids'''].squeeze()
__SCREAMING_SNAKE_CASE : Any = target_inputs['''input_ids'''].squeeze()
__SCREAMING_SNAKE_CASE : Dict = source_inputs['''attention_mask'''].squeeze()
return {
"input_ids": source_ids,
"attention_mask": src_mask,
"decoder_input_ids": target_ids,
}
@staticmethod
def SCREAMING_SNAKE_CASE_ ( _lowerCamelCase :Any ):
return [len(_lowerCamelCase ) for x in Path(_lowerCamelCase ).open().readlines()]
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :List[str] ):
__SCREAMING_SNAKE_CASE : int = torch.stack([x['''input_ids'''] for x in batch] )
__SCREAMING_SNAKE_CASE : str = torch.stack([x['''attention_mask'''] for x in batch] )
__SCREAMING_SNAKE_CASE : int = torch.stack([x['''decoder_input_ids'''] for x in batch] )
__SCREAMING_SNAKE_CASE : str = (
self.tokenizer.generator.pad_token_id
if isinstance(self.tokenizer , _lowerCamelCase )
else self.tokenizer.pad_token_id
)
__SCREAMING_SNAKE_CASE : Union[str, Any] = (
self.tokenizer.question_encoder.pad_token_id
if isinstance(self.tokenizer , _lowerCamelCase )
else self.tokenizer.pad_token_id
)
__SCREAMING_SNAKE_CASE : List[str] = trim_batch(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = trim_batch(_lowerCamelCase , _lowerCamelCase , attention_mask=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = {
'''input_ids''': source_ids,
'''attention_mask''': source_mask,
'''decoder_input_ids''': y,
}
return batch
_lowerCamelCase = getLogger(__name__)
def lowerCAmelCase_ ( lowercase_ : List[List] ):
'''simple docstring'''
return list(itertools.chain.from_iterable(lowercase_ ) )
def lowerCAmelCase_ ( lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = get_git_info()
save_json(lowercase_ , os.path.join(lowercase_ , '''git_log.json''' ) )
def lowerCAmelCase_ ( lowercase_ : Any , lowercase_ : Optional[int] , lowercase_ : str=4 , **lowercase_ : List[str] ):
'''simple docstring'''
with open(lowercase_ , '''w''' ) as f:
json.dump(lowercase_ , lowercase_ , indent=lowercase_ , **lowercase_ )
def lowerCAmelCase_ ( lowercase_ : Union[str, Any] ):
'''simple docstring'''
with open(lowercase_ ) as f:
return json.load(lowercase_ )
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = git.Repo(search_parent_directories=lowercase_ )
__SCREAMING_SNAKE_CASE : List[str] = {
'''repo_id''': str(lowercase_ ),
'''repo_sha''': str(repo.head.object.hexsha ),
'''repo_branch''': str(repo.active_branch ),
'''hostname''': str(socket.gethostname() ),
}
return repo_infos
def lowerCAmelCase_ ( lowercase_ : Callable , lowercase_ : Iterable ):
'''simple docstring'''
return list(map(lowercase_ , lowercase_ ) )
def lowerCAmelCase_ ( lowercase_ : Any , lowercase_ : Any ):
'''simple docstring'''
with open(lowercase_ , '''wb''' ) as f:
return pickle.dump(lowercase_ , lowercase_ )
def lowerCAmelCase_ ( lowercase_ : Any ):
'''simple docstring'''
def remove_articles(lowercase_ : Dict ):
return re.sub(r'''\b(a|an|the)\b''' , ''' ''' , lowercase_ )
def white_space_fix(lowercase_ : Optional[int] ):
return " ".join(text.split() )
def remove_punc(lowercase_ : Any ):
__SCREAMING_SNAKE_CASE : Optional[int] = set(string.punctuation )
return "".join(ch for ch in text if ch not in exclude )
def lower(lowercase_ : int ):
return text.lower()
return white_space_fix(remove_articles(remove_punc(lower(lowercase_ ) ) ) )
def lowerCAmelCase_ ( lowercase_ : Optional[int] , lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = normalize_answer(lowercase_ ).split()
__SCREAMING_SNAKE_CASE : Any = normalize_answer(lowercase_ ).split()
__SCREAMING_SNAKE_CASE : Tuple = Counter(lowercase_ ) & Counter(lowercase_ )
__SCREAMING_SNAKE_CASE : Tuple = sum(common.values() )
if num_same == 0:
return 0
__SCREAMING_SNAKE_CASE : Any = 1.0 * num_same / len(lowercase_ )
__SCREAMING_SNAKE_CASE : List[str] = 1.0 * num_same / len(lowercase_ )
__SCREAMING_SNAKE_CASE : Optional[int] = (2 * precision * recall) / (precision + recall)
return fa
def lowerCAmelCase_ ( lowercase_ : str , lowercase_ : Union[str, Any] ):
'''simple docstring'''
return normalize_answer(lowercase_ ) == normalize_answer(lowercase_ )
def lowerCAmelCase_ ( lowercase_ : List[str] , lowercase_ : List[str] ):
'''simple docstring'''
assert len(lowercase_ ) == len(lowercase_ )
__SCREAMING_SNAKE_CASE : Union[str, Any] = 0
for hypo, pred in zip(lowercase_ , lowercase_ ):
em += exact_match_score(lowercase_ , lowercase_ )
if len(lowercase_ ) > 0:
em /= len(lowercase_ )
return {"em": em}
def lowerCAmelCase_ ( lowercase_ : str ):
'''simple docstring'''
return model_prefix.startswith('''rag''' )
def lowerCAmelCase_ ( lowercase_ : Optional[Any] , lowercase_ : Tuple , lowercase_ : Optional[int] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = {p: p for p in extra_params}
# T5 models don't have `dropout` param, they have `dropout_rate` instead
__SCREAMING_SNAKE_CASE : Any = '''dropout_rate'''
for p in extra_params:
if getattr(lowercase_ , lowercase_ , lowercase_ ):
if not hasattr(lowercase_ , lowercase_ ) and not hasattr(lowercase_ , equivalent_param[p] ):
logger.info('''config doesn\'t have a `{}` attribute'''.format(lowercase_ ) )
delattr(lowercase_ , lowercase_ )
continue
__SCREAMING_SNAKE_CASE : Optional[int] = p if hasattr(lowercase_ , lowercase_ ) else equivalent_param[p]
setattr(lowercase_ , lowercase_ , getattr(lowercase_ , lowercase_ ) )
delattr(lowercase_ , lowercase_ )
return hparams, config
| 674 | 1 |
"""simple docstring"""
import gc
import random
import unittest
import numpy as np
import torch
from PIL import Image
from diffusers import (
DDIMScheduler,
KandinskyVaaControlnetImgaImgPipeline,
KandinskyVaaPriorEmbaEmbPipeline,
UNetaDConditionModel,
VQModel,
)
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..test_pipelines_common import PipelineTesterMixin, assert_mean_pixel_difference
enable_full_determinism()
class snake_case ( __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = KandinskyVaaControlnetImgaImgPipeline
lowerCamelCase__ = ['''image_embeds''', '''negative_image_embeds''', '''image''', '''hint''']
lowerCamelCase__ = ['''image_embeds''', '''negative_image_embeds''', '''image''', '''hint''']
lowerCamelCase__ = [
'''generator''',
'''height''',
'''width''',
'''strength''',
'''guidance_scale''',
'''num_inference_steps''',
'''return_dict''',
'''guidance_scale''',
'''num_images_per_prompt''',
'''output_type''',
'''return_dict''',
]
lowerCamelCase__ = False
@property
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
return 3_2
@property
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
return 3_2
@property
def SCREAMING_SNAKE_CASE_ ( self :int ):
return self.time_input_dim
@property
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
return self.time_input_dim * 4
@property
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
return 1_0_0
@property
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : int = {
'''in_channels''': 8,
# Out channels is double in channels because predicts mean and variance
'''out_channels''': 8,
'''addition_embed_type''': '''image_hint''',
'''down_block_types''': ('''ResnetDownsampleBlock2D''', '''SimpleCrossAttnDownBlock2D'''),
'''up_block_types''': ('''SimpleCrossAttnUpBlock2D''', '''ResnetUpsampleBlock2D'''),
'''mid_block_type''': '''UNetMidBlock2DSimpleCrossAttn''',
'''block_out_channels''': (self.block_out_channels_a, self.block_out_channels_a * 2),
'''layers_per_block''': 1,
'''encoder_hid_dim''': self.text_embedder_hidden_size,
'''encoder_hid_dim_type''': '''image_proj''',
'''cross_attention_dim''': self.cross_attention_dim,
'''attention_head_dim''': 4,
'''resnet_time_scale_shift''': '''scale_shift''',
'''class_embed_type''': None,
}
__SCREAMING_SNAKE_CASE : int = UNetaDConditionModel(**_lowerCamelCase )
return model
@property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
return {
"block_out_channels": [3_2, 3_2, 6_4, 6_4],
"down_block_types": [
"DownEncoderBlock2D",
"DownEncoderBlock2D",
"DownEncoderBlock2D",
"AttnDownEncoderBlock2D",
],
"in_channels": 3,
"latent_channels": 4,
"layers_per_block": 1,
"norm_num_groups": 8,
"norm_type": "spatial",
"num_vq_embeddings": 1_2,
"out_channels": 3,
"up_block_types": ["AttnUpDecoderBlock2D", "UpDecoderBlock2D", "UpDecoderBlock2D", "UpDecoderBlock2D"],
"vq_embed_dim": 4,
}
@property
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Any = VQModel(**self.dummy_movq_kwargs )
return model
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : int = self.dummy_unet
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.dummy_movq
__SCREAMING_SNAKE_CASE : Optional[Any] = {
'''num_train_timesteps''': 1_0_0_0,
'''beta_schedule''': '''linear''',
'''beta_start''': 0.0_0_0_8_5,
'''beta_end''': 0.0_1_2,
'''clip_sample''': False,
'''set_alpha_to_one''': False,
'''steps_offset''': 0,
'''prediction_type''': '''epsilon''',
'''thresholding''': False,
}
__SCREAMING_SNAKE_CASE : Optional[Any] = DDIMScheduler(**_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = {
'''unet''': unet,
'''scheduler''': scheduler,
'''movq''': movq,
}
return components
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] , _lowerCamelCase :Any , _lowerCamelCase :Optional[Any]=0 ):
__SCREAMING_SNAKE_CASE : Any = floats_tensor((1, self.text_embedder_hidden_size) , rng=random.Random(_lowerCamelCase ) ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = floats_tensor((1, self.text_embedder_hidden_size) , rng=random.Random(seed + 1 ) ).to(
_lowerCamelCase )
# create init_image
__SCREAMING_SNAKE_CASE : Optional[Any] = floats_tensor((1, 3, 6_4, 6_4) , rng=random.Random(_lowerCamelCase ) ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = image.cpu().permute(0 , 2 , 3 , 1 )[0]
__SCREAMING_SNAKE_CASE : Optional[int] = Image.fromarray(np.uinta(_lowerCamelCase ) ).convert('''RGB''' ).resize((2_5_6, 2_5_6) )
# create hint
__SCREAMING_SNAKE_CASE : Tuple = floats_tensor((1, 3, 6_4, 6_4) , rng=random.Random(_lowerCamelCase ) ).to(_lowerCamelCase )
if str(_lowerCamelCase ).startswith('''mps''' ):
__SCREAMING_SNAKE_CASE : str = torch.manual_seed(_lowerCamelCase )
else:
__SCREAMING_SNAKE_CASE : Tuple = torch.Generator(device=_lowerCamelCase ).manual_seed(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = {
'''image''': init_image,
'''image_embeds''': image_embeds,
'''negative_image_embeds''': negative_image_embeds,
'''hint''': hint,
'''generator''': generator,
'''height''': 6_4,
'''width''': 6_4,
'''num_inference_steps''': 1_0,
'''guidance_scale''': 7.0,
'''strength''': 0.2,
'''output_type''': '''np''',
}
return inputs
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : int = '''cpu'''
__SCREAMING_SNAKE_CASE : str = self.get_dummy_components()
__SCREAMING_SNAKE_CASE : Optional[int] = self.pipeline_class(**_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = pipe(**self.get_dummy_inputs(_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : List[Any] = output.images
__SCREAMING_SNAKE_CASE : Dict = pipe(
**self.get_dummy_inputs(_lowerCamelCase ) , return_dict=_lowerCamelCase , )[0]
__SCREAMING_SNAKE_CASE : Dict = image[0, -3:, -3:, -1]
__SCREAMING_SNAKE_CASE : List[Any] = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 6_4, 6_4, 3)
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.array(
[0.5_4_9_8_5_0_3_4, 0.5_5_5_0_9_3_6_5, 0.5_2_5_6_1_5_0_4, 0.5_5_7_0_4_9_4, 0.5_5_9_3_8_1_8, 0.5_2_6_3_9_7_9, 0.5_0_2_8_5_6_4_3, 0.5_0_6_9_8_4_6, 0.5_1_1_9_6_7_3_6] )
assert (
np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
), f''' expected_slice {expected_slice}, but got {image_slice.flatten()}'''
assert (
np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1e-2
), f''' expected_slice {expected_slice}, but got {image_from_tuple_slice.flatten()}'''
@slow
@require_torch_gpu
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/kandinskyv22/kandinskyv22_controlnet_img2img_robotcat_fp16.npy''' )
__SCREAMING_SNAKE_CASE : str = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main''' '''/kandinsky/cat.png''' )
__SCREAMING_SNAKE_CASE : Any = init_image.resize((5_1_2, 5_1_2) )
__SCREAMING_SNAKE_CASE : str = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/kandinskyv22/hint_image_cat.png''' )
__SCREAMING_SNAKE_CASE : Tuple = torch.from_numpy(np.array(_lowerCamelCase ) ).float() / 2_5_5.0
__SCREAMING_SNAKE_CASE : Any = hint.permute(2 , 0 , 1 ).unsqueeze(0 )
__SCREAMING_SNAKE_CASE : Dict = '''A robot, 4k photo'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = KandinskyVaaPriorEmbaEmbPipeline.from_pretrained(
'''kandinsky-community/kandinsky-2-2-prior''' , torch_dtype=torch.floataa )
pipe_prior.to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = KandinskyVaaControlnetImgaImgPipeline.from_pretrained(
'''kandinsky-community/kandinsky-2-2-controlnet-depth''' , torch_dtype=torch.floataa )
__SCREAMING_SNAKE_CASE : Union[str, Any] = pipeline.to(_lowerCamelCase )
pipeline.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = torch.Generator(device='''cpu''' ).manual_seed(0 )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[str] = pipe_prior(
_lowerCamelCase , image=_lowerCamelCase , strength=0.8_5 , generator=_lowerCamelCase , negative_prompt='''''' , ).to_tuple()
__SCREAMING_SNAKE_CASE : Union[str, Any] = pipeline(
image=_lowerCamelCase , image_embeds=_lowerCamelCase , negative_image_embeds=_lowerCamelCase , hint=_lowerCamelCase , generator=_lowerCamelCase , num_inference_steps=1_0_0 , height=5_1_2 , width=5_1_2 , strength=0.5 , output_type='''np''' , )
__SCREAMING_SNAKE_CASE : Optional[Any] = output.images[0]
assert image.shape == (5_1_2, 5_1_2, 3)
assert_mean_pixel_difference(_lowerCamelCase , _lowerCamelCase )
| 674 |
"""simple docstring"""
from collections.abc import Callable
import numpy as np
def lowerCAmelCase_ ( lowercase_ : Callable , lowercase_ : float , lowercase_ : float , lowercase_ : float , lowercase_ : float ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : int = int(np.ceil((x_end - xa) / step_size ) )
__SCREAMING_SNAKE_CASE : Dict = np.zeros((n + 1,) )
__SCREAMING_SNAKE_CASE : List[Any] = ya
__SCREAMING_SNAKE_CASE : Dict = xa
for k in range(lowercase_ ):
__SCREAMING_SNAKE_CASE : str = y[k] + step_size * ode_func(lowercase_ , y[k] )
__SCREAMING_SNAKE_CASE : int = y[k] + (
(step_size / 2) * (ode_func(lowercase_ , y[k] ) + ode_func(x + step_size , lowercase_ ))
)
x += step_size
return y
if __name__ == "__main__":
import doctest
doctest.testmod()
| 674 | 1 |
"""simple docstring"""
import argparse
import torch
from torch import nn
from transformers import MBartConfig, MBartForConditionalGeneration
def lowerCAmelCase_ ( lowercase_ : Dict ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[Any] = [
'''encoder.version''',
'''decoder.version''',
'''model.encoder.version''',
'''model.decoder.version''',
'''_float_tensor''',
'''decoder.output_projection.weight''',
]
for k in ignore_keys:
state_dict.pop(lowercase_ , lowercase_ )
def lowerCAmelCase_ ( lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[int] = emb.weight.shape
__SCREAMING_SNAKE_CASE : List[str] = nn.Linear(lowercase_ , lowercase_ , bias=lowercase_ )
__SCREAMING_SNAKE_CASE : List[str] = emb.weight.data
return lin_layer
def lowerCAmelCase_ ( lowercase_ : Any , lowercase_ : Union[str, Any]="facebook/mbart-large-en-ro" , lowercase_ : Any=False , lowercase_ : int=False ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[str] = torch.load(lowercase_ , map_location='''cpu''' )['''model''']
remove_ignore_keys_(lowercase_ )
__SCREAMING_SNAKE_CASE : List[str] = state_dict['''encoder.embed_tokens.weight'''].shape[0]
__SCREAMING_SNAKE_CASE : Any = MBartConfig.from_pretrained(lowercase_ , vocab_size=lowercase_ )
if mbart_aa and finetuned:
__SCREAMING_SNAKE_CASE : str = '''relu'''
__SCREAMING_SNAKE_CASE : Optional[Any] = state_dict['''decoder.embed_tokens.weight''']
__SCREAMING_SNAKE_CASE : Any = MBartForConditionalGeneration(lowercase_ )
model.model.load_state_dict(lowercase_ )
if finetuned:
__SCREAMING_SNAKE_CASE : Optional[Any] = make_linear_from_emb(model.model.shared )
return model
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''fairseq_path''', type=str, help='''bart.large, bart.large.cnn or a path to a model.pt on local filesystem.'''
)
parser.add_argument('''pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model.''')
parser.add_argument(
'''--hf_config''',
default='''facebook/mbart-large-cc25''',
type=str,
help='''Which huggingface architecture to use: mbart-large''',
)
parser.add_argument('''--mbart_50''', action='''store_true''', help='''whether the model is mMART-50 checkpoint''')
parser.add_argument('''--finetuned''', action='''store_true''', help='''whether the model is a fine-tuned checkpoint''')
_lowerCamelCase = parser.parse_args()
_lowerCamelCase = convert_fairseq_mbart_checkpoint_from_disk(
args.fairseq_path, hf_config_path=args.hf_config, finetuned=args.finetuned, mbart_aa=args.mbart_aa
)
model.save_pretrained(args.pytorch_dump_folder_path)
| 674 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...file_utils import _LazyModule, is_torch_available
from ...utils import OptionalDependencyNotAvailable
_lowerCamelCase = {
'''configuration_gpt_neox_japanese''': ['''GPT_NEOX_JAPANESE_PRETRAINED_CONFIG_ARCHIVE_MAP''', '''GPTNeoXJapaneseConfig'''],
'''tokenization_gpt_neox_japanese''': ['''GPTNeoXJapaneseTokenizer'''],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCamelCase = [
'''GPT_NEOX_JAPANESE_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''GPTNeoXJapaneseForCausalLM''',
'''GPTNeoXJapaneseLayer''',
'''GPTNeoXJapaneseModel''',
'''GPTNeoXJapanesePreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_gpt_neox_japanese import GPT_NEOX_JAPANESE_PRETRAINED_CONFIG_ARCHIVE_MAP, GPTNeoXJapaneseConfig
from .tokenization_gpt_neox_japanese import GPTNeoXJapaneseTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_gpt_neox_japanese import (
GPT_NEOX_JAPANESE_PRETRAINED_MODEL_ARCHIVE_LIST,
GPTNeoXJapaneseForCausalLM,
GPTNeoXJapaneseLayer,
GPTNeoXJapaneseModel,
GPTNeoXJapanesePreTrainedModel,
)
else:
import sys
_lowerCamelCase = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 674 | 1 |
"""simple docstring"""
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv('''TEST_SAGEMAKER''' , '''False''' ) ) is not True , reason='''Skipping test because should only be run when releasing minor transformers version''' , )
@pytest.mark.usefixtures('''sm_env''' )
@parameterized_class(
[
{
'''framework''': '''pytorch''',
'''script''': '''run_glue_model_parallelism.py''',
'''model_name_or_path''': '''roberta-large''',
'''instance_type''': '''ml.p3dn.24xlarge''',
'''results''': {'''train_runtime''': 1600, '''eval_accuracy''': 0.3, '''eval_loss''': 1.2},
},
{
'''framework''': '''pytorch''',
'''script''': '''run_glue.py''',
'''model_name_or_path''': '''roberta-large''',
'''instance_type''': '''ml.p3dn.24xlarge''',
'''results''': {'''train_runtime''': 1600, '''eval_accuracy''': 0.3, '''eval_loss''': 1.2},
},
] )
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
if self.framework == "pytorch":
subprocess.run(
f'''cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py'''.split() , encoding='''utf-8''' , check=_lowerCamelCase , )
assert hasattr(self , '''env''' )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] , _lowerCamelCase :Union[str, Any] ):
# configuration for running training on smdistributed Model Parallel
__SCREAMING_SNAKE_CASE : List[str] = {
'''enabled''': True,
'''processes_per_host''': 8,
}
__SCREAMING_SNAKE_CASE : int = {
'''enabled''': True,
'''parameters''': {
'''microbatches''': 4,
'''placement_strategy''': '''spread''',
'''pipeline''': '''interleaved''',
'''optimize''': '''speed''',
'''partitions''': 4,
'''ddp''': True,
},
}
__SCREAMING_SNAKE_CASE : List[Any] = {'''smdistributed''': {'''modelparallel''': smp_options}, '''mpi''': mpi_options}
__SCREAMING_SNAKE_CASE : Union[str, Any] = '''trainer''' if self.script == '''run_glue.py''' else '''smtrainer'''
# creates estimator
return HuggingFace(
entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f'''{self.env.base_job_name}-{instance_count}-smp-{name_extension}''' , instance_count=_lowerCamelCase , instance_type=self.instance_type , debugger_hook_config=_lowerCamelCase , hyperparameters={
**self.env.hyperparameters,
'''model_name_or_path''': self.model_name_or_path,
'''max_steps''': 5_0_0,
} , metric_definitions=self.env.metric_definitions , distribution=_lowerCamelCase , py_version='''py36''' , )
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :Optional[int] ):
TrainingJobAnalytics(_lowerCamelCase ).export_csv(f'''{self.env.test_path}/{job_name}_metrics.csv''' )
@parameterized.expand([(1,)] )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :Dict ):
# create estimator
__SCREAMING_SNAKE_CASE : str = self.create_estimator(_lowerCamelCase )
# run training
estimator.fit()
# result dataframe
__SCREAMING_SNAKE_CASE : int = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
__SCREAMING_SNAKE_CASE : int = list(result_metrics_df[result_metrics_df.metric_name == '''eval_accuracy''']['''value'''] )
__SCREAMING_SNAKE_CASE : Any = list(result_metrics_df[result_metrics_df.metric_name == '''eval_loss''']['''value'''] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
__SCREAMING_SNAKE_CASE : str = (
Session().describe_training_job(estimator.latest_training_job.name ).get('''TrainingTimeInSeconds''' , 9_9_9_9_9_9 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results['''eval_accuracy'''] for t in eval_accuracy )
assert all(t <= self.results['''eval_loss'''] for t in eval_loss )
# dump tests result into json file to share in PR
with open(f'''{estimator.latest_training_job.name}.json''' , '''w''' ) as outfile:
json.dump({'''train_time''': train_runtime, '''eval_accuracy''': eval_accuracy, '''eval_loss''': eval_loss} , _lowerCamelCase )
| 674 |
"""simple docstring"""
from __future__ import annotations
from typing import Any
class snake_case :
def __init__( self :Optional[Any] , _lowerCamelCase :int ):
__SCREAMING_SNAKE_CASE : int = num_of_nodes
__SCREAMING_SNAKE_CASE : list[list[int]] = []
__SCREAMING_SNAKE_CASE : dict[int, int] = {}
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :int , _lowerCamelCase :int , _lowerCamelCase :int ):
self.m_edges.append([u_node, v_node, weight] )
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :int ):
if self.m_component[u_node] == u_node:
return u_node
return self.find_component(self.m_component[u_node] )
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :int ):
if self.m_component[u_node] != u_node:
for k in self.m_component:
__SCREAMING_SNAKE_CASE : Optional[Any] = self.find_component(_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :list[int] , _lowerCamelCase :int , _lowerCamelCase :int ):
if component_size[u_node] <= component_size[v_node]:
__SCREAMING_SNAKE_CASE : List[Any] = v_node
component_size[v_node] += component_size[u_node]
self.set_component(_lowerCamelCase )
elif component_size[u_node] >= component_size[v_node]:
__SCREAMING_SNAKE_CASE : Dict = self.find_component(_lowerCamelCase )
component_size[u_node] += component_size[v_node]
self.set_component(_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Optional[int] = []
__SCREAMING_SNAKE_CASE : str = 0
__SCREAMING_SNAKE_CASE : list[Any] = [-1] * self.m_num_of_nodes
# A list of components (initialized to all of the nodes)
for node in range(self.m_num_of_nodes ):
self.m_component.update({node: node} )
component_size.append(1 )
__SCREAMING_SNAKE_CASE : str = self.m_num_of_nodes
while num_of_components > 1:
for edge in self.m_edges:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = edge
__SCREAMING_SNAKE_CASE : Optional[Any] = self.m_component[u]
__SCREAMING_SNAKE_CASE : int = self.m_component[v]
if u_component != v_component:
for component in (u_component, v_component):
if (
minimum_weight_edge[component] == -1
or minimum_weight_edge[component][2] > w
):
__SCREAMING_SNAKE_CASE : Optional[Any] = [u, v, w]
for edge in minimum_weight_edge:
if isinstance(_lowerCamelCase , _lowerCamelCase ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : int = edge
__SCREAMING_SNAKE_CASE : Tuple = self.m_component[u]
__SCREAMING_SNAKE_CASE : int = self.m_component[v]
if u_component != v_component:
mst_weight += w
self.union(_lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
print(f'''Added edge [{u} - {v}]\nAdded weight: {w}\n''' )
num_of_components -= 1
__SCREAMING_SNAKE_CASE : Optional[Any] = [-1] * self.m_num_of_nodes
print(f'''The total weight of the minimal spanning tree is: {mst_weight}''' )
def lowerCAmelCase_ ( ):
'''simple docstring'''
if __name__ == "__main__":
import doctest
doctest.testmod()
| 674 | 1 |
"""simple docstring"""
import argparse
import torch
from transformers import MobileBertConfig, MobileBertForPreTraining, load_tf_weights_in_mobilebert
from transformers.utils import logging
logging.set_verbosity_info()
def lowerCAmelCase_ ( lowercase_ : List[Any] , lowercase_ : List[Any] , lowercase_ : Optional[int] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = MobileBertConfig.from_json_file(lowercase_ )
print(F'''Building PyTorch model from configuration: {config}''' )
__SCREAMING_SNAKE_CASE : Any = MobileBertForPreTraining(lowercase_ )
# Load weights from tf checkpoint
__SCREAMING_SNAKE_CASE : str = load_tf_weights_in_mobilebert(lowercase_ , lowercase_ , lowercase_ )
# Save pytorch-model
print(F'''Save PyTorch model to {pytorch_dump_path}''' )
torch.save(model.state_dict() , lowercase_ )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--tf_checkpoint_path''', default=None, type=str, required=True, help='''Path to the TensorFlow checkpoint path.'''
)
parser.add_argument(
'''--mobilebert_config_file''',
default=None,
type=str,
required=True,
help=(
'''The config json file corresponding to the pre-trained MobileBERT model. \n'''
'''This specifies the model architecture.'''
),
)
parser.add_argument(
'''--pytorch_dump_path''', default=None, type=str, required=True, help='''Path to the output PyTorch model.'''
)
_lowerCamelCase = parser.parse_args()
convert_tf_checkpoint_to_pytorch(args.tf_checkpoint_path, args.mobilebert_config_file, args.pytorch_dump_path)
| 674 |
"""simple docstring"""
import argparse
import pickle
import numpy as np
import torch
from torch import nn
from transformers import ReformerConfig, ReformerModelWithLMHead
from transformers.utils import logging
logging.set_verbosity_info()
def lowerCAmelCase_ ( lowercase_ : int , lowercase_ : Any , lowercase_ : int=None ):
'''simple docstring'''
assert torch_layer.weight.shape == weight.shape, F'''{torch_layer} layer.weight does not match'''
__SCREAMING_SNAKE_CASE : str = nn.Parameter(lowercase_ )
if bias is not None:
assert torch_layer.bias.shape == bias.shape, F'''{torch_layer} layer.bias does not match'''
__SCREAMING_SNAKE_CASE : Tuple = nn.Parameter(lowercase_ )
def lowerCAmelCase_ ( lowercase_ : Tuple , lowercase_ : int , lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = np.asarray(weights[0] )
__SCREAMING_SNAKE_CASE : Optional[int] = np.asarray(weights[1] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(weights[2] )
set_param(
torch_layer.self_attention.query_key , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.value , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.output.dense , torch.tensor(lowercase_ ).view(-1 , lowercase_ ).contiguous().transpose(0 , 1 ) , )
def lowerCAmelCase_ ( lowercase_ : Dict , lowercase_ : List[str] , lowercase_ : List[str] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(weights[0] )
__SCREAMING_SNAKE_CASE : Any = np.asarray(weights[1] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(weights[2] )
__SCREAMING_SNAKE_CASE : Tuple = np.asarray(weights[3] )
set_param(
torch_layer.self_attention.query , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.key , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.self_attention.value , torch.tensor(lowercase_ ).transpose(1 , 2 ).contiguous().view(-1 , lowercase_ ) , )
set_param(
torch_layer.output.dense , torch.tensor(lowercase_ ).view(-1 , lowercase_ ).contiguous().transpose(0 , 1 ) , )
def lowerCAmelCase_ ( lowercase_ : int , lowercase_ : List[str] , lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = weights[0][0][0]
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(layer_norm_a[0] )
__SCREAMING_SNAKE_CASE : List[Any] = np.asarray(layer_norm_a[1] )
set_param(
torch_block.attention.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# lsh weights + output
__SCREAMING_SNAKE_CASE : Tuple = weights[0][1]
if len(lowercase_ ) < 4:
set_layer_weights_in_torch_lsh(lowercase_ , torch_block.attention , lowercase_ )
else:
set_layer_weights_in_torch_local(lowercase_ , torch_block.attention , lowercase_ )
# intermediate weighs
__SCREAMING_SNAKE_CASE : Any = weights[2][0][1][2]
# Chunked Feed Forward
if len(lowercase_ ) == 4:
__SCREAMING_SNAKE_CASE : List[str] = intermediate_weights[2]
# layernorm 2
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(intermediate_weights[0][0] )
__SCREAMING_SNAKE_CASE : Union[str, Any] = np.asarray(intermediate_weights[0][1] )
set_param(
torch_block.feed_forward.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# intermediate dense
__SCREAMING_SNAKE_CASE : int = np.asarray(intermediate_weights[1][0] )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(intermediate_weights[1][1] )
set_param(
torch_block.feed_forward.dense.dense , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
# intermediate out
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(intermediate_weights[4][0] )
__SCREAMING_SNAKE_CASE : Any = np.asarray(intermediate_weights[4][1] )
set_param(
torch_block.feed_forward.output.dense , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
def lowerCAmelCase_ ( lowercase_ : str , lowercase_ : Optional[Any] , lowercase_ : Optional[Any] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = torch_model.reformer
# word embeds
__SCREAMING_SNAKE_CASE : int = np.asarray(weights[1] )
set_param(
torch_model_reformer.embeddings.word_embeddings , torch.tensor(lowercase_ ) , )
if isinstance(weights[3] , lowercase_ ):
__SCREAMING_SNAKE_CASE : int = torch_model_reformer.embeddings.position_embeddings
for emb_idx in range(len(position_embeddings.weights ) ):
__SCREAMING_SNAKE_CASE : Dict = np.asarray(weights[3][emb_idx][0] )
assert (
position_embeddings.weights[emb_idx].shape == emb_weights.shape
), F'''{position_embeddings[emb_idx]} emb does not match'''
__SCREAMING_SNAKE_CASE : str = nn.Parameter(torch.tensor(lowercase_ ) )
__SCREAMING_SNAKE_CASE : List[Any] = weights[5]
assert len(torch_model_reformer.encoder.layers ) * 4 == len(
lowercase_ ), "HF and trax model do not have the same number of layers"
for layer_idx, layer in enumerate(torch_model_reformer.encoder.layers ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = trax_layer_weights[4 * layer_idx : 4 * (layer_idx + 1)]
set_block_weights_in_torch(lowercase_ , lowercase_ , lowercase_ )
# output layer norm
__SCREAMING_SNAKE_CASE : List[str] = np.asarray(weights[7][0] )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(weights[7][1] )
set_param(
torch_model_reformer.encoder.layer_norm , torch.tensor(lowercase_ ) , torch.tensor(lowercase_ ) , )
# output embeddings
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(weights[9][0] )
__SCREAMING_SNAKE_CASE : List[Any] = np.asarray(weights[9][1] )
set_param(
torch_model.lm_head.decoder , torch.tensor(lowercase_ ).transpose(0 , 1 ).contiguous() , torch.tensor(lowercase_ ) , )
def lowerCAmelCase_ ( lowercase_ : List[str] , lowercase_ : Any , lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = ReformerConfig.from_json_file(lowercase_ )
print(F'''Building PyTorch model from configuration: {config}''' )
__SCREAMING_SNAKE_CASE : List[str] = ReformerModelWithLMHead(lowercase_ )
with open(lowercase_ , '''rb''' ) as f:
__SCREAMING_SNAKE_CASE : int = pickle.load(lowercase_ )['''weights''']
set_model_weights_in_torch(lowercase_ , lowercase_ , config.hidden_size )
# Save pytorch-model
print(F'''Save PyTorch model to {pytorch_dump_path}''' )
torch.save(model.state_dict() , lowercase_ )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--trax_model_pkl_path''', default=None, type=str, required=True, help='''Path to the TensorFlow checkpoint path.'''
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
required=True,
help=(
'''The config json file corresponding to the pre-trained Reformer model. \n'''
'''This specifies the model architecture.'''
),
)
parser.add_argument(
'''--pytorch_dump_path''', default=None, type=str, required=True, help='''Path to the output PyTorch model.'''
)
_lowerCamelCase = parser.parse_args()
convert_trax_checkpoint_to_pytorch(args.trax_model_pkl_path, args.config_file, args.pytorch_dump_path)
| 674 | 1 |
"""simple docstring"""
import math
import sys
def lowerCAmelCase_ ( lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = ''''''
try:
with open(lowercase_ , '''rb''' ) as binary_file:
__SCREAMING_SNAKE_CASE : Dict = binary_file.read()
for dat in data:
__SCREAMING_SNAKE_CASE : Optional[int] = F'''{dat:08b}'''
result += curr_byte
return result
except OSError:
print('''File not accessible''' )
sys.exit()
def lowerCAmelCase_ ( lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[str] = {'''0''': '''0''', '''1''': '''1'''}
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = '''''', ''''''
__SCREAMING_SNAKE_CASE : Tuple = len(lowercase_ )
for i in range(len(lowercase_ ) ):
curr_string += data_bits[i]
if curr_string not in lexicon:
continue
__SCREAMING_SNAKE_CASE : int = lexicon[curr_string]
result += last_match_id
__SCREAMING_SNAKE_CASE : str = last_match_id + '''0'''
if math.loga(lowercase_ ).is_integer():
__SCREAMING_SNAKE_CASE : Optional[Any] = {}
for curr_key in list(lowercase_ ):
__SCREAMING_SNAKE_CASE : List[Any] = lexicon.pop(lowercase_ )
__SCREAMING_SNAKE_CASE : List[Any] = new_lex
__SCREAMING_SNAKE_CASE : Dict = last_match_id + '''1'''
index += 1
__SCREAMING_SNAKE_CASE : List[Any] = ''''''
return result
def lowerCAmelCase_ ( lowercase_ : str , lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : str = 8
try:
with open(lowercase_ , '''wb''' ) as opened_file:
__SCREAMING_SNAKE_CASE : str = [
to_write[i : i + byte_length]
for i in range(0 , len(lowercase_ ) , lowercase_ )
]
if len(result_byte_array[-1] ) % byte_length == 0:
result_byte_array.append('''10000000''' )
else:
result_byte_array[-1] += "1" + "0" * (
byte_length - len(result_byte_array[-1] ) - 1
)
for elem in result_byte_array[:-1]:
opened_file.write(int(lowercase_ , 2 ).to_bytes(1 , byteorder='''big''' ) )
except OSError:
print('''File not accessible''' )
sys.exit()
def lowerCAmelCase_ ( lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = 0
for letter in data_bits:
if letter == "1":
break
counter += 1
__SCREAMING_SNAKE_CASE : Tuple = data_bits[counter:]
__SCREAMING_SNAKE_CASE : int = data_bits[counter + 1 :]
return data_bits
def lowerCAmelCase_ ( lowercase_ : str , lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[Any] = read_file_binary(lowercase_ )
__SCREAMING_SNAKE_CASE : List[Any] = remove_prefix(lowercase_ )
__SCREAMING_SNAKE_CASE : Tuple = decompress_data(lowercase_ )
write_file_binary(lowercase_ , lowercase_ )
if __name__ == "__main__":
compress(sys.argv[1], sys.argv[2])
| 674 |
"""simple docstring"""
from typing import Callable, Optional, Union
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = {
'''microsoft/xprophetnet-large-wiki100-cased''': (
'''https://huggingface.co/microsoft/xprophetnet-large-wiki100-cased/resolve/main/config.json'''
),
}
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = '''xlm-prophetnet'''
lowerCamelCase__ = ['''past_key_values''']
lowerCamelCase__ = {
'''num_attention_heads''': '''num_encoder_attention_heads''',
}
def __init__( self :List[str] , _lowerCamelCase :Optional[float] = 0.1 , _lowerCamelCase :Optional[Union[str, Callable]] = "gelu" , _lowerCamelCase :Optional[int] = 3_0_5_2_2 , _lowerCamelCase :Optional[int] = 1_0_2_4 , _lowerCamelCase :Optional[int] = 4_0_9_6 , _lowerCamelCase :Optional[int] = 1_2 , _lowerCamelCase :Optional[int] = 1_6 , _lowerCamelCase :Optional[int] = 4_0_9_6 , _lowerCamelCase :Optional[int] = 1_2 , _lowerCamelCase :Optional[int] = 1_6 , _lowerCamelCase :Optional[float] = 0.1 , _lowerCamelCase :Optional[float] = 0.1 , _lowerCamelCase :Optional[int] = 5_1_2 , _lowerCamelCase :Optional[float] = 0.0_2 , _lowerCamelCase :Optional[bool] = True , _lowerCamelCase :Optional[bool] = True , _lowerCamelCase :Optional[int] = 0 , _lowerCamelCase :Optional[int] = 2 , _lowerCamelCase :Optional[int] = 3_2 , _lowerCamelCase :Optional[int] = 1_2_8 , _lowerCamelCase :Optional[bool] = False , _lowerCamelCase :Optional[float] = 0.0 , _lowerCamelCase :Optional[bool] = True , _lowerCamelCase :Optional[int] = 0 , _lowerCamelCase :Optional[int] = 1 , _lowerCamelCase :Optional[int] = 2 , **_lowerCamelCase :int , ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = vocab_size
__SCREAMING_SNAKE_CASE : Optional[int] = hidden_size
__SCREAMING_SNAKE_CASE : List[Any] = encoder_ffn_dim
__SCREAMING_SNAKE_CASE : str = num_encoder_layers
__SCREAMING_SNAKE_CASE : Optional[Any] = num_encoder_attention_heads
__SCREAMING_SNAKE_CASE : str = decoder_ffn_dim
__SCREAMING_SNAKE_CASE : List[Any] = num_decoder_layers
__SCREAMING_SNAKE_CASE : List[str] = num_decoder_attention_heads
__SCREAMING_SNAKE_CASE : Dict = max_position_embeddings
__SCREAMING_SNAKE_CASE : Any = init_std # Normal(0, this parameter)
__SCREAMING_SNAKE_CASE : Any = activation_function
# parameters for xlmprophetnet
__SCREAMING_SNAKE_CASE : List[Any] = ngram
__SCREAMING_SNAKE_CASE : int = num_buckets
__SCREAMING_SNAKE_CASE : List[str] = relative_max_distance
__SCREAMING_SNAKE_CASE : str = disable_ngram_loss
__SCREAMING_SNAKE_CASE : Optional[int] = eps
# 3 Types of Dropout
__SCREAMING_SNAKE_CASE : int = attention_dropout
__SCREAMING_SNAKE_CASE : Optional[Any] = activation_dropout
__SCREAMING_SNAKE_CASE : Dict = dropout
__SCREAMING_SNAKE_CASE : Any = use_cache
super().__init__(
pad_token_id=_lowerCamelCase , bos_token_id=_lowerCamelCase , eos_token_id=_lowerCamelCase , is_encoder_decoder=_lowerCamelCase , add_cross_attention=_lowerCamelCase , decoder_start_token_id=_lowerCamelCase , **_lowerCamelCase , )
@property
def SCREAMING_SNAKE_CASE_ ( self :int ):
return self.num_encoder_layers + self.num_decoder_layers
@num_hidden_layers.setter
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :List[Any] ):
raise NotImplementedError(
'''This model does not support the setting of `num_hidden_layers`. Please set `num_encoder_layers` and'''
''' `num_decoder_layers`.''' )
| 674 | 1 |
"""simple docstring"""
def lowerCAmelCase_ ( lowercase_ : int ):
'''simple docstring'''
assert isinstance(lowercase_ , lowercase_ ), F'''The input value of [n={number}] is not an integer'''
if number == 1:
return 2
elif number < 1:
__SCREAMING_SNAKE_CASE : Union[str, Any] = F'''The input value of [n={number}] has to be > 0'''
raise ValueError(lowercase_ )
else:
__SCREAMING_SNAKE_CASE : Tuple = sylvester(number - 1 )
__SCREAMING_SNAKE_CASE : Any = num - 1
__SCREAMING_SNAKE_CASE : Tuple = num
return lower * upper + 1
if __name__ == "__main__":
print(f'The 8th number in Sylvester\'s sequence: {sylvester(8)}')
| 674 |
"""simple docstring"""
import ast
import os
import re
import shutil
import tempfile
import unittest
from unittest import mock
import torch
from accelerate.test_utils.examples import compare_against_test
from accelerate.test_utils.testing import TempDirTestCase, require_trackers, run_command, slow
from accelerate.utils import write_basic_config
# DataLoaders built from `test_samples/MRPC` for quick testing
# Should mock `{script_name}.get_dataloaders` via:
# @mock.patch("{script_name}.get_dataloaders", mocked_dataloaders)
_lowerCamelCase = [
'''cross_validation.py''',
'''gradient_accumulation.py''',
'''local_sgd.py''',
'''multi_process_metrics.py''',
'''memory.py''',
'''automatic_gradient_accumulation.py''',
'''fsdp_with_peak_mem_tracking.py''',
'''deepspeed_with_config_support.py''',
'''megatron_lm_gpt_pretraining.py''',
]
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :str , _lowerCamelCase :bool , _lowerCamelCase :str = None , _lowerCamelCase :list = None ):
__SCREAMING_SNAKE_CASE : List[str] = None
__SCREAMING_SNAKE_CASE : Optional[Any] = os.path.abspath(os.path.join('''examples''' , '''by_feature''' ) )
__SCREAMING_SNAKE_CASE : Union[str, Any] = os.path.abspath('''examples''' )
for item in os.listdir(_lowerCamelCase ):
if item not in EXCLUDE_EXAMPLES:
__SCREAMING_SNAKE_CASE : List[Any] = os.path.join(_lowerCamelCase , _lowerCamelCase )
if os.path.isfile(_lowerCamelCase ) and ".py" in item_path:
with self.subTest(
tested_script=_lowerCamelCase , feature_script=_lowerCamelCase , tested_section='''main()''' if parser_only else '''training_function()''' , ):
__SCREAMING_SNAKE_CASE : Tuple = compare_against_test(
os.path.join(_lowerCamelCase , _lowerCamelCase ) , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = '''\n'''.join(_lowerCamelCase )
if special_strings is not None:
for string in special_strings:
__SCREAMING_SNAKE_CASE : List[Any] = diff.replace(_lowerCamelCase , '''''' )
self.assertEqual(_lowerCamelCase , '''''' )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
self.one_complete_example('''complete_nlp_example.py''' , _lowerCamelCase )
self.one_complete_example('''complete_nlp_example.py''' , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = os.path.abspath(os.path.join('''examples''' , '''cv_example.py''' ) )
__SCREAMING_SNAKE_CASE : Optional[int] = [
''' ''' * 1_6 + '''{\n\n''',
''' ''' * 2_0 + '''"accuracy": eval_metric["accuracy"],\n\n''',
''' ''' * 2_0 + '''"f1": eval_metric["f1"],\n\n''',
''' ''' * 2_0 + '''"train_loss": total_loss.item() / len(train_dataloader),\n\n''',
''' ''' * 2_0 + '''"epoch": epoch,\n\n''',
''' ''' * 1_6 + '''},\n\n''',
''' ''' * 1_6 + '''step=epoch,\n''',
''' ''' * 1_2,
''' ''' * 8 + '''for step, batch in enumerate(active_dataloader):\n''',
]
self.one_complete_example('''complete_cv_example.py''' , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
self.one_complete_example('''complete_cv_example.py''' , _lowerCamelCase , _lowerCamelCase , _lowerCamelCase )
@mock.patch.dict(os.environ , {'''TESTING_MOCKED_DATALOADERS''': '''1'''} )
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = False
@classmethod
def SCREAMING_SNAKE_CASE_ ( cls :Dict ):
super().setUpClass()
__SCREAMING_SNAKE_CASE : Dict = tempfile.mkdtemp()
__SCREAMING_SNAKE_CASE : str = os.path.join(cls._tmpdir , '''default_config.yml''' )
write_basic_config(save_location=cls.configPath )
__SCREAMING_SNAKE_CASE : List[Any] = ['''accelerate''', '''launch''', '''--config_file''', cls.configPath]
@classmethod
def SCREAMING_SNAKE_CASE_ ( cls :Dict ):
super().tearDownClass()
shutil.rmtree(cls._tmpdir )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : str = f'''
examples/by_feature/checkpointing.py
--checkpointing_steps epoch
--output_dir {self.tmpdir}
'''.split()
run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(self.tmpdir , '''epoch_0''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Optional[Any] = f'''
examples/by_feature/checkpointing.py
--checkpointing_steps 1
--output_dir {self.tmpdir}
'''.split()
__SCREAMING_SNAKE_CASE : Optional[int] = run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(self.tmpdir , '''step_2''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Optional[int] = f'''
examples/by_feature/checkpointing.py
--resume_from_checkpoint {os.path.join(self.tmpdir , 'epoch_0' )}
'''.split()
__SCREAMING_SNAKE_CASE : Any = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
self.assertNotIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Optional[int] = f'''
examples/by_feature/checkpointing.py
--resume_from_checkpoint {os.path.join(self.tmpdir , 'step_2' )}
'''.split()
__SCREAMING_SNAKE_CASE : List[str] = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
if torch.cuda.is_available():
__SCREAMING_SNAKE_CASE : List[Any] = torch.cuda.device_count()
else:
__SCREAMING_SNAKE_CASE : Optional[int] = 1
if num_processes > 1:
self.assertNotIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
else:
self.assertIn('''epoch 0:''' , _lowerCamelCase )
self.assertIn('''epoch 1:''' , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = '''
examples/by_feature/cross_validation.py
--num_folds 2
'''.split()
with mock.patch.dict(os.environ , {'''TESTING_MOCKED_DATALOADERS''': '''0'''} ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = run_command(self._launch_args + testargs , return_stdout=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = re.findall('''({.+})''' , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = [r for r in results if '''accuracy''' in r][-1]
__SCREAMING_SNAKE_CASE : Tuple = ast.literal_eval(_lowerCamelCase )
self.assertGreaterEqual(results['''accuracy'''] , 0.7_5 )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = ['''examples/by_feature/multi_process_metrics.py''']
run_command(self._launch_args + testargs )
@require_trackers
@mock.patch.dict(os.environ , {'''WANDB_MODE''': '''offline'''} )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
with tempfile.TemporaryDirectory() as tmpdir:
__SCREAMING_SNAKE_CASE : int = f'''
examples/by_feature/tracking.py
--with_tracking
--project_dir {tmpdir}
'''.split()
run_command(self._launch_args + testargs )
self.assertTrue(os.path.exists(os.path.join(_lowerCamelCase , '''tracking''' ) ) )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = ['''examples/by_feature/gradient_accumulation.py''']
run_command(self._launch_args + testargs )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[Any] = ['''examples/by_feature/local_sgd.py''']
run_command(self._launch_args + testargs )
| 674 | 1 |
"""simple docstring"""
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = [31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31]
__SCREAMING_SNAKE_CASE : str = 6
__SCREAMING_SNAKE_CASE : int = 1
__SCREAMING_SNAKE_CASE : List[Any] = 1901
__SCREAMING_SNAKE_CASE : Optional[int] = 0
while year < 2001:
day += 7
if (year % 4 == 0 and year % 100 != 0) or (year % 400 == 0):
if day > days_per_month[month - 1] and month != 2:
month += 1
__SCREAMING_SNAKE_CASE : Optional[Any] = day - days_per_month[month - 2]
elif day > 29 and month == 2:
month += 1
__SCREAMING_SNAKE_CASE : Optional[Any] = day - 29
else:
if day > days_per_month[month - 1]:
month += 1
__SCREAMING_SNAKE_CASE : Optional[int] = day - days_per_month[month - 2]
if month > 12:
year += 1
__SCREAMING_SNAKE_CASE : Any = 1
if year < 2001 and day == 1:
sundays += 1
return sundays
if __name__ == "__main__":
print(solution())
| 674 |
"""simple docstring"""
import argparse
import logging
import os
import time
import timeit
import datasets
import numpy as np
import pycuda.autoinit # noqa: F401
import pycuda.driver as cuda
import tensorrt as trt
import torch
from absl import logging as absl_logging
from accelerate import Accelerator
from datasets import load_dataset, load_metric
from torch.utils.data import DataLoader
from utils_qa import postprocess_qa_predictions
import transformers
from transformers import AutoTokenizer, EvalPrediction, default_data_collator, set_seed
from transformers.trainer_pt_utils import nested_concat, nested_truncate
_lowerCamelCase = trt.Logger(trt.Logger.WARNING)
_lowerCamelCase = absl_logging.get_absl_logger()
absl_logger.setLevel(logging.WARNING)
_lowerCamelCase = logging.getLogger(__name__)
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--onnx_model_path''',
default=None,
type=str,
required=True,
help='''Path to ONNX model: ''',
)
parser.add_argument(
'''--output_dir''',
default=None,
type=str,
required=True,
help='''The output directory where the model checkpoints and predictions will be written.''',
)
# Other parameters
parser.add_argument(
'''--tokenizer_name''',
default='''''',
type=str,
required=True,
help='''Pretrained tokenizer name or path if not the same as model_name''',
)
parser.add_argument(
'''--version_2_with_negative''',
action='''store_true''',
help='''If true, the SQuAD examples contain some that do not have an answer.''',
)
parser.add_argument(
'''--null_score_diff_threshold''',
type=float,
default=0.0,
help='''If null_score - best_non_null is greater than the threshold predict null.''',
)
parser.add_argument(
'''--max_seq_length''',
default=3_84,
type=int,
help=(
'''The maximum total input sequence length after WordPiece tokenization. Sequences '''
'''longer than this will be truncated, and sequences shorter than this will be padded.'''
),
)
parser.add_argument(
'''--doc_stride''',
default=1_28,
type=int,
help='''When splitting up a long document into chunks, how much stride to take between chunks.''',
)
parser.add_argument('''--per_device_eval_batch_size''', default=8, type=int, help='''Batch size per GPU/CPU for evaluation.''')
parser.add_argument(
'''--n_best_size''',
default=20,
type=int,
help='''The total number of n-best predictions to generate in the nbest_predictions.json output file.''',
)
parser.add_argument(
'''--max_answer_length''',
default=30,
type=int,
help=(
'''The maximum length of an answer that can be generated. This is needed because the start '''
'''and end predictions are not conditioned on one another.'''
),
)
parser.add_argument('''--seed''', type=int, default=42, help='''random seed for initialization''')
parser.add_argument(
'''--dataset_name''',
type=str,
default=None,
required=True,
help='''The name of the dataset to use (via the datasets library).''',
)
parser.add_argument(
'''--dataset_config_name''',
type=str,
default=None,
help='''The configuration name of the dataset to use (via the datasets library).''',
)
parser.add_argument(
'''--preprocessing_num_workers''', type=int, default=4, help='''A csv or a json file containing the training data.'''
)
parser.add_argument('''--overwrite_cache''', action='''store_true''', help='''Overwrite the cached training and evaluation sets''')
parser.add_argument(
'''--fp16''',
action='''store_true''',
help='''Whether to use 16-bit (mixed) precision instead of 32-bit''',
)
parser.add_argument(
'''--int8''',
action='''store_true''',
help='''Whether to use INT8''',
)
_lowerCamelCase = parser.parse_args()
if args.tokenizer_name:
_lowerCamelCase = AutoTokenizer.from_pretrained(args.tokenizer_name, use_fast=True)
else:
raise ValueError(
'''You are instantiating a new tokenizer from scratch. This is not supported by this script.'''
'''You can do it from another script, save it, and load it from here, using --tokenizer_name.'''
)
logger.info('''Training/evaluation parameters %s''', args)
_lowerCamelCase = args.per_device_eval_batch_size
_lowerCamelCase = (args.eval_batch_size, args.max_seq_length)
# TRT Engine properties
_lowerCamelCase = True
_lowerCamelCase = '''temp_engine/bert-fp32.engine'''
if args.fpaa:
_lowerCamelCase = '''temp_engine/bert-fp16.engine'''
if args.inta:
_lowerCamelCase = '''temp_engine/bert-int8.engine'''
# import ONNX file
if not os.path.exists('''temp_engine'''):
os.makedirs('''temp_engine''')
_lowerCamelCase = 1 << (int)(trt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH)
with trt.Builder(TRT_LOGGER) as builder, builder.create_network(EXPLICIT_BATCH) as network, trt.OnnxParser(
network, TRT_LOGGER
) as parser:
with open(args.onnx_model_path, '''rb''') as model:
if not parser.parse(model.read()):
for error in range(parser.num_errors):
print(parser.get_error(error))
# Query input names and shapes from parsed TensorRT network
_lowerCamelCase = [network.get_input(i) for i in range(network.num_inputs)]
_lowerCamelCase = [_input.name for _input in network_inputs] # ex: ["actual_input1"]
with builder.create_builder_config() as config:
_lowerCamelCase = 1 << 50
if STRICT_TYPES:
config.set_flag(trt.BuilderFlag.STRICT_TYPES)
if args.fpaa:
config.set_flag(trt.BuilderFlag.FPaa)
if args.inta:
config.set_flag(trt.BuilderFlag.INTa)
_lowerCamelCase = builder.create_optimization_profile()
config.add_optimization_profile(profile)
for i in range(len(input_names)):
profile.set_shape(input_names[i], INPUT_SHAPE, INPUT_SHAPE, INPUT_SHAPE)
_lowerCamelCase = builder.build_engine(network, config)
# serialize_engine and store in file (can be directly loaded and deserialized):
with open(engine_name, '''wb''') as f:
f.write(engine.serialize())
def lowerCAmelCase_ ( lowercase_ : List[Any] , lowercase_ : Tuple , lowercase_ : List[Any] , lowercase_ : List[Any] , lowercase_ : str , lowercase_ : Tuple , lowercase_ : Any , lowercase_ : Optional[int] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = np.asarray(inputs['''input_ids'''] , dtype=np.intaa )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(inputs['''attention_mask'''] , dtype=np.intaa )
__SCREAMING_SNAKE_CASE : Optional[Any] = np.asarray(inputs['''token_type_ids'''] , dtype=np.intaa )
# Copy inputs
cuda.memcpy_htod_async(d_inputs[0] , input_ids.ravel() , lowercase_ )
cuda.memcpy_htod_async(d_inputs[1] , attention_mask.ravel() , lowercase_ )
cuda.memcpy_htod_async(d_inputs[2] , token_type_ids.ravel() , lowercase_ )
# start time
__SCREAMING_SNAKE_CASE : Tuple = time.time()
# Run inference
context.execute_async(
bindings=[int(lowercase_ ) for d_inp in d_inputs] + [int(lowercase_ ), int(lowercase_ )] , stream_handle=stream.handle )
# Transfer predictions back from GPU
cuda.memcpy_dtoh_async(lowercase_ , lowercase_ , lowercase_ )
cuda.memcpy_dtoh_async(lowercase_ , lowercase_ , lowercase_ )
# Synchronize the stream and take time
stream.synchronize()
# end time
__SCREAMING_SNAKE_CASE : List[str] = time.time()
__SCREAMING_SNAKE_CASE : int = end_time - start_time
__SCREAMING_SNAKE_CASE : int = (h_outputa, h_outputa)
# print(outputs)
return outputs, infer_time
# Initialize the accelerator. We will let the accelerator handle device placement for us in this example.
_lowerCamelCase = Accelerator()
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format='''%(asctime)s - %(levelname)s - %(name)s - %(message)s''',
datefmt='''%m/%d/%Y %H:%M:%S''',
level=logging.INFO,
)
# Setup logging, we only want one process per machine to log things on the screen.
# accelerator.is_local_main_process is only True for one process per machine.
logger.setLevel(logging.INFO if accelerator.is_local_main_process else logging.ERROR)
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_info()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
# If passed along, set the training seed now.
if args.seed is not None:
set_seed(args.seed)
# Get the datasets: you can either provide your own CSV/JSON/TXT training and evaluation files (see below)
# or just provide the name of one of the public datasets available on the hub at https://huggingface.co/datasets/
# (the dataset will be downloaded automatically from the datasets Hub).
#
# For CSV/JSON files, this script will use the column called 'text' or the first column if no column called
# 'text' is found. You can easily tweak this behavior (see below).
if args.dataset_name is not None:
# Downloading and loading a dataset from the hub.
_lowerCamelCase = load_dataset(args.dataset_name, args.dataset_config_name)
else:
raise ValueError('''Evaluation requires a dataset name''')
# See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
# https://huggingface.co/docs/datasets/loading_datasets.html.
# Preprocessing the datasets.
# Preprocessing is slighlty different for training and evaluation.
_lowerCamelCase = raw_datasets['''validation'''].column_names
_lowerCamelCase = '''question''' if '''question''' in column_names else column_names[0]
_lowerCamelCase = '''context''' if '''context''' in column_names else column_names[1]
_lowerCamelCase = '''answers''' if '''answers''' in column_names else column_names[2]
# Padding side determines if we do (question|context) or (context|question).
_lowerCamelCase = tokenizer.padding_side == '''right'''
if args.max_seq_length > tokenizer.model_max_length:
logger.warning(
f'The max_seq_length passed ({args.max_seq_length}) is larger than the maximum length for the'
f'model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}.'
)
_lowerCamelCase = min(args.max_seq_length, tokenizer.model_max_length)
def lowerCAmelCase_ ( lowercase_ : Tuple ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = [q.lstrip() for q in examples[question_column_name]]
# Tokenize our examples with truncation and maybe padding, but keep the overflows using a stride. This results
# in one example possible giving several features when a context is long, each of those features having a
# context that overlaps a bit the context of the previous feature.
__SCREAMING_SNAKE_CASE : Optional[Any] = tokenizer(
examples[question_column_name if pad_on_right else context_column_name] , examples[context_column_name if pad_on_right else question_column_name] , truncation='''only_second''' if pad_on_right else '''only_first''' , max_length=lowercase_ , stride=args.doc_stride , return_overflowing_tokens=lowercase_ , return_offsets_mapping=lowercase_ , padding='''max_length''' , )
# Since one example might give us several features if it has a long context, we need a map from a feature to
# its corresponding example. This key gives us just that.
__SCREAMING_SNAKE_CASE : Optional[int] = tokenized_examples.pop('''overflow_to_sample_mapping''' )
# For evaluation, we will need to convert our predictions to substrings of the context, so we keep the
# corresponding example_id and we will store the offset mappings.
__SCREAMING_SNAKE_CASE : Any = []
for i in range(len(tokenized_examples['''input_ids'''] ) ):
# Grab the sequence corresponding to that example (to know what is the context and what is the question).
__SCREAMING_SNAKE_CASE : int = tokenized_examples.sequence_ids(lowercase_ )
__SCREAMING_SNAKE_CASE : str = 1 if pad_on_right else 0
# One example can give several spans, this is the index of the example containing this span of text.
__SCREAMING_SNAKE_CASE : str = sample_mapping[i]
tokenized_examples["example_id"].append(examples['''id'''][sample_index] )
# Set to None the offset_mapping that are not part of the context so it's easy to determine if a token
# position is part of the context or not.
__SCREAMING_SNAKE_CASE : List[str] = [
(o if sequence_ids[k] == context_index else None)
for k, o in enumerate(tokenized_examples['''offset_mapping'''][i] )
]
return tokenized_examples
_lowerCamelCase = raw_datasets['''validation''']
# Validation Feature Creation
_lowerCamelCase = eval_examples.map(
prepare_validation_features,
batched=True,
num_proc=args.preprocessing_num_workers,
remove_columns=column_names,
load_from_cache_file=not args.overwrite_cache,
desc='''Running tokenizer on validation dataset''',
)
_lowerCamelCase = default_data_collator
_lowerCamelCase = eval_dataset.remove_columns(['''example_id''', '''offset_mapping'''])
_lowerCamelCase = DataLoader(
eval_dataset_for_model, collate_fn=data_collator, batch_size=args.per_device_eval_batch_size
)
def lowerCAmelCase_ ( lowercase_ : Dict , lowercase_ : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : List[Any]="eval" ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = postprocess_qa_predictions(
examples=lowercase_ , features=lowercase_ , predictions=lowercase_ , version_2_with_negative=args.version_2_with_negative , n_best_size=args.n_best_size , max_answer_length=args.max_answer_length , null_score_diff_threshold=args.null_score_diff_threshold , output_dir=args.output_dir , prefix=lowercase_ , )
# Format the result to the format the metric expects.
if args.version_2_with_negative:
__SCREAMING_SNAKE_CASE : Union[str, Any] = [
{'''id''': k, '''prediction_text''': v, '''no_answer_probability''': 0.0} for k, v in predictions.items()
]
else:
__SCREAMING_SNAKE_CASE : int = [{'''id''': k, '''prediction_text''': v} for k, v in predictions.items()]
__SCREAMING_SNAKE_CASE : Any = [{'''id''': ex['''id'''], '''answers''': ex[answer_column_name]} for ex in examples]
return EvalPrediction(predictions=lowercase_ , label_ids=lowercase_ )
_lowerCamelCase = load_metric('''squad_v2''' if args.version_2_with_negative else '''squad''')
# Evaluation!
logger.info('''Loading ONNX model %s for evaluation''', args.onnx_model_path)
with open(engine_name, '''rb''') as f, trt.Runtime(TRT_LOGGER) as runtime, runtime.deserialize_cuda_engine(
f.read()
) as engine, engine.create_execution_context() as context:
# setup for TRT inferrence
for i in range(len(input_names)):
context.set_binding_shape(i, INPUT_SHAPE)
assert context.all_binding_shapes_specified
def lowerCAmelCase_ ( lowercase_ : Any ):
'''simple docstring'''
return trt.volume(engine.get_binding_shape(lowercase_ ) ) * engine.get_binding_dtype(lowercase_ ).itemsize
# Allocate device memory for inputs and outputs.
_lowerCamelCase = [cuda.mem_alloc(binding_nbytes(binding)) for binding in engine if engine.binding_is_input(binding)]
# Allocate output buffer
_lowerCamelCase = cuda.pagelocked_empty(tuple(context.get_binding_shape(3)), dtype=np.floataa)
_lowerCamelCase = cuda.pagelocked_empty(tuple(context.get_binding_shape(4)), dtype=np.floataa)
_lowerCamelCase = cuda.mem_alloc(h_outputa.nbytes)
_lowerCamelCase = cuda.mem_alloc(h_outputa.nbytes)
# Create a stream in which to copy inputs/outputs and run inference.
_lowerCamelCase = cuda.Stream()
# Evaluation
logger.info('''***** Running Evaluation *****''')
logger.info(f' Num examples = {len(eval_dataset)}')
logger.info(f' Batch size = {args.per_device_eval_batch_size}')
_lowerCamelCase = 0.0
_lowerCamelCase = 0
_lowerCamelCase = timeit.default_timer()
_lowerCamelCase = None
for step, batch in enumerate(eval_dataloader):
_lowerCamelCase , _lowerCamelCase = model_infer(batch, context, d_inputs, h_outputa, h_outputa, d_outputa, d_outputa, stream)
total_time += infer_time
niter += 1
_lowerCamelCase , _lowerCamelCase = outputs
_lowerCamelCase = torch.tensor(start_logits)
_lowerCamelCase = torch.tensor(end_logits)
# necessary to pad predictions and labels for being gathered
_lowerCamelCase = accelerator.pad_across_processes(start_logits, dim=1, pad_index=-1_00)
_lowerCamelCase = accelerator.pad_across_processes(end_logits, dim=1, pad_index=-1_00)
_lowerCamelCase = (accelerator.gather(start_logits).cpu().numpy(), accelerator.gather(end_logits).cpu().numpy())
_lowerCamelCase = logits if all_preds is None else nested_concat(all_preds, logits, padding_index=-1_00)
if all_preds is not None:
_lowerCamelCase = nested_truncate(all_preds, len(eval_dataset))
_lowerCamelCase = timeit.default_timer() - start_time
logger.info(''' Evaluation done in total %f secs (%f sec per example)''', evalTime, evalTime / len(eval_dataset))
# Inference time from TRT
logger.info('''Average Inference Time = {:.3f} ms'''.format(total_time * 10_00 / niter))
logger.info('''Total Inference Time = {:.3f} ms'''.format(total_time * 10_00))
logger.info('''Total Number of Inference = %d''', niter)
_lowerCamelCase = post_processing_function(eval_examples, eval_dataset, all_preds)
_lowerCamelCase = metric.compute(predictions=prediction.predictions, references=prediction.label_ids)
logger.info(f'Evaluation metrics: {eval_metric}')
| 674 | 1 |
"""simple docstring"""
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = {'''openai-gpt''': '''https://huggingface.co/openai-gpt/resolve/main/config.json'''}
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = '''openai-gpt'''
lowerCamelCase__ = {
'''max_position_embeddings''': '''n_positions''',
'''hidden_size''': '''n_embd''',
'''num_attention_heads''': '''n_head''',
'''num_hidden_layers''': '''n_layer''',
}
def __init__( self :str , _lowerCamelCase :Union[str, Any]=4_0_4_7_8 , _lowerCamelCase :Dict=5_1_2 , _lowerCamelCase :Dict=7_6_8 , _lowerCamelCase :Tuple=1_2 , _lowerCamelCase :List[Any]=1_2 , _lowerCamelCase :Union[str, Any]="gelu" , _lowerCamelCase :List[Any]=0.1 , _lowerCamelCase :List[Any]=0.1 , _lowerCamelCase :Tuple=0.1 , _lowerCamelCase :Optional[Any]=1e-5 , _lowerCamelCase :Union[str, Any]=0.0_2 , _lowerCamelCase :int="cls_index" , _lowerCamelCase :List[str]=True , _lowerCamelCase :str=None , _lowerCamelCase :List[str]=True , _lowerCamelCase :str=0.1 , **_lowerCamelCase :List[Any] , ):
__SCREAMING_SNAKE_CASE : Tuple = vocab_size
__SCREAMING_SNAKE_CASE : Any = n_positions
__SCREAMING_SNAKE_CASE : Dict = n_embd
__SCREAMING_SNAKE_CASE : Union[str, Any] = n_layer
__SCREAMING_SNAKE_CASE : Union[str, Any] = n_head
__SCREAMING_SNAKE_CASE : Optional[Any] = afn
__SCREAMING_SNAKE_CASE : Tuple = resid_pdrop
__SCREAMING_SNAKE_CASE : int = embd_pdrop
__SCREAMING_SNAKE_CASE : int = attn_pdrop
__SCREAMING_SNAKE_CASE : List[Any] = layer_norm_epsilon
__SCREAMING_SNAKE_CASE : Optional[Any] = initializer_range
__SCREAMING_SNAKE_CASE : List[str] = summary_type
__SCREAMING_SNAKE_CASE : str = summary_use_proj
__SCREAMING_SNAKE_CASE : List[Any] = summary_activation
__SCREAMING_SNAKE_CASE : Tuple = summary_first_dropout
__SCREAMING_SNAKE_CASE : List[Any] = summary_proj_to_labels
super().__init__(**_lowerCamelCase )
| 674 |
"""simple docstring"""
from __future__ import annotations
import unittest
from transformers import is_tf_available, is_torch_available
from transformers.testing_utils import DUMMY_UNKNOWN_IDENTIFIER, SMALL_MODEL_IDENTIFIER, is_pt_tf_cross_test, slow
if is_tf_available():
from transformers import (
AutoConfig,
BertConfig,
GPTaConfig,
TaConfig,
TFAutoModel,
TFAutoModelForCausalLM,
TFAutoModelForMaskedLM,
TFAutoModelForPreTraining,
TFAutoModelForQuestionAnswering,
TFAutoModelForSeqaSeqLM,
TFAutoModelForSequenceClassification,
TFAutoModelWithLMHead,
TFBertForMaskedLM,
TFBertForPreTraining,
TFBertForQuestionAnswering,
TFBertForSequenceClassification,
TFBertModel,
TFGPTaLMHeadModel,
TFRobertaForMaskedLM,
TFTaForConditionalGeneration,
)
from transformers.models.bert.modeling_tf_bert import TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST
from transformers.models.gpta.modeling_tf_gpta import TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST
from transformers.models.ta.modeling_tf_ta import TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST
if is_torch_available():
from transformers import (
AutoModel,
AutoModelForCausalLM,
AutoModelForMaskedLM,
AutoModelForPreTraining,
AutoModelForQuestionAnswering,
AutoModelForSeqaSeqLM,
AutoModelForSequenceClassification,
AutoModelWithLMHead,
BertForMaskedLM,
BertForPreTraining,
BertForQuestionAnswering,
BertForSequenceClassification,
BertModel,
GPTaLMHeadModel,
RobertaForMaskedLM,
TaForConditionalGeneration,
)
@is_pt_tf_cross_test
class snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : List[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = TFAutoModel.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModel.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : List[str] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelForPreTraining.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = AutoModelForPreTraining.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Any ):
for model_name in TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Optional[int] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = TFAutoModelForCausalLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = TFAutoModelForCausalLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelForCausalLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = AutoModelForCausalLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Any = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Union[str, Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = TFAutoModelForMaskedLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[str] = TFAutoModelForMaskedLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelForMaskedLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = AutoModelForMaskedLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :int ):
for model_name in TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : str = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelForSeqaSeqLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : str = TFAutoModelForSeqaSeqLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = AutoModelForSeqaSeqLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = AutoModelForSeqaSeqLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = TFAutoModelForSequenceClassification.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModelForSequenceClassification.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = TFAutoModelForQuestionAnswering.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = AutoModelForQuestionAnswering.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
__SCREAMING_SNAKE_CASE : Union[str, Any] = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
| 674 | 1 |
"""simple docstring"""
import shutil
import tempfile
import unittest
from unittest.mock import patch
from transformers import (
DefaultFlowCallback,
IntervalStrategy,
PrinterCallback,
ProgressCallback,
Trainer,
TrainerCallback,
TrainingArguments,
is_torch_available,
)
from transformers.testing_utils import require_torch
if is_torch_available():
from transformers.trainer import DEFAULT_CALLBACKS
from .test_trainer import RegressionDataset, RegressionModelConfig, RegressionPreTrainedModel
class snake_case ( __UpperCAmelCase ):
def __init__( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Optional[int] = []
def SCREAMING_SNAKE_CASE_ ( self :List[Any] , _lowerCamelCase :List[Any] , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :Union[str, Any] , **_lowerCamelCase :Optional[int] ):
self.events.append('''on_init_end''' )
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :Tuple , _lowerCamelCase :List[Any] , _lowerCamelCase :Dict , **_lowerCamelCase :Dict ):
self.events.append('''on_train_begin''' )
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :Dict , _lowerCamelCase :Optional[int] , _lowerCamelCase :int , **_lowerCamelCase :Any ):
self.events.append('''on_train_end''' )
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :Dict , _lowerCamelCase :List[Any] , **_lowerCamelCase :List[Any] ):
self.events.append('''on_epoch_begin''' )
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Dict , **_lowerCamelCase :List[str] ):
self.events.append('''on_epoch_end''' )
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :Optional[int] , _lowerCamelCase :List[Any] , _lowerCamelCase :Optional[int] , **_lowerCamelCase :List[Any] ):
self.events.append('''on_step_begin''' )
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :List[str] , _lowerCamelCase :int , _lowerCamelCase :Optional[int] , **_lowerCamelCase :str ):
self.events.append('''on_step_end''' )
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :List[Any] , _lowerCamelCase :Optional[int] , **_lowerCamelCase :int ):
self.events.append('''on_evaluate''' )
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :Tuple , _lowerCamelCase :Tuple , _lowerCamelCase :Tuple , **_lowerCamelCase :Dict ):
self.events.append('''on_predict''' )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] , _lowerCamelCase :Tuple , _lowerCamelCase :int , _lowerCamelCase :int , **_lowerCamelCase :str ):
self.events.append('''on_save''' )
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :Dict , _lowerCamelCase :List[Any] , _lowerCamelCase :str , **_lowerCamelCase :int ):
self.events.append('''on_log''' )
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :List[str] , _lowerCamelCase :int , _lowerCamelCase :List[str] , **_lowerCamelCase :List[str] ):
self.events.append('''on_prediction_step''' )
@require_torch
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : Optional[Any] = tempfile.mkdtemp()
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
shutil.rmtree(self.output_dir )
def SCREAMING_SNAKE_CASE_ ( self :List[str] , _lowerCamelCase :Optional[int]=0 , _lowerCamelCase :Optional[int]=0 , _lowerCamelCase :Optional[Any]=6_4 , _lowerCamelCase :Optional[int]=6_4 , _lowerCamelCase :Optional[int]=None , _lowerCamelCase :Dict=False , **_lowerCamelCase :int ):
# disable_tqdm in TrainingArguments has a flaky default since it depends on the level of logging. We make sure
# its set to False since the tests later on depend on its value.
__SCREAMING_SNAKE_CASE : Tuple = RegressionDataset(length=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = RegressionDataset(length=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = RegressionModelConfig(a=_lowerCamelCase , b=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = RegressionPreTrainedModel(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = TrainingArguments(self.output_dir , disable_tqdm=_lowerCamelCase , report_to=[] , **_lowerCamelCase )
return Trainer(
_lowerCamelCase , _lowerCamelCase , train_dataset=_lowerCamelCase , eval_dataset=_lowerCamelCase , callbacks=_lowerCamelCase , )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Tuple , _lowerCamelCase :Optional[int] ):
self.assertEqual(len(_lowerCamelCase ) , len(_lowerCamelCase ) )
# Order doesn't matter
__SCREAMING_SNAKE_CASE : Union[str, Any] = sorted(_lowerCamelCase , key=lambda _lowerCamelCase : cb.__name__ if isinstance(_lowerCamelCase , _lowerCamelCase ) else cb.__class__.__name__ )
__SCREAMING_SNAKE_CASE : List[Any] = sorted(_lowerCamelCase , key=lambda _lowerCamelCase : cb.__name__ if isinstance(_lowerCamelCase , _lowerCamelCase ) else cb.__class__.__name__ )
for cba, cba in zip(_lowerCamelCase , _lowerCamelCase ):
if isinstance(_lowerCamelCase , _lowerCamelCase ) and isinstance(_lowerCamelCase , _lowerCamelCase ):
self.assertEqual(_lowerCamelCase , _lowerCamelCase )
elif isinstance(_lowerCamelCase , _lowerCamelCase ) and not isinstance(_lowerCamelCase , _lowerCamelCase ):
self.assertEqual(_lowerCamelCase , cba.__class__ )
elif not isinstance(_lowerCamelCase , _lowerCamelCase ) and isinstance(_lowerCamelCase , _lowerCamelCase ):
self.assertEqual(cba.__class__ , _lowerCamelCase )
else:
self.assertEqual(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :int ):
__SCREAMING_SNAKE_CASE : Optional[Any] = ['''on_init_end''', '''on_train_begin''']
__SCREAMING_SNAKE_CASE : Dict = 0
__SCREAMING_SNAKE_CASE : str = len(trainer.get_eval_dataloader() )
__SCREAMING_SNAKE_CASE : Tuple = ['''on_prediction_step'''] * len(trainer.get_eval_dataloader() ) + ['''on_log''', '''on_evaluate''']
for _ in range(trainer.state.num_train_epochs ):
expected_events.append('''on_epoch_begin''' )
for _ in range(_lowerCamelCase ):
step += 1
expected_events += ["on_step_begin", "on_step_end"]
if step % trainer.args.logging_steps == 0:
expected_events.append('''on_log''' )
if trainer.args.evaluation_strategy == IntervalStrategy.STEPS and step % trainer.args.eval_steps == 0:
expected_events += evaluation_events.copy()
if step % trainer.args.save_steps == 0:
expected_events.append('''on_save''' )
expected_events.append('''on_epoch_end''' )
if trainer.args.evaluation_strategy == IntervalStrategy.EPOCH:
expected_events += evaluation_events.copy()
expected_events += ["on_log", "on_train_end"]
return expected_events
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE : int = self.get_trainer()
__SCREAMING_SNAKE_CASE : Optional[Any] = DEFAULT_CALLBACKS.copy() + [ProgressCallback]
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
# Callbacks passed at init are added to the default callbacks
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.get_trainer(callbacks=[MyTestTrainerCallback] )
expected_callbacks.append(_lowerCamelCase )
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
# TrainingArguments.disable_tqdm controls if use ProgressCallback or PrinterCallback
__SCREAMING_SNAKE_CASE : Any = self.get_trainer(disable_tqdm=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = DEFAULT_CALLBACKS.copy() + [PrinterCallback]
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[Any] = DEFAULT_CALLBACKS.copy() + [ProgressCallback]
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_trainer()
# We can add, pop, or remove by class name
trainer.remove_callback(_lowerCamelCase )
expected_callbacks.remove(_lowerCamelCase )
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_trainer()
__SCREAMING_SNAKE_CASE : Union[str, Any] = trainer.pop_callback(_lowerCamelCase )
self.assertEqual(cb.__class__ , _lowerCamelCase )
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
trainer.add_callback(_lowerCamelCase )
expected_callbacks.insert(0 , _lowerCamelCase )
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
# We can also add, pop, or remove by instance
__SCREAMING_SNAKE_CASE : Tuple = self.get_trainer()
__SCREAMING_SNAKE_CASE : Optional[Any] = trainer.callback_handler.callbacks[0]
trainer.remove_callback(_lowerCamelCase )
expected_callbacks.remove(_lowerCamelCase )
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = self.get_trainer()
__SCREAMING_SNAKE_CASE : int = trainer.callback_handler.callbacks[0]
__SCREAMING_SNAKE_CASE : Tuple = trainer.pop_callback(_lowerCamelCase )
self.assertEqual(_lowerCamelCase , _lowerCamelCase )
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
trainer.add_callback(_lowerCamelCase )
expected_callbacks.insert(0 , _lowerCamelCase )
self.check_callbacks_equality(trainer.callback_handler.callbacks , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
import warnings
# XXX: for now ignore scatter_gather warnings in this test since it's not relevant to what's being tested
warnings.simplefilter(action='''ignore''' , category=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_trainer(callbacks=[MyTestTrainerCallback] )
trainer.train()
__SCREAMING_SNAKE_CASE : Dict = trainer.callback_handler.callbacks[-2].events
self.assertEqual(_lowerCamelCase , self.get_expected_events(_lowerCamelCase ) )
# Independent log/save/eval
__SCREAMING_SNAKE_CASE : List[str] = self.get_trainer(callbacks=[MyTestTrainerCallback] , logging_steps=5 )
trainer.train()
__SCREAMING_SNAKE_CASE : List[str] = trainer.callback_handler.callbacks[-2].events
self.assertEqual(_lowerCamelCase , self.get_expected_events(_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : int = self.get_trainer(callbacks=[MyTestTrainerCallback] , save_steps=5 )
trainer.train()
__SCREAMING_SNAKE_CASE : List[str] = trainer.callback_handler.callbacks[-2].events
self.assertEqual(_lowerCamelCase , self.get_expected_events(_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_trainer(callbacks=[MyTestTrainerCallback] , eval_steps=5 , evaluation_strategy='''steps''' )
trainer.train()
__SCREAMING_SNAKE_CASE : Union[str, Any] = trainer.callback_handler.callbacks[-2].events
self.assertEqual(_lowerCamelCase , self.get_expected_events(_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Optional[Any] = self.get_trainer(callbacks=[MyTestTrainerCallback] , evaluation_strategy='''epoch''' )
trainer.train()
__SCREAMING_SNAKE_CASE : int = trainer.callback_handler.callbacks[-2].events
self.assertEqual(_lowerCamelCase , self.get_expected_events(_lowerCamelCase ) )
# A bit of everything
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_trainer(
callbacks=[MyTestTrainerCallback] , logging_steps=3 , save_steps=1_0 , eval_steps=5 , evaluation_strategy='''steps''' , )
trainer.train()
__SCREAMING_SNAKE_CASE : str = trainer.callback_handler.callbacks[-2].events
self.assertEqual(_lowerCamelCase , self.get_expected_events(_lowerCamelCase ) )
# warning should be emitted for duplicated callbacks
with patch('''transformers.trainer_callback.logger.warning''' ) as warn_mock:
__SCREAMING_SNAKE_CASE : str = self.get_trainer(
callbacks=[MyTestTrainerCallback, MyTestTrainerCallback] , )
assert str(_lowerCamelCase ) in warn_mock.call_args[0][0]
| 674 |
"""simple docstring"""
import os
import re
import warnings
from shutil import copyfile
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
if TYPE_CHECKING:
from ...tokenization_utils_base import TextInput
from ...utils import logging
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCamelCase = {
'''vocab_file''': {
'''t5-small''': '''https://huggingface.co/t5-small/resolve/main/spiece.model''',
'''t5-base''': '''https://huggingface.co/t5-base/resolve/main/spiece.model''',
'''t5-large''': '''https://huggingface.co/t5-large/resolve/main/spiece.model''',
'''t5-3b''': '''https://huggingface.co/t5-3b/resolve/main/spiece.model''',
'''t5-11b''': '''https://huggingface.co/t5-11b/resolve/main/spiece.model''',
}
}
# TODO(PVP) - this should be removed in Transformers v5
_lowerCamelCase = {
'''t5-small''': 5_12,
'''t5-base''': 5_12,
'''t5-large''': 5_12,
'''t5-3b''': 5_12,
'''t5-11b''': 5_12,
}
_lowerCamelCase = '''▁'''
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = VOCAB_FILES_NAMES
lowerCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
lowerCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
lowerCamelCase__ = ['''input_ids''', '''attention_mask''']
def __init__( self :int , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Union[str, Any]="</s>" , _lowerCamelCase :List[Any]="<unk>" , _lowerCamelCase :Union[str, Any]="<pad>" , _lowerCamelCase :int=1_0_0 , _lowerCamelCase :Union[str, Any]=None , _lowerCamelCase :Optional[Dict[str, Any]] = None , _lowerCamelCase :int=True , **_lowerCamelCase :List[Any] , ):
# Add extra_ids to the special token list
if extra_ids > 0 and additional_special_tokens is None:
__SCREAMING_SNAKE_CASE : Union[str, Any] = [f'''<extra_id_{i}>''' for i in range(_lowerCamelCase )]
elif extra_ids > 0 and additional_special_tokens is not None:
# Check that we have the right number of extra_id special tokens
__SCREAMING_SNAKE_CASE : Optional[int] = len(set(filter(lambda _lowerCamelCase : bool('''extra_id''' in str(_lowerCamelCase ) ) , _lowerCamelCase ) ) )
if extra_tokens != extra_ids:
raise ValueError(
f'''Both extra_ids ({extra_ids}) and additional_special_tokens ({additional_special_tokens}) are'''
''' provided to T5Tokenizer. In this case the additional_special_tokens must include the extra_ids'''
''' tokens''' )
if legacy:
logger.warning_once(
f'''You are using the legacy behaviour of the {self.__class__}. This means that tokens that come after special tokens will not be properly handled. We recommend you to'''
''' read the related pull request available at https://github.com/huggingface/transformers/pull/24565''' )
__SCREAMING_SNAKE_CASE : Optional[Any] = legacy
__SCREAMING_SNAKE_CASE : Optional[int] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
eos_token=_lowerCamelCase , unk_token=_lowerCamelCase , pad_token=_lowerCamelCase , extra_ids=_lowerCamelCase , additional_special_tokens=_lowerCamelCase , sp_model_kwargs=self.sp_model_kwargs , legacy=_lowerCamelCase , **_lowerCamelCase , )
__SCREAMING_SNAKE_CASE : Tuple = vocab_file
__SCREAMING_SNAKE_CASE : List[str] = extra_ids
__SCREAMING_SNAKE_CASE : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(_lowerCamelCase )
@staticmethod
def SCREAMING_SNAKE_CASE_ ( _lowerCamelCase :str , _lowerCamelCase :Union[str, Any] , _lowerCamelCase :int ):
if pretrained_model_name_or_path in TaTokenizer.max_model_input_sizes:
__SCREAMING_SNAKE_CASE : Any = TaTokenizer.max_model_input_sizes[pretrained_model_name_or_path]
if init_max_model_length is not None and init_max_model_length != max_model_length:
return init_max_model_length
elif init_max_model_length is None:
warnings.warn(
'''This tokenizer was incorrectly instantiated with a model max length of'''
f''' {deprecated_max_model_length} which will be corrected in Transformers v5.\nFor now, this'''
''' behavior is kept to avoid breaking backwards compatibility when padding/encoding with'''
''' `truncation is True`.\n- Be aware that you SHOULD NOT rely on'''
f''' {pretrained_model_name_or_path} automatically truncating your input to'''
f''' {deprecated_max_model_length} when padding/encoding.\n- If you want to encode/pad to sequences'''
f''' longer than {deprecated_max_model_length} you can either instantiate this tokenizer with'''
''' `model_max_length` or pass `max_length` when encoding/padding.\n- To avoid this warning, please'''
''' instantiate this tokenizer with `model_max_length` set to your preferred value.''' , _lowerCamelCase , )
return max_model_length
@property
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
return self.sp_model.get_piece_size() + self._extra_ids
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : str = {self.convert_ids_to_tokens(_lowerCamelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :List[int] , _lowerCamelCase :Optional[List[int]] = None , _lowerCamelCase :bool = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_lowerCamelCase , token_ids_a=_lowerCamelCase , already_has_special_tokens=_lowerCamelCase )
# normal case: some special tokens
if token_ids_a is None:
return ([0] * len(_lowerCamelCase )) + [1]
return ([0] * len(_lowerCamelCase )) + [1] + ([0] * len(_lowerCamelCase )) + [1]
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
return list(
set(filter(lambda _lowerCamelCase : bool(re.search(r'''<extra_id_\d+>''' , _lowerCamelCase ) ) is not None , self.additional_special_tokens ) ) )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
return [self._convert_token_to_id(_lowerCamelCase ) for token in self.get_sentinel_tokens()]
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :List[int] ):
if len(_lowerCamelCase ) > 0 and token_ids[-1] == self.eos_token_id:
warnings.warn(
f'''This sequence already has {self.eos_token}. In future versions this behavior may lead to duplicated'''
''' eos tokens being added.''' )
return token_ids
else:
return token_ids + [self.eos_token_id]
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :List[int] , _lowerCamelCase :Optional[List[int]] = None ):
__SCREAMING_SNAKE_CASE : List[str] = [self.eos_token_id]
if token_ids_a is None:
return len(token_ids_a + eos ) * [0]
return len(token_ids_a + eos + token_ids_a + eos ) * [0]
def SCREAMING_SNAKE_CASE_ ( self :List[Any] , _lowerCamelCase :List[int] , _lowerCamelCase :Optional[List[int]] = None ):
__SCREAMING_SNAKE_CASE : Optional[Any] = self._add_eos_if_not_present(_lowerCamelCase )
if token_ids_a is None:
return token_ids_a
else:
__SCREAMING_SNAKE_CASE : Union[str, Any] = self._add_eos_if_not_present(_lowerCamelCase )
return token_ids_a + token_ids_a
def __getstate__( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Any = self.__dict__.copy()
__SCREAMING_SNAKE_CASE : List[str] = None
return state
def __setstate__( self :Optional[Any] , _lowerCamelCase :List[str] ):
__SCREAMING_SNAKE_CASE : Tuple = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs''' ):
__SCREAMING_SNAKE_CASE : Optional[int] = {}
__SCREAMING_SNAKE_CASE : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :"TextInput" , **_lowerCamelCase :str ):
# Replace the SPIECE_UNDERLINE with a space to make sure SPIECE_UNDERLINE is only used at
# the beginning of the text
if not self.legacy:
__SCREAMING_SNAKE_CASE : Dict = SPIECE_UNDERLINE + text.replace(_lowerCamelCase , ''' ''' )
return super().tokenize(_lowerCamelCase , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :List[Any] , **_lowerCamelCase :Dict ):
if not self.legacy:
__SCREAMING_SNAKE_CASE : str = text.startswith(_lowerCamelCase )
if is_first:
__SCREAMING_SNAKE_CASE : str = text[1:]
__SCREAMING_SNAKE_CASE : Tuple = self.sp_model.encode(_lowerCamelCase , out_type=_lowerCamelCase )
if not self.legacy and not is_first and not text.startswith(''' ''' ) and tokens[0].startswith(_lowerCamelCase ):
__SCREAMING_SNAKE_CASE : Optional[int] = ([tokens[0][1:]] if len(tokens[0] ) > 1 else []) + tokens[1:]
return tokens
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Optional[Any] ):
if token.startswith('''<extra_id_''' ):
__SCREAMING_SNAKE_CASE : Tuple = re.match(r'''<extra_id_(\d+)>''' , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = int(match.group(1 ) )
return self.vocab_size - num - 1
return self.sp_model.piece_to_id(_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :Optional[int] ):
if index < self.sp_model.get_piece_size():
__SCREAMING_SNAKE_CASE : List[Any] = self.sp_model.IdToPiece(_lowerCamelCase )
else:
__SCREAMING_SNAKE_CASE : Dict = f'''<extra_id_{self.vocab_size - 1 - index}>'''
return token
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :Any ):
__SCREAMING_SNAKE_CASE : str = []
__SCREAMING_SNAKE_CASE : Dict = ''''''
__SCREAMING_SNAKE_CASE : Dict = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_lowerCamelCase ) + token
__SCREAMING_SNAKE_CASE : List[str] = True
__SCREAMING_SNAKE_CASE : str = []
else:
current_sub_tokens.append(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = False
out_string += self.sp_model.decode(_lowerCamelCase )
return out_string.strip()
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :str , _lowerCamelCase :Optional[str] = None ):
if not os.path.isdir(_lowerCamelCase ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__SCREAMING_SNAKE_CASE : List[str] = os.path.join(
_lowerCamelCase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_lowerCamelCase ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _lowerCamelCase )
elif not os.path.isfile(self.vocab_file ):
with open(_lowerCamelCase , '''wb''' ) as fi:
__SCREAMING_SNAKE_CASE : Any = self.sp_model.serialized_model_proto()
fi.write(_lowerCamelCase )
return (out_vocab_file,)
| 674 | 1 |
"""simple docstring"""
from __future__ import annotations
import unittest
from transformers import is_tf_available, is_torch_available
from transformers.testing_utils import DUMMY_UNKNOWN_IDENTIFIER, SMALL_MODEL_IDENTIFIER, is_pt_tf_cross_test, slow
if is_tf_available():
from transformers import (
AutoConfig,
BertConfig,
GPTaConfig,
TaConfig,
TFAutoModel,
TFAutoModelForCausalLM,
TFAutoModelForMaskedLM,
TFAutoModelForPreTraining,
TFAutoModelForQuestionAnswering,
TFAutoModelForSeqaSeqLM,
TFAutoModelForSequenceClassification,
TFAutoModelWithLMHead,
TFBertForMaskedLM,
TFBertForPreTraining,
TFBertForQuestionAnswering,
TFBertForSequenceClassification,
TFBertModel,
TFGPTaLMHeadModel,
TFRobertaForMaskedLM,
TFTaForConditionalGeneration,
)
from transformers.models.bert.modeling_tf_bert import TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST
from transformers.models.gpta.modeling_tf_gpta import TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST
from transformers.models.ta.modeling_tf_ta import TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST
if is_torch_available():
from transformers import (
AutoModel,
AutoModelForCausalLM,
AutoModelForMaskedLM,
AutoModelForPreTraining,
AutoModelForQuestionAnswering,
AutoModelForSeqaSeqLM,
AutoModelForSequenceClassification,
AutoModelWithLMHead,
BertForMaskedLM,
BertForPreTraining,
BertForQuestionAnswering,
BertForSequenceClassification,
BertModel,
GPTaLMHeadModel,
RobertaForMaskedLM,
TaForConditionalGeneration,
)
@is_pt_tf_cross_test
class snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : List[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = TFAutoModel.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModel.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : List[str] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelForPreTraining.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = AutoModelForPreTraining.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Any ):
for model_name in TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Optional[int] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = TFAutoModelForCausalLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = TFAutoModelForCausalLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelForCausalLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = AutoModelForCausalLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Any = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : Union[str, Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = TFAutoModelForMaskedLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[str] = TFAutoModelForMaskedLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelForMaskedLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = AutoModelForMaskedLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :int ):
for model_name in TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__SCREAMING_SNAKE_CASE : str = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelForSeqaSeqLM.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : str = TFAutoModelForSeqaSeqLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = AutoModelForSeqaSeqLM.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = AutoModelForSeqaSeqLM.from_pretrained(
_lowerCamelCase , output_loading_info=_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = TFAutoModelForSequenceClassification.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoModelForSequenceClassification.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
__SCREAMING_SNAKE_CASE : Optional[Any] = AutoConfig.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = TFAutoModelForQuestionAnswering.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = AutoModelForQuestionAnswering.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Tuple = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
__SCREAMING_SNAKE_CASE : Tuple = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : Optional[Any] = TFAutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_pt=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
__SCREAMING_SNAKE_CASE : Union[str, Any] = AutoModelWithLMHead.from_pretrained(_lowerCamelCase , from_tf=_lowerCamelCase )
self.assertIsInstance(_lowerCamelCase , _lowerCamelCase )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=_lowerCamelCase ) , 1_4_4_1_0 )
| 674 |
"""simple docstring"""
import os
import pytest
import yaml
from datasets.features.features import Features, Value
from datasets.info import DatasetInfo, DatasetInfosDict
@pytest.mark.parametrize(
'''files''' , [
['''full:README.md''', '''dataset_infos.json'''],
['''empty:README.md''', '''dataset_infos.json'''],
['''dataset_infos.json'''],
['''full:README.md'''],
] , )
def lowerCAmelCase_ ( lowercase_ : List[str] , lowercase_ : str ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = tmp_path_factory.mktemp('''dset_infos_dir''' )
if "full:README.md" in files:
with open(dataset_infos_dir / '''README.md''' , '''w''' ) as f:
f.write('''---\ndataset_info:\n dataset_size: 42\n---''' )
if "empty:README.md" in files:
with open(dataset_infos_dir / '''README.md''' , '''w''' ) as f:
f.write('''''' )
# we want to support dataset_infos.json for backward compatibility
if "dataset_infos.json" in files:
with open(dataset_infos_dir / '''dataset_infos.json''' , '''w''' ) as f:
f.write('''{"default": {"dataset_size": 42}}''' )
__SCREAMING_SNAKE_CASE : Optional[Any] = DatasetInfosDict.from_directory(lowercase_ )
assert dataset_infos
assert dataset_infos["default"].dataset_size == 42
@pytest.mark.parametrize(
'''dataset_info''' , [
DatasetInfo(),
DatasetInfo(
description='''foo''' , features=Features({'''a''': Value('''int32''' )} ) , builder_name='''builder''' , config_name='''config''' , version='''1.0.0''' , splits=[{'''name''': '''train'''}] , download_size=42 , ),
] , )
def lowerCAmelCase_ ( lowercase_ : Any , lowercase_ : DatasetInfo ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[str] = str(lowercase_ )
dataset_info.write_to_directory(lowercase_ )
__SCREAMING_SNAKE_CASE : Dict = DatasetInfo.from_directory(lowercase_ )
assert dataset_info == reloaded
assert os.path.exists(os.path.join(lowercase_ , '''dataset_info.json''' ) )
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = DatasetInfo(
description='''foo''' , citation='''bar''' , homepage='''https://foo.bar''' , license='''CC0''' , features=Features({'''a''': Value('''int32''' )} ) , post_processed={} , supervised_keys=() , task_templates=[] , builder_name='''builder''' , config_name='''config''' , version='''1.0.0''' , splits=[{'''name''': '''train''', '''num_examples''': 42}] , download_checksums={} , download_size=1337 , post_processing_size=442 , dataset_size=1234 , size_in_bytes=1337 + 442 + 1234 , )
__SCREAMING_SNAKE_CASE : Optional[int] = dataset_info._to_yaml_dict()
assert sorted(lowercase_ ) == sorted(DatasetInfo._INCLUDED_INFO_IN_YAML )
for key in DatasetInfo._INCLUDED_INFO_IN_YAML:
assert key in dataset_info_yaml_dict
assert isinstance(dataset_info_yaml_dict[key] , (list, dict, int, str) )
__SCREAMING_SNAKE_CASE : int = yaml.safe_dump(lowercase_ )
__SCREAMING_SNAKE_CASE : Union[str, Any] = yaml.safe_load(lowercase_ )
assert dataset_info_yaml_dict == reloaded
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = DatasetInfo()
__SCREAMING_SNAKE_CASE : Union[str, Any] = dataset_info._to_yaml_dict()
assert dataset_info_yaml_dict == {}
@pytest.mark.parametrize(
'''dataset_infos_dict''' , [
DatasetInfosDict(),
DatasetInfosDict({'''default''': DatasetInfo()} ),
DatasetInfosDict({'''my_config_name''': DatasetInfo()} ),
DatasetInfosDict(
{
'''default''': DatasetInfo(
description='''foo''' , features=Features({'''a''': Value('''int32''' )} ) , builder_name='''builder''' , config_name='''config''' , version='''1.0.0''' , splits=[{'''name''': '''train'''}] , download_size=42 , )
} ),
DatasetInfosDict(
{
'''v1''': DatasetInfo(dataset_size=42 ),
'''v2''': DatasetInfo(dataset_size=1337 ),
} ),
] , )
def lowerCAmelCase_ ( lowercase_ : List[Any] , lowercase_ : DatasetInfosDict ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = str(lowercase_ )
dataset_infos_dict.write_to_directory(lowercase_ )
__SCREAMING_SNAKE_CASE : Optional[Any] = DatasetInfosDict.from_directory(lowercase_ )
# the config_name of the dataset_infos_dict take over the attribute
for config_name, dataset_info in dataset_infos_dict.items():
__SCREAMING_SNAKE_CASE : Optional[int] = config_name
# the yaml representation doesn't include fields like description or citation
# so we just test that we can recover what we can from the yaml
__SCREAMING_SNAKE_CASE : Tuple = DatasetInfo._from_yaml_dict(dataset_info._to_yaml_dict() )
assert dataset_infos_dict == reloaded
if dataset_infos_dict:
assert os.path.exists(os.path.join(lowercase_ , '''README.md''' ) )
| 674 | 1 |
"""simple docstring"""
import gc
import random
import unittest
import numpy as np
import torch
from transformers import XLMRobertaTokenizer
from diffusers import (
AltDiffusionImgaImgPipeline,
AutoencoderKL,
PNDMScheduler,
UNetaDConditionModel,
)
from diffusers.image_processor import VaeImageProcessor
from diffusers.pipelines.alt_diffusion.modeling_roberta_series import (
RobertaSeriesConfig,
RobertaSeriesModelWithTransformation,
)
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
enable_full_determinism()
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Any ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
@property
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = 1
__SCREAMING_SNAKE_CASE : Dict = 3
__SCREAMING_SNAKE_CASE : Dict = (3_2, 3_2)
__SCREAMING_SNAKE_CASE : Any = floats_tensor((batch_size, num_channels) + sizes , rng=random.Random(0 ) ).to(_lowerCamelCase )
return image
@property
def SCREAMING_SNAKE_CASE_ ( self :int ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Union[str, Any] = UNetaDConditionModel(
block_out_channels=(3_2, 6_4) , layers_per_block=2 , sample_size=3_2 , in_channels=4 , out_channels=4 , down_block_types=('''DownBlock2D''', '''CrossAttnDownBlock2D''') , up_block_types=('''CrossAttnUpBlock2D''', '''UpBlock2D''') , cross_attention_dim=3_2 , )
return model
@property
def SCREAMING_SNAKE_CASE_ ( self :str ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : int = AutoencoderKL(
block_out_channels=[3_2, 6_4] , in_channels=3 , out_channels=3 , down_block_types=['''DownEncoderBlock2D''', '''DownEncoderBlock2D'''] , up_block_types=['''UpDecoderBlock2D''', '''UpDecoderBlock2D'''] , latent_channels=4 , )
return model
@property
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : int = RobertaSeriesConfig(
hidden_size=3_2 , project_dim=3_2 , intermediate_size=3_7 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=5_0_0_6 , )
return RobertaSeriesModelWithTransformation(_lowerCamelCase )
@property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
def extract(*_lowerCamelCase :str , **_lowerCamelCase :Optional[Any] ):
class snake_case :
def __init__( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.ones([0] )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , _lowerCamelCase :int ):
self.pixel_values.to(_lowerCamelCase )
return self
return Out()
return extract
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Dict = '''cpu''' # ensure determinism for the device-dependent torch.Generator
__SCREAMING_SNAKE_CASE : Any = self.dummy_cond_unet
__SCREAMING_SNAKE_CASE : List[str] = PNDMScheduler(skip_prk_steps=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = self.dummy_vae
__SCREAMING_SNAKE_CASE : Tuple = self.dummy_text_encoder
__SCREAMING_SNAKE_CASE : Optional[Any] = XLMRobertaTokenizer.from_pretrained('''hf-internal-testing/tiny-xlm-roberta''' )
__SCREAMING_SNAKE_CASE : int = 7_7
__SCREAMING_SNAKE_CASE : List[str] = self.dummy_image.to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = init_image / 2 + 0.5
# make sure here that pndm scheduler skips prk
__SCREAMING_SNAKE_CASE : Union[str, Any] = AltDiffusionImgaImgPipeline(
unet=_lowerCamelCase , scheduler=_lowerCamelCase , vae=_lowerCamelCase , text_encoder=_lowerCamelCase , tokenizer=_lowerCamelCase , safety_checker=_lowerCamelCase , feature_extractor=self.dummy_extractor , )
__SCREAMING_SNAKE_CASE : Optional[int] = VaeImageProcessor(vae_scale_factor=alt_pipe.vae_scale_factor , do_normalize=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = alt_pipe.to(_lowerCamelCase )
alt_pipe.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = '''A painting of a squirrel eating a burger'''
__SCREAMING_SNAKE_CASE : List[str] = torch.Generator(device=_lowerCamelCase ).manual_seed(0 )
__SCREAMING_SNAKE_CASE : List[Any] = alt_pipe(
[prompt] , generator=_lowerCamelCase , guidance_scale=6.0 , num_inference_steps=2 , output_type='''np''' , image=_lowerCamelCase , )
__SCREAMING_SNAKE_CASE : Any = output.images
__SCREAMING_SNAKE_CASE : str = torch.Generator(device=_lowerCamelCase ).manual_seed(0 )
__SCREAMING_SNAKE_CASE : Optional[int] = alt_pipe(
[prompt] , generator=_lowerCamelCase , guidance_scale=6.0 , num_inference_steps=2 , output_type='''np''' , image=_lowerCamelCase , return_dict=_lowerCamelCase , )[0]
__SCREAMING_SNAKE_CASE : Dict = image[0, -3:, -3:, -1]
__SCREAMING_SNAKE_CASE : Dict = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 3_2, 3_2, 3)
__SCREAMING_SNAKE_CASE : Tuple = np.array([0.4_4_2_7, 0.3_7_3_1, 0.4_2_4_9, 0.4_9_4_1, 0.4_5_4_6, 0.4_1_4_8, 0.4_1_9_3, 0.4_6_6_6, 0.4_4_9_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5e-3
assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 5e-3
@unittest.skipIf(torch_device != '''cuda''' , '''This test requires a GPU''' )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : str = self.dummy_cond_unet
__SCREAMING_SNAKE_CASE : Optional[Any] = PNDMScheduler(skip_prk_steps=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = self.dummy_vae
__SCREAMING_SNAKE_CASE : str = self.dummy_text_encoder
__SCREAMING_SNAKE_CASE : Dict = XLMRobertaTokenizer.from_pretrained('''hf-internal-testing/tiny-xlm-roberta''' )
__SCREAMING_SNAKE_CASE : List[str] = 7_7
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.dummy_image.to(_lowerCamelCase )
# put models in fp16
__SCREAMING_SNAKE_CASE : Union[str, Any] = unet.half()
__SCREAMING_SNAKE_CASE : str = vae.half()
__SCREAMING_SNAKE_CASE : Dict = bert.half()
# make sure here that pndm scheduler skips prk
__SCREAMING_SNAKE_CASE : Optional[int] = AltDiffusionImgaImgPipeline(
unet=_lowerCamelCase , scheduler=_lowerCamelCase , vae=_lowerCamelCase , text_encoder=_lowerCamelCase , tokenizer=_lowerCamelCase , safety_checker=_lowerCamelCase , feature_extractor=self.dummy_extractor , )
__SCREAMING_SNAKE_CASE : str = VaeImageProcessor(vae_scale_factor=alt_pipe.vae_scale_factor , do_normalize=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = alt_pipe.to(_lowerCamelCase )
alt_pipe.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = '''A painting of a squirrel eating a burger'''
__SCREAMING_SNAKE_CASE : Any = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Any = alt_pipe(
[prompt] , generator=_lowerCamelCase , num_inference_steps=2 , output_type='''np''' , image=_lowerCamelCase , ).images
assert image.shape == (1, 3_2, 3_2, 3)
@unittest.skipIf(torch_device != '''cuda''' , '''This test requires a GPU''' )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : int = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/img2img/sketch-mountains-input.jpg''' )
# resize to resolution that is divisible by 8 but not 16 or 32
__SCREAMING_SNAKE_CASE : Any = init_image.resize((7_6_0, 5_0_4) )
__SCREAMING_SNAKE_CASE : List[Any] = '''BAAI/AltDiffusion'''
__SCREAMING_SNAKE_CASE : Optional[Any] = AltDiffusionImgaImgPipeline.from_pretrained(
_lowerCamelCase , safety_checker=_lowerCamelCase , )
pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
pipe.enable_attention_slicing()
__SCREAMING_SNAKE_CASE : Optional[int] = '''A fantasy landscape, trending on artstation'''
__SCREAMING_SNAKE_CASE : Dict = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : List[Any] = pipe(
prompt=_lowerCamelCase , image=_lowerCamelCase , strength=0.7_5 , guidance_scale=7.5 , generator=_lowerCamelCase , output_type='''np''' , )
__SCREAMING_SNAKE_CASE : str = output.images[0]
__SCREAMING_SNAKE_CASE : Union[str, Any] = image[2_5_5:2_5_8, 3_8_3:3_8_6, -1]
assert image.shape == (5_0_4, 7_6_0, 3)
__SCREAMING_SNAKE_CASE : str = np.array([0.9_3_5_8, 0.9_3_9_7, 0.9_5_9_9, 0.9_9_0_1, 1.0_0_0_0, 1.0_0_0_0, 0.9_8_8_2, 1.0_0_0_0, 1.0_0_0_0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@slow
@require_torch_gpu
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :int ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Any = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/img2img/sketch-mountains-input.jpg''' )
__SCREAMING_SNAKE_CASE : List[str] = init_image.resize((7_6_8, 5_1_2) )
__SCREAMING_SNAKE_CASE : Any = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/img2img/fantasy_landscape_alt.npy''' )
__SCREAMING_SNAKE_CASE : List[str] = '''BAAI/AltDiffusion'''
__SCREAMING_SNAKE_CASE : Dict = AltDiffusionImgaImgPipeline.from_pretrained(
_lowerCamelCase , safety_checker=_lowerCamelCase , )
pipe.to(_lowerCamelCase )
pipe.set_progress_bar_config(disable=_lowerCamelCase )
pipe.enable_attention_slicing()
__SCREAMING_SNAKE_CASE : List[str] = '''A fantasy landscape, trending on artstation'''
__SCREAMING_SNAKE_CASE : Optional[int] = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Optional[Any] = pipe(
prompt=_lowerCamelCase , image=_lowerCamelCase , strength=0.7_5 , guidance_scale=7.5 , generator=_lowerCamelCase , output_type='''np''' , )
__SCREAMING_SNAKE_CASE : List[str] = output.images[0]
assert image.shape == (5_1_2, 7_6_8, 3)
# img2img is flaky across GPUs even in fp32, so using MAE here
assert np.abs(expected_image - image ).max() < 1e-2
| 674 |
"""simple docstring"""
from typing import Optional, Union
import torch
from torch import nn
from ...configuration_utils import ConfigMixin, register_to_config
from ...models.modeling_utils import ModelMixin
class snake_case ( __UpperCAmelCase , __UpperCAmelCase ):
@register_to_config
def __init__( self :List[str] , _lowerCamelCase :int = 7_6_8 , ):
super().__init__()
__SCREAMING_SNAKE_CASE : Union[str, Any] = nn.Parameter(torch.zeros(1 , _lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : List[Any] = nn.Parameter(torch.ones(1 , _lowerCamelCase ) )
def SCREAMING_SNAKE_CASE_ ( self :Any , _lowerCamelCase :Optional[Union[str, torch.device]] = None , _lowerCamelCase :Optional[torch.dtype] = None , ):
__SCREAMING_SNAKE_CASE : Optional[int] = nn.Parameter(self.mean.to(_lowerCamelCase ).to(_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Tuple = nn.Parameter(self.std.to(_lowerCamelCase ).to(_lowerCamelCase ) )
return self
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :Tuple ):
__SCREAMING_SNAKE_CASE : Tuple = (embeds - self.mean) * 1.0 / self.std
return embeds
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[Any] = (embeds * self.std) + self.mean
return embeds
| 674 | 1 |
"""simple docstring"""
import argparse
import json
from pathlib import Path
import requests
import torch
from huggingface_hub import cached_download, hf_hub_download, hf_hub_url
from PIL import Image
from transformers import DetaConfig, DetaForObjectDetection, DetaImageProcessor, SwinConfig
from transformers.utils import logging
logging.set_verbosity_info()
_lowerCamelCase = logging.get_logger(__name__)
def lowerCAmelCase_ ( lowercase_ : List[Any] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : int = SwinConfig(
embed_dim=192 , depths=(2, 2, 18, 2) , num_heads=(6, 12, 24, 48) , window_size=12 , out_features=['''stage2''', '''stage3''', '''stage4'''] , )
__SCREAMING_SNAKE_CASE : Dict = DetaConfig(
backbone_config=lowercase_ , num_queries=900 , encoder_ffn_dim=2048 , decoder_ffn_dim=2048 , num_feature_levels=5 , assign_first_stage=lowercase_ , with_box_refine=lowercase_ , two_stage=lowercase_ , )
# set labels
__SCREAMING_SNAKE_CASE : Any = '''huggingface/label-files'''
if "o365" in model_name:
__SCREAMING_SNAKE_CASE : int = 366
__SCREAMING_SNAKE_CASE : Any = '''object365-id2label.json'''
else:
__SCREAMING_SNAKE_CASE : List[str] = 91
__SCREAMING_SNAKE_CASE : str = '''coco-detection-id2label.json'''
__SCREAMING_SNAKE_CASE : Dict = num_labels
__SCREAMING_SNAKE_CASE : Union[str, Any] = json.load(open(cached_download(hf_hub_url(lowercase_ , lowercase_ , repo_type='''dataset''' ) ) , '''r''' ) )
__SCREAMING_SNAKE_CASE : Optional[Any] = {int(lowercase_ ): v for k, v in idalabel.items()}
__SCREAMING_SNAKE_CASE : Optional[int] = idalabel
__SCREAMING_SNAKE_CASE : str = {v: k for k, v in idalabel.items()}
return config
def lowerCAmelCase_ ( lowercase_ : Any ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[Any] = []
# stem
# fmt: off
rename_keys.append(('''backbone.0.body.patch_embed.proj.weight''', '''model.backbone.model.embeddings.patch_embeddings.projection.weight''') )
rename_keys.append(('''backbone.0.body.patch_embed.proj.bias''', '''model.backbone.model.embeddings.patch_embeddings.projection.bias''') )
rename_keys.append(('''backbone.0.body.patch_embed.norm.weight''', '''model.backbone.model.embeddings.norm.weight''') )
rename_keys.append(('''backbone.0.body.patch_embed.norm.bias''', '''model.backbone.model.embeddings.norm.bias''') )
# stages
for i in range(len(config.backbone_config.depths ) ):
for j in range(config.backbone_config.depths[i] ):
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.norm1.weight''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.layernorm_before.weight''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.norm1.bias''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.layernorm_before.bias''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.attn.relative_position_bias_table''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.attention.self.relative_position_bias_table''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.attn.relative_position_index''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.attention.self.relative_position_index''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.attn.proj.weight''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.attention.output.dense.weight''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.attn.proj.bias''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.attention.output.dense.bias''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.norm2.weight''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.layernorm_after.weight''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.norm2.bias''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.layernorm_after.bias''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.mlp.fc1.weight''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.intermediate.dense.weight''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.mlp.fc1.bias''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.intermediate.dense.bias''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.mlp.fc2.weight''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.output.dense.weight''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.blocks.{j}.mlp.fc2.bias''', F'''model.backbone.model.encoder.layers.{i}.blocks.{j}.output.dense.bias''') )
if i < 3:
rename_keys.append((F'''backbone.0.body.layers.{i}.downsample.reduction.weight''', F'''model.backbone.model.encoder.layers.{i}.downsample.reduction.weight''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.downsample.norm.weight''', F'''model.backbone.model.encoder.layers.{i}.downsample.norm.weight''') )
rename_keys.append((F'''backbone.0.body.layers.{i}.downsample.norm.bias''', F'''model.backbone.model.encoder.layers.{i}.downsample.norm.bias''') )
rename_keys.append(('''backbone.0.body.norm1.weight''', '''model.backbone.model.hidden_states_norms.stage2.weight''') )
rename_keys.append(('''backbone.0.body.norm1.bias''', '''model.backbone.model.hidden_states_norms.stage2.bias''') )
rename_keys.append(('''backbone.0.body.norm2.weight''', '''model.backbone.model.hidden_states_norms.stage3.weight''') )
rename_keys.append(('''backbone.0.body.norm2.bias''', '''model.backbone.model.hidden_states_norms.stage3.bias''') )
rename_keys.append(('''backbone.0.body.norm3.weight''', '''model.backbone.model.hidden_states_norms.stage4.weight''') )
rename_keys.append(('''backbone.0.body.norm3.bias''', '''model.backbone.model.hidden_states_norms.stage4.bias''') )
# transformer encoder
for i in range(config.encoder_layers ):
rename_keys.append((F'''transformer.encoder.layers.{i}.self_attn.sampling_offsets.weight''', F'''model.encoder.layers.{i}.self_attn.sampling_offsets.weight''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.self_attn.sampling_offsets.bias''', F'''model.encoder.layers.{i}.self_attn.sampling_offsets.bias''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.self_attn.attention_weights.weight''', F'''model.encoder.layers.{i}.self_attn.attention_weights.weight''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.self_attn.attention_weights.bias''', F'''model.encoder.layers.{i}.self_attn.attention_weights.bias''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.self_attn.value_proj.weight''', F'''model.encoder.layers.{i}.self_attn.value_proj.weight''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.self_attn.value_proj.bias''', F'''model.encoder.layers.{i}.self_attn.value_proj.bias''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.self_attn.output_proj.weight''', F'''model.encoder.layers.{i}.self_attn.output_proj.weight''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.self_attn.output_proj.bias''', F'''model.encoder.layers.{i}.self_attn.output_proj.bias''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.norm1.weight''', F'''model.encoder.layers.{i}.self_attn_layer_norm.weight''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.norm1.bias''', F'''model.encoder.layers.{i}.self_attn_layer_norm.bias''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.linear1.weight''', F'''model.encoder.layers.{i}.fc1.weight''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.linear1.bias''', F'''model.encoder.layers.{i}.fc1.bias''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.linear2.weight''', F'''model.encoder.layers.{i}.fc2.weight''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.linear2.bias''', F'''model.encoder.layers.{i}.fc2.bias''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.norm2.weight''', F'''model.encoder.layers.{i}.final_layer_norm.weight''') )
rename_keys.append((F'''transformer.encoder.layers.{i}.norm2.bias''', F'''model.encoder.layers.{i}.final_layer_norm.bias''') )
# transformer decoder
for i in range(config.decoder_layers ):
rename_keys.append((F'''transformer.decoder.layers.{i}.cross_attn.sampling_offsets.weight''', F'''model.decoder.layers.{i}.encoder_attn.sampling_offsets.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.cross_attn.sampling_offsets.bias''', F'''model.decoder.layers.{i}.encoder_attn.sampling_offsets.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.cross_attn.attention_weights.weight''', F'''model.decoder.layers.{i}.encoder_attn.attention_weights.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.cross_attn.attention_weights.bias''', F'''model.decoder.layers.{i}.encoder_attn.attention_weights.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.cross_attn.value_proj.weight''', F'''model.decoder.layers.{i}.encoder_attn.value_proj.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.cross_attn.value_proj.bias''', F'''model.decoder.layers.{i}.encoder_attn.value_proj.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.cross_attn.output_proj.weight''', F'''model.decoder.layers.{i}.encoder_attn.output_proj.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.cross_attn.output_proj.bias''', F'''model.decoder.layers.{i}.encoder_attn.output_proj.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.norm1.weight''', F'''model.decoder.layers.{i}.encoder_attn_layer_norm.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.norm1.bias''', F'''model.decoder.layers.{i}.encoder_attn_layer_norm.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.self_attn.out_proj.weight''', F'''model.decoder.layers.{i}.self_attn.out_proj.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.self_attn.out_proj.bias''', F'''model.decoder.layers.{i}.self_attn.out_proj.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.norm2.weight''', F'''model.decoder.layers.{i}.self_attn_layer_norm.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.norm2.bias''', F'''model.decoder.layers.{i}.self_attn_layer_norm.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.linear1.weight''', F'''model.decoder.layers.{i}.fc1.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.linear1.bias''', F'''model.decoder.layers.{i}.fc1.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.linear2.weight''', F'''model.decoder.layers.{i}.fc2.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.linear2.bias''', F'''model.decoder.layers.{i}.fc2.bias''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.norm3.weight''', F'''model.decoder.layers.{i}.final_layer_norm.weight''') )
rename_keys.append((F'''transformer.decoder.layers.{i}.norm3.bias''', F'''model.decoder.layers.{i}.final_layer_norm.bias''') )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( lowercase_ : List[Any] , lowercase_ : str , lowercase_ : Tuple ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[Any] = dct.pop(lowercase_ )
__SCREAMING_SNAKE_CASE : Tuple = val
def lowerCAmelCase_ ( lowercase_ : Dict , lowercase_ : Optional[int] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[str] = [int(backbone_config.embed_dim * 2**i ) for i in range(len(backbone_config.depths ) )]
for i in range(len(backbone_config.depths ) ):
__SCREAMING_SNAKE_CASE : str = num_features[i]
for j in range(backbone_config.depths[i] ):
# fmt: off
# read in weights + bias of input projection layer (in original implementation, this is a single matrix + bias)
__SCREAMING_SNAKE_CASE : List[str] = state_dict.pop(F'''backbone.0.body.layers.{i}.blocks.{j}.attn.qkv.weight''' )
__SCREAMING_SNAKE_CASE : int = state_dict.pop(F'''backbone.0.body.layers.{i}.blocks.{j}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
__SCREAMING_SNAKE_CASE : str = in_proj_weight[:dim, :]
__SCREAMING_SNAKE_CASE : int = in_proj_bias[: dim]
__SCREAMING_SNAKE_CASE : Union[str, Any] = in_proj_weight[
dim : dim * 2, :
]
__SCREAMING_SNAKE_CASE : Optional[Any] = in_proj_bias[
dim : dim * 2
]
__SCREAMING_SNAKE_CASE : Dict = in_proj_weight[
-dim :, :
]
__SCREAMING_SNAKE_CASE : List[str] = in_proj_bias[-dim :]
# fmt: on
def lowerCAmelCase_ ( lowercase_ : Optional[Any] , lowercase_ : Optional[Any] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = config.d_model
for i in range(config.decoder_layers ):
# read in weights + bias of input projection layer of self-attention
__SCREAMING_SNAKE_CASE : int = state_dict.pop(F'''transformer.decoder.layers.{i}.self_attn.in_proj_weight''' )
__SCREAMING_SNAKE_CASE : Union[str, Any] = state_dict.pop(F'''transformer.decoder.layers.{i}.self_attn.in_proj_bias''' )
# next, add query, keys and values (in that order) to the state dict
__SCREAMING_SNAKE_CASE : str = in_proj_weight[:hidden_size, :]
__SCREAMING_SNAKE_CASE : List[Any] = in_proj_bias[:hidden_size]
__SCREAMING_SNAKE_CASE : Optional[int] = in_proj_weight[
hidden_size : hidden_size * 2, :
]
__SCREAMING_SNAKE_CASE : Union[str, Any] = in_proj_bias[hidden_size : hidden_size * 2]
__SCREAMING_SNAKE_CASE : Optional[int] = in_proj_weight[-hidden_size:, :]
__SCREAMING_SNAKE_CASE : Tuple = in_proj_bias[-hidden_size:]
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
__SCREAMING_SNAKE_CASE : Optional[Any] = Image.open(requests.get(lowercase_ , stream=lowercase_ ).raw )
return im
@torch.no_grad()
def lowerCAmelCase_ ( lowercase_ : Any , lowercase_ : Tuple , lowercase_ : List[Any] ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = get_deta_config(lowercase_ )
# load original state dict
if model_name == "deta-swin-large":
__SCREAMING_SNAKE_CASE : Union[str, Any] = hf_hub_download(repo_id='''nielsr/deta-checkpoints''' , filename='''adet_swin_ft.pth''' )
elif model_name == "deta-swin-large-o365":
__SCREAMING_SNAKE_CASE : Any = hf_hub_download(repo_id='''jozhang97/deta-swin-l-o365''' , filename='''deta_swin_pt_o365.pth''' )
else:
raise ValueError(F'''Model name {model_name} not supported''' )
__SCREAMING_SNAKE_CASE : List[Any] = torch.load(lowercase_ , map_location='''cpu''' )['''model''']
# original state dict
for name, param in state_dict.items():
print(lowercase_ , param.shape )
# rename keys
__SCREAMING_SNAKE_CASE : List[Any] = create_rename_keys(lowercase_ )
for src, dest in rename_keys:
rename_key(lowercase_ , lowercase_ , lowercase_ )
read_in_swin_q_k_v(lowercase_ , config.backbone_config )
read_in_decoder_q_k_v(lowercase_ , lowercase_ )
# fix some prefixes
for key in state_dict.copy().keys():
if "transformer.decoder.class_embed" in key or "transformer.decoder.bbox_embed" in key:
__SCREAMING_SNAKE_CASE : str = state_dict.pop(lowercase_ )
__SCREAMING_SNAKE_CASE : Optional[Any] = val
if "input_proj" in key:
__SCREAMING_SNAKE_CASE : Any = state_dict.pop(lowercase_ )
__SCREAMING_SNAKE_CASE : Union[str, Any] = val
if "level_embed" in key or "pos_trans" in key or "pix_trans" in key or "enc_output" in key:
__SCREAMING_SNAKE_CASE : Optional[int] = state_dict.pop(lowercase_ )
__SCREAMING_SNAKE_CASE : Dict = val
# finally, create HuggingFace model and load state dict
__SCREAMING_SNAKE_CASE : int = DetaForObjectDetection(lowercase_ )
model.load_state_dict(lowercase_ )
model.eval()
__SCREAMING_SNAKE_CASE : Optional[int] = '''cuda''' if torch.cuda.is_available() else '''cpu'''
model.to(lowercase_ )
# load image processor
__SCREAMING_SNAKE_CASE : int = DetaImageProcessor(format='''coco_detection''' )
# verify our conversion on image
__SCREAMING_SNAKE_CASE : Tuple = prepare_img()
__SCREAMING_SNAKE_CASE : Optional[int] = processor(images=lowercase_ , return_tensors='''pt''' )
__SCREAMING_SNAKE_CASE : Optional[int] = encoding['''pixel_values''']
__SCREAMING_SNAKE_CASE : Dict = model(pixel_values.to(lowercase_ ) )
# verify logits
print('''Logits:''' , outputs.logits[0, :3, :3] )
print('''Boxes:''' , outputs.pred_boxes[0, :3, :3] )
if model_name == "deta-swin-large":
__SCREAMING_SNAKE_CASE : int = torch.tensor(
[[-7.6_308, -2.8_485, -5.3_737], [-7.2_037, -4.5_505, -4.8_027], [-7.2_943, -4.2_611, -4.6_617]] )
__SCREAMING_SNAKE_CASE : Tuple = torch.tensor([[0.4_987, 0.4_969, 0.9_999], [0.2_549, 0.5_498, 0.4_805], [0.5_498, 0.2_757, 0.0_569]] )
elif model_name == "deta-swin-large-o365":
__SCREAMING_SNAKE_CASE : str = torch.tensor(
[[-8.0_122, -3.5_720, -4.9_717], [-8.1_547, -3.6_886, -4.6_389], [-7.6_610, -3.6_194, -5.0_134]] )
__SCREAMING_SNAKE_CASE : Any = torch.tensor([[0.2_523, 0.5_549, 0.4_881], [0.7_715, 0.4_149, 0.4_601], [0.5_503, 0.2_753, 0.0_575]] )
assert torch.allclose(outputs.logits[0, :3, :3] , expected_logits.to(lowercase_ ) , atol=1E-4 )
assert torch.allclose(outputs.pred_boxes[0, :3, :3] , expected_boxes.to(lowercase_ ) , atol=1E-4 )
print('''Everything ok!''' )
if pytorch_dump_folder_path:
# Save model and processor
logger.info(F'''Saving PyTorch model and processor to {pytorch_dump_folder_path}...''' )
Path(lowercase_ ).mkdir(exist_ok=lowercase_ )
model.save_pretrained(lowercase_ )
processor.save_pretrained(lowercase_ )
# Push to hub
if push_to_hub:
print('''Pushing model and processor to hub...''' )
model.push_to_hub(F'''jozhang97/{model_name}''' )
processor.push_to_hub(F'''jozhang97/{model_name}''' )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
parser.add_argument(
'''--model_name''',
type=str,
default='''deta-swin-large''',
choices=['''deta-swin-large''', '''deta-swin-large-o365'''],
help='''Name of the model you\'d like to convert.''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''',
default=None,
type=str,
help='''Path to the folder to output PyTorch model.''',
)
parser.add_argument(
'''--push_to_hub''', action='''store_true''', help='''Whether or not to push the converted model to the 🤗 hub.'''
)
_lowerCamelCase = parser.parse_args()
convert_deta_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 674 |
"""simple docstring"""
import argparse
import torch
from transformers import BertConfig, BertForPreTraining, load_tf_weights_in_bert
from transformers.utils import logging
logging.set_verbosity_info()
def lowerCAmelCase_ ( lowercase_ : Dict , lowercase_ : int , lowercase_ : Any ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Tuple = BertConfig.from_json_file(lowercase_ )
print(F'''Building PyTorch model from configuration: {config}''' )
__SCREAMING_SNAKE_CASE : Optional[int] = BertForPreTraining(lowercase_ )
# Load weights from tf checkpoint
load_tf_weights_in_bert(lowercase_ , lowercase_ , lowercase_ )
# Save pytorch-model
print(F'''Save PyTorch model to {pytorch_dump_path}''' )
torch.save(model.state_dict() , lowercase_ )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--tf_checkpoint_path''', default=None, type=str, required=True, help='''Path to the TensorFlow checkpoint path.'''
)
parser.add_argument(
'''--bert_config_file''',
default=None,
type=str,
required=True,
help=(
'''The config json file corresponding to the pre-trained BERT model. \n'''
'''This specifies the model architecture.'''
),
)
parser.add_argument(
'''--pytorch_dump_path''', default=None, type=str, required=True, help='''Path to the output PyTorch model.'''
)
_lowerCamelCase = parser.parse_args()
convert_tf_checkpoint_to_pytorch(args.tf_checkpoint_path, args.bert_config_file, args.pytorch_dump_path)
| 674 | 1 |
"""simple docstring"""
from __future__ import annotations
def lowerCAmelCase_ ( lowercase_ : str , lowercase_ : list[str] | None = None ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Dict = word_bank or []
# create a table
__SCREAMING_SNAKE_CASE : int = len(lowercase_ ) + 1
__SCREAMING_SNAKE_CASE : list[list[list[str]]] = []
for _ in range(lowercase_ ):
table.append([] )
# seed value
__SCREAMING_SNAKE_CASE : Optional[Any] = [[]] # because empty string has empty combination
# iterate through the indices
for i in range(lowercase_ ):
# condition
if table[i] != []:
for word in word_bank:
# slice condition
if target[i : i + len(lowercase_ )] == word:
__SCREAMING_SNAKE_CASE : list[list[str]] = [
[word, *way] for way in table[i]
]
# adds the word to every combination the current position holds
# now,push that combination to the table[i+len(word)]
table[i + len(lowercase_ )] += new_combinations
# combinations are in reverse order so reverse for better output
for combination in table[len(lowercase_ )]:
combination.reverse()
return table[len(lowercase_ )]
if __name__ == "__main__":
print(all_construct('''jwajalapa''', ['''jwa''', '''j''', '''w''', '''a''', '''la''', '''lapa''']))
print(all_construct('''rajamati''', ['''s''', '''raj''', '''amat''', '''raja''', '''ma''', '''i''', '''t''']))
print(
all_construct(
'''hexagonosaurus''',
['''h''', '''ex''', '''hex''', '''ag''', '''ago''', '''ru''', '''auru''', '''rus''', '''go''', '''no''', '''o''', '''s'''],
)
)
| 674 |
"""simple docstring"""
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import numpy as np
import torch
from ..models.clipseg import CLIPSegForImageSegmentation
from ..utils import is_vision_available, requires_backends
from .base import PipelineTool
if is_vision_available():
from PIL import Image
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = (
'''This is a tool that creates a segmentation mask of an image according to a label. It cannot create an image.'''
'''It takes two arguments named `image` which should be the original image, and `label` which should be a text '''
'''describing the elements what should be identified in the segmentation mask. The tool returns the mask.'''
)
lowerCamelCase__ = '''CIDAS/clipseg-rd64-refined'''
lowerCamelCase__ = '''image_segmenter'''
lowerCamelCase__ = CLIPSegForImageSegmentation
lowerCamelCase__ = ['''image''', '''text''']
lowerCamelCase__ = ['''image''']
def __init__( self :Dict , *_lowerCamelCase :Union[str, Any] , **_lowerCamelCase :Tuple ):
requires_backends(self , ['''vision'''] )
super().__init__(*_lowerCamelCase , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :"Image" , _lowerCamelCase :str ):
return self.pre_processor(text=[label] , images=[image] , padding=_lowerCamelCase , return_tensors='''pt''' )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] , _lowerCamelCase :Optional[int] ):
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[Any] = self.model(**_lowerCamelCase ).logits
return logits
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Tuple ):
__SCREAMING_SNAKE_CASE : Optional[int] = outputs.cpu().detach().numpy()
__SCREAMING_SNAKE_CASE : str = 0
__SCREAMING_SNAKE_CASE : str = 1
return Image.fromarray((array * 2_5_5).astype(np.uinta ) )
| 674 | 1 |
"""simple docstring"""
import argparse
import os
import shutil
import torch
from emmental.modules import MagnitudeBinarizer, ThresholdBinarizer, TopKBinarizer
def lowerCAmelCase_ ( lowercase_ : Tuple ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : int = args.pruning_method
__SCREAMING_SNAKE_CASE : Dict = args.threshold
__SCREAMING_SNAKE_CASE : str = args.model_name_or_path.rstrip('''/''' )
__SCREAMING_SNAKE_CASE : List[Any] = args.target_model_path
print(F'''Load fine-pruned model from {model_name_or_path}''' )
__SCREAMING_SNAKE_CASE : Optional[int] = torch.load(os.path.join(lowercase_ , '''pytorch_model.bin''' ) )
__SCREAMING_SNAKE_CASE : List[Any] = {}
for name, tensor in model.items():
if "embeddings" in name or "LayerNorm" in name or "pooler" in name:
__SCREAMING_SNAKE_CASE : List[Any] = tensor
print(F'''Copied layer {name}''' )
elif "classifier" in name or "qa_output" in name:
__SCREAMING_SNAKE_CASE : Tuple = tensor
print(F'''Copied layer {name}''' )
elif "bias" in name:
__SCREAMING_SNAKE_CASE : str = tensor
print(F'''Copied layer {name}''' )
else:
if pruning_method == "magnitude":
__SCREAMING_SNAKE_CASE : Tuple = MagnitudeBinarizer.apply(inputs=lowercase_ , threshold=lowercase_ )
__SCREAMING_SNAKE_CASE : Any = tensor * mask
print(F'''Pruned layer {name}''' )
elif pruning_method == "topK":
if "mask_scores" in name:
continue
__SCREAMING_SNAKE_CASE : List[str] = name[:-6]
__SCREAMING_SNAKE_CASE : str = model[F'''{prefix_}mask_scores''']
__SCREAMING_SNAKE_CASE : List[str] = TopKBinarizer.apply(lowercase_ , lowercase_ )
__SCREAMING_SNAKE_CASE : Optional[Any] = tensor * mask
print(F'''Pruned layer {name}''' )
elif pruning_method == "sigmoied_threshold":
if "mask_scores" in name:
continue
__SCREAMING_SNAKE_CASE : Optional[Any] = name[:-6]
__SCREAMING_SNAKE_CASE : Optional[Any] = model[F'''{prefix_}mask_scores''']
__SCREAMING_SNAKE_CASE : Optional[int] = ThresholdBinarizer.apply(lowercase_ , lowercase_ , lowercase_ )
__SCREAMING_SNAKE_CASE : Optional[int] = tensor * mask
print(F'''Pruned layer {name}''' )
elif pruning_method == "l0":
if "mask_scores" in name:
continue
__SCREAMING_SNAKE_CASE : Tuple = name[:-6]
__SCREAMING_SNAKE_CASE : int = model[F'''{prefix_}mask_scores''']
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Union[str, Any] = -0.1, 1.1
__SCREAMING_SNAKE_CASE : List[str] = torch.sigmoid(lowercase_ )
__SCREAMING_SNAKE_CASE : Optional[Any] = s * (r - l) + l
__SCREAMING_SNAKE_CASE : Dict = s_bar.clamp(min=0.0 , max=1.0 )
__SCREAMING_SNAKE_CASE : Optional[Any] = tensor * mask
print(F'''Pruned layer {name}''' )
else:
raise ValueError('''Unknown pruning method''' )
if target_model_path is None:
__SCREAMING_SNAKE_CASE : Union[str, Any] = os.path.join(
os.path.dirname(lowercase_ ) , F'''bertarized_{os.path.basename(lowercase_ )}''' )
if not os.path.isdir(lowercase_ ):
shutil.copytree(lowercase_ , lowercase_ )
print(F'''\nCreated folder {target_model_path}''' )
torch.save(lowercase_ , os.path.join(lowercase_ , '''pytorch_model.bin''' ) )
print('''\nPruned model saved! See you later!''' )
if __name__ == "__main__":
_lowerCamelCase = argparse.ArgumentParser()
parser.add_argument(
'''--pruning_method''',
choices=['''l0''', '''magnitude''', '''topK''', '''sigmoied_threshold'''],
type=str,
required=True,
help=(
'''Pruning Method (l0 = L0 regularization, magnitude = Magnitude pruning, topK = Movement pruning,'''
''' sigmoied_threshold = Soft movement pruning)'''
),
)
parser.add_argument(
'''--threshold''',
type=float,
required=False,
help=(
'''For `magnitude` and `topK`, it is the level of remaining weights (in %) in the fine-pruned model.'''
'''For `sigmoied_threshold`, it is the threshold \tau against which the (sigmoied) scores are compared.'''
'''Not needed for `l0`'''
),
)
parser.add_argument(
'''--model_name_or_path''',
type=str,
required=True,
help='''Folder containing the model that was previously fine-pruned''',
)
parser.add_argument(
'''--target_model_path''',
default=None,
type=str,
required=False,
help='''Folder containing the model that was previously fine-pruned''',
)
_lowerCamelCase = parser.parse_args()
main(args)
| 674 |
"""simple docstring"""
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
from transformers import BertTokenizerFast
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES, BertTokenizer
from transformers.testing_utils import require_tokenizers, require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import VisionTextDualEncoderProcessor, ViTImageProcessor
@require_tokenizers
@require_vision
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = tempfile.mkdtemp()
# fmt: off
__SCREAMING_SNAKE_CASE : Optional[int] = ['''[UNK]''', '''[CLS]''', '''[SEP]''', '''[PAD]''', '''[MASK]''', '''want''', '''##want''', '''##ed''', '''wa''', '''un''', '''runn''', '''##ing''', ''',''', '''low''', '''lowest''']
# fmt: on
__SCREAMING_SNAKE_CASE : List[Any] = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
__SCREAMING_SNAKE_CASE : Optional[int] = {
'''do_resize''': True,
'''size''': {'''height''': 1_8, '''width''': 1_8},
'''do_normalize''': True,
'''image_mean''': [0.5, 0.5, 0.5],
'''image_std''': [0.5, 0.5, 0.5],
}
__SCREAMING_SNAKE_CASE : List[Any] = os.path.join(self.tmpdirname , _lowerCamelCase )
with open(self.image_processor_file , '''w''' , encoding='''utf-8''' ) as fp:
json.dump(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] , **_lowerCamelCase :List[str] ):
return BertTokenizer.from_pretrained(self.tmpdirname , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] , **_lowerCamelCase :Optional[int] ):
return ViTImageProcessor.from_pretrained(self.tmpdirname , **_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
shutil.rmtree(self.tmpdirname )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = [np.random.randint(2_5_5 , size=(3, 3_0, 4_0_0) , dtype=np.uinta )]
__SCREAMING_SNAKE_CASE : Tuple = [Image.fromarray(np.moveaxis(_lowerCamelCase , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Optional[Any] = self.get_image_processor()
__SCREAMING_SNAKE_CASE : Optional[Any] = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
processor.save_pretrained(self.tmpdirname )
__SCREAMING_SNAKE_CASE : int = VisionTextDualEncoderProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor.image_processor , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : str = VisionTextDualEncoderProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
__SCREAMING_SNAKE_CASE : Optional[int] = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''' )
__SCREAMING_SNAKE_CASE : List[str] = self.get_image_processor(do_normalize=_lowerCamelCase , padding_value=1.0 )
__SCREAMING_SNAKE_CASE : int = VisionTextDualEncoderProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=_lowerCamelCase , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.get_image_processor()
__SCREAMING_SNAKE_CASE : Dict = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Tuple = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = self.prepare_image_inputs()
__SCREAMING_SNAKE_CASE : Optional[int] = image_processor(_lowerCamelCase , return_tensors='''np''' )
__SCREAMING_SNAKE_CASE : Tuple = processor(images=_lowerCamelCase , return_tensors='''np''' )
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1e-2 )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : Any = self.get_image_processor()
__SCREAMING_SNAKE_CASE : Any = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Tuple = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = '''lower newer'''
__SCREAMING_SNAKE_CASE : Optional[int] = processor(text=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = tokenizer(_lowerCamelCase )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE : str = self.get_image_processor()
__SCREAMING_SNAKE_CASE : int = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : int = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = '''lower newer'''
__SCREAMING_SNAKE_CASE : int = self.prepare_image_inputs()
__SCREAMING_SNAKE_CASE : int = processor(text=_lowerCamelCase , images=_lowerCamelCase )
self.assertListEqual(list(inputs.keys() ) , ['''input_ids''', '''token_type_ids''', '''attention_mask''', '''pixel_values'''] )
# test if it raises when no input is passed
with self.assertRaises(_lowerCamelCase ):
processor()
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : Dict = self.get_image_processor()
__SCREAMING_SNAKE_CASE : List[Any] = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Union[str, Any] = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
__SCREAMING_SNAKE_CASE : Tuple = processor.batch_decode(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = tokenizer.batch_decode(_lowerCamelCase )
self.assertListEqual(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : str = self.get_image_processor()
__SCREAMING_SNAKE_CASE : List[str] = self.get_tokenizer()
__SCREAMING_SNAKE_CASE : Optional[int] = VisionTextDualEncoderProcessor(tokenizer=_lowerCamelCase , image_processor=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = '''lower newer'''
__SCREAMING_SNAKE_CASE : Dict = self.prepare_image_inputs()
__SCREAMING_SNAKE_CASE : str = processor(text=_lowerCamelCase , images=_lowerCamelCase )
self.assertListEqual(list(inputs.keys() ) , processor.model_input_names )
| 674 | 1 |
"""simple docstring"""
from collections import OrderedDict
from typing import Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
_lowerCamelCase = logging.get_logger(__name__)
_lowerCamelCase = {
'''microsoft/resnet-50''': '''https://huggingface.co/microsoft/resnet-50/blob/main/config.json''',
}
class snake_case ( __UpperCAmelCase , __UpperCAmelCase ):
lowerCamelCase__ = '''resnet'''
lowerCamelCase__ = ['''basic''', '''bottleneck''']
def __init__( self :Tuple , _lowerCamelCase :Tuple=3 , _lowerCamelCase :Any=6_4 , _lowerCamelCase :Optional[Any]=[2_5_6, 5_1_2, 1_0_2_4, 2_0_4_8] , _lowerCamelCase :List[Any]=[3, 4, 6, 3] , _lowerCamelCase :List[str]="bottleneck" , _lowerCamelCase :List[Any]="relu" , _lowerCamelCase :Optional[Any]=False , _lowerCamelCase :Tuple=None , _lowerCamelCase :str=None , **_lowerCamelCase :Optional[int] , ):
super().__init__(**_lowerCamelCase )
if layer_type not in self.layer_types:
raise ValueError(f'''layer_type={layer_type} is not one of {','.join(self.layer_types )}''' )
__SCREAMING_SNAKE_CASE : List[Any] = num_channels
__SCREAMING_SNAKE_CASE : List[str] = embedding_size
__SCREAMING_SNAKE_CASE : Union[str, Any] = hidden_sizes
__SCREAMING_SNAKE_CASE : List[str] = depths
__SCREAMING_SNAKE_CASE : Optional[int] = layer_type
__SCREAMING_SNAKE_CASE : Dict = hidden_act
__SCREAMING_SNAKE_CASE : int = downsample_in_first_stage
__SCREAMING_SNAKE_CASE : int = ['''stem'''] + [f'''stage{idx}''' for idx in range(1 , len(_lowerCamelCase ) + 1 )]
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[str] = get_aligned_output_features_output_indices(
out_features=_lowerCamelCase , out_indices=_lowerCamelCase , stage_names=self.stage_names )
class snake_case ( __UpperCAmelCase ):
lowerCamelCase__ = version.parse('''1.11''' )
@property
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
return OrderedDict(
[
('''pixel_values''', {0: '''batch''', 1: '''num_channels''', 2: '''height''', 3: '''width'''}),
] )
@property
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
return 1e-3
| 674 |
"""simple docstring"""
import inspect
import unittest
import numpy as np
from tests.test_modeling_common import floats_tensor
from transformers import DetrConfig, MaskFormerConfig, SwinConfig, is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device
from transformers.utils import cached_property
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import MaskFormerForInstanceSegmentation, MaskFormerModel
if is_vision_available():
from transformers import MaskFormerImageProcessor
if is_vision_available():
from PIL import Image
class snake_case :
def __init__( self :Optional[int] , _lowerCamelCase :int , _lowerCamelCase :Any=2 , _lowerCamelCase :List[str]=True , _lowerCamelCase :str=False , _lowerCamelCase :Tuple=1_0 , _lowerCamelCase :str=3 , _lowerCamelCase :str=3_2 * 4 , _lowerCamelCase :Dict=3_2 * 6 , _lowerCamelCase :str=4 , _lowerCamelCase :Any=3_2 , ):
__SCREAMING_SNAKE_CASE : List[str] = parent
__SCREAMING_SNAKE_CASE : Tuple = batch_size
__SCREAMING_SNAKE_CASE : Optional[Any] = is_training
__SCREAMING_SNAKE_CASE : Dict = use_auxiliary_loss
__SCREAMING_SNAKE_CASE : List[str] = num_queries
__SCREAMING_SNAKE_CASE : Optional[int] = num_channels
__SCREAMING_SNAKE_CASE : List[Any] = min_size
__SCREAMING_SNAKE_CASE : int = max_size
__SCREAMING_SNAKE_CASE : Any = num_labels
__SCREAMING_SNAKE_CASE : Union[str, Any] = mask_feature_size
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : str = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to(
_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = torch.ones([self.batch_size, self.min_size, self.max_size] , device=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = (
torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] , device=_lowerCamelCase ) > 0.5
).float()
__SCREAMING_SNAKE_CASE : Dict = (torch.rand((self.batch_size, self.num_labels) , device=_lowerCamelCase ) > 0.5).long()
__SCREAMING_SNAKE_CASE : str = self.get_config()
return config, pixel_values, pixel_mask, mask_labels, class_labels
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
return MaskFormerConfig.from_backbone_and_decoder_configs(
backbone_config=SwinConfig(
depths=[1, 1, 1, 1] , ) , decoder_config=DetrConfig(
decoder_ffn_dim=1_2_8 , num_queries=self.num_queries , decoder_attention_heads=2 , d_model=self.mask_feature_size , ) , mask_feature_size=self.mask_feature_size , fpn_feature_size=self.mask_feature_size , num_channels=self.num_channels , num_labels=self.num_labels , )
def SCREAMING_SNAKE_CASE_ ( self :str ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = self.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE : Any = {'''pixel_values''': pixel_values, '''pixel_mask''': pixel_mask}
return config, inputs_dict
def SCREAMING_SNAKE_CASE_ ( self :int , _lowerCamelCase :int , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = output.encoder_hidden_states
__SCREAMING_SNAKE_CASE : int = output.pixel_decoder_hidden_states
__SCREAMING_SNAKE_CASE : Optional[int] = output.transformer_decoder_hidden_states
self.parent.assertTrue(len(_lowerCamelCase ) , len(config.backbone_config.depths ) )
self.parent.assertTrue(len(_lowerCamelCase ) , len(config.backbone_config.depths ) )
self.parent.assertTrue(len(_lowerCamelCase ) , config.decoder_config.decoder_layers )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Dict , _lowerCamelCase :str , _lowerCamelCase :Any , _lowerCamelCase :Optional[Any]=False ):
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[str] = MaskFormerModel(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
__SCREAMING_SNAKE_CASE : str = model(pixel_values=_lowerCamelCase , pixel_mask=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = model(_lowerCamelCase , output_hidden_states=_lowerCamelCase )
# the correct shape of output.transformer_decoder_hidden_states ensure the correcteness of the
# encoder and pixel decoder
self.parent.assertEqual(
output.transformer_decoder_last_hidden_state.shape , (self.batch_size, self.num_queries, self.mask_feature_size) , )
# let's ensure the other two hidden state exists
self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(output.encoder_last_hidden_state is not None )
if output_hidden_states:
self.check_output_hidden_state(_lowerCamelCase , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :Optional[Any] , _lowerCamelCase :Optional[int] , _lowerCamelCase :Optional[Any] , _lowerCamelCase :List[str] , _lowerCamelCase :Dict ):
__SCREAMING_SNAKE_CASE : Optional[Any] = MaskFormerForInstanceSegmentation(config=_lowerCamelCase )
model.to(_lowerCamelCase )
model.eval()
def comm_check_on_output(_lowerCamelCase :Optional[Any] ):
# let's still check that all the required stuff is there
self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.encoder_last_hidden_state is not None )
# okay, now we need to check the logits shape
# due to the encoder compression, masks have a //4 spatial size
self.parent.assertEqual(
result.masks_queries_logits.shape , (self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) , )
# + 1 for null class
self.parent.assertEqual(
result.class_queries_logits.shape , (self.batch_size, self.num_queries, self.num_labels + 1) )
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[str] = model(pixel_values=_lowerCamelCase , pixel_mask=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = model(_lowerCamelCase )
comm_check_on_output(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = model(
pixel_values=_lowerCamelCase , pixel_mask=_lowerCamelCase , mask_labels=_lowerCamelCase , class_labels=_lowerCamelCase )
comm_check_on_output(_lowerCamelCase )
self.parent.assertTrue(result.loss is not None )
self.parent.assertEqual(result.loss.shape , torch.Size([1] ) )
@require_torch
class snake_case ( __UpperCAmelCase , __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = (MaskFormerModel, MaskFormerForInstanceSegmentation) if is_torch_available() else ()
lowerCamelCase__ = (
{'''feature-extraction''': MaskFormerModel, '''image-segmentation''': MaskFormerForInstanceSegmentation}
if is_torch_available()
else {}
)
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
lowerCamelCase__ = False
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
__SCREAMING_SNAKE_CASE : Optional[Any] = MaskFormerModelTester(self )
__SCREAMING_SNAKE_CASE : str = ConfigTester(self , config_class=_lowerCamelCase , has_text_modality=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(_lowerCamelCase , **_lowerCamelCase , output_hidden_states=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
__SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_maskformer_instance_segmentation_head_model(*_lowerCamelCase )
@unittest.skip(reason='''MaskFormer does not use inputs_embeds''' )
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
pass
@unittest.skip(reason='''MaskFormer does not have a get_input_embeddings method''' )
def SCREAMING_SNAKE_CASE_ ( self :int ):
pass
@unittest.skip(reason='''MaskFormer is not a generative model''' )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
pass
@unittest.skip(reason='''MaskFormer does not use token embeddings''' )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
pass
@require_torch_multi_gpu
@unittest.skip(
reason='''MaskFormer has some layers using `add_module` which doesn\'t work well with `nn.DataParallel`''' )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
pass
@unittest.skip('''Will be fixed soon by reducing the size of the model used for common tests.''' )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
pass
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Tuple = model_class(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
__SCREAMING_SNAKE_CASE : List[str] = [*signature.parameters.keys()]
__SCREAMING_SNAKE_CASE : Optional[Any] = ['''pixel_values''']
self.assertListEqual(arg_names[:1] , _lowerCamelCase )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
for model_name in ["facebook/maskformer-swin-small-coco"]:
__SCREAMING_SNAKE_CASE : Tuple = MaskFormerModel.from_pretrained(_lowerCamelCase )
self.assertIsNotNone(_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Dict = (self.model_tester.min_size,) * 2
__SCREAMING_SNAKE_CASE : Dict = {
'''pixel_values''': torch.randn((2, 3, *size) , device=_lowerCamelCase ),
'''mask_labels''': torch.randn((2, 1_0, *size) , device=_lowerCamelCase ),
'''class_labels''': torch.zeros(2 , 1_0 , device=_lowerCamelCase ).long(),
}
__SCREAMING_SNAKE_CASE : Dict = MaskFormerForInstanceSegmentation(MaskFormerConfig() ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[Any] = model(**_lowerCamelCase )
self.assertTrue(outputs.loss is not None )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(_lowerCamelCase , **_lowerCamelCase , output_hidden_states=_lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__SCREAMING_SNAKE_CASE : Dict = model_class(_lowerCamelCase ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Optional[int] = model(**_lowerCamelCase , output_attentions=_lowerCamelCase )
self.assertTrue(outputs.attentions is not None )
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
if not self.model_tester.is_training:
return
# only MaskFormerForInstanceSegmentation has the loss
__SCREAMING_SNAKE_CASE : List[Any] = self.all_model_classes[1]
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.model_tester.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE : int = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.train()
__SCREAMING_SNAKE_CASE : Any = model(_lowerCamelCase , mask_labels=_lowerCamelCase , class_labels=_lowerCamelCase ).loss
loss.backward()
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# only MaskFormerForInstanceSegmentation has the loss
__SCREAMING_SNAKE_CASE : Tuple = self.all_model_classes[1]
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.model_tester.prepare_config_and_inputs()
__SCREAMING_SNAKE_CASE : Tuple = True
__SCREAMING_SNAKE_CASE : List[str] = True
__SCREAMING_SNAKE_CASE : Any = model_class(_lowerCamelCase )
model.to(_lowerCamelCase )
model.train()
__SCREAMING_SNAKE_CASE : Optional[int] = model(_lowerCamelCase , mask_labels=_lowerCamelCase , class_labels=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = outputs.encoder_hidden_states[0]
encoder_hidden_states.retain_grad()
__SCREAMING_SNAKE_CASE : str = outputs.pixel_decoder_hidden_states[0]
pixel_decoder_hidden_states.retain_grad()
# we requires_grad=True in inputs_embeds (line 2152), the original implementation don't
__SCREAMING_SNAKE_CASE : Optional[int] = outputs.transformer_decoder_hidden_states[0]
transformer_decoder_hidden_states.retain_grad()
__SCREAMING_SNAKE_CASE : int = outputs.attentions[0]
attentions.retain_grad()
outputs.loss.backward(retain_graph=_lowerCamelCase )
self.assertIsNotNone(encoder_hidden_states.grad )
self.assertIsNotNone(pixel_decoder_hidden_states.grad )
self.assertIsNotNone(transformer_decoder_hidden_states.grad )
self.assertIsNotNone(attentions.grad )
_lowerCamelCase = 1e-4
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : List[Any] = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
return image
@require_vision
@slow
class snake_case ( unittest.TestCase ):
@cached_property
def SCREAMING_SNAKE_CASE_ ( self :str ):
return (
MaskFormerImageProcessor.from_pretrained('''facebook/maskformer-swin-small-coco''' )
if is_vision_available()
else None
)
def SCREAMING_SNAKE_CASE_ ( self :List[Any] ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = MaskFormerModel.from_pretrained('''facebook/maskformer-swin-small-coco''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = self.default_image_processor
__SCREAMING_SNAKE_CASE : Tuple = prepare_img()
__SCREAMING_SNAKE_CASE : List[Any] = image_processor(_lowerCamelCase , return_tensors='''pt''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = inputs['''pixel_values'''].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 3_2) == 0 and (inputs_shape[-2] % 3_2) == 0 )
# check size
self.assertEqual(_lowerCamelCase , (1, 3, 8_0_0, 1_0_8_8) )
with torch.no_grad():
__SCREAMING_SNAKE_CASE : Tuple = model(**_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.tensor(
[[-0.0_4_8_2, 0.9_2_2_8, 0.4_9_5_1], [-0.2_5_4_7, 0.8_0_1_7, 0.8_5_2_7], [-0.0_0_6_9, 0.3_3_8_5, -0.0_0_8_9]] ).to(_lowerCamelCase )
self.assertTrue(
torch.allclose(
outputs.encoder_last_hidden_state[0, 0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Any = torch.tensor(
[[-0.8_4_2_2, -0.8_4_3_4, -0.9_7_1_8], [-1.0_1_4_4, -0.5_5_6_5, -0.4_1_9_5], [-1.0_0_3_8, -0.4_4_8_4, -0.1_9_6_1]] ).to(_lowerCamelCase )
self.assertTrue(
torch.allclose(
outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
__SCREAMING_SNAKE_CASE : Optional[int] = torch.tensor(
[[0.2_8_5_2, -0.0_1_5_9, 0.9_7_3_5], [0.6_2_5_4, 0.1_8_5_8, 0.8_5_2_9], [-0.0_6_8_0, -0.4_1_1_6, 1.8_4_1_3]] ).to(_lowerCamelCase )
self.assertTrue(
torch.allclose(
outputs.transformer_decoder_last_hidden_state[0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
__SCREAMING_SNAKE_CASE : Tuple = (
MaskFormerForInstanceSegmentation.from_pretrained('''facebook/maskformer-swin-small-coco''' )
.to(_lowerCamelCase )
.eval()
)
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.default_image_processor
__SCREAMING_SNAKE_CASE : Tuple = prepare_img()
__SCREAMING_SNAKE_CASE : str = image_processor(_lowerCamelCase , return_tensors='''pt''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = inputs['''pixel_values'''].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 3_2) == 0 and (inputs_shape[-2] % 3_2) == 0 )
# check size
self.assertEqual(_lowerCamelCase , (1, 3, 8_0_0, 1_0_8_8) )
with torch.no_grad():
__SCREAMING_SNAKE_CASE : str = model(**_lowerCamelCase )
# masks_queries_logits
__SCREAMING_SNAKE_CASE : Any = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape , (1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) , )
__SCREAMING_SNAKE_CASE : List[Any] = [
[-1.3_7_3_7_1_2_4, -1.7_7_2_4_9_3_7, -1.9_3_6_4_2_3_3],
[-1.5_9_7_7_2_8_1, -1.9_8_6_7_9_3_9, -2.1_5_2_3_6_9_5],
[-1.5_7_9_5_3_9_8, -1.9_2_6_9_8_3_2, -2.0_9_3_9_4_2],
]
__SCREAMING_SNAKE_CASE : Any = torch.tensor(_lowerCamelCase ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
# class_queries_logits
__SCREAMING_SNAKE_CASE : List[Any] = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape , (1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor(
[
[1.65_12e00, -5.25_72e00, -3.35_19e00],
[3.61_69e-02, -5.90_25e00, -2.93_13e00],
[1.07_66e-04, -7.76_30e00, -5.12_63e00],
] ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
def SCREAMING_SNAKE_CASE_ ( self :Any ):
__SCREAMING_SNAKE_CASE : Tuple = (
MaskFormerForInstanceSegmentation.from_pretrained('''facebook/maskformer-resnet101-coco-stuff''' )
.to(_lowerCamelCase )
.eval()
)
__SCREAMING_SNAKE_CASE : Tuple = self.default_image_processor
__SCREAMING_SNAKE_CASE : str = prepare_img()
__SCREAMING_SNAKE_CASE : List[Any] = image_processor(_lowerCamelCase , return_tensors='''pt''' ).to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[Any] = inputs['''pixel_values'''].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 3_2) == 0 and (inputs_shape[-2] % 3_2) == 0 )
# check size
self.assertEqual(_lowerCamelCase , (1, 3, 8_0_0, 1_0_8_8) )
with torch.no_grad():
__SCREAMING_SNAKE_CASE : List[Any] = model(**_lowerCamelCase )
# masks_queries_logits
__SCREAMING_SNAKE_CASE : Union[str, Any] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape , (1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) , )
__SCREAMING_SNAKE_CASE : List[str] = [[-0.9_0_4_6, -2.6_3_6_6, -4.6_0_6_2], [-3.4_1_7_9, -5.7_8_9_0, -8.8_0_5_7], [-4.9_1_7_9, -7.6_5_6_0, -1_0.7_7_1_1]]
__SCREAMING_SNAKE_CASE : Any = torch.tensor(_lowerCamelCase ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
# class_queries_logits
__SCREAMING_SNAKE_CASE : int = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape , (1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor(
[[4.7_1_8_8, -3.2_5_8_5, -2.8_8_5_7], [6.6_8_7_1, -2.9_1_8_1, -1.2_4_8_7], [7.2_4_4_9, -2.2_7_6_4, -2.1_8_7_4]] ).to(_lowerCamelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] , _lowerCamelCase , atol=_lowerCamelCase ) )
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : int = (
MaskFormerForInstanceSegmentation.from_pretrained('''facebook/maskformer-swin-small-coco''' )
.to(_lowerCamelCase )
.eval()
)
__SCREAMING_SNAKE_CASE : Any = self.default_image_processor
__SCREAMING_SNAKE_CASE : int = image_processor(
[np.zeros((3, 8_0_0, 1_3_3_3) ), np.zeros((3, 8_0_0, 1_3_3_3) )] , segmentation_maps=[np.zeros((3_8_4, 3_8_4) ).astype(np.floataa ), np.zeros((3_8_4, 3_8_4) ).astype(np.floataa )] , return_tensors='''pt''' , )
__SCREAMING_SNAKE_CASE : Dict = inputs['''pixel_values'''].to(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Dict = [el.to(_lowerCamelCase ) for el in inputs['''mask_labels''']]
__SCREAMING_SNAKE_CASE : str = [el.to(_lowerCamelCase ) for el in inputs['''class_labels''']]
with torch.no_grad():
__SCREAMING_SNAKE_CASE : int = model(**_lowerCamelCase )
self.assertTrue(outputs.loss is not None )
| 674 | 1 |
"""simple docstring"""
import unittest
import numpy as np
import torch
from diffusers import PNDMPipeline, PNDMScheduler, UNetaDModel
from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device
enable_full_determinism()
class snake_case ( unittest.TestCase ):
@property
def SCREAMING_SNAKE_CASE_ ( self :Any ):
torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : str = UNetaDModel(
block_out_channels=(3_2, 6_4) , layers_per_block=2 , sample_size=3_2 , in_channels=3 , out_channels=3 , down_block_types=('''DownBlock2D''', '''AttnDownBlock2D''') , up_block_types=('''AttnUpBlock2D''', '''UpBlock2D''') , )
return model
def SCREAMING_SNAKE_CASE_ ( self :int ):
__SCREAMING_SNAKE_CASE : Optional[int] = self.dummy_uncond_unet
__SCREAMING_SNAKE_CASE : int = PNDMScheduler()
__SCREAMING_SNAKE_CASE : Optional[int] = PNDMPipeline(unet=_lowerCamelCase , scheduler=_lowerCamelCase )
pndm.to(_lowerCamelCase )
pndm.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Optional[Any] = pndm(generator=_lowerCamelCase , num_inference_steps=2_0 , output_type='''numpy''' ).images
__SCREAMING_SNAKE_CASE : str = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : str = pndm(generator=_lowerCamelCase , num_inference_steps=2_0 , output_type='''numpy''' , return_dict=_lowerCamelCase )[0]
__SCREAMING_SNAKE_CASE : List[str] = image[0, -3:, -3:, -1]
__SCREAMING_SNAKE_CASE : Dict = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 3_2, 3_2, 3)
__SCREAMING_SNAKE_CASE : Tuple = np.array([1.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1e-2
@slow
@require_torch
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : List[str] = '''google/ddpm-cifar10-32'''
__SCREAMING_SNAKE_CASE : Tuple = UNetaDModel.from_pretrained(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = PNDMScheduler()
__SCREAMING_SNAKE_CASE : Optional[int] = PNDMPipeline(unet=_lowerCamelCase , scheduler=_lowerCamelCase )
pndm.to(_lowerCamelCase )
pndm.set_progress_bar_config(disable=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = torch.manual_seed(0 )
__SCREAMING_SNAKE_CASE : Tuple = pndm(generator=_lowerCamelCase , output_type='''numpy''' ).images
__SCREAMING_SNAKE_CASE : Dict = image[0, -3:, -3:, -1]
assert image.shape == (1, 3_2, 3_2, 3)
__SCREAMING_SNAKE_CASE : Dict = np.array([0.1_5_6_4, 0.1_4_6_4_5, 0.1_4_0_6, 0.1_4_7_1_5, 0.1_2_4_2_5, 0.1_4_0_4_5, 0.1_3_1_1_5, 0.1_2_1_7_5, 0.1_2_5] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
| 674 |
"""simple docstring"""
import contextlib
import importlib
import io
import unittest
import transformers
# Try to import everything from transformers to ensure every object can be loaded.
from transformers import * # noqa F406
from transformers.testing_utils import DUMMY_UNKNOWN_IDENTIFIER, require_flax, require_tf, require_torch
from transformers.utils import ContextManagers, find_labels, is_flax_available, is_tf_available, is_torch_available
if is_torch_available():
from transformers import BertForPreTraining, BertForQuestionAnswering, BertForSequenceClassification
if is_tf_available():
from transformers import TFBertForPreTraining, TFBertForQuestionAnswering, TFBertForSequenceClassification
if is_flax_available():
from transformers import FlaxBertForPreTraining, FlaxBertForQuestionAnswering, FlaxBertForSequenceClassification
_lowerCamelCase = DUMMY_UNKNOWN_IDENTIFIER
# An actual model hosted on huggingface.co
_lowerCamelCase = '''main'''
# Default branch name
_lowerCamelCase = '''f2c752cfc5c0ab6f4bdec59acea69eefbee381c2'''
# One particular commit (not the top of `main`)
_lowerCamelCase = '''aaaaaaa'''
# This commit does not exist, so we should 404.
_lowerCamelCase = '''d9e9f15bc825e4b2c9249e9578f884bbcb5e3684'''
# Sha-1 of config.json on the top of `main`, for checking purposes
_lowerCamelCase = '''4b243c475af8d0a7754e87d7d096c92e5199ec2fe168a2ee7998e3b8e9bcb1d3'''
@contextlib.contextmanager
def lowerCAmelCase_ ( ):
'''simple docstring'''
print('''Welcome!''' )
yield
print('''Bye!''' )
@contextlib.contextmanager
def lowerCAmelCase_ ( ):
'''simple docstring'''
print('''Bonjour!''' )
yield
print('''Au revoir!''' )
class snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_ ( self :Union[str, Any] ):
# If the spec is missing, importlib would not be able to import the module dynamically.
assert transformers.__spec__ is not None
assert importlib.util.find_spec('''transformers''' ) is not None
class snake_case ( unittest.TestCase ):
@unittest.mock.patch('''sys.stdout''' , new_callable=io.StringIO )
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :Optional[int] ):
with ContextManagers([] ):
print('''Transformers are awesome!''' )
# The print statement adds a new line at the end of the output
self.assertEqual(mock_stdout.getvalue() , '''Transformers are awesome!\n''' )
@unittest.mock.patch('''sys.stdout''' , new_callable=io.StringIO )
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :Optional[int] ):
with ContextManagers([context_en()] ):
print('''Transformers are awesome!''' )
# The output should be wrapped with an English welcome and goodbye
self.assertEqual(mock_stdout.getvalue() , '''Welcome!\nTransformers are awesome!\nBye!\n''' )
@unittest.mock.patch('''sys.stdout''' , new_callable=io.StringIO )
def SCREAMING_SNAKE_CASE_ ( self :Dict , _lowerCamelCase :List[str] ):
with ContextManagers([context_fr(), context_en()] ):
print('''Transformers are awesome!''' )
# The output should be wrapped with an English and French welcome and goodbye
self.assertEqual(mock_stdout.getvalue() , '''Bonjour!\nWelcome!\nTransformers are awesome!\nBye!\nAu revoir!\n''' )
@require_torch
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels'''] )
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels''', '''next_sentence_label'''] )
self.assertEqual(find_labels(_lowerCamelCase ) , ['''start_positions''', '''end_positions'''] )
class snake_case ( __UpperCAmelCase ):
pass
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels'''] )
@require_tf
def SCREAMING_SNAKE_CASE_ ( self :int ):
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels'''] )
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels''', '''next_sentence_label'''] )
self.assertEqual(find_labels(_lowerCamelCase ) , ['''start_positions''', '''end_positions'''] )
class snake_case ( __UpperCAmelCase ):
pass
self.assertEqual(find_labels(_lowerCamelCase ) , ['''labels'''] )
@require_flax
def SCREAMING_SNAKE_CASE_ ( self :Dict ):
# Flax models don't have labels
self.assertEqual(find_labels(_lowerCamelCase ) , [] )
self.assertEqual(find_labels(_lowerCamelCase ) , [] )
self.assertEqual(find_labels(_lowerCamelCase ) , [] )
class snake_case ( __UpperCAmelCase ):
pass
self.assertEqual(find_labels(_lowerCamelCase ) , [] )
| 674 | 1 |
"""simple docstring"""
import doctest
import glob
import importlib
import inspect
import os
import re
from contextlib import contextmanager
from functools import wraps
from unittest.mock import patch
import numpy as np
import pytest
from absl.testing import parameterized
import datasets
from datasets import load_metric
from .utils import for_all_test_methods, local, slow
# mark all tests as integration
_lowerCamelCase = pytest.mark.integration
_lowerCamelCase = {'''comet'''}
_lowerCamelCase = importlib.util.find_spec('''fairseq''') is not None
_lowerCamelCase = {'''code_eval'''}
_lowerCamelCase = os.name == '''nt'''
_lowerCamelCase = {'''bertscore''', '''frugalscore''', '''perplexity'''}
_lowerCamelCase = importlib.util.find_spec('''transformers''') is not None
def lowerCAmelCase_ ( lowercase_ : Tuple ):
'''simple docstring'''
@wraps(lowercase_ )
def wrapper(self : Optional[Any] , lowercase_ : Dict ):
if not _has_fairseq and metric_name in REQUIRE_FAIRSEQ:
self.skipTest('''"test requires Fairseq"''' )
else:
test_case(self , lowercase_ )
return wrapper
def lowerCAmelCase_ ( lowercase_ : int ):
'''simple docstring'''
@wraps(lowercase_ )
def wrapper(self : List[Any] , lowercase_ : str ):
if not _has_transformers and metric_name in REQUIRE_TRANSFORMERS:
self.skipTest('''"test requires transformers"''' )
else:
test_case(self , lowercase_ )
return wrapper
def lowerCAmelCase_ ( lowercase_ : Any ):
'''simple docstring'''
@wraps(lowercase_ )
def wrapper(self : List[str] , lowercase_ : Union[str, Any] ):
if _on_windows and metric_name in UNSUPPORTED_ON_WINDOWS:
self.skipTest('''"test not supported on Windows"''' )
else:
test_case(self , lowercase_ )
return wrapper
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Any = [metric_dir.split(os.sep )[-2] for metric_dir in glob.glob('''./metrics/*/''' )]
return [{"testcase_name": x, "metric_name": x} for x in metrics if x != "gleu"] # gleu is unfinished
@parameterized.named_parameters(get_local_metric_names() )
@for_all_test_methods(
__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
@local
class snake_case ( parameterized.TestCase ):
lowerCamelCase__ = {}
lowerCamelCase__ = None
@pytest.mark.filterwarnings('''ignore:metric_module_factory is deprecated:FutureWarning''' )
@pytest.mark.filterwarnings('''ignore:load_metric is deprecated:FutureWarning''' )
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :Optional[int] ):
__SCREAMING_SNAKE_CASE : Dict = '''[...]'''
__SCREAMING_SNAKE_CASE : Optional[int] = importlib.import_module(
datasets.load.metric_module_factory(os.path.join('''metrics''' , _lowerCamelCase ) ).module_path )
__SCREAMING_SNAKE_CASE : List[str] = datasets.load.import_main_class(metric_module.__name__ , dataset=_lowerCamelCase )
# check parameters
__SCREAMING_SNAKE_CASE : int = inspect.signature(metric._compute ).parameters
self.assertTrue(all(p.kind != p.VAR_KEYWORD for p in parameters.values() ) ) # no **kwargs
# run doctest
with self.patch_intensive_calls(_lowerCamelCase , metric_module.__name__ ):
with self.use_local_metrics():
try:
__SCREAMING_SNAKE_CASE : Tuple = doctest.testmod(_lowerCamelCase , verbose=_lowerCamelCase , raise_on_error=_lowerCamelCase )
except doctest.UnexpectedException as e:
raise e.exc_info[1] # raise the exception that doctest caught
self.assertEqual(results.failed , 0 )
self.assertGreater(results.attempted , 1 )
@slow
def SCREAMING_SNAKE_CASE_ ( self :str , _lowerCamelCase :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : List[Any] = '''[...]'''
__SCREAMING_SNAKE_CASE : List[Any] = importlib.import_module(
datasets.load.metric_module_factory(os.path.join('''metrics''' , _lowerCamelCase ) ).module_path )
# run doctest
with self.use_local_metrics():
__SCREAMING_SNAKE_CASE : Any = doctest.testmod(_lowerCamelCase , verbose=_lowerCamelCase , raise_on_error=_lowerCamelCase )
self.assertEqual(results.failed , 0 )
self.assertGreater(results.attempted , 1 )
@contextmanager
def SCREAMING_SNAKE_CASE_ ( self :List[Any] , _lowerCamelCase :str , _lowerCamelCase :Optional[Any] ):
if metric_name in self.INTENSIVE_CALLS_PATCHER:
with self.INTENSIVE_CALLS_PATCHER[metric_name](_lowerCamelCase ):
yield
else:
yield
@contextmanager
def SCREAMING_SNAKE_CASE_ ( self :Any ):
def load_local_metric(_lowerCamelCase :Optional[Any] , *_lowerCamelCase :Dict , **_lowerCamelCase :Optional[Any] ):
return load_metric(os.path.join('''metrics''' , _lowerCamelCase ) , *_lowerCamelCase , **_lowerCamelCase )
with patch('''datasets.load_metric''' ) as mock_load_metric:
__SCREAMING_SNAKE_CASE : Optional[Any] = load_local_metric
yield
@classmethod
def SCREAMING_SNAKE_CASE_ ( cls :List[str] , _lowerCamelCase :Tuple ):
def wrapper(_lowerCamelCase :Union[str, Any] ):
__SCREAMING_SNAKE_CASE : Tuple = contextmanager(_lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = patcher
return patcher
return wrapper
@LocalMetricTest.register_intensive_calls_patcher('''bleurt''' )
def lowerCAmelCase_ ( lowercase_ : List[str] ):
'''simple docstring'''
import tensorflow.compat.va as tf
from bleurt.score import Predictor
tf.flags.DEFINE_string('''sv''' , '''''' , '''''' ) # handle pytest cli flags
class snake_case ( __UpperCAmelCase ):
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :Optional[int] ):
assert len(input_dict['''input_ids'''] ) == 2
return np.array([1.0_3, 1.0_4] )
# mock predict_fn which is supposed to do a forward pass with a bleurt model
with patch('''bleurt.score._create_predictor''' ) as mock_create_predictor:
__SCREAMING_SNAKE_CASE : Dict = MockedPredictor()
yield
@LocalMetricTest.register_intensive_calls_patcher('''bertscore''' )
def lowerCAmelCase_ ( lowercase_ : Optional[int] ):
'''simple docstring'''
import torch
def bert_cos_score_idf(lowercase_ : Optional[Any] , lowercase_ : Any , *lowercase_ : int , **lowercase_ : List[str] ):
return torch.tensor([[1.0, 1.0, 1.0]] * len(lowercase_ ) )
# mock get_model which is supposed to do download a bert model
# mock bert_cos_score_idf which is supposed to do a forward pass with a bert model
with patch('''bert_score.scorer.get_model''' ), patch(
'''bert_score.scorer.bert_cos_score_idf''' ) as mock_bert_cos_score_idf:
__SCREAMING_SNAKE_CASE : List[str] = bert_cos_score_idf
yield
@LocalMetricTest.register_intensive_calls_patcher('''comet''' )
def lowerCAmelCase_ ( lowercase_ : Optional[int] ):
'''simple docstring'''
def load_from_checkpoint(lowercase_ : Tuple ):
class snake_case :
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] , _lowerCamelCase :int , *_lowerCamelCase :List[str] , **_lowerCamelCase :List[str] ):
assert len(_lowerCamelCase ) == 2
__SCREAMING_SNAKE_CASE : int = [0.1_9, 0.9_2]
return scores, sum(_lowerCamelCase ) / len(_lowerCamelCase )
return Model()
# mock load_from_checkpoint which is supposed to do download a bert model
# mock load_from_checkpoint which is supposed to do download a bert model
with patch('''comet.download_model''' ) as mock_download_model:
__SCREAMING_SNAKE_CASE : str = None
with patch('''comet.load_from_checkpoint''' ) as mock_load_from_checkpoint:
__SCREAMING_SNAKE_CASE : Tuple = load_from_checkpoint
yield
def lowerCAmelCase_ ( ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : int = load_metric(os.path.join('''metrics''' , '''seqeval''' ) )
__SCREAMING_SNAKE_CASE : List[str] = '''ERROR'''
__SCREAMING_SNAKE_CASE : Union[str, Any] = F'''Scheme should be one of [IOB1, IOB2, IOE1, IOE2, IOBES, BILOU], got {wrong_scheme}'''
with pytest.raises(lowercase_ , match=re.escape(lowercase_ ) ):
metric.compute(predictions=[] , references=[] , scheme=lowercase_ )
| 674 |
"""simple docstring"""
import json
import pathlib
import unittest
import numpy as np
from transformers.testing_utils import require_torch, require_vision, slow
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import YolosImageProcessor
class snake_case ( unittest.TestCase ):
def __init__( self :List[Any] , _lowerCamelCase :Dict , _lowerCamelCase :Tuple=7 , _lowerCamelCase :Dict=3 , _lowerCamelCase :Optional[Any]=3_0 , _lowerCamelCase :List[str]=4_0_0 , _lowerCamelCase :Union[str, Any]=True , _lowerCamelCase :Union[str, Any]=None , _lowerCamelCase :List[Any]=True , _lowerCamelCase :Any=[0.5, 0.5, 0.5] , _lowerCamelCase :Dict=[0.5, 0.5, 0.5] , _lowerCamelCase :Dict=True , _lowerCamelCase :str=1 / 2_5_5 , _lowerCamelCase :Union[str, Any]=True , ):
# by setting size["longest_edge"] > max_resolution we're effectively not testing this :p
__SCREAMING_SNAKE_CASE : Optional[Any] = size if size is not None else {'''shortest_edge''': 1_8, '''longest_edge''': 1_3_3_3}
__SCREAMING_SNAKE_CASE : List[str] = parent
__SCREAMING_SNAKE_CASE : Dict = batch_size
__SCREAMING_SNAKE_CASE : str = num_channels
__SCREAMING_SNAKE_CASE : Union[str, Any] = min_resolution
__SCREAMING_SNAKE_CASE : Union[str, Any] = max_resolution
__SCREAMING_SNAKE_CASE : Tuple = do_resize
__SCREAMING_SNAKE_CASE : Union[str, Any] = size
__SCREAMING_SNAKE_CASE : int = do_normalize
__SCREAMING_SNAKE_CASE : List[Any] = image_mean
__SCREAMING_SNAKE_CASE : Tuple = image_std
__SCREAMING_SNAKE_CASE : Dict = do_rescale
__SCREAMING_SNAKE_CASE : Optional[int] = rescale_factor
__SCREAMING_SNAKE_CASE : List[Any] = do_pad
def SCREAMING_SNAKE_CASE_ ( self :Optional[Any] ):
return {
"do_resize": self.do_resize,
"size": self.size,
"do_normalize": self.do_normalize,
"image_mean": self.image_mean,
"image_std": self.image_std,
"do_rescale": self.do_rescale,
"rescale_factor": self.rescale_factor,
"do_pad": self.do_pad,
}
def SCREAMING_SNAKE_CASE_ ( self :Tuple , _lowerCamelCase :Dict , _lowerCamelCase :List[Any]=False ):
if not batched:
__SCREAMING_SNAKE_CASE : str = image_inputs[0]
if isinstance(_lowerCamelCase , Image.Image ):
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[int] = image.size
else:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = image.shape[1], image.shape[2]
if w < h:
__SCREAMING_SNAKE_CASE : str = int(self.size['''shortest_edge'''] * h / w )
__SCREAMING_SNAKE_CASE : int = self.size['''shortest_edge''']
elif w > h:
__SCREAMING_SNAKE_CASE : Optional[Any] = self.size['''shortest_edge''']
__SCREAMING_SNAKE_CASE : int = int(self.size['''shortest_edge'''] * w / h )
else:
__SCREAMING_SNAKE_CASE : List[str] = self.size['''shortest_edge''']
__SCREAMING_SNAKE_CASE : List[str] = self.size['''shortest_edge''']
else:
__SCREAMING_SNAKE_CASE : Optional[Any] = []
for image in image_inputs:
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = self.get_expected_values([image] )
expected_values.append((expected_height, expected_width) )
__SCREAMING_SNAKE_CASE : Optional[int] = max(_lowerCamelCase , key=lambda _lowerCamelCase : item[0] )[0]
__SCREAMING_SNAKE_CASE : int = max(_lowerCamelCase , key=lambda _lowerCamelCase : item[1] )[1]
return expected_height, expected_width
@require_torch
@require_vision
class snake_case ( __UpperCAmelCase , unittest.TestCase ):
lowerCamelCase__ = YolosImageProcessor if is_vision_available() else None
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
__SCREAMING_SNAKE_CASE : List[Any] = YolosImageProcessingTester(self )
@property
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
return self.image_processor_tester.prepare_image_processor_dict()
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : Any = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(_lowerCamelCase , '''image_mean''' ) )
self.assertTrue(hasattr(_lowerCamelCase , '''image_std''' ) )
self.assertTrue(hasattr(_lowerCamelCase , '''do_normalize''' ) )
self.assertTrue(hasattr(_lowerCamelCase , '''do_resize''' ) )
self.assertTrue(hasattr(_lowerCamelCase , '''size''' ) )
def SCREAMING_SNAKE_CASE_ ( self :Tuple ):
__SCREAMING_SNAKE_CASE : Union[str, Any] = self.image_processing_class.from_dict(self.image_processor_dict )
self.assertEqual(image_processor.size , {'''shortest_edge''': 1_8, '''longest_edge''': 1_3_3_3} )
self.assertEqual(image_processor.do_pad , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Tuple = self.image_processing_class.from_dict(
self.image_processor_dict , size=4_2 , max_size=8_4 , pad_and_return_pixel_mask=_lowerCamelCase )
self.assertEqual(image_processor.size , {'''shortest_edge''': 4_2, '''longest_edge''': 8_4} )
self.assertEqual(image_processor.do_pad , _lowerCamelCase )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
pass
def SCREAMING_SNAKE_CASE_ ( self :int ):
# Initialize image_processing
__SCREAMING_SNAKE_CASE : Optional[int] = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
__SCREAMING_SNAKE_CASE : List[str] = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowerCamelCase )
for image in image_inputs:
self.assertIsInstance(_lowerCamelCase , Image.Image )
# Test not batched input
__SCREAMING_SNAKE_CASE : str = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Dict = self.image_processor_tester.get_expected_values(_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Tuple = self.image_processor_tester.get_expected_values(_lowerCamelCase , batched=_lowerCamelCase )
__SCREAMING_SNAKE_CASE : Union[str, Any] = image_processing(_lowerCamelCase , return_tensors='''pt''' ).pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# Initialize image_processing
__SCREAMING_SNAKE_CASE : Optional[Any] = self.image_processing_class(**self.image_processor_dict )
# create random numpy tensors
__SCREAMING_SNAKE_CASE : Optional[int] = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowerCamelCase , numpify=_lowerCamelCase )
for image in image_inputs:
self.assertIsInstance(_lowerCamelCase , np.ndarray )
# Test not batched input
__SCREAMING_SNAKE_CASE : Dict = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[str] = self.image_processor_tester.get_expected_values(_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
__SCREAMING_SNAKE_CASE : List[Any] = image_processing(_lowerCamelCase , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Any = self.image_processor_tester.get_expected_values(_lowerCamelCase , batched=_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def SCREAMING_SNAKE_CASE_ ( self :Any ):
# Initialize image_processing
__SCREAMING_SNAKE_CASE : str = self.image_processing_class(**self.image_processor_dict )
# create random PyTorch tensors
__SCREAMING_SNAKE_CASE : Any = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowerCamelCase , torchify=_lowerCamelCase )
for image in image_inputs:
self.assertIsInstance(_lowerCamelCase , torch.Tensor )
# Test not batched input
__SCREAMING_SNAKE_CASE : List[str] = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : Optional[Any] = self.image_processor_tester.get_expected_values(_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
__SCREAMING_SNAKE_CASE : Optional[int] = image_processing(_lowerCamelCase , return_tensors='''pt''' ).pixel_values
__SCREAMING_SNAKE_CASE , __SCREAMING_SNAKE_CASE : List[Any] = self.image_processor_tester.get_expected_values(_lowerCamelCase , batched=_lowerCamelCase )
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def SCREAMING_SNAKE_CASE_ ( self :List[str] ):
# Initialize image_processings
__SCREAMING_SNAKE_CASE : Dict = self.image_processing_class(**self.image_processor_dict )
__SCREAMING_SNAKE_CASE : Tuple = self.image_processing_class(do_resize=_lowerCamelCase , do_normalize=_lowerCamelCase , do_rescale=_lowerCamelCase )
# create random PyTorch tensors
__SCREAMING_SNAKE_CASE : int = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowerCamelCase , torchify=_lowerCamelCase )
for image in image_inputs:
self.assertIsInstance(_lowerCamelCase , torch.Tensor )
# Test whether the method "pad" and calling the image processor return the same tensors
__SCREAMING_SNAKE_CASE : Union[str, Any] = image_processing_a.pad(_lowerCamelCase , return_tensors='''pt''' )
__SCREAMING_SNAKE_CASE : Union[str, Any] = image_processing_a(_lowerCamelCase , return_tensors='''pt''' )
self.assertTrue(
torch.allclose(encoded_images_with_method['''pixel_values'''] , encoded_images['''pixel_values'''] , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE_ ( self :int ):
# prepare image and target
__SCREAMING_SNAKE_CASE : Tuple = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
with open('''./tests/fixtures/tests_samples/COCO/coco_annotations.txt''' , '''r''' ) as f:
__SCREAMING_SNAKE_CASE : Tuple = json.loads(f.read() )
__SCREAMING_SNAKE_CASE : Union[str, Any] = {'''image_id''': 3_9_7_6_9, '''annotations''': target}
# encode them
__SCREAMING_SNAKE_CASE : List[Any] = YolosImageProcessor.from_pretrained('''hustvl/yolos-small''' )
__SCREAMING_SNAKE_CASE : Optional[Any] = image_processing(images=_lowerCamelCase , annotations=_lowerCamelCase , return_tensors='''pt''' )
# verify pixel values
__SCREAMING_SNAKE_CASE : Optional[int] = torch.Size([1, 3, 8_0_0, 1_0_6_6] )
self.assertEqual(encoding['''pixel_values'''].shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor([0.2_7_9_6, 0.3_1_3_8, 0.3_4_8_1] )
self.assertTrue(torch.allclose(encoding['''pixel_values'''][0, 0, 0, :3] , _lowerCamelCase , atol=1e-4 ) )
# verify area
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.tensor([5_8_8_7.9_6_0_0, 1_1_2_5_0.2_0_6_1, 4_8_9_3_5_3.8_4_3_8, 8_3_7_1_2_2.7_5_0_0, 1_4_7_9_6_7.5_1_5_6, 1_6_5_7_3_2.3_4_3_8] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''area'''] , _lowerCamelCase ) )
# verify boxes
__SCREAMING_SNAKE_CASE : List[str] = torch.Size([6, 4] )
self.assertEqual(encoding['''labels'''][0]['''boxes'''].shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : Any = torch.tensor([0.5_5_0_3, 0.2_7_6_5, 0.0_6_0_4, 0.2_2_1_5] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''boxes'''][0] , _lowerCamelCase , atol=1e-3 ) )
# verify image_id
__SCREAMING_SNAKE_CASE : Optional[Any] = torch.tensor([3_9_7_6_9] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''image_id'''] , _lowerCamelCase ) )
# verify is_crowd
__SCREAMING_SNAKE_CASE : str = torch.tensor([0, 0, 0, 0, 0, 0] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''iscrowd'''] , _lowerCamelCase ) )
# verify class_labels
__SCREAMING_SNAKE_CASE : Dict = torch.tensor([7_5, 7_5, 6_3, 6_5, 1_7, 1_7] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''class_labels'''] , _lowerCamelCase ) )
# verify orig_size
__SCREAMING_SNAKE_CASE : Union[str, Any] = torch.tensor([4_8_0, 6_4_0] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''orig_size'''] , _lowerCamelCase ) )
# verify size
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor([8_0_0, 1_0_6_6] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''size'''] , _lowerCamelCase ) )
@slow
def SCREAMING_SNAKE_CASE_ ( self :Optional[int] ):
# prepare image, target and masks_path
__SCREAMING_SNAKE_CASE : List[str] = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
with open('''./tests/fixtures/tests_samples/COCO/coco_panoptic_annotations.txt''' , '''r''' ) as f:
__SCREAMING_SNAKE_CASE : Union[str, Any] = json.loads(f.read() )
__SCREAMING_SNAKE_CASE : Dict = {'''file_name''': '''000000039769.png''', '''image_id''': 3_9_7_6_9, '''segments_info''': target}
__SCREAMING_SNAKE_CASE : Optional[int] = pathlib.Path('''./tests/fixtures/tests_samples/COCO/coco_panoptic''' )
# encode them
__SCREAMING_SNAKE_CASE : Any = YolosImageProcessor(format='''coco_panoptic''' )
__SCREAMING_SNAKE_CASE : Dict = image_processing(images=_lowerCamelCase , annotations=_lowerCamelCase , masks_path=_lowerCamelCase , return_tensors='''pt''' )
# verify pixel values
__SCREAMING_SNAKE_CASE : str = torch.Size([1, 3, 8_0_0, 1_0_6_6] )
self.assertEqual(encoding['''pixel_values'''].shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : str = torch.tensor([0.2_7_9_6, 0.3_1_3_8, 0.3_4_8_1] )
self.assertTrue(torch.allclose(encoding['''pixel_values'''][0, 0, 0, :3] , _lowerCamelCase , atol=1e-4 ) )
# verify area
__SCREAMING_SNAKE_CASE : Any = torch.tensor([1_4_7_9_7_9.6_8_7_5, 1_6_5_5_2_7.0_4_6_9, 4_8_4_6_3_8.5_9_3_8, 1_1_2_9_2.9_3_7_5, 5_8_7_9.6_5_6_2, 7_6_3_4.1_1_4_7] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''area'''] , _lowerCamelCase ) )
# verify boxes
__SCREAMING_SNAKE_CASE : List[str] = torch.Size([6, 4] )
self.assertEqual(encoding['''labels'''][0]['''boxes'''].shape , _lowerCamelCase )
__SCREAMING_SNAKE_CASE : int = torch.tensor([0.2_6_2_5, 0.5_4_3_7, 0.4_6_8_8, 0.8_6_2_5] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''boxes'''][0] , _lowerCamelCase , atol=1e-3 ) )
# verify image_id
__SCREAMING_SNAKE_CASE : Dict = torch.tensor([3_9_7_6_9] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''image_id'''] , _lowerCamelCase ) )
# verify is_crowd
__SCREAMING_SNAKE_CASE : List[Any] = torch.tensor([0, 0, 0, 0, 0, 0] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''iscrowd'''] , _lowerCamelCase ) )
# verify class_labels
__SCREAMING_SNAKE_CASE : Optional[Any] = torch.tensor([1_7, 1_7, 6_3, 7_5, 7_5, 9_3] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''class_labels'''] , _lowerCamelCase ) )
# verify masks
__SCREAMING_SNAKE_CASE : Optional[Any] = 8_2_2_8_7_3
self.assertEqual(encoding['''labels'''][0]['''masks'''].sum().item() , _lowerCamelCase )
# verify orig_size
__SCREAMING_SNAKE_CASE : List[str] = torch.tensor([4_8_0, 6_4_0] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''orig_size'''] , _lowerCamelCase ) )
# verify size
__SCREAMING_SNAKE_CASE : Any = torch.tensor([8_0_0, 1_0_6_6] )
self.assertTrue(torch.allclose(encoding['''labels'''][0]['''size'''] , _lowerCamelCase ) )
| 674 | 1 |
"""simple docstring"""
import webbrowser
from sys import argv
from urllib.parse import parse_qs, quote
import requests
from bsa import BeautifulSoup
from fake_useragent import UserAgent
if __name__ == "__main__":
_lowerCamelCase = '''%20'''.join(argv[1:]) if len(argv) > 1 else quote(str(input('''Search: ''')))
print('''Googling.....''')
_lowerCamelCase = f'https://www.google.com/search?q={query}&num=100'
_lowerCamelCase = requests.get(
url,
headers={'''User-Agent''': str(UserAgent().random)},
)
try:
_lowerCamelCase = (
BeautifulSoup(res.text, '''html.parser''')
.find('''div''', attrs={'''class''': '''yuRUbf'''})
.find('''a''')
.get('''href''')
)
except AttributeError:
_lowerCamelCase = parse_qs(
BeautifulSoup(res.text, '''html.parser''')
.find('''div''', attrs={'''class''': '''kCrYT'''})
.find('''a''')
.get('''href''')
)['''url'''][0]
webbrowser.open(link)
| 674 |
"""simple docstring"""
from __future__ import annotations
def lowerCAmelCase_ ( lowercase_ : int ):
'''simple docstring'''
__SCREAMING_SNAKE_CASE : Optional[int] = str(lowercase_ )
return len(lowercase_ ) == 9 and set(lowercase_ ) == set('''123456789''' )
def lowerCAmelCase_ ( ):
'''simple docstring'''
for base_num in range(9999 , 4999 , -1 ):
__SCREAMING_SNAKE_CASE : List[str] = 10_0002 * base_num
if is_9_pandigital(lowercase_ ):
return candidate
for base_num in range(333 , 99 , -1 ):
__SCREAMING_SNAKE_CASE : List[Any] = 100_2003 * base_num
if is_9_pandigital(lowercase_ ):
return candidate
return None
if __name__ == "__main__":
print(f'{solution() = }')
| 674 | 1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.