code
stringlengths 87
55.2k
| code_codestyle
int64 0
349
| style_context
stringlengths 135
49.1k
| style_context_codestyle
int64 0
349
| label
int64 0
1
|
|---|---|---|---|---|
import unittest
from transformers import SPIECE_UNDERLINE, ReformerTokenizer, ReformerTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
_UpperCAmelCase : Optional[Any] = get_tests_dir("""fixtures/test_sentencepiece.model""")
@require_sentencepiece
@require_tokenizers
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = ReformerTokenizer
UpperCAmelCase__ = ReformerTokenizerFast
UpperCAmelCase__ = True
UpperCAmelCase__ = False
UpperCAmelCase__ = True
def A_ ( self : Union[str, Any] ) -> List[str]:
super().setUp()
lowerCamelCase__ : int = ReformerTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
tokenizer.save_pretrained(self.tmpdirname )
def A_ ( self : Tuple ) -> int:
lowerCamelCase__ : Tuple = '<s>'
lowerCamelCase__ : Optional[Any] = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCAmelCase ) , UpperCAmelCase )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : Optional[Any] ) -> List[str]:
lowerCamelCase__ : Union[str, Any] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , '<unk>' )
self.assertEqual(vocab_keys[1] , '<s>' )
self.assertEqual(vocab_keys[-1] , 'j' )
self.assertEqual(len(UpperCAmelCase ) , 1000 )
def A_ ( self : List[str] ) -> Union[str, Any]:
self.assertEqual(self.get_tokenizer().vocab_size , 1000 )
def A_ ( self : Optional[int] ) -> List[str]:
if not self.test_rust_tokenizer:
return
lowerCamelCase__ : List[Any] = self.get_tokenizer()
lowerCamelCase__ : int = self.get_rust_tokenizer()
lowerCamelCase__ : Optional[int] = 'I was born in 92000, and this is falsé.'
lowerCamelCase__ : Union[str, Any] = tokenizer.tokenize(UpperCAmelCase )
lowerCamelCase__ : str = rust_tokenizer.tokenize(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : int = tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase )
lowerCamelCase__ : int = rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.get_rust_tokenizer()
lowerCamelCase__ : Tuple = tokenizer.encode(UpperCAmelCase )
lowerCamelCase__ : List[Any] = rust_tokenizer.encode(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Optional[Any] , UpperCAmelCase : Optional[Any]=15 ) -> List[str]:
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(F"""{tokenizer.__class__.__name__} ({pretrained_name})""" ):
lowerCamelCase__ : Tuple = self.rust_tokenizer_class.from_pretrained(UpperCAmelCase , **UpperCAmelCase )
# Simple input
lowerCamelCase__ : List[str] = 'This is a simple input'
lowerCamelCase__ : Tuple = ['This is a simple input 1', 'This is a simple input 2']
lowerCamelCase__ : Tuple = ('This is a simple input', 'This is a pair')
lowerCamelCase__ : int = [
('This is a simple input 1', 'This is a simple input 2'),
('This is a simple pair 1', 'This is a simple pair 2'),
]
# Simple input tests
self.assertRaises(UpperCAmelCase , tokenizer_r.encode , UpperCAmelCase , max_length=UpperCAmelCase , padding='max_length' )
# Simple input
self.assertRaises(UpperCAmelCase , tokenizer_r.encode_plus , UpperCAmelCase , max_length=UpperCAmelCase , padding='max_length' )
# Simple input
self.assertRaises(
UpperCAmelCase , tokenizer_r.batch_encode_plus , UpperCAmelCase , max_length=UpperCAmelCase , padding='max_length' , )
# Pair input
self.assertRaises(UpperCAmelCase , tokenizer_r.encode , UpperCAmelCase , max_length=UpperCAmelCase , padding='max_length' )
# Pair input
self.assertRaises(UpperCAmelCase , tokenizer_r.encode_plus , UpperCAmelCase , max_length=UpperCAmelCase , padding='max_length' )
# Pair input
self.assertRaises(
UpperCAmelCase , tokenizer_r.batch_encode_plus , UpperCAmelCase , max_length=UpperCAmelCase , padding='max_length' , )
def A_ ( self : Dict ) -> Tuple:
pass
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Any = ReformerTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : str = tokenizer.tokenize('This is a test' )
self.assertListEqual(UpperCAmelCase , ['▁This', '▁is', '▁a', '▁t', 'est'] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(UpperCAmelCase ) , [285, 46, 10, 170, 382] , )
lowerCamelCase__ : Tuple = tokenizer.tokenize('I was born in 92000, and this is falsé.' )
self.assertListEqual(
UpperCAmelCase , [
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',
'.',
] , )
lowerCamelCase__ : List[str] = tokenizer.convert_tokens_to_ids(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
lowerCamelCase__ : Union[str, Any] = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [
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>',
'.',
] , )
@cached_property
def A_ ( self : str ) -> List[str]:
return ReformerTokenizer.from_pretrained('google/reformer-crime-and-punishment' )
@slow
def A_ ( self : List[Any] ) -> Any:
lowerCamelCase__ : str = 'Hello World!'
lowerCamelCase__ : Any = [126, 32, 262, 152, 38, 72, 287]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@slow
def A_ ( self : Any ) -> List[str]:
lowerCamelCase__ : Optional[int] = (
'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'
)
lowerCamelCase__ : Dict = [
108,
265,
24,
111,
4,
258,
156,
35,
28,
275,
3,
259,
297,
260,
84,
4,
35,
110,
44,
8,
259,
91,
268,
21,
11,
209,
274,
109,
266,
277,
117,
86,
93,
315,
258,
278,
258,
277,
258,
0,
258,
288,
258,
319,
258,
0,
258,
0,
258,
0,
258,
0,
258,
287,
258,
315,
258,
289,
258,
278,
99,
269,
266,
262,
8,
259,
241,
4,
217,
230,
268,
266,
55,
168,
106,
75,
193,
266,
223,
27,
49,
26,
282,
25,
264,
299,
19,
26,
0,
258,
277,
117,
86,
93,
176,
183,
270,
11,
262,
42,
61,
265,
]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@require_torch
@slow
def A_ ( self : Dict ) -> Optional[int]:
import torch
from transformers import ReformerConfig, ReformerModel
# Build sequence
lowerCamelCase__ : Tuple = list(self.big_tokenizer.get_vocab().keys() )[:10]
lowerCamelCase__ : int = ' '.join(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.big_tokenizer.encode_plus(UpperCAmelCase , return_tensors='pt' )
lowerCamelCase__ : List[str] = self.big_tokenizer.batch_encode_plus([sequence, sequence] , return_tensors='pt' )
lowerCamelCase__ : Any = ReformerConfig()
# The input gets padded during training so adjust the axial position encodings from the pretrained model value of (512, 1024)
lowerCamelCase__ : int = encoded_sequence['input_ids'].shape
lowerCamelCase__ : str = ReformerModel(UpperCAmelCase )
# Reformer has config.vocab_size == tokenizer.vocab_size == len(tokenizer) - 1 = 320; len(tokenizer) is 321 (including a pad token with id 320)
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**UpperCAmelCase )
model(**UpperCAmelCase )
@slow
def A_ ( self : int ) -> str:
# fmt: off
lowerCamelCase__ : Union[str, Any] = {'input_ids': [[108, 265, 24, 111, 4, 258, 156, 7, 51, 279, 58, 7, 76, 25, 69, 278], [140, 243, 264, 134, 17, 267, 77, 263, 22, 262, 297, 258, 304, 177, 279, 266, 14, 89, 13, 35, 261, 299, 272, 137, 275, 278]], '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]]} # noqa: E501
# fmt: on
# This tokenizer does not know some characters like ")".
# That is the reason why we use very simple texts here.
# Also see https://github.com/huggingface/transformers/pull/11737#issuecomment-850769064
lowerCamelCase__ : Dict = [
'This is a very simple sentence.',
'The quick brown fox jumps over the lazy dog.',
]
self.tokenizer_integration_test_util(
expected_encoding=UpperCAmelCase , model_name='google/reformer-crime-and-punishment' , revision='0e6c3decb8211d49bf881013425dc8b0448b3f5a' , padding=UpperCAmelCase , sequences=UpperCAmelCase , )
| 50
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_base import BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import PaddingStrategy, logging
from .tokenization_realm import RealmTokenizer
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
_UpperCAmelCase : Any = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : int = {
"""vocab_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/vocab.txt"""
),
"""google/realm-orqa-nq-openqa""": """https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-nq-reader""": """https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-openqa""": """https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-reader""": """https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/tokenizer.jsont"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-openqa""": (
"""https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-reader""": (
"""https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-openqa""": (
"""https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-reader""": (
"""https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : int = {
"""google/realm-cc-news-pretrained-embedder""": 5_12,
"""google/realm-cc-news-pretrained-encoder""": 5_12,
"""google/realm-cc-news-pretrained-scorer""": 5_12,
"""google/realm-cc-news-pretrained-openqa""": 5_12,
"""google/realm-orqa-nq-openqa""": 5_12,
"""google/realm-orqa-nq-reader""": 5_12,
"""google/realm-orqa-wq-openqa""": 5_12,
"""google/realm-orqa-wq-reader""": 5_12,
}
_UpperCAmelCase : Any = {
"""google/realm-cc-news-pretrained-embedder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-encoder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-scorer""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-reader""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-reader""": {"""do_lower_case""": True},
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_INIT_CONFIGURATION
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = RealmTokenizer
def __init__( self : Optional[int] , UpperCAmelCase : Tuple=None , UpperCAmelCase : Any=None , UpperCAmelCase : List[Any]=True , UpperCAmelCase : Optional[Any]="[UNK]" , UpperCAmelCase : Any="[SEP]" , UpperCAmelCase : Tuple="[PAD]" , UpperCAmelCase : List[Any]="[CLS]" , UpperCAmelCase : Union[str, Any]="[MASK]" , UpperCAmelCase : Optional[Any]=True , UpperCAmelCase : Any=None , **UpperCAmelCase : Optional[int] , ) -> str:
super().__init__(
UpperCAmelCase , tokenizer_file=UpperCAmelCase , do_lower_case=UpperCAmelCase , unk_token=UpperCAmelCase , sep_token=UpperCAmelCase , pad_token=UpperCAmelCase , cls_token=UpperCAmelCase , mask_token=UpperCAmelCase , tokenize_chinese_chars=UpperCAmelCase , strip_accents=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get('lowercase' , UpperCAmelCase ) != do_lower_case
or normalizer_state.get('strip_accents' , UpperCAmelCase ) != strip_accents
or normalizer_state.get('handle_chinese_chars' , UpperCAmelCase ) != tokenize_chinese_chars
):
lowerCamelCase__ : Optional[int] = getattr(UpperCAmelCase , normalizer_state.pop('type' ) )
lowerCamelCase__ : Optional[Any] = do_lower_case
lowerCamelCase__ : str = strip_accents
lowerCamelCase__ : Optional[Any] = tokenize_chinese_chars
lowerCamelCase__ : int = normalizer_class(**UpperCAmelCase )
lowerCamelCase__ : str = do_lower_case
def A_ ( self : Optional[int] , UpperCAmelCase : int , **UpperCAmelCase : int ) -> List[Any]:
lowerCamelCase__ : List[Any] = PaddingStrategy.MAX_LENGTH
lowerCamelCase__ : Optional[int] = text
lowerCamelCase__ : Dict = kwargs.pop('text_pair' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = kwargs.pop('return_tensors' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = {
'input_ids': [],
'attention_mask': [],
'token_type_ids': [],
}
for idx, candidate_text in enumerate(UpperCAmelCase ):
if batch_text_pair is not None:
lowerCamelCase__ : Tuple = batch_text_pair[idx]
else:
lowerCamelCase__ : Dict = None
lowerCamelCase__ : Optional[int] = super().__call__(UpperCAmelCase , UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Any = encoded_candidates.get('input_ids' )
lowerCamelCase__ : Union[str, Any] = encoded_candidates.get('attention_mask' )
lowerCamelCase__ : Tuple = encoded_candidates.get('token_type_ids' )
if encoded_input_ids is not None:
output_data["input_ids"].append(UpperCAmelCase )
if encoded_attention_mask is not None:
output_data["attention_mask"].append(UpperCAmelCase )
if encoded_token_type_ids is not None:
output_data["token_type_ids"].append(UpperCAmelCase )
lowerCamelCase__ : int = {key: item for key, item in output_data.items() if len(UpperCAmelCase ) != 0}
return BatchEncoding(UpperCAmelCase , tensor_type=UpperCAmelCase )
def A_ ( self : int , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Optional[int]=None ) -> List[str]:
lowerCamelCase__ : Tuple = [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 A_ ( self : Tuple , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : List[Any] = [self.sep_token_id]
lowerCamelCase__ : Optional[int] = [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 A_ ( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
lowerCamelCase__ : int = self._tokenizer.model.save(UpperCAmelCase , name=UpperCAmelCase )
return tuple(UpperCAmelCase )
| 50
| 1
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_pegasus import PegasusTokenizer
else:
_UpperCAmelCase : Dict = None
_UpperCAmelCase : Union[str, Any] = logging.get_logger(__name__)
_UpperCAmelCase : Optional[int] = """▁"""
_UpperCAmelCase : Optional[Any] = {"""vocab_file""": """spiece.model""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : Union[str, Any] = {
"""vocab_file""": {"""google/pegasus-xsum""": """https://huggingface.co/google/pegasus-xsum/resolve/main/spiece.model"""},
"""tokenizer_file""": {
"""google/pegasus-xsum""": """https://huggingface.co/google/pegasus-xsum/resolve/main/tokenizer.json"""
},
}
_UpperCAmelCase : List[Any] = {
"""google/pegasus-xsum""": 5_12,
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = PegasusTokenizer
UpperCAmelCase__ = ["""input_ids""", """attention_mask"""]
def __init__( self : int , UpperCAmelCase : Optional[int]=None , UpperCAmelCase : List[Any]=None , UpperCAmelCase : Any="<pad>" , UpperCAmelCase : Union[str, Any]="</s>" , UpperCAmelCase : Dict="<unk>" , UpperCAmelCase : Dict="<mask_2>" , UpperCAmelCase : List[str]="<mask_1>" , UpperCAmelCase : Dict=None , UpperCAmelCase : str=103 , **UpperCAmelCase : str , ) -> Any:
lowerCamelCase__ : Optional[int] = offset
if additional_special_tokens is not None:
if not isinstance(UpperCAmelCase , UpperCAmelCase ):
raise TypeError(
F"""additional_special_tokens should be of type {type(UpperCAmelCase )}, but is"""
F""" {type(UpperCAmelCase )}""" )
lowerCamelCase__ : Optional[Any] = (
([mask_token_sent] + additional_special_tokens)
if mask_token_sent not in additional_special_tokens and mask_token_sent is not None
else additional_special_tokens
)
# fill additional tokens with ..., <unk_token_102> in case not all additional tokens are already taken
additional_special_tokens_extended += [
F"""<unk_{i}>""" for i in range(len(UpperCAmelCase ) , self.offset - 1 )
]
if len(set(UpperCAmelCase ) ) != len(UpperCAmelCase ):
raise ValueError(
'Please make sure that the provided additional_special_tokens do not contain an incorrectly'
F""" shifted list of <unk_x> tokens. Found {additional_special_tokens_extended}.""" )
lowerCamelCase__ : Optional[Any] = additional_special_tokens_extended
else:
lowerCamelCase__ : Tuple = [mask_token_sent] if mask_token_sent is not None else []
additional_special_tokens += [F"""<unk_{i}>""" for i in range(2 , self.offset )]
super().__init__(
UpperCAmelCase , tokenizer_file=UpperCAmelCase , pad_token=UpperCAmelCase , eos_token=UpperCAmelCase , unk_token=UpperCAmelCase , mask_token=UpperCAmelCase , mask_token_sent=UpperCAmelCase , offset=UpperCAmelCase , additional_special_tokens=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : int = vocab_file
lowerCamelCase__ : Any = False if not self.vocab_file else True
def A_ ( self : Tuple , UpperCAmelCase : str ) -> Optional[int]:
lowerCamelCase__ : Optional[Any] = set(self.all_special_ids ) # call it once instead of inside list comp
all_special_ids.remove(self.unk_token_id ) # <unk> is only sometimes special
if all_special_ids != set(range(len(self.additional_special_tokens ) + 3 ) ):
raise ValueError(
'There should be 3 special tokens: mask_token, pad_token, and eos_token +'
F""" {len(self.additional_special_tokens )} additional_special_tokens, but got {all_special_ids}""" )
return [1 if x in all_special_ids else 0 for x in seq]
def A_ ( self : List[Any] , UpperCAmelCase : List , UpperCAmelCase : Optional[List] = None , UpperCAmelCase : bool = False ) -> List[int]:
if already_has_special_tokens:
return self._special_token_mask(UpperCAmelCase )
elif token_ids_a is None:
return self._special_token_mask(UpperCAmelCase ) + [1]
else:
return self._special_token_mask(token_ids_a + token_ids_a ) + [1]
def A_ ( self : str , UpperCAmelCase : Optional[Any] , UpperCAmelCase : int=None ) -> List[int]:
if token_ids_a is None:
return token_ids_a + [self.eos_token_id]
# We don't expect to process pairs, but leave the pair logic for API consistency
return token_ids_a + token_ids_a + [self.eos_token_id]
def A_ ( self : List[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not self.can_save_slow_tokenizer:
raise ValueError(
'Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '
'tokenizer.' )
if not os.path.isdir(UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCamelCase__ : List[Any] = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCAmelCase ):
copyfile(self.vocab_file , UpperCAmelCase )
return (out_vocab_file,)
| 50
|
import logging
import os
from dataclasses import dataclass
from typing import List, Optional, Union
import tqdm
from filelock import FileLock
from transformers import (
BartTokenizer,
BartTokenizerFast,
DataProcessor,
PreTrainedTokenizer,
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
is_tf_available,
is_torch_available,
)
_UpperCAmelCase : Optional[Any] = logging.getLogger(__name__)
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
if is_torch_available():
import torch
from torch.utils.data import Dataset
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
def __init__( self : int , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : List[str]=False , UpperCAmelCase : bool = False , ) -> List[str]:
lowerCamelCase__ : int = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = os.path.join(
UpperCAmelCase , 'cached_{}_{}_{}_{}'.format(
'dev' if evaluate else 'train' , tokenizer.__class__.__name__ , str(UpperCAmelCase ) , UpperCAmelCase , ) , )
lowerCamelCase__ : int = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = label_list[2], label_list[1]
lowerCamelCase__ : List[str] = label_list
# Make sure only the first process in distributed training processes the dataset,
# and the others will use the cache.
lowerCamelCase__ : str = cached_features_file + '.lock'
with FileLock(UpperCAmelCase ):
if os.path.exists(UpperCAmelCase ) and not overwrite_cache:
logger.info(F"""Loading features from cached file {cached_features_file}""" )
lowerCamelCase__ : int = torch.load(UpperCAmelCase )
else:
logger.info(F"""Creating features from dataset file at {data_dir}""" )
lowerCamelCase__ : str = (
processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
)
logger.info('Training examples: %s' , len(UpperCAmelCase ) )
lowerCamelCase__ : Dict = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
logger.info('Saving features into cached file %s' , UpperCAmelCase )
torch.save(self.features , UpperCAmelCase )
def __len__( self : Optional[int] ) -> Optional[Any]:
return len(self.features )
def __getitem__( self : Tuple , UpperCAmelCase : Dict ) -> InputFeatures:
return self.features[i]
def A_ ( self : int ) -> int:
return self.label_list
if is_tf_available():
import tensorflow as tf
class lowerCAmelCase :
UpperCAmelCase__ = 42
def __init__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = 128 , UpperCAmelCase : Any=False , UpperCAmelCase : bool = False , ) -> Union[str, Any]:
lowerCamelCase__ : Any = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : str = label_list[2], label_list[1]
lowerCamelCase__ : Optional[int] = label_list
lowerCamelCase__ : int = processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def gen():
for ex_index, ex in tqdm.tqdm(enumerate(self.features ) , desc='convert examples to features' ):
if ex_index % 10000 == 0:
logger.info('Writing example %d of %d' % (ex_index, len(UpperCAmelCase )) )
yield (
{
"example_id": 0,
"input_ids": ex.input_ids,
"attention_mask": ex.attention_mask,
"token_type_ids": ex.token_type_ids,
},
ex.label,
)
lowerCamelCase__ : Optional[int] = tf.data.Dataset.from_generator(
UpperCAmelCase , (
{
'example_id': tf.intaa,
'input_ids': tf.intaa,
'attention_mask': tf.intaa,
'token_type_ids': tf.intaa,
},
tf.intaa,
) , (
{
'example_id': tf.TensorShape([] ),
'input_ids': tf.TensorShape([None, None] ),
'attention_mask': tf.TensorShape([None, None] ),
'token_type_ids': tf.TensorShape([None, None] ),
},
tf.TensorShape([] ),
) , )
def A_ ( self : Any ) -> Any:
return self.dataset
def __len__( self : Tuple ) -> int:
return len(self.features )
def __getitem__( self : List[str] , UpperCAmelCase : Any ) -> InputFeatures:
return self.features[i]
def A_ ( self : Dict ) -> str:
return self.label_list
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : int , UpperCAmelCase : List[Any] ) -> int:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_train_set.txt' ) ) , 'train' )
def A_ ( self : Any , UpperCAmelCase : int ) -> List[Any]:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_evaluation_set.txt' ) ) , 'dev' )
def A_ ( self : Any ) -> List[Any]:
return ["contradiction", "entailment", "neutral"]
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : List[str] ) -> List[str]:
lowerCamelCase__ : List[str] = []
for i, line in enumerate(UpperCAmelCase ):
if i == 0:
continue
lowerCamelCase__ : Tuple = '%s-%s' % (set_type, line[0])
lowerCamelCase__ : str = line[5]
lowerCamelCase__ : Dict = line[6]
lowerCamelCase__ : int = line[7][2:] if line[7].startswith('ex' ) else line[7]
lowerCamelCase__ : Dict = line[0]
examples.append(InputExample(guid=UpperCAmelCase , text_a=UpperCAmelCase , text_b=UpperCAmelCase , label=UpperCAmelCase , pairID=UpperCAmelCase ) )
return examples
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ) -> Optional[int]:
lowerCamelCase__ : int = {label: i for i, label in enumerate(_UpperCAmelCase )}
lowerCamelCase__ : List[Any] = []
for ex_index, example in tqdm.tqdm(enumerate(_UpperCAmelCase ) , desc='convert examples to features' ):
if ex_index % 1_0000 == 0:
logger.info('Writing example %d' % (ex_index) )
lowerCamelCase__ : List[Any] = tokenizer(
example.text_a , example.text_b , add_special_tokens=_UpperCAmelCase , max_length=_UpperCAmelCase , padding='max_length' , truncation=_UpperCAmelCase , return_overflowing_tokens=_UpperCAmelCase , )
lowerCamelCase__ : List[str] = label_map[example.label] if example.label in label_map else 0
lowerCamelCase__ : Optional[int] = int(example.pairID )
features.append(InputFeatures(**_UpperCAmelCase , label=_UpperCAmelCase , pairID=_UpperCAmelCase ) )
for i, example in enumerate(examples[:5] ):
logger.info('*** Example ***' )
logger.info(F"""guid: {example}""" )
logger.info(F"""features: {features[i]}""" )
return features
_UpperCAmelCase : str = {
"""hans""": 3,
}
_UpperCAmelCase : List[Any] = {
"""hans""": HansProcessor,
}
| 50
| 1
|
import argparse
import torch
from torch import nn
from transformers import MaMaaaConfig, MaMaaaForConditionalGeneration
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Optional[int] = [
'encoder.version',
'decoder.version',
'model.encoder.version',
'model.decoder.version',
'decoder.output_projection.weight',
'_float_tensor',
'encoder.embed_positions._float_tensor',
'decoder.embed_positions._float_tensor',
]
for k in ignore_keys:
state_dict.pop(_UpperCAmelCase , _UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[Any]:
lowerCamelCase__ , lowerCamelCase__ : List[str] = emb.weight.shape
lowerCamelCase__ : Tuple = nn.Linear(_UpperCAmelCase , _UpperCAmelCase , bias=_UpperCAmelCase )
lowerCamelCase__ : Dict = emb.weight.data
return lin_layer
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Tuple = torch.load(_UpperCAmelCase , map_location='cpu' )
lowerCamelCase__ : List[str] = mam_aaa['args'] or mam_aaa['cfg']['model']
lowerCamelCase__ : Optional[int] = mam_aaa['model']
remove_ignore_keys_(_UpperCAmelCase )
lowerCamelCase__ : str = state_dict['encoder.embed_tokens.weight'].shape[0]
lowerCamelCase__ : Union[str, Any] = MaMaaaConfig(
vocab_size=_UpperCAmelCase , max_position_embeddings=1024 , encoder_layers=args.encoder_layers , decoder_layers=args.decoder_layers , encoder_attention_heads=args.encoder_attention_heads , decoder_attention_heads=args.decoder_attention_heads , encoder_ffn_dim=args.encoder_ffn_embed_dim , decoder_ffn_dim=args.decoder_ffn_embed_dim , d_model=args.encoder_embed_dim , encoder_layerdrop=args.encoder_layerdrop , decoder_layerdrop=args.decoder_layerdrop , dropout=args.dropout , attention_dropout=args.attention_dropout , activation_dropout=args.activation_dropout , activation_function='relu' , )
lowerCamelCase__ : Optional[Any] = state_dict['decoder.embed_tokens.weight']
lowerCamelCase__ : Union[str, Any] = MaMaaaForConditionalGeneration(_UpperCAmelCase )
model.model.load_state_dict(_UpperCAmelCase , strict=_UpperCAmelCase )
lowerCamelCase__ : List[str] = make_linear_from_emb(model.model.shared )
return model
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""fairseq_path""", type=str, help="""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.""")
_UpperCAmelCase : str = parser.parse_args()
_UpperCAmelCase : Optional[Any] = convert_fairseq_mamaaa_checkpoint_from_disk(args.fairseq_pathß)
model.save_pretrained(args.pytorch_dump_folder_path)
| 50
|
import unittest
from transformers import BertGenerationTokenizer
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
_UpperCAmelCase : Optional[Any] = """▁"""
_UpperCAmelCase : Optional[Any] = get_tests_dir("""fixtures/test_sentencepiece.model""")
@require_sentencepiece
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = BertGenerationTokenizer
UpperCAmelCase__ = False
UpperCAmelCase__ = True
def A_ ( self : List[Any] ) -> List[str]:
super().setUp()
lowerCamelCase__ : Dict = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
tokenizer.save_pretrained(self.tmpdirname )
def A_ ( self : Optional[Any] ) -> Dict:
lowerCamelCase__ : List[str] = '<s>'
lowerCamelCase__ : Optional[int] = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCAmelCase ) , UpperCAmelCase )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : List[str] ) -> Optional[int]:
lowerCamelCase__ : Optional[Any] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , '<unk>' )
self.assertEqual(vocab_keys[1] , '<s>' )
self.assertEqual(vocab_keys[-1] , '<pad>' )
self.assertEqual(len(UpperCAmelCase ) , 1002 )
def A_ ( self : List[Any] ) -> Optional[Any]:
self.assertEqual(self.get_tokenizer().vocab_size , 1000 )
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : List[str] = tokenizer.tokenize('This is a test' )
self.assertListEqual(UpperCAmelCase , ['▁This', '▁is', '▁a', '▁t', 'est'] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(UpperCAmelCase ) , [285, 46, 10, 170, 382] , )
lowerCamelCase__ : List[str] = tokenizer.tokenize('I was born in 92000, and this is falsé.' )
self.assertListEqual(
UpperCAmelCase , [
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',
'.',
] , )
lowerCamelCase__ : Optional[int] = tokenizer.convert_tokens_to_ids(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
lowerCamelCase__ : Optional[Any] = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [
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>',
'.',
] , )
@cached_property
def A_ ( self : Dict ) -> Tuple:
return BertGenerationTokenizer.from_pretrained('google/bert_for_seq_generation_L-24_bbc_encoder' )
@slow
def A_ ( self : Optional[int] ) -> List[str]:
lowerCamelCase__ : Union[str, Any] = 'Hello World!'
lowerCamelCase__ : Dict = [18536, 2260, 101]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@slow
def A_ ( self : Optional[Any] ) -> str:
lowerCamelCase__ : List[Any] = (
'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'
)
lowerCamelCase__ : Any = [
871,
419,
358,
946,
991,
2521,
452,
358,
1357,
387,
7751,
3536,
112,
985,
456,
126,
865,
938,
5400,
5734,
458,
1368,
467,
786,
2462,
5246,
1159,
633,
865,
4519,
457,
582,
852,
2557,
427,
916,
508,
405,
34324,
497,
391,
408,
11342,
1244,
385,
100,
938,
985,
456,
574,
362,
12597,
3200,
3129,
1172,
]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@require_torch
@slow
def A_ ( self : int ) -> Optional[Any]:
import torch
from transformers import BertGenerationConfig, BertGenerationEncoder
# Build sequence
lowerCamelCase__ : str = list(self.big_tokenizer.get_vocab().keys() )[:10]
lowerCamelCase__ : int = ' '.join(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.big_tokenizer.encode_plus(UpperCAmelCase , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.big_tokenizer.batch_encode_plus(
[sequence + ' ' + sequence] , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Tuple = BertGenerationConfig()
lowerCamelCase__ : Optional[Any] = BertGenerationEncoder(UpperCAmelCase )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**UpperCAmelCase )
model(**UpperCAmelCase )
@slow
def A_ ( self : Optional[int] ) -> List[Any]:
# fmt: off
lowerCamelCase__ : Any = {'input_ids': [[39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114], [448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 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], [484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 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]], '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, 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, 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=UpperCAmelCase , model_name='google/bert_for_seq_generation_L-24_bbc_encoder' , revision='c817d1fd1be2ffa69431227a1fe320544943d4db' , )
| 50
| 1
|
from __future__ import annotations
import numpy as np
from numpy import floataa
from numpy.typing import NDArray
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ) -> list[float]:
lowerCamelCase__ , lowerCamelCase__ : List[str] = coefficient_matrix.shape
lowerCamelCase__ , lowerCamelCase__ : Tuple = constant_matrix.shape
if rowsa != colsa:
lowerCamelCase__ : List[Any] = F"""Coefficient matrix dimensions must be nxn but received {rowsa}x{colsa}"""
raise ValueError(_UpperCAmelCase )
if colsa != 1:
lowerCamelCase__ : Tuple = F"""Constant matrix must be nx1 but received {rowsa}x{colsa}"""
raise ValueError(_UpperCAmelCase )
if rowsa != rowsa:
lowerCamelCase__ : Optional[Any] = (
'Coefficient and constant matrices dimensions must be nxn and nx1 but '
F"""received {rowsa}x{colsa} and {rowsa}x{colsa}"""
)
raise ValueError(_UpperCAmelCase )
if len(_UpperCAmelCase ) != rowsa:
lowerCamelCase__ : Tuple = (
'Number of initial values must be equal to number of rows in coefficient '
F"""matrix but received {len(_UpperCAmelCase )} and {rowsa}"""
)
raise ValueError(_UpperCAmelCase )
if iterations <= 0:
raise ValueError('Iterations must be at least 1' )
lowerCamelCase__ : NDArray[floataa] = np.concatenate(
(coefficient_matrix, constant_matrix) , axis=1 )
lowerCamelCase__ , lowerCamelCase__ : List[str] = table.shape
strictly_diagonally_dominant(_UpperCAmelCase )
# Iterates the whole matrix for given number of times
for _ in range(_UpperCAmelCase ):
lowerCamelCase__ : str = []
for row in range(_UpperCAmelCase ):
lowerCamelCase__ : Dict = 0
for col in range(_UpperCAmelCase ):
if col == row:
lowerCamelCase__ : int = table[row][col]
elif col == cols - 1:
lowerCamelCase__ : Optional[int] = table[row][col]
else:
temp += (-1) * table[row][col] * init_val[col]
lowerCamelCase__ : Optional[Any] = (temp + val) / denom
new_val.append(_UpperCAmelCase )
lowerCamelCase__ : List[str] = new_val
return [float(_UpperCAmelCase ) for i in new_val]
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
lowerCamelCase__ , lowerCamelCase__ : List[Any] = table.shape
lowerCamelCase__ : Optional[int] = True
for i in range(0 , _UpperCAmelCase ):
lowerCamelCase__ : Optional[Any] = 0
for j in range(0 , cols - 1 ):
if i == j:
continue
else:
total += table[i][j]
if table[i][i] <= total:
raise ValueError('Coefficient matrix is not strictly diagonally dominant' )
return is_diagonally_dominant
# Test Cases
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
|
import os
import tempfile
from functools import partial
from unittest import TestCase
from unittest.mock import patch
import numpy as np
import pytest
from datasets.arrow_dataset import Dataset
from datasets.search import ElasticSearchIndex, FaissIndex, MissingIndex
from .utils import require_elasticsearch, require_faiss
_UpperCAmelCase : str = pytest.mark.integration
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : List[Any] ) -> Union[str, Any]:
lowerCamelCase__ : List[Any] = Dataset.from_dict({'filename': ['my_name-train' + '_' + str(UpperCAmelCase ) for x in np.arange(30 ).tolist()]} )
return dset
def A_ ( self : Optional[Any] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
lowerCamelCase__ : List[Any] = dset.map(
lambda UpperCAmelCase , UpperCAmelCase : {"vecs": i * np.ones(5 , dtype=np.floataa )} , with_indices=UpperCAmelCase , keep_in_memory=UpperCAmelCase )
lowerCamelCase__ : Tuple = dset.add_faiss_index('vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
dset.drop_index('vecs' )
def A_ ( self : Union[str, Any] ) -> int:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT , )
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : List[str] ) -> Tuple:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , metric_type=faiss.METRIC_INNER_PRODUCT , )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
dset.save_faiss_index('vecs' , tmp_file.name )
dset.load_faiss_index('vecs2' , tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs2' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' )
dset.drop_index('vecs' )
self.assertRaises(UpperCAmelCase , partial(dset.get_nearest_examples , 'vecs2' , np.ones(5 , dtype=np.floataa ) ) )
def A_ ( self : Dict ) -> Dict:
from elasticsearch import Elasticsearch
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : List[Any] = {'acknowledged': True}
mocked_bulk.return_value([(True, None)] * 30 )
lowerCamelCase__ : int = {'hits': {'hits': [{'_score': 1, '_id': 29}]}}
lowerCamelCase__ : List[str] = Elasticsearch()
dset.add_elasticsearch_index('filename' , es_client=UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Dict = dset.get_nearest_examples('filename' , 'my_name-train_29' )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Any ) -> Dict:
import faiss
lowerCamelCase__ : Tuple = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
# add vectors
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsNotNone(index.faiss_index )
self.assertEqual(index.faiss_index.ntotal , 5 )
index.add_vectors(np.zeros((5, 5) , dtype=np.floataa ) )
self.assertEqual(index.faiss_index.ntotal , 10 )
# single query
lowerCamelCase__ : int = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Any = 1
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = index.search(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search , query.reshape(-1 , 1 ) )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
# batched queries
lowerCamelCase__ : str = np.eye(5 , dtype=np.floataa )[::-1]
lowerCamelCase__ , lowerCamelCase__ : Dict = index.search_batch(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search_batch , queries[0] )
lowerCamelCase__ : str = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[str] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([4, 3, 2, 1, 0] , UpperCAmelCase )
def A_ ( self : List[Any] ) -> List[Any]:
import faiss
lowerCamelCase__ : Any = FaissIndex(string_factory='Flat' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
lowerCamelCase__ : Tuple = FaissIndex(string_factory='LSH' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexLSH )
with self.assertRaises(UpperCAmelCase ):
lowerCamelCase__ : List[str] = FaissIndex(string_factory='Flat' , custom_index=faiss.IndexFlat(5 ) )
def A_ ( self : List[str] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Optional[Any] = faiss.IndexFlat(5 )
lowerCamelCase__ : int = FaissIndex(custom_index=UpperCAmelCase )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
def A_ ( self : Any ) -> Optional[int]:
import faiss
lowerCamelCase__ : Any = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
index.save(tmp_file.name )
lowerCamelCase__ : List[Any] = FaissIndex.load(tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ : List[str] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Tuple = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(UpperCAmelCase )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
@require_faiss
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Any:
import faiss
lowerCamelCase__ : Optional[Any] = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
lowerCamelCase__ : Optional[int] = 'index.faiss'
lowerCamelCase__ : Optional[Any] = F"""mock://{index_name}"""
index.save(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Tuple = FaissIndex.load(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Optional[int] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Dict = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(_UpperCAmelCase )
assert scores[0] > 0
assert indices[0] == 1
@require_elasticsearch
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Dict ) -> List[Any]:
from elasticsearch import Elasticsearch
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : Any = Elasticsearch()
lowerCamelCase__ : Tuple = {'acknowledged': True}
lowerCamelCase__ : Tuple = ElasticSearchIndex(es_client=UpperCAmelCase )
mocked_bulk.return_value([(True, None)] * 3 )
index.add_documents(['foo', 'bar', 'foobar'] )
# single query
lowerCamelCase__ : Optional[int] = 'foo'
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search(UpperCAmelCase )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# single query with timeout
lowerCamelCase__ : Any = 'foo'
lowerCamelCase__ : List[str] = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Tuple = index.search(UpperCAmelCase , request_timeout=30 )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# batched queries
lowerCamelCase__ : List[str] = ['foo', 'bar', 'foobar']
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : str = index.search_batch(UpperCAmelCase )
lowerCamelCase__ : List[str] = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[Any] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
# batched queries with timeout
lowerCamelCase__ : str = ['foo', 'bar', 'foobar']
lowerCamelCase__ : List[Any] = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search_batch(UpperCAmelCase , request_timeout=30 )
lowerCamelCase__ : Optional[Any] = [scores[0] for scores in total_scores]
lowerCamelCase__ : Dict = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
| 50
| 1
|
import argparse
import collections
import json
from pathlib import Path
import requests
import torch
import yaml
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import (
MobileViTImageProcessor,
MobileViTVaConfig,
MobileViTVaForImageClassification,
MobileViTVaForSemanticSegmentation,
)
from transformers.utils import logging
logging.set_verbosity_info()
_UpperCAmelCase : Union[str, Any] = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
print('Loading config file...' )
def flatten_yaml_as_dict(_UpperCAmelCase , _UpperCAmelCase="" , _UpperCAmelCase="." ):
lowerCamelCase__ : List[Any] = []
for k, v in d.items():
lowerCamelCase__ : Tuple = parent_key + sep + k if parent_key else k
if isinstance(_UpperCAmelCase , collections.abc.MutableMapping ):
items.extend(flatten_yaml_as_dict(_UpperCAmelCase , _UpperCAmelCase , sep=_UpperCAmelCase ).items() )
else:
items.append((new_key, v) )
return dict(_UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = argparse.Namespace()
with open(_UpperCAmelCase , 'r' ) as yaml_file:
try:
lowerCamelCase__ : Dict = yaml.load(_UpperCAmelCase , Loader=yaml.FullLoader )
lowerCamelCase__ : Optional[Any] = flatten_yaml_as_dict(_UpperCAmelCase )
for k, v in flat_cfg.items():
setattr(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
except yaml.YAMLError as exc:
logger.error('Error while loading config file: {}. Error message: {}'.format(_UpperCAmelCase , str(_UpperCAmelCase ) ) )
return config
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> str:
lowerCamelCase__ : List[Any] = MobileViTVaConfig()
lowerCamelCase__ : Any = False
# dataset
if task_name.startswith('imagenet1k_' ):
lowerCamelCase__ : Dict = 1000
if int(task_name.strip().split('_' )[-1] ) == 384:
lowerCamelCase__ : Optional[int] = 384
else:
lowerCamelCase__ : Optional[int] = 256
lowerCamelCase__ : Tuple = 'imagenet-1k-id2label.json'
elif task_name.startswith('imagenet21k_to_1k_' ):
lowerCamelCase__ : Any = 2_1000
if int(task_name.strip().split('_' )[-1] ) == 384:
lowerCamelCase__ : List[Any] = 384
else:
lowerCamelCase__ : Union[str, Any] = 256
lowerCamelCase__ : int = 'imagenet-22k-id2label.json'
elif task_name.startswith('ade20k_' ):
lowerCamelCase__ : List[str] = 151
lowerCamelCase__ : List[str] = 512
lowerCamelCase__ : Tuple = 'ade20k-id2label.json'
lowerCamelCase__ : List[Any] = True
elif task_name.startswith('voc_' ):
lowerCamelCase__ : str = 21
lowerCamelCase__ : Optional[Any] = 512
lowerCamelCase__ : Union[str, Any] = 'pascal-voc-id2label.json'
lowerCamelCase__ : Union[str, Any] = True
# orig_config
lowerCamelCase__ : Tuple = load_orig_config_file(_UpperCAmelCase )
assert getattr(_UpperCAmelCase , 'model.classification.name' , -1 ) == "mobilevit_v2", "Invalid model"
lowerCamelCase__ : Tuple = getattr(_UpperCAmelCase , 'model.classification.mitv2.width_multiplier' , 1.0 )
assert (
getattr(_UpperCAmelCase , 'model.classification.mitv2.attn_norm_layer' , -1 ) == "layer_norm_2d"
), "Norm layers other than layer_norm_2d is not supported"
lowerCamelCase__ : str = getattr(_UpperCAmelCase , 'model.classification.activation.name' , 'swish' )
# config.image_size == getattr(orig_config, 'sampler.bs.crop_size_width', 256)
if is_segmentation_model:
lowerCamelCase__ : Union[str, Any] = getattr(_UpperCAmelCase , 'model.segmentation.output_stride' , 16 )
if "_deeplabv3" in task_name:
lowerCamelCase__ : List[str] = getattr(_UpperCAmelCase , 'model.segmentation.deeplabv3.aspp_rates' , [12, 24, 36] )
lowerCamelCase__ : List[str] = getattr(_UpperCAmelCase , 'model.segmentation.deeplabv3.aspp_out_channels' , 512 )
lowerCamelCase__ : Any = getattr(_UpperCAmelCase , 'model.segmentation.deeplabv3.aspp_dropout' , 0.1 )
# id2label
lowerCamelCase__ : List[Any] = 'huggingface/label-files'
lowerCamelCase__ : List[Any] = json.load(open(hf_hub_download(_UpperCAmelCase , _UpperCAmelCase , repo_type='dataset' ) , 'r' ) )
lowerCamelCase__ : Optional[int] = {int(_UpperCAmelCase ): v for k, v in idalabel.items()}
lowerCamelCase__ : List[str] = idalabel
lowerCamelCase__ : Optional[int] = {v: k for k, v in idalabel.items()}
return config
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> int:
lowerCamelCase__ : Union[str, Any] = dct.pop(_UpperCAmelCase )
lowerCamelCase__ : Optional[int] = val
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=False ) -> str:
if base_model:
lowerCamelCase__ : Optional[Any] = ''
else:
lowerCamelCase__ : Tuple = 'mobilevitv2.'
lowerCamelCase__ : List[str] = []
for k in state_dict.keys():
if k[:8] == "encoder.":
lowerCamelCase__ : Optional[Any] = k[8:]
else:
lowerCamelCase__ : str = k
if ".block." in k:
lowerCamelCase__ : str = k_new.replace('.block.' , '.' )
if ".conv." in k:
lowerCamelCase__ : List[str] = k_new.replace('.conv.' , '.convolution.' )
if ".norm." in k:
lowerCamelCase__ : Optional[int] = k_new.replace('.norm.' , '.normalization.' )
if "conv_1." in k:
lowerCamelCase__ : Any = k_new.replace('conv_1.' , F"""{model_prefix}conv_stem.""" )
for i in [1, 2]:
if F"""layer_{i}.""" in k:
lowerCamelCase__ : int = k_new.replace(F"""layer_{i}.""" , F"""{model_prefix}encoder.layer.{i-1}.layer.""" )
if ".exp_1x1." in k:
lowerCamelCase__ : Optional[Any] = k_new.replace('.exp_1x1.' , '.expand_1x1.' )
if ".red_1x1." in k:
lowerCamelCase__ : List[Any] = k_new.replace('.red_1x1.' , '.reduce_1x1.' )
for i in [3, 4, 5]:
if F"""layer_{i}.0.""" in k:
lowerCamelCase__ : Dict = k_new.replace(F"""layer_{i}.0.""" , F"""{model_prefix}encoder.layer.{i-1}.downsampling_layer.""" )
if F"""layer_{i}.1.local_rep.0.""" in k:
lowerCamelCase__ : Any = k_new.replace(F"""layer_{i}.1.local_rep.0.""" , F"""{model_prefix}encoder.layer.{i-1}.conv_kxk.""" )
if F"""layer_{i}.1.local_rep.1.""" in k:
lowerCamelCase__ : List[Any] = k_new.replace(F"""layer_{i}.1.local_rep.1.""" , F"""{model_prefix}encoder.layer.{i-1}.conv_1x1.""" )
for i in [3, 4, 5]:
if i == 3:
lowerCamelCase__ : Dict = [0, 1]
elif i == 4:
lowerCamelCase__ : Optional[int] = [0, 1, 2, 3]
elif i == 5:
lowerCamelCase__ : List[str] = [0, 1, 2]
for j in j_in:
if F"""layer_{i}.1.global_rep.{j}.""" in k:
lowerCamelCase__ : List[Any] = k_new.replace(
F"""layer_{i}.1.global_rep.{j}.""" , F"""{model_prefix}encoder.layer.{i-1}.transformer.layer.{j}.""" )
if F"""layer_{i}.1.global_rep.{j+1}.""" in k:
lowerCamelCase__ : Optional[Any] = k_new.replace(
F"""layer_{i}.1.global_rep.{j+1}.""" , F"""{model_prefix}encoder.layer.{i-1}.layernorm.""" )
if F"""layer_{i}.1.conv_proj.""" in k:
lowerCamelCase__ : Any = k_new.replace(F"""layer_{i}.1.conv_proj.""" , F"""{model_prefix}encoder.layer.{i-1}.conv_projection.""" )
if "pre_norm_attn.0." in k:
lowerCamelCase__ : str = k_new.replace('pre_norm_attn.0.' , 'layernorm_before.' )
if "pre_norm_attn.1." in k:
lowerCamelCase__ : Optional[int] = k_new.replace('pre_norm_attn.1.' , 'attention.' )
if "pre_norm_ffn.0." in k:
lowerCamelCase__ : Union[str, Any] = k_new.replace('pre_norm_ffn.0.' , 'layernorm_after.' )
if "pre_norm_ffn.1." in k:
lowerCamelCase__ : Dict = k_new.replace('pre_norm_ffn.1.' , 'ffn.conv1.' )
if "pre_norm_ffn.3." in k:
lowerCamelCase__ : int = k_new.replace('pre_norm_ffn.3.' , 'ffn.conv2.' )
if "classifier.1." in k:
lowerCamelCase__ : int = k_new.replace('classifier.1.' , 'classifier.' )
if "seg_head." in k:
lowerCamelCase__ : Optional[int] = k_new.replace('seg_head.' , 'segmentation_head.' )
if ".aspp_layer." in k:
lowerCamelCase__ : Tuple = k_new.replace('.aspp_layer.' , '.' )
if ".aspp_pool." in k:
lowerCamelCase__ : List[str] = k_new.replace('.aspp_pool.' , '.' )
rename_keys.append((k, k_new) )
return rename_keys
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Any = []
for k in state_dict.keys():
if k.startswith('seg_head.aux_head.' ):
keys_to_ignore.append(_UpperCAmelCase )
for k in keys_to_ignore:
state_dict.pop(_UpperCAmelCase , _UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( ) -> Any:
lowerCamelCase__ : int = 'http://images.cocodataset.org/val2017/000000039769.jpg'
# url = "https://cdn.britannica.com/86/141086-050-9D7C75EE/Gulfstream-G450-business-jet-passengers.jpg"
lowerCamelCase__ : Optional[Any] = Image.open(requests.get(_UpperCAmelCase , stream=_UpperCAmelCase ).raw )
return im
@torch.no_grad()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Optional[Any] = get_mobilevitva_config(_UpperCAmelCase , _UpperCAmelCase )
# load original state_dict
lowerCamelCase__ : int = torch.load(_UpperCAmelCase , map_location='cpu' )
# load huggingface model
if task_name.startswith('ade20k_' ) or task_name.startswith('voc_' ):
lowerCamelCase__ : Tuple = MobileViTVaForSemanticSegmentation(_UpperCAmelCase ).eval()
lowerCamelCase__ : Dict = False
else:
lowerCamelCase__ : Union[str, Any] = MobileViTVaForImageClassification(_UpperCAmelCase ).eval()
lowerCamelCase__ : List[Any] = False
# remove and rename some keys of load the original model
lowerCamelCase__ : List[Any] = checkpoint
remove_unused_keys(_UpperCAmelCase )
lowerCamelCase__ : Any = create_rename_keys(_UpperCAmelCase , base_model=_UpperCAmelCase )
for rename_key_src, rename_key_dest in rename_keys:
rename_key(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
# load modified state_dict
model.load_state_dict(_UpperCAmelCase )
# Check outputs on an image, prepared by MobileViTImageProcessor
lowerCamelCase__ : Optional[Any] = MobileViTImageProcessor(crop_size=config.image_size , size=config.image_size + 32 )
lowerCamelCase__ : Optional[Any] = image_processor(images=prepare_img() , return_tensors='pt' )
lowerCamelCase__ : str = model(**_UpperCAmelCase )
# verify classification model
if task_name.startswith('imagenet' ):
lowerCamelCase__ : Any = outputs.logits
lowerCamelCase__ : str = logits.argmax(-1 ).item()
print('Predicted class:' , model.config.idalabel[predicted_class_idx] )
if task_name.startswith('imagenet1k_256' ) and config.width_multiplier == 1.0:
# expected_logits for base variant
lowerCamelCase__ : Union[str, Any] = torch.tensor([-1.6_336e00, -7.3_204e-02, -5.1_883e-01] )
assert torch.allclose(logits[0, :3] , _UpperCAmelCase , atol=1e-4 )
Path(_UpperCAmelCase ).mkdir(exist_ok=_UpperCAmelCase )
print(F"""Saving model {task_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__":
_UpperCAmelCase : Optional[int] = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--task""",
default="""imagenet1k_256""",
type=str,
help=(
"""Name of the task for which the MobileViTV2 model you'd like to convert is trained on . """
"""
Classification (ImageNet-1k)
- MobileViTV2 (256x256) : imagenet1k_256
- MobileViTV2 (Trained on 256x256 and Finetuned on 384x384) : imagenet1k_384
- MobileViTV2 (Trained on ImageNet-21k and Finetuned on ImageNet-1k 256x256) :
imagenet21k_to_1k_256
- MobileViTV2 (Trained on ImageNet-21k, Finetuned on ImageNet-1k 256x256, and Finetuned on
ImageNet-1k 384x384) : imagenet21k_to_1k_384
Segmentation
- ADE20K Dataset : ade20k_deeplabv3
- Pascal VOC 2012 Dataset: voc_deeplabv3
"""
),
choices=[
"""imagenet1k_256""",
"""imagenet1k_384""",
"""imagenet21k_to_1k_256""",
"""imagenet21k_to_1k_384""",
"""ade20k_deeplabv3""",
"""voc_deeplabv3""",
],
)
parser.add_argument(
"""--orig_checkpoint_path""", required=True, type=str, help="""Path to the original state dict (.pt file)."""
)
parser.add_argument("""--orig_config_path""", required=True, type=str, help="""Path to the original config file.""")
parser.add_argument(
"""--pytorch_dump_folder_path""", required=True, type=str, help="""Path to the output PyTorch model directory."""
)
_UpperCAmelCase : Any = parser.parse_args()
convert_mobilevitva_checkpoint(
args.task, args.orig_checkpoint_path, args.orig_config_path, args.pytorch_dump_folder_path
)
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> bool:
lowerCamelCase__ : List[str] = len(_UpperCAmelCase )
lowerCamelCase__ : str = [[False] * (required_sum + 1) for _ in range(arr_len + 1 )]
# for each arr value, a sum of zero(0) can be formed by not taking any element
# hence True/1
for i in range(arr_len + 1 ):
lowerCamelCase__ : Tuple = True
# sum is not zero and set is empty then false
for i in range(1 , required_sum + 1 ):
lowerCamelCase__ : Dict = False
for i in range(1 , arr_len + 1 ):
for j in range(1 , required_sum + 1 ):
if arr[i - 1] > j:
lowerCamelCase__ : str = subset[i - 1][j]
if arr[i - 1] <= j:
lowerCamelCase__ : Dict = subset[i - 1][j] or subset[i - 1][j - arr[i - 1]]
return subset[arr_len][required_sum]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
def SCREAMING_SNAKE_CASE ( ) -> List[Any]:
lowerCamelCase__ : Dict = []
lowerCamelCase__ : List[Any] = 1
while len(_UpperCAmelCase ) < 1e6:
constant.append(str(_UpperCAmelCase ) )
i += 1
lowerCamelCase__ : Optional[Any] = ''.join(_UpperCAmelCase )
return (
int(constant[0] )
* int(constant[9] )
* int(constant[99] )
* int(constant[999] )
* int(constant[9999] )
* int(constant[9_9999] )
* int(constant[99_9999] )
)
if __name__ == "__main__":
print(solution())
| 50
|
import torch
from transformers import PreTrainedModel, XLMRobertaConfig, XLMRobertaModel
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """M-CLIP"""
def __init__( self : Optional[Any] , UpperCAmelCase : Union[str, Any]=1024 , UpperCAmelCase : Tuple=768 , **UpperCAmelCase : Optional[int] ) -> Dict:
lowerCamelCase__ : Optional[int] = transformerDimSize
lowerCamelCase__ : Optional[Any] = imageDimSize
super().__init__(**UpperCAmelCase )
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = MCLIPConfig
def __init__( self : List[Any] , UpperCAmelCase : Dict , *UpperCAmelCase : Any , **UpperCAmelCase : Dict ) -> Dict:
super().__init__(UpperCAmelCase , *UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Tuple = XLMRobertaModel(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = torch.nn.Linear(
in_features=config.transformerDimensions , out_features=config.numDims )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Dict ) -> Tuple:
lowerCamelCase__ : Any = self.transformer(input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase )[0]
lowerCamelCase__ : int = (embs * attention_mask.unsqueeze(2 )).sum(dim=1 ) / attention_mask.sum(dim=1 )[:, None]
return self.LinearTransformation(UpperCAmelCase ), embs
| 50
| 1
|
import copy
import tempfile
import unittest
from transformers import MaMaaaConfig, is_torch_available
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow, torch_device
from transformers.utils import cached_property
from ...generation.test_utils import GenerationTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import MaMaaaForConditionalGeneration, MaMaaaModel, MaMaaaTokenizer
from transformers.models.mam_aaa.modeling_mam_aaa import MaMaaaDecoder, MaMaaaEncoder
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=None , _UpperCAmelCase=None , _UpperCAmelCase=None , _UpperCAmelCase=None , _UpperCAmelCase=None , ) -> str:
if attention_mask is None:
lowerCamelCase__ : int = input_ids.ne(config.pad_token_id )
if decoder_attention_mask is None:
lowerCamelCase__ : int = decoder_input_ids.ne(config.pad_token_id )
if head_mask is None:
lowerCamelCase__ : Tuple = torch.ones(config.encoder_layers , config.encoder_attention_heads , device=_UpperCAmelCase )
if decoder_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(config.decoder_layers , config.decoder_attention_heads , device=_UpperCAmelCase )
if cross_attn_head_mask is None:
lowerCamelCase__ : Tuple = torch.ones(config.decoder_layers , config.decoder_attention_heads , device=_UpperCAmelCase )
return {
"input_ids": input_ids,
"decoder_input_ids": decoder_input_ids,
"attention_mask": attention_mask,
"decoder_attention_mask": attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
"cross_attn_head_mask": cross_attn_head_mask,
}
class lowerCAmelCase :
def __init__( self : Dict , UpperCAmelCase : Dict , UpperCAmelCase : int=13 , UpperCAmelCase : Tuple=7 , UpperCAmelCase : int=True , UpperCAmelCase : List[Any]=False , UpperCAmelCase : List[str]=99 , UpperCAmelCase : Optional[int]=16 , UpperCAmelCase : Tuple=2 , UpperCAmelCase : str=4 , UpperCAmelCase : Optional[int]=4 , UpperCAmelCase : List[Any]="relu" , UpperCAmelCase : Dict=0.1 , UpperCAmelCase : Optional[int]=0.1 , UpperCAmelCase : List[Any]=0.0 , UpperCAmelCase : List[str]=0.0 , UpperCAmelCase : str=20 , UpperCAmelCase : List[Any]=2 , UpperCAmelCase : str=1 , UpperCAmelCase : List[str]=0 , ) -> Dict:
lowerCamelCase__ : Optional[int] = parent
lowerCamelCase__ : Tuple = batch_size
lowerCamelCase__ : int = seq_length
lowerCamelCase__ : Union[str, Any] = is_training
lowerCamelCase__ : Dict = use_labels
lowerCamelCase__ : Optional[int] = vocab_size
lowerCamelCase__ : Optional[Any] = hidden_size
lowerCamelCase__ : str = num_hidden_layers
lowerCamelCase__ : Dict = num_attention_heads
lowerCamelCase__ : str = intermediate_size
lowerCamelCase__ : str = hidden_act
lowerCamelCase__ : Dict = hidden_dropout_prob
lowerCamelCase__ : Any = attention_probs_dropout_prob
lowerCamelCase__ : Dict = encoder_layerdrop
lowerCamelCase__ : List[Any] = decoder_layerdrop
lowerCamelCase__ : Any = max_position_embeddings
lowerCamelCase__ : str = eos_token_id
lowerCamelCase__ : int = pad_token_id
lowerCamelCase__ : Union[str, Any] = bos_token_id
def A_ ( self : Any ) -> int:
lowerCamelCase__ : List[str] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
lowerCamelCase__ : str = self.eos_token_id # Eos Token
lowerCamelCase__ : int = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
# we need to clamp the input ids here to avoid having pad token in between
# this is because for M2M100 the position_ids are prepared such that
# all pad tokens have pos id = 2 and rest are between 2..seq_length
# and the seq_length here is seq_length - num_pad_tokens
# but when using past, there is no way of knowing if the past input ids had
# pad tokens in them, which results in incorrect seq_lenth and which in turn results in
# position_ids being off by num_pad_tokens in past input
lowerCamelCase__ : Any = input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Dict = decoder_input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Tuple = self.get_config()
lowerCamelCase__ : str = prepare_mam_aaa_inputs_dict(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
return config, inputs_dict
def A_ ( self : Optional[int] ) -> Union[str, Any]:
return MaMaaaConfig(
vocab_size=self.vocab_size , d_model=self.hidden_size , encoder_layers=self.num_hidden_layers , decoder_layers=self.num_hidden_layers , encoder_attention_heads=self.num_attention_heads , decoder_attention_heads=self.num_attention_heads , encoder_ffn_dim=self.intermediate_size , decoder_ffn_dim=self.intermediate_size , dropout=self.hidden_dropout_prob , attention_dropout=self.attention_probs_dropout_prob , encoder_layerdrop=self.encoder_layerdrop , decoder_layerdrop=self.decoder_layerdrop , 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 , )
def A_ ( self : int ) -> Optional[Any]:
lowerCamelCase__ , lowerCamelCase__ : Dict = self.prepare_config_and_inputs()
return config, inputs_dict
def A_ ( self : Tuple , UpperCAmelCase : Any , UpperCAmelCase : Union[str, Any] ) -> Any:
lowerCamelCase__ : Any = MaMaaaModel(config=UpperCAmelCase ).get_decoder().to(UpperCAmelCase ).eval()
lowerCamelCase__ : Tuple = inputs_dict['input_ids']
lowerCamelCase__ : Tuple = inputs_dict['attention_mask']
lowerCamelCase__ : str = inputs_dict['head_mask']
# first forward pass
lowerCamelCase__ : Dict = model(UpperCAmelCase , attention_mask=UpperCAmelCase , head_mask=UpperCAmelCase , use_cache=UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = outputs.to_tuple()
# create hypothetical multiple next token and extent to next_input_ids
lowerCamelCase__ : Tuple = ids_tensor((self.batch_size, 3) , config.vocab_size )
lowerCamelCase__ : Any = ids_tensor((self.batch_size, 3) , 2 )
# append to next input_ids and
lowerCamelCase__ : Optional[int] = torch.cat([input_ids, next_tokens] , dim=-1 )
lowerCamelCase__ : Optional[Any] = torch.cat([attention_mask, next_attn_mask] , dim=-1 )
lowerCamelCase__ : Tuple = model(UpperCAmelCase , attention_mask=UpperCAmelCase )['last_hidden_state']
lowerCamelCase__ : str = model(UpperCAmelCase , attention_mask=UpperCAmelCase , past_key_values=UpperCAmelCase )[
'last_hidden_state'
]
# select random slice
lowerCamelCase__ : int = ids_tensor((1,) , output_from_past.shape[-1] ).item()
lowerCamelCase__ : str = output_from_no_past[:, -3:, random_slice_idx].detach()
lowerCamelCase__ : Any = output_from_past[:, :, random_slice_idx].detach()
self.parent.assertTrue(output_from_past_slice.shape[1] == next_tokens.shape[1] )
# test that outputs are equal for slice
self.parent.assertTrue(torch.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-2 ) )
def A_ ( self : List[str] , UpperCAmelCase : List[str] , UpperCAmelCase : str ) -> List[str]:
lowerCamelCase__ : Optional[int] = MaMaaaModel(config=UpperCAmelCase ).to(UpperCAmelCase ).eval()
lowerCamelCase__ : Optional[int] = model(**UpperCAmelCase )
lowerCamelCase__ : Dict = outputs.encoder_last_hidden_state
lowerCamelCase__ : int = outputs.last_hidden_state
with tempfile.TemporaryDirectory() as tmpdirname:
lowerCamelCase__ : Optional[int] = model.get_encoder()
encoder.save_pretrained(UpperCAmelCase )
lowerCamelCase__ : List[Any] = MaMaaaEncoder.from_pretrained(UpperCAmelCase ).to(UpperCAmelCase )
lowerCamelCase__ : int = encoder(inputs_dict['input_ids'] , attention_mask=inputs_dict['attention_mask'] )[
0
]
self.parent.assertTrue((encoder_last_hidden_state_a - encoder_last_hidden_state).abs().max().item() < 1e-3 )
with tempfile.TemporaryDirectory() as tmpdirname:
lowerCamelCase__ : List[Any] = model.get_decoder()
decoder.save_pretrained(UpperCAmelCase )
lowerCamelCase__ : Any = MaMaaaDecoder.from_pretrained(UpperCAmelCase ).to(UpperCAmelCase )
lowerCamelCase__ : Any = decoder(
input_ids=inputs_dict['decoder_input_ids'] , attention_mask=inputs_dict['decoder_attention_mask'] , encoder_hidden_states=UpperCAmelCase , encoder_attention_mask=inputs_dict['attention_mask'] , )[0]
self.parent.assertTrue((last_hidden_state_a - last_hidden_state).abs().max().item() < 1e-3 )
@require_torch
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(
MaMaaaModel,
MaMaaaForConditionalGeneration,
)
if is_torch_available()
else ()
)
UpperCAmelCase__ = (MaMaaaForConditionalGeneration,) if is_torch_available() else ()
UpperCAmelCase__ = (
{
"""conversational""": MaMaaaForConditionalGeneration,
"""feature-extraction""": MaMaaaModel,
"""summarization""": MaMaaaForConditionalGeneration,
"""text2text-generation""": MaMaaaForConditionalGeneration,
"""translation""": MaMaaaForConditionalGeneration,
}
if is_torch_available()
else {}
)
UpperCAmelCase__ = True
UpperCAmelCase__ = True
UpperCAmelCase__ = False
UpperCAmelCase__ = False
def A_ ( self : List[str] , UpperCAmelCase : str , UpperCAmelCase : Optional[Any] , UpperCAmelCase : int , UpperCAmelCase : Any , UpperCAmelCase : List[str] ) -> Any:
if pipeline_test_casse_name == "TranslationPipelineTests":
# Get `ValueError: Translation requires a `src_lang` and a `tgt_lang` for this model`.
# `M2M100Config` was never used in pipeline tests: cannot create a simple tokenizer.
return True
return False
def A_ ( self : Any ) -> int:
lowerCamelCase__ : Union[str, Any] = MaMaaaModelTester(self )
lowerCamelCase__ : Dict = ConfigTester(self , config_class=UpperCAmelCase )
def A_ ( self : int ) -> List[Any]:
self.config_tester.run_common_tests()
def A_ ( self : Optional[int] ) -> Union[str, Any]:
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = self.model_tester.prepare_config_and_inputs()
for model_class in self.all_model_classes:
lowerCamelCase__ : Any = model_class(UpperCAmelCase )
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : List[Any] = model_class.from_pretrained(UpperCAmelCase , output_loading_info=UpperCAmelCase )
self.assertEqual(info['missing_keys'] , [] )
def A_ ( self : Optional[int] ) -> int:
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_decoder_model_past_large_inputs(*UpperCAmelCase )
def A_ ( self : Tuple ) -> List[str]:
lowerCamelCase__ : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_encoder_decoder_model_standalone(*UpperCAmelCase )
def A_ ( self : Optional[int] ) -> List[Any]:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in (MaMaaaModel, MaMaaaForConditionalGeneration):
lowerCamelCase__ : Union[str, Any] = model_class(UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : List[str] = copy.deepcopy(self._prepare_for_class(UpperCAmelCase , UpperCAmelCase ) )
if not self.is_encoder_decoder:
lowerCamelCase__ : Tuple = inputs['input_ids']
del inputs["input_ids"]
else:
lowerCamelCase__ : List[str] = inputs['input_ids']
lowerCamelCase__ : List[Any] = inputs.get('decoder_input_ids' , UpperCAmelCase )
del inputs["input_ids"]
inputs.pop('decoder_input_ids' , UpperCAmelCase )
lowerCamelCase__ : List[str] = model.get_input_embeddings()
if not self.is_encoder_decoder:
lowerCamelCase__ : List[Any] = wte(UpperCAmelCase )
else:
lowerCamelCase__ : int = wte(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = wte(UpperCAmelCase )
with torch.no_grad():
model(**UpperCAmelCase )[0]
def A_ ( self : Dict ) -> List[str]:
lowerCamelCase__ , lowerCamelCase__ : Tuple = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Dict = input_dict['input_ids']
lowerCamelCase__ : List[str] = input_ids.ne(1 ).to(UpperCAmelCase )
lowerCamelCase__ : List[str] = MaMaaaForConditionalGeneration(UpperCAmelCase ).eval().to(UpperCAmelCase )
if torch_device == "cuda":
model.half()
model.generate(UpperCAmelCase , attention_mask=UpperCAmelCase )
model.generate(num_beams=4 , do_sample=UpperCAmelCase , early_stopping=UpperCAmelCase , num_return_sequences=3 )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Optional[Any]:
return torch.tensor(_UpperCAmelCase , dtype=torch.long , device=_UpperCAmelCase )
_UpperCAmelCase : Union[str, Any] = 1E-4
@require_torch
@require_sentencepiece
@require_tokenizers
@slow
class lowerCAmelCase ( unittest.TestCase ):
@cached_property
def A_ ( self : Dict ) -> Optional[int]:
return MaMaaaTokenizer.from_pretrained('facebook/m2m100_418M' )
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Optional[int] = MaMaaaModel.from_pretrained('facebook/m2m100_418M' ).to(UpperCAmelCase )
lowerCamelCase__ : str = _long_tensor([[128028, 98, 12, 30527, 2732, 159, 7755, 61904, 39144, 38, 2]] )
lowerCamelCase__ : Any = _long_tensor([[2, 128028, 98, 12, 30527, 2732, 159, 7755, 61904, 39144, 38]] )
lowerCamelCase__ : Optional[int] = prepare_mam_aaa_inputs_dict(model.config , UpperCAmelCase , UpperCAmelCase )
with torch.no_grad():
lowerCamelCase__ : Tuple = model(**UpperCAmelCase )[0]
lowerCamelCase__ : Tuple = torch.Size((1, 11, 1024) )
self.assertEqual(output.shape , UpperCAmelCase )
# change to expected output here
lowerCamelCase__ : int = torch.tensor(
[[-0.7_7_8_0, -0.1_6_7_6, 0.1_0_3_8], [-6.7_5_5_6, -1.3_9_9_2, 0.0_5_6_7], [-7.5_3_8_3, -0.5_9_2_0, -0.2_7_7_9]] , device=UpperCAmelCase )
self.assertTrue(torch.allclose(output[:, :3, :3] , UpperCAmelCase , atol=UpperCAmelCase ) )
def A_ ( self : List[str] ) -> Dict:
lowerCamelCase__ : List[Any] = MaMaaaForConditionalGeneration.from_pretrained('facebook/m2m100_418M' ).to(UpperCAmelCase )
# change to intended input
lowerCamelCase__ : Union[str, Any] = _long_tensor([[128028, 98, 12, 30527, 2732, 159, 7755, 61904, 39144, 38, 2]] )
lowerCamelCase__ : Tuple = _long_tensor([[2, 128028, 98, 12, 30527, 2732, 159, 7755, 61904, 39144, 38]] )
lowerCamelCase__ : Optional[int] = prepare_mam_aaa_inputs_dict(model.config , UpperCAmelCase , UpperCAmelCase )
with torch.no_grad():
lowerCamelCase__ : str = model(**UpperCAmelCase )[0]
lowerCamelCase__ : Dict = torch.Size((1, 11, model.config.vocab_size) )
self.assertEqual(output.shape , UpperCAmelCase )
# change to expected output here
lowerCamelCase__ : Optional[int] = torch.tensor(
[[-1.0_4_4_8, -1.0_4_1_1, 3.7_9_9_2], [-3.2_1_9_1, -3.2_3_8_6, -1.3_4_5_1], [-3.6_2_1_0, -3.5_9_9_3, 0.4_9_2_5]] , device=UpperCAmelCase )
self.assertTrue(torch.allclose(output[:, :3, :3] , UpperCAmelCase , atol=UpperCAmelCase ) )
def A_ ( self : int ) -> str:
lowerCamelCase__ : List[str] = MaMaaaForConditionalGeneration.from_pretrained('facebook/m2m100_418M' ).to(UpperCAmelCase )
lowerCamelCase__ : str = MaMaaaTokenizer.from_pretrained('facebook/m2m100_418M' , src_lang='fr' , tgt_lang='en' )
lowerCamelCase__ : Dict = [
'L\'affaire NSA souligne l\'absence totale de débat sur le renseignement',
'Selon moi, il y a deux niveaux de réponse de la part du gouvernement français.',
'Lorsque François Hollande téléphone à Barack Obama ou quand le ministre des affaires étrangères Laurent'
' Fabius convoque l\'ambassadeur des Etats-Unis, ils réagissent à une vraie découverte, qui est celle de'
' l\'ampleur de la surveillance américaine sur l\'ensemble des communications en France.',
]
# The below article tests that we don't add any hypotheses outside of the top n_beams
lowerCamelCase__ : List[str] = tokenizer(UpperCAmelCase , padding=UpperCAmelCase , return_tensors='pt' )
lowerCamelCase__ : Tuple = model.generate(
input_ids=dct['input_ids'].to(UpperCAmelCase ) , attention_mask=dct['attention_mask'].to(UpperCAmelCase ) , num_beams=5 , forced_bos_token_id=tokenizer.get_lang_id('en' ) , )
lowerCamelCase__ : str = [
'The NSA case highlights the total absence of intelligence debate',
'I think there are two levels of response from the French government.',
'When François Hollande calls Barack Obama or when Foreign Minister Laurent Fabius calls the U.S.'
' Ambassador, they respond to a real discovery, which is that of the scale of U.S. surveillance on all'
' communications in France.',
]
lowerCamelCase__ : Tuple = tokenizer.batch_decode(
hypotheses_batch.tolist() , clean_up_tokenization_spaces=UpperCAmelCase , skip_special_tokens=UpperCAmelCase )
assert generated == expected_en
| 50
|
from itertools import count
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 50 ) -> int:
lowerCamelCase__ : Optional[Any] = [1] * min_block_length
for n in count(_UpperCAmelCase ):
fill_count_functions.append(1 )
for block_length in range(_UpperCAmelCase , n + 1 ):
for block_start in range(n - block_length ):
fill_count_functions[n] += fill_count_functions[
n - block_start - block_length - 1
]
fill_count_functions[n] += 1
if fill_count_functions[n] > 100_0000:
break
return n
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
import argparse
from tax import checkpoints
from transformers import AutoConfig, FlaxAutoModelForSeqaSeqLM
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : List[str] = AutoConfig.from_pretrained(_UpperCAmelCase )
lowerCamelCase__ : Any = FlaxAutoModelForSeqaSeqLM.from_config(config=_UpperCAmelCase )
lowerCamelCase__ : Optional[int] = checkpoints.load_tax_checkpoint(_UpperCAmelCase )
lowerCamelCase__ : Dict = 'wi_0' in tax_model['target']['encoder']['layers_0']['mlp']
if config.model_type == "t5":
lowerCamelCase__ : Optional[Any] = 'SelfAttention'
if config.model_type == "longt5" and config.encoder_attention_type == "local":
lowerCamelCase__ : Union[str, Any] = 'LocalSelfAttention'
elif config.model_type == "longt5" and config.encoder_attention_type == "transient-global":
lowerCamelCase__ : Tuple = 'TransientGlobalSelfAttention'
else:
raise ValueError(
'Given config is expected to have `model_type=\'t5\'`, or `model_type=\'longt5` with `encoder_attention_type`'
' attribute with a value from [\'local\', \'transient-global].' )
# Encoder
for layer_index in range(config.num_layers ):
lowerCamelCase__ : Union[str, Any] = F"""layers_{str(_UpperCAmelCase )}"""
# Self-Attention
lowerCamelCase__ : Any = tax_model['target']['encoder'][layer_name]['attention']['key']['kernel']
lowerCamelCase__ : List[str] = tax_model['target']['encoder'][layer_name]['attention']['out']['kernel']
lowerCamelCase__ : Union[str, Any] = tax_model['target']['encoder'][layer_name]['attention']['query']['kernel']
lowerCamelCase__ : Tuple = tax_model['target']['encoder'][layer_name]['attention']['value']['kernel']
# Global input layer norm
if config.model_type == "longt5" and config.encoder_attention_type == "transient-global":
lowerCamelCase__ : Optional[Any] = tax_model['target']['encoder'][layer_name]['attention']['T5LayerNorm_0']['scale']
# Layer Normalization
lowerCamelCase__ : int = tax_model['target']['encoder'][layer_name]['pre_attention_layer_norm']['scale']
if split_mlp_wi:
lowerCamelCase__ : Tuple = tax_model['target']['encoder'][layer_name]['mlp']['wi_0']['kernel']
lowerCamelCase__ : Optional[int] = tax_model['target']['encoder'][layer_name]['mlp']['wi_1']['kernel']
else:
lowerCamelCase__ : List[str] = tax_model['target']['encoder'][layer_name]['mlp']['wi']['kernel']
lowerCamelCase__ : int = tax_model['target']['encoder'][layer_name]['mlp']['wo']['kernel']
# Layer Normalization
lowerCamelCase__ : Optional[Any] = tax_model['target']['encoder'][layer_name]['pre_mlp_layer_norm']['scale']
# Assigning
lowerCamelCase__ : Optional[int] = flax_model.params['encoder']['block'][str(_UpperCAmelCase )]['layer']
lowerCamelCase__ : int = tax_attention_key
lowerCamelCase__ : str = tax_attention_out
lowerCamelCase__ : Optional[Any] = tax_attention_query
lowerCamelCase__ : Dict = tax_attention_value
lowerCamelCase__ : List[str] = tax_attention_layer_norm
# Global input layer norm
if config.model_type == "longt5" and config.encoder_attention_type == "transient-global":
lowerCamelCase__ : List[Any] = tax_global_layer_norm
if split_mlp_wi:
lowerCamelCase__ : List[str] = tax_mlp_wi_a
lowerCamelCase__ : Union[str, Any] = tax_mlp_wi_a
else:
lowerCamelCase__ : List[Any] = tax_mlp_wi
lowerCamelCase__ : List[str] = tax_mlp_wo
lowerCamelCase__ : Optional[int] = tax_mlp_layer_norm
lowerCamelCase__ : Any = flax_model_encoder_layer_block
# Only for layer 0:
lowerCamelCase__ : List[Any] = tax_model['target']['encoder']['relpos_bias']['rel_embedding'].T
lowerCamelCase__ : List[str] = tax_encoder_rel_embedding
# Side/global relative position_bias + layer norm
if config.model_type == "longt5" and config.encoder_attention_type == "transient-global":
lowerCamelCase__ : Optional[Any] = tax_model['target']['encoder']['side_relpos_bias']['rel_embedding'].T
lowerCamelCase__ : str = tax_encoder_global_rel_embedding
# Assigning
lowerCamelCase__ : Union[str, Any] = tax_model['target']['encoder']['encoder_norm']['scale']
lowerCamelCase__ : str = tax_encoder_norm
# Decoder
for layer_index in range(config.num_layers ):
lowerCamelCase__ : str = F"""layers_{str(_UpperCAmelCase )}"""
# Self-Attention
lowerCamelCase__ : Optional[int] = tax_model['target']['decoder'][layer_name]['self_attention']['key']['kernel']
lowerCamelCase__ : List[Any] = tax_model['target']['decoder'][layer_name]['self_attention']['out']['kernel']
lowerCamelCase__ : str = tax_model['target']['decoder'][layer_name]['self_attention']['query']['kernel']
lowerCamelCase__ : Tuple = tax_model['target']['decoder'][layer_name]['self_attention']['value']['kernel']
# Layer Normalization
lowerCamelCase__ : Union[str, Any] = tax_model['target']['decoder'][layer_name]['pre_self_attention_layer_norm'][
'scale'
]
# Encoder-Decoder-Attention
lowerCamelCase__ : Optional[Any] = tax_model['target']['decoder'][layer_name]['encoder_decoder_attention']
lowerCamelCase__ : str = tax_enc_dec_attention_module['key']['kernel']
lowerCamelCase__ : str = tax_enc_dec_attention_module['out']['kernel']
lowerCamelCase__ : int = tax_enc_dec_attention_module['query']['kernel']
lowerCamelCase__ : str = tax_enc_dec_attention_module['value']['kernel']
# Layer Normalization
lowerCamelCase__ : Dict = tax_model['target']['decoder'][layer_name]['pre_cross_attention_layer_norm']['scale']
# MLP
if split_mlp_wi:
lowerCamelCase__ : Optional[Any] = tax_model['target']['decoder'][layer_name]['mlp']['wi_0']['kernel']
lowerCamelCase__ : str = tax_model['target']['decoder'][layer_name]['mlp']['wi_1']['kernel']
else:
lowerCamelCase__ : List[str] = tax_model['target']['decoder'][layer_name]['mlp']['wi']['kernel']
lowerCamelCase__ : Union[str, Any] = tax_model['target']['decoder'][layer_name]['mlp']['wo']['kernel']
# Layer Normalization
lowerCamelCase__ : List[str] = tax_model['target']['decoder'][layer_name]['pre_mlp_layer_norm']['scale']
# Assigning
lowerCamelCase__ : Tuple = flax_model.params['decoder']['block'][str(_UpperCAmelCase )]['layer']
lowerCamelCase__ : Optional[Any] = tax_attention_key
lowerCamelCase__ : str = tax_attention_out
lowerCamelCase__ : int = tax_attention_query
lowerCamelCase__ : Optional[Any] = tax_attention_value
lowerCamelCase__ : Optional[Any] = tax_pre_attention_layer_norm
lowerCamelCase__ : str = tax_enc_dec_attention_key
lowerCamelCase__ : Dict = tax_enc_dec_attention_out
lowerCamelCase__ : Optional[Any] = tax_enc_dec_attention_query
lowerCamelCase__ : Optional[int] = tax_enc_dec_attention_value
lowerCamelCase__ : Tuple = tax_cross_layer_norm
if split_mlp_wi:
lowerCamelCase__ : List[Any] = tax_mlp_wi_a
lowerCamelCase__ : Union[str, Any] = tax_mlp_wi_a
else:
lowerCamelCase__ : str = tax_mlp_wi
lowerCamelCase__ : Optional[int] = tax_mlp_wo
lowerCamelCase__ : Dict = txa_mlp_layer_norm
lowerCamelCase__ : List[Any] = flax_model_decoder_layer_block
# Decoder Normalization
lowerCamelCase__ : str = tax_model['target']['decoder']['decoder_norm']['scale']
lowerCamelCase__ : List[Any] = txa_decoder_norm
# Only for layer 0:
lowerCamelCase__ : Any = tax_model['target']['decoder']['relpos_bias']['rel_embedding'].T
lowerCamelCase__ : List[str] = tax_decoder_rel_embedding
# Token Embeddings
lowerCamelCase__ : Union[str, Any] = tax_model['target']['token_embedder']['embedding']
lowerCamelCase__ : Optional[Any] = txa_token_embeddings
# LM Head (only in v1.1 and LongT5 checkpoints)
if "logits_dense" in tax_model["target"]["decoder"]:
lowerCamelCase__ : List[str] = tax_model['target']['decoder']['logits_dense']['kernel']
flax_model.save_pretrained(_UpperCAmelCase )
print('T5X Model was sucessfully converted!' )
if __name__ == "__main__":
_UpperCAmelCase : int = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--t5x_checkpoint_path""", default=None, type=str, required=True, help="""Path the T5X checkpoint."""
)
parser.add_argument("""--config_name""", default=None, type=str, required=True, help="""Config name of LongT5/T5 model.""")
parser.add_argument(
"""--flax_dump_folder_path""", default=None, type=str, required=True, help="""Path to the output FLAX model."""
)
_UpperCAmelCase : Tuple = parser.parse_args()
convert_tax_checkpoint_to_flax(args.tax_checkpoint_path, args.config_name, args.flax_dump_folder_path)
| 50
|
from __future__ import annotations
from typing import Any
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> None:
create_state_space_tree(_UpperCAmelCase , [] , 0 )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> None:
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__":
_UpperCAmelCase : list[Any] = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["""A""", """B""", """C"""])
generate_all_subsequences(seq)
| 50
| 1
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Optional[Any] = set()
# edges = list of graph's edges
lowerCamelCase__ : List[str] = get_edges(_UpperCAmelCase )
# While there are still elements in edges list, take an arbitrary edge
# (from_node, to_node) and add his extremity to chosen_vertices and then
# remove all arcs adjacent to the from_node and to_node
while edges:
lowerCamelCase__ , lowerCamelCase__ : str = edges.pop()
chosen_vertices.add(_UpperCAmelCase )
chosen_vertices.add(_UpperCAmelCase )
for edge in edges.copy():
if from_node in edge or to_node in edge:
edges.discard(_UpperCAmelCase )
return chosen_vertices
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Union[str, Any] = set()
for from_node, to_nodes in graph.items():
for to_node in to_nodes:
edges.add((from_node, to_node) )
return edges
if __name__ == "__main__":
import doctest
doctest.testmod()
# graph = {0: [1, 3], 1: [0, 3], 2: [0, 3, 4], 3: [0, 1, 2], 4: [2, 3]}
# print(f"Matching vertex cover:\n{matching_min_vertex_cover(graph)}")
| 50
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from tokenizers import processors
from ...tokenization_utils import AddedToken, BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_nllb import NllbTokenizer
else:
_UpperCAmelCase : int = None
_UpperCAmelCase : Dict = logging.get_logger(__name__)
_UpperCAmelCase : Optional[int] = {"""vocab_file""": """sentencepiece.bpe.model""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : List[Any] = {
"""vocab_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/sentencepiece.bpe.model"""
),
},
"""tokenizer_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : List[str] = {
"""facebook/nllb-large-en-ro""": 10_24,
"""facebook/nllb-200-distilled-600M""": 10_24,
}
# fmt: off
_UpperCAmelCase : Optional[int] = ["""ace_Arab""", """ace_Latn""", """acm_Arab""", """acq_Arab""", """aeb_Arab""", """afr_Latn""", """ajp_Arab""", """aka_Latn""", """amh_Ethi""", """apc_Arab""", """arb_Arab""", """ars_Arab""", """ary_Arab""", """arz_Arab""", """asm_Beng""", """ast_Latn""", """awa_Deva""", """ayr_Latn""", """azb_Arab""", """azj_Latn""", """bak_Cyrl""", """bam_Latn""", """ban_Latn""", """bel_Cyrl""", """bem_Latn""", """ben_Beng""", """bho_Deva""", """bjn_Arab""", """bjn_Latn""", """bod_Tibt""", """bos_Latn""", """bug_Latn""", """bul_Cyrl""", """cat_Latn""", """ceb_Latn""", """ces_Latn""", """cjk_Latn""", """ckb_Arab""", """crh_Latn""", """cym_Latn""", """dan_Latn""", """deu_Latn""", """dik_Latn""", """dyu_Latn""", """dzo_Tibt""", """ell_Grek""", """eng_Latn""", """epo_Latn""", """est_Latn""", """eus_Latn""", """ewe_Latn""", """fao_Latn""", """pes_Arab""", """fij_Latn""", """fin_Latn""", """fon_Latn""", """fra_Latn""", """fur_Latn""", """fuv_Latn""", """gla_Latn""", """gle_Latn""", """glg_Latn""", """grn_Latn""", """guj_Gujr""", """hat_Latn""", """hau_Latn""", """heb_Hebr""", """hin_Deva""", """hne_Deva""", """hrv_Latn""", """hun_Latn""", """hye_Armn""", """ibo_Latn""", """ilo_Latn""", """ind_Latn""", """isl_Latn""", """ita_Latn""", """jav_Latn""", """jpn_Jpan""", """kab_Latn""", """kac_Latn""", """kam_Latn""", """kan_Knda""", """kas_Arab""", """kas_Deva""", """kat_Geor""", """knc_Arab""", """knc_Latn""", """kaz_Cyrl""", """kbp_Latn""", """kea_Latn""", """khm_Khmr""", """kik_Latn""", """kin_Latn""", """kir_Cyrl""", """kmb_Latn""", """kon_Latn""", """kor_Hang""", """kmr_Latn""", """lao_Laoo""", """lvs_Latn""", """lij_Latn""", """lim_Latn""", """lin_Latn""", """lit_Latn""", """lmo_Latn""", """ltg_Latn""", """ltz_Latn""", """lua_Latn""", """lug_Latn""", """luo_Latn""", """lus_Latn""", """mag_Deva""", """mai_Deva""", """mal_Mlym""", """mar_Deva""", """min_Latn""", """mkd_Cyrl""", """plt_Latn""", """mlt_Latn""", """mni_Beng""", """khk_Cyrl""", """mos_Latn""", """mri_Latn""", """zsm_Latn""", """mya_Mymr""", """nld_Latn""", """nno_Latn""", """nob_Latn""", """npi_Deva""", """nso_Latn""", """nus_Latn""", """nya_Latn""", """oci_Latn""", """gaz_Latn""", """ory_Orya""", """pag_Latn""", """pan_Guru""", """pap_Latn""", """pol_Latn""", """por_Latn""", """prs_Arab""", """pbt_Arab""", """quy_Latn""", """ron_Latn""", """run_Latn""", """rus_Cyrl""", """sag_Latn""", """san_Deva""", """sat_Beng""", """scn_Latn""", """shn_Mymr""", """sin_Sinh""", """slk_Latn""", """slv_Latn""", """smo_Latn""", """sna_Latn""", """snd_Arab""", """som_Latn""", """sot_Latn""", """spa_Latn""", """als_Latn""", """srd_Latn""", """srp_Cyrl""", """ssw_Latn""", """sun_Latn""", """swe_Latn""", """swh_Latn""", """szl_Latn""", """tam_Taml""", """tat_Cyrl""", """tel_Telu""", """tgk_Cyrl""", """tgl_Latn""", """tha_Thai""", """tir_Ethi""", """taq_Latn""", """taq_Tfng""", """tpi_Latn""", """tsn_Latn""", """tso_Latn""", """tuk_Latn""", """tum_Latn""", """tur_Latn""", """twi_Latn""", """tzm_Tfng""", """uig_Arab""", """ukr_Cyrl""", """umb_Latn""", """urd_Arab""", """uzn_Latn""", """vec_Latn""", """vie_Latn""", """war_Latn""", """wol_Latn""", """xho_Latn""", """ydd_Hebr""", """yor_Latn""", """yue_Hant""", """zho_Hans""", """zho_Hant""", """zul_Latn"""]
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = ["""input_ids""", """attention_mask"""]
UpperCAmelCase__ = NllbTokenizer
UpperCAmelCase__ = []
UpperCAmelCase__ = []
def __init__( self : Tuple , UpperCAmelCase : int=None , UpperCAmelCase : Any=None , UpperCAmelCase : str="<s>" , UpperCAmelCase : Optional[Any]="</s>" , UpperCAmelCase : str="</s>" , UpperCAmelCase : Tuple="<s>" , UpperCAmelCase : Optional[Any]="<unk>" , UpperCAmelCase : List[str]="<pad>" , UpperCAmelCase : Union[str, Any]="<mask>" , UpperCAmelCase : Tuple=None , UpperCAmelCase : int=None , UpperCAmelCase : Dict=None , UpperCAmelCase : Any=False , **UpperCAmelCase : Optional[int] , ) -> Tuple:
# Mask token behave like a normal word, i.e. include the space before it
lowerCamelCase__ : List[Any] = AddedToken(UpperCAmelCase , lstrip=UpperCAmelCase , rstrip=UpperCAmelCase ) if isinstance(UpperCAmelCase , UpperCAmelCase ) else mask_token
lowerCamelCase__ : Union[str, Any] = legacy_behaviour
super().__init__(
vocab_file=UpperCAmelCase , tokenizer_file=UpperCAmelCase , bos_token=UpperCAmelCase , eos_token=UpperCAmelCase , sep_token=UpperCAmelCase , cls_token=UpperCAmelCase , unk_token=UpperCAmelCase , pad_token=UpperCAmelCase , mask_token=UpperCAmelCase , src_lang=UpperCAmelCase , tgt_lang=UpperCAmelCase , additional_special_tokens=UpperCAmelCase , legacy_behaviour=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = vocab_file
lowerCamelCase__ : Dict = False if not self.vocab_file else True
lowerCamelCase__ : Optional[Any] = FAIRSEQ_LANGUAGE_CODES.copy()
if additional_special_tokens is not None:
# Only add those special tokens if they are not already there.
_additional_special_tokens.extend(
[t for t in additional_special_tokens if t not in _additional_special_tokens] )
self.add_special_tokens({'additional_special_tokens': _additional_special_tokens} )
lowerCamelCase__ : str = {
lang_code: self.convert_tokens_to_ids(UpperCAmelCase ) for lang_code in FAIRSEQ_LANGUAGE_CODES
}
lowerCamelCase__ : int = src_lang if src_lang is not None else 'eng_Latn'
lowerCamelCase__ : List[Any] = self.convert_tokens_to_ids(self._src_lang )
lowerCamelCase__ : str = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def A_ ( self : int ) -> str:
return self._src_lang
@src_lang.setter
def A_ ( self : List[Any] , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Any = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
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 A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : Dict = [self.sep_token_id]
lowerCamelCase__ : List[str] = [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 + sep + token_ids_a + sep ) * [0]
def A_ ( self : int , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Optional[str] , UpperCAmelCase : Optional[str] , **UpperCAmelCase : List[str] ) -> Dict:
if src_lang is None or tgt_lang is None:
raise ValueError('Translation requires a `src_lang` and a `tgt_lang` for this model' )
lowerCamelCase__ : Optional[int] = src_lang
lowerCamelCase__ : Optional[int] = self(UpperCAmelCase , add_special_tokens=UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.convert_tokens_to_ids(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tgt_lang_id
return inputs
def A_ ( self : Dict , UpperCAmelCase : List[str] , UpperCAmelCase : str = "eng_Latn" , UpperCAmelCase : Optional[List[str]] = None , UpperCAmelCase : str = "fra_Latn" , **UpperCAmelCase : Dict , ) -> BatchEncoding:
lowerCamelCase__ : Any = src_lang
lowerCamelCase__ : int = tgt_lang
return super().prepare_seqaseq_batch(UpperCAmelCase , UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Union[str, Any] ) -> Optional[int]:
return self.set_src_lang_special_tokens(self.src_lang )
def A_ ( self : Any ) -> Union[str, Any]:
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def A_ ( self : str , UpperCAmelCase : Optional[Any] ) -> None:
lowerCamelCase__ : int = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : int = []
lowerCamelCase__ : str = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : int = [self.cur_lang_code]
lowerCamelCase__ : Tuple = [self.eos_token_id]
lowerCamelCase__ : Any = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : Optional[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : str = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : int , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Union[str, Any] = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : Dict = []
lowerCamelCase__ : Union[str, Any] = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : Any = [self.cur_lang_code]
lowerCamelCase__ : Optional[Any] = [self.eos_token_id]
lowerCamelCase__ : Union[str, Any] = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : List[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : Optional[int] = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : Union[str, Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not self.can_save_slow_tokenizer:
raise ValueError(
'Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '
'tokenizer.' )
if not os.path.isdir(UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory.""" )
return
lowerCamelCase__ : int = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCAmelCase ):
copyfile(self.vocab_file , UpperCAmelCase )
return (out_vocab_file,)
| 50
| 1
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
if not all(x.isalpha() for x in string ):
raise ValueError('String must only contain alphabetic characters.' )
lowerCamelCase__ : Tuple = sorted(string.lower() )
return len(_UpperCAmelCase ) == len(set(_UpperCAmelCase ) )
if __name__ == "__main__":
_UpperCAmelCase : Optional[Any] = input("""Enter a string """).strip()
_UpperCAmelCase : Optional[int] = is_isogram(input_str)
print(F"""{input_str} is {"an" if isogram else "not an"} isogram.""")
| 50
|
import argparse
import torch
from torch import nn
from transformers import MaMaaaConfig, MaMaaaForConditionalGeneration
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Optional[int] = [
'encoder.version',
'decoder.version',
'model.encoder.version',
'model.decoder.version',
'decoder.output_projection.weight',
'_float_tensor',
'encoder.embed_positions._float_tensor',
'decoder.embed_positions._float_tensor',
]
for k in ignore_keys:
state_dict.pop(_UpperCAmelCase , _UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[Any]:
lowerCamelCase__ , lowerCamelCase__ : List[str] = emb.weight.shape
lowerCamelCase__ : Tuple = nn.Linear(_UpperCAmelCase , _UpperCAmelCase , bias=_UpperCAmelCase )
lowerCamelCase__ : Dict = emb.weight.data
return lin_layer
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Tuple = torch.load(_UpperCAmelCase , map_location='cpu' )
lowerCamelCase__ : List[str] = mam_aaa['args'] or mam_aaa['cfg']['model']
lowerCamelCase__ : Optional[int] = mam_aaa['model']
remove_ignore_keys_(_UpperCAmelCase )
lowerCamelCase__ : str = state_dict['encoder.embed_tokens.weight'].shape[0]
lowerCamelCase__ : Union[str, Any] = MaMaaaConfig(
vocab_size=_UpperCAmelCase , max_position_embeddings=1024 , encoder_layers=args.encoder_layers , decoder_layers=args.decoder_layers , encoder_attention_heads=args.encoder_attention_heads , decoder_attention_heads=args.decoder_attention_heads , encoder_ffn_dim=args.encoder_ffn_embed_dim , decoder_ffn_dim=args.decoder_ffn_embed_dim , d_model=args.encoder_embed_dim , encoder_layerdrop=args.encoder_layerdrop , decoder_layerdrop=args.decoder_layerdrop , dropout=args.dropout , attention_dropout=args.attention_dropout , activation_dropout=args.activation_dropout , activation_function='relu' , )
lowerCamelCase__ : Optional[Any] = state_dict['decoder.embed_tokens.weight']
lowerCamelCase__ : Union[str, Any] = MaMaaaForConditionalGeneration(_UpperCAmelCase )
model.model.load_state_dict(_UpperCAmelCase , strict=_UpperCAmelCase )
lowerCamelCase__ : List[str] = make_linear_from_emb(model.model.shared )
return model
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""fairseq_path""", type=str, help="""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.""")
_UpperCAmelCase : str = parser.parse_args()
_UpperCAmelCase : Optional[Any] = convert_fairseq_mamaaa_checkpoint_from_disk(args.fairseq_pathß)
model.save_pretrained(args.pytorch_dump_folder_path)
| 50
| 1
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_torch_available,
)
_UpperCAmelCase : Optional[Any] = {
"""configuration_speecht5""": [
"""SPEECHT5_PRETRAINED_CONFIG_ARCHIVE_MAP""",
"""SPEECHT5_PRETRAINED_HIFIGAN_CONFIG_ARCHIVE_MAP""",
"""SpeechT5Config""",
"""SpeechT5HifiGanConfig""",
],
"""feature_extraction_speecht5""": ["""SpeechT5FeatureExtractor"""],
"""processing_speecht5""": ["""SpeechT5Processor"""],
}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_UpperCAmelCase : Union[str, Any] = ["""SpeechT5Tokenizer"""]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_UpperCAmelCase : Tuple = [
"""SPEECHT5_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""SpeechT5ForSpeechToText""",
"""SpeechT5ForSpeechToSpeech""",
"""SpeechT5ForTextToSpeech""",
"""SpeechT5Model""",
"""SpeechT5PreTrainedModel""",
"""SpeechT5HifiGan""",
]
if TYPE_CHECKING:
from .configuration_speechta import (
SPEECHT5_PRETRAINED_CONFIG_ARCHIVE_MAP,
SPEECHT5_PRETRAINED_HIFIGAN_CONFIG_ARCHIVE_MAP,
SpeechTaConfig,
SpeechTaHifiGanConfig,
)
from .feature_extraction_speechta import SpeechTaFeatureExtractor
from .processing_speechta import SpeechTaProcessor
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_speechta import SpeechTaTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_speechta import (
SPEECHT5_PRETRAINED_MODEL_ARCHIVE_LIST,
SpeechTaForSpeechToSpeech,
SpeechTaForSpeechToText,
SpeechTaForTextToSpeech,
SpeechTaHifiGan,
SpeechTaModel,
SpeechTaPreTrainedModel,
)
else:
import sys
_UpperCAmelCase : Any = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 50
|
import argparse
import json
import os
import time
import zipfile
from get_ci_error_statistics import download_artifact, get_artifacts_links
from transformers import logging
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Dict:
lowerCamelCase__ : str = set()
lowerCamelCase__ : Any = []
def parse_line(_UpperCAmelCase ):
for line in fp:
if isinstance(_UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Any = line.decode('UTF-8' )
if "warnings summary (final)" in line:
continue
# This means we are outside the body of a warning
elif not line.startswith(' ' ):
# process a single warning and move it to `selected_warnings`.
if len(_UpperCAmelCase ) > 0:
lowerCamelCase__ : str = '\n'.join(_UpperCAmelCase )
# Only keep the warnings specified in `targets`
if any(F""": {x}: """ in warning for x in targets ):
selected_warnings.add(_UpperCAmelCase )
buffer.clear()
continue
else:
lowerCamelCase__ : List[str] = line.strip()
buffer.append(_UpperCAmelCase )
if from_gh:
for filename in os.listdir(_UpperCAmelCase ):
lowerCamelCase__ : Dict = os.path.join(_UpperCAmelCase , _UpperCAmelCase )
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
else:
try:
with zipfile.ZipFile(_UpperCAmelCase ) as z:
for filename in z.namelist():
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with z.open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
except Exception:
logger.warning(
F"""{artifact_path} is either an invalid zip file or something else wrong. This file is skipped.""" )
return selected_warnings
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = set()
lowerCamelCase__ : Optional[int] = [os.path.join(_UpperCAmelCase , _UpperCAmelCase ) for p in os.listdir(_UpperCAmelCase ) if (p.endswith('.zip' ) or from_gh)]
for p in paths:
selected_warnings.update(extract_warnings_from_single_artifact(_UpperCAmelCase , _UpperCAmelCase ) )
return selected_warnings
if __name__ == "__main__":
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
return values.split(',' )
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""--workflow_run_id""", type=str, required=True, help="""A GitHub Actions workflow run id.""")
parser.add_argument(
"""--output_dir""",
type=str,
required=True,
help="""Where to store the downloaded artifacts and other result files.""",
)
parser.add_argument("""--token""", default=None, type=str, help="""A token that has actions:read permission.""")
# optional parameters
parser.add_argument(
"""--targets""",
default="""DeprecationWarning,UserWarning,FutureWarning""",
type=list_str,
help="""Comma-separated list of target warning(s) which we want to extract.""",
)
parser.add_argument(
"""--from_gh""",
action="""store_true""",
help="""If running from a GitHub action workflow and collecting warnings from its artifacts.""",
)
_UpperCAmelCase : Union[str, Any] = parser.parse_args()
_UpperCAmelCase : Dict = args.from_gh
if from_gh:
# The artifacts have to be downloaded using `actions/download-artifact@v3`
pass
else:
os.makedirs(args.output_dir, exist_ok=True)
# get download links
_UpperCAmelCase : Union[str, Any] = get_artifacts_links(args.workflow_run_id, token=args.token)
with open(os.path.join(args.output_dir, """artifacts.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(artifacts, fp, ensure_ascii=False, indent=4)
# download artifacts
for idx, (name, url) in enumerate(artifacts.items()):
print(name)
print(url)
print("""=""" * 80)
download_artifact(name, url, args.output_dir, args.token)
# Be gentle to GitHub
time.sleep(1)
# extract warnings from artifacts
_UpperCAmelCase : Dict = extract_warnings(args.output_dir, args.targets)
_UpperCAmelCase : Optional[Any] = sorted(selected_warnings)
with open(os.path.join(args.output_dir, """selected_warnings.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(selected_warnings, fp, ensure_ascii=False, indent=4)
| 50
| 1
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_UpperCAmelCase : str = logging.get_logger(__name__)
_UpperCAmelCase : Any = {
"""facebook/timesformer""": """https://huggingface.co/facebook/timesformer/resolve/main/config.json""",
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """timesformer"""
def __init__( self : Optional[int] , UpperCAmelCase : Any=224 , UpperCAmelCase : Union[str, Any]=16 , UpperCAmelCase : List[str]=3 , UpperCAmelCase : Optional[Any]=8 , UpperCAmelCase : Any=768 , UpperCAmelCase : str=12 , UpperCAmelCase : Dict=12 , UpperCAmelCase : List[Any]=3072 , UpperCAmelCase : Any="gelu" , UpperCAmelCase : Optional[Any]=0.0 , UpperCAmelCase : str=0.0 , UpperCAmelCase : Tuple=0.0_2 , UpperCAmelCase : str=1e-6 , UpperCAmelCase : List[Any]=True , UpperCAmelCase : Dict="divided_space_time" , UpperCAmelCase : Dict=0 , **UpperCAmelCase : Tuple , ) -> int:
super().__init__(**UpperCAmelCase )
lowerCamelCase__ : Tuple = image_size
lowerCamelCase__ : Dict = patch_size
lowerCamelCase__ : int = num_channels
lowerCamelCase__ : Dict = num_frames
lowerCamelCase__ : Union[str, Any] = hidden_size
lowerCamelCase__ : int = num_hidden_layers
lowerCamelCase__ : Optional[Any] = num_attention_heads
lowerCamelCase__ : Optional[int] = intermediate_size
lowerCamelCase__ : List[Any] = hidden_act
lowerCamelCase__ : List[str] = hidden_dropout_prob
lowerCamelCase__ : Tuple = attention_probs_dropout_prob
lowerCamelCase__ : Tuple = initializer_range
lowerCamelCase__ : Optional[Any] = layer_norm_eps
lowerCamelCase__ : Optional[int] = qkv_bias
lowerCamelCase__ : int = attention_type
lowerCamelCase__ : int = drop_path_rate
| 50
|
import flax.linen as nn
import jax
import jax.numpy as jnp
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : Any ) -> Any:
lowerCamelCase__ : str = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : int , UpperCAmelCase : Dict ) -> Optional[Any]:
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : str = hidden_states.shape
lowerCamelCase__ : Union[str, Any] = jax.image.resize(
UpperCAmelCase , shape=(batch, height * 2, width * 2, channels) , method='nearest' , )
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Tuple = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : str , UpperCAmelCase : Union[str, Any] ) -> Optional[Any]:
# pad = ((0, 0), (0, 1), (0, 1), (0, 0)) # pad height and width dim
# hidden_states = jnp.pad(hidden_states, pad_width=pad)
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = 0.0
UpperCAmelCase__ = None
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> Union[str, Any]:
lowerCamelCase__ : Optional[Any] = self.in_channels if self.out_channels is None else self.out_channels
lowerCamelCase__ : Tuple = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : int = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Union[str, Any] = nn.Dense(UpperCAmelCase , dtype=self.dtype )
lowerCamelCase__ : Union[str, Any] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : List[Any] = nn.Dropout(self.dropout_prob )
lowerCamelCase__ : Tuple = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Optional[Any] = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
lowerCamelCase__ : Union[str, Any] = None
if use_nin_shortcut:
lowerCamelCase__ : Dict = nn.Conv(
UpperCAmelCase , kernel_size=(1, 1) , strides=(1, 1) , padding='VALID' , dtype=self.dtype , )
def __call__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : int=True ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = hidden_states
lowerCamelCase__ : List[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[Any] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Any = self.conva(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.time_emb_proj(nn.swish(UpperCAmelCase ) )
lowerCamelCase__ : List[str] = jnp.expand_dims(jnp.expand_dims(UpperCAmelCase , 1 ) , 1 )
lowerCamelCase__ : List[str] = hidden_states + temb
lowerCamelCase__ : Optional[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[str] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = self.dropout(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = self.conva(UpperCAmelCase )
if self.conv_shortcut is not None:
lowerCamelCase__ : Dict = self.conv_shortcut(UpperCAmelCase )
return hidden_states + residual
| 50
| 1
|
import json
import os
from typing import Optional, Tuple
import regex as re
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_UpperCAmelCase : str = logging.get_logger(__name__)
_UpperCAmelCase : Optional[int] = {
"""vocab_file""": """vocab.json""",
"""merges_file""": """merges.txt""",
}
_UpperCAmelCase : List[str] = {
"""vocab_file""": {"""ctrl""": """https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-vocab.json"""},
"""merges_file""": {"""ctrl""": """https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-merges.txt"""},
}
_UpperCAmelCase : str = {
"""ctrl""": 2_56,
}
_UpperCAmelCase : Tuple = {
"""Pregnancy""": 16_86_29,
"""Christianity""": 76_75,
"""Explain""": 10_64_23,
"""Fitness""": 6_34_40,
"""Saving""": 6_31_63,
"""Ask""": 2_71_71,
"""Ass""": 9_59_85,
"""Joke""": 16_35_09,
"""Questions""": 4_56_22,
"""Thoughts""": 4_96_05,
"""Retail""": 5_23_42,
"""Feminism""": 16_43_38,
"""Writing""": 1_19_92,
"""Atheism""": 19_22_63,
"""Netflix""": 4_86_16,
"""Computing""": 3_96_39,
"""Opinion""": 4_32_13,
"""Alone""": 4_49_67,
"""Funny""": 5_89_17,
"""Gaming""": 4_03_58,
"""Human""": 40_88,
"""India""": 13_31,
"""Joker""": 7_71_38,
"""Diet""": 3_62_06,
"""Legal""": 1_18_59,
"""Norman""": 49_39,
"""Tip""": 7_26_89,
"""Weight""": 5_23_43,
"""Movies""": 4_62_73,
"""Running""": 2_34_25,
"""Science""": 20_90,
"""Horror""": 3_77_93,
"""Confession""": 6_05_72,
"""Finance""": 1_22_50,
"""Politics""": 1_63_60,
"""Scary""": 19_19_85,
"""Support""": 1_26_54,
"""Technologies""": 3_25_16,
"""Teenage""": 6_61_60,
"""Event""": 3_27_69,
"""Learned""": 6_74_60,
"""Notion""": 18_27_70,
"""Wikipedia""": 3_75_83,
"""Books""": 66_65,
"""Extract""": 7_60_50,
"""Confessions""": 10_27_01,
"""Conspiracy""": 7_59_32,
"""Links""": 6_36_74,
"""Narcissus""": 15_04_25,
"""Relationship""": 5_47_66,
"""Relationships""": 13_47_96,
"""Reviews""": 4_16_71,
"""News""": 42_56,
"""Translation""": 2_68_20,
"""multilingual""": 12_84_06,
}
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Optional[Any]:
lowerCamelCase__ : str = set()
lowerCamelCase__ : Tuple = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
lowerCamelCase__ : int = char
lowerCamelCase__ : Union[str, Any] = set(_UpperCAmelCase )
return pairs
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = CONTROL_CODES
def __init__( self : List[str] , UpperCAmelCase : List[str] , UpperCAmelCase : Any , UpperCAmelCase : str="<unk>" , **UpperCAmelCase : Union[str, Any] ) -> List[str]:
super().__init__(unk_token=UpperCAmelCase , **UpperCAmelCase )
with open(UpperCAmelCase , encoding='utf-8' ) as vocab_handle:
lowerCamelCase__ : Optional[int] = json.load(UpperCAmelCase )
lowerCamelCase__ : Any = {v: k for k, v in self.encoder.items()}
with open(UpperCAmelCase , encoding='utf-8' ) as merges_handle:
lowerCamelCase__ : Optional[Any] = merges_handle.read().split('\n' )[1:-1]
lowerCamelCase__ : Any = [tuple(merge.split() ) for merge in merges]
lowerCamelCase__ : List[Any] = dict(zip(UpperCAmelCase , range(len(UpperCAmelCase ) ) ) )
lowerCamelCase__ : List[Any] = {}
@property
def A_ ( self : Tuple ) -> Optional[int]:
return len(self.encoder )
def A_ ( self : Dict ) -> Union[str, Any]:
return dict(self.encoder , **self.added_tokens_encoder )
def A_ ( self : Union[str, Any] , UpperCAmelCase : List[Any] ) -> List[Any]:
if token in self.cache:
return self.cache[token]
lowerCamelCase__ : Union[str, Any] = tuple(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = tuple(list(word[:-1] ) + [word[-1] + '</w>'] )
lowerCamelCase__ : Optional[int] = get_pairs(UpperCAmelCase )
if not pairs:
return token
while True:
lowerCamelCase__ : str = min(UpperCAmelCase , key=lambda UpperCAmelCase : self.bpe_ranks.get(UpperCAmelCase , float('inf' ) ) )
if bigram not in self.bpe_ranks:
break
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = bigram
lowerCamelCase__ : List[Any] = []
lowerCamelCase__ : str = 0
while i < len(UpperCAmelCase ):
try:
lowerCamelCase__ : List[str] = word.index(UpperCAmelCase , UpperCAmelCase )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
lowerCamelCase__ : Optional[int] = j
if word[i] == first and i < len(UpperCAmelCase ) - 1 and word[i + 1] == second:
new_word.append(first + second )
i += 2
else:
new_word.append(word[i] )
i += 1
lowerCamelCase__ : Tuple = tuple(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = new_word
if len(UpperCAmelCase ) == 1:
break
else:
lowerCamelCase__ : Dict = get_pairs(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = '@@ '.join(UpperCAmelCase )
lowerCamelCase__ : str = word[:-4]
lowerCamelCase__ : Dict = word
return word
def A_ ( self : Optional[int] , UpperCAmelCase : str ) -> List[str]:
lowerCamelCase__ : Dict = []
lowerCamelCase__ : Dict = re.findall(R'\S+\n?' , UpperCAmelCase )
for token in words:
split_tokens.extend(list(self.bpe(UpperCAmelCase ).split(' ' ) ) )
return split_tokens
def A_ ( self : Optional[Any] , UpperCAmelCase : Tuple ) -> int:
return self.encoder.get(UpperCAmelCase , self.encoder.get(self.unk_token ) )
def A_ ( self : int , UpperCAmelCase : Union[str, Any] ) -> int:
return self.decoder.get(UpperCAmelCase , self.unk_token )
def A_ ( self : Optional[int] , UpperCAmelCase : Dict ) -> int:
lowerCamelCase__ : Optional[Any] = ' '.join(UpperCAmelCase ).replace('@@ ' , '' ).strip()
return out_string
def A_ ( self : Any , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not os.path.isdir(UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCamelCase__ : Union[str, Any] = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
lowerCamelCase__ : Tuple = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['merges_file'] )
with open(UpperCAmelCase , 'w' , encoding='utf-8' ) as f:
f.write(json.dumps(self.encoder , indent=2 , sort_keys=UpperCAmelCase , ensure_ascii=UpperCAmelCase ) + '\n' )
lowerCamelCase__ : List[str] = 0
with open(UpperCAmelCase , 'w' , encoding='utf-8' ) as writer:
writer.write('#version: 0.2\n' )
for bpe_tokens, token_index in sorted(self.bpe_ranks.items() , key=lambda UpperCAmelCase : kv[1] ):
if index != token_index:
logger.warning(
F"""Saving vocabulary to {merge_file}: BPE merge indices are not consecutive."""
' Please check that the tokenizer is not corrupted!' )
lowerCamelCase__ : Optional[int] = token_index
writer.write(' '.join(UpperCAmelCase ) + '\n' )
index += 1
return vocab_file, merge_file
# def decode(self, token_ids, skip_special_tokens=False, clean_up_tokenization_spaces=True):
# filtered_tokens = ' '.join(self.convert_ids_to_tokens(token_ids, skip_special_tokens=skip_special_tokens))
# tokens_generated_so_far = re.sub('(@@ )', '', string=filtered_tokens)
# tokens_generated_so_far = re.sub('(@@ ?$)', '', string=tokens_generated_so_far)
# return ''.join(tokens_generated_so_far)
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Optional[Any] = set()
# edges = list of graph's edges
lowerCamelCase__ : List[str] = get_edges(_UpperCAmelCase )
# While there are still elements in edges list, take an arbitrary edge
# (from_node, to_node) and add his extremity to chosen_vertices and then
# remove all arcs adjacent to the from_node and to_node
while edges:
lowerCamelCase__ , lowerCamelCase__ : str = edges.pop()
chosen_vertices.add(_UpperCAmelCase )
chosen_vertices.add(_UpperCAmelCase )
for edge in edges.copy():
if from_node in edge or to_node in edge:
edges.discard(_UpperCAmelCase )
return chosen_vertices
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Union[str, Any] = set()
for from_node, to_nodes in graph.items():
for to_node in to_nodes:
edges.add((from_node, to_node) )
return edges
if __name__ == "__main__":
import doctest
doctest.testmod()
# graph = {0: [1, 3], 1: [0, 3], 2: [0, 3, 4], 3: [0, 1, 2], 4: [2, 3]}
# print(f"Matching vertex cover:\n{matching_min_vertex_cover(graph)}")
| 50
| 1
|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
_UpperCAmelCase : int = {
"""configuration_nezha""": ["""NEZHA_PRETRAINED_CONFIG_ARCHIVE_MAP""", """NezhaConfig"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_UpperCAmelCase : Any = [
"""NEZHA_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""NezhaForNextSentencePrediction""",
"""NezhaForMaskedLM""",
"""NezhaForPreTraining""",
"""NezhaForMultipleChoice""",
"""NezhaForQuestionAnswering""",
"""NezhaForSequenceClassification""",
"""NezhaForTokenClassification""",
"""NezhaModel""",
"""NezhaPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_nezha import NEZHA_PRETRAINED_CONFIG_ARCHIVE_MAP, NezhaConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_nezha import (
NEZHA_PRETRAINED_MODEL_ARCHIVE_LIST,
NezhaForMaskedLM,
NezhaForMultipleChoice,
NezhaForNextSentencePrediction,
NezhaForPreTraining,
NezhaForQuestionAnswering,
NezhaForSequenceClassification,
NezhaForTokenClassification,
NezhaModel,
NezhaPreTrainedModel,
)
else:
import sys
_UpperCAmelCase : Any = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 50
|
from __future__ import annotations
import math
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
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(_UpperCAmelCase ) + 1 ) , 6 ):
if number % i == 0 or number % (i + 2) == 0:
return False
return True
_UpperCAmelCase : Any = [num for num in range(3, 10_00_01, 2) if not is_prime(num)]
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list[int]:
if not isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise ValueError('n must be an integer' )
if n <= 0:
raise ValueError('n must be >= 0' )
lowerCamelCase__ : int = []
for num in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : Union[str, Any] = 0
while 2 * i * i <= odd_composites[num]:
lowerCamelCase__ : Dict = odd_composites[num] - 2 * i * i
if is_prime(_UpperCAmelCase ):
break
i += 1
else:
list_nums.append(odd_composites[num] )
if len(_UpperCAmelCase ) == n:
return list_nums
return []
def SCREAMING_SNAKE_CASE ( ) -> int:
return compute_nums(1 )[0]
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
from __future__ import annotations
import math
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> list:
if len(_UpperCAmelCase ) != 2 or len(a[0] ) != 2 or len(_UpperCAmelCase ) != 2 or len(b[0] ) != 2:
raise Exception('Matrices are not 2x2' )
lowerCamelCase__ : Optional[int] = [
[a[0][0] * b[0][0] + a[0][1] * b[1][0], a[0][0] * b[0][1] + a[0][1] * b[1][1]],
[a[1][0] * b[0][0] + a[1][1] * b[1][0], a[1][0] * b[0][1] + a[1][1] * b[1][1]],
]
return new_matrix
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Optional[Any]:
return [
[matrix_a[row][col] + matrix_b[row][col] for col in range(len(matrix_a[row] ) )]
for row in range(len(_UpperCAmelCase ) )
]
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> int:
return [
[matrix_a[row][col] - matrix_b[row][col] for col in range(len(matrix_a[row] ) )]
for row in range(len(_UpperCAmelCase ) )
]
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> tuple[list, list, list, list]:
if len(_UpperCAmelCase ) % 2 != 0 or len(a[0] ) % 2 != 0:
raise Exception('Odd matrices are not supported!' )
lowerCamelCase__ : List[str] = len(_UpperCAmelCase )
lowerCamelCase__ : List[str] = matrix_length // 2
lowerCamelCase__ : Union[str, Any] = [[a[i][j] for j in range(_UpperCAmelCase , _UpperCAmelCase )] for i in range(_UpperCAmelCase )]
lowerCamelCase__ : Optional[int] = [
[a[i][j] for j in range(_UpperCAmelCase , _UpperCAmelCase )] for i in range(_UpperCAmelCase , _UpperCAmelCase )
]
lowerCamelCase__ : List[Any] = [[a[i][j] for j in range(_UpperCAmelCase )] for i in range(_UpperCAmelCase )]
lowerCamelCase__ : str = [[a[i][j] for j in range(_UpperCAmelCase )] for i in range(_UpperCAmelCase , _UpperCAmelCase )]
return top_left, top_right, bot_left, bot_right
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> tuple[int, int]:
return len(_UpperCAmelCase ), len(matrix[0] )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> None:
print('\n'.join(str(_UpperCAmelCase ) for line in matrix ) )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> list:
if matrix_dimensions(_UpperCAmelCase ) == (2, 2):
return default_matrix_multiplication(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = split_matrix(_UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : str = split_matrix(_UpperCAmelCase )
lowerCamelCase__ : List[Any] = actual_strassen(_UpperCAmelCase , matrix_subtraction(_UpperCAmelCase , _UpperCAmelCase ) )
lowerCamelCase__ : Tuple = actual_strassen(matrix_addition(_UpperCAmelCase , _UpperCAmelCase ) , _UpperCAmelCase )
lowerCamelCase__ : Optional[int] = actual_strassen(matrix_addition(_UpperCAmelCase , _UpperCAmelCase ) , _UpperCAmelCase )
lowerCamelCase__ : Optional[int] = actual_strassen(_UpperCAmelCase , matrix_subtraction(_UpperCAmelCase , _UpperCAmelCase ) )
lowerCamelCase__ : Optional[Any] = actual_strassen(matrix_addition(_UpperCAmelCase , _UpperCAmelCase ) , matrix_addition(_UpperCAmelCase , _UpperCAmelCase ) )
lowerCamelCase__ : List[Any] = actual_strassen(matrix_subtraction(_UpperCAmelCase , _UpperCAmelCase ) , matrix_addition(_UpperCAmelCase , _UpperCAmelCase ) )
lowerCamelCase__ : Union[str, Any] = actual_strassen(matrix_subtraction(_UpperCAmelCase , _UpperCAmelCase ) , matrix_addition(_UpperCAmelCase , _UpperCAmelCase ) )
lowerCamelCase__ : Dict = matrix_addition(matrix_subtraction(matrix_addition(_UpperCAmelCase , _UpperCAmelCase ) , _UpperCAmelCase ) , _UpperCAmelCase )
lowerCamelCase__ : Any = matrix_addition(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : Tuple = matrix_addition(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : int = matrix_subtraction(matrix_subtraction(matrix_addition(_UpperCAmelCase , _UpperCAmelCase ) , _UpperCAmelCase ) , _UpperCAmelCase )
# construct the new matrix from our 4 quadrants
lowerCamelCase__ : Dict = []
for i in range(len(_UpperCAmelCase ) ):
new_matrix.append(top_left[i] + top_right[i] )
for i in range(len(_UpperCAmelCase ) ):
new_matrix.append(bot_left[i] + bot_right[i] )
return new_matrix
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> list:
if matrix_dimensions(_UpperCAmelCase )[1] != matrix_dimensions(_UpperCAmelCase )[0]:
lowerCamelCase__ : Any = (
'Unable to multiply these matrices, please check the dimensions.\n'
F"""Matrix A: {matrixa}\n"""
F"""Matrix B: {matrixa}"""
)
raise Exception(_UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = matrix_dimensions(_UpperCAmelCase )
lowerCamelCase__ : Optional[int] = matrix_dimensions(_UpperCAmelCase )
if dimensiona[0] == dimensiona[1] and dimensiona[0] == dimensiona[1]:
return [matrixa, matrixa]
lowerCamelCase__ : Optional[Any] = max(*_UpperCAmelCase , *_UpperCAmelCase )
lowerCamelCase__ : Dict = int(math.pow(2 , math.ceil(math.loga(_UpperCAmelCase ) ) ) )
lowerCamelCase__ : str = matrixa
lowerCamelCase__ : List[Any] = matrixa
# Adding zeros to the matrices so that the arrays dimensions are the same and also
# power of 2
for i in range(0 , _UpperCAmelCase ):
if i < dimensiona[0]:
for _ in range(dimensiona[1] , _UpperCAmelCase ):
new_matrixa[i].append(0 )
else:
new_matrixa.append([0] * maxim )
if i < dimensiona[0]:
for _ in range(dimensiona[1] , _UpperCAmelCase ):
new_matrixa[i].append(0 )
else:
new_matrixa.append([0] * maxim )
lowerCamelCase__ : int = actual_strassen(_UpperCAmelCase , _UpperCAmelCase )
# Removing the additional zeros
for i in range(0 , _UpperCAmelCase ):
if i < dimensiona[0]:
for _ in range(dimensiona[1] , _UpperCAmelCase ):
final_matrix[i].pop()
else:
final_matrix.pop()
return final_matrix
if __name__ == "__main__":
_UpperCAmelCase : List[Any] = [
[2, 3, 4, 5],
[6, 4, 3, 1],
[2, 3, 6, 7],
[3, 1, 2, 4],
[2, 3, 4, 5],
[6, 4, 3, 1],
[2, 3, 6, 7],
[3, 1, 2, 4],
[2, 3, 4, 5],
[6, 2, 3, 1],
]
_UpperCAmelCase : Tuple = [[0, 2, 1, 1], [16, 2, 3, 3], [2, 2, 7, 7], [13, 11, 22, 4]]
print(strassen(matrixa, matrixa))
| 50
|
import argparse
from collections import OrderedDict
from pathlib import Path
import requests
import torch
from PIL import Image
from transformers import GLPNConfig, GLPNForDepthEstimation, GLPNImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
_UpperCAmelCase : Dict = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[str]:
lowerCamelCase__ : str = OrderedDict()
for key, value in state_dict.items():
if key.startswith('module.encoder' ):
lowerCamelCase__ : Optional[Any] = key.replace('module.encoder' , 'glpn.encoder' )
if key.startswith('module.decoder' ):
lowerCamelCase__ : List[str] = key.replace('module.decoder' , 'decoder.stages' )
if "patch_embed" in key:
# replace for example patch_embed1 by patch_embeddings.0
lowerCamelCase__ : Dict = key[key.find('patch_embed' ) + len('patch_embed' )]
lowerCamelCase__ : Tuple = key.replace(F"""patch_embed{idx}""" , F"""patch_embeddings.{int(_UpperCAmelCase )-1}""" )
if "norm" in key:
lowerCamelCase__ : str = key.replace('norm' , 'layer_norm' )
if "glpn.encoder.layer_norm" in key:
# replace for example layer_norm1 by layer_norm.0
lowerCamelCase__ : Dict = key[key.find('glpn.encoder.layer_norm' ) + len('glpn.encoder.layer_norm' )]
lowerCamelCase__ : str = key.replace(F"""layer_norm{idx}""" , F"""layer_norm.{int(_UpperCAmelCase )-1}""" )
if "layer_norm1" in key:
lowerCamelCase__ : Optional[int] = key.replace('layer_norm1' , 'layer_norm_1' )
if "layer_norm2" in key:
lowerCamelCase__ : Optional[int] = key.replace('layer_norm2' , 'layer_norm_2' )
if "block" in key:
# replace for example block1 by block.0
lowerCamelCase__ : List[Any] = key[key.find('block' ) + len('block' )]
lowerCamelCase__ : int = key.replace(F"""block{idx}""" , F"""block.{int(_UpperCAmelCase )-1}""" )
if "attn.q" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('attn.q' , 'attention.self.query' )
if "attn.proj" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('attn.proj' , 'attention.output.dense' )
if "attn" in key:
lowerCamelCase__ : Dict = key.replace('attn' , 'attention.self' )
if "fc1" in key:
lowerCamelCase__ : Dict = key.replace('fc1' , 'dense1' )
if "fc2" in key:
lowerCamelCase__ : Any = key.replace('fc2' , 'dense2' )
if "linear_pred" in key:
lowerCamelCase__ : Dict = key.replace('linear_pred' , 'classifier' )
if "linear_fuse" in key:
lowerCamelCase__ : Tuple = key.replace('linear_fuse.conv' , 'linear_fuse' )
lowerCamelCase__ : List[str] = key.replace('linear_fuse.bn' , 'batch_norm' )
if "linear_c" in key:
# replace for example linear_c4 by linear_c.3
lowerCamelCase__ : Optional[Any] = key[key.find('linear_c' ) + len('linear_c' )]
lowerCamelCase__ : Dict = key.replace(F"""linear_c{idx}""" , F"""linear_c.{int(_UpperCAmelCase )-1}""" )
if "bot_conv" in key:
lowerCamelCase__ : str = key.replace('bot_conv' , '0.convolution' )
if "skip_conv1" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('skip_conv1' , '1.convolution' )
if "skip_conv2" in key:
lowerCamelCase__ : List[Any] = key.replace('skip_conv2' , '2.convolution' )
if "fusion1" in key:
lowerCamelCase__ : Optional[int] = key.replace('fusion1' , '1.fusion' )
if "fusion2" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('fusion2' , '2.fusion' )
if "fusion3" in key:
lowerCamelCase__ : List[Any] = key.replace('fusion3' , '3.fusion' )
if "fusion" in key and "conv" in key:
lowerCamelCase__ : str = key.replace('conv' , 'convolutional_layer' )
if key.startswith('module.last_layer_depth' ):
lowerCamelCase__ : Dict = key.replace('module.last_layer_depth' , 'head.head' )
lowerCamelCase__ : str = value
return new_state_dict
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
# for each of the encoder blocks:
for i in range(config.num_encoder_blocks ):
for j in range(config.depths[i] ):
# read in weights + bias of keys and values (which is a single matrix in the original implementation)
lowerCamelCase__ : Any = state_dict.pop(F"""glpn.encoder.block.{i}.{j}.attention.self.kv.weight""" )
lowerCamelCase__ : Optional[Any] = state_dict.pop(F"""glpn.encoder.block.{i}.{j}.attention.self.kv.bias""" )
# next, add keys and values (in that order) to the state dict
lowerCamelCase__ : Optional[int] = kv_weight[
: config.hidden_sizes[i], :
]
lowerCamelCase__ : Optional[int] = kv_bias[: config.hidden_sizes[i]]
lowerCamelCase__ : Any = kv_weight[
config.hidden_sizes[i] :, :
]
lowerCamelCase__ : Dict = kv_bias[config.hidden_sizes[i] :]
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : List[str] = 'http://images.cocodataset.org/val2017/000000039769.jpg'
lowerCamelCase__ : Tuple = Image.open(requests.get(_UpperCAmelCase , stream=_UpperCAmelCase ).raw )
return image
@torch.no_grad()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=False , _UpperCAmelCase=None ) -> Optional[int]:
lowerCamelCase__ : str = GLPNConfig(hidden_sizes=[64, 128, 320, 512] , decoder_hidden_size=64 , depths=[3, 8, 27, 3] )
# load image processor (only resize + rescale)
lowerCamelCase__ : Union[str, Any] = GLPNImageProcessor()
# prepare image
lowerCamelCase__ : str = prepare_img()
lowerCamelCase__ : Tuple = image_processor(images=_UpperCAmelCase , return_tensors='pt' ).pixel_values
logger.info('Converting model...' )
# load original state dict
lowerCamelCase__ : Any = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
# rename keys
lowerCamelCase__ : str = rename_keys(_UpperCAmelCase )
# key and value matrices need special treatment
read_in_k_v(_UpperCAmelCase , _UpperCAmelCase )
# create HuggingFace model and load state dict
lowerCamelCase__ : Dict = GLPNForDepthEstimation(_UpperCAmelCase )
model.load_state_dict(_UpperCAmelCase )
model.eval()
# forward pass
lowerCamelCase__ : List[str] = model(_UpperCAmelCase )
lowerCamelCase__ : Tuple = outputs.predicted_depth
# verify output
if model_name is not None:
if "nyu" in model_name:
lowerCamelCase__ : List[Any] = torch.tensor(
[[4.4_147, 4.0_873, 4.0_673], [3.7_890, 3.2_881, 3.1_525], [3.7_674, 3.5_423, 3.4_913]] )
elif "kitti" in model_name:
lowerCamelCase__ : List[str] = torch.tensor(
[[3.4_291, 2.7_865, 2.5_151], [3.2_841, 2.7_021, 2.3_502], [3.1_147, 2.4_625, 2.2_481]] )
else:
raise ValueError(F"""Unknown model name: {model_name}""" )
lowerCamelCase__ : Tuple = torch.Size([1, 480, 640] )
assert predicted_depth.shape == expected_shape
assert torch.allclose(predicted_depth[0, :3, :3] , _UpperCAmelCase , atol=1e-4 )
print('Looks ok!' )
# finally, push to hub if required
if push_to_hub:
logger.info('Pushing model and image processor to the hub...' )
model.push_to_hub(
repo_path_or_name=Path(_UpperCAmelCase , _UpperCAmelCase ) , organization='nielsr' , commit_message='Add model' , use_temp_dir=_UpperCAmelCase , )
image_processor.push_to_hub(
repo_path_or_name=Path(_UpperCAmelCase , _UpperCAmelCase ) , organization='nielsr' , commit_message='Add image processor' , use_temp_dir=_UpperCAmelCase , )
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
parser.add_argument(
"""--checkpoint_path""",
default=None,
type=str,
help="""Path to the original PyTorch checkpoint (.pth file).""",
)
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 to upload the model to the HuggingFace hub."""
)
parser.add_argument(
"""--model_name""",
default="""glpn-kitti""",
type=str,
help="""Name of the model in case you're pushing to the hub.""",
)
_UpperCAmelCase : int = parser.parse_args()
convert_glpn_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 50
| 1
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> bool:
lowerCamelCase__ : List[str] = len(_UpperCAmelCase )
lowerCamelCase__ : str = [[False] * (required_sum + 1) for _ in range(arr_len + 1 )]
# for each arr value, a sum of zero(0) can be formed by not taking any element
# hence True/1
for i in range(arr_len + 1 ):
lowerCamelCase__ : Tuple = True
# sum is not zero and set is empty then false
for i in range(1 , required_sum + 1 ):
lowerCamelCase__ : Dict = False
for i in range(1 , arr_len + 1 ):
for j in range(1 , required_sum + 1 ):
if arr[i - 1] > j:
lowerCamelCase__ : str = subset[i - 1][j]
if arr[i - 1] <= j:
lowerCamelCase__ : Dict = subset[i - 1][j] or subset[i - 1][j - arr[i - 1]]
return subset[arr_len][required_sum]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
|
import tempfile
import unittest
from transformers import TaConfig, is_torch_available
from transformers.testing_utils import (
require_sentencepiece,
require_tokenizers,
require_torch,
slow,
torch_device,
)
from ...generation.test_utils import GenerationTesterMixin
from ...test_modeling_common import ModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import AutoTokenizer, UMTaForConditionalGeneration, UMTaForQuestionAnswering, UMTaModel
class lowerCAmelCase :
def __init__( self : Tuple , UpperCAmelCase : Dict , UpperCAmelCase : Union[str, Any]=99 , UpperCAmelCase : str=13 , UpperCAmelCase : List[str]=7 , UpperCAmelCase : str=9 , UpperCAmelCase : Optional[int]=True , UpperCAmelCase : str=True , UpperCAmelCase : Any=False , UpperCAmelCase : Union[str, Any]=32 , UpperCAmelCase : List[str]=5 , UpperCAmelCase : Tuple=4 , UpperCAmelCase : Union[str, Any]=37 , UpperCAmelCase : int=8 , UpperCAmelCase : List[str]=0.1 , UpperCAmelCase : Any=0.0_0_2 , UpperCAmelCase : Optional[Any]=1 , UpperCAmelCase : List[Any]=0 , UpperCAmelCase : Union[str, Any]=0 , UpperCAmelCase : Tuple=None , UpperCAmelCase : Optional[Any]=None , ) -> Union[str, Any]:
lowerCamelCase__ : int = parent
lowerCamelCase__ : Any = batch_size
lowerCamelCase__ : Optional[int] = encoder_seq_length
lowerCamelCase__ : int = decoder_seq_length
# For common tests
lowerCamelCase__ : List[str] = self.decoder_seq_length
lowerCamelCase__ : Optional[int] = is_training
lowerCamelCase__ : List[Any] = use_attention_mask
lowerCamelCase__ : Optional[Any] = use_labels
lowerCamelCase__ : Union[str, Any] = vocab_size
lowerCamelCase__ : Union[str, Any] = hidden_size
lowerCamelCase__ : Optional[Any] = num_hidden_layers
lowerCamelCase__ : Any = num_attention_heads
lowerCamelCase__ : str = d_ff
lowerCamelCase__ : Optional[Any] = relative_attention_num_buckets
lowerCamelCase__ : Any = dropout_rate
lowerCamelCase__ : Any = initializer_factor
lowerCamelCase__ : Union[str, Any] = eos_token_id
lowerCamelCase__ : List[str] = pad_token_id
lowerCamelCase__ : List[str] = decoder_start_token_id
lowerCamelCase__ : List[Any] = None
lowerCamelCase__ : Optional[Any] = decoder_layers
def A_ ( self : List[Any] ) -> int:
return TaConfig.from_pretrained('google/umt5-base' )
def A_ ( self : List[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[int] , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : Tuple=None , UpperCAmelCase : List[str]=None , UpperCAmelCase : str=None , UpperCAmelCase : Optional[Any]=None , UpperCAmelCase : Optional[Any]=None , ) -> List[str]:
if attention_mask is None:
lowerCamelCase__ : Optional[Any] = input_ids.ne(config.pad_token_id )
if decoder_attention_mask is None:
lowerCamelCase__ : Optional[Any] = decoder_input_ids.ne(config.pad_token_id )
if head_mask is None:
lowerCamelCase__ : int = torch.ones(config.num_hidden_layers , config.num_attention_heads , device=UpperCAmelCase )
if decoder_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
if cross_attn_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(
config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
return {
"input_ids": input_ids,
"decoder_input_ids": decoder_input_ids,
"attention_mask": attention_mask,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
"cross_attn_head_mask": cross_attn_head_mask,
}
def A_ ( self : str ) -> List[str]:
lowerCamelCase__ : Any = ids_tensor([self.batch_size, self.encoder_seq_length] , self.vocab_size )
lowerCamelCase__ : Optional[Any] = ids_tensor([self.batch_size, self.decoder_seq_length] , self.vocab_size )
# we need to clamp the input ids here to avoid having pad token in between
# this is because for NllbMoe the position_ids are prepared such that
# all pad tokens have pos id = 2 and rest are between 2..seq_length
# and the seq_length here is seq_length - num_pad_tokens
# but when using past, there is no way of knowing if the past input ids had
# pad tokens in them, which results in incorrect seq_lenth and which in turn results in
# position_ids being off by num_pad_tokens in past input
lowerCamelCase__ : List[str] = input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Union[str, Any] = decoder_input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Dict = self.get_config()
lowerCamelCase__ : Tuple = config.num_attention_heads
lowerCamelCase__ : Any = self.prepare_inputs_dict(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
return config, input_dict
def A_ ( self : Tuple ) -> Union[str, Any]:
lowerCamelCase__ , lowerCamelCase__ : Dict = self.prepare_config_and_inputs()
return config, inputs_dict
def A_ ( self : Optional[int] ) -> List[str]:
return TaConfig(
vocab_size=166 , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Union[str, Any] ) -> Dict:
return TaConfig(
vocab_size=self.vocab_size , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Any , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[str] , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Any , UpperCAmelCase : Dict , ) -> str:
lowerCamelCase__ : Dict = UMTaModel(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Optional[int] = model(
input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase , decoder_attention_mask=UpperCAmelCase , )
lowerCamelCase__ : Any = model(input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase )
lowerCamelCase__ : Dict = result.last_hidden_state
lowerCamelCase__ : Any = result.past_key_values
lowerCamelCase__ : List[Any] = result.encoder_last_hidden_state
self.parent.assertEqual(encoder_output.size() , (self.batch_size, self.encoder_seq_length, self.hidden_size) )
self.parent.assertEqual(decoder_output.size() , (self.batch_size, self.decoder_seq_length, self.hidden_size) )
# There should be `num_layers` key value embeddings stored in decoder_past
self.parent.assertEqual(len(UpperCAmelCase ) , config.num_layers )
# There should be a self attn key, a self attn value, a cross attn key and a cross attn value stored in each decoder_past tuple
self.parent.assertEqual(len(decoder_past[0] ) , 4 )
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[Any] , UpperCAmelCase : int , UpperCAmelCase : List[str] , UpperCAmelCase : Optional[Any] , ) -> Optional[int]:
lowerCamelCase__ : List[Any] = UMTaModel(config=UpperCAmelCase ).get_decoder().to(UpperCAmelCase ).eval()
# first forward pass
lowerCamelCase__ : Tuple = model(UpperCAmelCase , use_cache=UpperCAmelCase )
lowerCamelCase__ : List[Any] = model(UpperCAmelCase )
lowerCamelCase__ : int = model(UpperCAmelCase , use_cache=UpperCAmelCase )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) + 1 )
lowerCamelCase__ , lowerCamelCase__ : Dict = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
lowerCamelCase__ : Optional[int] = ids_tensor((self.batch_size, 1) , config.vocab_size )
# append to next input_ids and
lowerCamelCase__ : List[Any] = torch.cat([input_ids, next_tokens] , dim=-1 )
lowerCamelCase__ : List[str] = model(UpperCAmelCase )['last_hidden_state']
lowerCamelCase__ : str = model(UpperCAmelCase , past_key_values=UpperCAmelCase )['last_hidden_state']
# select random slice
lowerCamelCase__ : List[str] = ids_tensor((1,) , output_from_past.shape[-1] ).item()
lowerCamelCase__ : Tuple = output_from_no_past[:, -1, random_slice_idx].detach()
lowerCamelCase__ : List[str] = output_from_past[:, 0, random_slice_idx].detach()
# test that outputs are equal for slice
self.parent.assertTrue(torch.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-3 ) )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : List[str] , ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = UMTaModel(config=UpperCAmelCase ).to(UpperCAmelCase ).half().eval()
lowerCamelCase__ : Optional[int] = model(**UpperCAmelCase )['last_hidden_state']
self.parent.assertFalse(torch.isnan(UpperCAmelCase ).any().item() )
@require_torch
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(UMTaModel, UMTaForConditionalGeneration, UMTaForQuestionAnswering) if is_torch_available() else ()
)
UpperCAmelCase__ = (UMTaForConditionalGeneration,) if is_torch_available() else ()
UpperCAmelCase__ = (
{
"""conversational""": UMTaForConditionalGeneration,
"""feature-extraction""": UMTaModel,
"""summarization""": UMTaForConditionalGeneration,
"""text2text-generation""": UMTaForConditionalGeneration,
"""translation""": UMTaForConditionalGeneration,
"""question-answering""": UMTaForQuestionAnswering,
}
if is_torch_available()
else {}
)
UpperCAmelCase__ = True
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = True
UpperCAmelCase__ = True
# The small UMT5 model needs higher percentages for CPU/MP tests
UpperCAmelCase__ = [0.8, 0.9]
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = UMTaModelTester(self )
@unittest.skip('Test has a segmentation fault on torch 1.8.0' )
def A_ ( self : Tuple ) -> int:
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Tuple = UMTaModel(config_and_inputs[0] ).to(UpperCAmelCase )
with tempfile.TemporaryDirectory() as tmpdirname:
torch.onnx.export(
UpperCAmelCase , (config_and_inputs[1], config_and_inputs[3], config_and_inputs[2]) , F"""{tmpdirname}/t5_test.onnx""" , export_params=UpperCAmelCase , opset_version=9 , input_names=['input_ids', 'decoder_input_ids'] , )
@unittest.skipIf(torch_device == 'cpu' , 'Cant do half precision' )
def A_ ( self : Tuple ) -> Optional[Any]:
lowerCamelCase__ : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model_fpaa_forward(*UpperCAmelCase )
def A_ ( self : List[Any] ) -> str:
lowerCamelCase__ : int = ['encoder_attentions', 'decoder_attentions', 'cross_attentions']
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Any = config_and_inputs[0]
lowerCamelCase__ : Any = UMTaForConditionalGeneration(UpperCAmelCase ).eval()
model.to(UpperCAmelCase )
lowerCamelCase__ : Tuple = {
'head_mask': torch.zeros(config.num_layers , config.num_heads , device=UpperCAmelCase ),
'decoder_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
'cross_attn_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
}
for attn_name, (name, mask) in zip(UpperCAmelCase , head_masking.items() ):
lowerCamelCase__ : Union[str, Any] = {name: mask}
# Explicitly pass decoder_head_mask as it is required from T5 model when head_mask specified
if name == "head_mask":
lowerCamelCase__ : Union[str, Any] = torch.ones(
config.num_decoder_layers , config.num_heads , device=UpperCAmelCase )
lowerCamelCase__ : Tuple = model.generate(
config_and_inputs[1]['input_ids'] , num_beams=1 , max_length=3 , output_attentions=UpperCAmelCase , return_dict_in_generate=UpperCAmelCase , **UpperCAmelCase , )
# We check the state of decoder_attentions and cross_attentions just from the last step
lowerCamelCase__ : Union[str, Any] = out[attn_name] if attn_name == attention_names[0] else out[attn_name][-1]
self.assertEqual(sum([w.sum().item() for w in attn_weights] ) , 0.0 )
@unittest.skip('Does not work on the tiny model as we keep hitting edge cases.' )
def A_ ( self : Optional[Any] ) -> Optional[Any]:
pass
@require_torch
@require_sentencepiece
@require_tokenizers
class lowerCAmelCase ( unittest.TestCase ):
@slow
@unittest.skip(
'Unless we stop stripping left and right by default for all special tokens, the expected ids obtained here will not match the original ones. Wait for https://github.com/huggingface/transformers/pull/23909 to be merged' )
def A_ ( self : Any ) -> int:
lowerCamelCase__ : Optional[Any] = UMTaForConditionalGeneration.from_pretrained('google/umt5-small' , return_dict=UpperCAmelCase ).to(UpperCAmelCase )
lowerCamelCase__ : List[str] = AutoTokenizer.from_pretrained('google/umt5-small' , use_fast=UpperCAmelCase , legacy=UpperCAmelCase )
lowerCamelCase__ : Dict = [
'Bonjour monsieur <extra_id_0> bien <extra_id_1>.',
'No se como puedo <extra_id_0>.',
'This is the reason why we <extra_id_0> them.',
'The <extra_id_0> walks in <extra_id_1>, seats',
'A <extra_id_0> walks into a bar and orders a <extra_id_1> with <extra_id_2> pinch of <extra_id_3>.',
]
lowerCamelCase__ : Tuple = tokenizer(UpperCAmelCase , return_tensors='pt' , padding=UpperCAmelCase ).input_ids
# fmt: off
lowerCamelCase__ : Any = torch.tensor(
[
[ 38530, 210703, 256299, 1410, 256298, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 826, 321, 671, 25922, 256299, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 1460, 339, 312, 19014, 10620, 758, 256299, 2355,274, 1, 0, 0, 0, 0, 0, 0,0, 0],
[ 517, 256299, 14869, 281, 301, 256298, 275, 119983,1, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 320, 256299, 14869, 281, 2234, 289, 2275, 333,61391, 289, 256298, 543, 256297, 168714, 329, 256296,274, 1],
] )
# fmt: on
torch.testing.assert_allclose(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Optional[int] = model.generate(input_ids.to(UpperCAmelCase ) )
lowerCamelCase__ : List[Any] = [
'<pad><extra_id_0> et<extra_id_1> [eod] <extra_id_2><extra_id_55>.. [eod] 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 <extra_id_56>ajšietosto<extra_id_56>lleux<extra_id_19><extra_id_6>ajšie</s>',
'<pad><extra_id_0>.<extra_id_1>.,<0x0A>...spech <0x0A><extra_id_20> <extra_id_21></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> are not going to be a part of the world. We are not going to be a part of<extra_id_1> and<extra_id_2><0x0A><extra_id_48>.<extra_id_48></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> door<extra_id_1>, the door<extra_id_2> 피해[/</s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0>nyone who<extra_id_1> drink<extra_id_2> a<extra_id_3> alcohol<extra_id_4> A<extra_id_5> A. This<extra_id_6> I<extra_id_7><extra_id_52><extra_id_53></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
]
lowerCamelCase__ : Union[str, Any] = tokenizer.batch_decode(UpperCAmelCase )
self.assertEqual(UpperCAmelCase , UpperCAmelCase )
| 50
| 1
|
import random
import unittest
import torch
from diffusers import IFImgaImgSuperResolutionPipeline
from diffusers.utils import floats_tensor
from diffusers.utils.import_utils import is_xformers_available
from diffusers.utils.testing_utils import skip_mps, torch_device
from ..pipeline_params import TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS, TEXT_GUIDED_IMAGE_VARIATION_PARAMS
from ..test_pipelines_common import PipelineTesterMixin
from . import IFPipelineTesterMixin
@skip_mps
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = IFImgaImgSuperResolutionPipeline
UpperCAmelCase__ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {"""width""", """height"""}
UpperCAmelCase__ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({"""original_image"""} )
UpperCAmelCase__ = PipelineTesterMixin.required_optional_params - {"""latents"""}
def A_ ( self : int ) -> Tuple:
return self._get_superresolution_dummy_components()
def A_ ( self : List[Any] , UpperCAmelCase : List[str] , UpperCAmelCase : List[Any]=0 ) -> List[Any]:
if str(UpperCAmelCase ).startswith('mps' ):
lowerCamelCase__ : Tuple = torch.manual_seed(UpperCAmelCase )
else:
lowerCamelCase__ : Tuple = torch.Generator(device=UpperCAmelCase ).manual_seed(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = floats_tensor((1, 3, 32, 32) , rng=random.Random(UpperCAmelCase ) ).to(UpperCAmelCase )
lowerCamelCase__ : List[Any] = floats_tensor((1, 3, 16, 16) , rng=random.Random(UpperCAmelCase ) ).to(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = {
'prompt': 'A painting of a squirrel eating a burger',
'image': image,
'original_image': original_image,
'generator': generator,
'num_inference_steps': 2,
'output_type': 'numpy',
}
return inputs
@unittest.skipIf(
torch_device != 'cuda' or not is_xformers_available() , reason='XFormers attention is only available with CUDA and `xformers` installed' , )
def A_ ( self : int ) -> Tuple:
self._test_xformers_attention_forwardGenerator_pass(expected_max_diff=1e-3 )
def A_ ( self : Tuple ) -> List[Any]:
self._test_save_load_optional_components()
@unittest.skipIf(torch_device != 'cuda' , reason='float16 requires CUDA' )
def A_ ( self : str ) -> str:
# Due to non-determinism in save load of the hf-internal-testing/tiny-random-t5 text encoder
super().test_save_load_floataa(expected_max_diff=1e-1 )
def A_ ( self : List[Any] ) -> Optional[int]:
self._test_attention_slicing_forward_pass(expected_max_diff=1e-2 )
def A_ ( self : int ) -> Dict:
self._test_save_load_local()
def A_ ( self : str ) -> List[Any]:
self._test_inference_batch_single_identical(
expected_max_diff=1e-2 , )
| 50
|
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=None , _UpperCAmelCase="no" , _UpperCAmelCase="29500" ) -> Tuple:
lowerCamelCase__ : Dict = False
lowerCamelCase__ : Dict = False
if any(key.startswith('KAGGLE' ) for key in os.environ.keys() ):
lowerCamelCase__ : Optional[Any] = True
elif "IPython" in sys.modules:
lowerCamelCase__ : Optional[Any] = 'google.colab' in str(sys.modules['IPython'].get_ipython() )
try:
lowerCamelCase__ : List[str] = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
F"""Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}.""" )
if (in_colab or in_kaggle) and (os.environ.get('TPU_NAME' , _UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '
'your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if num_processes is None:
lowerCamelCase__ : Optional[Any] = 8
lowerCamelCase__ : List[str] = PrepareForLaunch(_UpperCAmelCase , distributed_type='TPU' )
print(F"""Launching a training on {num_processes} TPU cores.""" )
xmp.spawn(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on one CPU.' )
function(*_UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '
'inside your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if torch.cuda.is_initialized():
raise ValueError(
'To launch a multi-GPU training from your notebook, you need to avoid running any instruction '
'using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '
'function.' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port=_UpperCAmelCase , mixed_precision=_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = PrepareForLaunch(_UpperCAmelCase , distributed_type='MULTI_GPU' )
print(F"""Launching training on {num_processes} GPUs.""" )
try:
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '
'This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '
'Please review your imports and test them when running the `notebook_launcher()` to identify '
'which one is problematic.' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
lowerCamelCase__ : int = '1'
print('Launching training on MPS.' )
elif torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on CPU.' )
function(*_UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=2 ) -> Optional[Any]:
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port='29500' , accelerate_mixed_precision='no' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='yes' , ):
lowerCamelCase__ : Optional[Any] = PrepareForLaunch(_UpperCAmelCase , debug=_UpperCAmelCase )
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
| 50
| 1
|
from __future__ import annotations
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase = None ) -> list[list[str]]:
lowerCamelCase__ : int = word_bank or []
# create a table
lowerCamelCase__ : int = len(_UpperCAmelCase ) + 1
lowerCamelCase__ : list[list[list[str]]] = []
for _ in range(_UpperCAmelCase ):
table.append([] )
# seed value
lowerCamelCase__ : str = [[]] # because empty string has empty combination
# iterate through the indices
for i in range(_UpperCAmelCase ):
# condition
if table[i] != []:
for word in word_bank:
# slice condition
if target[i : i + len(_UpperCAmelCase )] == word:
lowerCamelCase__ : 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(_UpperCAmelCase )] += new_combinations
# combinations are in reverse order so reverse for better output
for combination in table[len(_UpperCAmelCase )]:
combination.reverse()
return table[len(_UpperCAmelCase )]
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"""],
)
)
| 50
|
from dataclasses import dataclass
from typing import Optional, Tuple, Union
import torch
import torch.nn as nn
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import BaseOutput
from .embeddings import GaussianFourierProjection, TimestepEmbedding, Timesteps
from .modeling_utils import ModelMixin
from .unet_ad_blocks import get_down_block, get_mid_block, get_out_block, get_up_block
@dataclass
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase ):
@register_to_config
def __init__( self : List[str] , UpperCAmelCase : int = 65536 , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 0 , UpperCAmelCase : str = "fourier" , UpperCAmelCase : bool = True , UpperCAmelCase : bool = False , UpperCAmelCase : float = 0.0 , UpperCAmelCase : Tuple[str] = ("DownBlock1DNoSkip", "DownBlock1D", "AttnDownBlock1D") , UpperCAmelCase : Tuple[str] = ("AttnUpBlock1D", "UpBlock1D", "UpBlock1DNoSkip") , UpperCAmelCase : Tuple[str] = "UNetMidBlock1D" , UpperCAmelCase : str = None , UpperCAmelCase : Tuple[int] = (32, 32, 64) , UpperCAmelCase : str = None , UpperCAmelCase : int = 8 , UpperCAmelCase : int = 1 , UpperCAmelCase : bool = False , ) -> List[Any]:
super().__init__()
lowerCamelCase__ : Optional[int] = sample_size
# time
if time_embedding_type == "fourier":
lowerCamelCase__ : Optional[Any] = GaussianFourierProjection(
embedding_size=8 , set_W_to_weight=UpperCAmelCase , log=UpperCAmelCase , flip_sin_to_cos=UpperCAmelCase )
lowerCamelCase__ : Any = 2 * block_out_channels[0]
elif time_embedding_type == "positional":
lowerCamelCase__ : List[Any] = Timesteps(
block_out_channels[0] , flip_sin_to_cos=UpperCAmelCase , downscale_freq_shift=UpperCAmelCase )
lowerCamelCase__ : Dict = block_out_channels[0]
if use_timestep_embedding:
lowerCamelCase__ : str = block_out_channels[0] * 4
lowerCamelCase__ : List[Any] = TimestepEmbedding(
in_channels=UpperCAmelCase , time_embed_dim=UpperCAmelCase , act_fn=UpperCAmelCase , out_dim=block_out_channels[0] , )
lowerCamelCase__ : Any = nn.ModuleList([] )
lowerCamelCase__ : Tuple = None
lowerCamelCase__ : List[str] = nn.ModuleList([] )
lowerCamelCase__ : Optional[int] = None
# down
lowerCamelCase__ : Optional[int] = in_channels
for i, down_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Union[str, Any] = output_channel
lowerCamelCase__ : Tuple = block_out_channels[i]
if i == 0:
input_channel += extra_in_channels
lowerCamelCase__ : Union[str, Any] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Optional[int] = get_down_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_downsample=not is_final_block or downsample_each_block , )
self.down_blocks.append(UpperCAmelCase )
# mid
lowerCamelCase__ : Optional[int] = get_mid_block(
UpperCAmelCase , in_channels=block_out_channels[-1] , mid_channels=block_out_channels[-1] , out_channels=block_out_channels[-1] , embed_dim=block_out_channels[0] , num_layers=UpperCAmelCase , add_downsample=UpperCAmelCase , )
# up
lowerCamelCase__ : Optional[int] = list(reversed(UpperCAmelCase ) )
lowerCamelCase__ : Optional[int] = reversed_block_out_channels[0]
if out_block_type is None:
lowerCamelCase__ : List[str] = out_channels
else:
lowerCamelCase__ : Any = block_out_channels[0]
for i, up_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Tuple = output_channel
lowerCamelCase__ : Union[str, Any] = (
reversed_block_out_channels[i + 1] if i < len(UpperCAmelCase ) - 1 else final_upsample_channels
)
lowerCamelCase__ : List[str] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Dict = get_up_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_upsample=not is_final_block , )
self.up_blocks.append(UpperCAmelCase )
lowerCamelCase__ : int = output_channel
# out
lowerCamelCase__ : int = norm_num_groups if norm_num_groups is not None else min(block_out_channels[0] // 4 , 32 )
lowerCamelCase__ : List[Any] = get_out_block(
out_block_type=UpperCAmelCase , num_groups_out=UpperCAmelCase , embed_dim=block_out_channels[0] , out_channels=UpperCAmelCase , act_fn=UpperCAmelCase , fc_dim=block_out_channels[-1] // 4 , )
def A_ ( self : List[Any] , UpperCAmelCase : torch.FloatTensor , UpperCAmelCase : Union[torch.Tensor, float, int] , UpperCAmelCase : bool = True , ) -> Union[UNetaDOutput, Tuple]:
lowerCamelCase__ : Optional[Any] = timestep
if not torch.is_tensor(UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = torch.tensor([timesteps] , dtype=torch.long , device=sample.device )
elif torch.is_tensor(UpperCAmelCase ) and len(timesteps.shape ) == 0:
lowerCamelCase__ : List[str] = timesteps[None].to(sample.device )
lowerCamelCase__ : Optional[int] = self.time_proj(UpperCAmelCase )
if self.config.use_timestep_embedding:
lowerCamelCase__ : str = self.time_mlp(UpperCAmelCase )
else:
lowerCamelCase__ : List[str] = timestep_embed[..., None]
lowerCamelCase__ : str = timestep_embed.repeat([1, 1, sample.shape[2]] ).to(sample.dtype )
lowerCamelCase__ : str = timestep_embed.broadcast_to((sample.shape[:1] + timestep_embed.shape[1:]) )
# 2. down
lowerCamelCase__ : str = ()
for downsample_block in self.down_blocks:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = downsample_block(hidden_states=UpperCAmelCase , temb=UpperCAmelCase )
down_block_res_samples += res_samples
# 3. mid
if self.mid_block:
lowerCamelCase__ : Optional[Any] = self.mid_block(UpperCAmelCase , UpperCAmelCase )
# 4. up
for i, upsample_block in enumerate(self.up_blocks ):
lowerCamelCase__ : Dict = down_block_res_samples[-1:]
lowerCamelCase__ : Optional[Any] = down_block_res_samples[:-1]
lowerCamelCase__ : Any = upsample_block(UpperCAmelCase , res_hidden_states_tuple=UpperCAmelCase , temb=UpperCAmelCase )
# 5. post-process
if self.out_block:
lowerCamelCase__ : Any = self.out_block(UpperCAmelCase , UpperCAmelCase )
if not return_dict:
return (sample,)
return UNetaDOutput(sample=UpperCAmelCase )
| 50
| 1
|
import pytest
from datasets.parallel import ParallelBackendConfig, parallel_backend
from datasets.utils.py_utils import map_nested
from .utils import require_dill_gt_0_3_2, require_joblibspark, require_not_windows
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Union[str, Any]: # picklable for multiprocessing
return i + 1
@require_dill_gt_0_3_2
@require_joblibspark
@require_not_windows
def SCREAMING_SNAKE_CASE ( ) -> int:
with parallel_backend('spark' ):
assert ParallelBackendConfig.backend_name == "spark"
lowerCamelCase__ : Union[str, Any] = [1, 2, 3]
with pytest.raises(_UpperCAmelCase ):
with parallel_backend('unsupported backend' ):
map_nested(_UpperCAmelCase , _UpperCAmelCase , num_proc=2 )
with pytest.raises(_UpperCAmelCase ):
with parallel_backend('unsupported backend' ):
map_nested(_UpperCAmelCase , _UpperCAmelCase , num_proc=-1 )
@require_dill_gt_0_3_2
@require_joblibspark
@require_not_windows
@pytest.mark.parametrize('num_proc' , [2, -1] )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Dict:
lowerCamelCase__ : Union[str, Any] = [1, 2]
lowerCamelCase__ : str = {'a': 1, 'b': 2}
lowerCamelCase__ : Optional[int] = {'a': [1, 2], 'b': [3, 4]}
lowerCamelCase__ : Tuple = {'a': {'1': 1}, 'b': 2}
lowerCamelCase__ : Dict = {'a': 1, 'b': 2, 'c': 3, 'd': 4}
lowerCamelCase__ : str = [2, 3]
lowerCamelCase__ : Tuple = {'a': 2, 'b': 3}
lowerCamelCase__ : Optional[Any] = {'a': [2, 3], 'b': [4, 5]}
lowerCamelCase__ : List[Any] = {'a': {'1': 2}, 'b': 3}
lowerCamelCase__ : List[Any] = {'a': 2, 'b': 3, 'c': 4, 'd': 5}
with parallel_backend('spark' ):
assert map_nested(_UpperCAmelCase , _UpperCAmelCase , num_proc=_UpperCAmelCase ) == expected_map_nested_sa
assert map_nested(_UpperCAmelCase , _UpperCAmelCase , num_proc=_UpperCAmelCase ) == expected_map_nested_sa
assert map_nested(_UpperCAmelCase , _UpperCAmelCase , num_proc=_UpperCAmelCase ) == expected_map_nested_sa
assert map_nested(_UpperCAmelCase , _UpperCAmelCase , num_proc=_UpperCAmelCase ) == expected_map_nested_sa
assert map_nested(_UpperCAmelCase , _UpperCAmelCase , num_proc=_UpperCAmelCase ) == expected_map_nested_sa
| 50
|
from __future__ import annotations
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> list[tuple[int, int]]:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = position
lowerCamelCase__ : Optional[Any] = [
(y + 1, x + 2),
(y - 1, x + 2),
(y + 1, x - 2),
(y - 1, x - 2),
(y + 2, x + 1),
(y + 2, x - 1),
(y - 2, x + 1),
(y - 2, x - 1),
]
lowerCamelCase__ : Dict = []
for position in positions:
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = position
if 0 <= y_test < n and 0 <= x_test < n:
permissible_positions.append(_UpperCAmelCase )
return permissible_positions
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
return not any(elem == 0 for row in board for elem in row )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> bool:
if is_complete(_UpperCAmelCase ):
return True
for position in get_valid_pos(_UpperCAmelCase , len(_UpperCAmelCase ) ):
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = position
if board[y][x] == 0:
lowerCamelCase__ : List[Any] = curr + 1
if open_knight_tour_helper(_UpperCAmelCase , _UpperCAmelCase , curr + 1 ):
return True
lowerCamelCase__ : Optional[Any] = 0
return False
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list[list[int]]:
lowerCamelCase__ : Any = [[0 for i in range(_UpperCAmelCase )] for j in range(_UpperCAmelCase )]
for i in range(_UpperCAmelCase ):
for j in range(_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = 1
if open_knight_tour_helper(_UpperCAmelCase , (i, j) , 1 ):
return board
lowerCamelCase__ : Optional[Any] = 0
lowerCamelCase__ : Any = F"""Open Kight Tour cannot be performed on a board of size {n}"""
raise ValueError(_UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_UpperCAmelCase : Any = logging.get_logger(__name__)
_UpperCAmelCase : str = {
"""facebook/s2t-small-librispeech-asr""": (
"""https://huggingface.co/facebook/s2t-small-librispeech-asr/resolve/main/config.json"""
),
# See all Speech2Text models at https://huggingface.co/models?filter=speech_to_text
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """speech_to_text"""
UpperCAmelCase__ = ["""past_key_values"""]
UpperCAmelCase__ = {"""num_attention_heads""": """encoder_attention_heads""", """hidden_size""": """d_model"""}
def __init__( self : List[Any] , UpperCAmelCase : str=10000 , UpperCAmelCase : int=12 , UpperCAmelCase : Optional[int]=2048 , UpperCAmelCase : Optional[int]=4 , UpperCAmelCase : List[Any]=6 , UpperCAmelCase : Dict=2048 , UpperCAmelCase : List[str]=4 , UpperCAmelCase : Optional[int]=0.0 , UpperCAmelCase : Optional[int]=0.0 , UpperCAmelCase : Tuple=True , UpperCAmelCase : Dict=True , UpperCAmelCase : List[str]="relu" , UpperCAmelCase : Any=256 , UpperCAmelCase : int=0.1 , UpperCAmelCase : Tuple=0.0 , UpperCAmelCase : List[str]=0.0 , UpperCAmelCase : Tuple=0.0_2 , UpperCAmelCase : Any=2 , UpperCAmelCase : Optional[Any]=True , UpperCAmelCase : int=1 , UpperCAmelCase : int=0 , UpperCAmelCase : int=2 , UpperCAmelCase : int=6000 , UpperCAmelCase : Optional[Any]=1024 , UpperCAmelCase : Any=2 , UpperCAmelCase : int=(5, 5) , UpperCAmelCase : int=1024 , UpperCAmelCase : Optional[int]=80 , UpperCAmelCase : Tuple=1 , **UpperCAmelCase : Dict , ) -> Any:
lowerCamelCase__ : Optional[int] = vocab_size
lowerCamelCase__ : List[str] = d_model
lowerCamelCase__ : int = encoder_ffn_dim
lowerCamelCase__ : Optional[Any] = encoder_layers
lowerCamelCase__ : Tuple = encoder_attention_heads
lowerCamelCase__ : str = decoder_ffn_dim
lowerCamelCase__ : Optional[Any] = decoder_layers
lowerCamelCase__ : Dict = decoder_attention_heads
lowerCamelCase__ : int = dropout
lowerCamelCase__ : str = attention_dropout
lowerCamelCase__ : Dict = activation_dropout
lowerCamelCase__ : int = activation_function
lowerCamelCase__ : Tuple = init_std
lowerCamelCase__ : Optional[int] = encoder_layerdrop
lowerCamelCase__ : Dict = decoder_layerdrop
lowerCamelCase__ : Union[str, Any] = use_cache
lowerCamelCase__ : Optional[Any] = encoder_layers
lowerCamelCase__ : Any = scale_embedding # scale factor will be sqrt(d_model) if True
lowerCamelCase__ : Optional[int] = max_source_positions
lowerCamelCase__ : Optional[int] = max_target_positions
lowerCamelCase__ : int = num_conv_layers
lowerCamelCase__ : Union[str, Any] = list(UpperCAmelCase )
lowerCamelCase__ : int = conv_channels
lowerCamelCase__ : Dict = input_feat_per_channel
lowerCamelCase__ : int = input_channels
if len(self.conv_kernel_sizes ) != self.num_conv_layers:
raise ValueError(
'Configuration for convolutional module is incorrect. '
'It is required that `len(config.conv_kernel_sizes)` == `config.num_conv_layers` '
F"""but is `len(config.conv_kernel_sizes) = {len(self.conv_kernel_sizes )}`, """
F"""`config.num_conv_layers = {self.num_conv_layers}`.""" )
super().__init__(
pad_token_id=UpperCAmelCase , bos_token_id=UpperCAmelCase , eos_token_id=UpperCAmelCase , is_encoder_decoder=UpperCAmelCase , decoder_start_token_id=UpperCAmelCase , **UpperCAmelCase , )
| 50
|
import argparse
import json
from collections import OrderedDict
import torch
from huggingface_hub import cached_download, hf_hub_url
from transformers import AutoImageProcessor, CvtConfig, CvtForImageClassification
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> int:
lowerCamelCase__ : Optional[int] = []
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.weight""",
F"""stage{idx}.patch_embed.proj.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.bias""",
F"""stage{idx}.patch_embed.proj.bias""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.weight""",
F"""stage{idx}.patch_embed.norm.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.bias""",
F"""stage{idx}.patch_embed.norm.bias""",
) )
return embed
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Tuple = []
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj.bias""",
) )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.weight""", F"""stage{idx}.blocks.{cnt}.norm1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.bias""", F"""stage{idx}.blocks.{cnt}.norm1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.weight""", F"""stage{idx}.blocks.{cnt}.norm2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.bias""", F"""stage{idx}.blocks.{cnt}.norm2.bias""") )
return attention_weights
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = []
token.append((F"""cvt.encoder.stages.{idx}.cls_token""", 'stage2.cls_token') )
return token
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : str = []
head.append(('layernorm.weight', 'norm.weight') )
head.append(('layernorm.bias', 'norm.bias') )
head.append(('classifier.weight', 'head.weight') )
head.append(('classifier.bias', 'head.bias') )
return head
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
lowerCamelCase__ : Tuple = 'imagenet-1k-id2label.json'
lowerCamelCase__ : Union[str, Any] = 1000
lowerCamelCase__ : Optional[Any] = 'huggingface/label-files'
lowerCamelCase__ : Any = num_labels
lowerCamelCase__ : Dict = json.load(open(cached_download(hf_hub_url(_UpperCAmelCase , _UpperCAmelCase , repo_type='dataset' ) ) , 'r' ) )
lowerCamelCase__ : int = {int(_UpperCAmelCase ): v for k, v in idalabel.items()}
lowerCamelCase__ : Tuple = idalabel
lowerCamelCase__ : List[Any] = {v: k for k, v in idalabel.items()}
lowerCamelCase__ : List[str] = CvtConfig(num_labels=_UpperCAmelCase , idalabel=_UpperCAmelCase , labelaid=_UpperCAmelCase )
# For depth size 13 (13 = 1+2+10)
if cvt_model.rsplit('/' , 1 )[-1][4:6] == "13":
lowerCamelCase__ : List[Any] = [1, 2, 10]
# For depth size 21 (21 = 1+4+16)
elif cvt_model.rsplit('/' , 1 )[-1][4:6] == "21":
lowerCamelCase__ : Dict = [1, 4, 16]
# For wide cvt (similar to wide-resnet) depth size 24 (w24 = 2 + 2 20)
else:
lowerCamelCase__ : Optional[Any] = [2, 2, 20]
lowerCamelCase__ : Optional[int] = [3, 12, 16]
lowerCamelCase__ : str = [192, 768, 1024]
lowerCamelCase__ : Any = CvtForImageClassification(_UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = AutoImageProcessor.from_pretrained('facebook/convnext-base-224-22k-1k' )
lowerCamelCase__ : Tuple = image_size
lowerCamelCase__ : List[str] = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
lowerCamelCase__ : Optional[int] = OrderedDict()
lowerCamelCase__ : Tuple = []
for idx in range(len(config.depth ) ):
if config.cls_token[idx]:
lowerCamelCase__ : Optional[Any] = list_of_state_dict + cls_token(_UpperCAmelCase )
lowerCamelCase__ : str = list_of_state_dict + embeddings(_UpperCAmelCase )
for cnt in range(config.depth[idx] ):
lowerCamelCase__ : str = list_of_state_dict + attention(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : int = list_of_state_dict + final()
for gg in list_of_state_dict:
print(_UpperCAmelCase )
for i in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : str = original_weights[list_of_state_dict[i][1]]
model.load_state_dict(_UpperCAmelCase )
model.save_pretrained(_UpperCAmelCase )
image_processor.save_pretrained(_UpperCAmelCase )
# Download the weights from zoo: https://1drv.ms/u/s!AhIXJn_J-blW9RzF3rMW7SsLHa8h?e=blQ0Al
if __name__ == "__main__":
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
parser.add_argument(
"""--cvt_model""",
default="""cvt-w24""",
type=str,
help="""Name of the cvt model you'd like to convert.""",
)
parser.add_argument(
"""--image_size""",
default=3_84,
type=int,
help="""Input Image Size""",
)
parser.add_argument(
"""--cvt_file_name""",
default=R"""cvtmodels\CvT-w24-384x384-IN-22k.pth""",
type=str,
help="""Input Image Size""",
)
parser.add_argument(
"""--pytorch_dump_folder_path""", default=None, type=str, help="""Path to the output PyTorch model directory."""
)
_UpperCAmelCase : List[str] = parser.parse_args()
convert_cvt_checkpoint(args.cvt_model, args.image_size, args.cvt_file_name, args.pytorch_dump_folder_path)
| 50
| 1
|
import itertools
import random
import unittest
import numpy as np
from transformers import is_speech_available
from transformers.testing_utils import require_torch, require_torchaudio
from ...test_sequence_feature_extraction_common import SequenceFeatureExtractionTestMixin
if is_speech_available():
from transformers import SpeechaTextFeatureExtractor
_UpperCAmelCase : Union[str, Any] = random.Random()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=1.0 , _UpperCAmelCase=None , _UpperCAmelCase=None ) -> str:
if rng is None:
lowerCamelCase__ : Optional[Any] = global_rng
lowerCamelCase__ : Tuple = []
for batch_idx in range(shape[0] ):
values.append([] )
for _ in range(shape[1] ):
values[-1].append(rng.random() * scale )
return values
@require_torch
@require_torchaudio
class lowerCAmelCase ( unittest.TestCase ):
def __init__( self : int , UpperCAmelCase : Any , UpperCAmelCase : Union[str, Any]=7 , UpperCAmelCase : Dict=400 , UpperCAmelCase : Tuple=2000 , UpperCAmelCase : Optional[Any]=24 , UpperCAmelCase : List[Any]=24 , UpperCAmelCase : Optional[Any]=0.0 , UpperCAmelCase : Union[str, Any]=16000 , UpperCAmelCase : Union[str, Any]=True , UpperCAmelCase : Dict=True , ) -> int:
lowerCamelCase__ : Union[str, Any] = parent
lowerCamelCase__ : Tuple = batch_size
lowerCamelCase__ : Any = min_seq_length
lowerCamelCase__ : List[str] = max_seq_length
lowerCamelCase__ : Dict = (self.max_seq_length - self.min_seq_length) // (self.batch_size - 1)
lowerCamelCase__ : Optional[Any] = feature_size
lowerCamelCase__ : Dict = num_mel_bins
lowerCamelCase__ : Union[str, Any] = padding_value
lowerCamelCase__ : str = sampling_rate
lowerCamelCase__ : Tuple = return_attention_mask
lowerCamelCase__ : Tuple = do_normalize
def A_ ( self : List[str] ) -> List[Any]:
return {
"feature_size": self.feature_size,
"num_mel_bins": self.num_mel_bins,
"padding_value": self.padding_value,
"sampling_rate": self.sampling_rate,
"return_attention_mask": self.return_attention_mask,
"do_normalize": self.do_normalize,
}
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[int]=False , UpperCAmelCase : List[Any]=False ) -> str:
def _flatten(UpperCAmelCase : Tuple ):
return list(itertools.chain(*UpperCAmelCase ) )
if equal_length:
lowerCamelCase__ : List[str] = [floats_list((self.max_seq_length, self.feature_size) ) for _ in range(self.batch_size )]
else:
# make sure that inputs increase in size
lowerCamelCase__ : List[Any] = [
floats_list((x, self.feature_size) )
for x in range(self.min_seq_length , self.max_seq_length , self.seq_length_diff )
]
if numpify:
lowerCamelCase__ : Union[str, Any] = [np.asarray(UpperCAmelCase ) for x in speech_inputs]
return speech_inputs
@require_torch
@require_torchaudio
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = SpeechaTextFeatureExtractor if is_speech_available() else None
def A_ ( self : Tuple ) -> str:
lowerCamelCase__ : Any = SpeechaTextFeatureExtractionTester(self )
def A_ ( self : int , UpperCAmelCase : Optional[Any] ) -> List[str]:
self.assertTrue(np.all(np.mean(UpperCAmelCase , axis=0 ) < 1e-3 ) )
self.assertTrue(np.all(np.abs(np.var(UpperCAmelCase , axis=0 ) - 1 ) < 1e-3 ) )
def A_ ( self : List[str] ) -> Dict:
# Tests that all call wrap to encode_plus and batch_encode_plus
lowerCamelCase__ : List[Any] = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
# create three inputs of length 800, 1000, and 1200
lowerCamelCase__ : str = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
lowerCamelCase__ : Any = [np.asarray(UpperCAmelCase ) for speech_input in speech_inputs]
# Test feature size
lowerCamelCase__ : Optional[int] = feature_extractor(UpperCAmelCase , padding=UpperCAmelCase , return_tensors='np' ).input_features
self.assertTrue(input_features.ndim == 3 )
self.assertTrue(input_features.shape[-1] == feature_extractor.feature_size )
# Test not batched input
lowerCamelCase__ : Optional[Any] = feature_extractor(speech_inputs[0] , return_tensors='np' ).input_features
lowerCamelCase__ : List[Any] = feature_extractor(np_speech_inputs[0] , return_tensors='np' ).input_features
self.assertTrue(np.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-3 ) )
# Test batched
lowerCamelCase__ : List[Any] = feature_extractor(UpperCAmelCase , return_tensors='np' ).input_features
lowerCamelCase__ : Union[str, Any] = feature_extractor(UpperCAmelCase , return_tensors='np' ).input_features
for enc_seq_a, enc_seq_a in zip(UpperCAmelCase , UpperCAmelCase ):
self.assertTrue(np.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-3 ) )
# Test 2-D numpy arrays are batched.
lowerCamelCase__ : Union[str, Any] = [floats_list((1, x) )[0] for x in (800, 800, 800)]
lowerCamelCase__ : str = np.asarray(UpperCAmelCase )
lowerCamelCase__ : List[str] = feature_extractor(UpperCAmelCase , return_tensors='np' ).input_features
lowerCamelCase__ : Tuple = feature_extractor(UpperCAmelCase , return_tensors='np' ).input_features
for enc_seq_a, enc_seq_a in zip(UpperCAmelCase , UpperCAmelCase ):
self.assertTrue(np.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-3 ) )
def A_ ( self : Any ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
lowerCamelCase__ : str = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
lowerCamelCase__ : str = ['longest', 'max_length', 'do_not_pad']
lowerCamelCase__ : Tuple = [None, 16, None]
for max_length, padding in zip(UpperCAmelCase , UpperCAmelCase ):
lowerCamelCase__ : Any = feature_extractor(
UpperCAmelCase , padding=UpperCAmelCase , max_length=UpperCAmelCase , return_attention_mask=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = inputs.input_features
lowerCamelCase__ : List[str] = inputs.attention_mask
lowerCamelCase__ : Union[str, Any] = [np.sum(UpperCAmelCase ) for x in attention_mask]
self._check_zero_mean_unit_variance(input_features[0][: fbank_feat_lengths[0]] )
self._check_zero_mean_unit_variance(input_features[1][: fbank_feat_lengths[1]] )
self._check_zero_mean_unit_variance(input_features[2][: fbank_feat_lengths[2]] )
def A_ ( self : int ) -> Any:
lowerCamelCase__ : Optional[Any] = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
lowerCamelCase__ : Optional[Any] = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
lowerCamelCase__ : Optional[Any] = ['longest', 'max_length', 'do_not_pad']
lowerCamelCase__ : str = [None, 16, None]
for max_length, padding in zip(UpperCAmelCase , UpperCAmelCase ):
lowerCamelCase__ : Any = feature_extractor(
UpperCAmelCase , max_length=UpperCAmelCase , padding=UpperCAmelCase , return_tensors='np' , return_attention_mask=UpperCAmelCase )
lowerCamelCase__ : Optional[int] = inputs.input_features
lowerCamelCase__ : List[Any] = inputs.attention_mask
lowerCamelCase__ : Any = [np.sum(UpperCAmelCase ) for x in attention_mask]
self._check_zero_mean_unit_variance(input_features[0][: fbank_feat_lengths[0]] )
self.assertTrue(input_features[0][fbank_feat_lengths[0] :].sum() < 1e-6 )
self._check_zero_mean_unit_variance(input_features[1][: fbank_feat_lengths[1]] )
self.assertTrue(input_features[0][fbank_feat_lengths[1] :].sum() < 1e-6 )
self._check_zero_mean_unit_variance(input_features[2][: fbank_feat_lengths[2]] )
def A_ ( self : int ) -> Tuple:
lowerCamelCase__ : str = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
lowerCamelCase__ : Union[str, Any] = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
lowerCamelCase__ : Dict = feature_extractor(
UpperCAmelCase , padding='max_length' , max_length=4 , truncation=UpperCAmelCase , return_tensors='np' , return_attention_mask=UpperCAmelCase , )
lowerCamelCase__ : Optional[Any] = inputs.input_features
lowerCamelCase__ : List[str] = inputs.attention_mask
lowerCamelCase__ : Union[str, Any] = np.sum(attention_mask == 1 , axis=1 )
self._check_zero_mean_unit_variance(input_features[0, : fbank_feat_lengths[0]] )
self._check_zero_mean_unit_variance(input_features[1] )
self._check_zero_mean_unit_variance(input_features[2] )
def A_ ( self : Optional[Any] ) -> Optional[Any]:
lowerCamelCase__ : List[Any] = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
lowerCamelCase__ : Union[str, Any] = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
lowerCamelCase__ : List[str] = feature_extractor(
UpperCAmelCase , padding='longest' , max_length=4 , truncation=UpperCAmelCase , return_tensors='np' , return_attention_mask=UpperCAmelCase , )
lowerCamelCase__ : int = inputs.input_features
lowerCamelCase__ : List[str] = inputs.attention_mask
lowerCamelCase__ : List[Any] = np.sum(attention_mask == 1 , axis=1 )
self._check_zero_mean_unit_variance(input_features[0, : fbank_feat_lengths[0]] )
self._check_zero_mean_unit_variance(input_features[1, : fbank_feat_lengths[1]] )
self._check_zero_mean_unit_variance(input_features[2] )
# make sure that if max_length < longest -> then pad to max_length
self.assertEqual(input_features.shape , (3, 4, 24) )
lowerCamelCase__ : int = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
lowerCamelCase__ : str = feature_extractor(
UpperCAmelCase , padding='longest' , max_length=16 , truncation=UpperCAmelCase , return_tensors='np' , return_attention_mask=UpperCAmelCase , )
lowerCamelCase__ : Optional[Any] = inputs.input_features
lowerCamelCase__ : Optional[int] = inputs.attention_mask
lowerCamelCase__ : Any = np.sum(attention_mask == 1 , axis=1 )
self._check_zero_mean_unit_variance(input_features[0, : fbank_feat_lengths[0]] )
self._check_zero_mean_unit_variance(input_features[1, : fbank_feat_lengths[1]] )
self._check_zero_mean_unit_variance(input_features[2] )
# make sure that if max_length < longest -> then pad to max_length
self.assertEqual(input_features.shape , (3, 6, 24) )
def A_ ( self : Optional[Any] ) -> Optional[int]:
import torch
lowerCamelCase__ : Union[str, Any] = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
lowerCamelCase__ : Dict = np.random.rand(100 , 32 ).astype(np.floataa )
lowerCamelCase__ : Optional[int] = np_speech_inputs.tolist()
for inputs in [py_speech_inputs, np_speech_inputs]:
lowerCamelCase__ : Union[str, Any] = feature_extractor.pad([{'input_features': inputs}] , return_tensors='np' )
self.assertTrue(np_processed.input_features.dtype == np.floataa )
lowerCamelCase__ : List[Any] = feature_extractor.pad([{'input_features': inputs}] , return_tensors='pt' )
self.assertTrue(pt_processed.input_features.dtype == torch.floataa )
def A_ ( self : int , UpperCAmelCase : List[str] ) -> Any:
from datasets import load_dataset
lowerCamelCase__ : str = load_dataset('hf-internal-testing/librispeech_asr_dummy' , 'clean' , split='validation' )
# automatic decoding with librispeech
lowerCamelCase__ : str = ds.sort('id' ).select(range(UpperCAmelCase ) )[:num_samples]['audio']
return [x["array"] for x in speech_samples]
def A_ ( self : str ) -> Tuple:
# fmt: off
lowerCamelCase__ : Optional[int] = np.array([
-1.5_7_4_5, -1.7_7_1_3, -1.7_0_2_0, -1.6_0_6_9, -1.2_2_5_0, -1.1_1_0_5, -0.9_0_7_2, -0.8_2_4_1,
-1.2_3_1_0, -0.8_0_9_8, -0.3_3_2_0, -0.4_1_0_1, -0.7_9_8_5, -0.4_9_9_6, -0.8_2_1_3, -0.9_1_2_8,
-1.0_4_2_0, -1.1_2_8_6, -1.0_4_4_0, -0.7_9_9_9, -0.8_4_0_5, -1.2_2_7_5, -1.5_4_4_3, -1.4_6_2_5,
] )
# fmt: on
lowerCamelCase__ : Any = self._load_datasamples(1 )
lowerCamelCase__ : List[str] = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
lowerCamelCase__ : Dict = feature_extractor(UpperCAmelCase , return_tensors='pt' ).input_features
self.assertEquals(input_features.shape , (1, 584, 24) )
self.assertTrue(np.allclose(input_features[0, 0, :30] , UpperCAmelCase , atol=1e-4 ) )
| 50
|
# Copyright 2021 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.
import argparse
import os
from accelerate.test_utils import execute_subprocess_async
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase=None ) -> Tuple:
if subparsers is not None:
lowerCamelCase__ : Any = subparsers.add_parser('test' )
else:
lowerCamelCase__ : int = argparse.ArgumentParser('Accelerate test command' )
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 SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = os.path.sep.join(__file__.split(os.path.sep )[:-2] + ['test_utils', 'scripts', 'test_script.py'] )
if args.config_file is None:
lowerCamelCase__ : List[str] = script_name
else:
lowerCamelCase__ : List[Any] = F"""--config_file={args.config_file} {script_name}"""
lowerCamelCase__ : str = ['accelerate-launch'] + test_args.split()
lowerCamelCase__ : Dict = execute_subprocess_async(_UpperCAmelCase , env=os.environ.copy() )
if result.returncode == 0:
print('Test is a success! You are ready for your distributed training!' )
def SCREAMING_SNAKE_CASE ( ) -> Any:
lowerCamelCase__ : Any = test_command_parser()
lowerCamelCase__ : List[Any] = parser.parse_args()
test_command(_UpperCAmelCase )
if __name__ == "__main__":
main()
| 50
| 1
|
import unittest
from transformers import DebertaVaTokenizer, DebertaVaTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, slow
from ...test_tokenization_common import TokenizerTesterMixin
_UpperCAmelCase : Any = get_tests_dir("""fixtures/spiece.model""")
@require_sentencepiece
@require_tokenizers
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = DebertaVaTokenizer
UpperCAmelCase__ = DebertaVaTokenizerFast
UpperCAmelCase__ = True
UpperCAmelCase__ = True
def A_ ( self : Dict ) -> int:
super().setUp()
# We have a SentencePiece fixture for testing
lowerCamelCase__ : Optional[Any] = DebertaVaTokenizer(UpperCAmelCase , unk_token='<unk>' )
tokenizer.save_pretrained(self.tmpdirname )
def A_ ( self : Dict , UpperCAmelCase : Tuple ) -> int:
lowerCamelCase__ : Dict = 'this is a test'
lowerCamelCase__ : Union[str, Any] = 'this is a test'
return input_text, output_text
def A_ ( self : Union[str, Any] ) -> Optional[Any]:
lowerCamelCase__ : str = '<pad>'
lowerCamelCase__ : int = 0
self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCAmelCase ) , UpperCAmelCase )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : Union[str, Any] ) -> Optional[int]:
lowerCamelCase__ : Tuple = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , '<pad>' )
self.assertEqual(vocab_keys[1] , '<unk>' )
self.assertEqual(vocab_keys[-1] , '[PAD]' )
self.assertEqual(len(UpperCAmelCase ) , 30001 )
def A_ ( self : Optional[Any] ) -> int:
self.assertEqual(self.get_tokenizer().vocab_size , 30000 )
def A_ ( self : str ) -> List[str]:
# fmt: off
lowerCamelCase__ : Union[str, Any] = ' \tHeLLo!how \n Are yoU? '
lowerCamelCase__ : int = ['▁hello', '!', 'how', '▁are', '▁you', '?']
# fmt: on
lowerCamelCase__ : Any = DebertaVaTokenizer(UpperCAmelCase , do_lower_case=UpperCAmelCase )
lowerCamelCase__ : Optional[int] = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Tuple = DebertaVaTokenizerFast(UpperCAmelCase , do_lower_case=UpperCAmelCase )
lowerCamelCase__ : int = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
@unittest.skip('There is an inconsistency between slow and fast tokenizer due to a bug in the fast one.' )
def A_ ( self : Optional[int] ) -> int:
pass
@unittest.skip('There is an inconsistency between slow and fast tokenizer due to a bug in the fast one.' )
def A_ ( self : Dict ) -> List[str]:
pass
def A_ ( self : str ) -> Any:
# fmt: off
lowerCamelCase__ : List[Any] = 'I was born in 92000, and this is falsé.'
lowerCamelCase__ : int = ['▁', '<unk>', '▁was', '▁born', '▁in', '▁9', '2000', '▁', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '▁', '.', ]
# fmt: on
lowerCamelCase__ : Union[str, Any] = DebertaVaTokenizer(UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : List[Any] = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : int = DebertaVaTokenizerFast(UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : Tuple = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : int ) -> Tuple:
# fmt: off
lowerCamelCase__ : str = 'I was born in 92000, and this is falsé.'
lowerCamelCase__ : Dict = ['▁i', '▁was', '▁born', '▁in', '▁9', '2000', '▁', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '▁', '.', ]
# fmt: on
lowerCamelCase__ : str = DebertaVaTokenizer(UpperCAmelCase , do_lower_case=UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = DebertaVaTokenizerFast(UpperCAmelCase , do_lower_case=UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Union[str, Any] ) -> str:
# fmt: off
lowerCamelCase__ : Any = 'I was born in 92000, and this is falsé.'
lowerCamelCase__ : Tuple = ['▁i', '▁was', '▁born', '▁in', '▁9', '2000', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '.', ]
# fmt: on
lowerCamelCase__ : Tuple = DebertaVaTokenizer(UpperCAmelCase , do_lower_case=UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : List[Any] = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Tuple = DebertaVaTokenizerFast(UpperCAmelCase , do_lower_case=UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : List[Any] = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : List[str] ) -> Tuple:
# fmt: off
lowerCamelCase__ : Dict = 'I was born in 92000, and this is falsé.'
lowerCamelCase__ : List[str] = ['▁', '<unk>', '▁was', '▁born', '▁in', '▁9', '2000', '▁', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '▁', '.', ]
# fmt: on
lowerCamelCase__ : List[Any] = DebertaVaTokenizer(UpperCAmelCase , do_lower_case=UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : Tuple = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = DebertaVaTokenizerFast(UpperCAmelCase , do_lower_case=UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : Dict = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Optional[int] ) -> Union[str, Any]:
# fmt: off
lowerCamelCase__ : str = ' \tHeLLo!how \n Are yoU? '
lowerCamelCase__ : Union[str, Any] = ['▁', '<unk>', 'e', '<unk>', 'o', '!', 'how', '▁', '<unk>', 're', '▁yo', '<unk>', '?']
# fmt: on
lowerCamelCase__ : str = DebertaVaTokenizer(UpperCAmelCase , do_lower_case=UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : int = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = DebertaVaTokenizerFast(UpperCAmelCase , do_lower_case=UpperCAmelCase , split_by_punct=UpperCAmelCase )
lowerCamelCase__ : Tuple = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Dict ) -> int:
lowerCamelCase__ : Optional[Any] = self.get_tokenizer()
lowerCamelCase__ : Optional[int] = self.get_rust_tokenizer()
lowerCamelCase__ : List[Any] = 'I was born in 92000, and this is falsé.'
lowerCamelCase__ : List[Any] = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
lowerCamelCase__ : List[str] = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase ) )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Dict = tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Any = self.get_rust_tokenizer()
lowerCamelCase__ : Dict = tokenizer.encode(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = rust_tokenizer.encode(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : List[Any] ) -> int:
lowerCamelCase__ : List[Any] = 'This is a test'
lowerCamelCase__ : Optional[int] = [13, 1, 4398, 25, 21, 1289]
lowerCamelCase__ : Dict = ['▁', 'T', 'his', '▁is', '▁a', '▁test']
lowerCamelCase__ : str = ['▁', '<unk>', 'his', '▁is', '▁a', '▁test']
lowerCamelCase__ : Dict = DebertaVaTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : Optional[int] = DebertaVaTokenizerFast(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : Tuple = tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Tuple = tokenizer.tokenize(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Tuple = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : List[str] = rust_tokenizer.tokenize(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = rust_tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
# fmt: off
lowerCamelCase__ : Optional[int] = 'I was born in 92000, and this is falsé.'
lowerCamelCase__ : Dict = [13, 1, 23, 386, 19, 561, 3050, 15, 17, 48, 25, 8256, 18, 1, 9]
lowerCamelCase__ : List[str] = ['▁', 'I', '▁was', '▁born', '▁in', '▁9', '2000', ',', '▁and', '▁this', '▁is', '▁fal', 's', 'é', '.', ]
lowerCamelCase__ : Dict = ['▁', '<unk>', '▁was', '▁born', '▁in', '▁9', '2000', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '.', ]
# fmt: on
lowerCamelCase__ : Any = tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Dict = tokenizer.tokenize(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Optional[int] = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : List[Any] = rust_tokenizer.encode(UpperCAmelCase , add_special_tokens=UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : List[str] = rust_tokenizer.tokenize(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : List[Any] = rust_tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : int ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = DebertaVaTokenizer(UpperCAmelCase )
lowerCamelCase__ : Dict = tokenizer.encode('sequence builders' )
lowerCamelCase__ : str = tokenizer.encode('multi-sequence build' )
lowerCamelCase__ : Optional[Any] = tokenizer.build_inputs_with_special_tokens(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tokenizer.build_inputs_with_special_tokens(UpperCAmelCase , UpperCAmelCase )
self.assertEqual([tokenizer.cls_token_id] + text + [tokenizer.sep_token_id] , UpperCAmelCase )
self.assertEqual(
[tokenizer.cls_token_id] + text + [tokenizer.sep_token_id] + text_a + [tokenizer.sep_token_id] , UpperCAmelCase , )
@slow
def A_ ( self : Tuple ) -> int:
# fmt: off
lowerCamelCase__ : str = {'input_ids': [[1, 39867, 36, 19390, 486, 27, 35052, 81436, 18, 60685, 1225, 7, 35052, 81436, 18, 9367, 16899, 18, 15937, 53, 594, 773, 18, 16287, 30465, 36, 15937, 6, 41139, 38, 36979, 60763, 191, 6, 34132, 99, 6, 50538, 390, 43230, 6, 34132, 2779, 20850, 14, 699, 1072, 1194, 36, 382, 10901, 53, 7, 699, 1072, 2084, 36, 20422, 630, 53, 19, 105, 3049, 1896, 1053, 16899, 1506, 11, 37978, 4243, 7, 1237, 31869, 200, 16566, 654, 6, 35052, 81436, 7, 55630, 13593, 4, 2], [1, 26, 15011, 13, 667, 8, 1053, 18, 23611, 1237, 72356, 12820, 34, 104134, 1209, 35, 13313, 6627, 21, 202, 347, 7, 164, 2399, 11, 46, 4485, 4, 2, 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, 5, 1232, 2864, 15785, 14951, 105, 5, 8581, 1250, 4, 2, 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]], 'token_type_ids': [[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, 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, 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]], '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, 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, 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=UpperCAmelCase , model_name='microsoft/deberta-v2-xlarge' , revision='ad6e42c1532ddf3a15c39246b63f5559d558b670' , )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 100_0000 ) -> int:
lowerCamelCase__ : int = limit + 1
lowerCamelCase__ : Optional[Any] = [0] * limit
for first_term in range(1 , _UpperCAmelCase ):
for n in range(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Optional[Any] = first_term + n / first_term
if common_difference % 4: # d must be divisble by 4
continue
else:
common_difference /= 4
if (
first_term > common_difference
and first_term < 4 * common_difference
): # since x,y,z are positive integers
frequency[n] += 1 # so z>0 and a>d ,also 4d<a
lowerCamelCase__ : List[str] = sum(1 for x in frequency[1:limit] if x == 10 )
return count
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
from __future__ import annotations
from typing import Any
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> None:
create_state_space_tree(_UpperCAmelCase , [] , 0 )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> None:
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__":
_UpperCAmelCase : list[Any] = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["""A""", """B""", """C"""])
generate_all_subsequences(seq)
| 50
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_base import BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import PaddingStrategy, logging
from .tokenization_realm import RealmTokenizer
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
_UpperCAmelCase : Any = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : int = {
"""vocab_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/vocab.txt"""
),
"""google/realm-orqa-nq-openqa""": """https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-nq-reader""": """https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-openqa""": """https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-reader""": """https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/tokenizer.jsont"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-openqa""": (
"""https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-reader""": (
"""https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-openqa""": (
"""https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-reader""": (
"""https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : int = {
"""google/realm-cc-news-pretrained-embedder""": 5_12,
"""google/realm-cc-news-pretrained-encoder""": 5_12,
"""google/realm-cc-news-pretrained-scorer""": 5_12,
"""google/realm-cc-news-pretrained-openqa""": 5_12,
"""google/realm-orqa-nq-openqa""": 5_12,
"""google/realm-orqa-nq-reader""": 5_12,
"""google/realm-orqa-wq-openqa""": 5_12,
"""google/realm-orqa-wq-reader""": 5_12,
}
_UpperCAmelCase : Any = {
"""google/realm-cc-news-pretrained-embedder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-encoder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-scorer""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-reader""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-reader""": {"""do_lower_case""": True},
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_INIT_CONFIGURATION
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = RealmTokenizer
def __init__( self : Optional[int] , UpperCAmelCase : Tuple=None , UpperCAmelCase : Any=None , UpperCAmelCase : List[Any]=True , UpperCAmelCase : Optional[Any]="[UNK]" , UpperCAmelCase : Any="[SEP]" , UpperCAmelCase : Tuple="[PAD]" , UpperCAmelCase : List[Any]="[CLS]" , UpperCAmelCase : Union[str, Any]="[MASK]" , UpperCAmelCase : Optional[Any]=True , UpperCAmelCase : Any=None , **UpperCAmelCase : Optional[int] , ) -> str:
super().__init__(
UpperCAmelCase , tokenizer_file=UpperCAmelCase , do_lower_case=UpperCAmelCase , unk_token=UpperCAmelCase , sep_token=UpperCAmelCase , pad_token=UpperCAmelCase , cls_token=UpperCAmelCase , mask_token=UpperCAmelCase , tokenize_chinese_chars=UpperCAmelCase , strip_accents=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get('lowercase' , UpperCAmelCase ) != do_lower_case
or normalizer_state.get('strip_accents' , UpperCAmelCase ) != strip_accents
or normalizer_state.get('handle_chinese_chars' , UpperCAmelCase ) != tokenize_chinese_chars
):
lowerCamelCase__ : Optional[int] = getattr(UpperCAmelCase , normalizer_state.pop('type' ) )
lowerCamelCase__ : Optional[Any] = do_lower_case
lowerCamelCase__ : str = strip_accents
lowerCamelCase__ : Optional[Any] = tokenize_chinese_chars
lowerCamelCase__ : int = normalizer_class(**UpperCAmelCase )
lowerCamelCase__ : str = do_lower_case
def A_ ( self : Optional[int] , UpperCAmelCase : int , **UpperCAmelCase : int ) -> List[Any]:
lowerCamelCase__ : List[Any] = PaddingStrategy.MAX_LENGTH
lowerCamelCase__ : Optional[int] = text
lowerCamelCase__ : Dict = kwargs.pop('text_pair' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = kwargs.pop('return_tensors' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = {
'input_ids': [],
'attention_mask': [],
'token_type_ids': [],
}
for idx, candidate_text in enumerate(UpperCAmelCase ):
if batch_text_pair is not None:
lowerCamelCase__ : Tuple = batch_text_pair[idx]
else:
lowerCamelCase__ : Dict = None
lowerCamelCase__ : Optional[int] = super().__call__(UpperCAmelCase , UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Any = encoded_candidates.get('input_ids' )
lowerCamelCase__ : Union[str, Any] = encoded_candidates.get('attention_mask' )
lowerCamelCase__ : Tuple = encoded_candidates.get('token_type_ids' )
if encoded_input_ids is not None:
output_data["input_ids"].append(UpperCAmelCase )
if encoded_attention_mask is not None:
output_data["attention_mask"].append(UpperCAmelCase )
if encoded_token_type_ids is not None:
output_data["token_type_ids"].append(UpperCAmelCase )
lowerCamelCase__ : int = {key: item for key, item in output_data.items() if len(UpperCAmelCase ) != 0}
return BatchEncoding(UpperCAmelCase , tensor_type=UpperCAmelCase )
def A_ ( self : int , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Optional[int]=None ) -> List[str]:
lowerCamelCase__ : Tuple = [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 A_ ( self : Tuple , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : List[Any] = [self.sep_token_id]
lowerCamelCase__ : Optional[int] = [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 A_ ( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
lowerCamelCase__ : int = self._tokenizer.model.save(UpperCAmelCase , name=UpperCAmelCase )
return tuple(UpperCAmelCase )
| 50
| 1
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import logging
from .tokenization_convbert import ConvBertTokenizer
_UpperCAmelCase : Any = logging.get_logger(__name__)
_UpperCAmelCase : Optional[Any] = {"""vocab_file""": """vocab.txt"""}
_UpperCAmelCase : Union[str, Any] = {
"""vocab_file""": {
"""YituTech/conv-bert-base""": """https://huggingface.co/YituTech/conv-bert-base/resolve/main/vocab.txt""",
"""YituTech/conv-bert-medium-small""": (
"""https://huggingface.co/YituTech/conv-bert-medium-small/resolve/main/vocab.txt"""
),
"""YituTech/conv-bert-small""": """https://huggingface.co/YituTech/conv-bert-small/resolve/main/vocab.txt""",
}
}
_UpperCAmelCase : Dict = {
"""YituTech/conv-bert-base""": 5_12,
"""YituTech/conv-bert-medium-small""": 5_12,
"""YituTech/conv-bert-small""": 5_12,
}
_UpperCAmelCase : List[str] = {
"""YituTech/conv-bert-base""": {"""do_lower_case""": True},
"""YituTech/conv-bert-medium-small""": {"""do_lower_case""": True},
"""YituTech/conv-bert-small""": {"""do_lower_case""": True},
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_INIT_CONFIGURATION
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = ConvBertTokenizer
def __init__( self : Optional[Any] , UpperCAmelCase : int=None , UpperCAmelCase : Dict=None , UpperCAmelCase : Optional[Any]=True , UpperCAmelCase : Tuple="[UNK]" , UpperCAmelCase : Optional[int]="[SEP]" , UpperCAmelCase : List[Any]="[PAD]" , UpperCAmelCase : List[Any]="[CLS]" , UpperCAmelCase : Union[str, Any]="[MASK]" , UpperCAmelCase : Any=True , UpperCAmelCase : int=None , **UpperCAmelCase : Union[str, Any] , ) -> List[str]:
super().__init__(
UpperCAmelCase , tokenizer_file=UpperCAmelCase , do_lower_case=UpperCAmelCase , unk_token=UpperCAmelCase , sep_token=UpperCAmelCase , pad_token=UpperCAmelCase , cls_token=UpperCAmelCase , mask_token=UpperCAmelCase , tokenize_chinese_chars=UpperCAmelCase , strip_accents=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : int = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get('lowercase' , UpperCAmelCase ) != do_lower_case
or normalizer_state.get('strip_accents' , UpperCAmelCase ) != strip_accents
or normalizer_state.get('handle_chinese_chars' , UpperCAmelCase ) != tokenize_chinese_chars
):
lowerCamelCase__ : Union[str, Any] = getattr(UpperCAmelCase , normalizer_state.pop('type' ) )
lowerCamelCase__ : Dict = do_lower_case
lowerCamelCase__ : Dict = strip_accents
lowerCamelCase__ : Union[str, Any] = tokenize_chinese_chars
lowerCamelCase__ : Tuple = normalizer_class(**UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = do_lower_case
def A_ ( self : Union[str, Any] , UpperCAmelCase : str , UpperCAmelCase : Any=None ) -> Dict:
lowerCamelCase__ : List[Any] = [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 A_ ( self : Union[str, Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : List[Any] = [self.sep_token_id]
lowerCamelCase__ : List[str] = [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 A_ ( self : List[str] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
lowerCamelCase__ : List[Any] = self._tokenizer.model.save(UpperCAmelCase , name=UpperCAmelCase )
return tuple(UpperCAmelCase )
| 50
|
import logging
import os
from dataclasses import dataclass
from typing import List, Optional, Union
import tqdm
from filelock import FileLock
from transformers import (
BartTokenizer,
BartTokenizerFast,
DataProcessor,
PreTrainedTokenizer,
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
is_tf_available,
is_torch_available,
)
_UpperCAmelCase : Optional[Any] = logging.getLogger(__name__)
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
if is_torch_available():
import torch
from torch.utils.data import Dataset
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
def __init__( self : int , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : List[str]=False , UpperCAmelCase : bool = False , ) -> List[str]:
lowerCamelCase__ : int = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = os.path.join(
UpperCAmelCase , 'cached_{}_{}_{}_{}'.format(
'dev' if evaluate else 'train' , tokenizer.__class__.__name__ , str(UpperCAmelCase ) , UpperCAmelCase , ) , )
lowerCamelCase__ : int = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = label_list[2], label_list[1]
lowerCamelCase__ : List[str] = label_list
# Make sure only the first process in distributed training processes the dataset,
# and the others will use the cache.
lowerCamelCase__ : str = cached_features_file + '.lock'
with FileLock(UpperCAmelCase ):
if os.path.exists(UpperCAmelCase ) and not overwrite_cache:
logger.info(F"""Loading features from cached file {cached_features_file}""" )
lowerCamelCase__ : int = torch.load(UpperCAmelCase )
else:
logger.info(F"""Creating features from dataset file at {data_dir}""" )
lowerCamelCase__ : str = (
processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
)
logger.info('Training examples: %s' , len(UpperCAmelCase ) )
lowerCamelCase__ : Dict = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
logger.info('Saving features into cached file %s' , UpperCAmelCase )
torch.save(self.features , UpperCAmelCase )
def __len__( self : Optional[int] ) -> Optional[Any]:
return len(self.features )
def __getitem__( self : Tuple , UpperCAmelCase : Dict ) -> InputFeatures:
return self.features[i]
def A_ ( self : int ) -> int:
return self.label_list
if is_tf_available():
import tensorflow as tf
class lowerCAmelCase :
UpperCAmelCase__ = 42
def __init__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = 128 , UpperCAmelCase : Any=False , UpperCAmelCase : bool = False , ) -> Union[str, Any]:
lowerCamelCase__ : Any = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : str = label_list[2], label_list[1]
lowerCamelCase__ : Optional[int] = label_list
lowerCamelCase__ : int = processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def gen():
for ex_index, ex in tqdm.tqdm(enumerate(self.features ) , desc='convert examples to features' ):
if ex_index % 10000 == 0:
logger.info('Writing example %d of %d' % (ex_index, len(UpperCAmelCase )) )
yield (
{
"example_id": 0,
"input_ids": ex.input_ids,
"attention_mask": ex.attention_mask,
"token_type_ids": ex.token_type_ids,
},
ex.label,
)
lowerCamelCase__ : Optional[int] = tf.data.Dataset.from_generator(
UpperCAmelCase , (
{
'example_id': tf.intaa,
'input_ids': tf.intaa,
'attention_mask': tf.intaa,
'token_type_ids': tf.intaa,
},
tf.intaa,
) , (
{
'example_id': tf.TensorShape([] ),
'input_ids': tf.TensorShape([None, None] ),
'attention_mask': tf.TensorShape([None, None] ),
'token_type_ids': tf.TensorShape([None, None] ),
},
tf.TensorShape([] ),
) , )
def A_ ( self : Any ) -> Any:
return self.dataset
def __len__( self : Tuple ) -> int:
return len(self.features )
def __getitem__( self : List[str] , UpperCAmelCase : Any ) -> InputFeatures:
return self.features[i]
def A_ ( self : Dict ) -> str:
return self.label_list
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : int , UpperCAmelCase : List[Any] ) -> int:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_train_set.txt' ) ) , 'train' )
def A_ ( self : Any , UpperCAmelCase : int ) -> List[Any]:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_evaluation_set.txt' ) ) , 'dev' )
def A_ ( self : Any ) -> List[Any]:
return ["contradiction", "entailment", "neutral"]
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : List[str] ) -> List[str]:
lowerCamelCase__ : List[str] = []
for i, line in enumerate(UpperCAmelCase ):
if i == 0:
continue
lowerCamelCase__ : Tuple = '%s-%s' % (set_type, line[0])
lowerCamelCase__ : str = line[5]
lowerCamelCase__ : Dict = line[6]
lowerCamelCase__ : int = line[7][2:] if line[7].startswith('ex' ) else line[7]
lowerCamelCase__ : Dict = line[0]
examples.append(InputExample(guid=UpperCAmelCase , text_a=UpperCAmelCase , text_b=UpperCAmelCase , label=UpperCAmelCase , pairID=UpperCAmelCase ) )
return examples
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ) -> Optional[int]:
lowerCamelCase__ : int = {label: i for i, label in enumerate(_UpperCAmelCase )}
lowerCamelCase__ : List[Any] = []
for ex_index, example in tqdm.tqdm(enumerate(_UpperCAmelCase ) , desc='convert examples to features' ):
if ex_index % 1_0000 == 0:
logger.info('Writing example %d' % (ex_index) )
lowerCamelCase__ : List[Any] = tokenizer(
example.text_a , example.text_b , add_special_tokens=_UpperCAmelCase , max_length=_UpperCAmelCase , padding='max_length' , truncation=_UpperCAmelCase , return_overflowing_tokens=_UpperCAmelCase , )
lowerCamelCase__ : List[str] = label_map[example.label] if example.label in label_map else 0
lowerCamelCase__ : Optional[int] = int(example.pairID )
features.append(InputFeatures(**_UpperCAmelCase , label=_UpperCAmelCase , pairID=_UpperCAmelCase ) )
for i, example in enumerate(examples[:5] ):
logger.info('*** Example ***' )
logger.info(F"""guid: {example}""" )
logger.info(F"""features: {features[i]}""" )
return features
_UpperCAmelCase : str = {
"""hans""": 3,
}
_UpperCAmelCase : List[Any] = {
"""hans""": HansProcessor,
}
| 50
| 1
|
import argparse
import json
from collections import OrderedDict
from functools import partial
from pathlib import Path
import timm
import torch
from huggingface_hub import hf_hub_download
from transformers import LevitConfig, LevitForImageClassificationWithTeacher, LevitImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
_UpperCAmelCase : Tuple = logging.get_logger()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase = True ) -> int:
print(F"""Converting {name}...""" )
with torch.no_grad():
if hidden_sizes == 128:
if name[-1] == "S":
lowerCamelCase__ : Union[str, Any] = timm.create_model('levit_128s' , pretrained=_UpperCAmelCase )
else:
lowerCamelCase__ : int = timm.create_model('levit_128' , pretrained=_UpperCAmelCase )
if hidden_sizes == 192:
lowerCamelCase__ : List[str] = timm.create_model('levit_192' , pretrained=_UpperCAmelCase )
if hidden_sizes == 256:
lowerCamelCase__ : List[Any] = timm.create_model('levit_256' , pretrained=_UpperCAmelCase )
if hidden_sizes == 384:
lowerCamelCase__ : int = timm.create_model('levit_384' , pretrained=_UpperCAmelCase )
from_model.eval()
lowerCamelCase__ : Optional[Any] = LevitForImageClassificationWithTeacher(_UpperCAmelCase ).eval()
lowerCamelCase__ : List[str] = OrderedDict()
lowerCamelCase__ : str = from_model.state_dict()
lowerCamelCase__ : Optional[Any] = list(from_model.state_dict().keys() )
lowerCamelCase__ : str = list(our_model.state_dict().keys() )
print(len(_UpperCAmelCase ) , len(_UpperCAmelCase ) )
for i in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : List[Any] = weights[og_keys[i]]
our_model.load_state_dict(_UpperCAmelCase )
lowerCamelCase__ : int = torch.randn((2, 3, 224, 224) )
lowerCamelCase__ : Optional[int] = from_model(_UpperCAmelCase )
lowerCamelCase__ : Any = our_model(_UpperCAmelCase ).logits
assert torch.allclose(_UpperCAmelCase , _UpperCAmelCase ), "The model logits don't match the original one."
lowerCamelCase__ : Optional[Any] = name
print(_UpperCAmelCase )
if push_to_hub:
our_model.save_pretrained(save_directory / checkpoint_name )
lowerCamelCase__ : int = LevitImageProcessor()
image_processor.save_pretrained(save_directory / checkpoint_name )
print(F"""Pushed {checkpoint_name}""" )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase = None , _UpperCAmelCase = True ) -> List[str]:
lowerCamelCase__ : Optional[int] = 'imagenet-1k-id2label.json'
lowerCamelCase__ : Tuple = 1000
lowerCamelCase__ : Union[str, Any] = (1, num_labels)
lowerCamelCase__ : List[str] = 'huggingface/label-files'
lowerCamelCase__ : int = num_labels
lowerCamelCase__ : Tuple = json.load(open(hf_hub_download(_UpperCAmelCase , _UpperCAmelCase , repo_type='dataset' ) , 'r' ) )
lowerCamelCase__ : List[Any] = {int(_UpperCAmelCase ): v for k, v in idalabel.items()}
lowerCamelCase__ : List[Any] = idalabel
lowerCamelCase__ : Union[str, Any] = {v: k for k, v in idalabel.items()}
lowerCamelCase__ : Union[str, Any] = partial(_UpperCAmelCase , num_labels=_UpperCAmelCase , idalabel=_UpperCAmelCase , labelaid=_UpperCAmelCase )
lowerCamelCase__ : int = {
'levit-128S': 128,
'levit-128': 128,
'levit-192': 192,
'levit-256': 256,
'levit-384': 384,
}
lowerCamelCase__ : int = {
'levit-128S': ImageNetPreTrainedConfig(
hidden_sizes=[128, 256, 384] , num_attention_heads=[4, 6, 8] , depths=[2, 3, 4] , key_dim=[16, 16, 16] , drop_path_rate=0 , ),
'levit-128': ImageNetPreTrainedConfig(
hidden_sizes=[128, 256, 384] , num_attention_heads=[4, 8, 12] , depths=[4, 4, 4] , key_dim=[16, 16, 16] , drop_path_rate=0 , ),
'levit-192': ImageNetPreTrainedConfig(
hidden_sizes=[192, 288, 384] , num_attention_heads=[3, 5, 6] , depths=[4, 4, 4] , key_dim=[32, 32, 32] , drop_path_rate=0 , ),
'levit-256': ImageNetPreTrainedConfig(
hidden_sizes=[256, 384, 512] , num_attention_heads=[4, 6, 8] , depths=[4, 4, 4] , key_dim=[32, 32, 32] , drop_path_rate=0 , ),
'levit-384': ImageNetPreTrainedConfig(
hidden_sizes=[384, 512, 768] , num_attention_heads=[6, 9, 12] , depths=[4, 4, 4] , key_dim=[32, 32, 32] , drop_path_rate=0.1 , ),
}
if model_name:
convert_weight_and_push(
names_to_hidden_sizes[model_name] , _UpperCAmelCase , names_to_config[model_name] , _UpperCAmelCase , _UpperCAmelCase )
else:
for model_name, config in names_to_config.items():
convert_weight_and_push(names_to_hidden_sizes[model_name] , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
return config, expected_shape
if __name__ == "__main__":
_UpperCAmelCase : Union[str, Any] = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--model_name""",
default=None,
type=str,
help="""The name of the model you wish to convert, it must be one of the supported Levit* architecture,""",
)
parser.add_argument(
"""--pytorch_dump_folder_path""",
default="""levit-dump-folder/""",
type=Path,
required=False,
help="""Path to the output PyTorch model directory.""",
)
parser.add_argument("""--push_to_hub""", action="""store_true""", help="""Push model and image processor to the hub""")
parser.add_argument(
"""--no-push_to_hub""",
dest="""push_to_hub""",
action="""store_false""",
help="""Do not push model and image processor to the hub""",
)
_UpperCAmelCase : Optional[Any] = parser.parse_args()
_UpperCAmelCase : Path = args.pytorch_dump_folder_path
pytorch_dump_folder_path.mkdir(exist_ok=True, parents=True)
convert_weights_and_push(pytorch_dump_folder_path, args.model_name, args.push_to_hub)
| 50
|
import unittest
from transformers import BertGenerationTokenizer
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
_UpperCAmelCase : Optional[Any] = """▁"""
_UpperCAmelCase : Optional[Any] = get_tests_dir("""fixtures/test_sentencepiece.model""")
@require_sentencepiece
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = BertGenerationTokenizer
UpperCAmelCase__ = False
UpperCAmelCase__ = True
def A_ ( self : List[Any] ) -> List[str]:
super().setUp()
lowerCamelCase__ : Dict = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
tokenizer.save_pretrained(self.tmpdirname )
def A_ ( self : Optional[Any] ) -> Dict:
lowerCamelCase__ : List[str] = '<s>'
lowerCamelCase__ : Optional[int] = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCAmelCase ) , UpperCAmelCase )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : List[str] ) -> Optional[int]:
lowerCamelCase__ : Optional[Any] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , '<unk>' )
self.assertEqual(vocab_keys[1] , '<s>' )
self.assertEqual(vocab_keys[-1] , '<pad>' )
self.assertEqual(len(UpperCAmelCase ) , 1002 )
def A_ ( self : List[Any] ) -> Optional[Any]:
self.assertEqual(self.get_tokenizer().vocab_size , 1000 )
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : List[str] = tokenizer.tokenize('This is a test' )
self.assertListEqual(UpperCAmelCase , ['▁This', '▁is', '▁a', '▁t', 'est'] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(UpperCAmelCase ) , [285, 46, 10, 170, 382] , )
lowerCamelCase__ : List[str] = tokenizer.tokenize('I was born in 92000, and this is falsé.' )
self.assertListEqual(
UpperCAmelCase , [
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',
'.',
] , )
lowerCamelCase__ : Optional[int] = tokenizer.convert_tokens_to_ids(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
lowerCamelCase__ : Optional[Any] = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [
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>',
'.',
] , )
@cached_property
def A_ ( self : Dict ) -> Tuple:
return BertGenerationTokenizer.from_pretrained('google/bert_for_seq_generation_L-24_bbc_encoder' )
@slow
def A_ ( self : Optional[int] ) -> List[str]:
lowerCamelCase__ : Union[str, Any] = 'Hello World!'
lowerCamelCase__ : Dict = [18536, 2260, 101]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@slow
def A_ ( self : Optional[Any] ) -> str:
lowerCamelCase__ : List[Any] = (
'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'
)
lowerCamelCase__ : Any = [
871,
419,
358,
946,
991,
2521,
452,
358,
1357,
387,
7751,
3536,
112,
985,
456,
126,
865,
938,
5400,
5734,
458,
1368,
467,
786,
2462,
5246,
1159,
633,
865,
4519,
457,
582,
852,
2557,
427,
916,
508,
405,
34324,
497,
391,
408,
11342,
1244,
385,
100,
938,
985,
456,
574,
362,
12597,
3200,
3129,
1172,
]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@require_torch
@slow
def A_ ( self : int ) -> Optional[Any]:
import torch
from transformers import BertGenerationConfig, BertGenerationEncoder
# Build sequence
lowerCamelCase__ : str = list(self.big_tokenizer.get_vocab().keys() )[:10]
lowerCamelCase__ : int = ' '.join(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.big_tokenizer.encode_plus(UpperCAmelCase , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.big_tokenizer.batch_encode_plus(
[sequence + ' ' + sequence] , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Tuple = BertGenerationConfig()
lowerCamelCase__ : Optional[Any] = BertGenerationEncoder(UpperCAmelCase )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**UpperCAmelCase )
model(**UpperCAmelCase )
@slow
def A_ ( self : Optional[int] ) -> List[Any]:
# fmt: off
lowerCamelCase__ : Any = {'input_ids': [[39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114], [448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 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], [484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 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]], '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, 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, 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=UpperCAmelCase , model_name='google/bert_for_seq_generation_L-24_bbc_encoder' , revision='c817d1fd1be2ffa69431227a1fe320544943d4db' , )
| 50
| 1
|
import copy
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ..auto import CONFIG_MAPPING
_UpperCAmelCase : int = logging.get_logger(__name__)
_UpperCAmelCase : str = {
"""SenseTime/deformable-detr""": """https://huggingface.co/sensetime/deformable-detr/resolve/main/config.json""",
# See all Deformable DETR models at https://huggingface.co/models?filter=deformable-detr
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """deformable_detr"""
UpperCAmelCase__ = {
"""hidden_size""": """d_model""",
"""num_attention_heads""": """encoder_attention_heads""",
}
def __init__( self : str , UpperCAmelCase : Any=True , UpperCAmelCase : str=None , UpperCAmelCase : Dict=3 , UpperCAmelCase : Union[str, Any]=300 , UpperCAmelCase : Dict=1024 , UpperCAmelCase : Optional[int]=6 , UpperCAmelCase : Optional[int]=1024 , UpperCAmelCase : Optional[int]=8 , UpperCAmelCase : List[str]=6 , UpperCAmelCase : int=1024 , UpperCAmelCase : str=8 , UpperCAmelCase : Optional[int]=0.0 , UpperCAmelCase : str=True , UpperCAmelCase : List[Any]="relu" , UpperCAmelCase : str=256 , UpperCAmelCase : Dict=0.1 , UpperCAmelCase : List[Any]=0.0 , UpperCAmelCase : Optional[Any]=0.0 , UpperCAmelCase : List[str]=0.0_2 , UpperCAmelCase : int=1.0 , UpperCAmelCase : int=True , UpperCAmelCase : List[str]=False , UpperCAmelCase : str="sine" , UpperCAmelCase : int="resnet50" , UpperCAmelCase : Any=True , UpperCAmelCase : int=False , UpperCAmelCase : Union[str, Any]=4 , UpperCAmelCase : Union[str, Any]=4 , UpperCAmelCase : Tuple=4 , UpperCAmelCase : Any=False , UpperCAmelCase : Optional[int]=300 , UpperCAmelCase : List[str]=False , UpperCAmelCase : int=1 , UpperCAmelCase : Optional[Any]=5 , UpperCAmelCase : Optional[int]=2 , UpperCAmelCase : Union[str, Any]=1 , UpperCAmelCase : Optional[Any]=1 , UpperCAmelCase : List[str]=5 , UpperCAmelCase : Dict=2 , UpperCAmelCase : Any=0.1 , UpperCAmelCase : List[Any]=0.2_5 , UpperCAmelCase : Dict=False , **UpperCAmelCase : Optional[Any] , ) -> int:
if backbone_config is not None and use_timm_backbone:
raise ValueError('You can\'t specify both `backbone_config` and `use_timm_backbone`.' )
if not use_timm_backbone:
if backbone_config is None:
logger.info('`backbone_config` is `None`. Initializing the config with the default `ResNet` backbone.' )
lowerCamelCase__ : Optional[Any] = CONFIG_MAPPING['resnet'](out_features=['stage4'] )
elif isinstance(UpperCAmelCase , UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = backbone_config.get('model_type' )
lowerCamelCase__ : int = CONFIG_MAPPING[backbone_model_type]
lowerCamelCase__ : Union[str, Any] = config_class.from_dict(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = use_timm_backbone
lowerCamelCase__ : List[Any] = backbone_config
lowerCamelCase__ : Union[str, Any] = num_channels
lowerCamelCase__ : str = num_queries
lowerCamelCase__ : Any = max_position_embeddings
lowerCamelCase__ : Optional[Any] = d_model
lowerCamelCase__ : List[str] = encoder_ffn_dim
lowerCamelCase__ : Optional[Any] = encoder_layers
lowerCamelCase__ : Tuple = encoder_attention_heads
lowerCamelCase__ : Optional[int] = decoder_ffn_dim
lowerCamelCase__ : List[Any] = decoder_layers
lowerCamelCase__ : Optional[Any] = decoder_attention_heads
lowerCamelCase__ : str = dropout
lowerCamelCase__ : Any = attention_dropout
lowerCamelCase__ : Any = activation_dropout
lowerCamelCase__ : Tuple = activation_function
lowerCamelCase__ : Union[str, Any] = init_std
lowerCamelCase__ : str = init_xavier_std
lowerCamelCase__ : Optional[int] = encoder_layerdrop
lowerCamelCase__ : Union[str, Any] = auxiliary_loss
lowerCamelCase__ : List[Any] = position_embedding_type
lowerCamelCase__ : List[Any] = backbone
lowerCamelCase__ : int = use_pretrained_backbone
lowerCamelCase__ : List[str] = dilation
# deformable attributes
lowerCamelCase__ : Any = num_feature_levels
lowerCamelCase__ : Optional[int] = encoder_n_points
lowerCamelCase__ : Union[str, Any] = decoder_n_points
lowerCamelCase__ : Union[str, Any] = two_stage
lowerCamelCase__ : str = two_stage_num_proposals
lowerCamelCase__ : Union[str, Any] = with_box_refine
if two_stage is True and with_box_refine is False:
raise ValueError('If two_stage is True, with_box_refine must be True.' )
# Hungarian matcher
lowerCamelCase__ : int = class_cost
lowerCamelCase__ : str = bbox_cost
lowerCamelCase__ : Optional[int] = giou_cost
# Loss coefficients
lowerCamelCase__ : int = mask_loss_coefficient
lowerCamelCase__ : int = dice_loss_coefficient
lowerCamelCase__ : Tuple = bbox_loss_coefficient
lowerCamelCase__ : Union[str, Any] = giou_loss_coefficient
lowerCamelCase__ : Any = eos_coefficient
lowerCamelCase__ : Any = focal_alpha
lowerCamelCase__ : Tuple = disable_custom_kernels
super().__init__(is_encoder_decoder=UpperCAmelCase , **UpperCAmelCase )
@property
def A_ ( self : str ) -> int:
return self.encoder_attention_heads
@property
def A_ ( self : str ) -> int:
return self.d_model
def A_ ( self : Optional[Any] ) -> List[Any]:
lowerCamelCase__ : int = copy.deepcopy(self.__dict__ )
if self.backbone_config is not None:
lowerCamelCase__ : Union[str, Any] = self.backbone_config.to_dict()
lowerCamelCase__ : Dict = self.__class__.model_type
return output
| 50
|
import os
import tempfile
from functools import partial
from unittest import TestCase
from unittest.mock import patch
import numpy as np
import pytest
from datasets.arrow_dataset import Dataset
from datasets.search import ElasticSearchIndex, FaissIndex, MissingIndex
from .utils import require_elasticsearch, require_faiss
_UpperCAmelCase : str = pytest.mark.integration
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : List[Any] ) -> Union[str, Any]:
lowerCamelCase__ : List[Any] = Dataset.from_dict({'filename': ['my_name-train' + '_' + str(UpperCAmelCase ) for x in np.arange(30 ).tolist()]} )
return dset
def A_ ( self : Optional[Any] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
lowerCamelCase__ : List[Any] = dset.map(
lambda UpperCAmelCase , UpperCAmelCase : {"vecs": i * np.ones(5 , dtype=np.floataa )} , with_indices=UpperCAmelCase , keep_in_memory=UpperCAmelCase )
lowerCamelCase__ : Tuple = dset.add_faiss_index('vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
dset.drop_index('vecs' )
def A_ ( self : Union[str, Any] ) -> int:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT , )
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : List[str] ) -> Tuple:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , metric_type=faiss.METRIC_INNER_PRODUCT , )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
dset.save_faiss_index('vecs' , tmp_file.name )
dset.load_faiss_index('vecs2' , tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs2' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' )
dset.drop_index('vecs' )
self.assertRaises(UpperCAmelCase , partial(dset.get_nearest_examples , 'vecs2' , np.ones(5 , dtype=np.floataa ) ) )
def A_ ( self : Dict ) -> Dict:
from elasticsearch import Elasticsearch
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : List[Any] = {'acknowledged': True}
mocked_bulk.return_value([(True, None)] * 30 )
lowerCamelCase__ : int = {'hits': {'hits': [{'_score': 1, '_id': 29}]}}
lowerCamelCase__ : List[str] = Elasticsearch()
dset.add_elasticsearch_index('filename' , es_client=UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Dict = dset.get_nearest_examples('filename' , 'my_name-train_29' )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Any ) -> Dict:
import faiss
lowerCamelCase__ : Tuple = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
# add vectors
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsNotNone(index.faiss_index )
self.assertEqual(index.faiss_index.ntotal , 5 )
index.add_vectors(np.zeros((5, 5) , dtype=np.floataa ) )
self.assertEqual(index.faiss_index.ntotal , 10 )
# single query
lowerCamelCase__ : int = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Any = 1
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = index.search(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search , query.reshape(-1 , 1 ) )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
# batched queries
lowerCamelCase__ : str = np.eye(5 , dtype=np.floataa )[::-1]
lowerCamelCase__ , lowerCamelCase__ : Dict = index.search_batch(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search_batch , queries[0] )
lowerCamelCase__ : str = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[str] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([4, 3, 2, 1, 0] , UpperCAmelCase )
def A_ ( self : List[Any] ) -> List[Any]:
import faiss
lowerCamelCase__ : Any = FaissIndex(string_factory='Flat' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
lowerCamelCase__ : Tuple = FaissIndex(string_factory='LSH' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexLSH )
with self.assertRaises(UpperCAmelCase ):
lowerCamelCase__ : List[str] = FaissIndex(string_factory='Flat' , custom_index=faiss.IndexFlat(5 ) )
def A_ ( self : List[str] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Optional[Any] = faiss.IndexFlat(5 )
lowerCamelCase__ : int = FaissIndex(custom_index=UpperCAmelCase )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
def A_ ( self : Any ) -> Optional[int]:
import faiss
lowerCamelCase__ : Any = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
index.save(tmp_file.name )
lowerCamelCase__ : List[Any] = FaissIndex.load(tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ : List[str] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Tuple = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(UpperCAmelCase )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
@require_faiss
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Any:
import faiss
lowerCamelCase__ : Optional[Any] = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
lowerCamelCase__ : Optional[int] = 'index.faiss'
lowerCamelCase__ : Optional[Any] = F"""mock://{index_name}"""
index.save(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Tuple = FaissIndex.load(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Optional[int] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Dict = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(_UpperCAmelCase )
assert scores[0] > 0
assert indices[0] == 1
@require_elasticsearch
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Dict ) -> List[Any]:
from elasticsearch import Elasticsearch
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : Any = Elasticsearch()
lowerCamelCase__ : Tuple = {'acknowledged': True}
lowerCamelCase__ : Tuple = ElasticSearchIndex(es_client=UpperCAmelCase )
mocked_bulk.return_value([(True, None)] * 3 )
index.add_documents(['foo', 'bar', 'foobar'] )
# single query
lowerCamelCase__ : Optional[int] = 'foo'
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search(UpperCAmelCase )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# single query with timeout
lowerCamelCase__ : Any = 'foo'
lowerCamelCase__ : List[str] = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Tuple = index.search(UpperCAmelCase , request_timeout=30 )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# batched queries
lowerCamelCase__ : List[str] = ['foo', 'bar', 'foobar']
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : str = index.search_batch(UpperCAmelCase )
lowerCamelCase__ : List[str] = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[Any] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
# batched queries with timeout
lowerCamelCase__ : str = ['foo', 'bar', 'foobar']
lowerCamelCase__ : List[Any] = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search_batch(UpperCAmelCase , request_timeout=30 )
lowerCamelCase__ : Optional[Any] = [scores[0] for scores in total_scores]
lowerCamelCase__ : Dict = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
| 50
| 1
|
from dataclasses import dataclass
from typing import Optional, Tuple, Union
import torch
import torch.nn as nn
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import BaseOutput
from .embeddings import GaussianFourierProjection, TimestepEmbedding, Timesteps
from .modeling_utils import ModelMixin
from .unet_ad_blocks import get_down_block, get_mid_block, get_out_block, get_up_block
@dataclass
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase ):
@register_to_config
def __init__( self : List[str] , UpperCAmelCase : int = 65536 , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 0 , UpperCAmelCase : str = "fourier" , UpperCAmelCase : bool = True , UpperCAmelCase : bool = False , UpperCAmelCase : float = 0.0 , UpperCAmelCase : Tuple[str] = ("DownBlock1DNoSkip", "DownBlock1D", "AttnDownBlock1D") , UpperCAmelCase : Tuple[str] = ("AttnUpBlock1D", "UpBlock1D", "UpBlock1DNoSkip") , UpperCAmelCase : Tuple[str] = "UNetMidBlock1D" , UpperCAmelCase : str = None , UpperCAmelCase : Tuple[int] = (32, 32, 64) , UpperCAmelCase : str = None , UpperCAmelCase : int = 8 , UpperCAmelCase : int = 1 , UpperCAmelCase : bool = False , ) -> List[Any]:
super().__init__()
lowerCamelCase__ : Optional[int] = sample_size
# time
if time_embedding_type == "fourier":
lowerCamelCase__ : Optional[Any] = GaussianFourierProjection(
embedding_size=8 , set_W_to_weight=UpperCAmelCase , log=UpperCAmelCase , flip_sin_to_cos=UpperCAmelCase )
lowerCamelCase__ : Any = 2 * block_out_channels[0]
elif time_embedding_type == "positional":
lowerCamelCase__ : List[Any] = Timesteps(
block_out_channels[0] , flip_sin_to_cos=UpperCAmelCase , downscale_freq_shift=UpperCAmelCase )
lowerCamelCase__ : Dict = block_out_channels[0]
if use_timestep_embedding:
lowerCamelCase__ : str = block_out_channels[0] * 4
lowerCamelCase__ : List[Any] = TimestepEmbedding(
in_channels=UpperCAmelCase , time_embed_dim=UpperCAmelCase , act_fn=UpperCAmelCase , out_dim=block_out_channels[0] , )
lowerCamelCase__ : Any = nn.ModuleList([] )
lowerCamelCase__ : Tuple = None
lowerCamelCase__ : List[str] = nn.ModuleList([] )
lowerCamelCase__ : Optional[int] = None
# down
lowerCamelCase__ : Optional[int] = in_channels
for i, down_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Union[str, Any] = output_channel
lowerCamelCase__ : Tuple = block_out_channels[i]
if i == 0:
input_channel += extra_in_channels
lowerCamelCase__ : Union[str, Any] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Optional[int] = get_down_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_downsample=not is_final_block or downsample_each_block , )
self.down_blocks.append(UpperCAmelCase )
# mid
lowerCamelCase__ : Optional[int] = get_mid_block(
UpperCAmelCase , in_channels=block_out_channels[-1] , mid_channels=block_out_channels[-1] , out_channels=block_out_channels[-1] , embed_dim=block_out_channels[0] , num_layers=UpperCAmelCase , add_downsample=UpperCAmelCase , )
# up
lowerCamelCase__ : Optional[int] = list(reversed(UpperCAmelCase ) )
lowerCamelCase__ : Optional[int] = reversed_block_out_channels[0]
if out_block_type is None:
lowerCamelCase__ : List[str] = out_channels
else:
lowerCamelCase__ : Any = block_out_channels[0]
for i, up_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Tuple = output_channel
lowerCamelCase__ : Union[str, Any] = (
reversed_block_out_channels[i + 1] if i < len(UpperCAmelCase ) - 1 else final_upsample_channels
)
lowerCamelCase__ : List[str] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Dict = get_up_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_upsample=not is_final_block , )
self.up_blocks.append(UpperCAmelCase )
lowerCamelCase__ : int = output_channel
# out
lowerCamelCase__ : int = norm_num_groups if norm_num_groups is not None else min(block_out_channels[0] // 4 , 32 )
lowerCamelCase__ : List[Any] = get_out_block(
out_block_type=UpperCAmelCase , num_groups_out=UpperCAmelCase , embed_dim=block_out_channels[0] , out_channels=UpperCAmelCase , act_fn=UpperCAmelCase , fc_dim=block_out_channels[-1] // 4 , )
def A_ ( self : List[Any] , UpperCAmelCase : torch.FloatTensor , UpperCAmelCase : Union[torch.Tensor, float, int] , UpperCAmelCase : bool = True , ) -> Union[UNetaDOutput, Tuple]:
lowerCamelCase__ : Optional[Any] = timestep
if not torch.is_tensor(UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = torch.tensor([timesteps] , dtype=torch.long , device=sample.device )
elif torch.is_tensor(UpperCAmelCase ) and len(timesteps.shape ) == 0:
lowerCamelCase__ : List[str] = timesteps[None].to(sample.device )
lowerCamelCase__ : Optional[int] = self.time_proj(UpperCAmelCase )
if self.config.use_timestep_embedding:
lowerCamelCase__ : str = self.time_mlp(UpperCAmelCase )
else:
lowerCamelCase__ : List[str] = timestep_embed[..., None]
lowerCamelCase__ : str = timestep_embed.repeat([1, 1, sample.shape[2]] ).to(sample.dtype )
lowerCamelCase__ : str = timestep_embed.broadcast_to((sample.shape[:1] + timestep_embed.shape[1:]) )
# 2. down
lowerCamelCase__ : str = ()
for downsample_block in self.down_blocks:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = downsample_block(hidden_states=UpperCAmelCase , temb=UpperCAmelCase )
down_block_res_samples += res_samples
# 3. mid
if self.mid_block:
lowerCamelCase__ : Optional[Any] = self.mid_block(UpperCAmelCase , UpperCAmelCase )
# 4. up
for i, upsample_block in enumerate(self.up_blocks ):
lowerCamelCase__ : Dict = down_block_res_samples[-1:]
lowerCamelCase__ : Optional[Any] = down_block_res_samples[:-1]
lowerCamelCase__ : Any = upsample_block(UpperCAmelCase , res_hidden_states_tuple=UpperCAmelCase , temb=UpperCAmelCase )
# 5. post-process
if self.out_block:
lowerCamelCase__ : Any = self.out_block(UpperCAmelCase , UpperCAmelCase )
if not return_dict:
return (sample,)
return UNetaDOutput(sample=UpperCAmelCase )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> bool:
lowerCamelCase__ : List[str] = len(_UpperCAmelCase )
lowerCamelCase__ : str = [[False] * (required_sum + 1) for _ in range(arr_len + 1 )]
# for each arr value, a sum of zero(0) can be formed by not taking any element
# hence True/1
for i in range(arr_len + 1 ):
lowerCamelCase__ : Tuple = True
# sum is not zero and set is empty then false
for i in range(1 , required_sum + 1 ):
lowerCamelCase__ : Dict = False
for i in range(1 , arr_len + 1 ):
for j in range(1 , required_sum + 1 ):
if arr[i - 1] > j:
lowerCamelCase__ : str = subset[i - 1][j]
if arr[i - 1] <= j:
lowerCamelCase__ : Dict = subset[i - 1][j] or subset[i - 1][j - arr[i - 1]]
return subset[arr_len][required_sum]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
from ...utils import (
OptionalDependencyNotAvailable,
is_flax_available,
is_torch_available,
is_transformers_available,
)
try:
if not (is_transformers_available() and is_torch_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ...utils.dummy_torch_and_transformers_objects import * # noqa F403
else:
from .multicontrolnet import MultiControlNetModel
from .pipeline_controlnet import StableDiffusionControlNetPipeline
from .pipeline_controlnet_imgaimg import StableDiffusionControlNetImgaImgPipeline
from .pipeline_controlnet_inpaint import StableDiffusionControlNetInpaintPipeline
if is_transformers_available() and is_flax_available():
from .pipeline_flax_controlnet import FlaxStableDiffusionControlNetPipeline
| 50
|
import torch
from transformers import PreTrainedModel, XLMRobertaConfig, XLMRobertaModel
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """M-CLIP"""
def __init__( self : Optional[Any] , UpperCAmelCase : Union[str, Any]=1024 , UpperCAmelCase : Tuple=768 , **UpperCAmelCase : Optional[int] ) -> Dict:
lowerCamelCase__ : Optional[int] = transformerDimSize
lowerCamelCase__ : Optional[Any] = imageDimSize
super().__init__(**UpperCAmelCase )
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = MCLIPConfig
def __init__( self : List[Any] , UpperCAmelCase : Dict , *UpperCAmelCase : Any , **UpperCAmelCase : Dict ) -> Dict:
super().__init__(UpperCAmelCase , *UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Tuple = XLMRobertaModel(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = torch.nn.Linear(
in_features=config.transformerDimensions , out_features=config.numDims )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Dict ) -> Tuple:
lowerCamelCase__ : Any = self.transformer(input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase )[0]
lowerCamelCase__ : int = (embs * attention_mask.unsqueeze(2 )).sum(dim=1 ) / attention_mask.sum(dim=1 )[:, None]
return self.LinearTransformation(UpperCAmelCase ), embs
| 50
| 1
|
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_UpperCAmelCase : List[str] = logging.get_logger(__name__)
_UpperCAmelCase : Union[str, Any] = {
"""facebook/xlm-roberta-xl""": """https://huggingface.co/facebook/xlm-roberta-xl/resolve/main/config.json""",
"""facebook/xlm-roberta-xxl""": """https://huggingface.co/facebook/xlm-roberta-xxl/resolve/main/config.json""",
# See all XLM-RoBERTa-XL models at https://huggingface.co/models?filter=xlm-roberta-xl
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """xlm-roberta-xl"""
def __init__( self : List[str] , UpperCAmelCase : Dict=250880 , UpperCAmelCase : int=2560 , UpperCAmelCase : Tuple=36 , UpperCAmelCase : Dict=32 , UpperCAmelCase : Dict=10240 , UpperCAmelCase : List[Any]="gelu" , UpperCAmelCase : Optional[Any]=0.1 , UpperCAmelCase : Optional[int]=0.1 , UpperCAmelCase : int=514 , UpperCAmelCase : Any=1 , UpperCAmelCase : Dict=0.0_2 , UpperCAmelCase : Union[str, Any]=1e-05 , UpperCAmelCase : List[Any]=1 , UpperCAmelCase : Any=0 , UpperCAmelCase : List[str]=2 , UpperCAmelCase : List[Any]="absolute" , UpperCAmelCase : Tuple=True , UpperCAmelCase : Optional[Any]=None , **UpperCAmelCase : Any , ) -> List[str]:
super().__init__(pad_token_id=UpperCAmelCase , bos_token_id=UpperCAmelCase , eos_token_id=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Dict = vocab_size
lowerCamelCase__ : str = hidden_size
lowerCamelCase__ : str = num_hidden_layers
lowerCamelCase__ : List[Any] = num_attention_heads
lowerCamelCase__ : List[str] = hidden_act
lowerCamelCase__ : Dict = intermediate_size
lowerCamelCase__ : Any = hidden_dropout_prob
lowerCamelCase__ : List[str] = attention_probs_dropout_prob
lowerCamelCase__ : Tuple = max_position_embeddings
lowerCamelCase__ : Any = type_vocab_size
lowerCamelCase__ : Optional[Any] = initializer_range
lowerCamelCase__ : List[str] = layer_norm_eps
lowerCamelCase__ : Any = position_embedding_type
lowerCamelCase__ : int = use_cache
lowerCamelCase__ : int = classifier_dropout
class lowerCAmelCase ( __UpperCamelCase ):
@property
def A_ ( self : Optional[Any] ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCamelCase__ : Optional[Any] = {0: 'batch', 1: 'choice', 2: 'sequence'}
else:
lowerCamelCase__ : Dict = {0: 'batch', 1: 'sequence'}
return OrderedDict(
[
('input_ids', dynamic_axis),
('attention_mask', dynamic_axis),
] )
| 50
|
from itertools import count
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 50 ) -> int:
lowerCamelCase__ : Optional[Any] = [1] * min_block_length
for n in count(_UpperCAmelCase ):
fill_count_functions.append(1 )
for block_length in range(_UpperCAmelCase , n + 1 ):
for block_start in range(n - block_length ):
fill_count_functions[n] += fill_count_functions[
n - block_start - block_length - 1
]
fill_count_functions[n] += 1
if fill_count_functions[n] > 100_0000:
break
return n
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
import argparse
from torch import nn
# transformers_old should correspond to branch `save_old_prophetnet_model_structure` here
# original prophetnet_checkpoints are saved under `patrickvonplaten/..._old` respectively
from transformers_old.modeling_prophetnet import (
ProphetNetForConditionalGeneration as ProphetNetForConditionalGenerationOld,
)
from transformers_old.modeling_xlm_prophetnet import (
XLMProphetNetForConditionalGeneration as XLMProphetNetForConditionalGenerationOld,
)
from transformers import ProphetNetForConditionalGeneration, XLMProphetNetForConditionalGeneration, logging
_UpperCAmelCase : str = logging.get_logger(__name__)
logging.set_verbosity_info()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Dict:
if "xprophetnet" in prophetnet_checkpoint_path:
lowerCamelCase__ : Union[str, Any] = XLMProphetNetForConditionalGenerationOld.from_pretrained(_UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = XLMProphetNetForConditionalGeneration.from_pretrained(
_UpperCAmelCase , output_loading_info=_UpperCAmelCase )
else:
lowerCamelCase__ : str = ProphetNetForConditionalGenerationOld.from_pretrained(_UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : int = ProphetNetForConditionalGeneration.from_pretrained(
_UpperCAmelCase , output_loading_info=_UpperCAmelCase )
lowerCamelCase__ : List[str] = ['key_proj', 'value_proj', 'query_proj']
lowerCamelCase__ : Union[str, Any] = {
'self_attn': 'ngram_self_attn',
'cross_attn': 'encoder_attn',
'cross_attn_layer_norm': 'encoder_attn_layer_norm',
'feed_forward_layer_norm': 'final_layer_norm',
'feed_forward': '',
'intermediate': 'fc1',
'output': 'fc2',
'key_proj': 'k_proj',
'query_proj': 'q_proj',
'value_proj': 'v_proj',
'word_embeddings': 'embed_tokens',
'embeddings_layer_norm': 'emb_layer_norm',
'relative_pos_embeddings': 'relative_linear',
'ngram_embeddings': 'ngram_input_embed',
'position_embeddings': 'embed_positions',
}
for key in loading_info["missing_keys"]:
lowerCamelCase__ : Optional[Any] = key.split('.' )
if attributes[0] == "lm_head":
lowerCamelCase__ : Optional[Any] = prophet
lowerCamelCase__ : Union[str, Any] = prophet_old
else:
lowerCamelCase__ : int = prophet.prophetnet
lowerCamelCase__ : Optional[int] = prophet_old.model
lowerCamelCase__ : Dict = False
for attribute in attributes:
if attribute in mapping:
lowerCamelCase__ : Any = mapping[attribute]
if not hasattr(_UpperCAmelCase , _UpperCAmelCase ) and len(_UpperCAmelCase ) > 0:
lowerCamelCase__ : Any = attribute
elif hasattr(_UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Union[str, Any] = attribute
if attribute == "weight":
assert old_model.weight.shape == model.weight.shape, "Shapes have to match!"
lowerCamelCase__ : Union[str, Any] = old_model.weight
logger.info(F"""{attribute} is initialized.""" )
lowerCamelCase__ : Optional[Any] = True
break
elif attribute == "bias":
assert old_model.bias.shape == model.bias.shape, "Shapes have to match!"
lowerCamelCase__ : List[Any] = old_model.bias
logger.info(F"""{attribute} is initialized""" )
lowerCamelCase__ : Union[str, Any] = True
break
elif attribute in special_keys and hasattr(_UpperCAmelCase , 'in_proj_weight' ):
lowerCamelCase__ : Union[str, Any] = old_model.in_proj_weight.shape[0] // 3
lowerCamelCase__ : Dict = getattr(_UpperCAmelCase , _UpperCAmelCase )
param.weight.shape == old_model.in_proj_weight[:embed_dim, :].shape, "Shapes have to match"
param.bias.shape == old_model.in_proj_bias[:embed_dim].shape, "Shapes have to match"
if attribute == "query_proj":
lowerCamelCase__ : Tuple = nn.Parameter(old_model.in_proj_weight[:embed_dim, :] )
lowerCamelCase__ : List[str] = nn.Parameter(old_model.in_proj_bias[:embed_dim] )
elif attribute == "key_proj":
lowerCamelCase__ : Optional[int] = nn.Parameter(old_model.in_proj_weight[embed_dim : 2 * embed_dim, :] )
lowerCamelCase__ : Optional[Any] = nn.Parameter(old_model.in_proj_bias[embed_dim : 2 * embed_dim] )
elif attribute == "value_proj":
lowerCamelCase__ : Optional[int] = nn.Parameter(old_model.in_proj_weight[2 * embed_dim :, :] )
lowerCamelCase__ : Any = nn.Parameter(old_model.in_proj_bias[2 * embed_dim :] )
lowerCamelCase__ : Optional[Any] = True
break
elif attribute == "position_embeddings":
assert (
model.position_embeddings.weight.shape[-1] == old_model.embed_positions.weight.shape[-1]
), "Hidden size has to match"
assert model.position_embeddings.weight.shape[0] == 512, "We want 512 position_embeddings."
lowerCamelCase__ : Any = nn.Parameter(old_model.embed_positions.weight[:512, :] )
lowerCamelCase__ : Optional[Any] = True
break
if attribute.isdigit():
lowerCamelCase__ : Optional[Any] = model[int(_UpperCAmelCase )]
lowerCamelCase__ : List[str] = old_model[int(_UpperCAmelCase )]
else:
lowerCamelCase__ : int = getattr(_UpperCAmelCase , _UpperCAmelCase )
if old_attribute == "":
lowerCamelCase__ : Optional[int] = old_model
else:
if not hasattr(_UpperCAmelCase , _UpperCAmelCase ):
raise ValueError(F"""{old_model} does not have {old_attribute}""" )
lowerCamelCase__ : Optional[int] = getattr(_UpperCAmelCase , _UpperCAmelCase )
if not is_key_init:
raise ValueError(F"""{key} was not correctly initialized!""" )
print(F"""Saving model to {pytorch_dump_folder_path}""" )
prophet.save_pretrained(_UpperCAmelCase )
if __name__ == "__main__":
_UpperCAmelCase : Dict = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--prophetnet_checkpoint_path""", default=None, type=str, required=True, help="""Path the official PyTorch dump."""
)
parser.add_argument(
"""--pytorch_dump_folder_path""", default=None, type=str, required=True, help="""Path to the output PyTorch model."""
)
_UpperCAmelCase : str = parser.parse_args()
convert_prophetnet_checkpoint_to_pytorch(args.prophetnet_checkpoint_path, args.pytorch_dump_folder_path)
| 50
|
from __future__ import annotations
from typing import Any
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> None:
create_state_space_tree(_UpperCAmelCase , [] , 0 )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> None:
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__":
_UpperCAmelCase : list[Any] = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["""A""", """B""", """C"""])
generate_all_subsequences(seq)
| 50
| 1
|
_UpperCAmelCase : List[Any] = {
"""a""": """AAAAA""",
"""b""": """AAAAB""",
"""c""": """AAABA""",
"""d""": """AAABB""",
"""e""": """AABAA""",
"""f""": """AABAB""",
"""g""": """AABBA""",
"""h""": """AABBB""",
"""i""": """ABAAA""",
"""j""": """BBBAA""",
"""k""": """ABAAB""",
"""l""": """ABABA""",
"""m""": """ABABB""",
"""n""": """ABBAA""",
"""o""": """ABBAB""",
"""p""": """ABBBA""",
"""q""": """ABBBB""",
"""r""": """BAAAA""",
"""s""": """BAAAB""",
"""t""": """BAABA""",
"""u""": """BAABB""",
"""v""": """BBBAB""",
"""w""": """BABAA""",
"""x""": """BABAB""",
"""y""": """BABBA""",
"""z""": """BABBB""",
""" """: """ """,
}
_UpperCAmelCase : int = {value: key for key, value in encode_dict.items()}
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Optional[int] = ''
for letter in word.lower():
if letter.isalpha() or letter == " ":
encoded += encode_dict[letter]
else:
raise Exception('encode() accepts only letters of the alphabet and spaces' )
return encoded
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
if set(_UpperCAmelCase ) - {"A", "B", " "} != set():
raise Exception('decode() accepts only \'A\', \'B\' and spaces' )
lowerCamelCase__ : Any = ''
for word in coded.split():
while len(_UpperCAmelCase ) != 0:
decoded += decode_dict[word[:5]]
lowerCamelCase__ : Union[str, Any] = word[5:]
decoded += " "
return decoded.strip()
if __name__ == "__main__":
from doctest import testmod
testmod()
| 50
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from tokenizers import processors
from ...tokenization_utils import AddedToken, BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_nllb import NllbTokenizer
else:
_UpperCAmelCase : int = None
_UpperCAmelCase : Dict = logging.get_logger(__name__)
_UpperCAmelCase : Optional[int] = {"""vocab_file""": """sentencepiece.bpe.model""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : List[Any] = {
"""vocab_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/sentencepiece.bpe.model"""
),
},
"""tokenizer_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : List[str] = {
"""facebook/nllb-large-en-ro""": 10_24,
"""facebook/nllb-200-distilled-600M""": 10_24,
}
# fmt: off
_UpperCAmelCase : Optional[int] = ["""ace_Arab""", """ace_Latn""", """acm_Arab""", """acq_Arab""", """aeb_Arab""", """afr_Latn""", """ajp_Arab""", """aka_Latn""", """amh_Ethi""", """apc_Arab""", """arb_Arab""", """ars_Arab""", """ary_Arab""", """arz_Arab""", """asm_Beng""", """ast_Latn""", """awa_Deva""", """ayr_Latn""", """azb_Arab""", """azj_Latn""", """bak_Cyrl""", """bam_Latn""", """ban_Latn""", """bel_Cyrl""", """bem_Latn""", """ben_Beng""", """bho_Deva""", """bjn_Arab""", """bjn_Latn""", """bod_Tibt""", """bos_Latn""", """bug_Latn""", """bul_Cyrl""", """cat_Latn""", """ceb_Latn""", """ces_Latn""", """cjk_Latn""", """ckb_Arab""", """crh_Latn""", """cym_Latn""", """dan_Latn""", """deu_Latn""", """dik_Latn""", """dyu_Latn""", """dzo_Tibt""", """ell_Grek""", """eng_Latn""", """epo_Latn""", """est_Latn""", """eus_Latn""", """ewe_Latn""", """fao_Latn""", """pes_Arab""", """fij_Latn""", """fin_Latn""", """fon_Latn""", """fra_Latn""", """fur_Latn""", """fuv_Latn""", """gla_Latn""", """gle_Latn""", """glg_Latn""", """grn_Latn""", """guj_Gujr""", """hat_Latn""", """hau_Latn""", """heb_Hebr""", """hin_Deva""", """hne_Deva""", """hrv_Latn""", """hun_Latn""", """hye_Armn""", """ibo_Latn""", """ilo_Latn""", """ind_Latn""", """isl_Latn""", """ita_Latn""", """jav_Latn""", """jpn_Jpan""", """kab_Latn""", """kac_Latn""", """kam_Latn""", """kan_Knda""", """kas_Arab""", """kas_Deva""", """kat_Geor""", """knc_Arab""", """knc_Latn""", """kaz_Cyrl""", """kbp_Latn""", """kea_Latn""", """khm_Khmr""", """kik_Latn""", """kin_Latn""", """kir_Cyrl""", """kmb_Latn""", """kon_Latn""", """kor_Hang""", """kmr_Latn""", """lao_Laoo""", """lvs_Latn""", """lij_Latn""", """lim_Latn""", """lin_Latn""", """lit_Latn""", """lmo_Latn""", """ltg_Latn""", """ltz_Latn""", """lua_Latn""", """lug_Latn""", """luo_Latn""", """lus_Latn""", """mag_Deva""", """mai_Deva""", """mal_Mlym""", """mar_Deva""", """min_Latn""", """mkd_Cyrl""", """plt_Latn""", """mlt_Latn""", """mni_Beng""", """khk_Cyrl""", """mos_Latn""", """mri_Latn""", """zsm_Latn""", """mya_Mymr""", """nld_Latn""", """nno_Latn""", """nob_Latn""", """npi_Deva""", """nso_Latn""", """nus_Latn""", """nya_Latn""", """oci_Latn""", """gaz_Latn""", """ory_Orya""", """pag_Latn""", """pan_Guru""", """pap_Latn""", """pol_Latn""", """por_Latn""", """prs_Arab""", """pbt_Arab""", """quy_Latn""", """ron_Latn""", """run_Latn""", """rus_Cyrl""", """sag_Latn""", """san_Deva""", """sat_Beng""", """scn_Latn""", """shn_Mymr""", """sin_Sinh""", """slk_Latn""", """slv_Latn""", """smo_Latn""", """sna_Latn""", """snd_Arab""", """som_Latn""", """sot_Latn""", """spa_Latn""", """als_Latn""", """srd_Latn""", """srp_Cyrl""", """ssw_Latn""", """sun_Latn""", """swe_Latn""", """swh_Latn""", """szl_Latn""", """tam_Taml""", """tat_Cyrl""", """tel_Telu""", """tgk_Cyrl""", """tgl_Latn""", """tha_Thai""", """tir_Ethi""", """taq_Latn""", """taq_Tfng""", """tpi_Latn""", """tsn_Latn""", """tso_Latn""", """tuk_Latn""", """tum_Latn""", """tur_Latn""", """twi_Latn""", """tzm_Tfng""", """uig_Arab""", """ukr_Cyrl""", """umb_Latn""", """urd_Arab""", """uzn_Latn""", """vec_Latn""", """vie_Latn""", """war_Latn""", """wol_Latn""", """xho_Latn""", """ydd_Hebr""", """yor_Latn""", """yue_Hant""", """zho_Hans""", """zho_Hant""", """zul_Latn"""]
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = ["""input_ids""", """attention_mask"""]
UpperCAmelCase__ = NllbTokenizer
UpperCAmelCase__ = []
UpperCAmelCase__ = []
def __init__( self : Tuple , UpperCAmelCase : int=None , UpperCAmelCase : Any=None , UpperCAmelCase : str="<s>" , UpperCAmelCase : Optional[Any]="</s>" , UpperCAmelCase : str="</s>" , UpperCAmelCase : Tuple="<s>" , UpperCAmelCase : Optional[Any]="<unk>" , UpperCAmelCase : List[str]="<pad>" , UpperCAmelCase : Union[str, Any]="<mask>" , UpperCAmelCase : Tuple=None , UpperCAmelCase : int=None , UpperCAmelCase : Dict=None , UpperCAmelCase : Any=False , **UpperCAmelCase : Optional[int] , ) -> Tuple:
# Mask token behave like a normal word, i.e. include the space before it
lowerCamelCase__ : List[Any] = AddedToken(UpperCAmelCase , lstrip=UpperCAmelCase , rstrip=UpperCAmelCase ) if isinstance(UpperCAmelCase , UpperCAmelCase ) else mask_token
lowerCamelCase__ : Union[str, Any] = legacy_behaviour
super().__init__(
vocab_file=UpperCAmelCase , tokenizer_file=UpperCAmelCase , bos_token=UpperCAmelCase , eos_token=UpperCAmelCase , sep_token=UpperCAmelCase , cls_token=UpperCAmelCase , unk_token=UpperCAmelCase , pad_token=UpperCAmelCase , mask_token=UpperCAmelCase , src_lang=UpperCAmelCase , tgt_lang=UpperCAmelCase , additional_special_tokens=UpperCAmelCase , legacy_behaviour=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = vocab_file
lowerCamelCase__ : Dict = False if not self.vocab_file else True
lowerCamelCase__ : Optional[Any] = FAIRSEQ_LANGUAGE_CODES.copy()
if additional_special_tokens is not None:
# Only add those special tokens if they are not already there.
_additional_special_tokens.extend(
[t for t in additional_special_tokens if t not in _additional_special_tokens] )
self.add_special_tokens({'additional_special_tokens': _additional_special_tokens} )
lowerCamelCase__ : str = {
lang_code: self.convert_tokens_to_ids(UpperCAmelCase ) for lang_code in FAIRSEQ_LANGUAGE_CODES
}
lowerCamelCase__ : int = src_lang if src_lang is not None else 'eng_Latn'
lowerCamelCase__ : List[Any] = self.convert_tokens_to_ids(self._src_lang )
lowerCamelCase__ : str = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def A_ ( self : int ) -> str:
return self._src_lang
@src_lang.setter
def A_ ( self : List[Any] , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Any = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
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 A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : Dict = [self.sep_token_id]
lowerCamelCase__ : List[str] = [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 + sep + token_ids_a + sep ) * [0]
def A_ ( self : int , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Optional[str] , UpperCAmelCase : Optional[str] , **UpperCAmelCase : List[str] ) -> Dict:
if src_lang is None or tgt_lang is None:
raise ValueError('Translation requires a `src_lang` and a `tgt_lang` for this model' )
lowerCamelCase__ : Optional[int] = src_lang
lowerCamelCase__ : Optional[int] = self(UpperCAmelCase , add_special_tokens=UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.convert_tokens_to_ids(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tgt_lang_id
return inputs
def A_ ( self : Dict , UpperCAmelCase : List[str] , UpperCAmelCase : str = "eng_Latn" , UpperCAmelCase : Optional[List[str]] = None , UpperCAmelCase : str = "fra_Latn" , **UpperCAmelCase : Dict , ) -> BatchEncoding:
lowerCamelCase__ : Any = src_lang
lowerCamelCase__ : int = tgt_lang
return super().prepare_seqaseq_batch(UpperCAmelCase , UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Union[str, Any] ) -> Optional[int]:
return self.set_src_lang_special_tokens(self.src_lang )
def A_ ( self : Any ) -> Union[str, Any]:
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def A_ ( self : str , UpperCAmelCase : Optional[Any] ) -> None:
lowerCamelCase__ : int = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : int = []
lowerCamelCase__ : str = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : int = [self.cur_lang_code]
lowerCamelCase__ : Tuple = [self.eos_token_id]
lowerCamelCase__ : Any = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : Optional[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : str = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : int , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Union[str, Any] = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : Dict = []
lowerCamelCase__ : Union[str, Any] = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : Any = [self.cur_lang_code]
lowerCamelCase__ : Optional[Any] = [self.eos_token_id]
lowerCamelCase__ : Union[str, Any] = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : List[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : Optional[int] = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : Union[str, Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not self.can_save_slow_tokenizer:
raise ValueError(
'Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '
'tokenizer.' )
if not os.path.isdir(UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory.""" )
return
lowerCamelCase__ : int = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCAmelCase ):
copyfile(self.vocab_file , UpperCAmelCase )
return (out_vocab_file,)
| 50
| 1
|
from PIL import Image
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Image:
def brightness(_UpperCAmelCase ) -> float:
return 128 + level + (c - 128)
if not -255.0 <= level <= 255.0:
raise ValueError('level must be between -255.0 (black) and 255.0 (white)' )
return img.point(_UpperCAmelCase )
if __name__ == "__main__":
# Load image
with Image.open("""image_data/lena.jpg""") as img:
# Change brightness to 100
_UpperCAmelCase : Optional[Any] = change_brightness(img, 1_00)
brigt_img.save("""image_data/lena_brightness.png""", format="""png""")
| 50
|
import argparse
import torch
from torch import nn
from transformers import MaMaaaConfig, MaMaaaForConditionalGeneration
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Optional[int] = [
'encoder.version',
'decoder.version',
'model.encoder.version',
'model.decoder.version',
'decoder.output_projection.weight',
'_float_tensor',
'encoder.embed_positions._float_tensor',
'decoder.embed_positions._float_tensor',
]
for k in ignore_keys:
state_dict.pop(_UpperCAmelCase , _UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[Any]:
lowerCamelCase__ , lowerCamelCase__ : List[str] = emb.weight.shape
lowerCamelCase__ : Tuple = nn.Linear(_UpperCAmelCase , _UpperCAmelCase , bias=_UpperCAmelCase )
lowerCamelCase__ : Dict = emb.weight.data
return lin_layer
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Tuple = torch.load(_UpperCAmelCase , map_location='cpu' )
lowerCamelCase__ : List[str] = mam_aaa['args'] or mam_aaa['cfg']['model']
lowerCamelCase__ : Optional[int] = mam_aaa['model']
remove_ignore_keys_(_UpperCAmelCase )
lowerCamelCase__ : str = state_dict['encoder.embed_tokens.weight'].shape[0]
lowerCamelCase__ : Union[str, Any] = MaMaaaConfig(
vocab_size=_UpperCAmelCase , max_position_embeddings=1024 , encoder_layers=args.encoder_layers , decoder_layers=args.decoder_layers , encoder_attention_heads=args.encoder_attention_heads , decoder_attention_heads=args.decoder_attention_heads , encoder_ffn_dim=args.encoder_ffn_embed_dim , decoder_ffn_dim=args.decoder_ffn_embed_dim , d_model=args.encoder_embed_dim , encoder_layerdrop=args.encoder_layerdrop , decoder_layerdrop=args.decoder_layerdrop , dropout=args.dropout , attention_dropout=args.attention_dropout , activation_dropout=args.activation_dropout , activation_function='relu' , )
lowerCamelCase__ : Optional[Any] = state_dict['decoder.embed_tokens.weight']
lowerCamelCase__ : Union[str, Any] = MaMaaaForConditionalGeneration(_UpperCAmelCase )
model.model.load_state_dict(_UpperCAmelCase , strict=_UpperCAmelCase )
lowerCamelCase__ : List[str] = make_linear_from_emb(model.model.shared )
return model
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""fairseq_path""", type=str, help="""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.""")
_UpperCAmelCase : str = parser.parse_args()
_UpperCAmelCase : Optional[Any] = convert_fairseq_mamaaa_checkpoint_from_disk(args.fairseq_pathß)
model.save_pretrained(args.pytorch_dump_folder_path)
| 50
| 1
|
from __future__ import annotations
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> tuple[int, int]:
if b == 0:
return (1, 0)
((lowerCamelCase__) , (lowerCamelCase__)) : Tuple = extended_euclid(_UpperCAmelCase , a % b )
lowerCamelCase__ : str = a // b
return (y, x - k * y)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> int:
((lowerCamelCase__) , (lowerCamelCase__)) : List[Any] = extended_euclid(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = na * na
lowerCamelCase__ : Optional[Any] = ra * x * na + ra * y * na
return (n % m + m) % m
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> int:
((lowerCamelCase__) , (lowerCamelCase__)) : List[Any] = extended_euclid(_UpperCAmelCase , _UpperCAmelCase )
if b < 0:
lowerCamelCase__ : List[str] = (b % n + n) % n
return b
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> int:
lowerCamelCase__ , lowerCamelCase__ : Tuple = invert_modulo(_UpperCAmelCase , _UpperCAmelCase ), invert_modulo(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : str = na * na
lowerCamelCase__ : str = ra * x * na + ra * y * na
return (n % m + m) % m
if __name__ == "__main__":
from doctest import testmod
testmod(name="""chinese_remainder_theorem""", verbose=True)
testmod(name="""chinese_remainder_theorem2""", verbose=True)
testmod(name="""invert_modulo""", verbose=True)
testmod(name="""extended_euclid""", verbose=True)
| 50
|
import argparse
import json
import os
import time
import zipfile
from get_ci_error_statistics import download_artifact, get_artifacts_links
from transformers import logging
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Dict:
lowerCamelCase__ : str = set()
lowerCamelCase__ : Any = []
def parse_line(_UpperCAmelCase ):
for line in fp:
if isinstance(_UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Any = line.decode('UTF-8' )
if "warnings summary (final)" in line:
continue
# This means we are outside the body of a warning
elif not line.startswith(' ' ):
# process a single warning and move it to `selected_warnings`.
if len(_UpperCAmelCase ) > 0:
lowerCamelCase__ : str = '\n'.join(_UpperCAmelCase )
# Only keep the warnings specified in `targets`
if any(F""": {x}: """ in warning for x in targets ):
selected_warnings.add(_UpperCAmelCase )
buffer.clear()
continue
else:
lowerCamelCase__ : List[str] = line.strip()
buffer.append(_UpperCAmelCase )
if from_gh:
for filename in os.listdir(_UpperCAmelCase ):
lowerCamelCase__ : Dict = os.path.join(_UpperCAmelCase , _UpperCAmelCase )
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
else:
try:
with zipfile.ZipFile(_UpperCAmelCase ) as z:
for filename in z.namelist():
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with z.open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
except Exception:
logger.warning(
F"""{artifact_path} is either an invalid zip file or something else wrong. This file is skipped.""" )
return selected_warnings
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = set()
lowerCamelCase__ : Optional[int] = [os.path.join(_UpperCAmelCase , _UpperCAmelCase ) for p in os.listdir(_UpperCAmelCase ) if (p.endswith('.zip' ) or from_gh)]
for p in paths:
selected_warnings.update(extract_warnings_from_single_artifact(_UpperCAmelCase , _UpperCAmelCase ) )
return selected_warnings
if __name__ == "__main__":
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
return values.split(',' )
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""--workflow_run_id""", type=str, required=True, help="""A GitHub Actions workflow run id.""")
parser.add_argument(
"""--output_dir""",
type=str,
required=True,
help="""Where to store the downloaded artifacts and other result files.""",
)
parser.add_argument("""--token""", default=None, type=str, help="""A token that has actions:read permission.""")
# optional parameters
parser.add_argument(
"""--targets""",
default="""DeprecationWarning,UserWarning,FutureWarning""",
type=list_str,
help="""Comma-separated list of target warning(s) which we want to extract.""",
)
parser.add_argument(
"""--from_gh""",
action="""store_true""",
help="""If running from a GitHub action workflow and collecting warnings from its artifacts.""",
)
_UpperCAmelCase : Union[str, Any] = parser.parse_args()
_UpperCAmelCase : Dict = args.from_gh
if from_gh:
# The artifacts have to be downloaded using `actions/download-artifact@v3`
pass
else:
os.makedirs(args.output_dir, exist_ok=True)
# get download links
_UpperCAmelCase : Union[str, Any] = get_artifacts_links(args.workflow_run_id, token=args.token)
with open(os.path.join(args.output_dir, """artifacts.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(artifacts, fp, ensure_ascii=False, indent=4)
# download artifacts
for idx, (name, url) in enumerate(artifacts.items()):
print(name)
print(url)
print("""=""" * 80)
download_artifact(name, url, args.output_dir, args.token)
# Be gentle to GitHub
time.sleep(1)
# extract warnings from artifacts
_UpperCAmelCase : Dict = extract_warnings(args.output_dir, args.targets)
_UpperCAmelCase : Optional[Any] = sorted(selected_warnings)
with open(os.path.join(args.output_dir, """selected_warnings.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(selected_warnings, fp, ensure_ascii=False, indent=4)
| 50
| 1
|
# Lint as: python3
import sys
from collections.abc import Mapping
from typing import TYPE_CHECKING
import numpy as np
import pyarrow as pa
from .. import config
from ..utils.py_utils import map_nested
from .formatting import TensorFormatter
if TYPE_CHECKING:
import torch
class lowerCAmelCase ( TensorFormatter[Mapping, """torch.Tensor""", Mapping] ):
def __init__( self : str , UpperCAmelCase : str=None , **UpperCAmelCase : int ) -> Optional[Any]:
super().__init__(features=UpperCAmelCase )
lowerCamelCase__ : int = torch_tensor_kwargs
import torch # noqa import torch at initialization
def A_ ( self : Union[str, Any] , UpperCAmelCase : List[Any] ) -> Optional[Any]:
import torch
if isinstance(UpperCAmelCase , UpperCAmelCase ) and column:
if all(
isinstance(UpperCAmelCase , torch.Tensor ) and x.shape == column[0].shape and x.dtype == column[0].dtype
for x in column ):
return torch.stack(UpperCAmelCase )
return column
def A_ ( self : str , UpperCAmelCase : Dict ) -> List[Any]:
import torch
if isinstance(UpperCAmelCase , (str, bytes, type(UpperCAmelCase )) ):
return value
elif isinstance(UpperCAmelCase , (np.character, np.ndarray) ) and np.issubdtype(value.dtype , np.character ):
return value.tolist()
lowerCamelCase__ : Union[str, Any] = {}
if isinstance(UpperCAmelCase , (np.number, np.ndarray) ) and np.issubdtype(value.dtype , np.integer ):
lowerCamelCase__ : Optional[Any] = {'dtype': torch.intaa}
elif isinstance(UpperCAmelCase , (np.number, np.ndarray) ) and np.issubdtype(value.dtype , np.floating ):
lowerCamelCase__ : List[Any] = {'dtype': torch.floataa}
elif config.PIL_AVAILABLE and "PIL" in sys.modules:
import PIL.Image
if isinstance(UpperCAmelCase , PIL.Image.Image ):
lowerCamelCase__ : List[str] = np.asarray(UpperCAmelCase )
return torch.tensor(UpperCAmelCase , **{**default_dtype, **self.torch_tensor_kwargs} )
def A_ ( self : str , UpperCAmelCase : Tuple ) -> Optional[Any]:
import torch
# support for torch, tf, jax etc.
if hasattr(UpperCAmelCase , '__array__' ) and not isinstance(UpperCAmelCase , torch.Tensor ):
lowerCamelCase__ : Optional[int] = data_struct.__array__()
# support for nested types like struct of list of struct
if isinstance(UpperCAmelCase , np.ndarray ):
if data_struct.dtype == object: # torch tensors cannot be instantied from an array of objects
return self._consolidate([self.recursive_tensorize(UpperCAmelCase ) for substruct in data_struct] )
elif isinstance(UpperCAmelCase , (list, tuple) ):
return self._consolidate([self.recursive_tensorize(UpperCAmelCase ) for substruct in data_struct] )
return self._tensorize(UpperCAmelCase )
def A_ ( self : List[str] , UpperCAmelCase : dict ) -> List[str]:
return map_nested(self._recursive_tensorize , UpperCAmelCase , map_list=UpperCAmelCase )
def A_ ( self : Tuple , UpperCAmelCase : pa.Table ) -> Mapping:
lowerCamelCase__ : Tuple = self.numpy_arrow_extractor().extract_row(UpperCAmelCase )
lowerCamelCase__ : List[Any] = self.python_features_decoder.decode_row(UpperCAmelCase )
return self.recursive_tensorize(UpperCAmelCase )
def A_ ( self : List[str] , UpperCAmelCase : pa.Table ) -> "torch.Tensor":
lowerCamelCase__ : Optional[int] = self.numpy_arrow_extractor().extract_column(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.python_features_decoder.decode_column(UpperCAmelCase , pa_table.column_names[0] )
lowerCamelCase__ : int = self.recursive_tensorize(UpperCAmelCase )
lowerCamelCase__ : List[Any] = self._consolidate(UpperCAmelCase )
return column
def A_ ( self : Optional[int] , UpperCAmelCase : pa.Table ) -> Mapping:
lowerCamelCase__ : Optional[int] = self.numpy_arrow_extractor().extract_batch(UpperCAmelCase )
lowerCamelCase__ : List[Any] = self.python_features_decoder.decode_batch(UpperCAmelCase )
lowerCamelCase__ : Any = self.recursive_tensorize(UpperCAmelCase )
for column_name in batch:
lowerCamelCase__ : Optional[Any] = self._consolidate(batch[column_name] )
return batch
| 50
|
import flax.linen as nn
import jax
import jax.numpy as jnp
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : Any ) -> Any:
lowerCamelCase__ : str = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : int , UpperCAmelCase : Dict ) -> Optional[Any]:
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : str = hidden_states.shape
lowerCamelCase__ : Union[str, Any] = jax.image.resize(
UpperCAmelCase , shape=(batch, height * 2, width * 2, channels) , method='nearest' , )
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Tuple = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : str , UpperCAmelCase : Union[str, Any] ) -> Optional[Any]:
# pad = ((0, 0), (0, 1), (0, 1), (0, 0)) # pad height and width dim
# hidden_states = jnp.pad(hidden_states, pad_width=pad)
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = 0.0
UpperCAmelCase__ = None
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> Union[str, Any]:
lowerCamelCase__ : Optional[Any] = self.in_channels if self.out_channels is None else self.out_channels
lowerCamelCase__ : Tuple = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : int = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Union[str, Any] = nn.Dense(UpperCAmelCase , dtype=self.dtype )
lowerCamelCase__ : Union[str, Any] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : List[Any] = nn.Dropout(self.dropout_prob )
lowerCamelCase__ : Tuple = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Optional[Any] = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
lowerCamelCase__ : Union[str, Any] = None
if use_nin_shortcut:
lowerCamelCase__ : Dict = nn.Conv(
UpperCAmelCase , kernel_size=(1, 1) , strides=(1, 1) , padding='VALID' , dtype=self.dtype , )
def __call__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : int=True ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = hidden_states
lowerCamelCase__ : List[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[Any] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Any = self.conva(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.time_emb_proj(nn.swish(UpperCAmelCase ) )
lowerCamelCase__ : List[str] = jnp.expand_dims(jnp.expand_dims(UpperCAmelCase , 1 ) , 1 )
lowerCamelCase__ : List[str] = hidden_states + temb
lowerCamelCase__ : Optional[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[str] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = self.dropout(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = self.conva(UpperCAmelCase )
if self.conv_shortcut is not None:
lowerCamelCase__ : Dict = self.conv_shortcut(UpperCAmelCase )
return hidden_states + residual
| 50
| 1
|
from typing import List, Optional, Union
import torch
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,
)
_UpperCAmelCase : Any = logging.get_logger(__name__) # pylint: disable=invalid-name
_UpperCAmelCase : Tuple = """
Examples:
```py
>>> from diffusers import KandinskyV22Pipeline, KandinskyV22PriorPipeline
>>> import torch
>>> pipe_prior = KandinskyV22PriorPipeline.from_pretrained(\"kandinsky-community/kandinsky-2-2-prior\")
>>> pipe_prior.to(\"cuda\")
>>> prompt = \"red cat, 4k photo\"
>>> out = pipe_prior(prompt)
>>> image_emb = out.image_embeds
>>> zero_image_emb = out.negative_image_embeds
>>> pipe = KandinskyV22Pipeline.from_pretrained(\"kandinsky-community/kandinsky-2-2-decoder\")
>>> pipe.to(\"cuda\")
>>> image = pipe(
... image_embeds=image_emb,
... negative_image_embeds=zero_image_emb,
... height=768,
... width=768,
... num_inference_steps=50,
... ).images
>>> image[0].save(\"cat.png\")
```
"""
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=8 ) -> Dict:
lowerCamelCase__ : Tuple = height // scale_factor**2
if height % scale_factor**2 != 0:
new_height += 1
lowerCamelCase__ : Dict = width // scale_factor**2
if width % scale_factor**2 != 0:
new_width += 1
return new_height * scale_factor, new_width * scale_factor
class lowerCAmelCase ( __UpperCamelCase ):
def __init__( self : Tuple , UpperCAmelCase : UNetaDConditionModel , UpperCAmelCase : DDPMScheduler , UpperCAmelCase : VQModel , ) -> str:
super().__init__()
self.register_modules(
unet=UpperCAmelCase , scheduler=UpperCAmelCase , movq=UpperCAmelCase , )
lowerCamelCase__ : Optional[Any] = 2 ** (len(self.movq.config.block_out_channels ) - 1)
def A_ ( self : int , UpperCAmelCase : Optional[int] , UpperCAmelCase : Optional[int] , UpperCAmelCase : List[Any] , UpperCAmelCase : List[str] , UpperCAmelCase : int , UpperCAmelCase : Tuple ) -> List[str]:
if latents is None:
lowerCamelCase__ : Dict = randn_tensor(UpperCAmelCase , generator=UpperCAmelCase , device=UpperCAmelCase , dtype=UpperCAmelCase )
else:
if latents.shape != shape:
raise ValueError(F"""Unexpected latents shape, got {latents.shape}, expected {shape}""" )
lowerCamelCase__ : str = latents.to(UpperCAmelCase )
lowerCamelCase__ : List[Any] = latents * scheduler.init_noise_sigma
return latents
def A_ ( self : Tuple , UpperCAmelCase : Tuple=0 ) -> List[Any]:
if is_accelerate_available():
from accelerate import cpu_offload
else:
raise ImportError('Please install accelerate via `pip install accelerate`' )
lowerCamelCase__ : Any = torch.device(F"""cuda:{gpu_id}""" )
lowerCamelCase__ : List[Any] = [
self.unet,
self.movq,
]
for cpu_offloaded_model in models:
if cpu_offloaded_model is not None:
cpu_offload(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Optional[Any] , UpperCAmelCase : int=0 ) -> List[Any]:
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.' )
lowerCamelCase__ : Union[str, Any] = torch.device(F"""cuda:{gpu_id}""" )
if self.device.type != "cpu":
self.to('cpu' , silence_dtype_warnings=UpperCAmelCase )
torch.cuda.empty_cache() # otherwise we don't see the memory savings (but they probably exist)
lowerCamelCase__ : Union[str, Any] = None
for cpu_offloaded_model in [self.unet, self.movq]:
lowerCamelCase__ , lowerCamelCase__ : str = cpu_offload_with_hook(UpperCAmelCase , UpperCAmelCase , prev_module_hook=UpperCAmelCase )
# We'll offload the last model manually.
lowerCamelCase__ : Optional[int] = hook
@property
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._execution_device
def A_ ( self : List[str] ) -> Any:
if not hasattr(self.unet , '_hf_hook' ):
return self.device
for module in self.unet.modules():
if (
hasattr(UpperCAmelCase , '_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(UpperCAmelCase )
def __call__( self : str , UpperCAmelCase : Union[torch.FloatTensor, List[torch.FloatTensor]] , UpperCAmelCase : Union[torch.FloatTensor, List[torch.FloatTensor]] , UpperCAmelCase : int = 512 , UpperCAmelCase : int = 512 , UpperCAmelCase : int = 100 , UpperCAmelCase : float = 4.0 , UpperCAmelCase : int = 1 , UpperCAmelCase : Optional[Union[torch.Generator, List[torch.Generator]]] = None , UpperCAmelCase : Optional[torch.FloatTensor] = None , UpperCAmelCase : Optional[str] = "pil" , UpperCAmelCase : bool = True , ) -> Optional[int]:
lowerCamelCase__ : Any = self._execution_device
lowerCamelCase__ : Dict = guidance_scale > 1.0
if isinstance(UpperCAmelCase , UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = torch.cat(UpperCAmelCase , dim=0 )
lowerCamelCase__ : int = image_embeds.shape[0] * num_images_per_prompt
if isinstance(UpperCAmelCase , UpperCAmelCase ):
lowerCamelCase__ : int = torch.cat(UpperCAmelCase , dim=0 )
if do_classifier_free_guidance:
lowerCamelCase__ : int = image_embeds.repeat_interleave(UpperCAmelCase , dim=0 )
lowerCamelCase__ : Tuple = negative_image_embeds.repeat_interleave(UpperCAmelCase , dim=0 )
lowerCamelCase__ : Optional[int] = torch.cat([negative_image_embeds, image_embeds] , dim=0 ).to(dtype=self.unet.dtype , device=UpperCAmelCase )
self.scheduler.set_timesteps(UpperCAmelCase , device=UpperCAmelCase )
lowerCamelCase__ : Dict = self.scheduler.timesteps
lowerCamelCase__ : Dict = self.unet.config.in_channels
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = downscale_height_and_width(UpperCAmelCase , UpperCAmelCase , self.movq_scale_factor )
# create initial latent
lowerCamelCase__ : Union[str, Any] = self.prepare_latents(
(batch_size, num_channels_latents, height, width) , image_embeds.dtype , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , self.scheduler , )
for i, t in enumerate(self.progress_bar(UpperCAmelCase ) ):
# expand the latents if we are doing classifier free guidance
lowerCamelCase__ : List[str] = torch.cat([latents] * 2 ) if do_classifier_free_guidance else latents
lowerCamelCase__ : str = {'image_embeds': image_embeds}
lowerCamelCase__ : int = self.unet(
sample=UpperCAmelCase , timestep=UpperCAmelCase , encoder_hidden_states=UpperCAmelCase , added_cond_kwargs=UpperCAmelCase , return_dict=UpperCAmelCase , )[0]
if do_classifier_free_guidance:
lowerCamelCase__ , lowerCamelCase__ : Any = noise_pred.split(latents.shape[1] , dim=1 )
lowerCamelCase__ , lowerCamelCase__ : List[Any] = noise_pred.chunk(2 )
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = variance_pred.chunk(2 )
lowerCamelCase__ : str = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
lowerCamelCase__ : List[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"]
):
lowerCamelCase__ , lowerCamelCase__ : List[Any] = noise_pred.split(latents.shape[1] , dim=1 )
# compute the previous noisy sample x_t -> x_t-1
lowerCamelCase__ : Optional[int] = self.scheduler.step(
UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , generator=UpperCAmelCase , )[0]
# post-processing
lowerCamelCase__ : Dict = self.movq.decode(UpperCAmelCase , force_not_quantize=UpperCAmelCase )['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"]:
lowerCamelCase__ : Optional[int] = image * 0.5 + 0.5
lowerCamelCase__ : Optional[int] = image.clamp(0 , 1 )
lowerCamelCase__ : List[str] = image.cpu().permute(0 , 2 , 3 , 1 ).float().numpy()
if output_type == "pil":
lowerCamelCase__ : Dict = self.numpy_to_pil(UpperCAmelCase )
if not return_dict:
return (image,)
return ImagePipelineOutput(images=UpperCAmelCase )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Optional[Any] = set()
# edges = list of graph's edges
lowerCamelCase__ : List[str] = get_edges(_UpperCAmelCase )
# While there are still elements in edges list, take an arbitrary edge
# (from_node, to_node) and add his extremity to chosen_vertices and then
# remove all arcs adjacent to the from_node and to_node
while edges:
lowerCamelCase__ , lowerCamelCase__ : str = edges.pop()
chosen_vertices.add(_UpperCAmelCase )
chosen_vertices.add(_UpperCAmelCase )
for edge in edges.copy():
if from_node in edge or to_node in edge:
edges.discard(_UpperCAmelCase )
return chosen_vertices
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Union[str, Any] = set()
for from_node, to_nodes in graph.items():
for to_node in to_nodes:
edges.add((from_node, to_node) )
return edges
if __name__ == "__main__":
import doctest
doctest.testmod()
# graph = {0: [1, 3], 1: [0, 3], 2: [0, 3, 4], 3: [0, 1, 2], 4: [2, 3]}
# print(f"Matching vertex cover:\n{matching_min_vertex_cover(graph)}")
| 50
| 1
|
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers import CLIPTokenizer, CLIPTokenizerFast
from transformers.models.clip.tokenization_clip import VOCAB_FILES_NAMES
from transformers.testing_utils import require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import CLIPImageProcessor, CLIPProcessor
@require_vision
class lowerCAmelCase ( unittest.TestCase ):
def A_ ( self : List[Any] ) -> str:
lowerCamelCase__ : int = tempfile.mkdtemp()
# fmt: off
lowerCamelCase__ : Optional[Any] = ['l', 'o', 'w', 'e', 'r', 's', 't', 'i', 'd', 'n', 'lo', 'l</w>', 'w</w>', 'r</w>', 't</w>', 'low</w>', 'er</w>', 'lowest</w>', 'newer</w>', 'wider', '<unk>', '<|startoftext|>', '<|endoftext|>']
# fmt: on
lowerCamelCase__ : int = dict(zip(UpperCAmelCase , range(len(UpperCAmelCase ) ) ) )
lowerCamelCase__ : Any = ['#version: 0.2', 'l o', 'lo w</w>', 'e r</w>', '']
lowerCamelCase__ : List[str] = {'unk_token': '<unk>'}
lowerCamelCase__ : Tuple = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['vocab_file'] )
lowerCamelCase__ : List[str] = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['merges_file'] )
with open(self.vocab_file , 'w' , encoding='utf-8' ) as fp:
fp.write(json.dumps(UpperCAmelCase ) + '\n' )
with open(self.merges_file , 'w' , encoding='utf-8' ) as fp:
fp.write('\n'.join(UpperCAmelCase ) )
lowerCamelCase__ : Optional[Any] = {
'do_resize': True,
'size': 20,
'do_center_crop': True,
'crop_size': 18,
'do_normalize': True,
'image_mean': [0.4_8_1_4_5_4_6_6, 0.4_5_7_8_2_7_5, 0.4_0_8_2_1_0_7_3],
'image_std': [0.2_6_8_6_2_9_5_4, 0.2_6_1_3_0_2_5_8, 0.2_7_5_7_7_7_1_1],
}
lowerCamelCase__ : int = os.path.join(self.tmpdirname , UpperCAmelCase )
with open(self.image_processor_file , 'w' , encoding='utf-8' ) as fp:
json.dump(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Optional[int] , **UpperCAmelCase : Tuple ) -> Optional[int]:
return CLIPTokenizer.from_pretrained(self.tmpdirname , **UpperCAmelCase )
def A_ ( self : Union[str, Any] , **UpperCAmelCase : Any ) -> str:
return CLIPTokenizerFast.from_pretrained(self.tmpdirname , **UpperCAmelCase )
def A_ ( self : str , **UpperCAmelCase : Dict ) -> Optional[int]:
return CLIPImageProcessor.from_pretrained(self.tmpdirname , **UpperCAmelCase )
def A_ ( self : int ) -> Any:
shutil.rmtree(self.tmpdirname )
def A_ ( self : List[str] ) -> str:
lowerCamelCase__ : List[str] = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta )]
lowerCamelCase__ : List[Any] = [Image.fromarray(np.moveaxis(UpperCAmelCase , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def A_ ( self : int ) -> int:
lowerCamelCase__ : str = self.get_tokenizer()
lowerCamelCase__ : Optional[Any] = self.get_rust_tokenizer()
lowerCamelCase__ : str = self.get_image_processor()
lowerCamelCase__ : List[Any] = CLIPProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
processor_slow.save_pretrained(self.tmpdirname )
lowerCamelCase__ : Dict = CLIPProcessor.from_pretrained(self.tmpdirname , use_fast=UpperCAmelCase )
lowerCamelCase__ : Any = CLIPProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
processor_fast.save_pretrained(self.tmpdirname )
lowerCamelCase__ : Tuple = CLIPProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor_slow.tokenizer.get_vocab() , tokenizer_slow.get_vocab() )
self.assertEqual(processor_fast.tokenizer.get_vocab() , tokenizer_fast.get_vocab() )
self.assertEqual(tokenizer_slow.get_vocab() , tokenizer_fast.get_vocab() )
self.assertIsInstance(processor_slow.tokenizer , UpperCAmelCase )
self.assertIsInstance(processor_fast.tokenizer , UpperCAmelCase )
self.assertEqual(processor_slow.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertEqual(processor_fast.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor_slow.image_processor , UpperCAmelCase )
self.assertIsInstance(processor_fast.image_processor , UpperCAmelCase )
def A_ ( self : int ) -> Any:
lowerCamelCase__ : Any = CLIPProcessor(tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
lowerCamelCase__ : Optional[Any] = self.get_tokenizer(bos_token='(BOS)' , eos_token='(EOS)' )
lowerCamelCase__ : Any = self.get_image_processor(do_normalize=UpperCAmelCase , padding_value=1.0 )
lowerCamelCase__ : Any = CLIPProcessor.from_pretrained(
self.tmpdirname , bos_token='(BOS)' , eos_token='(EOS)' , do_normalize=UpperCAmelCase , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , UpperCAmelCase )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , UpperCAmelCase )
def A_ ( self : Any ) -> List[Any]:
lowerCamelCase__ : int = self.get_image_processor()
lowerCamelCase__ : Union[str, Any] = self.get_tokenizer()
lowerCamelCase__ : Dict = CLIPProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : Optional[int] = self.prepare_image_inputs()
lowerCamelCase__ : str = image_processor(UpperCAmelCase , return_tensors='np' )
lowerCamelCase__ : Tuple = processor(images=UpperCAmelCase , return_tensors='np' )
for key in input_image_proc.keys():
self.assertAlmostEqual(input_image_proc[key].sum() , input_processor[key].sum() , delta=1e-2 )
def A_ ( self : List[str] ) -> Optional[int]:
lowerCamelCase__ : List[Any] = self.get_image_processor()
lowerCamelCase__ : Any = self.get_tokenizer()
lowerCamelCase__ : List[Any] = CLIPProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = 'lower newer'
lowerCamelCase__ : Dict = processor(text=UpperCAmelCase )
lowerCamelCase__ : Tuple = tokenizer(UpperCAmelCase )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def A_ ( self : Union[str, Any] ) -> str:
lowerCamelCase__ : str = self.get_image_processor()
lowerCamelCase__ : Union[str, Any] = self.get_tokenizer()
lowerCamelCase__ : Optional[Any] = CLIPProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = 'lower newer'
lowerCamelCase__ : Union[str, Any] = self.prepare_image_inputs()
lowerCamelCase__ : List[Any] = processor(text=UpperCAmelCase , images=UpperCAmelCase )
self.assertListEqual(list(inputs.keys() ) , ['input_ids', 'attention_mask', 'pixel_values'] )
# test if it raises when no input is passed
with pytest.raises(UpperCAmelCase ):
processor()
def A_ ( self : Optional[Any] ) -> Tuple:
lowerCamelCase__ : int = self.get_image_processor()
lowerCamelCase__ : Union[str, Any] = self.get_tokenizer()
lowerCamelCase__ : Tuple = CLIPProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : List[str] = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
lowerCamelCase__ : Optional[Any] = processor.batch_decode(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = tokenizer.batch_decode(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Optional[int] ) -> Dict:
lowerCamelCase__ : Any = self.get_image_processor()
lowerCamelCase__ : Tuple = self.get_tokenizer()
lowerCamelCase__ : Tuple = CLIPProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = 'lower newer'
lowerCamelCase__ : List[str] = self.prepare_image_inputs()
lowerCamelCase__ : str = processor(text=UpperCAmelCase , images=UpperCAmelCase )
self.assertListEqual(list(inputs.keys() ) , processor.model_input_names )
| 50
|
from __future__ import annotations
import math
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
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(_UpperCAmelCase ) + 1 ) , 6 ):
if number % i == 0 or number % (i + 2) == 0:
return False
return True
_UpperCAmelCase : Any = [num for num in range(3, 10_00_01, 2) if not is_prime(num)]
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list[int]:
if not isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise ValueError('n must be an integer' )
if n <= 0:
raise ValueError('n must be >= 0' )
lowerCamelCase__ : int = []
for num in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : Union[str, Any] = 0
while 2 * i * i <= odd_composites[num]:
lowerCamelCase__ : Dict = odd_composites[num] - 2 * i * i
if is_prime(_UpperCAmelCase ):
break
i += 1
else:
list_nums.append(odd_composites[num] )
if len(_UpperCAmelCase ) == n:
return list_nums
return []
def SCREAMING_SNAKE_CASE ( ) -> int:
return compute_nums(1 )[0]
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
from typing import Optional
import pyspark
from .. import Features, NamedSplit
from ..download import DownloadMode
from ..packaged_modules.spark.spark import Spark
from .abc import AbstractDatasetReader
class lowerCAmelCase ( __UpperCamelCase ):
def __init__( self : Optional[int] , UpperCAmelCase : pyspark.sql.DataFrame , UpperCAmelCase : Optional[NamedSplit] = None , UpperCAmelCase : Optional[Features] = None , UpperCAmelCase : bool = True , UpperCAmelCase : str = None , UpperCAmelCase : bool = False , UpperCAmelCase : str = None , UpperCAmelCase : bool = True , UpperCAmelCase : str = "arrow" , **UpperCAmelCase : List[Any] , ) -> List[Any]:
super().__init__(
split=UpperCAmelCase , features=UpperCAmelCase , cache_dir=UpperCAmelCase , keep_in_memory=UpperCAmelCase , streaming=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : str = load_from_cache_file
lowerCamelCase__ : Dict = file_format
lowerCamelCase__ : Union[str, Any] = Spark(
df=UpperCAmelCase , features=UpperCAmelCase , cache_dir=UpperCAmelCase , working_dir=UpperCAmelCase , **UpperCAmelCase , )
def A_ ( self : int ) -> List[Any]:
if self.streaming:
return self.builder.as_streaming_dataset(split=self.split )
lowerCamelCase__ : List[Any] = None if self._load_from_cache_file else DownloadMode.FORCE_REDOWNLOAD
self.builder.download_and_prepare(
download_mode=UpperCAmelCase , file_format=self._file_format , )
return self.builder.as_dataset(split=self.split )
| 50
|
import argparse
from collections import OrderedDict
from pathlib import Path
import requests
import torch
from PIL import Image
from transformers import GLPNConfig, GLPNForDepthEstimation, GLPNImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
_UpperCAmelCase : Dict = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[str]:
lowerCamelCase__ : str = OrderedDict()
for key, value in state_dict.items():
if key.startswith('module.encoder' ):
lowerCamelCase__ : Optional[Any] = key.replace('module.encoder' , 'glpn.encoder' )
if key.startswith('module.decoder' ):
lowerCamelCase__ : List[str] = key.replace('module.decoder' , 'decoder.stages' )
if "patch_embed" in key:
# replace for example patch_embed1 by patch_embeddings.0
lowerCamelCase__ : Dict = key[key.find('patch_embed' ) + len('patch_embed' )]
lowerCamelCase__ : Tuple = key.replace(F"""patch_embed{idx}""" , F"""patch_embeddings.{int(_UpperCAmelCase )-1}""" )
if "norm" in key:
lowerCamelCase__ : str = key.replace('norm' , 'layer_norm' )
if "glpn.encoder.layer_norm" in key:
# replace for example layer_norm1 by layer_norm.0
lowerCamelCase__ : Dict = key[key.find('glpn.encoder.layer_norm' ) + len('glpn.encoder.layer_norm' )]
lowerCamelCase__ : str = key.replace(F"""layer_norm{idx}""" , F"""layer_norm.{int(_UpperCAmelCase )-1}""" )
if "layer_norm1" in key:
lowerCamelCase__ : Optional[int] = key.replace('layer_norm1' , 'layer_norm_1' )
if "layer_norm2" in key:
lowerCamelCase__ : Optional[int] = key.replace('layer_norm2' , 'layer_norm_2' )
if "block" in key:
# replace for example block1 by block.0
lowerCamelCase__ : List[Any] = key[key.find('block' ) + len('block' )]
lowerCamelCase__ : int = key.replace(F"""block{idx}""" , F"""block.{int(_UpperCAmelCase )-1}""" )
if "attn.q" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('attn.q' , 'attention.self.query' )
if "attn.proj" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('attn.proj' , 'attention.output.dense' )
if "attn" in key:
lowerCamelCase__ : Dict = key.replace('attn' , 'attention.self' )
if "fc1" in key:
lowerCamelCase__ : Dict = key.replace('fc1' , 'dense1' )
if "fc2" in key:
lowerCamelCase__ : Any = key.replace('fc2' , 'dense2' )
if "linear_pred" in key:
lowerCamelCase__ : Dict = key.replace('linear_pred' , 'classifier' )
if "linear_fuse" in key:
lowerCamelCase__ : Tuple = key.replace('linear_fuse.conv' , 'linear_fuse' )
lowerCamelCase__ : List[str] = key.replace('linear_fuse.bn' , 'batch_norm' )
if "linear_c" in key:
# replace for example linear_c4 by linear_c.3
lowerCamelCase__ : Optional[Any] = key[key.find('linear_c' ) + len('linear_c' )]
lowerCamelCase__ : Dict = key.replace(F"""linear_c{idx}""" , F"""linear_c.{int(_UpperCAmelCase )-1}""" )
if "bot_conv" in key:
lowerCamelCase__ : str = key.replace('bot_conv' , '0.convolution' )
if "skip_conv1" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('skip_conv1' , '1.convolution' )
if "skip_conv2" in key:
lowerCamelCase__ : List[Any] = key.replace('skip_conv2' , '2.convolution' )
if "fusion1" in key:
lowerCamelCase__ : Optional[int] = key.replace('fusion1' , '1.fusion' )
if "fusion2" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('fusion2' , '2.fusion' )
if "fusion3" in key:
lowerCamelCase__ : List[Any] = key.replace('fusion3' , '3.fusion' )
if "fusion" in key and "conv" in key:
lowerCamelCase__ : str = key.replace('conv' , 'convolutional_layer' )
if key.startswith('module.last_layer_depth' ):
lowerCamelCase__ : Dict = key.replace('module.last_layer_depth' , 'head.head' )
lowerCamelCase__ : str = value
return new_state_dict
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
# for each of the encoder blocks:
for i in range(config.num_encoder_blocks ):
for j in range(config.depths[i] ):
# read in weights + bias of keys and values (which is a single matrix in the original implementation)
lowerCamelCase__ : Any = state_dict.pop(F"""glpn.encoder.block.{i}.{j}.attention.self.kv.weight""" )
lowerCamelCase__ : Optional[Any] = state_dict.pop(F"""glpn.encoder.block.{i}.{j}.attention.self.kv.bias""" )
# next, add keys and values (in that order) to the state dict
lowerCamelCase__ : Optional[int] = kv_weight[
: config.hidden_sizes[i], :
]
lowerCamelCase__ : Optional[int] = kv_bias[: config.hidden_sizes[i]]
lowerCamelCase__ : Any = kv_weight[
config.hidden_sizes[i] :, :
]
lowerCamelCase__ : Dict = kv_bias[config.hidden_sizes[i] :]
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : List[str] = 'http://images.cocodataset.org/val2017/000000039769.jpg'
lowerCamelCase__ : Tuple = Image.open(requests.get(_UpperCAmelCase , stream=_UpperCAmelCase ).raw )
return image
@torch.no_grad()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=False , _UpperCAmelCase=None ) -> Optional[int]:
lowerCamelCase__ : str = GLPNConfig(hidden_sizes=[64, 128, 320, 512] , decoder_hidden_size=64 , depths=[3, 8, 27, 3] )
# load image processor (only resize + rescale)
lowerCamelCase__ : Union[str, Any] = GLPNImageProcessor()
# prepare image
lowerCamelCase__ : str = prepare_img()
lowerCamelCase__ : Tuple = image_processor(images=_UpperCAmelCase , return_tensors='pt' ).pixel_values
logger.info('Converting model...' )
# load original state dict
lowerCamelCase__ : Any = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
# rename keys
lowerCamelCase__ : str = rename_keys(_UpperCAmelCase )
# key and value matrices need special treatment
read_in_k_v(_UpperCAmelCase , _UpperCAmelCase )
# create HuggingFace model and load state dict
lowerCamelCase__ : Dict = GLPNForDepthEstimation(_UpperCAmelCase )
model.load_state_dict(_UpperCAmelCase )
model.eval()
# forward pass
lowerCamelCase__ : List[str] = model(_UpperCAmelCase )
lowerCamelCase__ : Tuple = outputs.predicted_depth
# verify output
if model_name is not None:
if "nyu" in model_name:
lowerCamelCase__ : List[Any] = torch.tensor(
[[4.4_147, 4.0_873, 4.0_673], [3.7_890, 3.2_881, 3.1_525], [3.7_674, 3.5_423, 3.4_913]] )
elif "kitti" in model_name:
lowerCamelCase__ : List[str] = torch.tensor(
[[3.4_291, 2.7_865, 2.5_151], [3.2_841, 2.7_021, 2.3_502], [3.1_147, 2.4_625, 2.2_481]] )
else:
raise ValueError(F"""Unknown model name: {model_name}""" )
lowerCamelCase__ : Tuple = torch.Size([1, 480, 640] )
assert predicted_depth.shape == expected_shape
assert torch.allclose(predicted_depth[0, :3, :3] , _UpperCAmelCase , atol=1e-4 )
print('Looks ok!' )
# finally, push to hub if required
if push_to_hub:
logger.info('Pushing model and image processor to the hub...' )
model.push_to_hub(
repo_path_or_name=Path(_UpperCAmelCase , _UpperCAmelCase ) , organization='nielsr' , commit_message='Add model' , use_temp_dir=_UpperCAmelCase , )
image_processor.push_to_hub(
repo_path_or_name=Path(_UpperCAmelCase , _UpperCAmelCase ) , organization='nielsr' , commit_message='Add image processor' , use_temp_dir=_UpperCAmelCase , )
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
parser.add_argument(
"""--checkpoint_path""",
default=None,
type=str,
help="""Path to the original PyTorch checkpoint (.pth file).""",
)
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 to upload the model to the HuggingFace hub."""
)
parser.add_argument(
"""--model_name""",
default="""glpn-kitti""",
type=str,
help="""Name of the model in case you're pushing to the hub.""",
)
_UpperCAmelCase : int = parser.parse_args()
convert_glpn_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 50
| 1
|
from typing import TYPE_CHECKING
from ....utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
_UpperCAmelCase : Any = {"""configuration_mmbt""": ["""MMBTConfig"""]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_UpperCAmelCase : int = ["""MMBTForClassification""", """MMBTModel""", """ModalEmbeddings"""]
if TYPE_CHECKING:
from .configuration_mmbt import MMBTConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mmbt import MMBTForClassification, MMBTModel, ModalEmbeddings
else:
import sys
_UpperCAmelCase : Tuple = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 50
|
import tempfile
import unittest
from transformers import TaConfig, is_torch_available
from transformers.testing_utils import (
require_sentencepiece,
require_tokenizers,
require_torch,
slow,
torch_device,
)
from ...generation.test_utils import GenerationTesterMixin
from ...test_modeling_common import ModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import AutoTokenizer, UMTaForConditionalGeneration, UMTaForQuestionAnswering, UMTaModel
class lowerCAmelCase :
def __init__( self : Tuple , UpperCAmelCase : Dict , UpperCAmelCase : Union[str, Any]=99 , UpperCAmelCase : str=13 , UpperCAmelCase : List[str]=7 , UpperCAmelCase : str=9 , UpperCAmelCase : Optional[int]=True , UpperCAmelCase : str=True , UpperCAmelCase : Any=False , UpperCAmelCase : Union[str, Any]=32 , UpperCAmelCase : List[str]=5 , UpperCAmelCase : Tuple=4 , UpperCAmelCase : Union[str, Any]=37 , UpperCAmelCase : int=8 , UpperCAmelCase : List[str]=0.1 , UpperCAmelCase : Any=0.0_0_2 , UpperCAmelCase : Optional[Any]=1 , UpperCAmelCase : List[Any]=0 , UpperCAmelCase : Union[str, Any]=0 , UpperCAmelCase : Tuple=None , UpperCAmelCase : Optional[Any]=None , ) -> Union[str, Any]:
lowerCamelCase__ : int = parent
lowerCamelCase__ : Any = batch_size
lowerCamelCase__ : Optional[int] = encoder_seq_length
lowerCamelCase__ : int = decoder_seq_length
# For common tests
lowerCamelCase__ : List[str] = self.decoder_seq_length
lowerCamelCase__ : Optional[int] = is_training
lowerCamelCase__ : List[Any] = use_attention_mask
lowerCamelCase__ : Optional[Any] = use_labels
lowerCamelCase__ : Union[str, Any] = vocab_size
lowerCamelCase__ : Union[str, Any] = hidden_size
lowerCamelCase__ : Optional[Any] = num_hidden_layers
lowerCamelCase__ : Any = num_attention_heads
lowerCamelCase__ : str = d_ff
lowerCamelCase__ : Optional[Any] = relative_attention_num_buckets
lowerCamelCase__ : Any = dropout_rate
lowerCamelCase__ : Any = initializer_factor
lowerCamelCase__ : Union[str, Any] = eos_token_id
lowerCamelCase__ : List[str] = pad_token_id
lowerCamelCase__ : List[str] = decoder_start_token_id
lowerCamelCase__ : List[Any] = None
lowerCamelCase__ : Optional[Any] = decoder_layers
def A_ ( self : List[Any] ) -> int:
return TaConfig.from_pretrained('google/umt5-base' )
def A_ ( self : List[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[int] , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : Tuple=None , UpperCAmelCase : List[str]=None , UpperCAmelCase : str=None , UpperCAmelCase : Optional[Any]=None , UpperCAmelCase : Optional[Any]=None , ) -> List[str]:
if attention_mask is None:
lowerCamelCase__ : Optional[Any] = input_ids.ne(config.pad_token_id )
if decoder_attention_mask is None:
lowerCamelCase__ : Optional[Any] = decoder_input_ids.ne(config.pad_token_id )
if head_mask is None:
lowerCamelCase__ : int = torch.ones(config.num_hidden_layers , config.num_attention_heads , device=UpperCAmelCase )
if decoder_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
if cross_attn_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(
config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
return {
"input_ids": input_ids,
"decoder_input_ids": decoder_input_ids,
"attention_mask": attention_mask,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
"cross_attn_head_mask": cross_attn_head_mask,
}
def A_ ( self : str ) -> List[str]:
lowerCamelCase__ : Any = ids_tensor([self.batch_size, self.encoder_seq_length] , self.vocab_size )
lowerCamelCase__ : Optional[Any] = ids_tensor([self.batch_size, self.decoder_seq_length] , self.vocab_size )
# we need to clamp the input ids here to avoid having pad token in between
# this is because for NllbMoe the position_ids are prepared such that
# all pad tokens have pos id = 2 and rest are between 2..seq_length
# and the seq_length here is seq_length - num_pad_tokens
# but when using past, there is no way of knowing if the past input ids had
# pad tokens in them, which results in incorrect seq_lenth and which in turn results in
# position_ids being off by num_pad_tokens in past input
lowerCamelCase__ : List[str] = input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Union[str, Any] = decoder_input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Dict = self.get_config()
lowerCamelCase__ : Tuple = config.num_attention_heads
lowerCamelCase__ : Any = self.prepare_inputs_dict(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
return config, input_dict
def A_ ( self : Tuple ) -> Union[str, Any]:
lowerCamelCase__ , lowerCamelCase__ : Dict = self.prepare_config_and_inputs()
return config, inputs_dict
def A_ ( self : Optional[int] ) -> List[str]:
return TaConfig(
vocab_size=166 , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Union[str, Any] ) -> Dict:
return TaConfig(
vocab_size=self.vocab_size , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Any , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[str] , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Any , UpperCAmelCase : Dict , ) -> str:
lowerCamelCase__ : Dict = UMTaModel(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Optional[int] = model(
input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase , decoder_attention_mask=UpperCAmelCase , )
lowerCamelCase__ : Any = model(input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase )
lowerCamelCase__ : Dict = result.last_hidden_state
lowerCamelCase__ : Any = result.past_key_values
lowerCamelCase__ : List[Any] = result.encoder_last_hidden_state
self.parent.assertEqual(encoder_output.size() , (self.batch_size, self.encoder_seq_length, self.hidden_size) )
self.parent.assertEqual(decoder_output.size() , (self.batch_size, self.decoder_seq_length, self.hidden_size) )
# There should be `num_layers` key value embeddings stored in decoder_past
self.parent.assertEqual(len(UpperCAmelCase ) , config.num_layers )
# There should be a self attn key, a self attn value, a cross attn key and a cross attn value stored in each decoder_past tuple
self.parent.assertEqual(len(decoder_past[0] ) , 4 )
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[Any] , UpperCAmelCase : int , UpperCAmelCase : List[str] , UpperCAmelCase : Optional[Any] , ) -> Optional[int]:
lowerCamelCase__ : List[Any] = UMTaModel(config=UpperCAmelCase ).get_decoder().to(UpperCAmelCase ).eval()
# first forward pass
lowerCamelCase__ : Tuple = model(UpperCAmelCase , use_cache=UpperCAmelCase )
lowerCamelCase__ : List[Any] = model(UpperCAmelCase )
lowerCamelCase__ : int = model(UpperCAmelCase , use_cache=UpperCAmelCase )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) + 1 )
lowerCamelCase__ , lowerCamelCase__ : Dict = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
lowerCamelCase__ : Optional[int] = ids_tensor((self.batch_size, 1) , config.vocab_size )
# append to next input_ids and
lowerCamelCase__ : List[Any] = torch.cat([input_ids, next_tokens] , dim=-1 )
lowerCamelCase__ : List[str] = model(UpperCAmelCase )['last_hidden_state']
lowerCamelCase__ : str = model(UpperCAmelCase , past_key_values=UpperCAmelCase )['last_hidden_state']
# select random slice
lowerCamelCase__ : List[str] = ids_tensor((1,) , output_from_past.shape[-1] ).item()
lowerCamelCase__ : Tuple = output_from_no_past[:, -1, random_slice_idx].detach()
lowerCamelCase__ : List[str] = output_from_past[:, 0, random_slice_idx].detach()
# test that outputs are equal for slice
self.parent.assertTrue(torch.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-3 ) )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : List[str] , ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = UMTaModel(config=UpperCAmelCase ).to(UpperCAmelCase ).half().eval()
lowerCamelCase__ : Optional[int] = model(**UpperCAmelCase )['last_hidden_state']
self.parent.assertFalse(torch.isnan(UpperCAmelCase ).any().item() )
@require_torch
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(UMTaModel, UMTaForConditionalGeneration, UMTaForQuestionAnswering) if is_torch_available() else ()
)
UpperCAmelCase__ = (UMTaForConditionalGeneration,) if is_torch_available() else ()
UpperCAmelCase__ = (
{
"""conversational""": UMTaForConditionalGeneration,
"""feature-extraction""": UMTaModel,
"""summarization""": UMTaForConditionalGeneration,
"""text2text-generation""": UMTaForConditionalGeneration,
"""translation""": UMTaForConditionalGeneration,
"""question-answering""": UMTaForQuestionAnswering,
}
if is_torch_available()
else {}
)
UpperCAmelCase__ = True
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = True
UpperCAmelCase__ = True
# The small UMT5 model needs higher percentages for CPU/MP tests
UpperCAmelCase__ = [0.8, 0.9]
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = UMTaModelTester(self )
@unittest.skip('Test has a segmentation fault on torch 1.8.0' )
def A_ ( self : Tuple ) -> int:
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Tuple = UMTaModel(config_and_inputs[0] ).to(UpperCAmelCase )
with tempfile.TemporaryDirectory() as tmpdirname:
torch.onnx.export(
UpperCAmelCase , (config_and_inputs[1], config_and_inputs[3], config_and_inputs[2]) , F"""{tmpdirname}/t5_test.onnx""" , export_params=UpperCAmelCase , opset_version=9 , input_names=['input_ids', 'decoder_input_ids'] , )
@unittest.skipIf(torch_device == 'cpu' , 'Cant do half precision' )
def A_ ( self : Tuple ) -> Optional[Any]:
lowerCamelCase__ : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model_fpaa_forward(*UpperCAmelCase )
def A_ ( self : List[Any] ) -> str:
lowerCamelCase__ : int = ['encoder_attentions', 'decoder_attentions', 'cross_attentions']
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Any = config_and_inputs[0]
lowerCamelCase__ : Any = UMTaForConditionalGeneration(UpperCAmelCase ).eval()
model.to(UpperCAmelCase )
lowerCamelCase__ : Tuple = {
'head_mask': torch.zeros(config.num_layers , config.num_heads , device=UpperCAmelCase ),
'decoder_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
'cross_attn_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
}
for attn_name, (name, mask) in zip(UpperCAmelCase , head_masking.items() ):
lowerCamelCase__ : Union[str, Any] = {name: mask}
# Explicitly pass decoder_head_mask as it is required from T5 model when head_mask specified
if name == "head_mask":
lowerCamelCase__ : Union[str, Any] = torch.ones(
config.num_decoder_layers , config.num_heads , device=UpperCAmelCase )
lowerCamelCase__ : Tuple = model.generate(
config_and_inputs[1]['input_ids'] , num_beams=1 , max_length=3 , output_attentions=UpperCAmelCase , return_dict_in_generate=UpperCAmelCase , **UpperCAmelCase , )
# We check the state of decoder_attentions and cross_attentions just from the last step
lowerCamelCase__ : Union[str, Any] = out[attn_name] if attn_name == attention_names[0] else out[attn_name][-1]
self.assertEqual(sum([w.sum().item() for w in attn_weights] ) , 0.0 )
@unittest.skip('Does not work on the tiny model as we keep hitting edge cases.' )
def A_ ( self : Optional[Any] ) -> Optional[Any]:
pass
@require_torch
@require_sentencepiece
@require_tokenizers
class lowerCAmelCase ( unittest.TestCase ):
@slow
@unittest.skip(
'Unless we stop stripping left and right by default for all special tokens, the expected ids obtained here will not match the original ones. Wait for https://github.com/huggingface/transformers/pull/23909 to be merged' )
def A_ ( self : Any ) -> int:
lowerCamelCase__ : Optional[Any] = UMTaForConditionalGeneration.from_pretrained('google/umt5-small' , return_dict=UpperCAmelCase ).to(UpperCAmelCase )
lowerCamelCase__ : List[str] = AutoTokenizer.from_pretrained('google/umt5-small' , use_fast=UpperCAmelCase , legacy=UpperCAmelCase )
lowerCamelCase__ : Dict = [
'Bonjour monsieur <extra_id_0> bien <extra_id_1>.',
'No se como puedo <extra_id_0>.',
'This is the reason why we <extra_id_0> them.',
'The <extra_id_0> walks in <extra_id_1>, seats',
'A <extra_id_0> walks into a bar and orders a <extra_id_1> with <extra_id_2> pinch of <extra_id_3>.',
]
lowerCamelCase__ : Tuple = tokenizer(UpperCAmelCase , return_tensors='pt' , padding=UpperCAmelCase ).input_ids
# fmt: off
lowerCamelCase__ : Any = torch.tensor(
[
[ 38530, 210703, 256299, 1410, 256298, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 826, 321, 671, 25922, 256299, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 1460, 339, 312, 19014, 10620, 758, 256299, 2355,274, 1, 0, 0, 0, 0, 0, 0,0, 0],
[ 517, 256299, 14869, 281, 301, 256298, 275, 119983,1, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 320, 256299, 14869, 281, 2234, 289, 2275, 333,61391, 289, 256298, 543, 256297, 168714, 329, 256296,274, 1],
] )
# fmt: on
torch.testing.assert_allclose(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Optional[int] = model.generate(input_ids.to(UpperCAmelCase ) )
lowerCamelCase__ : List[Any] = [
'<pad><extra_id_0> et<extra_id_1> [eod] <extra_id_2><extra_id_55>.. [eod] 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 <extra_id_56>ajšietosto<extra_id_56>lleux<extra_id_19><extra_id_6>ajšie</s>',
'<pad><extra_id_0>.<extra_id_1>.,<0x0A>...spech <0x0A><extra_id_20> <extra_id_21></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> are not going to be a part of the world. We are not going to be a part of<extra_id_1> and<extra_id_2><0x0A><extra_id_48>.<extra_id_48></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> door<extra_id_1>, the door<extra_id_2> 피해[/</s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0>nyone who<extra_id_1> drink<extra_id_2> a<extra_id_3> alcohol<extra_id_4> A<extra_id_5> A. This<extra_id_6> I<extra_id_7><extra_id_52><extra_id_53></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
]
lowerCamelCase__ : Union[str, Any] = tokenizer.batch_decode(UpperCAmelCase )
self.assertEqual(UpperCAmelCase , UpperCAmelCase )
| 50
| 1
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_UpperCAmelCase : Any = logging.get_logger(__name__)
_UpperCAmelCase : Dict = {
"""vinvino02/glpn-kitti""": """https://huggingface.co/vinvino02/glpn-kitti/resolve/main/config.json""",
# See all GLPN models at https://huggingface.co/models?filter=glpn
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """glpn"""
def __init__( self : int , UpperCAmelCase : List[str]=3 , UpperCAmelCase : Tuple=4 , UpperCAmelCase : List[Any]=[2, 2, 2, 2] , UpperCAmelCase : int=[8, 4, 2, 1] , UpperCAmelCase : Tuple=[32, 64, 160, 256] , UpperCAmelCase : str=[7, 3, 3, 3] , UpperCAmelCase : Union[str, Any]=[4, 2, 2, 2] , UpperCAmelCase : List[Any]=[1, 2, 5, 8] , UpperCAmelCase : Tuple=[4, 4, 4, 4] , UpperCAmelCase : Optional[int]="gelu" , UpperCAmelCase : Any=0.0 , UpperCAmelCase : Any=0.0 , UpperCAmelCase : int=0.0_2 , UpperCAmelCase : str=0.1 , UpperCAmelCase : List[str]=1e-6 , UpperCAmelCase : Dict=64 , UpperCAmelCase : Optional[int]=10 , UpperCAmelCase : Any=-1 , **UpperCAmelCase : List[Any] , ) -> str:
super().__init__(**UpperCAmelCase )
lowerCamelCase__ : int = num_channels
lowerCamelCase__ : str = num_encoder_blocks
lowerCamelCase__ : Any = depths
lowerCamelCase__ : int = sr_ratios
lowerCamelCase__ : List[Any] = hidden_sizes
lowerCamelCase__ : List[str] = patch_sizes
lowerCamelCase__ : int = strides
lowerCamelCase__ : List[str] = mlp_ratios
lowerCamelCase__ : Union[str, Any] = num_attention_heads
lowerCamelCase__ : List[Any] = hidden_act
lowerCamelCase__ : Dict = hidden_dropout_prob
lowerCamelCase__ : Dict = attention_probs_dropout_prob
lowerCamelCase__ : Union[str, Any] = initializer_range
lowerCamelCase__ : Optional[Any] = drop_path_rate
lowerCamelCase__ : int = layer_norm_eps
lowerCamelCase__ : List[Any] = decoder_hidden_size
lowerCamelCase__ : List[Any] = max_depth
lowerCamelCase__ : Tuple = head_in_index
| 50
|
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=None , _UpperCAmelCase="no" , _UpperCAmelCase="29500" ) -> Tuple:
lowerCamelCase__ : Dict = False
lowerCamelCase__ : Dict = False
if any(key.startswith('KAGGLE' ) for key in os.environ.keys() ):
lowerCamelCase__ : Optional[Any] = True
elif "IPython" in sys.modules:
lowerCamelCase__ : Optional[Any] = 'google.colab' in str(sys.modules['IPython'].get_ipython() )
try:
lowerCamelCase__ : List[str] = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
F"""Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}.""" )
if (in_colab or in_kaggle) and (os.environ.get('TPU_NAME' , _UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '
'your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if num_processes is None:
lowerCamelCase__ : Optional[Any] = 8
lowerCamelCase__ : List[str] = PrepareForLaunch(_UpperCAmelCase , distributed_type='TPU' )
print(F"""Launching a training on {num_processes} TPU cores.""" )
xmp.spawn(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on one CPU.' )
function(*_UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '
'inside your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if torch.cuda.is_initialized():
raise ValueError(
'To launch a multi-GPU training from your notebook, you need to avoid running any instruction '
'using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '
'function.' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port=_UpperCAmelCase , mixed_precision=_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = PrepareForLaunch(_UpperCAmelCase , distributed_type='MULTI_GPU' )
print(F"""Launching training on {num_processes} GPUs.""" )
try:
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '
'This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '
'Please review your imports and test them when running the `notebook_launcher()` to identify '
'which one is problematic.' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
lowerCamelCase__ : int = '1'
print('Launching training on MPS.' )
elif torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on CPU.' )
function(*_UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=2 ) -> Optional[Any]:
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port='29500' , accelerate_mixed_precision='no' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='yes' , ):
lowerCamelCase__ : Optional[Any] = PrepareForLaunch(_UpperCAmelCase , debug=_UpperCAmelCase )
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
| 50
| 1
|
import logging
import os
from dataclasses import dataclass
from typing import List, Optional, Union
import tqdm
from filelock import FileLock
from transformers import (
BartTokenizer,
BartTokenizerFast,
DataProcessor,
PreTrainedTokenizer,
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
is_tf_available,
is_torch_available,
)
_UpperCAmelCase : Optional[Any] = logging.getLogger(__name__)
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
if is_torch_available():
import torch
from torch.utils.data import Dataset
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
def __init__( self : int , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : List[str]=False , UpperCAmelCase : bool = False , ) -> List[str]:
lowerCamelCase__ : int = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = os.path.join(
UpperCAmelCase , 'cached_{}_{}_{}_{}'.format(
'dev' if evaluate else 'train' , tokenizer.__class__.__name__ , str(UpperCAmelCase ) , UpperCAmelCase , ) , )
lowerCamelCase__ : int = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = label_list[2], label_list[1]
lowerCamelCase__ : List[str] = label_list
# Make sure only the first process in distributed training processes the dataset,
# and the others will use the cache.
lowerCamelCase__ : str = cached_features_file + '.lock'
with FileLock(UpperCAmelCase ):
if os.path.exists(UpperCAmelCase ) and not overwrite_cache:
logger.info(F"""Loading features from cached file {cached_features_file}""" )
lowerCamelCase__ : int = torch.load(UpperCAmelCase )
else:
logger.info(F"""Creating features from dataset file at {data_dir}""" )
lowerCamelCase__ : str = (
processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
)
logger.info('Training examples: %s' , len(UpperCAmelCase ) )
lowerCamelCase__ : Dict = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
logger.info('Saving features into cached file %s' , UpperCAmelCase )
torch.save(self.features , UpperCAmelCase )
def __len__( self : Optional[int] ) -> Optional[Any]:
return len(self.features )
def __getitem__( self : Tuple , UpperCAmelCase : Dict ) -> InputFeatures:
return self.features[i]
def A_ ( self : int ) -> int:
return self.label_list
if is_tf_available():
import tensorflow as tf
class lowerCAmelCase :
UpperCAmelCase__ = 42
def __init__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = 128 , UpperCAmelCase : Any=False , UpperCAmelCase : bool = False , ) -> Union[str, Any]:
lowerCamelCase__ : Any = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : str = label_list[2], label_list[1]
lowerCamelCase__ : Optional[int] = label_list
lowerCamelCase__ : int = processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def gen():
for ex_index, ex in tqdm.tqdm(enumerate(self.features ) , desc='convert examples to features' ):
if ex_index % 10000 == 0:
logger.info('Writing example %d of %d' % (ex_index, len(UpperCAmelCase )) )
yield (
{
"example_id": 0,
"input_ids": ex.input_ids,
"attention_mask": ex.attention_mask,
"token_type_ids": ex.token_type_ids,
},
ex.label,
)
lowerCamelCase__ : Optional[int] = tf.data.Dataset.from_generator(
UpperCAmelCase , (
{
'example_id': tf.intaa,
'input_ids': tf.intaa,
'attention_mask': tf.intaa,
'token_type_ids': tf.intaa,
},
tf.intaa,
) , (
{
'example_id': tf.TensorShape([] ),
'input_ids': tf.TensorShape([None, None] ),
'attention_mask': tf.TensorShape([None, None] ),
'token_type_ids': tf.TensorShape([None, None] ),
},
tf.TensorShape([] ),
) , )
def A_ ( self : Any ) -> Any:
return self.dataset
def __len__( self : Tuple ) -> int:
return len(self.features )
def __getitem__( self : List[str] , UpperCAmelCase : Any ) -> InputFeatures:
return self.features[i]
def A_ ( self : Dict ) -> str:
return self.label_list
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : int , UpperCAmelCase : List[Any] ) -> int:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_train_set.txt' ) ) , 'train' )
def A_ ( self : Any , UpperCAmelCase : int ) -> List[Any]:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_evaluation_set.txt' ) ) , 'dev' )
def A_ ( self : Any ) -> List[Any]:
return ["contradiction", "entailment", "neutral"]
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : List[str] ) -> List[str]:
lowerCamelCase__ : List[str] = []
for i, line in enumerate(UpperCAmelCase ):
if i == 0:
continue
lowerCamelCase__ : Tuple = '%s-%s' % (set_type, line[0])
lowerCamelCase__ : str = line[5]
lowerCamelCase__ : Dict = line[6]
lowerCamelCase__ : int = line[7][2:] if line[7].startswith('ex' ) else line[7]
lowerCamelCase__ : Dict = line[0]
examples.append(InputExample(guid=UpperCAmelCase , text_a=UpperCAmelCase , text_b=UpperCAmelCase , label=UpperCAmelCase , pairID=UpperCAmelCase ) )
return examples
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ) -> Optional[int]:
lowerCamelCase__ : int = {label: i for i, label in enumerate(_UpperCAmelCase )}
lowerCamelCase__ : List[Any] = []
for ex_index, example in tqdm.tqdm(enumerate(_UpperCAmelCase ) , desc='convert examples to features' ):
if ex_index % 1_0000 == 0:
logger.info('Writing example %d' % (ex_index) )
lowerCamelCase__ : List[Any] = tokenizer(
example.text_a , example.text_b , add_special_tokens=_UpperCAmelCase , max_length=_UpperCAmelCase , padding='max_length' , truncation=_UpperCAmelCase , return_overflowing_tokens=_UpperCAmelCase , )
lowerCamelCase__ : List[str] = label_map[example.label] if example.label in label_map else 0
lowerCamelCase__ : Optional[int] = int(example.pairID )
features.append(InputFeatures(**_UpperCAmelCase , label=_UpperCAmelCase , pairID=_UpperCAmelCase ) )
for i, example in enumerate(examples[:5] ):
logger.info('*** Example ***' )
logger.info(F"""guid: {example}""" )
logger.info(F"""features: {features[i]}""" )
return features
_UpperCAmelCase : str = {
"""hans""": 3,
}
_UpperCAmelCase : List[Any] = {
"""hans""": HansProcessor,
}
| 50
|
from dataclasses import dataclass
from typing import Optional, Tuple, Union
import torch
import torch.nn as nn
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import BaseOutput
from .embeddings import GaussianFourierProjection, TimestepEmbedding, Timesteps
from .modeling_utils import ModelMixin
from .unet_ad_blocks import get_down_block, get_mid_block, get_out_block, get_up_block
@dataclass
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase ):
@register_to_config
def __init__( self : List[str] , UpperCAmelCase : int = 65536 , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 0 , UpperCAmelCase : str = "fourier" , UpperCAmelCase : bool = True , UpperCAmelCase : bool = False , UpperCAmelCase : float = 0.0 , UpperCAmelCase : Tuple[str] = ("DownBlock1DNoSkip", "DownBlock1D", "AttnDownBlock1D") , UpperCAmelCase : Tuple[str] = ("AttnUpBlock1D", "UpBlock1D", "UpBlock1DNoSkip") , UpperCAmelCase : Tuple[str] = "UNetMidBlock1D" , UpperCAmelCase : str = None , UpperCAmelCase : Tuple[int] = (32, 32, 64) , UpperCAmelCase : str = None , UpperCAmelCase : int = 8 , UpperCAmelCase : int = 1 , UpperCAmelCase : bool = False , ) -> List[Any]:
super().__init__()
lowerCamelCase__ : Optional[int] = sample_size
# time
if time_embedding_type == "fourier":
lowerCamelCase__ : Optional[Any] = GaussianFourierProjection(
embedding_size=8 , set_W_to_weight=UpperCAmelCase , log=UpperCAmelCase , flip_sin_to_cos=UpperCAmelCase )
lowerCamelCase__ : Any = 2 * block_out_channels[0]
elif time_embedding_type == "positional":
lowerCamelCase__ : List[Any] = Timesteps(
block_out_channels[0] , flip_sin_to_cos=UpperCAmelCase , downscale_freq_shift=UpperCAmelCase )
lowerCamelCase__ : Dict = block_out_channels[0]
if use_timestep_embedding:
lowerCamelCase__ : str = block_out_channels[0] * 4
lowerCamelCase__ : List[Any] = TimestepEmbedding(
in_channels=UpperCAmelCase , time_embed_dim=UpperCAmelCase , act_fn=UpperCAmelCase , out_dim=block_out_channels[0] , )
lowerCamelCase__ : Any = nn.ModuleList([] )
lowerCamelCase__ : Tuple = None
lowerCamelCase__ : List[str] = nn.ModuleList([] )
lowerCamelCase__ : Optional[int] = None
# down
lowerCamelCase__ : Optional[int] = in_channels
for i, down_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Union[str, Any] = output_channel
lowerCamelCase__ : Tuple = block_out_channels[i]
if i == 0:
input_channel += extra_in_channels
lowerCamelCase__ : Union[str, Any] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Optional[int] = get_down_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_downsample=not is_final_block or downsample_each_block , )
self.down_blocks.append(UpperCAmelCase )
# mid
lowerCamelCase__ : Optional[int] = get_mid_block(
UpperCAmelCase , in_channels=block_out_channels[-1] , mid_channels=block_out_channels[-1] , out_channels=block_out_channels[-1] , embed_dim=block_out_channels[0] , num_layers=UpperCAmelCase , add_downsample=UpperCAmelCase , )
# up
lowerCamelCase__ : Optional[int] = list(reversed(UpperCAmelCase ) )
lowerCamelCase__ : Optional[int] = reversed_block_out_channels[0]
if out_block_type is None:
lowerCamelCase__ : List[str] = out_channels
else:
lowerCamelCase__ : Any = block_out_channels[0]
for i, up_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Tuple = output_channel
lowerCamelCase__ : Union[str, Any] = (
reversed_block_out_channels[i + 1] if i < len(UpperCAmelCase ) - 1 else final_upsample_channels
)
lowerCamelCase__ : List[str] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Dict = get_up_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_upsample=not is_final_block , )
self.up_blocks.append(UpperCAmelCase )
lowerCamelCase__ : int = output_channel
# out
lowerCamelCase__ : int = norm_num_groups if norm_num_groups is not None else min(block_out_channels[0] // 4 , 32 )
lowerCamelCase__ : List[Any] = get_out_block(
out_block_type=UpperCAmelCase , num_groups_out=UpperCAmelCase , embed_dim=block_out_channels[0] , out_channels=UpperCAmelCase , act_fn=UpperCAmelCase , fc_dim=block_out_channels[-1] // 4 , )
def A_ ( self : List[Any] , UpperCAmelCase : torch.FloatTensor , UpperCAmelCase : Union[torch.Tensor, float, int] , UpperCAmelCase : bool = True , ) -> Union[UNetaDOutput, Tuple]:
lowerCamelCase__ : Optional[Any] = timestep
if not torch.is_tensor(UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = torch.tensor([timesteps] , dtype=torch.long , device=sample.device )
elif torch.is_tensor(UpperCAmelCase ) and len(timesteps.shape ) == 0:
lowerCamelCase__ : List[str] = timesteps[None].to(sample.device )
lowerCamelCase__ : Optional[int] = self.time_proj(UpperCAmelCase )
if self.config.use_timestep_embedding:
lowerCamelCase__ : str = self.time_mlp(UpperCAmelCase )
else:
lowerCamelCase__ : List[str] = timestep_embed[..., None]
lowerCamelCase__ : str = timestep_embed.repeat([1, 1, sample.shape[2]] ).to(sample.dtype )
lowerCamelCase__ : str = timestep_embed.broadcast_to((sample.shape[:1] + timestep_embed.shape[1:]) )
# 2. down
lowerCamelCase__ : str = ()
for downsample_block in self.down_blocks:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = downsample_block(hidden_states=UpperCAmelCase , temb=UpperCAmelCase )
down_block_res_samples += res_samples
# 3. mid
if self.mid_block:
lowerCamelCase__ : Optional[Any] = self.mid_block(UpperCAmelCase , UpperCAmelCase )
# 4. up
for i, upsample_block in enumerate(self.up_blocks ):
lowerCamelCase__ : Dict = down_block_res_samples[-1:]
lowerCamelCase__ : Optional[Any] = down_block_res_samples[:-1]
lowerCamelCase__ : Any = upsample_block(UpperCAmelCase , res_hidden_states_tuple=UpperCAmelCase , temb=UpperCAmelCase )
# 5. post-process
if self.out_block:
lowerCamelCase__ : Any = self.out_block(UpperCAmelCase , UpperCAmelCase )
if not return_dict:
return (sample,)
return UNetaDOutput(sample=UpperCAmelCase )
| 50
| 1
|
from typing import Any
class lowerCAmelCase :
def __init__( self : int , UpperCAmelCase : Any ) -> Dict:
lowerCamelCase__ : Optional[int] = data
lowerCamelCase__ : str = None
def __repr__( self : Any ) -> str:
return F"""Node({self.data})"""
class lowerCAmelCase :
def __init__( self : Tuple ) -> Any:
lowerCamelCase__ : Optional[Any] = None
def __iter__( self : Any ) -> Any:
lowerCamelCase__ : Any = self.head
while node:
yield node.data
lowerCamelCase__ : List[Any] = node.next
def __len__( self : Optional[Any] ) -> int:
return sum(1 for _ in self )
def __repr__( self : Tuple ) -> str:
return "->".join([str(UpperCAmelCase ) for item in self] )
def __getitem__( self : str , UpperCAmelCase : int ) -> Any:
if not 0 <= index < len(self ):
raise ValueError('list index out of range.' )
for i, node in enumerate(self ):
if i == index:
return node
return None
def __setitem__( self : Any , UpperCAmelCase : int , UpperCAmelCase : Any ) -> None:
if not 0 <= index < len(self ):
raise ValueError('list index out of range.' )
lowerCamelCase__ : List[str] = self.head
for _ in range(UpperCAmelCase ):
lowerCamelCase__ : Any = current.next
lowerCamelCase__ : Optional[int] = data
def A_ ( self : Any , UpperCAmelCase : Any ) -> None:
self.insert_nth(len(self ) , UpperCAmelCase )
def A_ ( self : List[str] , UpperCAmelCase : Any ) -> None:
self.insert_nth(0 , UpperCAmelCase )
def A_ ( self : Optional[Any] , UpperCAmelCase : int , UpperCAmelCase : Any ) -> None:
if not 0 <= index <= len(self ):
raise IndexError('list index out of range' )
lowerCamelCase__ : Union[str, Any] = Node(UpperCAmelCase )
if self.head is None:
lowerCamelCase__ : Dict = new_node
elif index == 0:
lowerCamelCase__ : Dict = self.head # link new_node to head
lowerCamelCase__ : Optional[int] = new_node
else:
lowerCamelCase__ : Optional[int] = self.head
for _ in range(index - 1 ):
lowerCamelCase__ : int = temp.next
lowerCamelCase__ : int = temp.next
lowerCamelCase__ : str = new_node
def A_ ( self : Tuple ) -> None: # print every node data
print(self )
def A_ ( self : Optional[int] ) -> Any:
return self.delete_nth(0 )
def A_ ( self : Any ) -> Any: # delete from tail
return self.delete_nth(len(self ) - 1 )
def A_ ( self : List[Any] , UpperCAmelCase : int = 0 ) -> Any:
if not 0 <= index <= len(self ) - 1: # test if index is valid
raise IndexError('List index out of range.' )
lowerCamelCase__ : Dict = self.head # default first node
if index == 0:
lowerCamelCase__ : str = self.head.next
else:
lowerCamelCase__ : Tuple = self.head
for _ in range(index - 1 ):
lowerCamelCase__ : Tuple = temp.next
lowerCamelCase__ : List[str] = temp.next
lowerCamelCase__ : Dict = temp.next.next
return delete_node.data
def A_ ( self : Optional[int] ) -> bool:
return self.head is None
def A_ ( self : Union[str, Any] ) -> None:
lowerCamelCase__ : str = None
lowerCamelCase__ : Optional[Any] = self.head
while current:
# Store the current node's next node.
lowerCamelCase__ : Union[str, Any] = current.next
# Make the current node's next point backwards
lowerCamelCase__ : Dict = prev
# Make the previous node be the current node
lowerCamelCase__ : Optional[Any] = current
# Make the current node the next node (to progress iteration)
lowerCamelCase__ : str = next_node
# Return prev in order to put the head at the end
lowerCamelCase__ : List[str] = prev
def SCREAMING_SNAKE_CASE ( ) -> None:
lowerCamelCase__ : Tuple = LinkedList()
assert linked_list.is_empty() is True
assert str(_UpperCAmelCase ) == ""
try:
linked_list.delete_head()
raise AssertionError # This should not happen.
except IndexError:
assert True # This should happen.
try:
linked_list.delete_tail()
raise AssertionError # This should not happen.
except IndexError:
assert True # This should happen.
for i in range(10 ):
assert len(_UpperCAmelCase ) == i
linked_list.insert_nth(_UpperCAmelCase , i + 1 )
assert str(_UpperCAmelCase ) == "->".join(str(_UpperCAmelCase ) for i in range(1 , 11 ) )
linked_list.insert_head(0 )
linked_list.insert_tail(11 )
assert str(_UpperCAmelCase ) == "->".join(str(_UpperCAmelCase ) for i in range(0 , 12 ) )
assert linked_list.delete_head() == 0
assert linked_list.delete_nth(9 ) == 10
assert linked_list.delete_tail() == 11
assert len(_UpperCAmelCase ) == 9
assert str(_UpperCAmelCase ) == "->".join(str(_UpperCAmelCase ) for i in range(1 , 10 ) )
assert all(linked_list[i] == i + 1 for i in range(0 , 9 ) ) is True
for i in range(0 , 9 ):
lowerCamelCase__ : List[Any] = -i
assert all(linked_list[i] == -i for i in range(0 , 9 ) ) is True
linked_list.reverse()
assert str(_UpperCAmelCase ) == "->".join(str(_UpperCAmelCase ) for i in range(-8 , 1 ) )
def SCREAMING_SNAKE_CASE ( ) -> None:
lowerCamelCase__ : Union[str, Any] = [
-9,
100,
Node(7734_5112 ),
'dlrow olleH',
7,
5555,
0,
-192.55_555,
'Hello, world!',
77.9,
Node(10 ),
None,
None,
12.20,
]
lowerCamelCase__ : List[Any] = LinkedList()
for i in test_input:
linked_list.insert_tail(_UpperCAmelCase )
# Check if it's empty or not
assert linked_list.is_empty() is False
assert (
str(_UpperCAmelCase ) == "-9->100->Node(77345112)->dlrow olleH->7->5555->0->"
"-192.55555->Hello, world!->77.9->Node(10)->None->None->12.2"
)
# Delete the head
lowerCamelCase__ : str = linked_list.delete_head()
assert result == -9
assert (
str(_UpperCAmelCase ) == "100->Node(77345112)->dlrow olleH->7->5555->0->-192.55555->"
"Hello, world!->77.9->Node(10)->None->None->12.2"
)
# Delete the tail
lowerCamelCase__ : List[Any] = linked_list.delete_tail()
assert result == 12.2
assert (
str(_UpperCAmelCase ) == "100->Node(77345112)->dlrow olleH->7->5555->0->-192.55555->"
"Hello, world!->77.9->Node(10)->None->None"
)
# Delete a node in specific location in linked list
lowerCamelCase__ : Any = linked_list.delete_nth(10 )
assert result is None
assert (
str(_UpperCAmelCase ) == "100->Node(77345112)->dlrow olleH->7->5555->0->-192.55555->"
"Hello, world!->77.9->Node(10)->None"
)
# Add a Node instance to its head
linked_list.insert_head(Node('Hello again, world!' ) )
assert (
str(_UpperCAmelCase )
== "Node(Hello again, world!)->100->Node(77345112)->dlrow olleH->"
"7->5555->0->-192.55555->Hello, world!->77.9->Node(10)->None"
)
# Add None to its tail
linked_list.insert_tail(_UpperCAmelCase )
assert (
str(_UpperCAmelCase )
== "Node(Hello again, world!)->100->Node(77345112)->dlrow olleH->"
"7->5555->0->-192.55555->Hello, world!->77.9->Node(10)->None->None"
)
# Reverse the linked list
linked_list.reverse()
assert (
str(_UpperCAmelCase )
== "None->None->Node(10)->77.9->Hello, world!->-192.55555->0->5555->"
"7->dlrow olleH->Node(77345112)->100->Node(Hello again, world!)"
)
def SCREAMING_SNAKE_CASE ( ) -> Any:
from doctest import testmod
testmod()
lowerCamelCase__ : str = LinkedList()
linked_list.insert_head(input('Inserting 1st at head ' ).strip() )
linked_list.insert_head(input('Inserting 2nd at head ' ).strip() )
print('\nPrint list:' )
linked_list.print_list()
linked_list.insert_tail(input('\nInserting 1st at tail ' ).strip() )
linked_list.insert_tail(input('Inserting 2nd at tail ' ).strip() )
print('\nPrint list:' )
linked_list.print_list()
print('\nDelete head' )
linked_list.delete_head()
print('Delete tail' )
linked_list.delete_tail()
print('\nPrint list:' )
linked_list.print_list()
print('\nReverse linked list' )
linked_list.reverse()
print('\nPrint list:' )
linked_list.print_list()
print('\nString representation of linked list:' )
print(_UpperCAmelCase )
print('\nReading/changing Node data using indexing:' )
print(F"""Element at Position 1: {linked_list[1]}""" )
lowerCamelCase__ : int = input('Enter New Value: ' ).strip()
print('New list:' )
print(_UpperCAmelCase )
print(F"""length of linked_list is : {len(_UpperCAmelCase )}""" )
if __name__ == "__main__":
main()
| 50
|
from __future__ import annotations
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> list[tuple[int, int]]:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = position
lowerCamelCase__ : Optional[Any] = [
(y + 1, x + 2),
(y - 1, x + 2),
(y + 1, x - 2),
(y - 1, x - 2),
(y + 2, x + 1),
(y + 2, x - 1),
(y - 2, x + 1),
(y - 2, x - 1),
]
lowerCamelCase__ : Dict = []
for position in positions:
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = position
if 0 <= y_test < n and 0 <= x_test < n:
permissible_positions.append(_UpperCAmelCase )
return permissible_positions
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
return not any(elem == 0 for row in board for elem in row )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> bool:
if is_complete(_UpperCAmelCase ):
return True
for position in get_valid_pos(_UpperCAmelCase , len(_UpperCAmelCase ) ):
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = position
if board[y][x] == 0:
lowerCamelCase__ : List[Any] = curr + 1
if open_knight_tour_helper(_UpperCAmelCase , _UpperCAmelCase , curr + 1 ):
return True
lowerCamelCase__ : Optional[Any] = 0
return False
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list[list[int]]:
lowerCamelCase__ : Any = [[0 for i in range(_UpperCAmelCase )] for j in range(_UpperCAmelCase )]
for i in range(_UpperCAmelCase ):
for j in range(_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = 1
if open_knight_tour_helper(_UpperCAmelCase , (i, j) , 1 ):
return board
lowerCamelCase__ : Optional[Any] = 0
lowerCamelCase__ : Any = F"""Open Kight Tour cannot be performed on a board of size {n}"""
raise ValueError(_UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
import json
import os
import unittest
from transformers.models.xlm.tokenization_xlm import VOCAB_FILES_NAMES, XLMTokenizer
from transformers.testing_utils import slow
from ...test_tokenization_common import TokenizerTesterMixin
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = XLMTokenizer
UpperCAmelCase__ = False
def A_ ( self : str ) -> List[str]:
super().setUp()
# Adapted from Sennrich et al. 2015 and https://github.com/rsennrich/subword-nmt
lowerCamelCase__ : Any = [
'l',
'o',
'w',
'e',
'r',
's',
't',
'i',
'd',
'n',
'w</w>',
'r</w>',
't</w>',
'lo',
'low',
'er</w>',
'low</w>',
'lowest</w>',
'newer</w>',
'wider</w>',
'<unk>',
]
lowerCamelCase__ : List[str] = dict(zip(UpperCAmelCase , range(len(UpperCAmelCase ) ) ) )
lowerCamelCase__ : Tuple = ['l o 123', 'lo w 1456', 'e r</w> 1789', '']
lowerCamelCase__ : Optional[Any] = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['vocab_file'] )
lowerCamelCase__ : int = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['merges_file'] )
with open(self.vocab_file , 'w' ) as fp:
fp.write(json.dumps(UpperCAmelCase ) )
with open(self.merges_file , 'w' ) as fp:
fp.write('\n'.join(UpperCAmelCase ) )
def A_ ( self : int , UpperCAmelCase : Dict ) -> Any:
lowerCamelCase__ : Tuple = 'lower newer'
lowerCamelCase__ : Tuple = 'lower newer'
return input_text, output_text
def A_ ( self : Optional[Any] ) -> Optional[int]:
lowerCamelCase__ : Dict = XLMTokenizer(self.vocab_file , self.merges_file )
lowerCamelCase__ : Tuple = 'lower'
lowerCamelCase__ : Any = ['low', 'er</w>']
lowerCamelCase__ : Dict = tokenizer.tokenize(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tokens + ['<unk>']
lowerCamelCase__ : Optional[int] = [14, 15, 20]
self.assertListEqual(tokenizer.convert_tokens_to_ids(UpperCAmelCase ) , UpperCAmelCase )
@slow
def A_ ( self : Optional[Any] ) -> List[Any]:
lowerCamelCase__ : str = XLMTokenizer.from_pretrained('xlm-mlm-en-2048' )
lowerCamelCase__ : List[str] = tokenizer.encode('sequence builders' , add_special_tokens=UpperCAmelCase )
lowerCamelCase__ : List[Any] = tokenizer.encode('multi-sequence build' , add_special_tokens=UpperCAmelCase )
lowerCamelCase__ : int = tokenizer.build_inputs_with_special_tokens(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = tokenizer.build_inputs_with_special_tokens(UpperCAmelCase , UpperCAmelCase )
assert encoded_sentence == [0] + text + [1]
assert encoded_pair == [0] + text + [1] + text_a + [1]
| 50
|
import argparse
import json
from collections import OrderedDict
import torch
from huggingface_hub import cached_download, hf_hub_url
from transformers import AutoImageProcessor, CvtConfig, CvtForImageClassification
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> int:
lowerCamelCase__ : Optional[int] = []
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.weight""",
F"""stage{idx}.patch_embed.proj.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.bias""",
F"""stage{idx}.patch_embed.proj.bias""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.weight""",
F"""stage{idx}.patch_embed.norm.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.bias""",
F"""stage{idx}.patch_embed.norm.bias""",
) )
return embed
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Tuple = []
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj.bias""",
) )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.weight""", F"""stage{idx}.blocks.{cnt}.norm1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.bias""", F"""stage{idx}.blocks.{cnt}.norm1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.weight""", F"""stage{idx}.blocks.{cnt}.norm2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.bias""", F"""stage{idx}.blocks.{cnt}.norm2.bias""") )
return attention_weights
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = []
token.append((F"""cvt.encoder.stages.{idx}.cls_token""", 'stage2.cls_token') )
return token
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : str = []
head.append(('layernorm.weight', 'norm.weight') )
head.append(('layernorm.bias', 'norm.bias') )
head.append(('classifier.weight', 'head.weight') )
head.append(('classifier.bias', 'head.bias') )
return head
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
lowerCamelCase__ : Tuple = 'imagenet-1k-id2label.json'
lowerCamelCase__ : Union[str, Any] = 1000
lowerCamelCase__ : Optional[Any] = 'huggingface/label-files'
lowerCamelCase__ : Any = num_labels
lowerCamelCase__ : Dict = json.load(open(cached_download(hf_hub_url(_UpperCAmelCase , _UpperCAmelCase , repo_type='dataset' ) ) , 'r' ) )
lowerCamelCase__ : int = {int(_UpperCAmelCase ): v for k, v in idalabel.items()}
lowerCamelCase__ : Tuple = idalabel
lowerCamelCase__ : List[Any] = {v: k for k, v in idalabel.items()}
lowerCamelCase__ : List[str] = CvtConfig(num_labels=_UpperCAmelCase , idalabel=_UpperCAmelCase , labelaid=_UpperCAmelCase )
# For depth size 13 (13 = 1+2+10)
if cvt_model.rsplit('/' , 1 )[-1][4:6] == "13":
lowerCamelCase__ : List[Any] = [1, 2, 10]
# For depth size 21 (21 = 1+4+16)
elif cvt_model.rsplit('/' , 1 )[-1][4:6] == "21":
lowerCamelCase__ : Dict = [1, 4, 16]
# For wide cvt (similar to wide-resnet) depth size 24 (w24 = 2 + 2 20)
else:
lowerCamelCase__ : Optional[Any] = [2, 2, 20]
lowerCamelCase__ : Optional[int] = [3, 12, 16]
lowerCamelCase__ : str = [192, 768, 1024]
lowerCamelCase__ : Any = CvtForImageClassification(_UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = AutoImageProcessor.from_pretrained('facebook/convnext-base-224-22k-1k' )
lowerCamelCase__ : Tuple = image_size
lowerCamelCase__ : List[str] = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
lowerCamelCase__ : Optional[int] = OrderedDict()
lowerCamelCase__ : Tuple = []
for idx in range(len(config.depth ) ):
if config.cls_token[idx]:
lowerCamelCase__ : Optional[Any] = list_of_state_dict + cls_token(_UpperCAmelCase )
lowerCamelCase__ : str = list_of_state_dict + embeddings(_UpperCAmelCase )
for cnt in range(config.depth[idx] ):
lowerCamelCase__ : str = list_of_state_dict + attention(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : int = list_of_state_dict + final()
for gg in list_of_state_dict:
print(_UpperCAmelCase )
for i in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : str = original_weights[list_of_state_dict[i][1]]
model.load_state_dict(_UpperCAmelCase )
model.save_pretrained(_UpperCAmelCase )
image_processor.save_pretrained(_UpperCAmelCase )
# Download the weights from zoo: https://1drv.ms/u/s!AhIXJn_J-blW9RzF3rMW7SsLHa8h?e=blQ0Al
if __name__ == "__main__":
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
parser.add_argument(
"""--cvt_model""",
default="""cvt-w24""",
type=str,
help="""Name of the cvt model you'd like to convert.""",
)
parser.add_argument(
"""--image_size""",
default=3_84,
type=int,
help="""Input Image Size""",
)
parser.add_argument(
"""--cvt_file_name""",
default=R"""cvtmodels\CvT-w24-384x384-IN-22k.pth""",
type=str,
help="""Input Image Size""",
)
parser.add_argument(
"""--pytorch_dump_folder_path""", default=None, type=str, help="""Path to the output PyTorch model directory."""
)
_UpperCAmelCase : List[str] = parser.parse_args()
convert_cvt_checkpoint(args.cvt_model, args.image_size, args.cvt_file_name, args.pytorch_dump_folder_path)
| 50
| 1
|
import argparse
import os
import evaluate
import torch
from datasets import load_dataset
from torch.optim import AdamW
from torch.utils.data import DataLoader
from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed
from accelerate import Accelerator, DistributedType
########################################################################
# This is a fully working simple example to use Accelerate,
# specifically showcasing the experiment tracking capability,
# and builds off the `nlp_example.py` script.
#
# This example trains a Bert base model on GLUE MRPC
# in any of the following settings (with the same script):
# - single CPU or single GPU
# - multi GPUS (using PyTorch distributed mode)
# - (multi) TPUs
# - fp16 (mixed-precision) or fp32 (normal precision)
#
# To help focus on the differences in the code, building `DataLoaders`
# was refactored into its own function.
# New additions from the base script can be found quickly by
# looking for the # New Code # tags
#
# To run it in each of these various modes, follow the instructions
# in the readme for examples:
# https://github.com/huggingface/accelerate/tree/main/examples
#
########################################################################
_UpperCAmelCase : List[Any] = 16
_UpperCAmelCase : Any = 32
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase = 16 ) -> List[str]:
lowerCamelCase__ : Any = AutoTokenizer.from_pretrained('bert-base-cased' )
lowerCamelCase__ : List[Any] = load_dataset('glue' , 'mrpc' )
def tokenize_function(_UpperCAmelCase ):
# max_length=None => use the model max length (it's actually the default)
lowerCamelCase__ : Optional[int] = tokenizer(examples['sentence1'] , examples['sentence2'] , truncation=_UpperCAmelCase , max_length=_UpperCAmelCase )
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
# starting with the main process first:
with accelerator.main_process_first():
lowerCamelCase__ : List[Any] = datasets.map(
_UpperCAmelCase , batched=_UpperCAmelCase , remove_columns=['idx', 'sentence1', 'sentence2'] , )
# We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the
# transformers library
lowerCamelCase__ : List[Any] = tokenized_datasets.rename_column('label' , 'labels' )
def collate_fn(_UpperCAmelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
lowerCamelCase__ : int = 128 if accelerator.distributed_type == DistributedType.TPU else None
# When using mixed precision we want round multiples of 8/16
if accelerator.mixed_precision == "fp8":
lowerCamelCase__ : int = 16
elif accelerator.mixed_precision != "no":
lowerCamelCase__ : Dict = 8
else:
lowerCamelCase__ : Optional[int] = None
return tokenizer.pad(
_UpperCAmelCase , padding='longest' , max_length=_UpperCAmelCase , pad_to_multiple_of=_UpperCAmelCase , return_tensors='pt' , )
# Instantiate dataloaders.
lowerCamelCase__ : Optional[int] = DataLoader(
tokenized_datasets['train'] , shuffle=_UpperCAmelCase , collate_fn=_UpperCAmelCase , batch_size=_UpperCAmelCase )
lowerCamelCase__ : str = DataLoader(
tokenized_datasets['validation'] , shuffle=_UpperCAmelCase , collate_fn=_UpperCAmelCase , batch_size=_UpperCAmelCase )
return train_dataloader, eval_dataloader
# For testing only
if os.environ.get("""TESTING_MOCKED_DATALOADERS""", None) == "1":
from accelerate.test_utils.training import mocked_dataloaders
_UpperCAmelCase : Optional[Any] = mocked_dataloaders # noqa: F811
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> str:
# For testing only
if os.environ.get('TESTING_MOCKED_DATALOADERS' , _UpperCAmelCase ) == "1":
lowerCamelCase__ : Optional[int] = 2
# Initialize Accelerator
# New Code #
# We pass in "all" to `log_with` to grab all available trackers in the environment
# Note: If using a custom `Tracker` class, should be passed in here such as:
# >>> log_with = ["all", MyCustomTrackerClassInstance()]
if args.with_tracking:
lowerCamelCase__ : List[Any] = Accelerator(
cpu=args.cpu , mixed_precision=args.mixed_precision , log_with='all' , project_dir=args.project_dir )
else:
lowerCamelCase__ : Optional[int] = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision )
# Sample hyper-parameters for learning rate, batch size, seed and a few other HPs
lowerCamelCase__ : Optional[Any] = config['lr']
lowerCamelCase__ : Tuple = int(config['num_epochs'] )
lowerCamelCase__ : Union[str, Any] = int(config['seed'] )
lowerCamelCase__ : str = int(config['batch_size'] )
set_seed(_UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = get_dataloaders(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : Optional[int] = evaluate.load('glue' , 'mrpc' )
# If the batch size is too big we use gradient accumulation
lowerCamelCase__ : Optional[Any] = 1
if batch_size > MAX_GPU_BATCH_SIZE and accelerator.distributed_type != DistributedType.TPU:
lowerCamelCase__ : Any = batch_size // MAX_GPU_BATCH_SIZE
lowerCamelCase__ : str = MAX_GPU_BATCH_SIZE
# Instantiate the model (we build the model here so that the seed also control new weights initialization)
lowerCamelCase__ : Union[str, Any] = AutoModelForSequenceClassification.from_pretrained('bert-base-cased' , return_dict=_UpperCAmelCase )
# We could avoid this line since the accelerator is set with `device_placement=True` (default value).
# Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer
# creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that).
lowerCamelCase__ : Optional[Any] = model.to(accelerator.device )
# Instantiate optimizer
lowerCamelCase__ : List[str] = AdamW(params=model.parameters() , lr=_UpperCAmelCase )
# Instantiate scheduler
lowerCamelCase__ : List[Any] = get_linear_schedule_with_warmup(
optimizer=_UpperCAmelCase , num_warmup_steps=100 , num_training_steps=(len(_UpperCAmelCase ) * num_epochs) // gradient_accumulation_steps , )
# Prepare everything
# There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the
# prepare method.
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = accelerator.prepare(
_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
# New Code #
# We need to initialize the trackers we use. Overall configurations can also be stored
if args.with_tracking:
lowerCamelCase__ : Optional[Any] = os.path.split(_UpperCAmelCase )[-1].split('.' )[0]
accelerator.init_trackers(_UpperCAmelCase , _UpperCAmelCase )
# Now we train the model
for epoch in range(_UpperCAmelCase ):
model.train()
# New Code #
# For our tracking example, we will log the total loss of each epoch
if args.with_tracking:
lowerCamelCase__ : Union[str, Any] = 0
for step, batch in enumerate(_UpperCAmelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
lowerCamelCase__ : Any = model(**_UpperCAmelCase )
lowerCamelCase__ : Dict = outputs.loss
# New Code #
if args.with_tracking:
total_loss += loss.detach().float()
lowerCamelCase__ : Dict = loss / gradient_accumulation_steps
accelerator.backward(_UpperCAmelCase )
if step % gradient_accumulation_steps == 0:
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
model.eval()
for step, batch in enumerate(_UpperCAmelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True` (the default).
batch.to(accelerator.device )
with torch.no_grad():
lowerCamelCase__ : str = model(**_UpperCAmelCase )
lowerCamelCase__ : List[str] = outputs.logits.argmax(dim=-1 )
lowerCamelCase__ , lowerCamelCase__ : Dict = accelerator.gather_for_metrics((predictions, batch['labels']) )
metric.add_batch(
predictions=_UpperCAmelCase , references=_UpperCAmelCase , )
lowerCamelCase__ : Tuple = metric.compute()
# Use accelerator.print to print only on the main process.
accelerator.print(F"""epoch {epoch}:""" , _UpperCAmelCase )
# New Code #
# To actually log, we call `Accelerator.log`
# The values passed can be of `str`, `int`, `float` or `dict` of `str` to `float`/`int`
if args.with_tracking:
accelerator.log(
{
'accuracy': eval_metric['accuracy'],
'f1': eval_metric['f1'],
'train_loss': total_loss.item() / len(_UpperCAmelCase ),
'epoch': epoch,
} , step=_UpperCAmelCase , )
# New Code #
# When a run is finished, you should call `accelerator.end_training()`
# to close all of the open trackers
if args.with_tracking:
accelerator.end_training()
def SCREAMING_SNAKE_CASE ( ) -> List[str]:
lowerCamelCase__ : Any = argparse.ArgumentParser(description='Simple example of training script.' )
parser.add_argument(
'--mixed_precision' , type=_UpperCAmelCase , default=_UpperCAmelCase , choices=['no', 'fp16', 'bf16', 'fp8'] , help='Whether to use mixed precision. Choose'
'between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10.'
'and an Nvidia Ampere GPU.' , )
parser.add_argument('--cpu' , action='store_true' , help='If passed, will train on the CPU.' )
parser.add_argument(
'--with_tracking' , action='store_true' , help='Whether to load in all available experiment trackers from the environment and use them for logging.' , )
parser.add_argument(
'--project_dir' , type=_UpperCAmelCase , default='logs' , help='Location on where to store experiment tracking logs` and relevent project information' , )
lowerCamelCase__ : Union[str, Any] = parser.parse_args()
lowerCamelCase__ : str = {'lr': 2e-5, 'num_epochs': 3, 'seed': 42, 'batch_size': 16}
training_function(_UpperCAmelCase , _UpperCAmelCase )
if __name__ == "__main__":
main()
| 50
|
# Copyright 2021 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.
import argparse
import os
from accelerate.test_utils import execute_subprocess_async
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase=None ) -> Tuple:
if subparsers is not None:
lowerCamelCase__ : Any = subparsers.add_parser('test' )
else:
lowerCamelCase__ : int = argparse.ArgumentParser('Accelerate test command' )
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 SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = os.path.sep.join(__file__.split(os.path.sep )[:-2] + ['test_utils', 'scripts', 'test_script.py'] )
if args.config_file is None:
lowerCamelCase__ : List[str] = script_name
else:
lowerCamelCase__ : List[Any] = F"""--config_file={args.config_file} {script_name}"""
lowerCamelCase__ : str = ['accelerate-launch'] + test_args.split()
lowerCamelCase__ : Dict = execute_subprocess_async(_UpperCAmelCase , env=os.environ.copy() )
if result.returncode == 0:
print('Test is a success! You are ready for your distributed training!' )
def SCREAMING_SNAKE_CASE ( ) -> Any:
lowerCamelCase__ : Any = test_command_parser()
lowerCamelCase__ : List[Any] = parser.parse_args()
test_command(_UpperCAmelCase )
if __name__ == "__main__":
main()
| 50
| 1
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_UpperCAmelCase : int = logging.get_logger(__name__)
_UpperCAmelCase : Tuple = {
"""google/vivit-b-16x2-kinetics400""": (
"""https://huggingface.co/google/vivit-b-16x2-kinetics400/resolve/main/config.json"""
),
# See all Vivit models at https://huggingface.co/models?filter=vivit
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """vivit"""
def __init__( self : Optional[int] , UpperCAmelCase : List[Any]=224 , UpperCAmelCase : List[str]=32 , UpperCAmelCase : Optional[int]=[2, 16, 16] , UpperCAmelCase : int=3 , UpperCAmelCase : int=768 , UpperCAmelCase : List[str]=12 , UpperCAmelCase : int=12 , UpperCAmelCase : List[Any]=3072 , UpperCAmelCase : Optional[Any]="gelu_fast" , UpperCAmelCase : Any=0.0 , UpperCAmelCase : int=0.0 , UpperCAmelCase : Optional[int]=0.0_2 , UpperCAmelCase : Tuple=1e-06 , UpperCAmelCase : Optional[int]=True , **UpperCAmelCase : Dict , ) -> str:
lowerCamelCase__ : Any = hidden_size
lowerCamelCase__ : List[str] = num_hidden_layers
lowerCamelCase__ : Dict = num_attention_heads
lowerCamelCase__ : List[Any] = intermediate_size
lowerCamelCase__ : Any = hidden_act
lowerCamelCase__ : Tuple = hidden_dropout_prob
lowerCamelCase__ : Union[str, Any] = attention_probs_dropout_prob
lowerCamelCase__ : Optional[int] = initializer_range
lowerCamelCase__ : Dict = layer_norm_eps
lowerCamelCase__ : int = image_size
lowerCamelCase__ : Any = num_frames
lowerCamelCase__ : Optional[Any] = tubelet_size
lowerCamelCase__ : Tuple = num_channels
lowerCamelCase__ : Any = qkv_bias
super().__init__(**UpperCAmelCase )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 100_0000 ) -> int:
lowerCamelCase__ : int = limit + 1
lowerCamelCase__ : Optional[Any] = [0] * limit
for first_term in range(1 , _UpperCAmelCase ):
for n in range(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Optional[Any] = first_term + n / first_term
if common_difference % 4: # d must be divisble by 4
continue
else:
common_difference /= 4
if (
first_term > common_difference
and first_term < 4 * common_difference
): # since x,y,z are positive integers
frequency[n] += 1 # so z>0 and a>d ,also 4d<a
lowerCamelCase__ : List[str] = sum(1 for x in frequency[1:limit] if x == 10 )
return count
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list:
if len(_UpperCAmelCase ) < 2:
return collection
def circle_sort_util(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> bool:
lowerCamelCase__ : Dict = False
if low == high:
return swapped
lowerCamelCase__ : List[str] = low
lowerCamelCase__ : List[str] = high
while left < right:
if collection[left] > collection[right]:
lowerCamelCase__ , lowerCamelCase__ : Any = (
collection[right],
collection[left],
)
lowerCamelCase__ : Optional[int] = True
left += 1
right -= 1
if left == right and collection[left] > collection[right + 1]:
lowerCamelCase__ , lowerCamelCase__ : List[Any] = (
collection[right + 1],
collection[left],
)
lowerCamelCase__ : List[str] = True
lowerCamelCase__ : Optional[int] = low + int((high - low) / 2 )
lowerCamelCase__ : Optional[Any] = circle_sort_util(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : List[str] = circle_sort_util(_UpperCAmelCase , mid + 1 , _UpperCAmelCase )
return swapped or left_swap or right_swap
lowerCamelCase__ : List[str] = True
while is_not_sorted is True:
lowerCamelCase__ : Dict = circle_sort_util(_UpperCAmelCase , 0 , len(_UpperCAmelCase ) - 1 )
return collection
if __name__ == "__main__":
_UpperCAmelCase : Optional[Any] = input("""Enter numbers separated by a comma:\n""").strip()
_UpperCAmelCase : List[str] = [int(item) for item in user_input.split(""",""")]
print(circle_sort(unsorted))
| 50
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_base import BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import PaddingStrategy, logging
from .tokenization_realm import RealmTokenizer
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
_UpperCAmelCase : Any = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : int = {
"""vocab_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/vocab.txt"""
),
"""google/realm-orqa-nq-openqa""": """https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-nq-reader""": """https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-openqa""": """https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-reader""": """https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/tokenizer.jsont"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-openqa""": (
"""https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-reader""": (
"""https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-openqa""": (
"""https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-reader""": (
"""https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : int = {
"""google/realm-cc-news-pretrained-embedder""": 5_12,
"""google/realm-cc-news-pretrained-encoder""": 5_12,
"""google/realm-cc-news-pretrained-scorer""": 5_12,
"""google/realm-cc-news-pretrained-openqa""": 5_12,
"""google/realm-orqa-nq-openqa""": 5_12,
"""google/realm-orqa-nq-reader""": 5_12,
"""google/realm-orqa-wq-openqa""": 5_12,
"""google/realm-orqa-wq-reader""": 5_12,
}
_UpperCAmelCase : Any = {
"""google/realm-cc-news-pretrained-embedder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-encoder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-scorer""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-reader""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-reader""": {"""do_lower_case""": True},
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_INIT_CONFIGURATION
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = RealmTokenizer
def __init__( self : Optional[int] , UpperCAmelCase : Tuple=None , UpperCAmelCase : Any=None , UpperCAmelCase : List[Any]=True , UpperCAmelCase : Optional[Any]="[UNK]" , UpperCAmelCase : Any="[SEP]" , UpperCAmelCase : Tuple="[PAD]" , UpperCAmelCase : List[Any]="[CLS]" , UpperCAmelCase : Union[str, Any]="[MASK]" , UpperCAmelCase : Optional[Any]=True , UpperCAmelCase : Any=None , **UpperCAmelCase : Optional[int] , ) -> str:
super().__init__(
UpperCAmelCase , tokenizer_file=UpperCAmelCase , do_lower_case=UpperCAmelCase , unk_token=UpperCAmelCase , sep_token=UpperCAmelCase , pad_token=UpperCAmelCase , cls_token=UpperCAmelCase , mask_token=UpperCAmelCase , tokenize_chinese_chars=UpperCAmelCase , strip_accents=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get('lowercase' , UpperCAmelCase ) != do_lower_case
or normalizer_state.get('strip_accents' , UpperCAmelCase ) != strip_accents
or normalizer_state.get('handle_chinese_chars' , UpperCAmelCase ) != tokenize_chinese_chars
):
lowerCamelCase__ : Optional[int] = getattr(UpperCAmelCase , normalizer_state.pop('type' ) )
lowerCamelCase__ : Optional[Any] = do_lower_case
lowerCamelCase__ : str = strip_accents
lowerCamelCase__ : Optional[Any] = tokenize_chinese_chars
lowerCamelCase__ : int = normalizer_class(**UpperCAmelCase )
lowerCamelCase__ : str = do_lower_case
def A_ ( self : Optional[int] , UpperCAmelCase : int , **UpperCAmelCase : int ) -> List[Any]:
lowerCamelCase__ : List[Any] = PaddingStrategy.MAX_LENGTH
lowerCamelCase__ : Optional[int] = text
lowerCamelCase__ : Dict = kwargs.pop('text_pair' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = kwargs.pop('return_tensors' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = {
'input_ids': [],
'attention_mask': [],
'token_type_ids': [],
}
for idx, candidate_text in enumerate(UpperCAmelCase ):
if batch_text_pair is not None:
lowerCamelCase__ : Tuple = batch_text_pair[idx]
else:
lowerCamelCase__ : Dict = None
lowerCamelCase__ : Optional[int] = super().__call__(UpperCAmelCase , UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Any = encoded_candidates.get('input_ids' )
lowerCamelCase__ : Union[str, Any] = encoded_candidates.get('attention_mask' )
lowerCamelCase__ : Tuple = encoded_candidates.get('token_type_ids' )
if encoded_input_ids is not None:
output_data["input_ids"].append(UpperCAmelCase )
if encoded_attention_mask is not None:
output_data["attention_mask"].append(UpperCAmelCase )
if encoded_token_type_ids is not None:
output_data["token_type_ids"].append(UpperCAmelCase )
lowerCamelCase__ : int = {key: item for key, item in output_data.items() if len(UpperCAmelCase ) != 0}
return BatchEncoding(UpperCAmelCase , tensor_type=UpperCAmelCase )
def A_ ( self : int , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Optional[int]=None ) -> List[str]:
lowerCamelCase__ : Tuple = [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 A_ ( self : Tuple , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : List[Any] = [self.sep_token_id]
lowerCamelCase__ : Optional[int] = [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 A_ ( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
lowerCamelCase__ : int = self._tokenizer.model.save(UpperCAmelCase , name=UpperCAmelCase )
return tuple(UpperCAmelCase )
| 50
| 1
|
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
_UpperCAmelCase : int = DUMMY_UNKNOWN_IDENTIFIER
# An actual model hosted on huggingface.co
_UpperCAmelCase : Optional[Any] = """main"""
# Default branch name
_UpperCAmelCase : Any = """f2c752cfc5c0ab6f4bdec59acea69eefbee381c2"""
# One particular commit (not the top of `main`)
_UpperCAmelCase : int = """aaaaaaa"""
# This commit does not exist, so we should 404.
_UpperCAmelCase : List[str] = """d9e9f15bc825e4b2c9249e9578f884bbcb5e3684"""
# Sha-1 of config.json on the top of `main`, for checking purposes
_UpperCAmelCase : Any = """4b243c475af8d0a7754e87d7d096c92e5199ec2fe168a2ee7998e3b8e9bcb1d3"""
@contextlib.contextmanager
def SCREAMING_SNAKE_CASE ( ) -> Tuple:
print('Welcome!' )
yield
print('Bye!' )
@contextlib.contextmanager
def SCREAMING_SNAKE_CASE ( ) -> Any:
print('Bonjour!' )
yield
print('Au revoir!' )
class lowerCAmelCase ( unittest.TestCase ):
def A_ ( self : List[str] ) -> str:
# 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 lowerCAmelCase ( unittest.TestCase ):
@unittest.mock.patch('sys.stdout' , new_callable=io.StringIO )
def A_ ( self : int , UpperCAmelCase : int ) -> 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 A_ ( self : str , UpperCAmelCase : Optional[Any] ) -> str:
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 A_ ( self : int , UpperCAmelCase : str ) -> 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 A_ ( self : Any ) -> Dict:
self.assertEqual(find_labels(UpperCAmelCase ) , ['labels'] )
self.assertEqual(find_labels(UpperCAmelCase ) , ['labels', 'next_sentence_label'] )
self.assertEqual(find_labels(UpperCAmelCase ) , ['start_positions', 'end_positions'] )
class lowerCAmelCase ( __UpperCamelCase ):
pass
self.assertEqual(find_labels(UpperCAmelCase ) , ['labels'] )
@require_tf
def A_ ( self : Any ) -> Any:
self.assertEqual(find_labels(UpperCAmelCase ) , ['labels'] )
self.assertEqual(find_labels(UpperCAmelCase ) , ['labels', 'next_sentence_label'] )
self.assertEqual(find_labels(UpperCAmelCase ) , ['start_positions', 'end_positions'] )
class lowerCAmelCase ( __UpperCamelCase ):
pass
self.assertEqual(find_labels(UpperCAmelCase ) , ['labels'] )
@require_flax
def A_ ( self : str ) -> Union[str, Any]:
# Flax models don't have labels
self.assertEqual(find_labels(UpperCAmelCase ) , [] )
self.assertEqual(find_labels(UpperCAmelCase ) , [] )
self.assertEqual(find_labels(UpperCAmelCase ) , [] )
class lowerCAmelCase ( __UpperCamelCase ):
pass
self.assertEqual(find_labels(UpperCAmelCase ) , [] )
| 50
|
import logging
import os
from dataclasses import dataclass
from typing import List, Optional, Union
import tqdm
from filelock import FileLock
from transformers import (
BartTokenizer,
BartTokenizerFast,
DataProcessor,
PreTrainedTokenizer,
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
is_tf_available,
is_torch_available,
)
_UpperCAmelCase : Optional[Any] = logging.getLogger(__name__)
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
if is_torch_available():
import torch
from torch.utils.data import Dataset
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
def __init__( self : int , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : List[str]=False , UpperCAmelCase : bool = False , ) -> List[str]:
lowerCamelCase__ : int = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = os.path.join(
UpperCAmelCase , 'cached_{}_{}_{}_{}'.format(
'dev' if evaluate else 'train' , tokenizer.__class__.__name__ , str(UpperCAmelCase ) , UpperCAmelCase , ) , )
lowerCamelCase__ : int = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = label_list[2], label_list[1]
lowerCamelCase__ : List[str] = label_list
# Make sure only the first process in distributed training processes the dataset,
# and the others will use the cache.
lowerCamelCase__ : str = cached_features_file + '.lock'
with FileLock(UpperCAmelCase ):
if os.path.exists(UpperCAmelCase ) and not overwrite_cache:
logger.info(F"""Loading features from cached file {cached_features_file}""" )
lowerCamelCase__ : int = torch.load(UpperCAmelCase )
else:
logger.info(F"""Creating features from dataset file at {data_dir}""" )
lowerCamelCase__ : str = (
processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
)
logger.info('Training examples: %s' , len(UpperCAmelCase ) )
lowerCamelCase__ : Dict = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
logger.info('Saving features into cached file %s' , UpperCAmelCase )
torch.save(self.features , UpperCAmelCase )
def __len__( self : Optional[int] ) -> Optional[Any]:
return len(self.features )
def __getitem__( self : Tuple , UpperCAmelCase : Dict ) -> InputFeatures:
return self.features[i]
def A_ ( self : int ) -> int:
return self.label_list
if is_tf_available():
import tensorflow as tf
class lowerCAmelCase :
UpperCAmelCase__ = 42
def __init__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = 128 , UpperCAmelCase : Any=False , UpperCAmelCase : bool = False , ) -> Union[str, Any]:
lowerCamelCase__ : Any = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : str = label_list[2], label_list[1]
lowerCamelCase__ : Optional[int] = label_list
lowerCamelCase__ : int = processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def gen():
for ex_index, ex in tqdm.tqdm(enumerate(self.features ) , desc='convert examples to features' ):
if ex_index % 10000 == 0:
logger.info('Writing example %d of %d' % (ex_index, len(UpperCAmelCase )) )
yield (
{
"example_id": 0,
"input_ids": ex.input_ids,
"attention_mask": ex.attention_mask,
"token_type_ids": ex.token_type_ids,
},
ex.label,
)
lowerCamelCase__ : Optional[int] = tf.data.Dataset.from_generator(
UpperCAmelCase , (
{
'example_id': tf.intaa,
'input_ids': tf.intaa,
'attention_mask': tf.intaa,
'token_type_ids': tf.intaa,
},
tf.intaa,
) , (
{
'example_id': tf.TensorShape([] ),
'input_ids': tf.TensorShape([None, None] ),
'attention_mask': tf.TensorShape([None, None] ),
'token_type_ids': tf.TensorShape([None, None] ),
},
tf.TensorShape([] ),
) , )
def A_ ( self : Any ) -> Any:
return self.dataset
def __len__( self : Tuple ) -> int:
return len(self.features )
def __getitem__( self : List[str] , UpperCAmelCase : Any ) -> InputFeatures:
return self.features[i]
def A_ ( self : Dict ) -> str:
return self.label_list
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : int , UpperCAmelCase : List[Any] ) -> int:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_train_set.txt' ) ) , 'train' )
def A_ ( self : Any , UpperCAmelCase : int ) -> List[Any]:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_evaluation_set.txt' ) ) , 'dev' )
def A_ ( self : Any ) -> List[Any]:
return ["contradiction", "entailment", "neutral"]
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : List[str] ) -> List[str]:
lowerCamelCase__ : List[str] = []
for i, line in enumerate(UpperCAmelCase ):
if i == 0:
continue
lowerCamelCase__ : Tuple = '%s-%s' % (set_type, line[0])
lowerCamelCase__ : str = line[5]
lowerCamelCase__ : Dict = line[6]
lowerCamelCase__ : int = line[7][2:] if line[7].startswith('ex' ) else line[7]
lowerCamelCase__ : Dict = line[0]
examples.append(InputExample(guid=UpperCAmelCase , text_a=UpperCAmelCase , text_b=UpperCAmelCase , label=UpperCAmelCase , pairID=UpperCAmelCase ) )
return examples
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ) -> Optional[int]:
lowerCamelCase__ : int = {label: i for i, label in enumerate(_UpperCAmelCase )}
lowerCamelCase__ : List[Any] = []
for ex_index, example in tqdm.tqdm(enumerate(_UpperCAmelCase ) , desc='convert examples to features' ):
if ex_index % 1_0000 == 0:
logger.info('Writing example %d' % (ex_index) )
lowerCamelCase__ : List[Any] = tokenizer(
example.text_a , example.text_b , add_special_tokens=_UpperCAmelCase , max_length=_UpperCAmelCase , padding='max_length' , truncation=_UpperCAmelCase , return_overflowing_tokens=_UpperCAmelCase , )
lowerCamelCase__ : List[str] = label_map[example.label] if example.label in label_map else 0
lowerCamelCase__ : Optional[int] = int(example.pairID )
features.append(InputFeatures(**_UpperCAmelCase , label=_UpperCAmelCase , pairID=_UpperCAmelCase ) )
for i, example in enumerate(examples[:5] ):
logger.info('*** Example ***' )
logger.info(F"""guid: {example}""" )
logger.info(F"""features: {features[i]}""" )
return features
_UpperCAmelCase : str = {
"""hans""": 3,
}
_UpperCAmelCase : List[Any] = {
"""hans""": HansProcessor,
}
| 50
| 1
|
from .data_collator import (
DataCollatorForLanguageModeling,
DataCollatorForPermutationLanguageModeling,
DataCollatorForSeqaSeq,
DataCollatorForSOP,
DataCollatorForTokenClassification,
DataCollatorForWholeWordMask,
DataCollatorWithPadding,
DefaultDataCollator,
default_data_collator,
)
from .metrics import glue_compute_metrics, xnli_compute_metrics
from .processors import (
DataProcessor,
InputExample,
InputFeatures,
SingleSentenceClassificationProcessor,
SquadExample,
SquadFeatures,
SquadVaProcessor,
SquadVaProcessor,
glue_convert_examples_to_features,
glue_output_modes,
glue_processors,
glue_tasks_num_labels,
squad_convert_examples_to_features,
xnli_output_modes,
xnli_processors,
xnli_tasks_num_labels,
)
| 50
|
import unittest
from transformers import BertGenerationTokenizer
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
_UpperCAmelCase : Optional[Any] = """▁"""
_UpperCAmelCase : Optional[Any] = get_tests_dir("""fixtures/test_sentencepiece.model""")
@require_sentencepiece
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = BertGenerationTokenizer
UpperCAmelCase__ = False
UpperCAmelCase__ = True
def A_ ( self : List[Any] ) -> List[str]:
super().setUp()
lowerCamelCase__ : Dict = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
tokenizer.save_pretrained(self.tmpdirname )
def A_ ( self : Optional[Any] ) -> Dict:
lowerCamelCase__ : List[str] = '<s>'
lowerCamelCase__ : Optional[int] = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCAmelCase ) , UpperCAmelCase )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : List[str] ) -> Optional[int]:
lowerCamelCase__ : Optional[Any] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , '<unk>' )
self.assertEqual(vocab_keys[1] , '<s>' )
self.assertEqual(vocab_keys[-1] , '<pad>' )
self.assertEqual(len(UpperCAmelCase ) , 1002 )
def A_ ( self : List[Any] ) -> Optional[Any]:
self.assertEqual(self.get_tokenizer().vocab_size , 1000 )
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : List[str] = tokenizer.tokenize('This is a test' )
self.assertListEqual(UpperCAmelCase , ['▁This', '▁is', '▁a', '▁t', 'est'] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(UpperCAmelCase ) , [285, 46, 10, 170, 382] , )
lowerCamelCase__ : List[str] = tokenizer.tokenize('I was born in 92000, and this is falsé.' )
self.assertListEqual(
UpperCAmelCase , [
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',
'.',
] , )
lowerCamelCase__ : Optional[int] = tokenizer.convert_tokens_to_ids(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
lowerCamelCase__ : Optional[Any] = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [
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>',
'.',
] , )
@cached_property
def A_ ( self : Dict ) -> Tuple:
return BertGenerationTokenizer.from_pretrained('google/bert_for_seq_generation_L-24_bbc_encoder' )
@slow
def A_ ( self : Optional[int] ) -> List[str]:
lowerCamelCase__ : Union[str, Any] = 'Hello World!'
lowerCamelCase__ : Dict = [18536, 2260, 101]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@slow
def A_ ( self : Optional[Any] ) -> str:
lowerCamelCase__ : List[Any] = (
'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'
)
lowerCamelCase__ : Any = [
871,
419,
358,
946,
991,
2521,
452,
358,
1357,
387,
7751,
3536,
112,
985,
456,
126,
865,
938,
5400,
5734,
458,
1368,
467,
786,
2462,
5246,
1159,
633,
865,
4519,
457,
582,
852,
2557,
427,
916,
508,
405,
34324,
497,
391,
408,
11342,
1244,
385,
100,
938,
985,
456,
574,
362,
12597,
3200,
3129,
1172,
]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@require_torch
@slow
def A_ ( self : int ) -> Optional[Any]:
import torch
from transformers import BertGenerationConfig, BertGenerationEncoder
# Build sequence
lowerCamelCase__ : str = list(self.big_tokenizer.get_vocab().keys() )[:10]
lowerCamelCase__ : int = ' '.join(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.big_tokenizer.encode_plus(UpperCAmelCase , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.big_tokenizer.batch_encode_plus(
[sequence + ' ' + sequence] , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Tuple = BertGenerationConfig()
lowerCamelCase__ : Optional[Any] = BertGenerationEncoder(UpperCAmelCase )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**UpperCAmelCase )
model(**UpperCAmelCase )
@slow
def A_ ( self : Optional[int] ) -> List[Any]:
# fmt: off
lowerCamelCase__ : Any = {'input_ids': [[39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114], [448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 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], [484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 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]], '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, 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, 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=UpperCAmelCase , model_name='google/bert_for_seq_generation_L-24_bbc_encoder' , revision='c817d1fd1be2ffa69431227a1fe320544943d4db' , )
| 50
| 1
|
from ... import PretrainedConfig
_UpperCAmelCase : List[str] = {
"""sijunhe/nezha-cn-base""": """https://huggingface.co/sijunhe/nezha-cn-base/resolve/main/config.json""",
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = NEZHA_PRETRAINED_CONFIG_ARCHIVE_MAP
UpperCAmelCase__ = """nezha"""
def __init__( self : int , UpperCAmelCase : Optional[Any]=21128 , UpperCAmelCase : List[str]=768 , UpperCAmelCase : Tuple=12 , UpperCAmelCase : str=12 , UpperCAmelCase : Tuple=3072 , UpperCAmelCase : List[str]="gelu" , UpperCAmelCase : Dict=0.1 , UpperCAmelCase : Optional[Any]=0.1 , UpperCAmelCase : Dict=512 , UpperCAmelCase : Optional[int]=64 , UpperCAmelCase : int=2 , UpperCAmelCase : List[Any]=0.0_2 , UpperCAmelCase : str=1e-12 , UpperCAmelCase : Optional[int]=0.1 , UpperCAmelCase : Optional[int]=0 , UpperCAmelCase : Optional[int]=2 , UpperCAmelCase : int=3 , UpperCAmelCase : str=True , **UpperCAmelCase : Tuple , ) -> Optional[int]:
super().__init__(pad_token_id=UpperCAmelCase , bos_token_id=UpperCAmelCase , eos_token_id=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Tuple = vocab_size
lowerCamelCase__ : Tuple = hidden_size
lowerCamelCase__ : Dict = num_hidden_layers
lowerCamelCase__ : Optional[Any] = num_attention_heads
lowerCamelCase__ : Optional[Any] = hidden_act
lowerCamelCase__ : int = intermediate_size
lowerCamelCase__ : Optional[int] = hidden_dropout_prob
lowerCamelCase__ : Union[str, Any] = attention_probs_dropout_prob
lowerCamelCase__ : int = max_position_embeddings
lowerCamelCase__ : Tuple = max_relative_position
lowerCamelCase__ : List[str] = type_vocab_size
lowerCamelCase__ : Union[str, Any] = initializer_range
lowerCamelCase__ : Union[str, Any] = layer_norm_eps
lowerCamelCase__ : Optional[int] = classifier_dropout
lowerCamelCase__ : str = use_cache
| 50
|
import os
import tempfile
from functools import partial
from unittest import TestCase
from unittest.mock import patch
import numpy as np
import pytest
from datasets.arrow_dataset import Dataset
from datasets.search import ElasticSearchIndex, FaissIndex, MissingIndex
from .utils import require_elasticsearch, require_faiss
_UpperCAmelCase : str = pytest.mark.integration
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : List[Any] ) -> Union[str, Any]:
lowerCamelCase__ : List[Any] = Dataset.from_dict({'filename': ['my_name-train' + '_' + str(UpperCAmelCase ) for x in np.arange(30 ).tolist()]} )
return dset
def A_ ( self : Optional[Any] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
lowerCamelCase__ : List[Any] = dset.map(
lambda UpperCAmelCase , UpperCAmelCase : {"vecs": i * np.ones(5 , dtype=np.floataa )} , with_indices=UpperCAmelCase , keep_in_memory=UpperCAmelCase )
lowerCamelCase__ : Tuple = dset.add_faiss_index('vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
dset.drop_index('vecs' )
def A_ ( self : Union[str, Any] ) -> int:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT , )
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : List[str] ) -> Tuple:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , metric_type=faiss.METRIC_INNER_PRODUCT , )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
dset.save_faiss_index('vecs' , tmp_file.name )
dset.load_faiss_index('vecs2' , tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs2' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' )
dset.drop_index('vecs' )
self.assertRaises(UpperCAmelCase , partial(dset.get_nearest_examples , 'vecs2' , np.ones(5 , dtype=np.floataa ) ) )
def A_ ( self : Dict ) -> Dict:
from elasticsearch import Elasticsearch
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : List[Any] = {'acknowledged': True}
mocked_bulk.return_value([(True, None)] * 30 )
lowerCamelCase__ : int = {'hits': {'hits': [{'_score': 1, '_id': 29}]}}
lowerCamelCase__ : List[str] = Elasticsearch()
dset.add_elasticsearch_index('filename' , es_client=UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Dict = dset.get_nearest_examples('filename' , 'my_name-train_29' )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Any ) -> Dict:
import faiss
lowerCamelCase__ : Tuple = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
# add vectors
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsNotNone(index.faiss_index )
self.assertEqual(index.faiss_index.ntotal , 5 )
index.add_vectors(np.zeros((5, 5) , dtype=np.floataa ) )
self.assertEqual(index.faiss_index.ntotal , 10 )
# single query
lowerCamelCase__ : int = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Any = 1
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = index.search(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search , query.reshape(-1 , 1 ) )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
# batched queries
lowerCamelCase__ : str = np.eye(5 , dtype=np.floataa )[::-1]
lowerCamelCase__ , lowerCamelCase__ : Dict = index.search_batch(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search_batch , queries[0] )
lowerCamelCase__ : str = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[str] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([4, 3, 2, 1, 0] , UpperCAmelCase )
def A_ ( self : List[Any] ) -> List[Any]:
import faiss
lowerCamelCase__ : Any = FaissIndex(string_factory='Flat' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
lowerCamelCase__ : Tuple = FaissIndex(string_factory='LSH' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexLSH )
with self.assertRaises(UpperCAmelCase ):
lowerCamelCase__ : List[str] = FaissIndex(string_factory='Flat' , custom_index=faiss.IndexFlat(5 ) )
def A_ ( self : List[str] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Optional[Any] = faiss.IndexFlat(5 )
lowerCamelCase__ : int = FaissIndex(custom_index=UpperCAmelCase )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
def A_ ( self : Any ) -> Optional[int]:
import faiss
lowerCamelCase__ : Any = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
index.save(tmp_file.name )
lowerCamelCase__ : List[Any] = FaissIndex.load(tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ : List[str] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Tuple = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(UpperCAmelCase )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
@require_faiss
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Any:
import faiss
lowerCamelCase__ : Optional[Any] = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
lowerCamelCase__ : Optional[int] = 'index.faiss'
lowerCamelCase__ : Optional[Any] = F"""mock://{index_name}"""
index.save(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Tuple = FaissIndex.load(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Optional[int] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Dict = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(_UpperCAmelCase )
assert scores[0] > 0
assert indices[0] == 1
@require_elasticsearch
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Dict ) -> List[Any]:
from elasticsearch import Elasticsearch
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : Any = Elasticsearch()
lowerCamelCase__ : Tuple = {'acknowledged': True}
lowerCamelCase__ : Tuple = ElasticSearchIndex(es_client=UpperCAmelCase )
mocked_bulk.return_value([(True, None)] * 3 )
index.add_documents(['foo', 'bar', 'foobar'] )
# single query
lowerCamelCase__ : Optional[int] = 'foo'
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search(UpperCAmelCase )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# single query with timeout
lowerCamelCase__ : Any = 'foo'
lowerCamelCase__ : List[str] = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Tuple = index.search(UpperCAmelCase , request_timeout=30 )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# batched queries
lowerCamelCase__ : List[str] = ['foo', 'bar', 'foobar']
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : str = index.search_batch(UpperCAmelCase )
lowerCamelCase__ : List[str] = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[Any] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
# batched queries with timeout
lowerCamelCase__ : str = ['foo', 'bar', 'foobar']
lowerCamelCase__ : List[Any] = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search_batch(UpperCAmelCase , request_timeout=30 )
lowerCamelCase__ : Optional[Any] = [scores[0] for scores in total_scores]
lowerCamelCase__ : Dict = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
| 50
| 1
|
import collections
import tempfile
import unittest
import numpy as np
from transformers.testing_utils import (
is_pt_flax_cross_test,
require_flax,
require_torch,
require_vision,
slow,
torch_device,
)
from transformers.utils import is_flax_available, is_torch_available, is_vision_available
from ...test_modeling_flax_common import floats_tensor, ids_tensor, random_attention_mask
from ..bert.test_modeling_flax_bert import FlaxBertModelTester
from ..clip.test_modeling_flax_clip import FlaxCLIPVisionModelTester
from ..vit.test_modeling_flax_vit import FlaxViTModelTester
if is_flax_available():
from transformers import (
FlaxBertModel,
FlaxCLIPVisionModel,
FlaxVisionTextDualEncoderModel,
FlaxViTModel,
VisionTextDualEncoderConfig,
VisionTextDualEncoderProcessor,
)
from transformers.modeling_flax_pytorch_utils import (
convert_pytorch_state_dict_to_flax,
load_flax_weights_in_pytorch_model,
)
if is_torch_available():
import torch
from transformers import VisionTextDualEncoderModel
if is_vision_available():
from PIL import Image
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Any:
if isinstance(_UpperCAmelCase , collections.abc.Iterable ):
return x
return (x, x)
@require_flax
class lowerCAmelCase :
def A_ ( self : Any , UpperCAmelCase : Dict , UpperCAmelCase : Optional[Any] ) -> Union[str, Any]:
pass
def A_ ( self : Dict ) -> int:
pass
def A_ ( self : List[Any] ) -> List[str]:
pass
def A_ ( self : Union[str, Any] , UpperCAmelCase : np.ndarray , UpperCAmelCase : np.ndarray , UpperCAmelCase : float ) -> Union[str, Any]:
lowerCamelCase__ : Optional[Any] = np.abs((a - b) ).max()
self.assertLessEqual(UpperCAmelCase , UpperCAmelCase , F"""Difference between torch and flax is {diff} (>= {tol}).""" )
def A_ ( self : Any , UpperCAmelCase : Dict , UpperCAmelCase : int , UpperCAmelCase : Optional[int] , UpperCAmelCase : List[Any] , UpperCAmelCase : Any=None , **UpperCAmelCase : Dict ) -> Tuple:
lowerCamelCase__ : str = VisionTextDualEncoderConfig.from_vision_text_configs(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = FlaxVisionTextDualEncoderModel(UpperCAmelCase )
lowerCamelCase__ : int = model(input_ids=UpperCAmelCase , pixel_values=UpperCAmelCase , attention_mask=UpperCAmelCase )
self.assertEqual(output['text_embeds'].shape , (input_ids.shape[0], config.projection_dim) )
self.assertEqual(output['image_embeds'].shape , (pixel_values.shape[0], config.projection_dim) )
def A_ ( self : Tuple , UpperCAmelCase : Any , UpperCAmelCase : List[Any] , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : str , UpperCAmelCase : Dict=None , **UpperCAmelCase : Optional[int] ) -> Dict:
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = self.get_vision_text_model(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = {'vision_model': vision_model, 'text_model': text_model}
lowerCamelCase__ : str = FlaxVisionTextDualEncoderModel.from_vision_text_pretrained(**UpperCAmelCase )
lowerCamelCase__ : Any = model(input_ids=UpperCAmelCase , pixel_values=UpperCAmelCase , attention_mask=UpperCAmelCase )
self.assertEqual(output['text_embeds'].shape , (input_ids.shape[0], model.config.projection_dim) )
self.assertEqual(output['image_embeds'].shape , (pixel_values.shape[0], model.config.projection_dim) )
def A_ ( self : Tuple , UpperCAmelCase : List[Any] , UpperCAmelCase : List[Any] , UpperCAmelCase : Optional[int] , UpperCAmelCase : int , UpperCAmelCase : Optional[int]=None , **UpperCAmelCase : List[str] ) -> Dict:
lowerCamelCase__ , lowerCamelCase__ : str = self.get_vision_text_model(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = {'vision_model': vision_model, 'text_model': text_model}
lowerCamelCase__ : Union[str, Any] = FlaxVisionTextDualEncoderModel.from_vision_text_pretrained(**UpperCAmelCase )
lowerCamelCase__ : Any = model(input_ids=UpperCAmelCase , pixel_values=UpperCAmelCase , attention_mask=UpperCAmelCase )
lowerCamelCase__ : Optional[int] = output[0]
with tempfile.TemporaryDirectory() as tmpdirname:
model.save_pretrained(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = FlaxVisionTextDualEncoderModel.from_pretrained(UpperCAmelCase )
lowerCamelCase__ : int = model(input_ids=UpperCAmelCase , pixel_values=UpperCAmelCase , attention_mask=UpperCAmelCase )
lowerCamelCase__ : List[Any] = after_output[0]
lowerCamelCase__ : Optional[Any] = np.amax(np.abs(out_a - out_a ) )
self.assertLessEqual(UpperCAmelCase , 1e-3 )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[int] , UpperCAmelCase : Dict , UpperCAmelCase : str , UpperCAmelCase : str , UpperCAmelCase : str=None , **UpperCAmelCase : List[Any] ) -> str:
lowerCamelCase__ , lowerCamelCase__ : Tuple = self.get_vision_text_model(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Tuple = {'vision_model': vision_model, 'text_model': text_model}
lowerCamelCase__ : Any = FlaxVisionTextDualEncoderModel.from_vision_text_pretrained(**UpperCAmelCase )
lowerCamelCase__ : int = model(
input_ids=UpperCAmelCase , pixel_values=UpperCAmelCase , attention_mask=UpperCAmelCase , output_attentions=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = output.vision_model_output.attentions
self.assertEqual(len(UpperCAmelCase ) , vision_config.num_hidden_layers )
# in ViT, the seq_len equals the number of patches + 1 (we add 1 for the [CLS] token)
lowerCamelCase__ : str = to_atuple(vision_model.config.image_size )
lowerCamelCase__ : Optional[int] = to_atuple(vision_model.config.patch_size )
lowerCamelCase__ : Tuple = (image_size[1] // patch_size[1]) * (image_size[0] // patch_size[0])
lowerCamelCase__ : Optional[Any] = num_patches + 1
self.assertEqual(vision_attentions[0].shape[-3:] , (vision_config.num_attention_heads, seq_len, seq_len) )
lowerCamelCase__ : List[str] = output.text_model_output.attentions
self.assertEqual(len(UpperCAmelCase ) , text_config.num_hidden_layers )
self.assertEqual(
text_attentions[0].shape[-3:] , (text_config.num_attention_heads, input_ids.shape[-1], input_ids.shape[-1]) , )
def A_ ( self : Any , UpperCAmelCase : int , UpperCAmelCase : Optional[Any] , UpperCAmelCase : List[Any] ) -> str:
pt_model.to(UpperCAmelCase )
pt_model.eval()
# prepare inputs
lowerCamelCase__ : Optional[Any] = inputs_dict
lowerCamelCase__ : Optional[int] = {k: torch.tensor(v.tolist() ) for k, v in flax_inputs.items()}
with torch.no_grad():
lowerCamelCase__ : Optional[int] = pt_model(**UpperCAmelCase ).to_tuple()
lowerCamelCase__ : Tuple = fx_model(**UpperCAmelCase ).to_tuple()
self.assertEqual(len(UpperCAmelCase ) , len(UpperCAmelCase ) , 'Output lengths differ between Flax and PyTorch' )
for fx_output, pt_output in zip(fx_outputs[:4] , pt_outputs[:4] ):
self.assert_almost_equals(UpperCAmelCase , pt_output.numpy() , 4e-2 )
# PT -> Flax
with tempfile.TemporaryDirectory() as tmpdirname:
pt_model.save_pretrained(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = FlaxVisionTextDualEncoderModel.from_pretrained(UpperCAmelCase , from_pt=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = fx_model_loaded(**UpperCAmelCase ).to_tuple()
self.assertEqual(len(UpperCAmelCase ) , len(UpperCAmelCase ) , 'Output lengths differ between Flax and PyTorch' )
for fx_output_loaded, pt_output in zip(fx_outputs_loaded[:4] , pt_outputs[:4] ):
self.assert_almost_equals(UpperCAmelCase , pt_output.numpy() , 4e-2 )
# Flax -> PT
with tempfile.TemporaryDirectory() as tmpdirname:
fx_model.save_pretrained(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = VisionTextDualEncoderModel.from_pretrained(UpperCAmelCase , from_flax=UpperCAmelCase )
pt_model_loaded.to(UpperCAmelCase )
pt_model_loaded.eval()
with torch.no_grad():
lowerCamelCase__ : Optional[int] = pt_model_loaded(**UpperCAmelCase ).to_tuple()
self.assertEqual(len(UpperCAmelCase ) , len(UpperCAmelCase ) , 'Output lengths differ between Flax and PyTorch' )
for fx_output, pt_output_loaded in zip(fx_outputs[:4] , pt_outputs_loaded[:4] ):
self.assert_almost_equals(UpperCAmelCase , pt_output_loaded.numpy() , 4e-2 )
def A_ ( self : str , UpperCAmelCase : Tuple , UpperCAmelCase : int , UpperCAmelCase : Optional[int] ) -> Dict:
lowerCamelCase__ : List[Any] = VisionTextDualEncoderConfig.from_vision_text_configs(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Dict = VisionTextDualEncoderModel(UpperCAmelCase )
lowerCamelCase__ : Tuple = FlaxVisionTextDualEncoderModel(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = convert_pytorch_state_dict_to_flax(pt_model.state_dict() , UpperCAmelCase )
lowerCamelCase__ : Dict = fx_state
self.check_pt_flax_equivalence(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Tuple , UpperCAmelCase : int , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Optional[int] ) -> str:
lowerCamelCase__ : str = VisionTextDualEncoderConfig.from_vision_text_configs(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Any = VisionTextDualEncoderModel(UpperCAmelCase )
lowerCamelCase__ : Dict = FlaxVisionTextDualEncoderModel(UpperCAmelCase )
lowerCamelCase__ : Dict = load_flax_weights_in_pytorch_model(UpperCAmelCase , fx_model.params )
self.check_pt_flax_equivalence(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def A_ ( self : List[Any] ) -> int:
lowerCamelCase__ : Optional[int] = self.prepare_config_and_inputs()
self.check_model_from_pretrained_configs(**UpperCAmelCase )
def A_ ( self : Any ) -> Any:
lowerCamelCase__ : List[str] = self.prepare_config_and_inputs()
self.check_vision_text_dual_encoder_from_pretrained(**UpperCAmelCase )
def A_ ( self : Any ) -> Optional[int]:
lowerCamelCase__ : Tuple = self.prepare_config_and_inputs()
self.check_save_load(**UpperCAmelCase )
def A_ ( self : Dict ) -> str:
lowerCamelCase__ : int = self.prepare_config_and_inputs()
self.check_vision_text_output_attention(**UpperCAmelCase )
@is_pt_flax_cross_test
def A_ ( self : int ) -> List[str]:
lowerCamelCase__ : str = self.prepare_config_and_inputs()
lowerCamelCase__ : Dict = config_inputs_dict.pop('vision_config' )
lowerCamelCase__ : List[Any] = config_inputs_dict.pop('text_config' )
lowerCamelCase__ : Optional[Any] = config_inputs_dict
self.check_equivalence_pt_to_flax(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
self.check_equivalence_flax_to_pt(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
@slow
def A_ ( self : Any ) -> Dict:
lowerCamelCase__ , lowerCamelCase__ : Tuple = self.get_pretrained_model_and_inputs()
lowerCamelCase__ : int = model_a(**UpperCAmelCase )
lowerCamelCase__ : List[Any] = outputs[0]
with tempfile.TemporaryDirectory() as tmp_dirname:
model_a.save_pretrained(UpperCAmelCase )
lowerCamelCase__ : Tuple = FlaxVisionTextDualEncoderModel.from_pretrained(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = model_a(**UpperCAmelCase )
lowerCamelCase__ : int = after_outputs[0]
lowerCamelCase__ : Optional[Any] = np.amax(np.abs(out_a - out_a ) )
self.assertLessEqual(UpperCAmelCase , 1e-5 )
@require_flax
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
def A_ ( self : Dict ) -> Tuple:
lowerCamelCase__ : Optional[int] = FlaxVisionTextDualEncoderModel.from_vision_text_pretrained(
'hf-internal-testing/tiny-random-vit' , 'hf-internal-testing/tiny-bert' , vision_from_pt=UpperCAmelCase , text_from_pt=UpperCAmelCase , )
lowerCamelCase__ : List[str] = 13
lowerCamelCase__ : Optional[int] = floats_tensor(
[
batch_size,
model.config.vision_config.num_channels,
model.config.vision_config.image_size,
model.config.vision_config.image_size,
] )
lowerCamelCase__ : Union[str, Any] = ids_tensor([batch_size, 4] , model.config.text_config.vocab_size )
lowerCamelCase__ : Union[str, Any] = random_attention_mask([batch_size, 4] )
lowerCamelCase__ : Tuple = {'pixel_values': pixel_values, 'input_ids': input_ids, 'attention_mask': attention_mask}
return model, inputs
def A_ ( self : Union[str, Any] , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : str ) -> Optional[int]:
lowerCamelCase__ : str = FlaxViTModel(UpperCAmelCase )
lowerCamelCase__ : int = FlaxBertModel(UpperCAmelCase )
return vision_model, text_model
def A_ ( self : Any ) -> Optional[int]:
lowerCamelCase__ : Any = FlaxViTModelTester(self )
lowerCamelCase__ : Tuple = FlaxBertModelTester(self )
lowerCamelCase__ : Union[str, Any] = vit_model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Tuple = bert_model_tester.prepare_config_and_inputs()
lowerCamelCase__ , lowerCamelCase__ : Dict = vision_config_and_inputs
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = text_config_and_inputs
# make sure that cross attention layers are added
return {
"text_config": text_config,
"vision_config": vision_config,
"pixel_values": pixel_values,
"attention_mask": attention_mask,
"input_ids": input_ids,
"token_type_ids": token_type_ids,
}
@require_torch
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
def A_ ( self : Tuple ) -> Union[str, Any]:
lowerCamelCase__ : List[str] = FlaxVisionTextDualEncoderModel.from_vision_text_pretrained(
'hf-internal-testing/tiny-random-clip' , 'hf-internal-testing/tiny-bert' , vision_from_pt=UpperCAmelCase , text_from_pt=UpperCAmelCase , )
lowerCamelCase__ : List[Any] = 13
lowerCamelCase__ : int = floats_tensor(
[
batch_size,
model.config.vision_config.num_channels,
model.config.vision_config.image_size,
model.config.vision_config.image_size,
] )
lowerCamelCase__ : Optional[Any] = ids_tensor([batch_size, 4] , model.config.text_config.vocab_size )
lowerCamelCase__ : Optional[int] = random_attention_mask([batch_size, 4] )
lowerCamelCase__ : Optional[Any] = {'pixel_values': pixel_values, 'input_ids': input_ids, 'attention_mask': attention_mask}
return model, inputs
def A_ ( self : Any , UpperCAmelCase : Tuple , UpperCAmelCase : Tuple ) -> Any:
lowerCamelCase__ : Any = FlaxCLIPVisionModel(UpperCAmelCase )
lowerCamelCase__ : Any = FlaxBertModel(UpperCAmelCase )
return vision_model, text_model
def A_ ( self : List[Any] ) -> List[Any]:
lowerCamelCase__ : Any = FlaxCLIPVisionModelTester(self )
lowerCamelCase__ : List[str] = FlaxBertModelTester(self )
lowerCamelCase__ : Optional[Any] = clip_model_tester.prepare_config_and_inputs()
lowerCamelCase__ : int = bert_model_tester.prepare_config_and_inputs()
lowerCamelCase__ , lowerCamelCase__ : Tuple = vision_config_and_inputs
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : int = text_config_and_inputs
# make sure that cross attention layers are added
return {
"text_config": text_config,
"vision_config": vision_config,
"pixel_values": pixel_values,
"attention_mask": attention_mask,
"input_ids": input_ids,
"token_type_ids": token_type_ids,
}
@require_flax
@require_vision
class lowerCAmelCase ( unittest.TestCase ):
@slow
def A_ ( self : int ) -> str:
lowerCamelCase__ : Tuple = FlaxVisionTextDualEncoderModel.from_pretrained('clip-italian/clip-italian' , logit_scale_init_value=1.0 )
lowerCamelCase__ : int = VisionTextDualEncoderProcessor.from_pretrained('clip-italian/clip-italian' )
lowerCamelCase__ : List[str] = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' )
lowerCamelCase__ : Optional[Any] = processor(
text=['una foto di un gatto', 'una foto di un cane'] , images=UpperCAmelCase , padding=UpperCAmelCase , return_tensors='np' )
lowerCamelCase__ : int = model(**UpperCAmelCase )
# verify the logits
self.assertEqual(outputs.logits_per_image.shape , (inputs.pixel_values.shape[0], inputs.input_ids.shape[0]) )
self.assertEqual(
outputs.logits_per_text.shape , (inputs.input_ids.shape[0], inputs.pixel_values.shape[0]) , )
lowerCamelCase__ : Optional[int] = np.array([[1.2_2_8_4_7_2_7, 0.3_1_0_4_1_2_2]] )
self.assertTrue(np.allclose(outputs.logits_per_image , UpperCAmelCase , atol=1e-3 ) )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> bool:
lowerCamelCase__ : List[str] = len(_UpperCAmelCase )
lowerCamelCase__ : str = [[False] * (required_sum + 1) for _ in range(arr_len + 1 )]
# for each arr value, a sum of zero(0) can be formed by not taking any element
# hence True/1
for i in range(arr_len + 1 ):
lowerCamelCase__ : Tuple = True
# sum is not zero and set is empty then false
for i in range(1 , required_sum + 1 ):
lowerCamelCase__ : Dict = False
for i in range(1 , arr_len + 1 ):
for j in range(1 , required_sum + 1 ):
if arr[i - 1] > j:
lowerCamelCase__ : str = subset[i - 1][j]
if arr[i - 1] <= j:
lowerCamelCase__ : Dict = subset[i - 1][j] or subset[i - 1][j - arr[i - 1]]
return subset[arr_len][required_sum]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
import os
import tempfile
from functools import partial
from unittest import TestCase
from unittest.mock import patch
import numpy as np
import pytest
from datasets.arrow_dataset import Dataset
from datasets.search import ElasticSearchIndex, FaissIndex, MissingIndex
from .utils import require_elasticsearch, require_faiss
_UpperCAmelCase : str = pytest.mark.integration
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : List[Any] ) -> Union[str, Any]:
lowerCamelCase__ : List[Any] = Dataset.from_dict({'filename': ['my_name-train' + '_' + str(UpperCAmelCase ) for x in np.arange(30 ).tolist()]} )
return dset
def A_ ( self : Optional[Any] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
lowerCamelCase__ : List[Any] = dset.map(
lambda UpperCAmelCase , UpperCAmelCase : {"vecs": i * np.ones(5 , dtype=np.floataa )} , with_indices=UpperCAmelCase , keep_in_memory=UpperCAmelCase )
lowerCamelCase__ : Tuple = dset.add_faiss_index('vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
dset.drop_index('vecs' )
def A_ ( self : Union[str, Any] ) -> int:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT , )
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : List[str] ) -> Tuple:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , metric_type=faiss.METRIC_INNER_PRODUCT , )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
dset.save_faiss_index('vecs' , tmp_file.name )
dset.load_faiss_index('vecs2' , tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs2' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' )
dset.drop_index('vecs' )
self.assertRaises(UpperCAmelCase , partial(dset.get_nearest_examples , 'vecs2' , np.ones(5 , dtype=np.floataa ) ) )
def A_ ( self : Dict ) -> Dict:
from elasticsearch import Elasticsearch
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : List[Any] = {'acknowledged': True}
mocked_bulk.return_value([(True, None)] * 30 )
lowerCamelCase__ : int = {'hits': {'hits': [{'_score': 1, '_id': 29}]}}
lowerCamelCase__ : List[str] = Elasticsearch()
dset.add_elasticsearch_index('filename' , es_client=UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Dict = dset.get_nearest_examples('filename' , 'my_name-train_29' )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Any ) -> Dict:
import faiss
lowerCamelCase__ : Tuple = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
# add vectors
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsNotNone(index.faiss_index )
self.assertEqual(index.faiss_index.ntotal , 5 )
index.add_vectors(np.zeros((5, 5) , dtype=np.floataa ) )
self.assertEqual(index.faiss_index.ntotal , 10 )
# single query
lowerCamelCase__ : int = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Any = 1
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = index.search(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search , query.reshape(-1 , 1 ) )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
# batched queries
lowerCamelCase__ : str = np.eye(5 , dtype=np.floataa )[::-1]
lowerCamelCase__ , lowerCamelCase__ : Dict = index.search_batch(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search_batch , queries[0] )
lowerCamelCase__ : str = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[str] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([4, 3, 2, 1, 0] , UpperCAmelCase )
def A_ ( self : List[Any] ) -> List[Any]:
import faiss
lowerCamelCase__ : Any = FaissIndex(string_factory='Flat' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
lowerCamelCase__ : Tuple = FaissIndex(string_factory='LSH' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexLSH )
with self.assertRaises(UpperCAmelCase ):
lowerCamelCase__ : List[str] = FaissIndex(string_factory='Flat' , custom_index=faiss.IndexFlat(5 ) )
def A_ ( self : List[str] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Optional[Any] = faiss.IndexFlat(5 )
lowerCamelCase__ : int = FaissIndex(custom_index=UpperCAmelCase )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
def A_ ( self : Any ) -> Optional[int]:
import faiss
lowerCamelCase__ : Any = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
index.save(tmp_file.name )
lowerCamelCase__ : List[Any] = FaissIndex.load(tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ : List[str] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Tuple = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(UpperCAmelCase )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
@require_faiss
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Any:
import faiss
lowerCamelCase__ : Optional[Any] = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
lowerCamelCase__ : Optional[int] = 'index.faiss'
lowerCamelCase__ : Optional[Any] = F"""mock://{index_name}"""
index.save(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Tuple = FaissIndex.load(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Optional[int] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Dict = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(_UpperCAmelCase )
assert scores[0] > 0
assert indices[0] == 1
@require_elasticsearch
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Dict ) -> List[Any]:
from elasticsearch import Elasticsearch
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : Any = Elasticsearch()
lowerCamelCase__ : Tuple = {'acknowledged': True}
lowerCamelCase__ : Tuple = ElasticSearchIndex(es_client=UpperCAmelCase )
mocked_bulk.return_value([(True, None)] * 3 )
index.add_documents(['foo', 'bar', 'foobar'] )
# single query
lowerCamelCase__ : Optional[int] = 'foo'
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search(UpperCAmelCase )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# single query with timeout
lowerCamelCase__ : Any = 'foo'
lowerCamelCase__ : List[str] = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Tuple = index.search(UpperCAmelCase , request_timeout=30 )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# batched queries
lowerCamelCase__ : List[str] = ['foo', 'bar', 'foobar']
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : str = index.search_batch(UpperCAmelCase )
lowerCamelCase__ : List[str] = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[Any] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
# batched queries with timeout
lowerCamelCase__ : str = ['foo', 'bar', 'foobar']
lowerCamelCase__ : List[Any] = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search_batch(UpperCAmelCase , request_timeout=30 )
lowerCamelCase__ : Optional[Any] = [scores[0] for scores in total_scores]
lowerCamelCase__ : Dict = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
| 50
|
import torch
from transformers import PreTrainedModel, XLMRobertaConfig, XLMRobertaModel
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """M-CLIP"""
def __init__( self : Optional[Any] , UpperCAmelCase : Union[str, Any]=1024 , UpperCAmelCase : Tuple=768 , **UpperCAmelCase : Optional[int] ) -> Dict:
lowerCamelCase__ : Optional[int] = transformerDimSize
lowerCamelCase__ : Optional[Any] = imageDimSize
super().__init__(**UpperCAmelCase )
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = MCLIPConfig
def __init__( self : List[Any] , UpperCAmelCase : Dict , *UpperCAmelCase : Any , **UpperCAmelCase : Dict ) -> Dict:
super().__init__(UpperCAmelCase , *UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Tuple = XLMRobertaModel(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = torch.nn.Linear(
in_features=config.transformerDimensions , out_features=config.numDims )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Dict ) -> Tuple:
lowerCamelCase__ : Any = self.transformer(input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase )[0]
lowerCamelCase__ : int = (embs * attention_mask.unsqueeze(2 )).sum(dim=1 ) / attention_mask.sum(dim=1 )[:, None]
return self.LinearTransformation(UpperCAmelCase ), embs
| 50
| 1
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list:
if n_term == "":
return []
lowerCamelCase__ : list = []
for temp in range(int(_UpperCAmelCase ) ):
series.append(F"""1/{temp + 1}""" if series else '1' )
return series
if __name__ == "__main__":
_UpperCAmelCase : Tuple = input("""Enter the last number (nth term) of the Harmonic Series""")
print("""Formula of Harmonic Series => 1+1/2+1/3 ..... 1/n""")
print(harmonic_series(nth_term))
| 50
|
from itertools import count
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 50 ) -> int:
lowerCamelCase__ : Optional[Any] = [1] * min_block_length
for n in count(_UpperCAmelCase ):
fill_count_functions.append(1 )
for block_length in range(_UpperCAmelCase , n + 1 ):
for block_start in range(n - block_length ):
fill_count_functions[n] += fill_count_functions[
n - block_start - block_length - 1
]
fill_count_functions[n] += 1
if fill_count_functions[n] > 100_0000:
break
return n
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
import flax.linen as nn
import jax
import jax.numpy as jnp
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : Any ) -> Any:
lowerCamelCase__ : str = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : int , UpperCAmelCase : Dict ) -> Optional[Any]:
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : str = hidden_states.shape
lowerCamelCase__ : Union[str, Any] = jax.image.resize(
UpperCAmelCase , shape=(batch, height * 2, width * 2, channels) , method='nearest' , )
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Tuple = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : str , UpperCAmelCase : Union[str, Any] ) -> Optional[Any]:
# pad = ((0, 0), (0, 1), (0, 1), (0, 0)) # pad height and width dim
# hidden_states = jnp.pad(hidden_states, pad_width=pad)
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = 0.0
UpperCAmelCase__ = None
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> Union[str, Any]:
lowerCamelCase__ : Optional[Any] = self.in_channels if self.out_channels is None else self.out_channels
lowerCamelCase__ : Tuple = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : int = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Union[str, Any] = nn.Dense(UpperCAmelCase , dtype=self.dtype )
lowerCamelCase__ : Union[str, Any] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : List[Any] = nn.Dropout(self.dropout_prob )
lowerCamelCase__ : Tuple = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Optional[Any] = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
lowerCamelCase__ : Union[str, Any] = None
if use_nin_shortcut:
lowerCamelCase__ : Dict = nn.Conv(
UpperCAmelCase , kernel_size=(1, 1) , strides=(1, 1) , padding='VALID' , dtype=self.dtype , )
def __call__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : int=True ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = hidden_states
lowerCamelCase__ : List[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[Any] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Any = self.conva(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.time_emb_proj(nn.swish(UpperCAmelCase ) )
lowerCamelCase__ : List[str] = jnp.expand_dims(jnp.expand_dims(UpperCAmelCase , 1 ) , 1 )
lowerCamelCase__ : List[str] = hidden_states + temb
lowerCamelCase__ : Optional[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[str] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = self.dropout(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = self.conva(UpperCAmelCase )
if self.conv_shortcut is not None:
lowerCamelCase__ : Dict = self.conv_shortcut(UpperCAmelCase )
return hidden_states + residual
| 50
|
from __future__ import annotations
from typing import Any
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> None:
create_state_space_tree(_UpperCAmelCase , [] , 0 )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> None:
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__":
_UpperCAmelCase : list[Any] = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["""A""", """B""", """C"""])
generate_all_subsequences(seq)
| 50
| 1
|
import argparse
from pathlib import Path
import fairseq
import torch
from fairseq.models.xmod import XMODModel as FairseqXmodModel
from packaging import version
from transformers import XmodConfig, XmodForMaskedLM, XmodForSequenceClassification
from transformers.utils import logging
if version.parse(fairseq.__version__) < version.parse("""0.12.2"""):
raise Exception("""requires fairseq >= 0.12.2""")
if version.parse(fairseq.__version__) > version.parse("""2"""):
raise Exception("""requires fairseq < v2""")
logging.set_verbosity_info()
_UpperCAmelCase : str = logging.get_logger(__name__)
_UpperCAmelCase : Dict = """Hello, World!"""
_UpperCAmelCase : Any = """en_XX"""
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Any:
lowerCamelCase__ : Optional[Any] = Path('data_bin' )
lowerCamelCase__ : Optional[Any] = FairseqXmodModel.from_pretrained(
model_name_or_path=str(Path(_UpperCAmelCase ).parent ) , checkpoint_file=Path(_UpperCAmelCase ).name , _name='xmod_base' , arch='xmod_base' , task='multilingual_masked_lm' , data_name_or_path=str(_UpperCAmelCase ) , bpe='sentencepiece' , sentencepiece_model=str(Path(_UpperCAmelCase ).parent / 'sentencepiece.bpe.model' ) , src_dict=str(data_dir / 'dict.txt' ) , )
xmod.eval() # disable dropout
print(_UpperCAmelCase )
lowerCamelCase__ : Any = xmod.model.encoder.sentence_encoder
lowerCamelCase__ : Optional[int] = XmodConfig(
vocab_size=xmod_sent_encoder.embed_tokens.num_embeddings , hidden_size=xmod.cfg.model.encoder_embed_dim , num_hidden_layers=xmod.cfg.model.encoder_layers , num_attention_heads=xmod.cfg.model.encoder_attention_heads , intermediate_size=xmod.cfg.model.encoder_ffn_embed_dim , max_position_embeddings=514 , type_vocab_size=1 , layer_norm_eps=1e-5 , pre_norm=xmod.cfg.model.encoder_normalize_before , adapter_reduction_factor=getattr(xmod.cfg.model , 'bottleneck' , 2 ) , adapter_layer_norm=xmod.cfg.model.adapter_layer_norm , adapter_reuse_layer_norm=xmod.cfg.model.adapter_reuse_layer_norm , ln_before_adapter=xmod.cfg.model.ln_before_adapter , languages=xmod.cfg.model.languages , )
if classification_head:
lowerCamelCase__ : Optional[int] = xmod.model.classification_heads['mnli'].out_proj.weight.shape[0]
print('Our X-MOD config:' , _UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = XmodForSequenceClassification(_UpperCAmelCase ) if classification_head else XmodForMaskedLM(_UpperCAmelCase )
model.eval()
# Now let's copy all the weights.
# Embeddings
lowerCamelCase__ : Any = xmod_sent_encoder.embed_tokens.weight
lowerCamelCase__ : str = xmod_sent_encoder.embed_positions.weight
lowerCamelCase__ : Optional[int] = torch.zeros_like(
model.roberta.embeddings.token_type_embeddings.weight ) # just zero them out b/c xmod doesn't use them.
lowerCamelCase__ : Any = xmod_sent_encoder.layernorm_embedding.weight
lowerCamelCase__ : str = xmod_sent_encoder.layernorm_embedding.bias
for i in range(config.num_hidden_layers ):
# Encoder: start of layer
lowerCamelCase__ : List[str] = model.roberta.encoder.layer[i]
lowerCamelCase__ : Dict = xmod_sent_encoder.layers[i]
# self attention
lowerCamelCase__ : Union[str, Any] = layer.attention.self
if not (
xmod_layer.self_attn.k_proj.weight.data.shape
== xmod_layer.self_attn.q_proj.weight.data.shape
== xmod_layer.self_attn.v_proj.weight.data.shape
== torch.Size((config.hidden_size, config.hidden_size) )
):
raise AssertionError('Dimensions of self-attention weights do not match.' )
lowerCamelCase__ : Optional[int] = xmod_layer.self_attn.q_proj.weight
lowerCamelCase__ : Any = xmod_layer.self_attn.q_proj.bias
lowerCamelCase__ : Tuple = xmod_layer.self_attn.k_proj.weight
lowerCamelCase__ : Tuple = xmod_layer.self_attn.k_proj.bias
lowerCamelCase__ : int = xmod_layer.self_attn.v_proj.weight
lowerCamelCase__ : Any = xmod_layer.self_attn.v_proj.bias
# self-attention output
lowerCamelCase__ : str = layer.attention.output
if self_output.dense.weight.shape != xmod_layer.self_attn.out_proj.weight.shape:
raise AssertionError('Dimensions of self-attention output weights do not match.' )
lowerCamelCase__ : Optional[Any] = xmod_layer.self_attn.out_proj.weight
lowerCamelCase__ : Tuple = xmod_layer.self_attn.out_proj.bias
lowerCamelCase__ : Dict = xmod_layer.self_attn_layer_norm.weight
lowerCamelCase__ : Tuple = xmod_layer.self_attn_layer_norm.bias
# intermediate
lowerCamelCase__ : Dict = layer.intermediate
if intermediate.dense.weight.shape != xmod_layer.fca.weight.shape:
raise AssertionError('Dimensions of intermediate weights do not match.' )
lowerCamelCase__ : Tuple = xmod_layer.fca.weight
lowerCamelCase__ : List[str] = xmod_layer.fca.bias
# output
lowerCamelCase__ : List[str] = layer.output
if bert_output.dense.weight.shape != xmod_layer.fca.weight.shape:
raise AssertionError('Dimensions of feed-forward weights do not match.' )
lowerCamelCase__ : Union[str, Any] = xmod_layer.fca.weight
lowerCamelCase__ : int = xmod_layer.fca.bias
lowerCamelCase__ : List[Any] = xmod_layer.final_layer_norm.weight
lowerCamelCase__ : Union[str, Any] = xmod_layer.final_layer_norm.bias
if bert_output.adapter_layer_norm is not None:
lowerCamelCase__ : Dict = xmod_layer.adapter_layer_norm.weight
lowerCamelCase__ : Tuple = xmod_layer.adapter_layer_norm.bias
if sorted(bert_output.adapter_modules.keys() ) != sorted(xmod_layer.adapter_modules.keys() ):
raise AssertionError('Lists of language adapters do not match.' )
for lang_code, adapter in xmod_layer.adapter_modules.items():
lowerCamelCase__ : Optional[int] = bert_output.adapter_modules[lang_code]
lowerCamelCase__ : List[Any] = xmod_layer.adapter_modules[lang_code]
lowerCamelCase__ : Dict = from_adapter.fca.weight
lowerCamelCase__ : Any = from_adapter.fca.bias
lowerCamelCase__ : Optional[Any] = from_adapter.fca.weight
lowerCamelCase__ : List[str] = from_adapter.fca.bias
# end of layer
if xmod_sent_encoder.layer_norm is not None:
lowerCamelCase__ : List[Any] = xmod_sent_encoder.layer_norm.weight
lowerCamelCase__ : List[str] = xmod_sent_encoder.layer_norm.bias
if classification_head:
lowerCamelCase__ : Union[str, Any] = xmod.model.classification_heads['mnli'].dense.weight
lowerCamelCase__ : Any = xmod.model.classification_heads['mnli'].dense.bias
lowerCamelCase__ : int = xmod.model.classification_heads['mnli'].out_proj.weight
lowerCamelCase__ : Union[str, Any] = xmod.model.classification_heads['mnli'].out_proj.bias
else:
# LM Head
lowerCamelCase__ : Optional[int] = xmod.model.encoder.lm_head.dense.weight
lowerCamelCase__ : Any = xmod.model.encoder.lm_head.dense.bias
lowerCamelCase__ : Dict = xmod.model.encoder.lm_head.layer_norm.weight
lowerCamelCase__ : str = xmod.model.encoder.lm_head.layer_norm.bias
lowerCamelCase__ : str = xmod.model.encoder.lm_head.weight
lowerCamelCase__ : Optional[Any] = xmod.model.encoder.lm_head.bias
# Let's check that we get the same results.
lowerCamelCase__ : Optional[Any] = xmod.encode(_UpperCAmelCase ).unsqueeze(0 ) # batch of size 1
model.roberta.set_default_language(_UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = model(_UpperCAmelCase )[0]
if classification_head:
lowerCamelCase__ : int = xmod.model.classification_heads['mnli'](xmod.extract_features(_UpperCAmelCase ) )
else:
lowerCamelCase__ : Any = xmod.model(_UpperCAmelCase , lang_id=[SAMPLE_LANGUAGE] )[0]
print(our_output.shape , their_output.shape )
lowerCamelCase__ : Any = torch.max(torch.abs(our_output - their_output ) ).item()
print(F"""max_absolute_diff = {max_absolute_diff}""" ) # ~ 1e-7
lowerCamelCase__ : Any = torch.allclose(_UpperCAmelCase , _UpperCAmelCase , atol=1e-3 )
print('Do both models output the same tensors?' , '🔥' if success else '💩' )
if not success:
raise Exception('Something went wRoNg' )
Path(_UpperCAmelCase ).mkdir(parents=_UpperCAmelCase , exist_ok=_UpperCAmelCase )
print(F"""Saving model to {pytorch_dump_folder_path}""" )
model.save_pretrained(_UpperCAmelCase )
if __name__ == "__main__":
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--xmod_checkpoint_path""", default=None, type=str, required=True, help="""Path the official PyTorch dump."""
)
parser.add_argument(
"""--pytorch_dump_folder_path""", default=None, type=str, required=True, help="""Path to the output PyTorch model."""
)
parser.add_argument(
"""--classification_head""", action="""store_true""", help="""Whether to convert a final classification head."""
)
_UpperCAmelCase : Tuple = parser.parse_args()
convert_xmod_checkpoint_to_pytorch(
args.xmod_checkpoint_path, args.pytorch_dump_folder_path, args.classification_head
)
| 50
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from tokenizers import processors
from ...tokenization_utils import AddedToken, BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_nllb import NllbTokenizer
else:
_UpperCAmelCase : int = None
_UpperCAmelCase : Dict = logging.get_logger(__name__)
_UpperCAmelCase : Optional[int] = {"""vocab_file""": """sentencepiece.bpe.model""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : List[Any] = {
"""vocab_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/sentencepiece.bpe.model"""
),
},
"""tokenizer_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : List[str] = {
"""facebook/nllb-large-en-ro""": 10_24,
"""facebook/nllb-200-distilled-600M""": 10_24,
}
# fmt: off
_UpperCAmelCase : Optional[int] = ["""ace_Arab""", """ace_Latn""", """acm_Arab""", """acq_Arab""", """aeb_Arab""", """afr_Latn""", """ajp_Arab""", """aka_Latn""", """amh_Ethi""", """apc_Arab""", """arb_Arab""", """ars_Arab""", """ary_Arab""", """arz_Arab""", """asm_Beng""", """ast_Latn""", """awa_Deva""", """ayr_Latn""", """azb_Arab""", """azj_Latn""", """bak_Cyrl""", """bam_Latn""", """ban_Latn""", """bel_Cyrl""", """bem_Latn""", """ben_Beng""", """bho_Deva""", """bjn_Arab""", """bjn_Latn""", """bod_Tibt""", """bos_Latn""", """bug_Latn""", """bul_Cyrl""", """cat_Latn""", """ceb_Latn""", """ces_Latn""", """cjk_Latn""", """ckb_Arab""", """crh_Latn""", """cym_Latn""", """dan_Latn""", """deu_Latn""", """dik_Latn""", """dyu_Latn""", """dzo_Tibt""", """ell_Grek""", """eng_Latn""", """epo_Latn""", """est_Latn""", """eus_Latn""", """ewe_Latn""", """fao_Latn""", """pes_Arab""", """fij_Latn""", """fin_Latn""", """fon_Latn""", """fra_Latn""", """fur_Latn""", """fuv_Latn""", """gla_Latn""", """gle_Latn""", """glg_Latn""", """grn_Latn""", """guj_Gujr""", """hat_Latn""", """hau_Latn""", """heb_Hebr""", """hin_Deva""", """hne_Deva""", """hrv_Latn""", """hun_Latn""", """hye_Armn""", """ibo_Latn""", """ilo_Latn""", """ind_Latn""", """isl_Latn""", """ita_Latn""", """jav_Latn""", """jpn_Jpan""", """kab_Latn""", """kac_Latn""", """kam_Latn""", """kan_Knda""", """kas_Arab""", """kas_Deva""", """kat_Geor""", """knc_Arab""", """knc_Latn""", """kaz_Cyrl""", """kbp_Latn""", """kea_Latn""", """khm_Khmr""", """kik_Latn""", """kin_Latn""", """kir_Cyrl""", """kmb_Latn""", """kon_Latn""", """kor_Hang""", """kmr_Latn""", """lao_Laoo""", """lvs_Latn""", """lij_Latn""", """lim_Latn""", """lin_Latn""", """lit_Latn""", """lmo_Latn""", """ltg_Latn""", """ltz_Latn""", """lua_Latn""", """lug_Latn""", """luo_Latn""", """lus_Latn""", """mag_Deva""", """mai_Deva""", """mal_Mlym""", """mar_Deva""", """min_Latn""", """mkd_Cyrl""", """plt_Latn""", """mlt_Latn""", """mni_Beng""", """khk_Cyrl""", """mos_Latn""", """mri_Latn""", """zsm_Latn""", """mya_Mymr""", """nld_Latn""", """nno_Latn""", """nob_Latn""", """npi_Deva""", """nso_Latn""", """nus_Latn""", """nya_Latn""", """oci_Latn""", """gaz_Latn""", """ory_Orya""", """pag_Latn""", """pan_Guru""", """pap_Latn""", """pol_Latn""", """por_Latn""", """prs_Arab""", """pbt_Arab""", """quy_Latn""", """ron_Latn""", """run_Latn""", """rus_Cyrl""", """sag_Latn""", """san_Deva""", """sat_Beng""", """scn_Latn""", """shn_Mymr""", """sin_Sinh""", """slk_Latn""", """slv_Latn""", """smo_Latn""", """sna_Latn""", """snd_Arab""", """som_Latn""", """sot_Latn""", """spa_Latn""", """als_Latn""", """srd_Latn""", """srp_Cyrl""", """ssw_Latn""", """sun_Latn""", """swe_Latn""", """swh_Latn""", """szl_Latn""", """tam_Taml""", """tat_Cyrl""", """tel_Telu""", """tgk_Cyrl""", """tgl_Latn""", """tha_Thai""", """tir_Ethi""", """taq_Latn""", """taq_Tfng""", """tpi_Latn""", """tsn_Latn""", """tso_Latn""", """tuk_Latn""", """tum_Latn""", """tur_Latn""", """twi_Latn""", """tzm_Tfng""", """uig_Arab""", """ukr_Cyrl""", """umb_Latn""", """urd_Arab""", """uzn_Latn""", """vec_Latn""", """vie_Latn""", """war_Latn""", """wol_Latn""", """xho_Latn""", """ydd_Hebr""", """yor_Latn""", """yue_Hant""", """zho_Hans""", """zho_Hant""", """zul_Latn"""]
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = ["""input_ids""", """attention_mask"""]
UpperCAmelCase__ = NllbTokenizer
UpperCAmelCase__ = []
UpperCAmelCase__ = []
def __init__( self : Tuple , UpperCAmelCase : int=None , UpperCAmelCase : Any=None , UpperCAmelCase : str="<s>" , UpperCAmelCase : Optional[Any]="</s>" , UpperCAmelCase : str="</s>" , UpperCAmelCase : Tuple="<s>" , UpperCAmelCase : Optional[Any]="<unk>" , UpperCAmelCase : List[str]="<pad>" , UpperCAmelCase : Union[str, Any]="<mask>" , UpperCAmelCase : Tuple=None , UpperCAmelCase : int=None , UpperCAmelCase : Dict=None , UpperCAmelCase : Any=False , **UpperCAmelCase : Optional[int] , ) -> Tuple:
# Mask token behave like a normal word, i.e. include the space before it
lowerCamelCase__ : List[Any] = AddedToken(UpperCAmelCase , lstrip=UpperCAmelCase , rstrip=UpperCAmelCase ) if isinstance(UpperCAmelCase , UpperCAmelCase ) else mask_token
lowerCamelCase__ : Union[str, Any] = legacy_behaviour
super().__init__(
vocab_file=UpperCAmelCase , tokenizer_file=UpperCAmelCase , bos_token=UpperCAmelCase , eos_token=UpperCAmelCase , sep_token=UpperCAmelCase , cls_token=UpperCAmelCase , unk_token=UpperCAmelCase , pad_token=UpperCAmelCase , mask_token=UpperCAmelCase , src_lang=UpperCAmelCase , tgt_lang=UpperCAmelCase , additional_special_tokens=UpperCAmelCase , legacy_behaviour=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = vocab_file
lowerCamelCase__ : Dict = False if not self.vocab_file else True
lowerCamelCase__ : Optional[Any] = FAIRSEQ_LANGUAGE_CODES.copy()
if additional_special_tokens is not None:
# Only add those special tokens if they are not already there.
_additional_special_tokens.extend(
[t for t in additional_special_tokens if t not in _additional_special_tokens] )
self.add_special_tokens({'additional_special_tokens': _additional_special_tokens} )
lowerCamelCase__ : str = {
lang_code: self.convert_tokens_to_ids(UpperCAmelCase ) for lang_code in FAIRSEQ_LANGUAGE_CODES
}
lowerCamelCase__ : int = src_lang if src_lang is not None else 'eng_Latn'
lowerCamelCase__ : List[Any] = self.convert_tokens_to_ids(self._src_lang )
lowerCamelCase__ : str = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def A_ ( self : int ) -> str:
return self._src_lang
@src_lang.setter
def A_ ( self : List[Any] , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Any = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
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 A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : Dict = [self.sep_token_id]
lowerCamelCase__ : List[str] = [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 + sep + token_ids_a + sep ) * [0]
def A_ ( self : int , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Optional[str] , UpperCAmelCase : Optional[str] , **UpperCAmelCase : List[str] ) -> Dict:
if src_lang is None or tgt_lang is None:
raise ValueError('Translation requires a `src_lang` and a `tgt_lang` for this model' )
lowerCamelCase__ : Optional[int] = src_lang
lowerCamelCase__ : Optional[int] = self(UpperCAmelCase , add_special_tokens=UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.convert_tokens_to_ids(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tgt_lang_id
return inputs
def A_ ( self : Dict , UpperCAmelCase : List[str] , UpperCAmelCase : str = "eng_Latn" , UpperCAmelCase : Optional[List[str]] = None , UpperCAmelCase : str = "fra_Latn" , **UpperCAmelCase : Dict , ) -> BatchEncoding:
lowerCamelCase__ : Any = src_lang
lowerCamelCase__ : int = tgt_lang
return super().prepare_seqaseq_batch(UpperCAmelCase , UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Union[str, Any] ) -> Optional[int]:
return self.set_src_lang_special_tokens(self.src_lang )
def A_ ( self : Any ) -> Union[str, Any]:
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def A_ ( self : str , UpperCAmelCase : Optional[Any] ) -> None:
lowerCamelCase__ : int = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : int = []
lowerCamelCase__ : str = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : int = [self.cur_lang_code]
lowerCamelCase__ : Tuple = [self.eos_token_id]
lowerCamelCase__ : Any = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : Optional[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : str = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : int , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Union[str, Any] = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : Dict = []
lowerCamelCase__ : Union[str, Any] = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : Any = [self.cur_lang_code]
lowerCamelCase__ : Optional[Any] = [self.eos_token_id]
lowerCamelCase__ : Union[str, Any] = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : List[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : Optional[int] = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : Union[str, Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not self.can_save_slow_tokenizer:
raise ValueError(
'Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '
'tokenizer.' )
if not os.path.isdir(UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory.""" )
return
lowerCamelCase__ : int = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCAmelCase ):
copyfile(self.vocab_file , UpperCAmelCase )
return (out_vocab_file,)
| 50
| 1
|
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
_UpperCAmelCase : Union[str, Any] = [
{"""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 SCREAMING_SNAKE_CASE ( _UpperCAmelCase=True ) -> Optional[Any]:
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=__UpperCamelCase ) )
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = None
UpperCAmelCase__ = None
def A_ ( self : List[Any] , UpperCAmelCase : List[Any] , UpperCAmelCase : Union[str, Any] ) -> Dict:
with TemporaryDirectory() as tmp_dir:
lowerCamelCase__ : List[Any] = dataset_module_factory(UpperCAmelCase , cache_dir=UpperCAmelCase )
lowerCamelCase__ : int = import_main_class(dataset_module.module_path , dataset=UpperCAmelCase )
lowerCamelCase__ : DatasetBuilder = builder_cls(
cache_dir=UpperCAmelCase , config_name=UpperCAmelCase , hash=dataset_module.hash , )
lowerCamelCase__ : int = '/'.join(
[
HF_GCP_BASE_URL,
builder_instance._relative_data_dir(with_hash=UpperCAmelCase ).replace(os.sep , '/' ),
config.DATASET_INFO_FILENAME,
] )
lowerCamelCase__ : Optional[int] = cached_path(UpperCAmelCase , cache_dir=UpperCAmelCase )
self.assertTrue(os.path.exists(UpperCAmelCase ) )
@pytest.mark.integration
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Tuple = tmp_path_factory.mktemp('test_hf_gcp' ) / 'test_wikipedia_simple'
lowerCamelCase__ : Tuple = dataset_module_factory('wikipedia' , cache_dir=_UpperCAmelCase )
lowerCamelCase__ : str = import_main_class(dataset_module.module_path )
lowerCamelCase__ : DatasetBuilder = builder_cls(
cache_dir=_UpperCAmelCase , 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__ : str = builder_instance.as_dataset()
assert ds
@pytest.mark.integration
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : str = dataset_module_factory('wikipedia' , cache_dir=_UpperCAmelCase )
lowerCamelCase__ : Dict = import_main_class(dataset_module.module_path , dataset=_UpperCAmelCase )
lowerCamelCase__ : DatasetBuilder = builder_cls(
cache_dir=_UpperCAmelCase , config_name='20220301.frr' , hash=dataset_module.hash , )
lowerCamelCase__ : Tuple = builder_instance.as_streaming_dataset()
assert ds
assert isinstance(_UpperCAmelCase , _UpperCAmelCase )
assert "train" in ds
assert isinstance(ds['train'] , _UpperCAmelCase )
assert next(iter(ds['train'] ) )
| 50
|
import argparse
import torch
from torch import nn
from transformers import MaMaaaConfig, MaMaaaForConditionalGeneration
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Optional[int] = [
'encoder.version',
'decoder.version',
'model.encoder.version',
'model.decoder.version',
'decoder.output_projection.weight',
'_float_tensor',
'encoder.embed_positions._float_tensor',
'decoder.embed_positions._float_tensor',
]
for k in ignore_keys:
state_dict.pop(_UpperCAmelCase , _UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[Any]:
lowerCamelCase__ , lowerCamelCase__ : List[str] = emb.weight.shape
lowerCamelCase__ : Tuple = nn.Linear(_UpperCAmelCase , _UpperCAmelCase , bias=_UpperCAmelCase )
lowerCamelCase__ : Dict = emb.weight.data
return lin_layer
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Tuple = torch.load(_UpperCAmelCase , map_location='cpu' )
lowerCamelCase__ : List[str] = mam_aaa['args'] or mam_aaa['cfg']['model']
lowerCamelCase__ : Optional[int] = mam_aaa['model']
remove_ignore_keys_(_UpperCAmelCase )
lowerCamelCase__ : str = state_dict['encoder.embed_tokens.weight'].shape[0]
lowerCamelCase__ : Union[str, Any] = MaMaaaConfig(
vocab_size=_UpperCAmelCase , max_position_embeddings=1024 , encoder_layers=args.encoder_layers , decoder_layers=args.decoder_layers , encoder_attention_heads=args.encoder_attention_heads , decoder_attention_heads=args.decoder_attention_heads , encoder_ffn_dim=args.encoder_ffn_embed_dim , decoder_ffn_dim=args.decoder_ffn_embed_dim , d_model=args.encoder_embed_dim , encoder_layerdrop=args.encoder_layerdrop , decoder_layerdrop=args.decoder_layerdrop , dropout=args.dropout , attention_dropout=args.attention_dropout , activation_dropout=args.activation_dropout , activation_function='relu' , )
lowerCamelCase__ : Optional[Any] = state_dict['decoder.embed_tokens.weight']
lowerCamelCase__ : Union[str, Any] = MaMaaaForConditionalGeneration(_UpperCAmelCase )
model.model.load_state_dict(_UpperCAmelCase , strict=_UpperCAmelCase )
lowerCamelCase__ : List[str] = make_linear_from_emb(model.model.shared )
return model
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""fairseq_path""", type=str, help="""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.""")
_UpperCAmelCase : str = parser.parse_args()
_UpperCAmelCase : Optional[Any] = convert_fairseq_mamaaa_checkpoint_from_disk(args.fairseq_pathß)
model.save_pretrained(args.pytorch_dump_folder_path)
| 50
| 1
|
import sys
import tempfile
import unittest
import unittest.mock as mock
from pathlib import Path
from huggingface_hub import HfFolder, delete_repo
from requests.exceptions import HTTPError
from transformers import AutoImageProcessor, ViTImageProcessor
from transformers.testing_utils import TOKEN, USER, get_tests_dir, is_staging_test
sys.path.append(str(Path(__file__).parent.parent / """utils"""))
from test_module.custom_image_processing import CustomImageProcessor # noqa E402
_UpperCAmelCase : Union[str, Any] = get_tests_dir("""fixtures""")
class lowerCAmelCase ( unittest.TestCase ):
def A_ ( self : Dict ) -> Tuple:
# A mock response for an HTTP head request to emulate server down
lowerCamelCase__ : Optional[Any] = mock.Mock()
lowerCamelCase__ : Dict = 500
lowerCamelCase__ : Any = {}
lowerCamelCase__ : Any = HTTPError
lowerCamelCase__ : int = {}
# Download this model to make sure it's in the cache.
lowerCamelCase__ : Any = ViTImageProcessor.from_pretrained('hf-internal-testing/tiny-random-vit' )
# Under the mock environment we get a 500 error when trying to reach the model.
with mock.patch('requests.Session.request' , return_value=UpperCAmelCase ) as mock_head:
lowerCamelCase__ : Tuple = ViTImageProcessor.from_pretrained('hf-internal-testing/tiny-random-vit' )
# This check we did call the fake head request
mock_head.assert_called()
def A_ ( self : Optional[int] ) -> Tuple:
# This test is for deprecated behavior and can be removed in v5
lowerCamelCase__ : Optional[Any] = ViTImageProcessor.from_pretrained(
'https://huggingface.co/hf-internal-testing/tiny-random-vit/resolve/main/preprocessor_config.json' )
def A_ ( self : str ) -> Tuple:
with self.assertRaises(UpperCAmelCase ):
# config is in subfolder, the following should not work without specifying the subfolder
lowerCamelCase__ : List[Any] = AutoImageProcessor.from_pretrained('hf-internal-testing/stable-diffusion-all-variants' )
lowerCamelCase__ : List[str] = AutoImageProcessor.from_pretrained(
'hf-internal-testing/stable-diffusion-all-variants' , subfolder='feature_extractor' )
self.assertIsNotNone(UpperCAmelCase )
@is_staging_test
class lowerCAmelCase ( unittest.TestCase ):
@classmethod
def A_ ( cls : str ) -> int:
lowerCamelCase__ : Union[str, Any] = TOKEN
HfFolder.save_token(UpperCAmelCase )
@classmethod
def A_ ( cls : Tuple ) -> Dict:
try:
delete_repo(token=cls._token , repo_id='test-image-processor' )
except HTTPError:
pass
try:
delete_repo(token=cls._token , repo_id='valid_org/test-image-processor-org' )
except HTTPError:
pass
try:
delete_repo(token=cls._token , repo_id='test-dynamic-image-processor' )
except HTTPError:
pass
def A_ ( self : Any ) -> Dict:
lowerCamelCase__ : Dict = ViTImageProcessor.from_pretrained(UpperCAmelCase )
image_processor.push_to_hub('test-image-processor' , use_auth_token=self._token )
lowerCamelCase__ : Any = ViTImageProcessor.from_pretrained(F"""{USER}/test-image-processor""" )
for k, v in image_processor.__dict__.items():
self.assertEqual(UpperCAmelCase , getattr(UpperCAmelCase , UpperCAmelCase ) )
# Reset repo
delete_repo(token=self._token , repo_id='test-image-processor' )
# Push to hub via save_pretrained
with tempfile.TemporaryDirectory() as tmp_dir:
image_processor.save_pretrained(
UpperCAmelCase , repo_id='test-image-processor' , push_to_hub=UpperCAmelCase , use_auth_token=self._token )
lowerCamelCase__ : Tuple = ViTImageProcessor.from_pretrained(F"""{USER}/test-image-processor""" )
for k, v in image_processor.__dict__.items():
self.assertEqual(UpperCAmelCase , getattr(UpperCAmelCase , UpperCAmelCase ) )
def A_ ( self : List[Any] ) -> Tuple:
lowerCamelCase__ : List[str] = ViTImageProcessor.from_pretrained(UpperCAmelCase )
image_processor.push_to_hub('valid_org/test-image-processor' , use_auth_token=self._token )
lowerCamelCase__ : List[str] = ViTImageProcessor.from_pretrained('valid_org/test-image-processor' )
for k, v in image_processor.__dict__.items():
self.assertEqual(UpperCAmelCase , getattr(UpperCAmelCase , UpperCAmelCase ) )
# Reset repo
delete_repo(token=self._token , repo_id='valid_org/test-image-processor' )
# Push to hub via save_pretrained
with tempfile.TemporaryDirectory() as tmp_dir:
image_processor.save_pretrained(
UpperCAmelCase , repo_id='valid_org/test-image-processor-org' , push_to_hub=UpperCAmelCase , use_auth_token=self._token )
lowerCamelCase__ : Tuple = ViTImageProcessor.from_pretrained('valid_org/test-image-processor-org' )
for k, v in image_processor.__dict__.items():
self.assertEqual(UpperCAmelCase , getattr(UpperCAmelCase , UpperCAmelCase ) )
def A_ ( self : List[str] ) -> List[Any]:
CustomImageProcessor.register_for_auto_class()
lowerCamelCase__ : Union[str, Any] = CustomImageProcessor.from_pretrained(UpperCAmelCase )
image_processor.push_to_hub('test-dynamic-image-processor' , use_auth_token=self._token )
# This has added the proper auto_map field to the config
self.assertDictEqual(
image_processor.auto_map , {'AutoImageProcessor': 'custom_image_processing.CustomImageProcessor'} , )
lowerCamelCase__ : int = AutoImageProcessor.from_pretrained(
F"""{USER}/test-dynamic-image-processor""" , trust_remote_code=UpperCAmelCase )
# Can't make an isinstance check because the new_image_processor is from the CustomImageProcessor class of a dynamic module
self.assertEqual(new_image_processor.__class__.__name__ , 'CustomImageProcessor' )
| 50
|
import argparse
import json
import os
import time
import zipfile
from get_ci_error_statistics import download_artifact, get_artifacts_links
from transformers import logging
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Dict:
lowerCamelCase__ : str = set()
lowerCamelCase__ : Any = []
def parse_line(_UpperCAmelCase ):
for line in fp:
if isinstance(_UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Any = line.decode('UTF-8' )
if "warnings summary (final)" in line:
continue
# This means we are outside the body of a warning
elif not line.startswith(' ' ):
# process a single warning and move it to `selected_warnings`.
if len(_UpperCAmelCase ) > 0:
lowerCamelCase__ : str = '\n'.join(_UpperCAmelCase )
# Only keep the warnings specified in `targets`
if any(F""": {x}: """ in warning for x in targets ):
selected_warnings.add(_UpperCAmelCase )
buffer.clear()
continue
else:
lowerCamelCase__ : List[str] = line.strip()
buffer.append(_UpperCAmelCase )
if from_gh:
for filename in os.listdir(_UpperCAmelCase ):
lowerCamelCase__ : Dict = os.path.join(_UpperCAmelCase , _UpperCAmelCase )
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
else:
try:
with zipfile.ZipFile(_UpperCAmelCase ) as z:
for filename in z.namelist():
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with z.open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
except Exception:
logger.warning(
F"""{artifact_path} is either an invalid zip file or something else wrong. This file is skipped.""" )
return selected_warnings
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = set()
lowerCamelCase__ : Optional[int] = [os.path.join(_UpperCAmelCase , _UpperCAmelCase ) for p in os.listdir(_UpperCAmelCase ) if (p.endswith('.zip' ) or from_gh)]
for p in paths:
selected_warnings.update(extract_warnings_from_single_artifact(_UpperCAmelCase , _UpperCAmelCase ) )
return selected_warnings
if __name__ == "__main__":
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
return values.split(',' )
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""--workflow_run_id""", type=str, required=True, help="""A GitHub Actions workflow run id.""")
parser.add_argument(
"""--output_dir""",
type=str,
required=True,
help="""Where to store the downloaded artifacts and other result files.""",
)
parser.add_argument("""--token""", default=None, type=str, help="""A token that has actions:read permission.""")
# optional parameters
parser.add_argument(
"""--targets""",
default="""DeprecationWarning,UserWarning,FutureWarning""",
type=list_str,
help="""Comma-separated list of target warning(s) which we want to extract.""",
)
parser.add_argument(
"""--from_gh""",
action="""store_true""",
help="""If running from a GitHub action workflow and collecting warnings from its artifacts.""",
)
_UpperCAmelCase : Union[str, Any] = parser.parse_args()
_UpperCAmelCase : Dict = args.from_gh
if from_gh:
# The artifacts have to be downloaded using `actions/download-artifact@v3`
pass
else:
os.makedirs(args.output_dir, exist_ok=True)
# get download links
_UpperCAmelCase : Union[str, Any] = get_artifacts_links(args.workflow_run_id, token=args.token)
with open(os.path.join(args.output_dir, """artifacts.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(artifacts, fp, ensure_ascii=False, indent=4)
# download artifacts
for idx, (name, url) in enumerate(artifacts.items()):
print(name)
print(url)
print("""=""" * 80)
download_artifact(name, url, args.output_dir, args.token)
# Be gentle to GitHub
time.sleep(1)
# extract warnings from artifacts
_UpperCAmelCase : Dict = extract_warnings(args.output_dir, args.targets)
_UpperCAmelCase : Optional[Any] = sorted(selected_warnings)
with open(os.path.join(args.output_dir, """selected_warnings.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(selected_warnings, fp, ensure_ascii=False, indent=4)
| 50
| 1
|
import argparse
import torch
from diffusers.pipelines.stable_diffusion.convert_from_ckpt import download_from_original_stable_diffusion_ckpt
if __name__ == "__main__":
_UpperCAmelCase : str = argparse.ArgumentParser()
parser.add_argument(
"""--checkpoint_path""", default=None, type=str, required=True, help="""Path to the checkpoint to convert."""
)
# !wget https://raw.githubusercontent.com/CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml
parser.add_argument(
"""--original_config_file""",
default=None,
type=str,
help="""The YAML config file corresponding to the original architecture.""",
)
parser.add_argument(
"""--num_in_channels""",
default=None,
type=int,
help="""The number of input channels. If `None` number of input channels will be automatically inferred.""",
)
parser.add_argument(
"""--scheduler_type""",
default="""pndm""",
type=str,
help="""Type of scheduler to use. Should be one of ['pndm', 'lms', 'ddim', 'euler', 'euler-ancestral', 'dpm']""",
)
parser.add_argument(
"""--pipeline_type""",
default=None,
type=str,
help=(
"""The pipeline type. One of 'FrozenOpenCLIPEmbedder', 'FrozenCLIPEmbedder', 'PaintByExample'"""
""". If `None` pipeline will be automatically inferred."""
),
)
parser.add_argument(
"""--image_size""",
default=None,
type=int,
help=(
"""The image size that the model was trained on. Use 512 for Stable Diffusion v1.X and Stable Siffusion v2"""
""" Base. Use 768 for Stable Diffusion v2."""
),
)
parser.add_argument(
"""--prediction_type""",
default=None,
type=str,
help=(
"""The prediction type that the model was trained on. Use 'epsilon' for Stable Diffusion v1.X and Stable"""
""" Diffusion v2 Base. Use 'v_prediction' for Stable Diffusion v2."""
),
)
parser.add_argument(
"""--extract_ema""",
action="""store_true""",
help=(
"""Only relevant for checkpoints that have both EMA and non-EMA weights. Whether to extract the EMA weights"""
""" or not. Defaults to `False`. Add `--extract_ema` to extract the EMA weights. EMA weights usually yield"""
""" higher quality images for inference. Non-EMA weights are usually better to continue fine-tuning."""
),
)
parser.add_argument(
"""--upcast_attention""",
action="""store_true""",
help=(
"""Whether the attention computation should always be upcasted. This is necessary when running stable"""
""" diffusion 2.1."""
),
)
parser.add_argument(
"""--from_safetensors""",
action="""store_true""",
help="""If `--checkpoint_path` is in `safetensors` format, load checkpoint with safetensors instead of PyTorch.""",
)
parser.add_argument(
"""--to_safetensors""",
action="""store_true""",
help="""Whether to store pipeline in safetensors format or not.""",
)
parser.add_argument("""--dump_path""", default=None, type=str, required=True, help="""Path to the output model.""")
parser.add_argument("""--device""", type=str, help="""Device to use (e.g. cpu, cuda:0, cuda:1, etc.)""")
parser.add_argument(
"""--stable_unclip""",
type=str,
default=None,
required=False,
help="""Set if this is a stable unCLIP model. One of 'txt2img' or 'img2img'.""",
)
parser.add_argument(
"""--stable_unclip_prior""",
type=str,
default=None,
required=False,
help="""Set if this is a stable unCLIP txt2img model. Selects which prior to use. If `--stable_unclip` is set to `txt2img`, the karlo prior (https://huggingface.co/kakaobrain/karlo-v1-alpha/tree/main/prior) is selected by default.""",
)
parser.add_argument(
"""--clip_stats_path""",
type=str,
help="""Path to the clip stats file. Only required if the stable unclip model's config specifies `model.params.noise_aug_config.params.clip_stats_path`.""",
required=False,
)
parser.add_argument(
"""--controlnet""", action="""store_true""", default=None, help="""Set flag if this is a controlnet checkpoint."""
)
parser.add_argument("""--half""", action="""store_true""", help="""Save weights in half precision.""")
parser.add_argument(
"""--vae_path""",
type=str,
default=None,
required=False,
help="""Set to a path, hub id to an already converted vae to not convert it again.""",
)
_UpperCAmelCase : List[Any] = parser.parse_args()
_UpperCAmelCase : str = download_from_original_stable_diffusion_ckpt(
checkpoint_path=args.checkpoint_path,
original_config_file=args.original_config_file,
image_size=args.image_size,
prediction_type=args.prediction_type,
model_type=args.pipeline_type,
extract_ema=args.extract_ema,
scheduler_type=args.scheduler_type,
num_in_channels=args.num_in_channels,
upcast_attention=args.upcast_attention,
from_safetensors=args.from_safetensors,
device=args.device,
stable_unclip=args.stable_unclip,
stable_unclip_prior=args.stable_unclip_prior,
clip_stats_path=args.clip_stats_path,
controlnet=args.controlnet,
vae_path=args.vae_path,
)
if args.half:
pipe.to(torch_dtype=torch.floataa)
if args.controlnet:
# only save the controlnet model
pipe.controlnet.save_pretrained(args.dump_path, safe_serialization=args.to_safetensors)
else:
pipe.save_pretrained(args.dump_path, safe_serialization=args.to_safetensors)
| 50
|
import flax.linen as nn
import jax
import jax.numpy as jnp
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : Any ) -> Any:
lowerCamelCase__ : str = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : int , UpperCAmelCase : Dict ) -> Optional[Any]:
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : str = hidden_states.shape
lowerCamelCase__ : Union[str, Any] = jax.image.resize(
UpperCAmelCase , shape=(batch, height * 2, width * 2, channels) , method='nearest' , )
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Tuple = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : str , UpperCAmelCase : Union[str, Any] ) -> Optional[Any]:
# pad = ((0, 0), (0, 1), (0, 1), (0, 0)) # pad height and width dim
# hidden_states = jnp.pad(hidden_states, pad_width=pad)
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = 0.0
UpperCAmelCase__ = None
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> Union[str, Any]:
lowerCamelCase__ : Optional[Any] = self.in_channels if self.out_channels is None else self.out_channels
lowerCamelCase__ : Tuple = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : int = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Union[str, Any] = nn.Dense(UpperCAmelCase , dtype=self.dtype )
lowerCamelCase__ : Union[str, Any] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : List[Any] = nn.Dropout(self.dropout_prob )
lowerCamelCase__ : Tuple = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Optional[Any] = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
lowerCamelCase__ : Union[str, Any] = None
if use_nin_shortcut:
lowerCamelCase__ : Dict = nn.Conv(
UpperCAmelCase , kernel_size=(1, 1) , strides=(1, 1) , padding='VALID' , dtype=self.dtype , )
def __call__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : int=True ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = hidden_states
lowerCamelCase__ : List[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[Any] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Any = self.conva(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.time_emb_proj(nn.swish(UpperCAmelCase ) )
lowerCamelCase__ : List[str] = jnp.expand_dims(jnp.expand_dims(UpperCAmelCase , 1 ) , 1 )
lowerCamelCase__ : List[str] = hidden_states + temb
lowerCamelCase__ : Optional[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[str] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = self.dropout(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = self.conva(UpperCAmelCase )
if self.conv_shortcut is not None:
lowerCamelCase__ : Dict = self.conv_shortcut(UpperCAmelCase )
return hidden_states + residual
| 50
| 1
|
import os
from typing import List, Optional, Union
from ...image_processing_utils import BatchFeature
from ...image_utils import ImageInput
from ...processing_utils import ProcessorMixin
from ...tokenization_utils_base import PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
from ...utils import TensorType
from ..auto import AutoTokenizer
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = ["""image_processor""", """tokenizer"""]
UpperCAmelCase__ = """BlipImageProcessor"""
UpperCAmelCase__ = """AutoTokenizer"""
def __init__( self : int , UpperCAmelCase : Tuple , UpperCAmelCase : str , UpperCAmelCase : Optional[Any] ) -> List[Any]:
super().__init__(UpperCAmelCase , UpperCAmelCase )
# add QFormer tokenizer
lowerCamelCase__ : int = qformer_tokenizer
def __call__( self : str , UpperCAmelCase : ImageInput = None , UpperCAmelCase : Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None , UpperCAmelCase : bool = True , UpperCAmelCase : Union[bool, str, PaddingStrategy] = False , UpperCAmelCase : Union[bool, str, TruncationStrategy] = None , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : int = 0 , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : Optional[bool] = None , UpperCAmelCase : bool = False , UpperCAmelCase : bool = False , UpperCAmelCase : bool = False , UpperCAmelCase : bool = False , UpperCAmelCase : bool = False , UpperCAmelCase : bool = True , UpperCAmelCase : Optional[Union[str, TensorType]] = None , **UpperCAmelCase : Tuple , ) -> BatchFeature:
if images is None and text is None:
raise ValueError('You have to specify at least images or text.' )
lowerCamelCase__ : Tuple = BatchFeature()
if text is not None:
lowerCamelCase__ : str = self.tokenizer(
text=UpperCAmelCase , add_special_tokens=UpperCAmelCase , padding=UpperCAmelCase , truncation=UpperCAmelCase , max_length=UpperCAmelCase , stride=UpperCAmelCase , pad_to_multiple_of=UpperCAmelCase , return_attention_mask=UpperCAmelCase , return_overflowing_tokens=UpperCAmelCase , return_special_tokens_mask=UpperCAmelCase , return_offsets_mapping=UpperCAmelCase , return_token_type_ids=UpperCAmelCase , return_length=UpperCAmelCase , verbose=UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase , )
encoding.update(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.qformer_tokenizer(
text=UpperCAmelCase , add_special_tokens=UpperCAmelCase , padding=UpperCAmelCase , truncation=UpperCAmelCase , max_length=UpperCAmelCase , stride=UpperCAmelCase , pad_to_multiple_of=UpperCAmelCase , return_attention_mask=UpperCAmelCase , return_overflowing_tokens=UpperCAmelCase , return_special_tokens_mask=UpperCAmelCase , return_offsets_mapping=UpperCAmelCase , return_token_type_ids=UpperCAmelCase , return_length=UpperCAmelCase , verbose=UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : Any = qformer_text_encoding.pop('input_ids' )
lowerCamelCase__ : List[Any] = qformer_text_encoding.pop('attention_mask' )
if images is not None:
lowerCamelCase__ : Optional[Any] = self.image_processor(UpperCAmelCase , return_tensors=UpperCAmelCase )
encoding.update(UpperCAmelCase )
return encoding
def A_ ( self : Dict , *UpperCAmelCase : str , **UpperCAmelCase : Any ) -> int:
return self.tokenizer.batch_decode(*UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Union[str, Any] , *UpperCAmelCase : List[Any] , **UpperCAmelCase : List[Any] ) -> int:
return self.tokenizer.decode(*UpperCAmelCase , **UpperCAmelCase )
@property
# Copied from transformers.models.blip.processing_blip.BlipProcessor.model_input_names
def A_ ( self : Dict ) -> str:
lowerCamelCase__ : Union[str, Any] = self.tokenizer.model_input_names
lowerCamelCase__ : Tuple = self.image_processor.model_input_names
return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names ) )
def A_ ( self : Any , UpperCAmelCase : List[str] , **UpperCAmelCase : Any ) -> Optional[int]:
if os.path.isfile(UpperCAmelCase ):
raise ValueError(F"""Provided path ({save_directory}) should be a directory, not a file""" )
os.makedirs(UpperCAmelCase , exist_ok=UpperCAmelCase )
lowerCamelCase__ : Any = os.path.join(UpperCAmelCase , 'qformer_tokenizer' )
self.qformer_tokenizer.save_pretrained(UpperCAmelCase )
return super().save_pretrained(UpperCAmelCase , **UpperCAmelCase )
@classmethod
def A_ ( cls : List[Any] , UpperCAmelCase : List[Any] , **UpperCAmelCase : Dict ) -> Dict:
lowerCamelCase__ : Union[str, Any] = AutoTokenizer.from_pretrained(UpperCAmelCase , subfolder='qformer_tokenizer' )
lowerCamelCase__ : Dict = cls._get_arguments_from_pretrained(UpperCAmelCase , **UpperCAmelCase )
args.append(UpperCAmelCase )
return cls(*UpperCAmelCase )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Optional[Any] = set()
# edges = list of graph's edges
lowerCamelCase__ : List[str] = get_edges(_UpperCAmelCase )
# While there are still elements in edges list, take an arbitrary edge
# (from_node, to_node) and add his extremity to chosen_vertices and then
# remove all arcs adjacent to the from_node and to_node
while edges:
lowerCamelCase__ , lowerCamelCase__ : str = edges.pop()
chosen_vertices.add(_UpperCAmelCase )
chosen_vertices.add(_UpperCAmelCase )
for edge in edges.copy():
if from_node in edge or to_node in edge:
edges.discard(_UpperCAmelCase )
return chosen_vertices
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Union[str, Any] = set()
for from_node, to_nodes in graph.items():
for to_node in to_nodes:
edges.add((from_node, to_node) )
return edges
if __name__ == "__main__":
import doctest
doctest.testmod()
# graph = {0: [1, 3], 1: [0, 3], 2: [0, 3, 4], 3: [0, 1, 2], 4: [2, 3]}
# print(f"Matching vertex cover:\n{matching_min_vertex_cover(graph)}")
| 50
| 1
|
from __future__ import annotations
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase = None , _UpperCAmelCase = None ) -> None:
if start is None:
lowerCamelCase__ : Optional[int] = 0
if end is None:
lowerCamelCase__ : Tuple = len(_UpperCAmelCase ) - 1
if start >= end:
return
lowerCamelCase__ : List[str] = (start + end) // 2
slowsort(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
slowsort(_UpperCAmelCase , mid + 1 , _UpperCAmelCase )
if sequence[end] < sequence[mid]:
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = sequence[mid], sequence[end]
slowsort(_UpperCAmelCase , _UpperCAmelCase , end - 1 )
if __name__ == "__main__":
from doctest import testmod
testmod()
| 50
|
from __future__ import annotations
import math
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
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(_UpperCAmelCase ) + 1 ) , 6 ):
if number % i == 0 or number % (i + 2) == 0:
return False
return True
_UpperCAmelCase : Any = [num for num in range(3, 10_00_01, 2) if not is_prime(num)]
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list[int]:
if not isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise ValueError('n must be an integer' )
if n <= 0:
raise ValueError('n must be >= 0' )
lowerCamelCase__ : int = []
for num in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : Union[str, Any] = 0
while 2 * i * i <= odd_composites[num]:
lowerCamelCase__ : Dict = odd_composites[num] - 2 * i * i
if is_prime(_UpperCAmelCase ):
break
i += 1
else:
list_nums.append(odd_composites[num] )
if len(_UpperCAmelCase ) == n:
return list_nums
return []
def SCREAMING_SNAKE_CASE ( ) -> int:
return compute_nums(1 )[0]
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
import warnings
from ...processing_utils import ProcessorMixin
from ...tokenization_utils_base import BatchEncoding
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = ["""image_processor""", """tokenizer"""]
UpperCAmelCase__ = """CLIPImageProcessor"""
UpperCAmelCase__ = ("""XLMRobertaTokenizer""", """XLMRobertaTokenizerFast""")
def __init__( self : str , UpperCAmelCase : Optional[int]=None , UpperCAmelCase : List[str]=None , **UpperCAmelCase : Any ) -> List[str]:
lowerCamelCase__ : Optional[Any] = None
if "feature_extractor" in kwargs:
warnings.warn(
'The `feature_extractor` argument is deprecated and will be removed in v5, use `image_processor`'
' instead.' , UpperCAmelCase , )
lowerCamelCase__ : str = kwargs.pop('feature_extractor' )
lowerCamelCase__ : Any = image_processor if image_processor is not None else feature_extractor
if image_processor is None:
raise ValueError('You need to specify an `image_processor`.' )
if tokenizer is None:
raise ValueError('You need to specify a `tokenizer`.' )
super().__init__(UpperCAmelCase , UpperCAmelCase )
def __call__( self : List[str] , UpperCAmelCase : Any=None , UpperCAmelCase : Dict=None , UpperCAmelCase : List[Any]=None , **UpperCAmelCase : int ) -> Optional[int]:
if text is None and images is None:
raise ValueError('You have to specify either text or images. Both cannot be none.' )
if text is not None:
lowerCamelCase__ : List[str] = self.tokenizer(UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
if images is not None:
lowerCamelCase__ : List[str] = self.image_processor(UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
if text is not None and images is not None:
lowerCamelCase__ : Optional[Any] = image_features.pixel_values
return encoding
elif text is not None:
return encoding
else:
return BatchEncoding(data=dict(**UpperCAmelCase ) , tensor_type=UpperCAmelCase )
def A_ ( self : int , *UpperCAmelCase : Union[str, Any] , **UpperCAmelCase : Dict ) -> Dict:
return self.tokenizer.batch_decode(*UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Dict , *UpperCAmelCase : Any , **UpperCAmelCase : Tuple ) -> Optional[int]:
return self.tokenizer.decode(*UpperCAmelCase , **UpperCAmelCase )
@property
def A_ ( self : Any ) -> List[Any]:
lowerCamelCase__ : Optional[int] = self.tokenizer.model_input_names
lowerCamelCase__ : Tuple = self.image_processor.model_input_names
return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names ) )
| 50
|
import argparse
from collections import OrderedDict
from pathlib import Path
import requests
import torch
from PIL import Image
from transformers import GLPNConfig, GLPNForDepthEstimation, GLPNImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
_UpperCAmelCase : Dict = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[str]:
lowerCamelCase__ : str = OrderedDict()
for key, value in state_dict.items():
if key.startswith('module.encoder' ):
lowerCamelCase__ : Optional[Any] = key.replace('module.encoder' , 'glpn.encoder' )
if key.startswith('module.decoder' ):
lowerCamelCase__ : List[str] = key.replace('module.decoder' , 'decoder.stages' )
if "patch_embed" in key:
# replace for example patch_embed1 by patch_embeddings.0
lowerCamelCase__ : Dict = key[key.find('patch_embed' ) + len('patch_embed' )]
lowerCamelCase__ : Tuple = key.replace(F"""patch_embed{idx}""" , F"""patch_embeddings.{int(_UpperCAmelCase )-1}""" )
if "norm" in key:
lowerCamelCase__ : str = key.replace('norm' , 'layer_norm' )
if "glpn.encoder.layer_norm" in key:
# replace for example layer_norm1 by layer_norm.0
lowerCamelCase__ : Dict = key[key.find('glpn.encoder.layer_norm' ) + len('glpn.encoder.layer_norm' )]
lowerCamelCase__ : str = key.replace(F"""layer_norm{idx}""" , F"""layer_norm.{int(_UpperCAmelCase )-1}""" )
if "layer_norm1" in key:
lowerCamelCase__ : Optional[int] = key.replace('layer_norm1' , 'layer_norm_1' )
if "layer_norm2" in key:
lowerCamelCase__ : Optional[int] = key.replace('layer_norm2' , 'layer_norm_2' )
if "block" in key:
# replace for example block1 by block.0
lowerCamelCase__ : List[Any] = key[key.find('block' ) + len('block' )]
lowerCamelCase__ : int = key.replace(F"""block{idx}""" , F"""block.{int(_UpperCAmelCase )-1}""" )
if "attn.q" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('attn.q' , 'attention.self.query' )
if "attn.proj" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('attn.proj' , 'attention.output.dense' )
if "attn" in key:
lowerCamelCase__ : Dict = key.replace('attn' , 'attention.self' )
if "fc1" in key:
lowerCamelCase__ : Dict = key.replace('fc1' , 'dense1' )
if "fc2" in key:
lowerCamelCase__ : Any = key.replace('fc2' , 'dense2' )
if "linear_pred" in key:
lowerCamelCase__ : Dict = key.replace('linear_pred' , 'classifier' )
if "linear_fuse" in key:
lowerCamelCase__ : Tuple = key.replace('linear_fuse.conv' , 'linear_fuse' )
lowerCamelCase__ : List[str] = key.replace('linear_fuse.bn' , 'batch_norm' )
if "linear_c" in key:
# replace for example linear_c4 by linear_c.3
lowerCamelCase__ : Optional[Any] = key[key.find('linear_c' ) + len('linear_c' )]
lowerCamelCase__ : Dict = key.replace(F"""linear_c{idx}""" , F"""linear_c.{int(_UpperCAmelCase )-1}""" )
if "bot_conv" in key:
lowerCamelCase__ : str = key.replace('bot_conv' , '0.convolution' )
if "skip_conv1" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('skip_conv1' , '1.convolution' )
if "skip_conv2" in key:
lowerCamelCase__ : List[Any] = key.replace('skip_conv2' , '2.convolution' )
if "fusion1" in key:
lowerCamelCase__ : Optional[int] = key.replace('fusion1' , '1.fusion' )
if "fusion2" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('fusion2' , '2.fusion' )
if "fusion3" in key:
lowerCamelCase__ : List[Any] = key.replace('fusion3' , '3.fusion' )
if "fusion" in key and "conv" in key:
lowerCamelCase__ : str = key.replace('conv' , 'convolutional_layer' )
if key.startswith('module.last_layer_depth' ):
lowerCamelCase__ : Dict = key.replace('module.last_layer_depth' , 'head.head' )
lowerCamelCase__ : str = value
return new_state_dict
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
# for each of the encoder blocks:
for i in range(config.num_encoder_blocks ):
for j in range(config.depths[i] ):
# read in weights + bias of keys and values (which is a single matrix in the original implementation)
lowerCamelCase__ : Any = state_dict.pop(F"""glpn.encoder.block.{i}.{j}.attention.self.kv.weight""" )
lowerCamelCase__ : Optional[Any] = state_dict.pop(F"""glpn.encoder.block.{i}.{j}.attention.self.kv.bias""" )
# next, add keys and values (in that order) to the state dict
lowerCamelCase__ : Optional[int] = kv_weight[
: config.hidden_sizes[i], :
]
lowerCamelCase__ : Optional[int] = kv_bias[: config.hidden_sizes[i]]
lowerCamelCase__ : Any = kv_weight[
config.hidden_sizes[i] :, :
]
lowerCamelCase__ : Dict = kv_bias[config.hidden_sizes[i] :]
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : List[str] = 'http://images.cocodataset.org/val2017/000000039769.jpg'
lowerCamelCase__ : Tuple = Image.open(requests.get(_UpperCAmelCase , stream=_UpperCAmelCase ).raw )
return image
@torch.no_grad()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=False , _UpperCAmelCase=None ) -> Optional[int]:
lowerCamelCase__ : str = GLPNConfig(hidden_sizes=[64, 128, 320, 512] , decoder_hidden_size=64 , depths=[3, 8, 27, 3] )
# load image processor (only resize + rescale)
lowerCamelCase__ : Union[str, Any] = GLPNImageProcessor()
# prepare image
lowerCamelCase__ : str = prepare_img()
lowerCamelCase__ : Tuple = image_processor(images=_UpperCAmelCase , return_tensors='pt' ).pixel_values
logger.info('Converting model...' )
# load original state dict
lowerCamelCase__ : Any = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
# rename keys
lowerCamelCase__ : str = rename_keys(_UpperCAmelCase )
# key and value matrices need special treatment
read_in_k_v(_UpperCAmelCase , _UpperCAmelCase )
# create HuggingFace model and load state dict
lowerCamelCase__ : Dict = GLPNForDepthEstimation(_UpperCAmelCase )
model.load_state_dict(_UpperCAmelCase )
model.eval()
# forward pass
lowerCamelCase__ : List[str] = model(_UpperCAmelCase )
lowerCamelCase__ : Tuple = outputs.predicted_depth
# verify output
if model_name is not None:
if "nyu" in model_name:
lowerCamelCase__ : List[Any] = torch.tensor(
[[4.4_147, 4.0_873, 4.0_673], [3.7_890, 3.2_881, 3.1_525], [3.7_674, 3.5_423, 3.4_913]] )
elif "kitti" in model_name:
lowerCamelCase__ : List[str] = torch.tensor(
[[3.4_291, 2.7_865, 2.5_151], [3.2_841, 2.7_021, 2.3_502], [3.1_147, 2.4_625, 2.2_481]] )
else:
raise ValueError(F"""Unknown model name: {model_name}""" )
lowerCamelCase__ : Tuple = torch.Size([1, 480, 640] )
assert predicted_depth.shape == expected_shape
assert torch.allclose(predicted_depth[0, :3, :3] , _UpperCAmelCase , atol=1e-4 )
print('Looks ok!' )
# finally, push to hub if required
if push_to_hub:
logger.info('Pushing model and image processor to the hub...' )
model.push_to_hub(
repo_path_or_name=Path(_UpperCAmelCase , _UpperCAmelCase ) , organization='nielsr' , commit_message='Add model' , use_temp_dir=_UpperCAmelCase , )
image_processor.push_to_hub(
repo_path_or_name=Path(_UpperCAmelCase , _UpperCAmelCase ) , organization='nielsr' , commit_message='Add image processor' , use_temp_dir=_UpperCAmelCase , )
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
parser.add_argument(
"""--checkpoint_path""",
default=None,
type=str,
help="""Path to the original PyTorch checkpoint (.pth file).""",
)
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 to upload the model to the HuggingFace hub."""
)
parser.add_argument(
"""--model_name""",
default="""glpn-kitti""",
type=str,
help="""Name of the model in case you're pushing to the hub.""",
)
_UpperCAmelCase : int = parser.parse_args()
convert_glpn_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 50
| 1
|
import math
import flax.linen as nn
import jax.numpy as jnp
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase = 1 , _UpperCAmelCase = 1 , _UpperCAmelCase = 1.0e4 , _UpperCAmelCase = False , _UpperCAmelCase = 1.0 , ) -> jnp.ndarray:
assert timesteps.ndim == 1, "Timesteps should be a 1d-array"
assert embedding_dim % 2 == 0, F"""Embedding dimension {embedding_dim} should be even"""
lowerCamelCase__ : int = float(embedding_dim // 2 )
lowerCamelCase__ : int = math.log(max_timescale / min_timescale ) / (num_timescales - freq_shift)
lowerCamelCase__ : Optional[int] = min_timescale * jnp.exp(jnp.arange(_UpperCAmelCase , dtype=jnp.floataa ) * -log_timescale_increment )
lowerCamelCase__ : Tuple = jnp.expand_dims(_UpperCAmelCase , 1 ) * jnp.expand_dims(_UpperCAmelCase , 0 )
# scale embeddings
lowerCamelCase__ : str = scale * emb
if flip_sin_to_cos:
lowerCamelCase__ : str = jnp.concatenate([jnp.cos(_UpperCAmelCase ), jnp.sin(_UpperCAmelCase )] , axis=1 )
else:
lowerCamelCase__ : Optional[int] = jnp.concatenate([jnp.sin(_UpperCAmelCase ), jnp.cos(_UpperCAmelCase )] , axis=1 )
lowerCamelCase__ : Union[str, Any] = jnp.reshape(_UpperCAmelCase , [jnp.shape(_UpperCAmelCase )[0], embedding_dim] )
return signal
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 32
UpperCAmelCase__ = jnp.floataa
@nn.compact
def __call__( self : Union[str, Any] , UpperCAmelCase : List[Any] ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = nn.Dense(self.time_embed_dim , dtype=self.dtype , name='linear_1' )(UpperCAmelCase )
lowerCamelCase__ : int = nn.silu(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = nn.Dense(self.time_embed_dim , dtype=self.dtype , name='linear_2' )(UpperCAmelCase )
return temb
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 32
UpperCAmelCase__ = False
UpperCAmelCase__ = 1
@nn.compact
def __call__( self : Any , UpperCAmelCase : Tuple ) -> Union[str, Any]:
return get_sinusoidal_embeddings(
UpperCAmelCase , embedding_dim=self.dim , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.freq_shift )
| 50
|
import tempfile
import unittest
from transformers import TaConfig, is_torch_available
from transformers.testing_utils import (
require_sentencepiece,
require_tokenizers,
require_torch,
slow,
torch_device,
)
from ...generation.test_utils import GenerationTesterMixin
from ...test_modeling_common import ModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import AutoTokenizer, UMTaForConditionalGeneration, UMTaForQuestionAnswering, UMTaModel
class lowerCAmelCase :
def __init__( self : Tuple , UpperCAmelCase : Dict , UpperCAmelCase : Union[str, Any]=99 , UpperCAmelCase : str=13 , UpperCAmelCase : List[str]=7 , UpperCAmelCase : str=9 , UpperCAmelCase : Optional[int]=True , UpperCAmelCase : str=True , UpperCAmelCase : Any=False , UpperCAmelCase : Union[str, Any]=32 , UpperCAmelCase : List[str]=5 , UpperCAmelCase : Tuple=4 , UpperCAmelCase : Union[str, Any]=37 , UpperCAmelCase : int=8 , UpperCAmelCase : List[str]=0.1 , UpperCAmelCase : Any=0.0_0_2 , UpperCAmelCase : Optional[Any]=1 , UpperCAmelCase : List[Any]=0 , UpperCAmelCase : Union[str, Any]=0 , UpperCAmelCase : Tuple=None , UpperCAmelCase : Optional[Any]=None , ) -> Union[str, Any]:
lowerCamelCase__ : int = parent
lowerCamelCase__ : Any = batch_size
lowerCamelCase__ : Optional[int] = encoder_seq_length
lowerCamelCase__ : int = decoder_seq_length
# For common tests
lowerCamelCase__ : List[str] = self.decoder_seq_length
lowerCamelCase__ : Optional[int] = is_training
lowerCamelCase__ : List[Any] = use_attention_mask
lowerCamelCase__ : Optional[Any] = use_labels
lowerCamelCase__ : Union[str, Any] = vocab_size
lowerCamelCase__ : Union[str, Any] = hidden_size
lowerCamelCase__ : Optional[Any] = num_hidden_layers
lowerCamelCase__ : Any = num_attention_heads
lowerCamelCase__ : str = d_ff
lowerCamelCase__ : Optional[Any] = relative_attention_num_buckets
lowerCamelCase__ : Any = dropout_rate
lowerCamelCase__ : Any = initializer_factor
lowerCamelCase__ : Union[str, Any] = eos_token_id
lowerCamelCase__ : List[str] = pad_token_id
lowerCamelCase__ : List[str] = decoder_start_token_id
lowerCamelCase__ : List[Any] = None
lowerCamelCase__ : Optional[Any] = decoder_layers
def A_ ( self : List[Any] ) -> int:
return TaConfig.from_pretrained('google/umt5-base' )
def A_ ( self : List[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[int] , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : Tuple=None , UpperCAmelCase : List[str]=None , UpperCAmelCase : str=None , UpperCAmelCase : Optional[Any]=None , UpperCAmelCase : Optional[Any]=None , ) -> List[str]:
if attention_mask is None:
lowerCamelCase__ : Optional[Any] = input_ids.ne(config.pad_token_id )
if decoder_attention_mask is None:
lowerCamelCase__ : Optional[Any] = decoder_input_ids.ne(config.pad_token_id )
if head_mask is None:
lowerCamelCase__ : int = torch.ones(config.num_hidden_layers , config.num_attention_heads , device=UpperCAmelCase )
if decoder_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
if cross_attn_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(
config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
return {
"input_ids": input_ids,
"decoder_input_ids": decoder_input_ids,
"attention_mask": attention_mask,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
"cross_attn_head_mask": cross_attn_head_mask,
}
def A_ ( self : str ) -> List[str]:
lowerCamelCase__ : Any = ids_tensor([self.batch_size, self.encoder_seq_length] , self.vocab_size )
lowerCamelCase__ : Optional[Any] = ids_tensor([self.batch_size, self.decoder_seq_length] , self.vocab_size )
# we need to clamp the input ids here to avoid having pad token in between
# this is because for NllbMoe the position_ids are prepared such that
# all pad tokens have pos id = 2 and rest are between 2..seq_length
# and the seq_length here is seq_length - num_pad_tokens
# but when using past, there is no way of knowing if the past input ids had
# pad tokens in them, which results in incorrect seq_lenth and which in turn results in
# position_ids being off by num_pad_tokens in past input
lowerCamelCase__ : List[str] = input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Union[str, Any] = decoder_input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Dict = self.get_config()
lowerCamelCase__ : Tuple = config.num_attention_heads
lowerCamelCase__ : Any = self.prepare_inputs_dict(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
return config, input_dict
def A_ ( self : Tuple ) -> Union[str, Any]:
lowerCamelCase__ , lowerCamelCase__ : Dict = self.prepare_config_and_inputs()
return config, inputs_dict
def A_ ( self : Optional[int] ) -> List[str]:
return TaConfig(
vocab_size=166 , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Union[str, Any] ) -> Dict:
return TaConfig(
vocab_size=self.vocab_size , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Any , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[str] , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Any , UpperCAmelCase : Dict , ) -> str:
lowerCamelCase__ : Dict = UMTaModel(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Optional[int] = model(
input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase , decoder_attention_mask=UpperCAmelCase , )
lowerCamelCase__ : Any = model(input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase )
lowerCamelCase__ : Dict = result.last_hidden_state
lowerCamelCase__ : Any = result.past_key_values
lowerCamelCase__ : List[Any] = result.encoder_last_hidden_state
self.parent.assertEqual(encoder_output.size() , (self.batch_size, self.encoder_seq_length, self.hidden_size) )
self.parent.assertEqual(decoder_output.size() , (self.batch_size, self.decoder_seq_length, self.hidden_size) )
# There should be `num_layers` key value embeddings stored in decoder_past
self.parent.assertEqual(len(UpperCAmelCase ) , config.num_layers )
# There should be a self attn key, a self attn value, a cross attn key and a cross attn value stored in each decoder_past tuple
self.parent.assertEqual(len(decoder_past[0] ) , 4 )
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[Any] , UpperCAmelCase : int , UpperCAmelCase : List[str] , UpperCAmelCase : Optional[Any] , ) -> Optional[int]:
lowerCamelCase__ : List[Any] = UMTaModel(config=UpperCAmelCase ).get_decoder().to(UpperCAmelCase ).eval()
# first forward pass
lowerCamelCase__ : Tuple = model(UpperCAmelCase , use_cache=UpperCAmelCase )
lowerCamelCase__ : List[Any] = model(UpperCAmelCase )
lowerCamelCase__ : int = model(UpperCAmelCase , use_cache=UpperCAmelCase )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) + 1 )
lowerCamelCase__ , lowerCamelCase__ : Dict = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
lowerCamelCase__ : Optional[int] = ids_tensor((self.batch_size, 1) , config.vocab_size )
# append to next input_ids and
lowerCamelCase__ : List[Any] = torch.cat([input_ids, next_tokens] , dim=-1 )
lowerCamelCase__ : List[str] = model(UpperCAmelCase )['last_hidden_state']
lowerCamelCase__ : str = model(UpperCAmelCase , past_key_values=UpperCAmelCase )['last_hidden_state']
# select random slice
lowerCamelCase__ : List[str] = ids_tensor((1,) , output_from_past.shape[-1] ).item()
lowerCamelCase__ : Tuple = output_from_no_past[:, -1, random_slice_idx].detach()
lowerCamelCase__ : List[str] = output_from_past[:, 0, random_slice_idx].detach()
# test that outputs are equal for slice
self.parent.assertTrue(torch.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-3 ) )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : List[str] , ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = UMTaModel(config=UpperCAmelCase ).to(UpperCAmelCase ).half().eval()
lowerCamelCase__ : Optional[int] = model(**UpperCAmelCase )['last_hidden_state']
self.parent.assertFalse(torch.isnan(UpperCAmelCase ).any().item() )
@require_torch
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(UMTaModel, UMTaForConditionalGeneration, UMTaForQuestionAnswering) if is_torch_available() else ()
)
UpperCAmelCase__ = (UMTaForConditionalGeneration,) if is_torch_available() else ()
UpperCAmelCase__ = (
{
"""conversational""": UMTaForConditionalGeneration,
"""feature-extraction""": UMTaModel,
"""summarization""": UMTaForConditionalGeneration,
"""text2text-generation""": UMTaForConditionalGeneration,
"""translation""": UMTaForConditionalGeneration,
"""question-answering""": UMTaForQuestionAnswering,
}
if is_torch_available()
else {}
)
UpperCAmelCase__ = True
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = True
UpperCAmelCase__ = True
# The small UMT5 model needs higher percentages for CPU/MP tests
UpperCAmelCase__ = [0.8, 0.9]
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = UMTaModelTester(self )
@unittest.skip('Test has a segmentation fault on torch 1.8.0' )
def A_ ( self : Tuple ) -> int:
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Tuple = UMTaModel(config_and_inputs[0] ).to(UpperCAmelCase )
with tempfile.TemporaryDirectory() as tmpdirname:
torch.onnx.export(
UpperCAmelCase , (config_and_inputs[1], config_and_inputs[3], config_and_inputs[2]) , F"""{tmpdirname}/t5_test.onnx""" , export_params=UpperCAmelCase , opset_version=9 , input_names=['input_ids', 'decoder_input_ids'] , )
@unittest.skipIf(torch_device == 'cpu' , 'Cant do half precision' )
def A_ ( self : Tuple ) -> Optional[Any]:
lowerCamelCase__ : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model_fpaa_forward(*UpperCAmelCase )
def A_ ( self : List[Any] ) -> str:
lowerCamelCase__ : int = ['encoder_attentions', 'decoder_attentions', 'cross_attentions']
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Any = config_and_inputs[0]
lowerCamelCase__ : Any = UMTaForConditionalGeneration(UpperCAmelCase ).eval()
model.to(UpperCAmelCase )
lowerCamelCase__ : Tuple = {
'head_mask': torch.zeros(config.num_layers , config.num_heads , device=UpperCAmelCase ),
'decoder_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
'cross_attn_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
}
for attn_name, (name, mask) in zip(UpperCAmelCase , head_masking.items() ):
lowerCamelCase__ : Union[str, Any] = {name: mask}
# Explicitly pass decoder_head_mask as it is required from T5 model when head_mask specified
if name == "head_mask":
lowerCamelCase__ : Union[str, Any] = torch.ones(
config.num_decoder_layers , config.num_heads , device=UpperCAmelCase )
lowerCamelCase__ : Tuple = model.generate(
config_and_inputs[1]['input_ids'] , num_beams=1 , max_length=3 , output_attentions=UpperCAmelCase , return_dict_in_generate=UpperCAmelCase , **UpperCAmelCase , )
# We check the state of decoder_attentions and cross_attentions just from the last step
lowerCamelCase__ : Union[str, Any] = out[attn_name] if attn_name == attention_names[0] else out[attn_name][-1]
self.assertEqual(sum([w.sum().item() for w in attn_weights] ) , 0.0 )
@unittest.skip('Does not work on the tiny model as we keep hitting edge cases.' )
def A_ ( self : Optional[Any] ) -> Optional[Any]:
pass
@require_torch
@require_sentencepiece
@require_tokenizers
class lowerCAmelCase ( unittest.TestCase ):
@slow
@unittest.skip(
'Unless we stop stripping left and right by default for all special tokens, the expected ids obtained here will not match the original ones. Wait for https://github.com/huggingface/transformers/pull/23909 to be merged' )
def A_ ( self : Any ) -> int:
lowerCamelCase__ : Optional[Any] = UMTaForConditionalGeneration.from_pretrained('google/umt5-small' , return_dict=UpperCAmelCase ).to(UpperCAmelCase )
lowerCamelCase__ : List[str] = AutoTokenizer.from_pretrained('google/umt5-small' , use_fast=UpperCAmelCase , legacy=UpperCAmelCase )
lowerCamelCase__ : Dict = [
'Bonjour monsieur <extra_id_0> bien <extra_id_1>.',
'No se como puedo <extra_id_0>.',
'This is the reason why we <extra_id_0> them.',
'The <extra_id_0> walks in <extra_id_1>, seats',
'A <extra_id_0> walks into a bar and orders a <extra_id_1> with <extra_id_2> pinch of <extra_id_3>.',
]
lowerCamelCase__ : Tuple = tokenizer(UpperCAmelCase , return_tensors='pt' , padding=UpperCAmelCase ).input_ids
# fmt: off
lowerCamelCase__ : Any = torch.tensor(
[
[ 38530, 210703, 256299, 1410, 256298, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 826, 321, 671, 25922, 256299, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 1460, 339, 312, 19014, 10620, 758, 256299, 2355,274, 1, 0, 0, 0, 0, 0, 0,0, 0],
[ 517, 256299, 14869, 281, 301, 256298, 275, 119983,1, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 320, 256299, 14869, 281, 2234, 289, 2275, 333,61391, 289, 256298, 543, 256297, 168714, 329, 256296,274, 1],
] )
# fmt: on
torch.testing.assert_allclose(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Optional[int] = model.generate(input_ids.to(UpperCAmelCase ) )
lowerCamelCase__ : List[Any] = [
'<pad><extra_id_0> et<extra_id_1> [eod] <extra_id_2><extra_id_55>.. [eod] 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 <extra_id_56>ajšietosto<extra_id_56>lleux<extra_id_19><extra_id_6>ajšie</s>',
'<pad><extra_id_0>.<extra_id_1>.,<0x0A>...spech <0x0A><extra_id_20> <extra_id_21></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> are not going to be a part of the world. We are not going to be a part of<extra_id_1> and<extra_id_2><0x0A><extra_id_48>.<extra_id_48></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> door<extra_id_1>, the door<extra_id_2> 피해[/</s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0>nyone who<extra_id_1> drink<extra_id_2> a<extra_id_3> alcohol<extra_id_4> A<extra_id_5> A. This<extra_id_6> I<extra_id_7><extra_id_52><extra_id_53></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
]
lowerCamelCase__ : Union[str, Any] = tokenizer.batch_decode(UpperCAmelCase )
self.assertEqual(UpperCAmelCase , UpperCAmelCase )
| 50
| 1
|
import os
from typing import Dict, List, Tuple, TypeVar, Union
_UpperCAmelCase : Dict = TypeVar("""T""")
_UpperCAmelCase : Dict = Union[List[T], Tuple[T, ...]]
_UpperCAmelCase : List[Any] = Union[T, List[T], Dict[str, T]]
_UpperCAmelCase : Any = Union[str, bytes, os.PathLike]
| 50
|
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=None , _UpperCAmelCase="no" , _UpperCAmelCase="29500" ) -> Tuple:
lowerCamelCase__ : Dict = False
lowerCamelCase__ : Dict = False
if any(key.startswith('KAGGLE' ) for key in os.environ.keys() ):
lowerCamelCase__ : Optional[Any] = True
elif "IPython" in sys.modules:
lowerCamelCase__ : Optional[Any] = 'google.colab' in str(sys.modules['IPython'].get_ipython() )
try:
lowerCamelCase__ : List[str] = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
F"""Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}.""" )
if (in_colab or in_kaggle) and (os.environ.get('TPU_NAME' , _UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '
'your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if num_processes is None:
lowerCamelCase__ : Optional[Any] = 8
lowerCamelCase__ : List[str] = PrepareForLaunch(_UpperCAmelCase , distributed_type='TPU' )
print(F"""Launching a training on {num_processes} TPU cores.""" )
xmp.spawn(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on one CPU.' )
function(*_UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '
'inside your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if torch.cuda.is_initialized():
raise ValueError(
'To launch a multi-GPU training from your notebook, you need to avoid running any instruction '
'using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '
'function.' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port=_UpperCAmelCase , mixed_precision=_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = PrepareForLaunch(_UpperCAmelCase , distributed_type='MULTI_GPU' )
print(F"""Launching training on {num_processes} GPUs.""" )
try:
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '
'This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '
'Please review your imports and test them when running the `notebook_launcher()` to identify '
'which one is problematic.' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
lowerCamelCase__ : int = '1'
print('Launching training on MPS.' )
elif torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on CPU.' )
function(*_UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=2 ) -> Optional[Any]:
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port='29500' , accelerate_mixed_precision='no' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='yes' , ):
lowerCamelCase__ : Optional[Any] = PrepareForLaunch(_UpperCAmelCase , debug=_UpperCAmelCase )
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
| 50
| 1
|
from __future__ import annotations
import inspect
import unittest
import numpy as np
from transformers import DeiTConfig
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 (
TFDeiTForImageClassification,
TFDeiTForImageClassificationWithTeacher,
TFDeiTForMaskedImageModeling,
TFDeiTModel,
)
from transformers.models.deit.modeling_tf_deit import TF_DEIT_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import DeiTImageProcessor
class lowerCAmelCase :
def __init__( self : List[str] , UpperCAmelCase : Dict , UpperCAmelCase : str=13 , UpperCAmelCase : Tuple=30 , UpperCAmelCase : List[Any]=2 , UpperCAmelCase : Union[str, Any]=3 , UpperCAmelCase : int=True , UpperCAmelCase : str=True , UpperCAmelCase : Tuple=32 , UpperCAmelCase : Optional[int]=2 , UpperCAmelCase : Optional[Any]=4 , UpperCAmelCase : Tuple=37 , UpperCAmelCase : Optional[Any]="gelu" , UpperCAmelCase : Dict=0.1 , UpperCAmelCase : Tuple=0.1 , UpperCAmelCase : Dict=10 , UpperCAmelCase : List[Any]=0.0_2 , UpperCAmelCase : int=3 , UpperCAmelCase : List[str]=None , UpperCAmelCase : int=2 , ) -> Dict:
lowerCamelCase__ : int = parent
lowerCamelCase__ : Optional[Any] = batch_size
lowerCamelCase__ : Dict = image_size
lowerCamelCase__ : Tuple = patch_size
lowerCamelCase__ : Tuple = num_channels
lowerCamelCase__ : Any = is_training
lowerCamelCase__ : Any = use_labels
lowerCamelCase__ : Optional[Any] = hidden_size
lowerCamelCase__ : Tuple = num_hidden_layers
lowerCamelCase__ : Optional[int] = num_attention_heads
lowerCamelCase__ : Optional[int] = intermediate_size
lowerCamelCase__ : Optional[Any] = hidden_act
lowerCamelCase__ : List[str] = hidden_dropout_prob
lowerCamelCase__ : Optional[Any] = attention_probs_dropout_prob
lowerCamelCase__ : str = type_sequence_label_size
lowerCamelCase__ : Union[str, Any] = initializer_range
lowerCamelCase__ : List[Any] = scope
lowerCamelCase__ : Optional[int] = encoder_stride
# in DeiT, the seq length equals the number of patches + 2 (we add 2 for the [CLS] and distilation tokens)
lowerCamelCase__ : Tuple = (image_size // patch_size) ** 2
lowerCamelCase__ : Tuple = num_patches + 2
def A_ ( self : List[str] ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
lowerCamelCase__ : Dict = None
if self.use_labels:
lowerCamelCase__ : Optional[Any] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
lowerCamelCase__ : str = self.get_config()
return config, pixel_values, labels
def A_ ( self : Optional[Any] ) -> List[Any]:
return DeiTConfig(
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=UpperCAmelCase , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , )
def A_ ( self : str , UpperCAmelCase : str , UpperCAmelCase : Optional[int] , UpperCAmelCase : Dict ) -> Union[str, Any]:
lowerCamelCase__ : List[Any] = TFDeiTModel(config=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = model(UpperCAmelCase )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
def A_ ( self : List[Any] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Dict , UpperCAmelCase : List[Any] ) -> Tuple:
lowerCamelCase__ : int = TFDeiTForMaskedImageModeling(config=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = model(UpperCAmelCase )
self.parent.assertEqual(
result.reconstruction.shape , (self.batch_size, self.num_channels, self.image_size, self.image_size) )
# test greyscale images
lowerCamelCase__ : Union[str, Any] = 1
lowerCamelCase__ : int = TFDeiTForMaskedImageModeling(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
lowerCamelCase__ : Optional[Any] = model(UpperCAmelCase )
self.parent.assertEqual(result.reconstruction.shape , (self.batch_size, 1, self.image_size, self.image_size) )
def A_ ( self : Tuple , UpperCAmelCase : Any , UpperCAmelCase : str , UpperCAmelCase : List[str] ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = self.type_sequence_label_size
lowerCamelCase__ : str = TFDeiTForImageClassification(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = model(UpperCAmelCase , labels=UpperCAmelCase )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# test greyscale images
lowerCamelCase__ : Optional[int] = 1
lowerCamelCase__ : Union[str, Any] = TFDeiTForImageClassification(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
lowerCamelCase__ : Dict = model(UpperCAmelCase , labels=UpperCAmelCase )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def A_ ( self : List[Any] ) -> Union[str, Any]:
lowerCamelCase__ : List[str] = self.prepare_config_and_inputs()
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : int = config_and_inputs
lowerCamelCase__ : Any = {'pixel_values': pixel_values}
return config, inputs_dict
@require_tf
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(
TFDeiTModel,
TFDeiTForImageClassification,
TFDeiTForImageClassificationWithTeacher,
TFDeiTForMaskedImageModeling,
)
if is_tf_available()
else ()
)
UpperCAmelCase__ = (
{
"""feature-extraction""": TFDeiTModel,
"""image-classification""": (TFDeiTForImageClassification, TFDeiTForImageClassificationWithTeacher),
}
if is_tf_available()
else {}
)
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = False
def A_ ( self : List[Any] ) -> List[str]:
lowerCamelCase__ : List[Any] = TFDeiTModelTester(self )
lowerCamelCase__ : Union[str, Any] = ConfigTester(self , config_class=UpperCAmelCase , has_text_modality=UpperCAmelCase , hidden_size=37 )
def A_ ( self : Tuple ) -> Union[str, Any]:
self.config_tester.run_common_tests()
@unittest.skip(reason='DeiT does not use inputs_embeds' )
def A_ ( self : int ) -> List[Any]:
pass
def A_ ( self : Optional[Any] ) -> Optional[Any]:
lowerCamelCase__ , lowerCamelCase__ : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCamelCase__ : Tuple = model_class(UpperCAmelCase )
self.assertIsInstance(model.get_input_embeddings() , (tf.keras.layers.Layer) )
lowerCamelCase__ : Tuple = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(UpperCAmelCase , tf.keras.layers.Dense ) )
def A_ ( self : List[str] ) -> str:
lowerCamelCase__ , lowerCamelCase__ : int = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCamelCase__ : List[Any] = model_class(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = inspect.signature(model.call )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
lowerCamelCase__ : Any = [*signature.parameters.keys()]
lowerCamelCase__ : List[Any] = ['pixel_values']
self.assertListEqual(arg_names[:1] , UpperCAmelCase )
def A_ ( self : List[Any] ) -> Dict:
lowerCamelCase__ : List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*UpperCAmelCase )
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_image_modeling(*UpperCAmelCase )
def A_ ( self : int ) -> str:
lowerCamelCase__ : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*UpperCAmelCase )
def A_ ( self : List[str] , UpperCAmelCase : Dict , UpperCAmelCase : Optional[int] , UpperCAmelCase : Any=False ) -> Union[str, Any]:
lowerCamelCase__ : List[str] = super()._prepare_for_class(UpperCAmelCase , UpperCAmelCase , return_labels=UpperCAmelCase )
if return_labels:
if "labels" in inputs_dict and "labels" not in inspect.signature(model_class.call ).parameters:
del inputs_dict["labels"]
return inputs_dict
@slow
def A_ ( self : Optional[int] ) -> Tuple:
for model_name in TF_DEIT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
lowerCamelCase__ : List[str] = TFDeiTModel.from_pretrained(UpperCAmelCase )
self.assertIsNotNone(UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( ) -> Optional[Any]:
lowerCamelCase__ : Union[str, Any] = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' )
return image
@require_tf
@require_vision
class lowerCAmelCase ( unittest.TestCase ):
@cached_property
def A_ ( self : Optional[int] ) -> List[Any]:
return (
DeiTImageProcessor.from_pretrained('facebook/deit-base-distilled-patch16-224' )
if is_vision_available()
else None
)
@slow
def A_ ( self : List[Any] ) -> List[str]:
lowerCamelCase__ : Optional[Any] = TFDeiTForImageClassificationWithTeacher.from_pretrained('facebook/deit-base-distilled-patch16-224' )
lowerCamelCase__ : int = self.default_image_processor
lowerCamelCase__ : Dict = prepare_img()
lowerCamelCase__ : Optional[Any] = image_processor(images=UpperCAmelCase , return_tensors='tf' )
# forward pass
lowerCamelCase__ : int = model(**UpperCAmelCase )
# verify the logits
lowerCamelCase__ : int = tf.TensorShape((1, 1000) )
self.assertEqual(outputs.logits.shape , UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tf.constant([-1.0_2_6_6, 0.1_9_1_2, -1.2_8_6_1] )
self.assertTrue(np.allclose(outputs.logits[0, :3] , UpperCAmelCase , atol=1e-4 ) )
| 50
|
from dataclasses import dataclass
from typing import Optional, Tuple, Union
import torch
import torch.nn as nn
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import BaseOutput
from .embeddings import GaussianFourierProjection, TimestepEmbedding, Timesteps
from .modeling_utils import ModelMixin
from .unet_ad_blocks import get_down_block, get_mid_block, get_out_block, get_up_block
@dataclass
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase ):
@register_to_config
def __init__( self : List[str] , UpperCAmelCase : int = 65536 , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 0 , UpperCAmelCase : str = "fourier" , UpperCAmelCase : bool = True , UpperCAmelCase : bool = False , UpperCAmelCase : float = 0.0 , UpperCAmelCase : Tuple[str] = ("DownBlock1DNoSkip", "DownBlock1D", "AttnDownBlock1D") , UpperCAmelCase : Tuple[str] = ("AttnUpBlock1D", "UpBlock1D", "UpBlock1DNoSkip") , UpperCAmelCase : Tuple[str] = "UNetMidBlock1D" , UpperCAmelCase : str = None , UpperCAmelCase : Tuple[int] = (32, 32, 64) , UpperCAmelCase : str = None , UpperCAmelCase : int = 8 , UpperCAmelCase : int = 1 , UpperCAmelCase : bool = False , ) -> List[Any]:
super().__init__()
lowerCamelCase__ : Optional[int] = sample_size
# time
if time_embedding_type == "fourier":
lowerCamelCase__ : Optional[Any] = GaussianFourierProjection(
embedding_size=8 , set_W_to_weight=UpperCAmelCase , log=UpperCAmelCase , flip_sin_to_cos=UpperCAmelCase )
lowerCamelCase__ : Any = 2 * block_out_channels[0]
elif time_embedding_type == "positional":
lowerCamelCase__ : List[Any] = Timesteps(
block_out_channels[0] , flip_sin_to_cos=UpperCAmelCase , downscale_freq_shift=UpperCAmelCase )
lowerCamelCase__ : Dict = block_out_channels[0]
if use_timestep_embedding:
lowerCamelCase__ : str = block_out_channels[0] * 4
lowerCamelCase__ : List[Any] = TimestepEmbedding(
in_channels=UpperCAmelCase , time_embed_dim=UpperCAmelCase , act_fn=UpperCAmelCase , out_dim=block_out_channels[0] , )
lowerCamelCase__ : Any = nn.ModuleList([] )
lowerCamelCase__ : Tuple = None
lowerCamelCase__ : List[str] = nn.ModuleList([] )
lowerCamelCase__ : Optional[int] = None
# down
lowerCamelCase__ : Optional[int] = in_channels
for i, down_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Union[str, Any] = output_channel
lowerCamelCase__ : Tuple = block_out_channels[i]
if i == 0:
input_channel += extra_in_channels
lowerCamelCase__ : Union[str, Any] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Optional[int] = get_down_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_downsample=not is_final_block or downsample_each_block , )
self.down_blocks.append(UpperCAmelCase )
# mid
lowerCamelCase__ : Optional[int] = get_mid_block(
UpperCAmelCase , in_channels=block_out_channels[-1] , mid_channels=block_out_channels[-1] , out_channels=block_out_channels[-1] , embed_dim=block_out_channels[0] , num_layers=UpperCAmelCase , add_downsample=UpperCAmelCase , )
# up
lowerCamelCase__ : Optional[int] = list(reversed(UpperCAmelCase ) )
lowerCamelCase__ : Optional[int] = reversed_block_out_channels[0]
if out_block_type is None:
lowerCamelCase__ : List[str] = out_channels
else:
lowerCamelCase__ : Any = block_out_channels[0]
for i, up_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Tuple = output_channel
lowerCamelCase__ : Union[str, Any] = (
reversed_block_out_channels[i + 1] if i < len(UpperCAmelCase ) - 1 else final_upsample_channels
)
lowerCamelCase__ : List[str] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Dict = get_up_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_upsample=not is_final_block , )
self.up_blocks.append(UpperCAmelCase )
lowerCamelCase__ : int = output_channel
# out
lowerCamelCase__ : int = norm_num_groups if norm_num_groups is not None else min(block_out_channels[0] // 4 , 32 )
lowerCamelCase__ : List[Any] = get_out_block(
out_block_type=UpperCAmelCase , num_groups_out=UpperCAmelCase , embed_dim=block_out_channels[0] , out_channels=UpperCAmelCase , act_fn=UpperCAmelCase , fc_dim=block_out_channels[-1] // 4 , )
def A_ ( self : List[Any] , UpperCAmelCase : torch.FloatTensor , UpperCAmelCase : Union[torch.Tensor, float, int] , UpperCAmelCase : bool = True , ) -> Union[UNetaDOutput, Tuple]:
lowerCamelCase__ : Optional[Any] = timestep
if not torch.is_tensor(UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = torch.tensor([timesteps] , dtype=torch.long , device=sample.device )
elif torch.is_tensor(UpperCAmelCase ) and len(timesteps.shape ) == 0:
lowerCamelCase__ : List[str] = timesteps[None].to(sample.device )
lowerCamelCase__ : Optional[int] = self.time_proj(UpperCAmelCase )
if self.config.use_timestep_embedding:
lowerCamelCase__ : str = self.time_mlp(UpperCAmelCase )
else:
lowerCamelCase__ : List[str] = timestep_embed[..., None]
lowerCamelCase__ : str = timestep_embed.repeat([1, 1, sample.shape[2]] ).to(sample.dtype )
lowerCamelCase__ : str = timestep_embed.broadcast_to((sample.shape[:1] + timestep_embed.shape[1:]) )
# 2. down
lowerCamelCase__ : str = ()
for downsample_block in self.down_blocks:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = downsample_block(hidden_states=UpperCAmelCase , temb=UpperCAmelCase )
down_block_res_samples += res_samples
# 3. mid
if self.mid_block:
lowerCamelCase__ : Optional[Any] = self.mid_block(UpperCAmelCase , UpperCAmelCase )
# 4. up
for i, upsample_block in enumerate(self.up_blocks ):
lowerCamelCase__ : Dict = down_block_res_samples[-1:]
lowerCamelCase__ : Optional[Any] = down_block_res_samples[:-1]
lowerCamelCase__ : Any = upsample_block(UpperCAmelCase , res_hidden_states_tuple=UpperCAmelCase , temb=UpperCAmelCase )
# 5. post-process
if self.out_block:
lowerCamelCase__ : Any = self.out_block(UpperCAmelCase , UpperCAmelCase )
if not return_dict:
return (sample,)
return UNetaDOutput(sample=UpperCAmelCase )
| 50
| 1
|
from string import ascii_uppercase
_UpperCAmelCase : List[str] = {str(ord(c) - 55): c for c in ascii_uppercase}
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> str:
if isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise TypeError('int() can\'t convert non-string with explicit base' )
if num < 0:
raise ValueError('parameter must be positive int' )
if isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise TypeError('\'str\' object cannot be interpreted as an integer' )
if isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise TypeError('\'float\' object cannot be interpreted as an integer' )
if base in (0, 1):
raise ValueError('base must be >= 2' )
if base > 36:
raise ValueError('base must be <= 36' )
lowerCamelCase__ : Optional[Any] = ''
lowerCamelCase__ : List[str] = 0
lowerCamelCase__ : Dict = 0
while div != 1:
lowerCamelCase__ , lowerCamelCase__ : List[str] = divmod(_UpperCAmelCase , _UpperCAmelCase )
if base >= 11 and 9 < mod < 36:
lowerCamelCase__ : Dict = ALPHABET_VALUES[str(_UpperCAmelCase )]
else:
lowerCamelCase__ : int = str(_UpperCAmelCase )
new_value += actual_value
lowerCamelCase__ : List[Any] = num // base
lowerCamelCase__ : Optional[int] = div
if div == 0:
return str(new_value[::-1] )
elif div == 1:
new_value += str(_UpperCAmelCase )
return str(new_value[::-1] )
return new_value[::-1]
if __name__ == "__main__":
import doctest
doctest.testmod()
for base in range(2, 37):
for num in range(10_00):
assert int(decimal_to_any(num, base), base) == num, (
num,
base,
decimal_to_any(num, base),
int(decimal_to_any(num, base), base),
)
| 50
|
from __future__ import annotations
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> list[tuple[int, int]]:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = position
lowerCamelCase__ : Optional[Any] = [
(y + 1, x + 2),
(y - 1, x + 2),
(y + 1, x - 2),
(y - 1, x - 2),
(y + 2, x + 1),
(y + 2, x - 1),
(y - 2, x + 1),
(y - 2, x - 1),
]
lowerCamelCase__ : Dict = []
for position in positions:
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = position
if 0 <= y_test < n and 0 <= x_test < n:
permissible_positions.append(_UpperCAmelCase )
return permissible_positions
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
return not any(elem == 0 for row in board for elem in row )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> bool:
if is_complete(_UpperCAmelCase ):
return True
for position in get_valid_pos(_UpperCAmelCase , len(_UpperCAmelCase ) ):
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = position
if board[y][x] == 0:
lowerCamelCase__ : List[Any] = curr + 1
if open_knight_tour_helper(_UpperCAmelCase , _UpperCAmelCase , curr + 1 ):
return True
lowerCamelCase__ : Optional[Any] = 0
return False
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list[list[int]]:
lowerCamelCase__ : Any = [[0 for i in range(_UpperCAmelCase )] for j in range(_UpperCAmelCase )]
for i in range(_UpperCAmelCase ):
for j in range(_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = 1
if open_knight_tour_helper(_UpperCAmelCase , (i, j) , 1 ):
return board
lowerCamelCase__ : Optional[Any] = 0
lowerCamelCase__ : Any = F"""Open Kight Tour cannot be performed on a board of size {n}"""
raise ValueError(_UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from tokenizers import processors
from ...tokenization_utils import AddedToken, BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_nllb import NllbTokenizer
else:
_UpperCAmelCase : int = None
_UpperCAmelCase : Dict = logging.get_logger(__name__)
_UpperCAmelCase : Optional[int] = {"""vocab_file""": """sentencepiece.bpe.model""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : List[Any] = {
"""vocab_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/sentencepiece.bpe.model"""
),
},
"""tokenizer_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : List[str] = {
"""facebook/nllb-large-en-ro""": 10_24,
"""facebook/nllb-200-distilled-600M""": 10_24,
}
# fmt: off
_UpperCAmelCase : Optional[int] = ["""ace_Arab""", """ace_Latn""", """acm_Arab""", """acq_Arab""", """aeb_Arab""", """afr_Latn""", """ajp_Arab""", """aka_Latn""", """amh_Ethi""", """apc_Arab""", """arb_Arab""", """ars_Arab""", """ary_Arab""", """arz_Arab""", """asm_Beng""", """ast_Latn""", """awa_Deva""", """ayr_Latn""", """azb_Arab""", """azj_Latn""", """bak_Cyrl""", """bam_Latn""", """ban_Latn""", """bel_Cyrl""", """bem_Latn""", """ben_Beng""", """bho_Deva""", """bjn_Arab""", """bjn_Latn""", """bod_Tibt""", """bos_Latn""", """bug_Latn""", """bul_Cyrl""", """cat_Latn""", """ceb_Latn""", """ces_Latn""", """cjk_Latn""", """ckb_Arab""", """crh_Latn""", """cym_Latn""", """dan_Latn""", """deu_Latn""", """dik_Latn""", """dyu_Latn""", """dzo_Tibt""", """ell_Grek""", """eng_Latn""", """epo_Latn""", """est_Latn""", """eus_Latn""", """ewe_Latn""", """fao_Latn""", """pes_Arab""", """fij_Latn""", """fin_Latn""", """fon_Latn""", """fra_Latn""", """fur_Latn""", """fuv_Latn""", """gla_Latn""", """gle_Latn""", """glg_Latn""", """grn_Latn""", """guj_Gujr""", """hat_Latn""", """hau_Latn""", """heb_Hebr""", """hin_Deva""", """hne_Deva""", """hrv_Latn""", """hun_Latn""", """hye_Armn""", """ibo_Latn""", """ilo_Latn""", """ind_Latn""", """isl_Latn""", """ita_Latn""", """jav_Latn""", """jpn_Jpan""", """kab_Latn""", """kac_Latn""", """kam_Latn""", """kan_Knda""", """kas_Arab""", """kas_Deva""", """kat_Geor""", """knc_Arab""", """knc_Latn""", """kaz_Cyrl""", """kbp_Latn""", """kea_Latn""", """khm_Khmr""", """kik_Latn""", """kin_Latn""", """kir_Cyrl""", """kmb_Latn""", """kon_Latn""", """kor_Hang""", """kmr_Latn""", """lao_Laoo""", """lvs_Latn""", """lij_Latn""", """lim_Latn""", """lin_Latn""", """lit_Latn""", """lmo_Latn""", """ltg_Latn""", """ltz_Latn""", """lua_Latn""", """lug_Latn""", """luo_Latn""", """lus_Latn""", """mag_Deva""", """mai_Deva""", """mal_Mlym""", """mar_Deva""", """min_Latn""", """mkd_Cyrl""", """plt_Latn""", """mlt_Latn""", """mni_Beng""", """khk_Cyrl""", """mos_Latn""", """mri_Latn""", """zsm_Latn""", """mya_Mymr""", """nld_Latn""", """nno_Latn""", """nob_Latn""", """npi_Deva""", """nso_Latn""", """nus_Latn""", """nya_Latn""", """oci_Latn""", """gaz_Latn""", """ory_Orya""", """pag_Latn""", """pan_Guru""", """pap_Latn""", """pol_Latn""", """por_Latn""", """prs_Arab""", """pbt_Arab""", """quy_Latn""", """ron_Latn""", """run_Latn""", """rus_Cyrl""", """sag_Latn""", """san_Deva""", """sat_Beng""", """scn_Latn""", """shn_Mymr""", """sin_Sinh""", """slk_Latn""", """slv_Latn""", """smo_Latn""", """sna_Latn""", """snd_Arab""", """som_Latn""", """sot_Latn""", """spa_Latn""", """als_Latn""", """srd_Latn""", """srp_Cyrl""", """ssw_Latn""", """sun_Latn""", """swe_Latn""", """swh_Latn""", """szl_Latn""", """tam_Taml""", """tat_Cyrl""", """tel_Telu""", """tgk_Cyrl""", """tgl_Latn""", """tha_Thai""", """tir_Ethi""", """taq_Latn""", """taq_Tfng""", """tpi_Latn""", """tsn_Latn""", """tso_Latn""", """tuk_Latn""", """tum_Latn""", """tur_Latn""", """twi_Latn""", """tzm_Tfng""", """uig_Arab""", """ukr_Cyrl""", """umb_Latn""", """urd_Arab""", """uzn_Latn""", """vec_Latn""", """vie_Latn""", """war_Latn""", """wol_Latn""", """xho_Latn""", """ydd_Hebr""", """yor_Latn""", """yue_Hant""", """zho_Hans""", """zho_Hant""", """zul_Latn"""]
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = ["""input_ids""", """attention_mask"""]
UpperCAmelCase__ = NllbTokenizer
UpperCAmelCase__ = []
UpperCAmelCase__ = []
def __init__( self : Tuple , UpperCAmelCase : int=None , UpperCAmelCase : Any=None , UpperCAmelCase : str="<s>" , UpperCAmelCase : Optional[Any]="</s>" , UpperCAmelCase : str="</s>" , UpperCAmelCase : Tuple="<s>" , UpperCAmelCase : Optional[Any]="<unk>" , UpperCAmelCase : List[str]="<pad>" , UpperCAmelCase : Union[str, Any]="<mask>" , UpperCAmelCase : Tuple=None , UpperCAmelCase : int=None , UpperCAmelCase : Dict=None , UpperCAmelCase : Any=False , **UpperCAmelCase : Optional[int] , ) -> Tuple:
# Mask token behave like a normal word, i.e. include the space before it
lowerCamelCase__ : List[Any] = AddedToken(UpperCAmelCase , lstrip=UpperCAmelCase , rstrip=UpperCAmelCase ) if isinstance(UpperCAmelCase , UpperCAmelCase ) else mask_token
lowerCamelCase__ : Union[str, Any] = legacy_behaviour
super().__init__(
vocab_file=UpperCAmelCase , tokenizer_file=UpperCAmelCase , bos_token=UpperCAmelCase , eos_token=UpperCAmelCase , sep_token=UpperCAmelCase , cls_token=UpperCAmelCase , unk_token=UpperCAmelCase , pad_token=UpperCAmelCase , mask_token=UpperCAmelCase , src_lang=UpperCAmelCase , tgt_lang=UpperCAmelCase , additional_special_tokens=UpperCAmelCase , legacy_behaviour=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = vocab_file
lowerCamelCase__ : Dict = False if not self.vocab_file else True
lowerCamelCase__ : Optional[Any] = FAIRSEQ_LANGUAGE_CODES.copy()
if additional_special_tokens is not None:
# Only add those special tokens if they are not already there.
_additional_special_tokens.extend(
[t for t in additional_special_tokens if t not in _additional_special_tokens] )
self.add_special_tokens({'additional_special_tokens': _additional_special_tokens} )
lowerCamelCase__ : str = {
lang_code: self.convert_tokens_to_ids(UpperCAmelCase ) for lang_code in FAIRSEQ_LANGUAGE_CODES
}
lowerCamelCase__ : int = src_lang if src_lang is not None else 'eng_Latn'
lowerCamelCase__ : List[Any] = self.convert_tokens_to_ids(self._src_lang )
lowerCamelCase__ : str = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def A_ ( self : int ) -> str:
return self._src_lang
@src_lang.setter
def A_ ( self : List[Any] , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Any = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
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 A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : Dict = [self.sep_token_id]
lowerCamelCase__ : List[str] = [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 + sep + token_ids_a + sep ) * [0]
def A_ ( self : int , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Optional[str] , UpperCAmelCase : Optional[str] , **UpperCAmelCase : List[str] ) -> Dict:
if src_lang is None or tgt_lang is None:
raise ValueError('Translation requires a `src_lang` and a `tgt_lang` for this model' )
lowerCamelCase__ : Optional[int] = src_lang
lowerCamelCase__ : Optional[int] = self(UpperCAmelCase , add_special_tokens=UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.convert_tokens_to_ids(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tgt_lang_id
return inputs
def A_ ( self : Dict , UpperCAmelCase : List[str] , UpperCAmelCase : str = "eng_Latn" , UpperCAmelCase : Optional[List[str]] = None , UpperCAmelCase : str = "fra_Latn" , **UpperCAmelCase : Dict , ) -> BatchEncoding:
lowerCamelCase__ : Any = src_lang
lowerCamelCase__ : int = tgt_lang
return super().prepare_seqaseq_batch(UpperCAmelCase , UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Union[str, Any] ) -> Optional[int]:
return self.set_src_lang_special_tokens(self.src_lang )
def A_ ( self : Any ) -> Union[str, Any]:
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def A_ ( self : str , UpperCAmelCase : Optional[Any] ) -> None:
lowerCamelCase__ : int = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : int = []
lowerCamelCase__ : str = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : int = [self.cur_lang_code]
lowerCamelCase__ : Tuple = [self.eos_token_id]
lowerCamelCase__ : Any = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : Optional[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : str = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : int , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Union[str, Any] = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : Dict = []
lowerCamelCase__ : Union[str, Any] = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : Any = [self.cur_lang_code]
lowerCamelCase__ : Optional[Any] = [self.eos_token_id]
lowerCamelCase__ : Union[str, Any] = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : List[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : Optional[int] = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : Union[str, Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not self.can_save_slow_tokenizer:
raise ValueError(
'Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '
'tokenizer.' )
if not os.path.isdir(UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory.""" )
return
lowerCamelCase__ : int = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCAmelCase ):
copyfile(self.vocab_file , UpperCAmelCase )
return (out_vocab_file,)
| 50
|
import argparse
import json
from collections import OrderedDict
import torch
from huggingface_hub import cached_download, hf_hub_url
from transformers import AutoImageProcessor, CvtConfig, CvtForImageClassification
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> int:
lowerCamelCase__ : Optional[int] = []
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.weight""",
F"""stage{idx}.patch_embed.proj.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.bias""",
F"""stage{idx}.patch_embed.proj.bias""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.weight""",
F"""stage{idx}.patch_embed.norm.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.bias""",
F"""stage{idx}.patch_embed.norm.bias""",
) )
return embed
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Tuple = []
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj.bias""",
) )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.weight""", F"""stage{idx}.blocks.{cnt}.norm1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.bias""", F"""stage{idx}.blocks.{cnt}.norm1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.weight""", F"""stage{idx}.blocks.{cnt}.norm2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.bias""", F"""stage{idx}.blocks.{cnt}.norm2.bias""") )
return attention_weights
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = []
token.append((F"""cvt.encoder.stages.{idx}.cls_token""", 'stage2.cls_token') )
return token
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : str = []
head.append(('layernorm.weight', 'norm.weight') )
head.append(('layernorm.bias', 'norm.bias') )
head.append(('classifier.weight', 'head.weight') )
head.append(('classifier.bias', 'head.bias') )
return head
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
lowerCamelCase__ : Tuple = 'imagenet-1k-id2label.json'
lowerCamelCase__ : Union[str, Any] = 1000
lowerCamelCase__ : Optional[Any] = 'huggingface/label-files'
lowerCamelCase__ : Any = num_labels
lowerCamelCase__ : Dict = json.load(open(cached_download(hf_hub_url(_UpperCAmelCase , _UpperCAmelCase , repo_type='dataset' ) ) , 'r' ) )
lowerCamelCase__ : int = {int(_UpperCAmelCase ): v for k, v in idalabel.items()}
lowerCamelCase__ : Tuple = idalabel
lowerCamelCase__ : List[Any] = {v: k for k, v in idalabel.items()}
lowerCamelCase__ : List[str] = CvtConfig(num_labels=_UpperCAmelCase , idalabel=_UpperCAmelCase , labelaid=_UpperCAmelCase )
# For depth size 13 (13 = 1+2+10)
if cvt_model.rsplit('/' , 1 )[-1][4:6] == "13":
lowerCamelCase__ : List[Any] = [1, 2, 10]
# For depth size 21 (21 = 1+4+16)
elif cvt_model.rsplit('/' , 1 )[-1][4:6] == "21":
lowerCamelCase__ : Dict = [1, 4, 16]
# For wide cvt (similar to wide-resnet) depth size 24 (w24 = 2 + 2 20)
else:
lowerCamelCase__ : Optional[Any] = [2, 2, 20]
lowerCamelCase__ : Optional[int] = [3, 12, 16]
lowerCamelCase__ : str = [192, 768, 1024]
lowerCamelCase__ : Any = CvtForImageClassification(_UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = AutoImageProcessor.from_pretrained('facebook/convnext-base-224-22k-1k' )
lowerCamelCase__ : Tuple = image_size
lowerCamelCase__ : List[str] = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
lowerCamelCase__ : Optional[int] = OrderedDict()
lowerCamelCase__ : Tuple = []
for idx in range(len(config.depth ) ):
if config.cls_token[idx]:
lowerCamelCase__ : Optional[Any] = list_of_state_dict + cls_token(_UpperCAmelCase )
lowerCamelCase__ : str = list_of_state_dict + embeddings(_UpperCAmelCase )
for cnt in range(config.depth[idx] ):
lowerCamelCase__ : str = list_of_state_dict + attention(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : int = list_of_state_dict + final()
for gg in list_of_state_dict:
print(_UpperCAmelCase )
for i in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : str = original_weights[list_of_state_dict[i][1]]
model.load_state_dict(_UpperCAmelCase )
model.save_pretrained(_UpperCAmelCase )
image_processor.save_pretrained(_UpperCAmelCase )
# Download the weights from zoo: https://1drv.ms/u/s!AhIXJn_J-blW9RzF3rMW7SsLHa8h?e=blQ0Al
if __name__ == "__main__":
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
parser.add_argument(
"""--cvt_model""",
default="""cvt-w24""",
type=str,
help="""Name of the cvt model you'd like to convert.""",
)
parser.add_argument(
"""--image_size""",
default=3_84,
type=int,
help="""Input Image Size""",
)
parser.add_argument(
"""--cvt_file_name""",
default=R"""cvtmodels\CvT-w24-384x384-IN-22k.pth""",
type=str,
help="""Input Image Size""",
)
parser.add_argument(
"""--pytorch_dump_folder_path""", default=None, type=str, help="""Path to the output PyTorch model directory."""
)
_UpperCAmelCase : List[str] = parser.parse_args()
convert_cvt_checkpoint(args.cvt_model, args.image_size, args.cvt_file_name, args.pytorch_dump_folder_path)
| 50
| 1
|
import unittest
from transformers import BigBirdConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, ids_tensor, random_attention_mask
if is_flax_available():
import jax
from transformers.models.big_bird.modeling_flax_big_bird import (
FlaxBigBirdForCausalLM,
FlaxBigBirdForMaskedLM,
FlaxBigBirdForMultipleChoice,
FlaxBigBirdForPreTraining,
FlaxBigBirdForQuestionAnswering,
FlaxBigBirdForSequenceClassification,
FlaxBigBirdForTokenClassification,
FlaxBigBirdModel,
)
class lowerCAmelCase ( unittest.TestCase ):
def __init__( self : Optional[Any] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Any=2 , UpperCAmelCase : str=56 , UpperCAmelCase : int=True , UpperCAmelCase : int=True , UpperCAmelCase : int=True , UpperCAmelCase : List[str]=True , UpperCAmelCase : str=99 , UpperCAmelCase : Union[str, Any]=32 , UpperCAmelCase : Tuple=2 , UpperCAmelCase : Optional[int]=2 , UpperCAmelCase : List[Any]=7 , UpperCAmelCase : Tuple="gelu_new" , UpperCAmelCase : List[str]=0.1 , UpperCAmelCase : List[str]=0.1 , UpperCAmelCase : List[str]=512 , UpperCAmelCase : Any=16 , UpperCAmelCase : Optional[int]=2 , UpperCAmelCase : List[Any]=0.0_2 , UpperCAmelCase : Any=4 , UpperCAmelCase : List[Any]="block_sparse" , UpperCAmelCase : Optional[Any]=True , UpperCAmelCase : int=False , UpperCAmelCase : Any=2 , UpperCAmelCase : Any=3 , ) -> str:
lowerCamelCase__ : Dict = parent
lowerCamelCase__ : List[Any] = batch_size
lowerCamelCase__ : Optional[int] = seq_length
lowerCamelCase__ : int = is_training
lowerCamelCase__ : Tuple = use_attention_mask
lowerCamelCase__ : Optional[Any] = use_token_type_ids
lowerCamelCase__ : Union[str, Any] = use_labels
lowerCamelCase__ : Optional[Any] = vocab_size
lowerCamelCase__ : Union[str, Any] = hidden_size
lowerCamelCase__ : List[str] = num_hidden_layers
lowerCamelCase__ : List[str] = num_attention_heads
lowerCamelCase__ : str = intermediate_size
lowerCamelCase__ : str = hidden_act
lowerCamelCase__ : Any = hidden_dropout_prob
lowerCamelCase__ : List[str] = attention_probs_dropout_prob
lowerCamelCase__ : Dict = max_position_embeddings
lowerCamelCase__ : Tuple = type_vocab_size
lowerCamelCase__ : List[str] = type_sequence_label_size
lowerCamelCase__ : Any = initializer_range
lowerCamelCase__ : Dict = num_choices
lowerCamelCase__ : int = rescale_embeddings
lowerCamelCase__ : Union[str, Any] = attention_type
lowerCamelCase__ : int = use_bias
lowerCamelCase__ : List[Any] = block_size
lowerCamelCase__ : str = num_random_blocks
def A_ ( self : str ) -> Dict:
lowerCamelCase__ : Any = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
lowerCamelCase__ : List[str] = None
if self.use_attention_mask:
lowerCamelCase__ : Tuple = random_attention_mask([self.batch_size, self.seq_length] )
lowerCamelCase__ : List[str] = None
if self.use_token_type_ids:
lowerCamelCase__ : str = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
lowerCamelCase__ : Any = BigBirdConfig(
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=UpperCAmelCase , initializer_range=self.initializer_range , attention_type=self.attention_type , block_size=self.block_size , num_random_blocks=self.num_random_blocks , use_bias=self.use_bias , rescale_embeddings=self.rescale_embeddings , )
return config, input_ids, token_type_ids, attention_mask
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Tuple = self.prepare_config_and_inputs()
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : List[str] = config_and_inputs
lowerCamelCase__ : Optional[int] = {
'input_ids': input_ids,
'token_type_ids': token_type_ids,
'attention_mask': attention_mask,
}
return config, inputs_dict
@require_flax
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(
FlaxBigBirdForCausalLM,
FlaxBigBirdModel,
FlaxBigBirdForPreTraining,
FlaxBigBirdForMaskedLM,
FlaxBigBirdForMultipleChoice,
FlaxBigBirdForQuestionAnswering,
FlaxBigBirdForSequenceClassification,
FlaxBigBirdForTokenClassification,
)
if is_flax_available()
else ()
)
UpperCAmelCase__ = False
UpperCAmelCase__ = False
def A_ ( self : int ) -> Any:
lowerCamelCase__ : Optional[Any] = FlaxBigBirdModelTester(self )
@slow
# copied from `test_modeling_flax_common` because it takes much longer than other models
def A_ ( self : List[str] ) -> int:
super().test_from_pretrained_save_pretrained()
@slow
# copied from `test_modeling_flax_common` because it takes much longer than other models
def A_ ( self : Union[str, Any] ) -> Dict:
super().test_from_pretrained_with_no_automatic_init()
@slow
# copied from `test_modeling_flax_common` because it takes much longer than other models
def A_ ( self : str ) -> List[Any]:
super().test_no_automatic_init()
@slow
# copied from `test_modeling_flax_common` because it takes much longer than other models
def A_ ( self : str ) -> int:
super().test_hidden_states_output()
@slow
def A_ ( self : List[Any] ) -> Optional[Any]:
for model_class_name in self.all_model_classes:
lowerCamelCase__ : str = model_class_name.from_pretrained('google/bigbird-roberta-base' )
self.assertIsNotNone(UpperCAmelCase )
def A_ ( self : Tuple ) -> Dict:
if self.test_attn_probs:
super().test_attention_outputs()
@slow
# copied from `test_modeling_flax_common` because it takes much longer than other models
def A_ ( self : Union[str, Any] ) -> Dict:
lowerCamelCase__ , lowerCamelCase__ : Dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
with self.subTest(model_class.__name__ ):
lowerCamelCase__ : str = self._prepare_for_class(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : List[Any] = model_class(UpperCAmelCase )
@jax.jit
def model_jitted(UpperCAmelCase : Dict , UpperCAmelCase : int=None , **UpperCAmelCase : str ):
return model(input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase , **UpperCAmelCase )
with self.subTest('JIT Enabled' ):
lowerCamelCase__ : Dict = model_jitted(**UpperCAmelCase ).to_tuple()
with self.subTest('JIT Disabled' ):
with jax.disable_jit():
lowerCamelCase__ : Dict = model_jitted(**UpperCAmelCase ).to_tuple()
self.assertEqual(len(UpperCAmelCase ) , len(UpperCAmelCase ) )
for jitted_output, output in zip(UpperCAmelCase , UpperCAmelCase ):
self.assertEqual(jitted_output.shape , output.shape )
def A_ ( self : str , UpperCAmelCase : Tuple , UpperCAmelCase : List[str] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Tuple=1e-5 , UpperCAmelCase : List[Any]="outputs" , UpperCAmelCase : Any=None ) -> Optional[int]:
# `bigbird_block_sparse_attention` in `FlaxBigBird` returns `attention_probs = None`, while in PyTorch version,
# an effort was done to return `attention_probs` (yet to be verified).
if name.startswith('outputs.attentions' ):
return
else:
super().check_pt_flax_outputs(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
| 50
|
# Copyright 2021 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.
import argparse
import os
from accelerate.test_utils import execute_subprocess_async
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase=None ) -> Tuple:
if subparsers is not None:
lowerCamelCase__ : Any = subparsers.add_parser('test' )
else:
lowerCamelCase__ : int = argparse.ArgumentParser('Accelerate test command' )
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 SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = os.path.sep.join(__file__.split(os.path.sep )[:-2] + ['test_utils', 'scripts', 'test_script.py'] )
if args.config_file is None:
lowerCamelCase__ : List[str] = script_name
else:
lowerCamelCase__ : List[Any] = F"""--config_file={args.config_file} {script_name}"""
lowerCamelCase__ : str = ['accelerate-launch'] + test_args.split()
lowerCamelCase__ : Dict = execute_subprocess_async(_UpperCAmelCase , env=os.environ.copy() )
if result.returncode == 0:
print('Test is a success! You are ready for your distributed training!' )
def SCREAMING_SNAKE_CASE ( ) -> Any:
lowerCamelCase__ : Any = test_command_parser()
lowerCamelCase__ : List[Any] = parser.parse_args()
test_command(_UpperCAmelCase )
if __name__ == "__main__":
main()
| 50
| 1
|
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=None , _UpperCAmelCase="no" , _UpperCAmelCase="29500" ) -> Tuple:
lowerCamelCase__ : Dict = False
lowerCamelCase__ : Dict = False
if any(key.startswith('KAGGLE' ) for key in os.environ.keys() ):
lowerCamelCase__ : Optional[Any] = True
elif "IPython" in sys.modules:
lowerCamelCase__ : Optional[Any] = 'google.colab' in str(sys.modules['IPython'].get_ipython() )
try:
lowerCamelCase__ : List[str] = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
F"""Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}.""" )
if (in_colab or in_kaggle) and (os.environ.get('TPU_NAME' , _UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '
'your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if num_processes is None:
lowerCamelCase__ : Optional[Any] = 8
lowerCamelCase__ : List[str] = PrepareForLaunch(_UpperCAmelCase , distributed_type='TPU' )
print(F"""Launching a training on {num_processes} TPU cores.""" )
xmp.spawn(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on one CPU.' )
function(*_UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '
'inside your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if torch.cuda.is_initialized():
raise ValueError(
'To launch a multi-GPU training from your notebook, you need to avoid running any instruction '
'using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '
'function.' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port=_UpperCAmelCase , mixed_precision=_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = PrepareForLaunch(_UpperCAmelCase , distributed_type='MULTI_GPU' )
print(F"""Launching training on {num_processes} GPUs.""" )
try:
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '
'This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '
'Please review your imports and test them when running the `notebook_launcher()` to identify '
'which one is problematic.' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
lowerCamelCase__ : int = '1'
print('Launching training on MPS.' )
elif torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on CPU.' )
function(*_UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=2 ) -> Optional[Any]:
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port='29500' , accelerate_mixed_precision='no' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='yes' , ):
lowerCamelCase__ : Optional[Any] = PrepareForLaunch(_UpperCAmelCase , debug=_UpperCAmelCase )
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 100_0000 ) -> int:
lowerCamelCase__ : int = limit + 1
lowerCamelCase__ : Optional[Any] = [0] * limit
for first_term in range(1 , _UpperCAmelCase ):
for n in range(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Optional[Any] = first_term + n / first_term
if common_difference % 4: # d must be divisble by 4
continue
else:
common_difference /= 4
if (
first_term > common_difference
and first_term < 4 * common_difference
): # since x,y,z are positive integers
frequency[n] += 1 # so z>0 and a>d ,also 4d<a
lowerCamelCase__ : List[str] = sum(1 for x in frequency[1:limit] if x == 10 )
return count
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> int:
if not isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise TypeError('only integers accepted as input' )
else:
lowerCamelCase__ : List[Any] = str(abs(_UpperCAmelCase ) )
lowerCamelCase__ : Optional[Any] = [list(_UpperCAmelCase ) for char in range(len(_UpperCAmelCase ) )]
for index in range(len(_UpperCAmelCase ) ):
num_transpositions[index].pop(_UpperCAmelCase )
return max(
int(''.join(list(_UpperCAmelCase ) ) ) for transposition in num_transpositions )
if __name__ == "__main__":
__import__("""doctest""").testmod()
| 50
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_base import BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import PaddingStrategy, logging
from .tokenization_realm import RealmTokenizer
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
_UpperCAmelCase : Any = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : int = {
"""vocab_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/vocab.txt"""
),
"""google/realm-orqa-nq-openqa""": """https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-nq-reader""": """https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-openqa""": """https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-reader""": """https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/tokenizer.jsont"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-openqa""": (
"""https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-reader""": (
"""https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-openqa""": (
"""https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-reader""": (
"""https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : int = {
"""google/realm-cc-news-pretrained-embedder""": 5_12,
"""google/realm-cc-news-pretrained-encoder""": 5_12,
"""google/realm-cc-news-pretrained-scorer""": 5_12,
"""google/realm-cc-news-pretrained-openqa""": 5_12,
"""google/realm-orqa-nq-openqa""": 5_12,
"""google/realm-orqa-nq-reader""": 5_12,
"""google/realm-orqa-wq-openqa""": 5_12,
"""google/realm-orqa-wq-reader""": 5_12,
}
_UpperCAmelCase : Any = {
"""google/realm-cc-news-pretrained-embedder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-encoder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-scorer""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-reader""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-reader""": {"""do_lower_case""": True},
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_INIT_CONFIGURATION
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = RealmTokenizer
def __init__( self : Optional[int] , UpperCAmelCase : Tuple=None , UpperCAmelCase : Any=None , UpperCAmelCase : List[Any]=True , UpperCAmelCase : Optional[Any]="[UNK]" , UpperCAmelCase : Any="[SEP]" , UpperCAmelCase : Tuple="[PAD]" , UpperCAmelCase : List[Any]="[CLS]" , UpperCAmelCase : Union[str, Any]="[MASK]" , UpperCAmelCase : Optional[Any]=True , UpperCAmelCase : Any=None , **UpperCAmelCase : Optional[int] , ) -> str:
super().__init__(
UpperCAmelCase , tokenizer_file=UpperCAmelCase , do_lower_case=UpperCAmelCase , unk_token=UpperCAmelCase , sep_token=UpperCAmelCase , pad_token=UpperCAmelCase , cls_token=UpperCAmelCase , mask_token=UpperCAmelCase , tokenize_chinese_chars=UpperCAmelCase , strip_accents=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get('lowercase' , UpperCAmelCase ) != do_lower_case
or normalizer_state.get('strip_accents' , UpperCAmelCase ) != strip_accents
or normalizer_state.get('handle_chinese_chars' , UpperCAmelCase ) != tokenize_chinese_chars
):
lowerCamelCase__ : Optional[int] = getattr(UpperCAmelCase , normalizer_state.pop('type' ) )
lowerCamelCase__ : Optional[Any] = do_lower_case
lowerCamelCase__ : str = strip_accents
lowerCamelCase__ : Optional[Any] = tokenize_chinese_chars
lowerCamelCase__ : int = normalizer_class(**UpperCAmelCase )
lowerCamelCase__ : str = do_lower_case
def A_ ( self : Optional[int] , UpperCAmelCase : int , **UpperCAmelCase : int ) -> List[Any]:
lowerCamelCase__ : List[Any] = PaddingStrategy.MAX_LENGTH
lowerCamelCase__ : Optional[int] = text
lowerCamelCase__ : Dict = kwargs.pop('text_pair' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = kwargs.pop('return_tensors' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = {
'input_ids': [],
'attention_mask': [],
'token_type_ids': [],
}
for idx, candidate_text in enumerate(UpperCAmelCase ):
if batch_text_pair is not None:
lowerCamelCase__ : Tuple = batch_text_pair[idx]
else:
lowerCamelCase__ : Dict = None
lowerCamelCase__ : Optional[int] = super().__call__(UpperCAmelCase , UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Any = encoded_candidates.get('input_ids' )
lowerCamelCase__ : Union[str, Any] = encoded_candidates.get('attention_mask' )
lowerCamelCase__ : Tuple = encoded_candidates.get('token_type_ids' )
if encoded_input_ids is not None:
output_data["input_ids"].append(UpperCAmelCase )
if encoded_attention_mask is not None:
output_data["attention_mask"].append(UpperCAmelCase )
if encoded_token_type_ids is not None:
output_data["token_type_ids"].append(UpperCAmelCase )
lowerCamelCase__ : int = {key: item for key, item in output_data.items() if len(UpperCAmelCase ) != 0}
return BatchEncoding(UpperCAmelCase , tensor_type=UpperCAmelCase )
def A_ ( self : int , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Optional[int]=None ) -> List[str]:
lowerCamelCase__ : Tuple = [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 A_ ( self : Tuple , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : List[Any] = [self.sep_token_id]
lowerCamelCase__ : Optional[int] = [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 A_ ( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
lowerCamelCase__ : int = self._tokenizer.model.save(UpperCAmelCase , name=UpperCAmelCase )
return tuple(UpperCAmelCase )
| 50
| 1
|
# Copyright 2022 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.
import argparse
import os
import subprocess
from packaging.version import Version, parse
from accelerate.commands.config.config_args import default_config_file, load_config_from_file
_UpperCAmelCase : Tuple = """Run commands across TPU VMs for initial setup before running `accelerate launch`."""
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase=None ) -> List[str]:
if subparsers is not None:
lowerCamelCase__ : Dict = subparsers.add_parser('tpu-config' , description=_description )
else:
lowerCamelCase__ : Optional[Any] = argparse.ArgumentParser('Accelerate tpu-config command' , description=_description )
# Core arguments
lowerCamelCase__ : Any = parser.add_argument_group(
'Config Arguments' , 'Arguments that can be configured through `accelerate config`.' )
config_args.add_argument(
'--config_file' , type=_UpperCAmelCase , default=_UpperCAmelCase , help='Path to the config file to use for accelerate.' , )
config_args.add_argument(
'--tpu_name' , default=_UpperCAmelCase , help='The name of the TPU to use. If not specified, will use the TPU specified in the config file.' , )
config_args.add_argument(
'--tpu_zone' , default=_UpperCAmelCase , help='The zone of the TPU to use. If not specified, will use the zone specified in the config file.' , )
lowerCamelCase__ : Optional[Any] = parser.add_argument_group('TPU Arguments' , 'Arguments for options ran inside the TPU.' )
pod_args.add_argument(
'--use_alpha' , action='store_true' , help='Whether to use `gcloud alpha` when running the TPU training script instead of `gcloud`.' , )
pod_args.add_argument(
'--command_file' , default=_UpperCAmelCase , help='The path to the file containing the commands to run on the pod on startup.' , )
pod_args.add_argument(
'--command' , action='append' , nargs='+' , help='A command to run on the pod. Can be passed multiple times.' , )
pod_args.add_argument(
'--install_accelerate' , action='store_true' , help='Whether to install accelerate on the pod. Defaults to False.' , )
pod_args.add_argument(
'--accelerate_version' , default='latest' , help='The version of accelerate to install on the pod. If not specified, will use the latest pypi version. Specify \'dev\' to install from GitHub.' , )
pod_args.add_argument(
'--debug' , action='store_true' , help='If set, will print the command that would be run instead of running it.' )
if subparsers is not None:
parser.set_defaults(func=_UpperCAmelCase )
return parser
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Any:
lowerCamelCase__ : Optional[int] = None
# Get the default from the config file if it exists.
if args.config_file is not None or os.path.isfile(_UpperCAmelCase ):
lowerCamelCase__ : List[str] = load_config_from_file(args.config_file )
if not args.command_file and defaults.command_file is not None and not args.command:
lowerCamelCase__ : Union[str, Any] = defaults.command_file
if not args.command and defaults.commands is not None:
lowerCamelCase__ : Optional[int] = defaults.commands
if not args.tpu_name:
lowerCamelCase__ : Any = defaults.tpu_name
if not args.tpu_zone:
lowerCamelCase__ : List[str] = defaults.tpu_zone
if args.accelerate_version == "dev":
lowerCamelCase__ : int = 'git+https://github.com/huggingface/accelerate.git'
elif args.accelerate_version == "latest":
lowerCamelCase__ : Union[str, Any] = 'accelerate -U'
elif isinstance(parse(args.accelerate_version ) , _UpperCAmelCase ):
lowerCamelCase__ : Dict = F"""accelerate=={args.accelerate_version}"""
if not args.command_file and not args.command:
raise ValueError('You must specify either a command file or a command to run on the pod.' )
if args.command_file:
with open(args.command_file , 'r' ) as f:
lowerCamelCase__ : Any = [f.read().splitlines()]
# To turn list of lists into list of strings
if isinstance(args.command[0] , _UpperCAmelCase ):
lowerCamelCase__ : List[Any] = [line for cmd in args.command for line in cmd]
# Default to the shared folder and install accelerate
lowerCamelCase__ : str = ['cd /usr/share']
if args.install_accelerate:
new_cmd += [F"""pip install {args.accelerate_version}"""]
new_cmd += args.command
lowerCamelCase__ : Dict = '; '.join(_UpperCAmelCase )
# Then send it to gcloud
# Eventually try to use google-api-core to do this instead of subprocess
lowerCamelCase__ : int = ['gcloud']
if args.use_alpha:
cmd += ["alpha"]
cmd += [
"compute",
"tpus",
"tpu-vm",
"ssh",
args.tpu_name,
"--zone",
args.tpu_zone,
"--command",
args.command,
"--worker",
"all",
]
if args.debug:
print(F"""Running {" ".join(_UpperCAmelCase )}""" )
return
subprocess.run(_UpperCAmelCase )
print('Successfully setup pod.' )
def SCREAMING_SNAKE_CASE ( ) -> List[str]:
lowerCamelCase__ : Any = tpu_command_parser()
lowerCamelCase__ : Tuple = parser.parse_args()
tpu_command_launcher(_UpperCAmelCase )
| 50
|
import logging
import os
from dataclasses import dataclass
from typing import List, Optional, Union
import tqdm
from filelock import FileLock
from transformers import (
BartTokenizer,
BartTokenizerFast,
DataProcessor,
PreTrainedTokenizer,
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
is_tf_available,
is_torch_available,
)
_UpperCAmelCase : Optional[Any] = logging.getLogger(__name__)
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
if is_torch_available():
import torch
from torch.utils.data import Dataset
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
def __init__( self : int , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : List[str]=False , UpperCAmelCase : bool = False , ) -> List[str]:
lowerCamelCase__ : int = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = os.path.join(
UpperCAmelCase , 'cached_{}_{}_{}_{}'.format(
'dev' if evaluate else 'train' , tokenizer.__class__.__name__ , str(UpperCAmelCase ) , UpperCAmelCase , ) , )
lowerCamelCase__ : int = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = label_list[2], label_list[1]
lowerCamelCase__ : List[str] = label_list
# Make sure only the first process in distributed training processes the dataset,
# and the others will use the cache.
lowerCamelCase__ : str = cached_features_file + '.lock'
with FileLock(UpperCAmelCase ):
if os.path.exists(UpperCAmelCase ) and not overwrite_cache:
logger.info(F"""Loading features from cached file {cached_features_file}""" )
lowerCamelCase__ : int = torch.load(UpperCAmelCase )
else:
logger.info(F"""Creating features from dataset file at {data_dir}""" )
lowerCamelCase__ : str = (
processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
)
logger.info('Training examples: %s' , len(UpperCAmelCase ) )
lowerCamelCase__ : Dict = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
logger.info('Saving features into cached file %s' , UpperCAmelCase )
torch.save(self.features , UpperCAmelCase )
def __len__( self : Optional[int] ) -> Optional[Any]:
return len(self.features )
def __getitem__( self : Tuple , UpperCAmelCase : Dict ) -> InputFeatures:
return self.features[i]
def A_ ( self : int ) -> int:
return self.label_list
if is_tf_available():
import tensorflow as tf
class lowerCAmelCase :
UpperCAmelCase__ = 42
def __init__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = 128 , UpperCAmelCase : Any=False , UpperCAmelCase : bool = False , ) -> Union[str, Any]:
lowerCamelCase__ : Any = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : str = label_list[2], label_list[1]
lowerCamelCase__ : Optional[int] = label_list
lowerCamelCase__ : int = processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def gen():
for ex_index, ex in tqdm.tqdm(enumerate(self.features ) , desc='convert examples to features' ):
if ex_index % 10000 == 0:
logger.info('Writing example %d of %d' % (ex_index, len(UpperCAmelCase )) )
yield (
{
"example_id": 0,
"input_ids": ex.input_ids,
"attention_mask": ex.attention_mask,
"token_type_ids": ex.token_type_ids,
},
ex.label,
)
lowerCamelCase__ : Optional[int] = tf.data.Dataset.from_generator(
UpperCAmelCase , (
{
'example_id': tf.intaa,
'input_ids': tf.intaa,
'attention_mask': tf.intaa,
'token_type_ids': tf.intaa,
},
tf.intaa,
) , (
{
'example_id': tf.TensorShape([] ),
'input_ids': tf.TensorShape([None, None] ),
'attention_mask': tf.TensorShape([None, None] ),
'token_type_ids': tf.TensorShape([None, None] ),
},
tf.TensorShape([] ),
) , )
def A_ ( self : Any ) -> Any:
return self.dataset
def __len__( self : Tuple ) -> int:
return len(self.features )
def __getitem__( self : List[str] , UpperCAmelCase : Any ) -> InputFeatures:
return self.features[i]
def A_ ( self : Dict ) -> str:
return self.label_list
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : int , UpperCAmelCase : List[Any] ) -> int:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_train_set.txt' ) ) , 'train' )
def A_ ( self : Any , UpperCAmelCase : int ) -> List[Any]:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_evaluation_set.txt' ) ) , 'dev' )
def A_ ( self : Any ) -> List[Any]:
return ["contradiction", "entailment", "neutral"]
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : List[str] ) -> List[str]:
lowerCamelCase__ : List[str] = []
for i, line in enumerate(UpperCAmelCase ):
if i == 0:
continue
lowerCamelCase__ : Tuple = '%s-%s' % (set_type, line[0])
lowerCamelCase__ : str = line[5]
lowerCamelCase__ : Dict = line[6]
lowerCamelCase__ : int = line[7][2:] if line[7].startswith('ex' ) else line[7]
lowerCamelCase__ : Dict = line[0]
examples.append(InputExample(guid=UpperCAmelCase , text_a=UpperCAmelCase , text_b=UpperCAmelCase , label=UpperCAmelCase , pairID=UpperCAmelCase ) )
return examples
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ) -> Optional[int]:
lowerCamelCase__ : int = {label: i for i, label in enumerate(_UpperCAmelCase )}
lowerCamelCase__ : List[Any] = []
for ex_index, example in tqdm.tqdm(enumerate(_UpperCAmelCase ) , desc='convert examples to features' ):
if ex_index % 1_0000 == 0:
logger.info('Writing example %d' % (ex_index) )
lowerCamelCase__ : List[Any] = tokenizer(
example.text_a , example.text_b , add_special_tokens=_UpperCAmelCase , max_length=_UpperCAmelCase , padding='max_length' , truncation=_UpperCAmelCase , return_overflowing_tokens=_UpperCAmelCase , )
lowerCamelCase__ : List[str] = label_map[example.label] if example.label in label_map else 0
lowerCamelCase__ : Optional[int] = int(example.pairID )
features.append(InputFeatures(**_UpperCAmelCase , label=_UpperCAmelCase , pairID=_UpperCAmelCase ) )
for i, example in enumerate(examples[:5] ):
logger.info('*** Example ***' )
logger.info(F"""guid: {example}""" )
logger.info(F"""features: {features[i]}""" )
return features
_UpperCAmelCase : str = {
"""hans""": 3,
}
_UpperCAmelCase : List[Any] = {
"""hans""": HansProcessor,
}
| 50
| 1
|
import argparse
import collections
import torch
from flax import traverse_util
from tax import checkpoints
from transformers import TaConfig, TaEncoderModel, TaForConditionalGeneration
from transformers.utils import logging
logging.set_verbosity_info()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase="attention" ) -> Dict:
lowerCamelCase__ : int = params[F"""{prefix}/layers_{i}/{layer_name}/key/kernel"""]
lowerCamelCase__ : int = params[F"""{prefix}/layers_{i}/{layer_name}/out/kernel"""]
lowerCamelCase__ : List[Any] = params[F"""{prefix}/layers_{i}/{layer_name}/query/kernel"""]
lowerCamelCase__ : str = params[F"""{prefix}/layers_{i}/{layer_name}/value/kernel"""]
return k, o, q, v
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=False ) -> List[Any]:
if split_mlp_wi:
lowerCamelCase__ : Optional[Any] = params[F"""{prefix}/layers_{i}/mlp/wi_0/kernel"""]
lowerCamelCase__ : Optional[int] = params[F"""{prefix}/layers_{i}/mlp/wi_1/kernel"""]
lowerCamelCase__ : Union[str, Any] = (wi_a, wi_a)
else:
lowerCamelCase__ : Tuple = params[F"""{prefix}/layers_{i}/mlp/wi/kernel"""]
lowerCamelCase__ : str = params[F"""{prefix}/layers_{i}/mlp/wo/kernel"""]
return wi, wo
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Any:
return params[F"""{prefix}/layers_{i}/{layer_name}/scale"""]
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , *, _UpperCAmelCase , _UpperCAmelCase ) -> Optional[Any]:
lowerCamelCase__ : List[str] = traverse_util.flatten_dict(variables['target'] )
lowerCamelCase__ : Optional[int] = {'/'.join(_UpperCAmelCase ): v for k, v in old.items()}
# v1.1 models have a gated GeLU with wi_0 and wi_1 instead of wi
lowerCamelCase__ : Optional[Any] = 'encoder/layers_0/mlp/wi_0/kernel' in old
print('Split MLP:' , _UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = collections.OrderedDict()
# Shared embeddings.
lowerCamelCase__ : Optional[Any] = old['token_embedder/embedding']
# Encoder.
for i in range(_UpperCAmelCase ):
# Block i, layer 0 (Self Attention).
lowerCamelCase__ : str = tax_layer_norm_lookup(_UpperCAmelCase , _UpperCAmelCase , 'encoder' , 'pre_attention_layer_norm' )
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : int = tax_attention_lookup(_UpperCAmelCase , _UpperCAmelCase , 'encoder' , 'attention' )
lowerCamelCase__ : Dict = layer_norm
lowerCamelCase__ : List[Any] = k.T
lowerCamelCase__ : Dict = o.T
lowerCamelCase__ : Optional[int] = q.T
lowerCamelCase__ : Union[str, Any] = v.T
# Block i, layer 1 (MLP).
lowerCamelCase__ : Dict = tax_layer_norm_lookup(_UpperCAmelCase , _UpperCAmelCase , 'encoder' , 'pre_mlp_layer_norm' )
lowerCamelCase__ , lowerCamelCase__ : Any = tax_mlp_lookup(_UpperCAmelCase , _UpperCAmelCase , 'encoder' , _UpperCAmelCase )
lowerCamelCase__ : List[Any] = layer_norm
if split_mlp_wi:
lowerCamelCase__ : str = wi[0].T
lowerCamelCase__ : Dict = wi[1].T
else:
lowerCamelCase__ : Optional[Any] = wi.T
lowerCamelCase__ : Tuple = wo.T
lowerCamelCase__ : int = old[
'encoder/relpos_bias/rel_embedding'
].T
lowerCamelCase__ : int = old['encoder/encoder_norm/scale']
if not is_encoder_only:
# Decoder.
for i in range(_UpperCAmelCase ):
# Block i, layer 0 (Self Attention).
lowerCamelCase__ : Optional[int] = tax_layer_norm_lookup(_UpperCAmelCase , _UpperCAmelCase , 'decoder' , 'pre_self_attention_layer_norm' )
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : Dict = tax_attention_lookup(_UpperCAmelCase , _UpperCAmelCase , 'decoder' , 'self_attention' )
lowerCamelCase__ : List[Any] = layer_norm
lowerCamelCase__ : List[Any] = k.T
lowerCamelCase__ : Union[str, Any] = o.T
lowerCamelCase__ : str = q.T
lowerCamelCase__ : int = v.T
# Block i, layer 1 (Cross Attention).
lowerCamelCase__ : str = tax_layer_norm_lookup(_UpperCAmelCase , _UpperCAmelCase , 'decoder' , 'pre_cross_attention_layer_norm' )
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = tax_attention_lookup(_UpperCAmelCase , _UpperCAmelCase , 'decoder' , 'encoder_decoder_attention' )
lowerCamelCase__ : str = layer_norm
lowerCamelCase__ : List[str] = k.T
lowerCamelCase__ : Tuple = o.T
lowerCamelCase__ : List[Any] = q.T
lowerCamelCase__ : int = v.T
# Block i, layer 2 (MLP).
lowerCamelCase__ : Optional[Any] = tax_layer_norm_lookup(_UpperCAmelCase , _UpperCAmelCase , 'decoder' , 'pre_mlp_layer_norm' )
lowerCamelCase__ , lowerCamelCase__ : List[Any] = tax_mlp_lookup(_UpperCAmelCase , _UpperCAmelCase , 'decoder' , _UpperCAmelCase )
lowerCamelCase__ : List[str] = layer_norm
if split_mlp_wi:
lowerCamelCase__ : str = wi[0].T
lowerCamelCase__ : str = wi[1].T
else:
lowerCamelCase__ : Any = wi.T
lowerCamelCase__ : Optional[Any] = wo.T
lowerCamelCase__ : int = old['decoder/decoder_norm/scale']
lowerCamelCase__ : Optional[int] = old[
'decoder/relpos_bias/rel_embedding'
].T
# LM Head (only in v1.1 checkpoints, in v1.0 embeddings are used instead)
if "decoder/logits_dense/kernel" in old:
lowerCamelCase__ : Dict = old['decoder/logits_dense/kernel'].T
return new
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Optional[Any]:
lowerCamelCase__ : int = collections.OrderedDict([(k, torch.from_numpy(v.copy() )) for (k, v) in converted_params.items()] )
# Add what is missing.
if "encoder.embed_tokens.weight" not in state_dict:
lowerCamelCase__ : Optional[int] = state_dict['shared.weight']
if not is_encoder_only:
if "decoder.embed_tokens.weight" not in state_dict:
lowerCamelCase__ : Union[str, Any] = state_dict['shared.weight']
if "lm_head.weight" not in state_dict: # For old 1.0 models.
print('Using shared word embeddings as lm_head.' )
lowerCamelCase__ : Optional[int] = state_dict['shared.weight']
return state_dict
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Any = checkpoints.load_tax_checkpoint(_UpperCAmelCase )
lowerCamelCase__ : Dict = convert_tax_to_pytorch(_UpperCAmelCase , num_layers=config.num_layers , is_encoder_only=_UpperCAmelCase )
lowerCamelCase__ : Tuple = make_state_dict(_UpperCAmelCase , _UpperCAmelCase )
model.load_state_dict(_UpperCAmelCase , strict=_UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase = False ) -> str:
lowerCamelCase__ : Optional[int] = TaConfig.from_json_file(_UpperCAmelCase )
print(F"""Building PyTorch model from configuration: {config}""" )
# Non-v1.1 checkpoints could also use T5Model, but this works for all.
# The v1.0 checkpoints will simply have an LM head that is the word embeddings.
if is_encoder_only:
lowerCamelCase__ : List[str] = TaEncoderModel(_UpperCAmelCase )
else:
lowerCamelCase__ : Tuple = TaForConditionalGeneration(_UpperCAmelCase )
# Load weights from tf checkpoint
load_tax_weights_in_ta(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
# Save pytorch-model
print(F"""Save PyTorch model to {pytorch_dump_path}""" )
model.save_pretrained(_UpperCAmelCase )
# Verify that we can load the checkpoint.
model.from_pretrained(_UpperCAmelCase )
print('Done' )
if __name__ == "__main__":
_UpperCAmelCase : List[Any] = argparse.ArgumentParser(description="""Converts a native T5X checkpoint into a PyTorch checkpoint.""")
# Required parameters
parser.add_argument(
"""--t5x_checkpoint_path""", default=None, type=str, required=True, help="""Path to the T5X checkpoint."""
)
parser.add_argument(
"""--config_file""",
default=None,
type=str,
required=True,
help="""The config json file corresponding to the pre-trained T5 model.\nThis specifies the model architecture.""",
)
parser.add_argument(
"""--pytorch_dump_path""", default=None, type=str, required=True, help="""Path to the output PyTorch model."""
)
parser.add_argument(
"""--is_encoder_only""", action="""store_true""", help="""Check if the model is encoder-decoder model""", default=False
)
_UpperCAmelCase : Tuple = parser.parse_args()
convert_tax_checkpoint_to_pytorch(
args.tax_checkpoint_path, args.config_file, args.pytorch_dump_path, args.is_encoder_only
)
| 50
|
import unittest
from transformers import BertGenerationTokenizer
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
_UpperCAmelCase : Optional[Any] = """▁"""
_UpperCAmelCase : Optional[Any] = get_tests_dir("""fixtures/test_sentencepiece.model""")
@require_sentencepiece
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = BertGenerationTokenizer
UpperCAmelCase__ = False
UpperCAmelCase__ = True
def A_ ( self : List[Any] ) -> List[str]:
super().setUp()
lowerCamelCase__ : Dict = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
tokenizer.save_pretrained(self.tmpdirname )
def A_ ( self : Optional[Any] ) -> Dict:
lowerCamelCase__ : List[str] = '<s>'
lowerCamelCase__ : Optional[int] = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCAmelCase ) , UpperCAmelCase )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : List[str] ) -> Optional[int]:
lowerCamelCase__ : Optional[Any] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , '<unk>' )
self.assertEqual(vocab_keys[1] , '<s>' )
self.assertEqual(vocab_keys[-1] , '<pad>' )
self.assertEqual(len(UpperCAmelCase ) , 1002 )
def A_ ( self : List[Any] ) -> Optional[Any]:
self.assertEqual(self.get_tokenizer().vocab_size , 1000 )
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : List[str] = tokenizer.tokenize('This is a test' )
self.assertListEqual(UpperCAmelCase , ['▁This', '▁is', '▁a', '▁t', 'est'] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(UpperCAmelCase ) , [285, 46, 10, 170, 382] , )
lowerCamelCase__ : List[str] = tokenizer.tokenize('I was born in 92000, and this is falsé.' )
self.assertListEqual(
UpperCAmelCase , [
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',
'.',
] , )
lowerCamelCase__ : Optional[int] = tokenizer.convert_tokens_to_ids(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
lowerCamelCase__ : Optional[Any] = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [
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>',
'.',
] , )
@cached_property
def A_ ( self : Dict ) -> Tuple:
return BertGenerationTokenizer.from_pretrained('google/bert_for_seq_generation_L-24_bbc_encoder' )
@slow
def A_ ( self : Optional[int] ) -> List[str]:
lowerCamelCase__ : Union[str, Any] = 'Hello World!'
lowerCamelCase__ : Dict = [18536, 2260, 101]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@slow
def A_ ( self : Optional[Any] ) -> str:
lowerCamelCase__ : List[Any] = (
'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'
)
lowerCamelCase__ : Any = [
871,
419,
358,
946,
991,
2521,
452,
358,
1357,
387,
7751,
3536,
112,
985,
456,
126,
865,
938,
5400,
5734,
458,
1368,
467,
786,
2462,
5246,
1159,
633,
865,
4519,
457,
582,
852,
2557,
427,
916,
508,
405,
34324,
497,
391,
408,
11342,
1244,
385,
100,
938,
985,
456,
574,
362,
12597,
3200,
3129,
1172,
]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@require_torch
@slow
def A_ ( self : int ) -> Optional[Any]:
import torch
from transformers import BertGenerationConfig, BertGenerationEncoder
# Build sequence
lowerCamelCase__ : str = list(self.big_tokenizer.get_vocab().keys() )[:10]
lowerCamelCase__ : int = ' '.join(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.big_tokenizer.encode_plus(UpperCAmelCase , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.big_tokenizer.batch_encode_plus(
[sequence + ' ' + sequence] , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Tuple = BertGenerationConfig()
lowerCamelCase__ : Optional[Any] = BertGenerationEncoder(UpperCAmelCase )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**UpperCAmelCase )
model(**UpperCAmelCase )
@slow
def A_ ( self : Optional[int] ) -> List[Any]:
# fmt: off
lowerCamelCase__ : Any = {'input_ids': [[39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114], [448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 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], [484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 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]], '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, 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, 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=UpperCAmelCase , model_name='google/bert_for_seq_generation_L-24_bbc_encoder' , revision='c817d1fd1be2ffa69431227a1fe320544943d4db' , )
| 50
| 1
|
import copy
import os
from typing import Union
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_UpperCAmelCase : Union[str, Any] = logging.get_logger(__name__)
_UpperCAmelCase : str = {
"""google/pix2struct-textcaps-base""": (
"""https://huggingface.co/google/pix2struct-textcaps-base/resolve/main/config.json"""
),
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """pix2struct_text_model"""
UpperCAmelCase__ = ["""past_key_values"""]
UpperCAmelCase__ = {
"""hidden_size""": """hidden_size""",
"""num_attention_heads""": """num_heads""",
"""num_hidden_layers""": """num_layers""",
}
def __init__( self : str , UpperCAmelCase : Any=50244 , UpperCAmelCase : Union[str, Any]=768 , UpperCAmelCase : List[Any]=64 , UpperCAmelCase : str=2048 , UpperCAmelCase : Optional[int]=12 , UpperCAmelCase : Optional[Any]=12 , UpperCAmelCase : Tuple=32 , UpperCAmelCase : Optional[Any]=128 , UpperCAmelCase : List[str]=0.1 , UpperCAmelCase : List[Any]=1e-6 , UpperCAmelCase : List[str]=1.0 , UpperCAmelCase : Any="gelu_new" , UpperCAmelCase : Optional[int]=0 , UpperCAmelCase : Optional[int]=False , UpperCAmelCase : List[Any]=0 , UpperCAmelCase : Optional[Any]=1 , UpperCAmelCase : Dict=False , UpperCAmelCase : List[str]=True , **UpperCAmelCase : Optional[int] , ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = vocab_size
lowerCamelCase__ : List[str] = hidden_size
lowerCamelCase__ : Optional[int] = d_kv
lowerCamelCase__ : Union[str, Any] = d_ff
lowerCamelCase__ : Dict = num_layers
lowerCamelCase__ : Tuple = num_heads
lowerCamelCase__ : str = relative_attention_num_buckets
lowerCamelCase__ : Optional[Any] = relative_attention_max_distance
lowerCamelCase__ : Tuple = dropout_rate
lowerCamelCase__ : str = layer_norm_epsilon
lowerCamelCase__ : str = initializer_factor
lowerCamelCase__ : Optional[Any] = use_cache
lowerCamelCase__ : Dict = eos_token_id
lowerCamelCase__ : Dict = decoder_start_token_id
# for backwards compatibility
lowerCamelCase__ : Tuple = dense_act_fn
super().__init__(
pad_token_id=UpperCAmelCase , eos_token_id=UpperCAmelCase , decoder_start_token_id=UpperCAmelCase , tie_word_embeddings=UpperCAmelCase , is_decoder=UpperCAmelCase , **UpperCAmelCase , )
@classmethod
def A_ ( cls : Tuple , UpperCAmelCase : Union[str, os.PathLike] , **UpperCAmelCase : str ) -> "PretrainedConfig":
cls._set_token_in_kwargs(UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = cls.get_config_dict(UpperCAmelCase , **UpperCAmelCase )
# get the text config dict if we are loading from Pix2StructConfig
if config_dict.get('model_type' ) == "pix2struct":
lowerCamelCase__ : Union[str, Any] = config_dict['text_config']
if "model_type" in config_dict and hasattr(cls , 'model_type' ) and config_dict["model_type"] != cls.model_type:
logger.warning(
F"""You are using a model of type {config_dict["model_type"]} to instantiate a model of type """
F"""{cls.model_type}. This is not supported for all configurations of models and can yield errors.""" )
return cls.from_dict(UpperCAmelCase , **UpperCAmelCase )
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """pix2struct_vision_model"""
def __init__( self : Optional[int] , UpperCAmelCase : int=768 , UpperCAmelCase : Optional[Any]=768 , UpperCAmelCase : Optional[Any]=2048 , UpperCAmelCase : str=64 , UpperCAmelCase : Union[str, Any]=12 , UpperCAmelCase : List[str]=12 , UpperCAmelCase : Any="gelu_new" , UpperCAmelCase : List[Any]=1e-6 , UpperCAmelCase : Any=0.0 , UpperCAmelCase : List[Any]=0.0 , UpperCAmelCase : Union[str, Any]=1e-10 , UpperCAmelCase : List[Any]=1.0 , UpperCAmelCase : Optional[int]=4096 , UpperCAmelCase : Union[str, Any]=32 , UpperCAmelCase : Union[str, Any]=128 , **UpperCAmelCase : List[str] , ) -> Tuple:
super().__init__(**UpperCAmelCase )
lowerCamelCase__ : List[str] = hidden_size
lowerCamelCase__ : Any = patch_embed_hidden_size
lowerCamelCase__ : Any = d_ff
lowerCamelCase__ : int = dropout_rate
lowerCamelCase__ : Optional[Any] = num_hidden_layers
lowerCamelCase__ : Optional[int] = num_attention_heads
lowerCamelCase__ : int = initializer_range
lowerCamelCase__ : Optional[int] = initializer_factor
lowerCamelCase__ : int = attention_dropout
lowerCamelCase__ : str = layer_norm_eps
lowerCamelCase__ : List[Any] = dense_act_fn
lowerCamelCase__ : int = seq_len
lowerCamelCase__ : str = relative_attention_num_buckets
lowerCamelCase__ : List[str] = relative_attention_max_distance
lowerCamelCase__ : Dict = d_kv
@classmethod
def A_ ( cls : Any , UpperCAmelCase : Union[str, os.PathLike] , **UpperCAmelCase : int ) -> "PretrainedConfig":
cls._set_token_in_kwargs(UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = cls.get_config_dict(UpperCAmelCase , **UpperCAmelCase )
# get the vision config dict if we are loading from Pix2StructConfig
if config_dict.get('model_type' ) == "pix2struct":
lowerCamelCase__ : Optional[Any] = config_dict['vision_config']
if "model_type" in config_dict and hasattr(cls , 'model_type' ) and config_dict["model_type"] != cls.model_type:
logger.warning(
F"""You are using a model of type {config_dict["model_type"]} to instantiate a model of type """
F"""{cls.model_type}. This is not supported for all configurations of models and can yield errors.""" )
return cls.from_dict(UpperCAmelCase , **UpperCAmelCase )
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """pix2struct"""
UpperCAmelCase__ = True
def __init__( self : str , UpperCAmelCase : Union[str, Any]=None , UpperCAmelCase : str=None , UpperCAmelCase : str=1.0 , UpperCAmelCase : str=0.0_2 , UpperCAmelCase : int=False , UpperCAmelCase : Optional[int]=False , UpperCAmelCase : List[Any]=True , **UpperCAmelCase : Dict , ) -> Optional[int]:
super().__init__(tie_word_embeddings=UpperCAmelCase , is_encoder_decoder=UpperCAmelCase , **UpperCAmelCase )
if text_config is None:
lowerCamelCase__ : List[Any] = {}
logger.info('text_config is None. Initializing the Pix2StructTextConfig with default values.' )
if vision_config is None:
lowerCamelCase__ : Tuple = {}
logger.info('vision_config is None. Initializing the Pix2StructVisionConfig with default values.' )
lowerCamelCase__ : List[Any] = PixaStructTextConfig(**UpperCAmelCase )
lowerCamelCase__ : List[Any] = PixaStructVisionConfig(**UpperCAmelCase )
lowerCamelCase__ : List[str] = self.text_config.decoder_start_token_id
lowerCamelCase__ : int = self.text_config.pad_token_id
lowerCamelCase__ : Any = self.text_config.eos_token_id
lowerCamelCase__ : Optional[int] = initializer_factor
lowerCamelCase__ : Any = initializer_range
lowerCamelCase__ : Union[str, Any] = self.initializer_range
lowerCamelCase__ : Optional[Any] = self.initializer_range
lowerCamelCase__ : List[str] = is_vqa
@classmethod
def A_ ( cls : Dict , UpperCAmelCase : PixaStructTextConfig , UpperCAmelCase : PixaStructVisionConfig , **UpperCAmelCase : Union[str, Any] ) -> str:
return cls(text_config=text_config.to_dict() , vision_config=vision_config.to_dict() , **UpperCAmelCase )
def A_ ( self : int ) -> Optional[int]:
lowerCamelCase__ : Dict = copy.deepcopy(self.__dict__ )
lowerCamelCase__ : List[Any] = self.text_config.to_dict()
lowerCamelCase__ : Any = self.vision_config.to_dict()
lowerCamelCase__ : List[Any] = self.__class__.model_type
return output
| 50
|
import os
import tempfile
from functools import partial
from unittest import TestCase
from unittest.mock import patch
import numpy as np
import pytest
from datasets.arrow_dataset import Dataset
from datasets.search import ElasticSearchIndex, FaissIndex, MissingIndex
from .utils import require_elasticsearch, require_faiss
_UpperCAmelCase : str = pytest.mark.integration
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : List[Any] ) -> Union[str, Any]:
lowerCamelCase__ : List[Any] = Dataset.from_dict({'filename': ['my_name-train' + '_' + str(UpperCAmelCase ) for x in np.arange(30 ).tolist()]} )
return dset
def A_ ( self : Optional[Any] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
lowerCamelCase__ : List[Any] = dset.map(
lambda UpperCAmelCase , UpperCAmelCase : {"vecs": i * np.ones(5 , dtype=np.floataa )} , with_indices=UpperCAmelCase , keep_in_memory=UpperCAmelCase )
lowerCamelCase__ : Tuple = dset.add_faiss_index('vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
dset.drop_index('vecs' )
def A_ ( self : Union[str, Any] ) -> int:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT , )
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : List[str] ) -> Tuple:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , metric_type=faiss.METRIC_INNER_PRODUCT , )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
dset.save_faiss_index('vecs' , tmp_file.name )
dset.load_faiss_index('vecs2' , tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs2' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' )
dset.drop_index('vecs' )
self.assertRaises(UpperCAmelCase , partial(dset.get_nearest_examples , 'vecs2' , np.ones(5 , dtype=np.floataa ) ) )
def A_ ( self : Dict ) -> Dict:
from elasticsearch import Elasticsearch
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : List[Any] = {'acknowledged': True}
mocked_bulk.return_value([(True, None)] * 30 )
lowerCamelCase__ : int = {'hits': {'hits': [{'_score': 1, '_id': 29}]}}
lowerCamelCase__ : List[str] = Elasticsearch()
dset.add_elasticsearch_index('filename' , es_client=UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Dict = dset.get_nearest_examples('filename' , 'my_name-train_29' )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Any ) -> Dict:
import faiss
lowerCamelCase__ : Tuple = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
# add vectors
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsNotNone(index.faiss_index )
self.assertEqual(index.faiss_index.ntotal , 5 )
index.add_vectors(np.zeros((5, 5) , dtype=np.floataa ) )
self.assertEqual(index.faiss_index.ntotal , 10 )
# single query
lowerCamelCase__ : int = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Any = 1
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = index.search(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search , query.reshape(-1 , 1 ) )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
# batched queries
lowerCamelCase__ : str = np.eye(5 , dtype=np.floataa )[::-1]
lowerCamelCase__ , lowerCamelCase__ : Dict = index.search_batch(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search_batch , queries[0] )
lowerCamelCase__ : str = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[str] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([4, 3, 2, 1, 0] , UpperCAmelCase )
def A_ ( self : List[Any] ) -> List[Any]:
import faiss
lowerCamelCase__ : Any = FaissIndex(string_factory='Flat' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
lowerCamelCase__ : Tuple = FaissIndex(string_factory='LSH' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexLSH )
with self.assertRaises(UpperCAmelCase ):
lowerCamelCase__ : List[str] = FaissIndex(string_factory='Flat' , custom_index=faiss.IndexFlat(5 ) )
def A_ ( self : List[str] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Optional[Any] = faiss.IndexFlat(5 )
lowerCamelCase__ : int = FaissIndex(custom_index=UpperCAmelCase )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
def A_ ( self : Any ) -> Optional[int]:
import faiss
lowerCamelCase__ : Any = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
index.save(tmp_file.name )
lowerCamelCase__ : List[Any] = FaissIndex.load(tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ : List[str] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Tuple = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(UpperCAmelCase )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
@require_faiss
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Any:
import faiss
lowerCamelCase__ : Optional[Any] = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
lowerCamelCase__ : Optional[int] = 'index.faiss'
lowerCamelCase__ : Optional[Any] = F"""mock://{index_name}"""
index.save(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Tuple = FaissIndex.load(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Optional[int] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Dict = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(_UpperCAmelCase )
assert scores[0] > 0
assert indices[0] == 1
@require_elasticsearch
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Dict ) -> List[Any]:
from elasticsearch import Elasticsearch
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : Any = Elasticsearch()
lowerCamelCase__ : Tuple = {'acknowledged': True}
lowerCamelCase__ : Tuple = ElasticSearchIndex(es_client=UpperCAmelCase )
mocked_bulk.return_value([(True, None)] * 3 )
index.add_documents(['foo', 'bar', 'foobar'] )
# single query
lowerCamelCase__ : Optional[int] = 'foo'
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search(UpperCAmelCase )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# single query with timeout
lowerCamelCase__ : Any = 'foo'
lowerCamelCase__ : List[str] = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Tuple = index.search(UpperCAmelCase , request_timeout=30 )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# batched queries
lowerCamelCase__ : List[str] = ['foo', 'bar', 'foobar']
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : str = index.search_batch(UpperCAmelCase )
lowerCamelCase__ : List[str] = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[Any] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
# batched queries with timeout
lowerCamelCase__ : str = ['foo', 'bar', 'foobar']
lowerCamelCase__ : List[Any] = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search_batch(UpperCAmelCase , request_timeout=30 )
lowerCamelCase__ : Optional[Any] = [scores[0] for scores in total_scores]
lowerCamelCase__ : Dict = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
| 50
| 1
|
import unittest
from transformers import BertGenerationTokenizer
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
_UpperCAmelCase : Optional[Any] = """▁"""
_UpperCAmelCase : Optional[Any] = get_tests_dir("""fixtures/test_sentencepiece.model""")
@require_sentencepiece
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = BertGenerationTokenizer
UpperCAmelCase__ = False
UpperCAmelCase__ = True
def A_ ( self : List[Any] ) -> List[str]:
super().setUp()
lowerCamelCase__ : Dict = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
tokenizer.save_pretrained(self.tmpdirname )
def A_ ( self : Optional[Any] ) -> Dict:
lowerCamelCase__ : List[str] = '<s>'
lowerCamelCase__ : Optional[int] = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCAmelCase ) , UpperCAmelCase )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : List[str] ) -> Optional[int]:
lowerCamelCase__ : Optional[Any] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , '<unk>' )
self.assertEqual(vocab_keys[1] , '<s>' )
self.assertEqual(vocab_keys[-1] , '<pad>' )
self.assertEqual(len(UpperCAmelCase ) , 1002 )
def A_ ( self : List[Any] ) -> Optional[Any]:
self.assertEqual(self.get_tokenizer().vocab_size , 1000 )
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : List[str] = tokenizer.tokenize('This is a test' )
self.assertListEqual(UpperCAmelCase , ['▁This', '▁is', '▁a', '▁t', 'est'] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(UpperCAmelCase ) , [285, 46, 10, 170, 382] , )
lowerCamelCase__ : List[str] = tokenizer.tokenize('I was born in 92000, and this is falsé.' )
self.assertListEqual(
UpperCAmelCase , [
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',
'.',
] , )
lowerCamelCase__ : Optional[int] = tokenizer.convert_tokens_to_ids(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
lowerCamelCase__ : Optional[Any] = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [
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>',
'.',
] , )
@cached_property
def A_ ( self : Dict ) -> Tuple:
return BertGenerationTokenizer.from_pretrained('google/bert_for_seq_generation_L-24_bbc_encoder' )
@slow
def A_ ( self : Optional[int] ) -> List[str]:
lowerCamelCase__ : Union[str, Any] = 'Hello World!'
lowerCamelCase__ : Dict = [18536, 2260, 101]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@slow
def A_ ( self : Optional[Any] ) -> str:
lowerCamelCase__ : List[Any] = (
'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'
)
lowerCamelCase__ : Any = [
871,
419,
358,
946,
991,
2521,
452,
358,
1357,
387,
7751,
3536,
112,
985,
456,
126,
865,
938,
5400,
5734,
458,
1368,
467,
786,
2462,
5246,
1159,
633,
865,
4519,
457,
582,
852,
2557,
427,
916,
508,
405,
34324,
497,
391,
408,
11342,
1244,
385,
100,
938,
985,
456,
574,
362,
12597,
3200,
3129,
1172,
]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@require_torch
@slow
def A_ ( self : int ) -> Optional[Any]:
import torch
from transformers import BertGenerationConfig, BertGenerationEncoder
# Build sequence
lowerCamelCase__ : str = list(self.big_tokenizer.get_vocab().keys() )[:10]
lowerCamelCase__ : int = ' '.join(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.big_tokenizer.encode_plus(UpperCAmelCase , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.big_tokenizer.batch_encode_plus(
[sequence + ' ' + sequence] , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Tuple = BertGenerationConfig()
lowerCamelCase__ : Optional[Any] = BertGenerationEncoder(UpperCAmelCase )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**UpperCAmelCase )
model(**UpperCAmelCase )
@slow
def A_ ( self : Optional[int] ) -> List[Any]:
# fmt: off
lowerCamelCase__ : Any = {'input_ids': [[39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114], [448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 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], [484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 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]], '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, 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, 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=UpperCAmelCase , model_name='google/bert_for_seq_generation_L-24_bbc_encoder' , revision='c817d1fd1be2ffa69431227a1fe320544943d4db' , )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> bool:
lowerCamelCase__ : List[str] = len(_UpperCAmelCase )
lowerCamelCase__ : str = [[False] * (required_sum + 1) for _ in range(arr_len + 1 )]
# for each arr value, a sum of zero(0) can be formed by not taking any element
# hence True/1
for i in range(arr_len + 1 ):
lowerCamelCase__ : Tuple = True
# sum is not zero and set is empty then false
for i in range(1 , required_sum + 1 ):
lowerCamelCase__ : Dict = False
for i in range(1 , arr_len + 1 ):
for j in range(1 , required_sum + 1 ):
if arr[i - 1] > j:
lowerCamelCase__ : str = subset[i - 1][j]
if arr[i - 1] <= j:
lowerCamelCase__ : Dict = subset[i - 1][j] or subset[i - 1][j - arr[i - 1]]
return subset[arr_len][required_sum]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
from string import ascii_uppercase
_UpperCAmelCase : Tuple = {char: i for i, char in enumerate(ascii_uppercase)}
_UpperCAmelCase : Optional[Any] = dict(enumerate(ascii_uppercase))
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> str:
lowerCamelCase__ : Optional[Any] = len(_UpperCAmelCase )
lowerCamelCase__ : List[Any] = 0
while True:
if x == i:
lowerCamelCase__ : Any = 0
if len(_UpperCAmelCase ) == len(_UpperCAmelCase ):
break
key += key[i]
i += 1
return key
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> str:
lowerCamelCase__ : Union[str, Any] = ''
lowerCamelCase__ : Union[str, Any] = 0
for letter in message:
if letter == " ":
cipher_text += " "
else:
lowerCamelCase__ : Any = (dicta[letter] - dicta[key_new[i]]) % 26
i += 1
cipher_text += dicta[x]
return cipher_text
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> str:
lowerCamelCase__ : Dict = ''
lowerCamelCase__ : Tuple = 0
for letter in cipher_text:
if letter == " ":
or_txt += " "
else:
lowerCamelCase__ : Tuple = (dicta[letter] + dicta[key_new[i]] + 26) % 26
i += 1
or_txt += dicta[x]
return or_txt
def SCREAMING_SNAKE_CASE ( ) -> None:
lowerCamelCase__ : Tuple = 'THE GERMAN ATTACK'
lowerCamelCase__ : Optional[Any] = 'SECRET'
lowerCamelCase__ : int = generate_key(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : int = cipher_text(_UpperCAmelCase , _UpperCAmelCase )
print(F"""Encrypted Text = {s}""" )
print(F"""Original Text = {original_text(_UpperCAmelCase , _UpperCAmelCase )}""" )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 50
|
import torch
from transformers import PreTrainedModel, XLMRobertaConfig, XLMRobertaModel
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """M-CLIP"""
def __init__( self : Optional[Any] , UpperCAmelCase : Union[str, Any]=1024 , UpperCAmelCase : Tuple=768 , **UpperCAmelCase : Optional[int] ) -> Dict:
lowerCamelCase__ : Optional[int] = transformerDimSize
lowerCamelCase__ : Optional[Any] = imageDimSize
super().__init__(**UpperCAmelCase )
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = MCLIPConfig
def __init__( self : List[Any] , UpperCAmelCase : Dict , *UpperCAmelCase : Any , **UpperCAmelCase : Dict ) -> Dict:
super().__init__(UpperCAmelCase , *UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Tuple = XLMRobertaModel(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = torch.nn.Linear(
in_features=config.transformerDimensions , out_features=config.numDims )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Dict ) -> Tuple:
lowerCamelCase__ : Any = self.transformer(input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase )[0]
lowerCamelCase__ : int = (embs * attention_mask.unsqueeze(2 )).sum(dim=1 ) / attention_mask.sum(dim=1 )[:, None]
return self.LinearTransformation(UpperCAmelCase ), embs
| 50
| 1
|
import argparse
import os
# New Code #
import evaluate
import torch
from datasets import load_dataset
from torch.optim import AdamW
from torch.utils.data import DataLoader
from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed
from accelerate import Accelerator, DistributedType
from accelerate.utils import find_executable_batch_size
########################################################################
# This is a fully working simple example to use Accelerate,
# specifically showcasing how to ensure out-of-memory errors never
# interrupt training, and builds off the `nlp_example.py` script.
#
# This example trains a Bert base model on GLUE MRPC
# in any of the following settings (with the same script):
# - single CPU or single GPU
# - multi GPUS (using PyTorch distributed mode)
# - (multi) TPUs
# - fp16 (mixed-precision) or fp32 (normal precision)
#
# New additions from the base script can be found quickly by
# looking for the # New Code # tags
#
# To run it in each of these various modes, follow the instructions
# in the readme for examples:
# https://github.com/huggingface/accelerate/tree/main/examples
#
########################################################################
_UpperCAmelCase : Tuple = 16
_UpperCAmelCase : Any = 32
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase = 16 ) -> Tuple:
lowerCamelCase__ : List[str] = AutoTokenizer.from_pretrained('bert-base-cased' )
lowerCamelCase__ : int = load_dataset('glue' , 'mrpc' )
def tokenize_function(_UpperCAmelCase ):
# max_length=None => use the model max length (it's actually the default)
lowerCamelCase__ : Optional[Any] = tokenizer(examples['sentence1'] , examples['sentence2'] , truncation=_UpperCAmelCase , max_length=_UpperCAmelCase )
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
# starting with the main process first:
with accelerator.main_process_first():
lowerCamelCase__ : List[str] = datasets.map(
_UpperCAmelCase , batched=_UpperCAmelCase , remove_columns=['idx', 'sentence1', 'sentence2'] , )
# We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the
# transformers library
lowerCamelCase__ : Tuple = tokenized_datasets.rename_column('label' , 'labels' )
def collate_fn(_UpperCAmelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
lowerCamelCase__ : Union[str, Any] = 128 if accelerator.distributed_type == DistributedType.TPU else None
# When using mixed precision we want round multiples of 8/16
if accelerator.mixed_precision == "fp8":
lowerCamelCase__ : Dict = 16
elif accelerator.mixed_precision != "no":
lowerCamelCase__ : List[Any] = 8
else:
lowerCamelCase__ : Optional[int] = None
return tokenizer.pad(
_UpperCAmelCase , padding='longest' , max_length=_UpperCAmelCase , pad_to_multiple_of=_UpperCAmelCase , return_tensors='pt' , )
# Instantiate dataloaders.
lowerCamelCase__ : Tuple = DataLoader(
tokenized_datasets['train'] , shuffle=_UpperCAmelCase , collate_fn=_UpperCAmelCase , batch_size=_UpperCAmelCase )
lowerCamelCase__ : Any = DataLoader(
tokenized_datasets['validation'] , shuffle=_UpperCAmelCase , collate_fn=_UpperCAmelCase , batch_size=_UpperCAmelCase )
return train_dataloader, eval_dataloader
# For testing only
if os.environ.get("""TESTING_MOCKED_DATALOADERS""", None) == "1":
from accelerate.test_utils.training import mocked_dataloaders
_UpperCAmelCase : Tuple = mocked_dataloaders # noqa: F811
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Union[str, Any]:
# For testing only
if os.environ.get('TESTING_MOCKED_DATALOADERS' , _UpperCAmelCase ) == "1":
lowerCamelCase__ : Any = 2
# Initialize accelerator
lowerCamelCase__ : str = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision )
# Sample hyper-parameters for learning rate, batch size, seed and a few other HPs
lowerCamelCase__ : str = config['lr']
lowerCamelCase__ : Dict = int(config['num_epochs'] )
lowerCamelCase__ : int = int(config['seed'] )
lowerCamelCase__ : int = int(config['batch_size'] )
lowerCamelCase__ : str = evaluate.load('glue' , 'mrpc' )
# New Code #
# We now can define an inner training loop function. It should take a batch size as the only parameter,
# and build the dataloaders in there.
# It also gets our decorator
@find_executable_batch_size(starting_batch_size=_UpperCAmelCase )
def inner_training_loop(_UpperCAmelCase ):
# And now just move everything below under this function
# We need to bring in the Accelerator object from earlier
nonlocal accelerator
# And reset all of its attributes that could hold onto any memory:
accelerator.free_memory()
# Then we can declare the model, optimizer, and everything else:
set_seed(_UpperCAmelCase )
# Instantiate the model (we build the model here so that the seed also control new weights initialization)
lowerCamelCase__ : Tuple = AutoModelForSequenceClassification.from_pretrained('bert-base-cased' , return_dict=_UpperCAmelCase )
# We could avoid this line since the accelerator is set with `device_placement=True` (default value).
# Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer
# creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that).
lowerCamelCase__ : List[Any] = model.to(accelerator.device )
# Instantiate optimizer
lowerCamelCase__ : List[Any] = AdamW(params=model.parameters() , lr=_UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Tuple = get_dataloaders(_UpperCAmelCase , _UpperCAmelCase )
# Instantiate scheduler
lowerCamelCase__ : Union[str, Any] = get_linear_schedule_with_warmup(
optimizer=_UpperCAmelCase , num_warmup_steps=100 , num_training_steps=(len(_UpperCAmelCase ) * num_epochs) , )
# Prepare everything
# There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the
# prepare method.
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = accelerator.prepare(
_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
# Now we train the model
for epoch in range(_UpperCAmelCase ):
model.train()
for step, batch in enumerate(_UpperCAmelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
lowerCamelCase__ : List[str] = model(**_UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = outputs.loss
accelerator.backward(_UpperCAmelCase )
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
model.eval()
for step, batch in enumerate(_UpperCAmelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
with torch.no_grad():
lowerCamelCase__ : List[str] = model(**_UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = outputs.logits.argmax(dim=-1 )
lowerCamelCase__ , lowerCamelCase__ : int = accelerator.gather_for_metrics((predictions, batch['labels']) )
metric.add_batch(
predictions=_UpperCAmelCase , references=_UpperCAmelCase , )
lowerCamelCase__ : Tuple = metric.compute()
# Use accelerator.print to print only on the main process.
accelerator.print(F"""epoch {epoch}:""" , _UpperCAmelCase )
# New Code #
# And call it at the end with no arguments
# Note: You could also refactor this outside of your training loop function
inner_training_loop()
def SCREAMING_SNAKE_CASE ( ) -> Dict:
lowerCamelCase__ : Dict = argparse.ArgumentParser(description='Simple example of training script.' )
parser.add_argument(
'--mixed_precision' , type=_UpperCAmelCase , default=_UpperCAmelCase , choices=['no', 'fp16', 'bf16', 'fp8'] , help='Whether to use mixed precision. Choose'
'between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10.'
'and an Nvidia Ampere GPU.' , )
parser.add_argument('--cpu' , action='store_true' , help='If passed, will train on the CPU.' )
lowerCamelCase__ : Union[str, Any] = parser.parse_args()
lowerCamelCase__ : int = {'lr': 2e-5, 'num_epochs': 3, 'seed': 42, 'batch_size': 16}
training_function(_UpperCAmelCase , _UpperCAmelCase )
if __name__ == "__main__":
main()
| 50
|
from itertools import count
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 50 ) -> int:
lowerCamelCase__ : Optional[Any] = [1] * min_block_length
for n in count(_UpperCAmelCase ):
fill_count_functions.append(1 )
for block_length in range(_UpperCAmelCase , n + 1 ):
for block_start in range(n - block_length ):
fill_count_functions[n] += fill_count_functions[
n - block_start - block_length - 1
]
fill_count_functions[n] += 1
if fill_count_functions[n] > 100_0000:
break
return n
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_UpperCAmelCase : int = logging.get_logger(__name__)
_UpperCAmelCase : int = {"""vocab_file""": """sentencepiece.model"""}
_UpperCAmelCase : Optional[Any] = {
"""vocab_file""": {
"""google/rembert""": """https://huggingface.co/google/rembert/resolve/main/sentencepiece.model""",
},
}
_UpperCAmelCase : Union[str, Any] = {
"""google/rembert""": 2_56,
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self : List[Any] , UpperCAmelCase : Any , UpperCAmelCase : Optional[Any]=False , UpperCAmelCase : Union[str, Any]=True , UpperCAmelCase : List[Any]=True , UpperCAmelCase : Dict="[CLS]" , UpperCAmelCase : Optional[Any]="[SEP]" , UpperCAmelCase : str="[UNK]" , UpperCAmelCase : Tuple="[SEP]" , UpperCAmelCase : Dict="[PAD]" , UpperCAmelCase : str="[CLS]" , UpperCAmelCase : str="[MASK]" , **UpperCAmelCase : str , ) -> Any:
super().__init__(
do_lower_case=UpperCAmelCase , remove_space=UpperCAmelCase , keep_accents=UpperCAmelCase , bos_token=UpperCAmelCase , eos_token=UpperCAmelCase , unk_token=UpperCAmelCase , sep_token=UpperCAmelCase , pad_token=UpperCAmelCase , cls_token=UpperCAmelCase , mask_token=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : Tuple = do_lower_case
lowerCamelCase__ : Tuple = remove_space
lowerCamelCase__ : Any = keep_accents
lowerCamelCase__ : Union[str, Any] = vocab_file
lowerCamelCase__ : str = spm.SentencePieceProcessor()
self.sp_model.Load(UpperCAmelCase )
@property
def A_ ( self : str ) -> int:
return len(self.sp_model )
def A_ ( self : Any ) -> str:
lowerCamelCase__ : Optional[int] = {self.convert_ids_to_tokens(UpperCAmelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self : Tuple ) -> Dict:
lowerCamelCase__ : Union[str, Any] = self.__dict__.copy()
lowerCamelCase__ : Any = None
return state
def __setstate__( self : List[str] , UpperCAmelCase : Any ) -> Tuple:
lowerCamelCase__ : List[Any] = d
lowerCamelCase__ : Optional[Any] = spm.SentencePieceProcessor()
self.sp_model.Load(self.vocab_file )
def A_ ( self : int , UpperCAmelCase : Optional[int] , UpperCAmelCase : Optional[Any]=False ) -> int:
lowerCamelCase__ : Any = self.sp_model.EncodeAsPieces(UpperCAmelCase )
return pieces
def A_ ( self : Optional[int] , UpperCAmelCase : Union[str, Any] ) -> List[Any]:
return self.sp_model.PieceToId(UpperCAmelCase )
def A_ ( self : Union[str, Any] , UpperCAmelCase : Tuple ) -> List[Any]:
return self.sp_model.IdToPiece(UpperCAmelCase )
def A_ ( self : Dict , UpperCAmelCase : Optional[Any] ) -> Optional[Any]:
lowerCamelCase__ : Any = self.sp_model.decode_pieces(UpperCAmelCase )
return out_string
def A_ ( self : int , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : int = [self.sep_token_id]
lowerCamelCase__ : Optional[int] = [self.cls_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 A_ ( self : List[str] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None , UpperCAmelCase : bool = False ) -> List[int]:
if already_has_special_tokens:
if token_ids_a is not None:
raise ValueError(
'You should not supply a second sequence if the provided sequence of '
'ids is already formatted with special tokens for the model.' )
return [1 if x in [self.sep_token_id, self.cls_token_id] else 0 for x in token_ids_a]
if token_ids_a is not None:
return [1] + ([0] * len(UpperCAmelCase )) + [1] + ([0] * len(UpperCAmelCase )) + [1]
return [1] + ([0] * len(UpperCAmelCase )) + [1]
def A_ ( self : List[str] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : Union[str, Any] = [self.sep_token_id]
lowerCamelCase__ : str = [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 A_ ( self : Tuple , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not os.path.isdir(UpperCAmelCase ):
logger.error('Vocabulary path ({}) should be a directory'.format(UpperCAmelCase ) )
return
lowerCamelCase__ : Tuple = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCAmelCase ):
copyfile(self.vocab_file , UpperCAmelCase )
return (out_vocab_file,)
| 50
|
from __future__ import annotations
from typing import Any
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> None:
create_state_space_tree(_UpperCAmelCase , [] , 0 )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> None:
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__":
_UpperCAmelCase : list[Any] = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["""A""", """B""", """C"""])
generate_all_subsequences(seq)
| 50
| 1
|
from manim import *
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Dict ) -> List[str]:
lowerCamelCase__ : List[Any] = Rectangle(height=0.5 , width=0.5 )
lowerCamelCase__ : Optional[Any] = Rectangle(height=0.4_6 , width=0.4_6 ).set_stroke(width=0 )
lowerCamelCase__ : Optional[int] = [mem.copy() for i in range(6 )]
lowerCamelCase__ : Union[str, Any] = [mem.copy() for i in range(6 )]
lowerCamelCase__ : Any = VGroup(*UpperCAmelCase ).arrange(UpperCAmelCase , buff=0 )
lowerCamelCase__ : Any = VGroup(*UpperCAmelCase ).arrange(UpperCAmelCase , buff=0 )
lowerCamelCase__ : int = VGroup(UpperCAmelCase , UpperCAmelCase ).arrange(UpperCAmelCase , buff=0 )
lowerCamelCase__ : List[Any] = Text('CPU' , font_size=24 )
lowerCamelCase__ : Union[str, Any] = Group(UpperCAmelCase , UpperCAmelCase ).arrange(UpperCAmelCase , buff=0.5 , aligned_edge=UpperCAmelCase )
cpu.move_to([-2.5, -0.5, 0] )
self.add(UpperCAmelCase )
lowerCamelCase__ : List[Any] = [mem.copy() for i in range(1 )]
lowerCamelCase__ : str = VGroup(*UpperCAmelCase ).arrange(UpperCAmelCase , buff=0 )
lowerCamelCase__ : int = Text('GPU' , font_size=24 )
lowerCamelCase__ : Optional[Any] = Group(UpperCAmelCase , UpperCAmelCase ).arrange(UpperCAmelCase , buff=0.5 , aligned_edge=UpperCAmelCase )
gpu.align_to(UpperCAmelCase , UpperCAmelCase )
gpu.set_x(gpu.get_x() - 1 )
self.add(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = [mem.copy() for i in range(6 )]
lowerCamelCase__ : Tuple = VGroup(*UpperCAmelCase ).arrange(UpperCAmelCase , buff=0 )
lowerCamelCase__ : Dict = Text('Model' , font_size=24 )
lowerCamelCase__ : Dict = Group(UpperCAmelCase , UpperCAmelCase ).arrange(UpperCAmelCase , buff=0.5 , aligned_edge=UpperCAmelCase )
model.move_to([3, -1.0, 0] )
self.play(
Create(UpperCAmelCase , run_time=1 ) , Create(UpperCAmelCase , run_time=1 ) , Create(UpperCAmelCase , run_time=1 ) , )
lowerCamelCase__ : Dict = MarkupText(
F"""First, an empty model skeleton is loaded\ninto <span fgcolor='{YELLOW}'>memory</span> without using much RAM.""" , font_size=24 , )
lowerCamelCase__ : Any = Square(side_length=2.2 )
key.move_to([-5, 2, 0] )
lowerCamelCase__ : str = MarkupText(
F"""<b>Key:</b>\n\n<span fgcolor='{YELLOW}'>●</span> Empty Model""" , font_size=18 , )
key_text.move_to([-5, 2.4, 0] )
step_a.move_to([2, 2, 0] )
self.play(Write(UpperCAmelCase , run_time=2.5 ) , Write(UpperCAmelCase ) , Write(UpperCAmelCase ) )
self.add(UpperCAmelCase )
lowerCamelCase__ : int = []
lowerCamelCase__ : str = []
lowerCamelCase__ : Dict = []
for i, rect in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Union[str, Any] = Rectangle(height=0.4_6 , width=0.4_6 ).set_stroke(width=0.0 ).set_fill(UpperCAmelCase , opacity=0.7 )
cpu_target.move_to(UpperCAmelCase )
cpu_target.generate_target()
lowerCamelCase__ : Dict = 0.4_6 / 4
lowerCamelCase__ : List[str] = 0.4_6 / 3
if i == 0:
cpu_target.target.next_to(cpu_left_col_base[0].get_corner(DOWN + LEFT ) , buff=0.0_2 , direction=UpperCAmelCase )
cpu_target.target.set_x(cpu_target.target.get_x() + 0.1 )
elif i == 3:
cpu_target.target.next_to(cpu_targs[0].target , direction=UpperCAmelCase , buff=0.0 )
else:
cpu_target.target.next_to(cpu_targs[i - 1].target , direction=UpperCAmelCase , buff=0.0 )
cpu_targs.append(UpperCAmelCase )
first_animations.append(rect.animate(run_time=0.5 ).set_stroke(UpperCAmelCase ) )
second_animations.append(MoveToTarget(UpperCAmelCase , run_time=1.5 ) )
self.play(*UpperCAmelCase )
self.play(*UpperCAmelCase )
self.wait()
| 50
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from tokenizers import processors
from ...tokenization_utils import AddedToken, BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_nllb import NllbTokenizer
else:
_UpperCAmelCase : int = None
_UpperCAmelCase : Dict = logging.get_logger(__name__)
_UpperCAmelCase : Optional[int] = {"""vocab_file""": """sentencepiece.bpe.model""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : List[Any] = {
"""vocab_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/sentencepiece.bpe.model"""
),
},
"""tokenizer_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : List[str] = {
"""facebook/nllb-large-en-ro""": 10_24,
"""facebook/nllb-200-distilled-600M""": 10_24,
}
# fmt: off
_UpperCAmelCase : Optional[int] = ["""ace_Arab""", """ace_Latn""", """acm_Arab""", """acq_Arab""", """aeb_Arab""", """afr_Latn""", """ajp_Arab""", """aka_Latn""", """amh_Ethi""", """apc_Arab""", """arb_Arab""", """ars_Arab""", """ary_Arab""", """arz_Arab""", """asm_Beng""", """ast_Latn""", """awa_Deva""", """ayr_Latn""", """azb_Arab""", """azj_Latn""", """bak_Cyrl""", """bam_Latn""", """ban_Latn""", """bel_Cyrl""", """bem_Latn""", """ben_Beng""", """bho_Deva""", """bjn_Arab""", """bjn_Latn""", """bod_Tibt""", """bos_Latn""", """bug_Latn""", """bul_Cyrl""", """cat_Latn""", """ceb_Latn""", """ces_Latn""", """cjk_Latn""", """ckb_Arab""", """crh_Latn""", """cym_Latn""", """dan_Latn""", """deu_Latn""", """dik_Latn""", """dyu_Latn""", """dzo_Tibt""", """ell_Grek""", """eng_Latn""", """epo_Latn""", """est_Latn""", """eus_Latn""", """ewe_Latn""", """fao_Latn""", """pes_Arab""", """fij_Latn""", """fin_Latn""", """fon_Latn""", """fra_Latn""", """fur_Latn""", """fuv_Latn""", """gla_Latn""", """gle_Latn""", """glg_Latn""", """grn_Latn""", """guj_Gujr""", """hat_Latn""", """hau_Latn""", """heb_Hebr""", """hin_Deva""", """hne_Deva""", """hrv_Latn""", """hun_Latn""", """hye_Armn""", """ibo_Latn""", """ilo_Latn""", """ind_Latn""", """isl_Latn""", """ita_Latn""", """jav_Latn""", """jpn_Jpan""", """kab_Latn""", """kac_Latn""", """kam_Latn""", """kan_Knda""", """kas_Arab""", """kas_Deva""", """kat_Geor""", """knc_Arab""", """knc_Latn""", """kaz_Cyrl""", """kbp_Latn""", """kea_Latn""", """khm_Khmr""", """kik_Latn""", """kin_Latn""", """kir_Cyrl""", """kmb_Latn""", """kon_Latn""", """kor_Hang""", """kmr_Latn""", """lao_Laoo""", """lvs_Latn""", """lij_Latn""", """lim_Latn""", """lin_Latn""", """lit_Latn""", """lmo_Latn""", """ltg_Latn""", """ltz_Latn""", """lua_Latn""", """lug_Latn""", """luo_Latn""", """lus_Latn""", """mag_Deva""", """mai_Deva""", """mal_Mlym""", """mar_Deva""", """min_Latn""", """mkd_Cyrl""", """plt_Latn""", """mlt_Latn""", """mni_Beng""", """khk_Cyrl""", """mos_Latn""", """mri_Latn""", """zsm_Latn""", """mya_Mymr""", """nld_Latn""", """nno_Latn""", """nob_Latn""", """npi_Deva""", """nso_Latn""", """nus_Latn""", """nya_Latn""", """oci_Latn""", """gaz_Latn""", """ory_Orya""", """pag_Latn""", """pan_Guru""", """pap_Latn""", """pol_Latn""", """por_Latn""", """prs_Arab""", """pbt_Arab""", """quy_Latn""", """ron_Latn""", """run_Latn""", """rus_Cyrl""", """sag_Latn""", """san_Deva""", """sat_Beng""", """scn_Latn""", """shn_Mymr""", """sin_Sinh""", """slk_Latn""", """slv_Latn""", """smo_Latn""", """sna_Latn""", """snd_Arab""", """som_Latn""", """sot_Latn""", """spa_Latn""", """als_Latn""", """srd_Latn""", """srp_Cyrl""", """ssw_Latn""", """sun_Latn""", """swe_Latn""", """swh_Latn""", """szl_Latn""", """tam_Taml""", """tat_Cyrl""", """tel_Telu""", """tgk_Cyrl""", """tgl_Latn""", """tha_Thai""", """tir_Ethi""", """taq_Latn""", """taq_Tfng""", """tpi_Latn""", """tsn_Latn""", """tso_Latn""", """tuk_Latn""", """tum_Latn""", """tur_Latn""", """twi_Latn""", """tzm_Tfng""", """uig_Arab""", """ukr_Cyrl""", """umb_Latn""", """urd_Arab""", """uzn_Latn""", """vec_Latn""", """vie_Latn""", """war_Latn""", """wol_Latn""", """xho_Latn""", """ydd_Hebr""", """yor_Latn""", """yue_Hant""", """zho_Hans""", """zho_Hant""", """zul_Latn"""]
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = ["""input_ids""", """attention_mask"""]
UpperCAmelCase__ = NllbTokenizer
UpperCAmelCase__ = []
UpperCAmelCase__ = []
def __init__( self : Tuple , UpperCAmelCase : int=None , UpperCAmelCase : Any=None , UpperCAmelCase : str="<s>" , UpperCAmelCase : Optional[Any]="</s>" , UpperCAmelCase : str="</s>" , UpperCAmelCase : Tuple="<s>" , UpperCAmelCase : Optional[Any]="<unk>" , UpperCAmelCase : List[str]="<pad>" , UpperCAmelCase : Union[str, Any]="<mask>" , UpperCAmelCase : Tuple=None , UpperCAmelCase : int=None , UpperCAmelCase : Dict=None , UpperCAmelCase : Any=False , **UpperCAmelCase : Optional[int] , ) -> Tuple:
# Mask token behave like a normal word, i.e. include the space before it
lowerCamelCase__ : List[Any] = AddedToken(UpperCAmelCase , lstrip=UpperCAmelCase , rstrip=UpperCAmelCase ) if isinstance(UpperCAmelCase , UpperCAmelCase ) else mask_token
lowerCamelCase__ : Union[str, Any] = legacy_behaviour
super().__init__(
vocab_file=UpperCAmelCase , tokenizer_file=UpperCAmelCase , bos_token=UpperCAmelCase , eos_token=UpperCAmelCase , sep_token=UpperCAmelCase , cls_token=UpperCAmelCase , unk_token=UpperCAmelCase , pad_token=UpperCAmelCase , mask_token=UpperCAmelCase , src_lang=UpperCAmelCase , tgt_lang=UpperCAmelCase , additional_special_tokens=UpperCAmelCase , legacy_behaviour=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = vocab_file
lowerCamelCase__ : Dict = False if not self.vocab_file else True
lowerCamelCase__ : Optional[Any] = FAIRSEQ_LANGUAGE_CODES.copy()
if additional_special_tokens is not None:
# Only add those special tokens if they are not already there.
_additional_special_tokens.extend(
[t for t in additional_special_tokens if t not in _additional_special_tokens] )
self.add_special_tokens({'additional_special_tokens': _additional_special_tokens} )
lowerCamelCase__ : str = {
lang_code: self.convert_tokens_to_ids(UpperCAmelCase ) for lang_code in FAIRSEQ_LANGUAGE_CODES
}
lowerCamelCase__ : int = src_lang if src_lang is not None else 'eng_Latn'
lowerCamelCase__ : List[Any] = self.convert_tokens_to_ids(self._src_lang )
lowerCamelCase__ : str = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def A_ ( self : int ) -> str:
return self._src_lang
@src_lang.setter
def A_ ( self : List[Any] , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Any = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
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 A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : Dict = [self.sep_token_id]
lowerCamelCase__ : List[str] = [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 + sep + token_ids_a + sep ) * [0]
def A_ ( self : int , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Optional[str] , UpperCAmelCase : Optional[str] , **UpperCAmelCase : List[str] ) -> Dict:
if src_lang is None or tgt_lang is None:
raise ValueError('Translation requires a `src_lang` and a `tgt_lang` for this model' )
lowerCamelCase__ : Optional[int] = src_lang
lowerCamelCase__ : Optional[int] = self(UpperCAmelCase , add_special_tokens=UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.convert_tokens_to_ids(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tgt_lang_id
return inputs
def A_ ( self : Dict , UpperCAmelCase : List[str] , UpperCAmelCase : str = "eng_Latn" , UpperCAmelCase : Optional[List[str]] = None , UpperCAmelCase : str = "fra_Latn" , **UpperCAmelCase : Dict , ) -> BatchEncoding:
lowerCamelCase__ : Any = src_lang
lowerCamelCase__ : int = tgt_lang
return super().prepare_seqaseq_batch(UpperCAmelCase , UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Union[str, Any] ) -> Optional[int]:
return self.set_src_lang_special_tokens(self.src_lang )
def A_ ( self : Any ) -> Union[str, Any]:
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def A_ ( self : str , UpperCAmelCase : Optional[Any] ) -> None:
lowerCamelCase__ : int = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : int = []
lowerCamelCase__ : str = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : int = [self.cur_lang_code]
lowerCamelCase__ : Tuple = [self.eos_token_id]
lowerCamelCase__ : Any = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : Optional[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : str = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : int , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Union[str, Any] = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : Dict = []
lowerCamelCase__ : Union[str, Any] = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : Any = [self.cur_lang_code]
lowerCamelCase__ : Optional[Any] = [self.eos_token_id]
lowerCamelCase__ : Union[str, Any] = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : List[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : Optional[int] = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : Union[str, Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not self.can_save_slow_tokenizer:
raise ValueError(
'Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '
'tokenizer.' )
if not os.path.isdir(UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory.""" )
return
lowerCamelCase__ : int = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCAmelCase ):
copyfile(self.vocab_file , UpperCAmelCase )
return (out_vocab_file,)
| 50
| 1
|
import csv
import tweepy
# Twitter API credentials
_UpperCAmelCase : Tuple = """"""
_UpperCAmelCase : Union[str, Any] = """"""
_UpperCAmelCase : int = """"""
_UpperCAmelCase : Optional[Any] = """"""
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> None:
# authorize twitter, initialize tweepy
lowerCamelCase__ : List[Any] = tweepy.OAuthHandler(_UpperCAmelCase , _UpperCAmelCase )
auth.set_access_token(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : List[Any] = tweepy.API(_UpperCAmelCase )
# initialize a list to hold all the tweepy Tweets
lowerCamelCase__ : Dict = []
# make initial request for most recent tweets (200 is the maximum allowed count)
lowerCamelCase__ : Tuple = api.user_timeline(screen_name=_UpperCAmelCase , count=200 )
# save most recent tweets
alltweets.extend(_UpperCAmelCase )
# save the id of the oldest tweet less one
lowerCamelCase__ : int = alltweets[-1].id - 1
# keep grabbing tweets until there are no tweets left to grab
while len(_UpperCAmelCase ) > 0:
print(F"""getting tweets before {oldest}""" )
# all subsequent requests use the max_id param to prevent duplicates
lowerCamelCase__ : Optional[Any] = api.user_timeline(
screen_name=_UpperCAmelCase , count=200 , max_id=_UpperCAmelCase )
# save most recent tweets
alltweets.extend(_UpperCAmelCase )
# update the id of the oldest tweet less one
lowerCamelCase__ : List[Any] = alltweets[-1].id - 1
print(F"""...{len(_UpperCAmelCase )} tweets downloaded so far""" )
# transform the tweepy tweets into a 2D array that will populate the csv
lowerCamelCase__ : Optional[Any] = [[tweet.id_str, tweet.created_at, tweet.text] for tweet in alltweets]
# write the csv
with open(F"""new_{screen_name}_tweets.csv""" , 'w' ) as f:
lowerCamelCase__ : Dict = csv.writer(_UpperCAmelCase )
writer.writerow(['id', 'created_at', 'text'] )
writer.writerows(_UpperCAmelCase )
if __name__ == "__main__":
# pass in the username of the account you want to download
get_all_tweets("""FirePing32""")
| 50
|
import argparse
import torch
from torch import nn
from transformers import MaMaaaConfig, MaMaaaForConditionalGeneration
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Optional[int] = [
'encoder.version',
'decoder.version',
'model.encoder.version',
'model.decoder.version',
'decoder.output_projection.weight',
'_float_tensor',
'encoder.embed_positions._float_tensor',
'decoder.embed_positions._float_tensor',
]
for k in ignore_keys:
state_dict.pop(_UpperCAmelCase , _UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[Any]:
lowerCamelCase__ , lowerCamelCase__ : List[str] = emb.weight.shape
lowerCamelCase__ : Tuple = nn.Linear(_UpperCAmelCase , _UpperCAmelCase , bias=_UpperCAmelCase )
lowerCamelCase__ : Dict = emb.weight.data
return lin_layer
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Tuple = torch.load(_UpperCAmelCase , map_location='cpu' )
lowerCamelCase__ : List[str] = mam_aaa['args'] or mam_aaa['cfg']['model']
lowerCamelCase__ : Optional[int] = mam_aaa['model']
remove_ignore_keys_(_UpperCAmelCase )
lowerCamelCase__ : str = state_dict['encoder.embed_tokens.weight'].shape[0]
lowerCamelCase__ : Union[str, Any] = MaMaaaConfig(
vocab_size=_UpperCAmelCase , max_position_embeddings=1024 , encoder_layers=args.encoder_layers , decoder_layers=args.decoder_layers , encoder_attention_heads=args.encoder_attention_heads , decoder_attention_heads=args.decoder_attention_heads , encoder_ffn_dim=args.encoder_ffn_embed_dim , decoder_ffn_dim=args.decoder_ffn_embed_dim , d_model=args.encoder_embed_dim , encoder_layerdrop=args.encoder_layerdrop , decoder_layerdrop=args.decoder_layerdrop , dropout=args.dropout , attention_dropout=args.attention_dropout , activation_dropout=args.activation_dropout , activation_function='relu' , )
lowerCamelCase__ : Optional[Any] = state_dict['decoder.embed_tokens.weight']
lowerCamelCase__ : Union[str, Any] = MaMaaaForConditionalGeneration(_UpperCAmelCase )
model.model.load_state_dict(_UpperCAmelCase , strict=_UpperCAmelCase )
lowerCamelCase__ : List[str] = make_linear_from_emb(model.model.shared )
return model
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""fairseq_path""", type=str, help="""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.""")
_UpperCAmelCase : str = parser.parse_args()
_UpperCAmelCase : Optional[Any] = convert_fairseq_mamaaa_checkpoint_from_disk(args.fairseq_pathß)
model.save_pretrained(args.pytorch_dump_folder_path)
| 50
| 1
|
from __future__ import annotations
from collections import deque
from collections.abc import Iterator
from dataclasses import dataclass
@dataclass
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = 42
class lowerCAmelCase :
def __init__( self : List[str] , UpperCAmelCase : int ) -> Dict:
lowerCamelCase__ : list[list[Edge]] = [[] for _ in range(UpperCAmelCase )]
lowerCamelCase__ : str = size
def __getitem__( self : List[str] , UpperCAmelCase : int ) -> Iterator[Edge]:
return iter(self._graph[vertex] )
@property
def A_ ( self : Dict ) -> List[Any]:
return self._size
def A_ ( self : Tuple , UpperCAmelCase : int , UpperCAmelCase : int , UpperCAmelCase : int ) -> List[Any]:
if weight not in (0, 1):
raise ValueError('Edge weight must be either 0 or 1.' )
if to_vertex < 0 or to_vertex >= self.size:
raise ValueError('Vertex indexes must be in [0; size).' )
self._graph[from_vertex].append(Edge(UpperCAmelCase , UpperCAmelCase ) )
def A_ ( self : Union[str, Any] , UpperCAmelCase : int , UpperCAmelCase : int ) -> int | None:
lowerCamelCase__ : Dict = deque([start_vertex] )
lowerCamelCase__ : list[int | None] = [None] * self.size
lowerCamelCase__ : int = 0
while queue:
lowerCamelCase__ : str = queue.popleft()
lowerCamelCase__ : List[Any] = distances[current_vertex]
if current_distance is None:
continue
for edge in self[current_vertex]:
lowerCamelCase__ : Any = current_distance + edge.weight
lowerCamelCase__ : List[Any] = distances[edge.destination_vertex]
if (
isinstance(UpperCAmelCase , UpperCAmelCase )
and new_distance >= dest_vertex_distance
):
continue
lowerCamelCase__ : List[Any] = new_distance
if edge.weight == 0:
queue.appendleft(edge.destination_vertex )
else:
queue.append(edge.destination_vertex )
if distances[finish_vertex] is None:
raise ValueError('No path from start_vertex to finish_vertex.' )
return distances[finish_vertex]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
|
import argparse
import json
import os
import time
import zipfile
from get_ci_error_statistics import download_artifact, get_artifacts_links
from transformers import logging
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Dict:
lowerCamelCase__ : str = set()
lowerCamelCase__ : Any = []
def parse_line(_UpperCAmelCase ):
for line in fp:
if isinstance(_UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Any = line.decode('UTF-8' )
if "warnings summary (final)" in line:
continue
# This means we are outside the body of a warning
elif not line.startswith(' ' ):
# process a single warning and move it to `selected_warnings`.
if len(_UpperCAmelCase ) > 0:
lowerCamelCase__ : str = '\n'.join(_UpperCAmelCase )
# Only keep the warnings specified in `targets`
if any(F""": {x}: """ in warning for x in targets ):
selected_warnings.add(_UpperCAmelCase )
buffer.clear()
continue
else:
lowerCamelCase__ : List[str] = line.strip()
buffer.append(_UpperCAmelCase )
if from_gh:
for filename in os.listdir(_UpperCAmelCase ):
lowerCamelCase__ : Dict = os.path.join(_UpperCAmelCase , _UpperCAmelCase )
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
else:
try:
with zipfile.ZipFile(_UpperCAmelCase ) as z:
for filename in z.namelist():
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with z.open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
except Exception:
logger.warning(
F"""{artifact_path} is either an invalid zip file or something else wrong. This file is skipped.""" )
return selected_warnings
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = set()
lowerCamelCase__ : Optional[int] = [os.path.join(_UpperCAmelCase , _UpperCAmelCase ) for p in os.listdir(_UpperCAmelCase ) if (p.endswith('.zip' ) or from_gh)]
for p in paths:
selected_warnings.update(extract_warnings_from_single_artifact(_UpperCAmelCase , _UpperCAmelCase ) )
return selected_warnings
if __name__ == "__main__":
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
return values.split(',' )
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""--workflow_run_id""", type=str, required=True, help="""A GitHub Actions workflow run id.""")
parser.add_argument(
"""--output_dir""",
type=str,
required=True,
help="""Where to store the downloaded artifacts and other result files.""",
)
parser.add_argument("""--token""", default=None, type=str, help="""A token that has actions:read permission.""")
# optional parameters
parser.add_argument(
"""--targets""",
default="""DeprecationWarning,UserWarning,FutureWarning""",
type=list_str,
help="""Comma-separated list of target warning(s) which we want to extract.""",
)
parser.add_argument(
"""--from_gh""",
action="""store_true""",
help="""If running from a GitHub action workflow and collecting warnings from its artifacts.""",
)
_UpperCAmelCase : Union[str, Any] = parser.parse_args()
_UpperCAmelCase : Dict = args.from_gh
if from_gh:
# The artifacts have to be downloaded using `actions/download-artifact@v3`
pass
else:
os.makedirs(args.output_dir, exist_ok=True)
# get download links
_UpperCAmelCase : Union[str, Any] = get_artifacts_links(args.workflow_run_id, token=args.token)
with open(os.path.join(args.output_dir, """artifacts.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(artifacts, fp, ensure_ascii=False, indent=4)
# download artifacts
for idx, (name, url) in enumerate(artifacts.items()):
print(name)
print(url)
print("""=""" * 80)
download_artifact(name, url, args.output_dir, args.token)
# Be gentle to GitHub
time.sleep(1)
# extract warnings from artifacts
_UpperCAmelCase : Dict = extract_warnings(args.output_dir, args.targets)
_UpperCAmelCase : Optional[Any] = sorted(selected_warnings)
with open(os.path.join(args.output_dir, """selected_warnings.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(selected_warnings, fp, ensure_ascii=False, indent=4)
| 50
| 1
|
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
_UpperCAmelCase : List[str] = False
_UpperCAmelCase : Dict = False
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> int:
return TrainCommand(_UpperCAmelCase )
class lowerCAmelCase ( __UpperCamelCase ):
@staticmethod
def A_ ( UpperCAmelCase : ArgumentParser ) -> Union[str, Any]:
lowerCamelCase__ : Optional[int] = parser.add_parser('train' , help='CLI tool to train a model on a task.' )
train_parser.add_argument(
'--train_data' , type=UpperCAmelCase , required=UpperCAmelCase , help='path to train (and optionally evaluation) dataset as a csv with tab separated labels and sentences.' , )
train_parser.add_argument(
'--column_label' , type=UpperCAmelCase , default=0 , help='Column of the dataset csv file with example labels.' )
train_parser.add_argument(
'--column_text' , type=UpperCAmelCase , default=1 , help='Column of the dataset csv file with example texts.' )
train_parser.add_argument(
'--column_id' , type=UpperCAmelCase , 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=UpperCAmelCase , default='' , help='path to validation dataset.' )
train_parser.add_argument(
'--validation_split' , type=UpperCAmelCase , 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=UpperCAmelCase , default='./' , help='path to saved the trained model.' )
train_parser.add_argument(
'--task' , type=UpperCAmelCase , default='text_classification' , help='Task to train the model on.' )
train_parser.add_argument(
'--model' , type=UpperCAmelCase , default='bert-base-uncased' , help='Model\'s name or path to stored model.' )
train_parser.add_argument('--train_batch_size' , type=UpperCAmelCase , default=32 , help='Batch size for training.' )
train_parser.add_argument('--valid_batch_size' , type=UpperCAmelCase , default=64 , help='Batch size for validation.' )
train_parser.add_argument('--learning_rate' , type=UpperCAmelCase , default=3e-5 , help='Learning rate.' )
train_parser.add_argument('--adam_epsilon' , type=UpperCAmelCase , default=1e-08 , help='Epsilon for Adam optimizer.' )
train_parser.set_defaults(func=UpperCAmelCase )
def __init__( self : Optional[int] , UpperCAmelCase : Namespace ) -> Dict:
lowerCamelCase__ : List[Any] = logging.get_logger('transformers-cli/training' )
lowerCamelCase__ : Optional[int] = 'tf' if is_tf_available() else 'torch'
os.makedirs(args.output , exist_ok=UpperCAmelCase )
lowerCamelCase__ : str = args.output
lowerCamelCase__ : List[str] = args.column_label
lowerCamelCase__ : int = args.column_text
lowerCamelCase__ : Any = args.column_id
self.logger.info(F"""Loading {args.task} pipeline for {args.model}""" )
if args.task == "text_classification":
lowerCamelCase__ : Optional[int] = 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}""" )
lowerCamelCase__ : Optional[Any] = 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 , )
lowerCamelCase__ : List[str] = None
if args.validation_data:
self.logger.info(F"""Loading validation dataset from {args.validation_data}""" )
lowerCamelCase__ : Union[str, Any] = 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 , )
lowerCamelCase__ : Optional[Any] = args.validation_split
lowerCamelCase__ : Any = args.train_batch_size
lowerCamelCase__ : Any = args.valid_batch_size
lowerCamelCase__ : Dict = args.learning_rate
lowerCamelCase__ : Union[str, Any] = args.adam_epsilon
def A_ ( self : Optional[Any] ) -> Optional[int]:
if self.framework == "tf":
return self.run_tf()
return self.run_torch()
def A_ ( self : int ) -> Optional[Any]:
raise NotImplementedError
def A_ ( self : str ) -> str:
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 )
| 50
|
import flax.linen as nn
import jax
import jax.numpy as jnp
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : Any ) -> Any:
lowerCamelCase__ : str = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : int , UpperCAmelCase : Dict ) -> Optional[Any]:
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : str = hidden_states.shape
lowerCamelCase__ : Union[str, Any] = jax.image.resize(
UpperCAmelCase , shape=(batch, height * 2, width * 2, channels) , method='nearest' , )
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Tuple = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : str , UpperCAmelCase : Union[str, Any] ) -> Optional[Any]:
# pad = ((0, 0), (0, 1), (0, 1), (0, 0)) # pad height and width dim
# hidden_states = jnp.pad(hidden_states, pad_width=pad)
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = 0.0
UpperCAmelCase__ = None
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> Union[str, Any]:
lowerCamelCase__ : Optional[Any] = self.in_channels if self.out_channels is None else self.out_channels
lowerCamelCase__ : Tuple = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : int = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Union[str, Any] = nn.Dense(UpperCAmelCase , dtype=self.dtype )
lowerCamelCase__ : Union[str, Any] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : List[Any] = nn.Dropout(self.dropout_prob )
lowerCamelCase__ : Tuple = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Optional[Any] = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
lowerCamelCase__ : Union[str, Any] = None
if use_nin_shortcut:
lowerCamelCase__ : Dict = nn.Conv(
UpperCAmelCase , kernel_size=(1, 1) , strides=(1, 1) , padding='VALID' , dtype=self.dtype , )
def __call__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : int=True ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = hidden_states
lowerCamelCase__ : List[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[Any] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Any = self.conva(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.time_emb_proj(nn.swish(UpperCAmelCase ) )
lowerCamelCase__ : List[str] = jnp.expand_dims(jnp.expand_dims(UpperCAmelCase , 1 ) , 1 )
lowerCamelCase__ : List[str] = hidden_states + temb
lowerCamelCase__ : Optional[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[str] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = self.dropout(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = self.conva(UpperCAmelCase )
if self.conv_shortcut is not None:
lowerCamelCase__ : Dict = self.conv_shortcut(UpperCAmelCase )
return hidden_states + residual
| 50
| 1
|
import unittest
import numpy as np
from transformers.testing_utils import require_flax, require_tf, require_torch
from transformers.utils import (
expand_dims,
flatten_dict,
is_flax_available,
is_tf_available,
is_torch_available,
reshape,
squeeze,
transpose,
)
if is_flax_available():
import jax.numpy as jnp
if is_tf_available():
import tensorflow as tf
if is_torch_available():
import torch
class lowerCAmelCase ( unittest.TestCase ):
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Union[str, Any] = {
'task_specific_params': {
'summarization': {'length_penalty': 1.0, 'max_length': 128, 'min_length': 12, 'num_beams': 4},
'summarization_cnn': {'length_penalty': 2.0, 'max_length': 142, 'min_length': 56, 'num_beams': 4},
'summarization_xsum': {'length_penalty': 1.0, 'max_length': 62, 'min_length': 11, 'num_beams': 6},
}
}
lowerCamelCase__ : Tuple = {
'task_specific_params.summarization.length_penalty': 1.0,
'task_specific_params.summarization.max_length': 128,
'task_specific_params.summarization.min_length': 12,
'task_specific_params.summarization.num_beams': 4,
'task_specific_params.summarization_cnn.length_penalty': 2.0,
'task_specific_params.summarization_cnn.max_length': 142,
'task_specific_params.summarization_cnn.min_length': 56,
'task_specific_params.summarization_cnn.num_beams': 4,
'task_specific_params.summarization_xsum.length_penalty': 1.0,
'task_specific_params.summarization_xsum.max_length': 62,
'task_specific_params.summarization_xsum.min_length': 11,
'task_specific_params.summarization_xsum.num_beams': 6,
}
self.assertEqual(flatten_dict(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : List[str] ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = np.random.randn(3 , 4 )
self.assertTrue(np.allclose(transpose(UpperCAmelCase ) , x.transpose() ) )
lowerCamelCase__ : str = np.random.randn(3 , 4 , 5 )
self.assertTrue(np.allclose(transpose(UpperCAmelCase , axes=(1, 2, 0) ) , x.transpose((1, 2, 0) ) ) )
@require_torch
def A_ ( self : Any ) -> int:
lowerCamelCase__ : List[Any] = np.random.randn(3 , 4 )
lowerCamelCase__ : Dict = torch.tensor(UpperCAmelCase )
self.assertTrue(np.allclose(transpose(UpperCAmelCase ) , transpose(UpperCAmelCase ).numpy() ) )
lowerCamelCase__ : Union[str, Any] = np.random.randn(3 , 4 , 5 )
lowerCamelCase__ : str = torch.tensor(UpperCAmelCase )
self.assertTrue(np.allclose(transpose(UpperCAmelCase , axes=(1, 2, 0) ) , transpose(UpperCAmelCase , axes=(1, 2, 0) ).numpy() ) )
@require_tf
def A_ ( self : int ) -> Union[str, Any]:
lowerCamelCase__ : int = np.random.randn(3 , 4 )
lowerCamelCase__ : Any = tf.constant(UpperCAmelCase )
self.assertTrue(np.allclose(transpose(UpperCAmelCase ) , transpose(UpperCAmelCase ).numpy() ) )
lowerCamelCase__ : Any = np.random.randn(3 , 4 , 5 )
lowerCamelCase__ : List[str] = tf.constant(UpperCAmelCase )
self.assertTrue(np.allclose(transpose(UpperCAmelCase , axes=(1, 2, 0) ) , transpose(UpperCAmelCase , axes=(1, 2, 0) ).numpy() ) )
@require_flax
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : List[str] = np.random.randn(3 , 4 )
lowerCamelCase__ : Optional[int] = jnp.array(UpperCAmelCase )
self.assertTrue(np.allclose(transpose(UpperCAmelCase ) , np.asarray(transpose(UpperCAmelCase ) ) ) )
lowerCamelCase__ : Any = np.random.randn(3 , 4 , 5 )
lowerCamelCase__ : List[str] = jnp.array(UpperCAmelCase )
self.assertTrue(np.allclose(transpose(UpperCAmelCase , axes=(1, 2, 0) ) , np.asarray(transpose(UpperCAmelCase , axes=(1, 2, 0) ) ) ) )
def A_ ( self : List[str] ) -> Dict:
lowerCamelCase__ : Tuple = np.random.randn(3 , 4 )
self.assertTrue(np.allclose(reshape(UpperCAmelCase , (4, 3) ) , np.reshape(UpperCAmelCase , (4, 3) ) ) )
lowerCamelCase__ : List[Any] = np.random.randn(3 , 4 , 5 )
self.assertTrue(np.allclose(reshape(UpperCAmelCase , (12, 5) ) , np.reshape(UpperCAmelCase , (12, 5) ) ) )
@require_torch
def A_ ( self : List[Any] ) -> int:
lowerCamelCase__ : Dict = np.random.randn(3 , 4 )
lowerCamelCase__ : Dict = torch.tensor(UpperCAmelCase )
self.assertTrue(np.allclose(reshape(UpperCAmelCase , (4, 3) ) , reshape(UpperCAmelCase , (4, 3) ).numpy() ) )
lowerCamelCase__ : Union[str, Any] = np.random.randn(3 , 4 , 5 )
lowerCamelCase__ : Union[str, Any] = torch.tensor(UpperCAmelCase )
self.assertTrue(np.allclose(reshape(UpperCAmelCase , (12, 5) ) , reshape(UpperCAmelCase , (12, 5) ).numpy() ) )
@require_tf
def A_ ( self : Tuple ) -> List[Any]:
lowerCamelCase__ : int = np.random.randn(3 , 4 )
lowerCamelCase__ : Dict = tf.constant(UpperCAmelCase )
self.assertTrue(np.allclose(reshape(UpperCAmelCase , (4, 3) ) , reshape(UpperCAmelCase , (4, 3) ).numpy() ) )
lowerCamelCase__ : List[str] = np.random.randn(3 , 4 , 5 )
lowerCamelCase__ : Dict = tf.constant(UpperCAmelCase )
self.assertTrue(np.allclose(reshape(UpperCAmelCase , (12, 5) ) , reshape(UpperCAmelCase , (12, 5) ).numpy() ) )
@require_flax
def A_ ( self : Optional[int] ) -> str:
lowerCamelCase__ : Optional[Any] = np.random.randn(3 , 4 )
lowerCamelCase__ : int = jnp.array(UpperCAmelCase )
self.assertTrue(np.allclose(reshape(UpperCAmelCase , (4, 3) ) , np.asarray(reshape(UpperCAmelCase , (4, 3) ) ) ) )
lowerCamelCase__ : str = np.random.randn(3 , 4 , 5 )
lowerCamelCase__ : Any = jnp.array(UpperCAmelCase )
self.assertTrue(np.allclose(reshape(UpperCAmelCase , (12, 5) ) , np.asarray(reshape(UpperCAmelCase , (12, 5) ) ) ) )
def A_ ( self : str ) -> List[str]:
lowerCamelCase__ : Union[str, Any] = np.random.randn(1 , 3 , 4 )
self.assertTrue(np.allclose(squeeze(UpperCAmelCase ) , np.squeeze(UpperCAmelCase ) ) )
lowerCamelCase__ : Any = np.random.randn(1 , 4 , 1 , 5 )
self.assertTrue(np.allclose(squeeze(UpperCAmelCase , axis=2 ) , np.squeeze(UpperCAmelCase , axis=2 ) ) )
@require_torch
def A_ ( self : List[str] ) -> List[str]:
lowerCamelCase__ : Tuple = np.random.randn(1 , 3 , 4 )
lowerCamelCase__ : Union[str, Any] = torch.tensor(UpperCAmelCase )
self.assertTrue(np.allclose(squeeze(UpperCAmelCase ) , squeeze(UpperCAmelCase ).numpy() ) )
lowerCamelCase__ : str = np.random.randn(1 , 4 , 1 , 5 )
lowerCamelCase__ : Optional[Any] = torch.tensor(UpperCAmelCase )
self.assertTrue(np.allclose(squeeze(UpperCAmelCase , axis=2 ) , squeeze(UpperCAmelCase , axis=2 ).numpy() ) )
@require_tf
def A_ ( self : Optional[int] ) -> List[str]:
lowerCamelCase__ : str = np.random.randn(1 , 3 , 4 )
lowerCamelCase__ : Optional[int] = tf.constant(UpperCAmelCase )
self.assertTrue(np.allclose(squeeze(UpperCAmelCase ) , squeeze(UpperCAmelCase ).numpy() ) )
lowerCamelCase__ : List[str] = np.random.randn(1 , 4 , 1 , 5 )
lowerCamelCase__ : Optional[Any] = tf.constant(UpperCAmelCase )
self.assertTrue(np.allclose(squeeze(UpperCAmelCase , axis=2 ) , squeeze(UpperCAmelCase , axis=2 ).numpy() ) )
@require_flax
def A_ ( self : Optional[Any] ) -> List[Any]:
lowerCamelCase__ : List[str] = np.random.randn(1 , 3 , 4 )
lowerCamelCase__ : Optional[Any] = jnp.array(UpperCAmelCase )
self.assertTrue(np.allclose(squeeze(UpperCAmelCase ) , np.asarray(squeeze(UpperCAmelCase ) ) ) )
lowerCamelCase__ : Optional[Any] = np.random.randn(1 , 4 , 1 , 5 )
lowerCamelCase__ : Tuple = jnp.array(UpperCAmelCase )
self.assertTrue(np.allclose(squeeze(UpperCAmelCase , axis=2 ) , np.asarray(squeeze(UpperCAmelCase , axis=2 ) ) ) )
def A_ ( self : Dict ) -> Optional[Any]:
lowerCamelCase__ : Any = np.random.randn(3 , 4 )
self.assertTrue(np.allclose(expand_dims(UpperCAmelCase , axis=1 ) , np.expand_dims(UpperCAmelCase , axis=1 ) ) )
@require_torch
def A_ ( self : Optional[Any] ) -> List[Any]:
lowerCamelCase__ : Any = np.random.randn(3 , 4 )
lowerCamelCase__ : Optional[int] = torch.tensor(UpperCAmelCase )
self.assertTrue(np.allclose(expand_dims(UpperCAmelCase , axis=1 ) , expand_dims(UpperCAmelCase , axis=1 ).numpy() ) )
@require_tf
def A_ ( self : int ) -> Optional[int]:
lowerCamelCase__ : Optional[int] = np.random.randn(3 , 4 )
lowerCamelCase__ : Any = tf.constant(UpperCAmelCase )
self.assertTrue(np.allclose(expand_dims(UpperCAmelCase , axis=1 ) , expand_dims(UpperCAmelCase , axis=1 ).numpy() ) )
@require_flax
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Optional[Any] = np.random.randn(3 , 4 )
lowerCamelCase__ : Optional[Any] = jnp.array(UpperCAmelCase )
self.assertTrue(np.allclose(expand_dims(UpperCAmelCase , axis=1 ) , np.asarray(expand_dims(UpperCAmelCase , axis=1 ) ) ) )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Optional[Any] = set()
# edges = list of graph's edges
lowerCamelCase__ : List[str] = get_edges(_UpperCAmelCase )
# While there are still elements in edges list, take an arbitrary edge
# (from_node, to_node) and add his extremity to chosen_vertices and then
# remove all arcs adjacent to the from_node and to_node
while edges:
lowerCamelCase__ , lowerCamelCase__ : str = edges.pop()
chosen_vertices.add(_UpperCAmelCase )
chosen_vertices.add(_UpperCAmelCase )
for edge in edges.copy():
if from_node in edge or to_node in edge:
edges.discard(_UpperCAmelCase )
return chosen_vertices
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> set:
lowerCamelCase__ : Union[str, Any] = set()
for from_node, to_nodes in graph.items():
for to_node in to_nodes:
edges.add((from_node, to_node) )
return edges
if __name__ == "__main__":
import doctest
doctest.testmod()
# graph = {0: [1, 3], 1: [0, 3], 2: [0, 3, 4], 3: [0, 1, 2], 4: [2, 3]}
# print(f"Matching vertex cover:\n{matching_min_vertex_cover(graph)}")
| 50
| 1
|
import argparse
import evaluate
import torch
from datasets import load_dataset
from torch.optim import AdamW
from torch.utils.data import DataLoader
from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed
from accelerate import Accelerator, DistributedType
########################################################################
# This is a fully working simple example to use Accelerate
#
# This example trains a Bert base model on GLUE MRPC
# in any of the following settings (with the same script):
# - single CPU or single GPU
# - multi GPUS (using PyTorch distributed mode)
# - (multi) TPUs
# - fp16 (mixed-precision) or fp32 (normal precision)
#
# To run it in each of these various modes, follow the instructions
# in the readme for examples:
# https://github.com/huggingface/accelerate/tree/main/examples
#
########################################################################
_UpperCAmelCase : str = 16
_UpperCAmelCase : str = 32
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase = 16 ) -> str:
lowerCamelCase__ : int = AutoTokenizer.from_pretrained('bert-base-cased' )
lowerCamelCase__ : Union[str, Any] = load_dataset('glue' , 'mrpc' )
def tokenize_function(_UpperCAmelCase ):
# max_length=None => use the model max length (it's actually the default)
lowerCamelCase__ : int = tokenizer(examples['sentence1'] , examples['sentence2'] , truncation=_UpperCAmelCase , max_length=_UpperCAmelCase )
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
# starting with the main process first:
with accelerator.main_process_first():
lowerCamelCase__ : Any = datasets.map(
_UpperCAmelCase , batched=_UpperCAmelCase , remove_columns=['idx', 'sentence1', 'sentence2'] , )
# We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the
# transformers library
lowerCamelCase__ : Dict = tokenized_datasets.rename_column('label' , 'labels' )
def collate_fn(_UpperCAmelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
lowerCamelCase__ : Dict = 128 if accelerator.distributed_type == DistributedType.TPU else None
# When using mixed precision we want round multiples of 8/16
if accelerator.mixed_precision == "fp8":
lowerCamelCase__ : List[Any] = 16
elif accelerator.mixed_precision != "no":
lowerCamelCase__ : Dict = 8
else:
lowerCamelCase__ : Any = None
return tokenizer.pad(
_UpperCAmelCase , padding='longest' , max_length=_UpperCAmelCase , pad_to_multiple_of=_UpperCAmelCase , return_tensors='pt' , )
# Instantiate dataloaders.
lowerCamelCase__ : List[str] = DataLoader(
tokenized_datasets['train'] , shuffle=_UpperCAmelCase , collate_fn=_UpperCAmelCase , batch_size=_UpperCAmelCase , drop_last=_UpperCAmelCase )
lowerCamelCase__ : str = DataLoader(
tokenized_datasets['validation'] , shuffle=_UpperCAmelCase , collate_fn=_UpperCAmelCase , batch_size=_UpperCAmelCase , drop_last=(accelerator.mixed_precision == 'fp8') , )
return train_dataloader, eval_dataloader
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Optional[Any]:
# Initialize accelerator
lowerCamelCase__ : Union[str, Any] = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision )
# Sample hyper-parameters for learning rate, batch size, seed and a few other HPs
lowerCamelCase__ : int = config['lr']
lowerCamelCase__ : Tuple = int(config['num_epochs'] )
lowerCamelCase__ : Union[str, Any] = int(config['seed'] )
lowerCamelCase__ : List[str] = int(config['batch_size'] )
lowerCamelCase__ : List[str] = evaluate.load('glue' , 'mrpc' )
# If the batch size is too big we use gradient accumulation
lowerCamelCase__ : Dict = 1
if batch_size > MAX_GPU_BATCH_SIZE and accelerator.distributed_type != DistributedType.TPU:
lowerCamelCase__ : Optional[Any] = batch_size // MAX_GPU_BATCH_SIZE
lowerCamelCase__ : Any = MAX_GPU_BATCH_SIZE
set_seed(_UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Dict = get_dataloaders(_UpperCAmelCase , _UpperCAmelCase )
# Instantiate the model (we build the model here so that the seed also control new weights initialization)
lowerCamelCase__ : Any = AutoModelForSequenceClassification.from_pretrained('bert-base-cased' , return_dict=_UpperCAmelCase )
# We could avoid this line since the accelerator is set with `device_placement=True` (default value).
# Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer
# creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that).
lowerCamelCase__ : Dict = model.to(accelerator.device )
# Instantiate optimizer
lowerCamelCase__ : int = AdamW(params=model.parameters() , lr=_UpperCAmelCase )
# Instantiate scheduler
lowerCamelCase__ : Dict = get_linear_schedule_with_warmup(
optimizer=_UpperCAmelCase , num_warmup_steps=100 , num_training_steps=(len(_UpperCAmelCase ) * num_epochs) // gradient_accumulation_steps , )
# Prepare everything
# There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the
# prepare method.
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : int = accelerator.prepare(
_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
# Now we train the model
for epoch in range(_UpperCAmelCase ):
model.train()
for step, batch in enumerate(_UpperCAmelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
lowerCamelCase__ : Optional[int] = model(**_UpperCAmelCase )
lowerCamelCase__ : Any = outputs.loss
lowerCamelCase__ : Optional[Any] = loss / gradient_accumulation_steps
accelerator.backward(_UpperCAmelCase )
if step % gradient_accumulation_steps == 0:
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
model.eval()
for step, batch in enumerate(_UpperCAmelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
with torch.no_grad():
lowerCamelCase__ : List[Any] = model(**_UpperCAmelCase )
lowerCamelCase__ : Any = outputs.logits.argmax(dim=-1 )
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = accelerator.gather_for_metrics((predictions, batch['labels']) )
metric.add_batch(
predictions=_UpperCAmelCase , references=_UpperCAmelCase , )
lowerCamelCase__ : List[Any] = metric.compute()
# Use accelerator.print to print only on the main process.
accelerator.print(F"""epoch {epoch}:""" , _UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( ) -> Optional[Any]:
lowerCamelCase__ : str = argparse.ArgumentParser(description='Simple example of training script.' )
parser.add_argument(
'--mixed_precision' , type=_UpperCAmelCase , default=_UpperCAmelCase , choices=['no', 'fp16', 'bf16', 'fp8'] , help='Whether to use mixed precision. Choose'
'between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10.'
'and an Nvidia Ampere GPU.' , )
parser.add_argument('--cpu' , action='store_true' , help='If passed, will train on the CPU.' )
lowerCamelCase__ : Tuple = parser.parse_args()
lowerCamelCase__ : Dict = {'lr': 2e-5, 'num_epochs': 3, 'seed': 42, 'batch_size': 16}
training_function(_UpperCAmelCase , _UpperCAmelCase )
if __name__ == "__main__":
main()
| 50
|
from __future__ import annotations
import math
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
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(_UpperCAmelCase ) + 1 ) , 6 ):
if number % i == 0 or number % (i + 2) == 0:
return False
return True
_UpperCAmelCase : Any = [num for num in range(3, 10_00_01, 2) if not is_prime(num)]
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list[int]:
if not isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise ValueError('n must be an integer' )
if n <= 0:
raise ValueError('n must be >= 0' )
lowerCamelCase__ : int = []
for num in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : Union[str, Any] = 0
while 2 * i * i <= odd_composites[num]:
lowerCamelCase__ : Dict = odd_composites[num] - 2 * i * i
if is_prime(_UpperCAmelCase ):
break
i += 1
else:
list_nums.append(odd_composites[num] )
if len(_UpperCAmelCase ) == n:
return list_nums
return []
def SCREAMING_SNAKE_CASE ( ) -> int:
return compute_nums(1 )[0]
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
_UpperCAmelCase : Dict = """
# Transformers 설치 방법
! pip install transformers datasets
# 마지막 릴리스 대신 소스에서 설치하려면, 위 명령을 주석으로 바꾸고 아래 명령을 해제하세요.
# ! pip install git+https://github.com/huggingface/transformers.git
"""
_UpperCAmelCase : Optional[int] = [{"""type""": """code""", """content""": INSTALL_CONTENT}]
_UpperCAmelCase : List[Any] = {
"""{processor_class}""": """FakeProcessorClass""",
"""{model_class}""": """FakeModelClass""",
"""{object_class}""": """FakeObjectClass""",
}
| 50
|
import argparse
from collections import OrderedDict
from pathlib import Path
import requests
import torch
from PIL import Image
from transformers import GLPNConfig, GLPNForDepthEstimation, GLPNImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
_UpperCAmelCase : Dict = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[str]:
lowerCamelCase__ : str = OrderedDict()
for key, value in state_dict.items():
if key.startswith('module.encoder' ):
lowerCamelCase__ : Optional[Any] = key.replace('module.encoder' , 'glpn.encoder' )
if key.startswith('module.decoder' ):
lowerCamelCase__ : List[str] = key.replace('module.decoder' , 'decoder.stages' )
if "patch_embed" in key:
# replace for example patch_embed1 by patch_embeddings.0
lowerCamelCase__ : Dict = key[key.find('patch_embed' ) + len('patch_embed' )]
lowerCamelCase__ : Tuple = key.replace(F"""patch_embed{idx}""" , F"""patch_embeddings.{int(_UpperCAmelCase )-1}""" )
if "norm" in key:
lowerCamelCase__ : str = key.replace('norm' , 'layer_norm' )
if "glpn.encoder.layer_norm" in key:
# replace for example layer_norm1 by layer_norm.0
lowerCamelCase__ : Dict = key[key.find('glpn.encoder.layer_norm' ) + len('glpn.encoder.layer_norm' )]
lowerCamelCase__ : str = key.replace(F"""layer_norm{idx}""" , F"""layer_norm.{int(_UpperCAmelCase )-1}""" )
if "layer_norm1" in key:
lowerCamelCase__ : Optional[int] = key.replace('layer_norm1' , 'layer_norm_1' )
if "layer_norm2" in key:
lowerCamelCase__ : Optional[int] = key.replace('layer_norm2' , 'layer_norm_2' )
if "block" in key:
# replace for example block1 by block.0
lowerCamelCase__ : List[Any] = key[key.find('block' ) + len('block' )]
lowerCamelCase__ : int = key.replace(F"""block{idx}""" , F"""block.{int(_UpperCAmelCase )-1}""" )
if "attn.q" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('attn.q' , 'attention.self.query' )
if "attn.proj" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('attn.proj' , 'attention.output.dense' )
if "attn" in key:
lowerCamelCase__ : Dict = key.replace('attn' , 'attention.self' )
if "fc1" in key:
lowerCamelCase__ : Dict = key.replace('fc1' , 'dense1' )
if "fc2" in key:
lowerCamelCase__ : Any = key.replace('fc2' , 'dense2' )
if "linear_pred" in key:
lowerCamelCase__ : Dict = key.replace('linear_pred' , 'classifier' )
if "linear_fuse" in key:
lowerCamelCase__ : Tuple = key.replace('linear_fuse.conv' , 'linear_fuse' )
lowerCamelCase__ : List[str] = key.replace('linear_fuse.bn' , 'batch_norm' )
if "linear_c" in key:
# replace for example linear_c4 by linear_c.3
lowerCamelCase__ : Optional[Any] = key[key.find('linear_c' ) + len('linear_c' )]
lowerCamelCase__ : Dict = key.replace(F"""linear_c{idx}""" , F"""linear_c.{int(_UpperCAmelCase )-1}""" )
if "bot_conv" in key:
lowerCamelCase__ : str = key.replace('bot_conv' , '0.convolution' )
if "skip_conv1" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('skip_conv1' , '1.convolution' )
if "skip_conv2" in key:
lowerCamelCase__ : List[Any] = key.replace('skip_conv2' , '2.convolution' )
if "fusion1" in key:
lowerCamelCase__ : Optional[int] = key.replace('fusion1' , '1.fusion' )
if "fusion2" in key:
lowerCamelCase__ : Union[str, Any] = key.replace('fusion2' , '2.fusion' )
if "fusion3" in key:
lowerCamelCase__ : List[Any] = key.replace('fusion3' , '3.fusion' )
if "fusion" in key and "conv" in key:
lowerCamelCase__ : str = key.replace('conv' , 'convolutional_layer' )
if key.startswith('module.last_layer_depth' ):
lowerCamelCase__ : Dict = key.replace('module.last_layer_depth' , 'head.head' )
lowerCamelCase__ : str = value
return new_state_dict
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
# for each of the encoder blocks:
for i in range(config.num_encoder_blocks ):
for j in range(config.depths[i] ):
# read in weights + bias of keys and values (which is a single matrix in the original implementation)
lowerCamelCase__ : Any = state_dict.pop(F"""glpn.encoder.block.{i}.{j}.attention.self.kv.weight""" )
lowerCamelCase__ : Optional[Any] = state_dict.pop(F"""glpn.encoder.block.{i}.{j}.attention.self.kv.bias""" )
# next, add keys and values (in that order) to the state dict
lowerCamelCase__ : Optional[int] = kv_weight[
: config.hidden_sizes[i], :
]
lowerCamelCase__ : Optional[int] = kv_bias[: config.hidden_sizes[i]]
lowerCamelCase__ : Any = kv_weight[
config.hidden_sizes[i] :, :
]
lowerCamelCase__ : Dict = kv_bias[config.hidden_sizes[i] :]
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : List[str] = 'http://images.cocodataset.org/val2017/000000039769.jpg'
lowerCamelCase__ : Tuple = Image.open(requests.get(_UpperCAmelCase , stream=_UpperCAmelCase ).raw )
return image
@torch.no_grad()
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=False , _UpperCAmelCase=None ) -> Optional[int]:
lowerCamelCase__ : str = GLPNConfig(hidden_sizes=[64, 128, 320, 512] , decoder_hidden_size=64 , depths=[3, 8, 27, 3] )
# load image processor (only resize + rescale)
lowerCamelCase__ : Union[str, Any] = GLPNImageProcessor()
# prepare image
lowerCamelCase__ : str = prepare_img()
lowerCamelCase__ : Tuple = image_processor(images=_UpperCAmelCase , return_tensors='pt' ).pixel_values
logger.info('Converting model...' )
# load original state dict
lowerCamelCase__ : Any = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
# rename keys
lowerCamelCase__ : str = rename_keys(_UpperCAmelCase )
# key and value matrices need special treatment
read_in_k_v(_UpperCAmelCase , _UpperCAmelCase )
# create HuggingFace model and load state dict
lowerCamelCase__ : Dict = GLPNForDepthEstimation(_UpperCAmelCase )
model.load_state_dict(_UpperCAmelCase )
model.eval()
# forward pass
lowerCamelCase__ : List[str] = model(_UpperCAmelCase )
lowerCamelCase__ : Tuple = outputs.predicted_depth
# verify output
if model_name is not None:
if "nyu" in model_name:
lowerCamelCase__ : List[Any] = torch.tensor(
[[4.4_147, 4.0_873, 4.0_673], [3.7_890, 3.2_881, 3.1_525], [3.7_674, 3.5_423, 3.4_913]] )
elif "kitti" in model_name:
lowerCamelCase__ : List[str] = torch.tensor(
[[3.4_291, 2.7_865, 2.5_151], [3.2_841, 2.7_021, 2.3_502], [3.1_147, 2.4_625, 2.2_481]] )
else:
raise ValueError(F"""Unknown model name: {model_name}""" )
lowerCamelCase__ : Tuple = torch.Size([1, 480, 640] )
assert predicted_depth.shape == expected_shape
assert torch.allclose(predicted_depth[0, :3, :3] , _UpperCAmelCase , atol=1e-4 )
print('Looks ok!' )
# finally, push to hub if required
if push_to_hub:
logger.info('Pushing model and image processor to the hub...' )
model.push_to_hub(
repo_path_or_name=Path(_UpperCAmelCase , _UpperCAmelCase ) , organization='nielsr' , commit_message='Add model' , use_temp_dir=_UpperCAmelCase , )
image_processor.push_to_hub(
repo_path_or_name=Path(_UpperCAmelCase , _UpperCAmelCase ) , organization='nielsr' , commit_message='Add image processor' , use_temp_dir=_UpperCAmelCase , )
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
parser.add_argument(
"""--checkpoint_path""",
default=None,
type=str,
help="""Path to the original PyTorch checkpoint (.pth file).""",
)
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 to upload the model to the HuggingFace hub."""
)
parser.add_argument(
"""--model_name""",
default="""glpn-kitti""",
type=str,
help="""Name of the model in case you're pushing to the hub.""",
)
_UpperCAmelCase : int = parser.parse_args()
convert_glpn_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 50
| 1
|
import json
import os
from typing import Optional, Tuple
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_UpperCAmelCase : List[str] = logging.get_logger(__name__)
_UpperCAmelCase : Tuple = {"""vocab_file""": """vocab.json"""}
_UpperCAmelCase : List[str] = {
"""vocab_file""": {
"""mgp-str""": """https://huggingface.co/alibaba-damo/mgp-str-base/blob/main/vocab.json""",
}
}
_UpperCAmelCase : str = {"""mgp-str""": 27}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self : List[Any] , UpperCAmelCase : List[Any] , UpperCAmelCase : Optional[int]="[GO]" , UpperCAmelCase : int="[GO]" , UpperCAmelCase : List[Any]="[s]" , UpperCAmelCase : str="[GO]" , **UpperCAmelCase : str ) -> List[str]:
super().__init__(
unk_token=UpperCAmelCase , bos_token=UpperCAmelCase , eos_token=UpperCAmelCase , pad_token=UpperCAmelCase , **UpperCAmelCase , )
with open(UpperCAmelCase , encoding='utf-8' ) as vocab_handle:
lowerCamelCase__ : List[Any] = json.load(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = {v: k for k, v in self.vocab.items()}
@property
def A_ ( self : List[str] ) -> Tuple:
return len(self.vocab )
def A_ ( self : Union[str, Any] ) -> Optional[int]:
return dict(self.vocab , **self.added_tokens_encoder )
def A_ ( self : Tuple , UpperCAmelCase : Union[str, Any] ) -> Dict:
lowerCamelCase__ : Dict = []
for s in text:
char_tokens.extend(UpperCAmelCase )
return char_tokens
def A_ ( self : List[Any] , UpperCAmelCase : Union[str, Any] ) -> Union[str, Any]:
return self.vocab.get(UpperCAmelCase , self.vocab.get(self.unk_token ) )
def A_ ( self : str , UpperCAmelCase : Dict ) -> Tuple:
return self.decoder.get(UpperCAmelCase )
def A_ ( self : int , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not os.path.isdir(UpperCAmelCase ):
logger.error('Vocabulary path ({}) should be a directory'.format(UpperCAmelCase ) )
return
lowerCamelCase__ : str = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
with open(UpperCAmelCase , 'w' , encoding='utf-8' ) as f:
f.write(json.dumps(self.vocab , indent=2 , sort_keys=UpperCAmelCase , ensure_ascii=UpperCAmelCase ) + '\n' )
return (vocab_file,)
| 50
|
import tempfile
import unittest
from transformers import TaConfig, is_torch_available
from transformers.testing_utils import (
require_sentencepiece,
require_tokenizers,
require_torch,
slow,
torch_device,
)
from ...generation.test_utils import GenerationTesterMixin
from ...test_modeling_common import ModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import AutoTokenizer, UMTaForConditionalGeneration, UMTaForQuestionAnswering, UMTaModel
class lowerCAmelCase :
def __init__( self : Tuple , UpperCAmelCase : Dict , UpperCAmelCase : Union[str, Any]=99 , UpperCAmelCase : str=13 , UpperCAmelCase : List[str]=7 , UpperCAmelCase : str=9 , UpperCAmelCase : Optional[int]=True , UpperCAmelCase : str=True , UpperCAmelCase : Any=False , UpperCAmelCase : Union[str, Any]=32 , UpperCAmelCase : List[str]=5 , UpperCAmelCase : Tuple=4 , UpperCAmelCase : Union[str, Any]=37 , UpperCAmelCase : int=8 , UpperCAmelCase : List[str]=0.1 , UpperCAmelCase : Any=0.0_0_2 , UpperCAmelCase : Optional[Any]=1 , UpperCAmelCase : List[Any]=0 , UpperCAmelCase : Union[str, Any]=0 , UpperCAmelCase : Tuple=None , UpperCAmelCase : Optional[Any]=None , ) -> Union[str, Any]:
lowerCamelCase__ : int = parent
lowerCamelCase__ : Any = batch_size
lowerCamelCase__ : Optional[int] = encoder_seq_length
lowerCamelCase__ : int = decoder_seq_length
# For common tests
lowerCamelCase__ : List[str] = self.decoder_seq_length
lowerCamelCase__ : Optional[int] = is_training
lowerCamelCase__ : List[Any] = use_attention_mask
lowerCamelCase__ : Optional[Any] = use_labels
lowerCamelCase__ : Union[str, Any] = vocab_size
lowerCamelCase__ : Union[str, Any] = hidden_size
lowerCamelCase__ : Optional[Any] = num_hidden_layers
lowerCamelCase__ : Any = num_attention_heads
lowerCamelCase__ : str = d_ff
lowerCamelCase__ : Optional[Any] = relative_attention_num_buckets
lowerCamelCase__ : Any = dropout_rate
lowerCamelCase__ : Any = initializer_factor
lowerCamelCase__ : Union[str, Any] = eos_token_id
lowerCamelCase__ : List[str] = pad_token_id
lowerCamelCase__ : List[str] = decoder_start_token_id
lowerCamelCase__ : List[Any] = None
lowerCamelCase__ : Optional[Any] = decoder_layers
def A_ ( self : List[Any] ) -> int:
return TaConfig.from_pretrained('google/umt5-base' )
def A_ ( self : List[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[int] , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : Tuple=None , UpperCAmelCase : List[str]=None , UpperCAmelCase : str=None , UpperCAmelCase : Optional[Any]=None , UpperCAmelCase : Optional[Any]=None , ) -> List[str]:
if attention_mask is None:
lowerCamelCase__ : Optional[Any] = input_ids.ne(config.pad_token_id )
if decoder_attention_mask is None:
lowerCamelCase__ : Optional[Any] = decoder_input_ids.ne(config.pad_token_id )
if head_mask is None:
lowerCamelCase__ : int = torch.ones(config.num_hidden_layers , config.num_attention_heads , device=UpperCAmelCase )
if decoder_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
if cross_attn_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(
config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
return {
"input_ids": input_ids,
"decoder_input_ids": decoder_input_ids,
"attention_mask": attention_mask,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
"cross_attn_head_mask": cross_attn_head_mask,
}
def A_ ( self : str ) -> List[str]:
lowerCamelCase__ : Any = ids_tensor([self.batch_size, self.encoder_seq_length] , self.vocab_size )
lowerCamelCase__ : Optional[Any] = ids_tensor([self.batch_size, self.decoder_seq_length] , self.vocab_size )
# we need to clamp the input ids here to avoid having pad token in between
# this is because for NllbMoe the position_ids are prepared such that
# all pad tokens have pos id = 2 and rest are between 2..seq_length
# and the seq_length here is seq_length - num_pad_tokens
# but when using past, there is no way of knowing if the past input ids had
# pad tokens in them, which results in incorrect seq_lenth and which in turn results in
# position_ids being off by num_pad_tokens in past input
lowerCamelCase__ : List[str] = input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Union[str, Any] = decoder_input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Dict = self.get_config()
lowerCamelCase__ : Tuple = config.num_attention_heads
lowerCamelCase__ : Any = self.prepare_inputs_dict(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
return config, input_dict
def A_ ( self : Tuple ) -> Union[str, Any]:
lowerCamelCase__ , lowerCamelCase__ : Dict = self.prepare_config_and_inputs()
return config, inputs_dict
def A_ ( self : Optional[int] ) -> List[str]:
return TaConfig(
vocab_size=166 , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Union[str, Any] ) -> Dict:
return TaConfig(
vocab_size=self.vocab_size , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Any , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[str] , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Any , UpperCAmelCase : Dict , ) -> str:
lowerCamelCase__ : Dict = UMTaModel(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Optional[int] = model(
input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase , decoder_attention_mask=UpperCAmelCase , )
lowerCamelCase__ : Any = model(input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase )
lowerCamelCase__ : Dict = result.last_hidden_state
lowerCamelCase__ : Any = result.past_key_values
lowerCamelCase__ : List[Any] = result.encoder_last_hidden_state
self.parent.assertEqual(encoder_output.size() , (self.batch_size, self.encoder_seq_length, self.hidden_size) )
self.parent.assertEqual(decoder_output.size() , (self.batch_size, self.decoder_seq_length, self.hidden_size) )
# There should be `num_layers` key value embeddings stored in decoder_past
self.parent.assertEqual(len(UpperCAmelCase ) , config.num_layers )
# There should be a self attn key, a self attn value, a cross attn key and a cross attn value stored in each decoder_past tuple
self.parent.assertEqual(len(decoder_past[0] ) , 4 )
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[Any] , UpperCAmelCase : int , UpperCAmelCase : List[str] , UpperCAmelCase : Optional[Any] , ) -> Optional[int]:
lowerCamelCase__ : List[Any] = UMTaModel(config=UpperCAmelCase ).get_decoder().to(UpperCAmelCase ).eval()
# first forward pass
lowerCamelCase__ : Tuple = model(UpperCAmelCase , use_cache=UpperCAmelCase )
lowerCamelCase__ : List[Any] = model(UpperCAmelCase )
lowerCamelCase__ : int = model(UpperCAmelCase , use_cache=UpperCAmelCase )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) + 1 )
lowerCamelCase__ , lowerCamelCase__ : Dict = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
lowerCamelCase__ : Optional[int] = ids_tensor((self.batch_size, 1) , config.vocab_size )
# append to next input_ids and
lowerCamelCase__ : List[Any] = torch.cat([input_ids, next_tokens] , dim=-1 )
lowerCamelCase__ : List[str] = model(UpperCAmelCase )['last_hidden_state']
lowerCamelCase__ : str = model(UpperCAmelCase , past_key_values=UpperCAmelCase )['last_hidden_state']
# select random slice
lowerCamelCase__ : List[str] = ids_tensor((1,) , output_from_past.shape[-1] ).item()
lowerCamelCase__ : Tuple = output_from_no_past[:, -1, random_slice_idx].detach()
lowerCamelCase__ : List[str] = output_from_past[:, 0, random_slice_idx].detach()
# test that outputs are equal for slice
self.parent.assertTrue(torch.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-3 ) )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : List[str] , ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = UMTaModel(config=UpperCAmelCase ).to(UpperCAmelCase ).half().eval()
lowerCamelCase__ : Optional[int] = model(**UpperCAmelCase )['last_hidden_state']
self.parent.assertFalse(torch.isnan(UpperCAmelCase ).any().item() )
@require_torch
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(UMTaModel, UMTaForConditionalGeneration, UMTaForQuestionAnswering) if is_torch_available() else ()
)
UpperCAmelCase__ = (UMTaForConditionalGeneration,) if is_torch_available() else ()
UpperCAmelCase__ = (
{
"""conversational""": UMTaForConditionalGeneration,
"""feature-extraction""": UMTaModel,
"""summarization""": UMTaForConditionalGeneration,
"""text2text-generation""": UMTaForConditionalGeneration,
"""translation""": UMTaForConditionalGeneration,
"""question-answering""": UMTaForQuestionAnswering,
}
if is_torch_available()
else {}
)
UpperCAmelCase__ = True
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = True
UpperCAmelCase__ = True
# The small UMT5 model needs higher percentages for CPU/MP tests
UpperCAmelCase__ = [0.8, 0.9]
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = UMTaModelTester(self )
@unittest.skip('Test has a segmentation fault on torch 1.8.0' )
def A_ ( self : Tuple ) -> int:
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Tuple = UMTaModel(config_and_inputs[0] ).to(UpperCAmelCase )
with tempfile.TemporaryDirectory() as tmpdirname:
torch.onnx.export(
UpperCAmelCase , (config_and_inputs[1], config_and_inputs[3], config_and_inputs[2]) , F"""{tmpdirname}/t5_test.onnx""" , export_params=UpperCAmelCase , opset_version=9 , input_names=['input_ids', 'decoder_input_ids'] , )
@unittest.skipIf(torch_device == 'cpu' , 'Cant do half precision' )
def A_ ( self : Tuple ) -> Optional[Any]:
lowerCamelCase__ : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model_fpaa_forward(*UpperCAmelCase )
def A_ ( self : List[Any] ) -> str:
lowerCamelCase__ : int = ['encoder_attentions', 'decoder_attentions', 'cross_attentions']
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Any = config_and_inputs[0]
lowerCamelCase__ : Any = UMTaForConditionalGeneration(UpperCAmelCase ).eval()
model.to(UpperCAmelCase )
lowerCamelCase__ : Tuple = {
'head_mask': torch.zeros(config.num_layers , config.num_heads , device=UpperCAmelCase ),
'decoder_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
'cross_attn_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
}
for attn_name, (name, mask) in zip(UpperCAmelCase , head_masking.items() ):
lowerCamelCase__ : Union[str, Any] = {name: mask}
# Explicitly pass decoder_head_mask as it is required from T5 model when head_mask specified
if name == "head_mask":
lowerCamelCase__ : Union[str, Any] = torch.ones(
config.num_decoder_layers , config.num_heads , device=UpperCAmelCase )
lowerCamelCase__ : Tuple = model.generate(
config_and_inputs[1]['input_ids'] , num_beams=1 , max_length=3 , output_attentions=UpperCAmelCase , return_dict_in_generate=UpperCAmelCase , **UpperCAmelCase , )
# We check the state of decoder_attentions and cross_attentions just from the last step
lowerCamelCase__ : Union[str, Any] = out[attn_name] if attn_name == attention_names[0] else out[attn_name][-1]
self.assertEqual(sum([w.sum().item() for w in attn_weights] ) , 0.0 )
@unittest.skip('Does not work on the tiny model as we keep hitting edge cases.' )
def A_ ( self : Optional[Any] ) -> Optional[Any]:
pass
@require_torch
@require_sentencepiece
@require_tokenizers
class lowerCAmelCase ( unittest.TestCase ):
@slow
@unittest.skip(
'Unless we stop stripping left and right by default for all special tokens, the expected ids obtained here will not match the original ones. Wait for https://github.com/huggingface/transformers/pull/23909 to be merged' )
def A_ ( self : Any ) -> int:
lowerCamelCase__ : Optional[Any] = UMTaForConditionalGeneration.from_pretrained('google/umt5-small' , return_dict=UpperCAmelCase ).to(UpperCAmelCase )
lowerCamelCase__ : List[str] = AutoTokenizer.from_pretrained('google/umt5-small' , use_fast=UpperCAmelCase , legacy=UpperCAmelCase )
lowerCamelCase__ : Dict = [
'Bonjour monsieur <extra_id_0> bien <extra_id_1>.',
'No se como puedo <extra_id_0>.',
'This is the reason why we <extra_id_0> them.',
'The <extra_id_0> walks in <extra_id_1>, seats',
'A <extra_id_0> walks into a bar and orders a <extra_id_1> with <extra_id_2> pinch of <extra_id_3>.',
]
lowerCamelCase__ : Tuple = tokenizer(UpperCAmelCase , return_tensors='pt' , padding=UpperCAmelCase ).input_ids
# fmt: off
lowerCamelCase__ : Any = torch.tensor(
[
[ 38530, 210703, 256299, 1410, 256298, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 826, 321, 671, 25922, 256299, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 1460, 339, 312, 19014, 10620, 758, 256299, 2355,274, 1, 0, 0, 0, 0, 0, 0,0, 0],
[ 517, 256299, 14869, 281, 301, 256298, 275, 119983,1, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 320, 256299, 14869, 281, 2234, 289, 2275, 333,61391, 289, 256298, 543, 256297, 168714, 329, 256296,274, 1],
] )
# fmt: on
torch.testing.assert_allclose(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Optional[int] = model.generate(input_ids.to(UpperCAmelCase ) )
lowerCamelCase__ : List[Any] = [
'<pad><extra_id_0> et<extra_id_1> [eod] <extra_id_2><extra_id_55>.. [eod] 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 <extra_id_56>ajšietosto<extra_id_56>lleux<extra_id_19><extra_id_6>ajšie</s>',
'<pad><extra_id_0>.<extra_id_1>.,<0x0A>...spech <0x0A><extra_id_20> <extra_id_21></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> are not going to be a part of the world. We are not going to be a part of<extra_id_1> and<extra_id_2><0x0A><extra_id_48>.<extra_id_48></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> door<extra_id_1>, the door<extra_id_2> 피해[/</s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0>nyone who<extra_id_1> drink<extra_id_2> a<extra_id_3> alcohol<extra_id_4> A<extra_id_5> A. This<extra_id_6> I<extra_id_7><extra_id_52><extra_id_53></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
]
lowerCamelCase__ : Union[str, Any] = tokenizer.batch_decode(UpperCAmelCase )
self.assertEqual(UpperCAmelCase , UpperCAmelCase )
| 50
| 1
|
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_UpperCAmelCase : int = logging.get_logger(__name__)
_UpperCAmelCase : int = {
"""andreasmadsen/efficient_mlm_m0.40""": (
"""https://huggingface.co/andreasmadsen/efficient_mlm_m0.40/resolve/main/config.json"""
),
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """roberta-prelayernorm"""
def __init__( self : Union[str, Any] , UpperCAmelCase : Union[str, Any]=50265 , UpperCAmelCase : List[str]=768 , UpperCAmelCase : Union[str, Any]=12 , UpperCAmelCase : int=12 , UpperCAmelCase : int=3072 , UpperCAmelCase : List[str]="gelu" , UpperCAmelCase : Optional[int]=0.1 , UpperCAmelCase : int=0.1 , UpperCAmelCase : Optional[int]=512 , UpperCAmelCase : Union[str, Any]=2 , UpperCAmelCase : Optional[Any]=0.0_2 , UpperCAmelCase : Any=1e-12 , UpperCAmelCase : Optional[Any]=1 , UpperCAmelCase : Tuple=0 , UpperCAmelCase : Union[str, Any]=2 , UpperCAmelCase : Optional[Any]="absolute" , UpperCAmelCase : Optional[int]=True , UpperCAmelCase : Union[str, Any]=None , **UpperCAmelCase : Any , ) -> Any:
super().__init__(pad_token_id=UpperCAmelCase , bos_token_id=UpperCAmelCase , eos_token_id=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Tuple = vocab_size
lowerCamelCase__ : List[Any] = hidden_size
lowerCamelCase__ : int = num_hidden_layers
lowerCamelCase__ : Optional[int] = num_attention_heads
lowerCamelCase__ : Union[str, Any] = hidden_act
lowerCamelCase__ : Any = intermediate_size
lowerCamelCase__ : Union[str, Any] = hidden_dropout_prob
lowerCamelCase__ : Union[str, Any] = attention_probs_dropout_prob
lowerCamelCase__ : Dict = max_position_embeddings
lowerCamelCase__ : Union[str, Any] = type_vocab_size
lowerCamelCase__ : List[Any] = initializer_range
lowerCamelCase__ : str = layer_norm_eps
lowerCamelCase__ : str = position_embedding_type
lowerCamelCase__ : Any = use_cache
lowerCamelCase__ : List[Any] = classifier_dropout
class lowerCAmelCase ( __UpperCamelCase ):
@property
def A_ ( self : Any ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCamelCase__ : Union[str, Any] = {0: 'batch', 1: 'choice', 2: 'sequence'}
else:
lowerCamelCase__ : Dict = {0: 'batch', 1: 'sequence'}
return OrderedDict(
[
('input_ids', dynamic_axis),
('attention_mask', dynamic_axis),
] )
| 50
|
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=None , _UpperCAmelCase="no" , _UpperCAmelCase="29500" ) -> Tuple:
lowerCamelCase__ : Dict = False
lowerCamelCase__ : Dict = False
if any(key.startswith('KAGGLE' ) for key in os.environ.keys() ):
lowerCamelCase__ : Optional[Any] = True
elif "IPython" in sys.modules:
lowerCamelCase__ : Optional[Any] = 'google.colab' in str(sys.modules['IPython'].get_ipython() )
try:
lowerCamelCase__ : List[str] = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
F"""Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}.""" )
if (in_colab or in_kaggle) and (os.environ.get('TPU_NAME' , _UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '
'your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if num_processes is None:
lowerCamelCase__ : Optional[Any] = 8
lowerCamelCase__ : List[str] = PrepareForLaunch(_UpperCAmelCase , distributed_type='TPU' )
print(F"""Launching a training on {num_processes} TPU cores.""" )
xmp.spawn(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on one CPU.' )
function(*_UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '
'inside your training function. Restart your notebook and make sure no cells initializes an '
'`Accelerator`.' )
if torch.cuda.is_initialized():
raise ValueError(
'To launch a multi-GPU training from your notebook, you need to avoid running any instruction '
'using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '
'function.' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port=_UpperCAmelCase , mixed_precision=_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = PrepareForLaunch(_UpperCAmelCase , distributed_type='MULTI_GPU' )
print(F"""Launching training on {num_processes} GPUs.""" )
try:
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '
'This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '
'Please review your imports and test them when running the `notebook_launcher()` to identify '
'which one is problematic.' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
lowerCamelCase__ : int = '1'
print('Launching training on MPS.' )
elif torch.cuda.is_available():
print('Launching training on one GPU.' )
else:
print('Launching training on CPU.' )
function(*_UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase=() , _UpperCAmelCase=2 ) -> Optional[Any]:
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=_UpperCAmelCase , master_addr='127.0.01' , master_port='29500' , accelerate_mixed_precision='no' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='yes' , ):
lowerCamelCase__ : Optional[Any] = PrepareForLaunch(_UpperCAmelCase , debug=_UpperCAmelCase )
start_processes(_UpperCAmelCase , args=_UpperCAmelCase , nprocs=_UpperCAmelCase , start_method='fork' )
| 50
| 1
|
import collections
import inspect
import unittest
from transformers import FocalNetConfig
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_backbone_common import BackboneTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, _config_zero_init, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import (
FocalNetBackbone,
FocalNetForImageClassification,
FocalNetForMaskedImageModeling,
FocalNetModel,
)
from transformers.models.focalnet.modeling_focalnet import FOCALNET_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class lowerCAmelCase :
def __init__( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : Any=13 , UpperCAmelCase : List[str]=32 , UpperCAmelCase : Any=2 , UpperCAmelCase : Union[str, Any]=3 , UpperCAmelCase : Union[str, Any]=16 , UpperCAmelCase : Tuple=[32, 64, 128] , UpperCAmelCase : Optional[Any]=[1, 2, 1] , UpperCAmelCase : Dict=[2, 2, 4] , UpperCAmelCase : Union[str, Any]=2 , UpperCAmelCase : str=2.0 , UpperCAmelCase : int=True , UpperCAmelCase : Tuple=0.0 , UpperCAmelCase : Dict=0.0 , UpperCAmelCase : int=0.1 , UpperCAmelCase : Tuple="gelu" , UpperCAmelCase : Dict=False , UpperCAmelCase : Union[str, Any]=True , UpperCAmelCase : Union[str, Any]=0.0_2 , UpperCAmelCase : Tuple=1e-5 , UpperCAmelCase : Optional[int]=True , UpperCAmelCase : Optional[Any]=None , UpperCAmelCase : List[str]=True , UpperCAmelCase : Union[str, Any]=10 , UpperCAmelCase : Union[str, Any]=8 , UpperCAmelCase : Dict=["stage1", "stage2"] , UpperCAmelCase : List[Any]=[1, 2] , ) -> Any:
lowerCamelCase__ : Dict = parent
lowerCamelCase__ : str = batch_size
lowerCamelCase__ : List[Any] = image_size
lowerCamelCase__ : Optional[Any] = patch_size
lowerCamelCase__ : Any = num_channels
lowerCamelCase__ : List[Any] = embed_dim
lowerCamelCase__ : Optional[Any] = hidden_sizes
lowerCamelCase__ : Any = depths
lowerCamelCase__ : List[str] = num_heads
lowerCamelCase__ : Any = window_size
lowerCamelCase__ : int = mlp_ratio
lowerCamelCase__ : str = qkv_bias
lowerCamelCase__ : Any = hidden_dropout_prob
lowerCamelCase__ : Tuple = attention_probs_dropout_prob
lowerCamelCase__ : str = drop_path_rate
lowerCamelCase__ : int = hidden_act
lowerCamelCase__ : Any = use_absolute_embeddings
lowerCamelCase__ : List[str] = patch_norm
lowerCamelCase__ : Dict = layer_norm_eps
lowerCamelCase__ : str = initializer_range
lowerCamelCase__ : str = is_training
lowerCamelCase__ : Tuple = scope
lowerCamelCase__ : int = use_labels
lowerCamelCase__ : Optional[int] = type_sequence_label_size
lowerCamelCase__ : str = encoder_stride
lowerCamelCase__ : Union[str, Any] = out_features
lowerCamelCase__ : Dict = out_indices
def A_ ( self : List[Any] ) -> Dict:
lowerCamelCase__ : Union[str, Any] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
lowerCamelCase__ : Any = None
if self.use_labels:
lowerCamelCase__ : List[Any] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
lowerCamelCase__ : Union[str, Any] = self.get_config()
return config, pixel_values, labels
def A_ ( self : Optional[int] ) -> Any:
return FocalNetConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , embed_dim=self.embed_dim , hidden_sizes=self.hidden_sizes , 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 A_ ( self : List[Any] , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : int , UpperCAmelCase : Tuple ) -> int:
lowerCamelCase__ : str = FocalNetModel(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Union[str, Any] = model(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
lowerCamelCase__ : 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 A_ ( self : Tuple , UpperCAmelCase : str , UpperCAmelCase : List[str] , UpperCAmelCase : int ) -> str:
lowerCamelCase__ : Tuple = FocalNetBackbone(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Union[str, Any] = model(UpperCAmelCase )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , len(config.out_features ) )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [self.batch_size, self.image_size, 8, 8] )
# 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
lowerCamelCase__ : Any = None
lowerCamelCase__ : int = FocalNetBackbone(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Optional[int] = model(UpperCAmelCase )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , 1 )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [self.batch_size, self.image_size * 2, 4, 4] )
# verify channels
self.parent.assertEqual(len(model.channels ) , 1 )
self.parent.assertListEqual(model.channels , [config.hidden_sizes[-1]] )
def A_ ( self : List[Any] , UpperCAmelCase : Tuple , UpperCAmelCase : List[Any] , UpperCAmelCase : List[Any] ) -> List[Any]:
lowerCamelCase__ : Dict = FocalNetForMaskedImageModeling(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Optional[int] = model(UpperCAmelCase )
self.parent.assertEqual(
result.reconstruction.shape , (self.batch_size, self.num_channels, self.image_size, self.image_size) )
# test greyscale images
lowerCamelCase__ : int = 1
lowerCamelCase__ : Optional[Any] = FocalNetForMaskedImageModeling(UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : List[str] = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
lowerCamelCase__ : Tuple = model(UpperCAmelCase )
self.parent.assertEqual(result.reconstruction.shape , (self.batch_size, 1, self.image_size, self.image_size) )
def A_ ( self : int , UpperCAmelCase : str , UpperCAmelCase : Dict , UpperCAmelCase : Dict ) -> int:
lowerCamelCase__ : Dict = self.type_sequence_label_size
lowerCamelCase__ : int = FocalNetForImageClassification(UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Dict = model(UpperCAmelCase , labels=UpperCAmelCase )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
# test greyscale images
lowerCamelCase__ : Optional[Any] = 1
lowerCamelCase__ : Optional[Any] = FocalNetForImageClassification(UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : int = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
lowerCamelCase__ : str = model(UpperCAmelCase )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def A_ ( self : Union[str, Any] ) -> int:
lowerCamelCase__ : Any = self.prepare_config_and_inputs()
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = config_and_inputs
lowerCamelCase__ : List[Any] = {'pixel_values': pixel_values}
return config, inputs_dict
@require_torch
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(
FocalNetModel,
FocalNetForImageClassification,
FocalNetForMaskedImageModeling,
FocalNetBackbone,
)
if is_torch_available()
else ()
)
UpperCAmelCase__ = (
{"""feature-extraction""": FocalNetModel, """image-classification""": FocalNetForImageClassification}
if is_torch_available()
else {}
)
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = False
def A_ ( self : Any ) -> List[str]:
lowerCamelCase__ : Tuple = FocalNetModelTester(self )
lowerCamelCase__ : Union[str, Any] = ConfigTester(self , config_class=UpperCAmelCase , embed_dim=37 , has_text_modality=UpperCAmelCase )
def A_ ( self : str ) -> Tuple:
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 A_ ( self : Any ) -> Dict:
return
def A_ ( self : List[Any] ) -> str:
lowerCamelCase__ : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*UpperCAmelCase )
def A_ ( self : Optional[Any] ) -> List[str]:
lowerCamelCase__ : Dict = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_backbone(*UpperCAmelCase )
def A_ ( self : Optional[int] ) -> int:
lowerCamelCase__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_image_modeling(*UpperCAmelCase )
def A_ ( self : str ) -> str:
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*UpperCAmelCase )
@unittest.skip(reason='FocalNet does not use inputs_embeds' )
def A_ ( self : Any ) -> List[str]:
pass
@unittest.skip(reason='FocalNet does not use feedforward chunking' )
def A_ ( self : Tuple ) -> Any:
pass
def A_ ( self : List[Any] ) -> List[Any]:
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes[:-1]:
lowerCamelCase__ : List[str] = model_class(UpperCAmelCase )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
lowerCamelCase__ : Optional[int] = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(UpperCAmelCase , nn.Linear ) )
def A_ ( self : Any ) -> str:
lowerCamelCase__ , lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes[:-1]:
lowerCamelCase__ : List[Any] = model_class(UpperCAmelCase )
lowerCamelCase__ : List[Any] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
lowerCamelCase__ : Union[str, Any] = [*signature.parameters.keys()]
lowerCamelCase__ : Tuple = ['pixel_values']
self.assertListEqual(arg_names[:1] , UpperCAmelCase )
def A_ ( self : Tuple , UpperCAmelCase : Dict , UpperCAmelCase : Tuple , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : Optional[int] ) -> List[Any]:
lowerCamelCase__ : Tuple = model_class(UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
with torch.no_grad():
lowerCamelCase__ : Tuple = model(**self._prepare_for_class(UpperCAmelCase , UpperCAmelCase ) )
lowerCamelCase__ : int = outputs.hidden_states
lowerCamelCase__ : Any = getattr(
self.model_tester , 'expected_num_hidden_layers' , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(UpperCAmelCase ) , UpperCAmelCase )
# FocalNet has a different seq_length
lowerCamelCase__ : Union[str, Any] = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
lowerCamelCase__ : List[str] = (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] , )
lowerCamelCase__ : Tuple = outputs.reshaped_hidden_states
self.assertEqual(len(UpperCAmelCase ) , UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : str = reshaped_hidden_states[0].shape
lowerCamelCase__ : str = (
reshaped_hidden_states[0].view(UpperCAmelCase , UpperCAmelCase , height * width ).permute(0 , 2 , 1 )
)
self.assertListEqual(
list(reshaped_hidden_states.shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def A_ ( self : int ) -> List[str]:
lowerCamelCase__ , lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs_for_common()
lowerCamelCase__ : Tuple = (
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[:-1]:
lowerCamelCase__ : Optional[int] = True
self.check_hidden_states_output(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
lowerCamelCase__ : Optional[Any] = True
self.check_hidden_states_output(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def A_ ( self : List[str] ) -> Dict:
lowerCamelCase__ , lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs_for_common()
lowerCamelCase__ : int = 3
lowerCamelCase__ : Any = (
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)
)
lowerCamelCase__ : Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
lowerCamelCase__ : List[str] = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
lowerCamelCase__ : Union[str, Any] = image_size[1] + patch_size[1] - (image_size[1] % patch_size[1])
for model_class in self.all_model_classes[:-1]:
lowerCamelCase__ : Optional[Any] = True
self.check_hidden_states_output(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , (padded_height, padded_width) )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
lowerCamelCase__ : List[Any] = True
self.check_hidden_states_output(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , (padded_height, padded_width) )
@slow
def A_ ( self : Dict ) -> Union[str, Any]:
for model_name in FOCALNET_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
lowerCamelCase__ : Tuple = FocalNetModel.from_pretrained(UpperCAmelCase )
self.assertIsNotNone(UpperCAmelCase )
def A_ ( self : List[Any] ) -> Dict:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
lowerCamelCase__ : Optional[int] = _config_zero_init(UpperCAmelCase )
for model_class in self.all_model_classes:
lowerCamelCase__ : Union[str, Any] = model_class(config=UpperCAmelCase )
for name, param in model.named_parameters():
if "embeddings" not in name and param.requires_grad:
self.assertIn(
((param.data.mean() * 1e9).round() / 1e9).item() , [0.0, 1.0] , msg=F"""Parameter {name} of model {model_class} seems not properly initialized""" , )
@require_vision
@require_torch
class lowerCAmelCase ( unittest.TestCase ):
@cached_property
def A_ ( self : str ) -> List[str]:
# TODO update organization
return AutoImageProcessor.from_pretrained('microsoft/focalnet-tiny' ) if is_vision_available() else None
@slow
def A_ ( self : List[Any] ) -> int:
lowerCamelCase__ : List[str] = FocalNetForImageClassification.from_pretrained('microsoft/focalnet-tiny' ).to(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = self.default_image_processor
lowerCamelCase__ : List[str] = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' )
lowerCamelCase__ : List[str] = image_processor(images=UpperCAmelCase , return_tensors='pt' ).to(UpperCAmelCase )
# forward pass
with torch.no_grad():
lowerCamelCase__ : List[Any] = model(**UpperCAmelCase )
# verify the logits
lowerCamelCase__ : List[str] = torch.Size((1, 1000) )
self.assertEqual(outputs.logits.shape , UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = torch.tensor([0.2_1_6_6, -0.4_3_6_8, 0.2_1_9_1] ).to(UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.logits[0, :3] , UpperCAmelCase , atol=1e-4 ) )
self.assertTrue(outputs.logits.argmax(dim=-1 ).item() , 281 )
@require_torch
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (FocalNetBackbone,) if is_torch_available() else ()
UpperCAmelCase__ = FocalNetConfig
UpperCAmelCase__ = False
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Any = FocalNetModelTester(self )
| 50
|
from dataclasses import dataclass
from typing import Optional, Tuple, Union
import torch
import torch.nn as nn
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import BaseOutput
from .embeddings import GaussianFourierProjection, TimestepEmbedding, Timesteps
from .modeling_utils import ModelMixin
from .unet_ad_blocks import get_down_block, get_mid_block, get_out_block, get_up_block
@dataclass
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase ):
@register_to_config
def __init__( self : List[str] , UpperCAmelCase : int = 65536 , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 2 , UpperCAmelCase : int = 0 , UpperCAmelCase : str = "fourier" , UpperCAmelCase : bool = True , UpperCAmelCase : bool = False , UpperCAmelCase : float = 0.0 , UpperCAmelCase : Tuple[str] = ("DownBlock1DNoSkip", "DownBlock1D", "AttnDownBlock1D") , UpperCAmelCase : Tuple[str] = ("AttnUpBlock1D", "UpBlock1D", "UpBlock1DNoSkip") , UpperCAmelCase : Tuple[str] = "UNetMidBlock1D" , UpperCAmelCase : str = None , UpperCAmelCase : Tuple[int] = (32, 32, 64) , UpperCAmelCase : str = None , UpperCAmelCase : int = 8 , UpperCAmelCase : int = 1 , UpperCAmelCase : bool = False , ) -> List[Any]:
super().__init__()
lowerCamelCase__ : Optional[int] = sample_size
# time
if time_embedding_type == "fourier":
lowerCamelCase__ : Optional[Any] = GaussianFourierProjection(
embedding_size=8 , set_W_to_weight=UpperCAmelCase , log=UpperCAmelCase , flip_sin_to_cos=UpperCAmelCase )
lowerCamelCase__ : Any = 2 * block_out_channels[0]
elif time_embedding_type == "positional":
lowerCamelCase__ : List[Any] = Timesteps(
block_out_channels[0] , flip_sin_to_cos=UpperCAmelCase , downscale_freq_shift=UpperCAmelCase )
lowerCamelCase__ : Dict = block_out_channels[0]
if use_timestep_embedding:
lowerCamelCase__ : str = block_out_channels[0] * 4
lowerCamelCase__ : List[Any] = TimestepEmbedding(
in_channels=UpperCAmelCase , time_embed_dim=UpperCAmelCase , act_fn=UpperCAmelCase , out_dim=block_out_channels[0] , )
lowerCamelCase__ : Any = nn.ModuleList([] )
lowerCamelCase__ : Tuple = None
lowerCamelCase__ : List[str] = nn.ModuleList([] )
lowerCamelCase__ : Optional[int] = None
# down
lowerCamelCase__ : Optional[int] = in_channels
for i, down_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Union[str, Any] = output_channel
lowerCamelCase__ : Tuple = block_out_channels[i]
if i == 0:
input_channel += extra_in_channels
lowerCamelCase__ : Union[str, Any] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Optional[int] = get_down_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_downsample=not is_final_block or downsample_each_block , )
self.down_blocks.append(UpperCAmelCase )
# mid
lowerCamelCase__ : Optional[int] = get_mid_block(
UpperCAmelCase , in_channels=block_out_channels[-1] , mid_channels=block_out_channels[-1] , out_channels=block_out_channels[-1] , embed_dim=block_out_channels[0] , num_layers=UpperCAmelCase , add_downsample=UpperCAmelCase , )
# up
lowerCamelCase__ : Optional[int] = list(reversed(UpperCAmelCase ) )
lowerCamelCase__ : Optional[int] = reversed_block_out_channels[0]
if out_block_type is None:
lowerCamelCase__ : List[str] = out_channels
else:
lowerCamelCase__ : Any = block_out_channels[0]
for i, up_block_type in enumerate(UpperCAmelCase ):
lowerCamelCase__ : Tuple = output_channel
lowerCamelCase__ : Union[str, Any] = (
reversed_block_out_channels[i + 1] if i < len(UpperCAmelCase ) - 1 else final_upsample_channels
)
lowerCamelCase__ : List[str] = i == len(UpperCAmelCase ) - 1
lowerCamelCase__ : Dict = get_up_block(
UpperCAmelCase , num_layers=UpperCAmelCase , in_channels=UpperCAmelCase , out_channels=UpperCAmelCase , temb_channels=block_out_channels[0] , add_upsample=not is_final_block , )
self.up_blocks.append(UpperCAmelCase )
lowerCamelCase__ : int = output_channel
# out
lowerCamelCase__ : int = norm_num_groups if norm_num_groups is not None else min(block_out_channels[0] // 4 , 32 )
lowerCamelCase__ : List[Any] = get_out_block(
out_block_type=UpperCAmelCase , num_groups_out=UpperCAmelCase , embed_dim=block_out_channels[0] , out_channels=UpperCAmelCase , act_fn=UpperCAmelCase , fc_dim=block_out_channels[-1] // 4 , )
def A_ ( self : List[Any] , UpperCAmelCase : torch.FloatTensor , UpperCAmelCase : Union[torch.Tensor, float, int] , UpperCAmelCase : bool = True , ) -> Union[UNetaDOutput, Tuple]:
lowerCamelCase__ : Optional[Any] = timestep
if not torch.is_tensor(UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = torch.tensor([timesteps] , dtype=torch.long , device=sample.device )
elif torch.is_tensor(UpperCAmelCase ) and len(timesteps.shape ) == 0:
lowerCamelCase__ : List[str] = timesteps[None].to(sample.device )
lowerCamelCase__ : Optional[int] = self.time_proj(UpperCAmelCase )
if self.config.use_timestep_embedding:
lowerCamelCase__ : str = self.time_mlp(UpperCAmelCase )
else:
lowerCamelCase__ : List[str] = timestep_embed[..., None]
lowerCamelCase__ : str = timestep_embed.repeat([1, 1, sample.shape[2]] ).to(sample.dtype )
lowerCamelCase__ : str = timestep_embed.broadcast_to((sample.shape[:1] + timestep_embed.shape[1:]) )
# 2. down
lowerCamelCase__ : str = ()
for downsample_block in self.down_blocks:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = downsample_block(hidden_states=UpperCAmelCase , temb=UpperCAmelCase )
down_block_res_samples += res_samples
# 3. mid
if self.mid_block:
lowerCamelCase__ : Optional[Any] = self.mid_block(UpperCAmelCase , UpperCAmelCase )
# 4. up
for i, upsample_block in enumerate(self.up_blocks ):
lowerCamelCase__ : Dict = down_block_res_samples[-1:]
lowerCamelCase__ : Optional[Any] = down_block_res_samples[:-1]
lowerCamelCase__ : Any = upsample_block(UpperCAmelCase , res_hidden_states_tuple=UpperCAmelCase , temb=UpperCAmelCase )
# 5. post-process
if self.out_block:
lowerCamelCase__ : Any = self.out_block(UpperCAmelCase , UpperCAmelCase )
if not return_dict:
return (sample,)
return UNetaDOutput(sample=UpperCAmelCase )
| 50
| 1
|
from transformers import DistilBertTokenizer, DistilBertTokenizerFast
from transformers.testing_utils import require_tokenizers, slow
from ..bert.test_tokenization_bert import BertTokenizationTest
@require_tokenizers
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = DistilBertTokenizer
UpperCAmelCase__ = DistilBertTokenizerFast
UpperCAmelCase__ = True
@slow
def A_ ( self : str ) -> Any:
lowerCamelCase__ : List[str] = DistilBertTokenizer.from_pretrained('distilbert-base-uncased' )
lowerCamelCase__ : Dict = tokenizer.encode('sequence builders' , add_special_tokens=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tokenizer.encode('multi-sequence build' , add_special_tokens=UpperCAmelCase )
lowerCamelCase__ : List[str] = tokenizer.build_inputs_with_special_tokens(UpperCAmelCase )
lowerCamelCase__ : List[Any] = tokenizer.build_inputs_with_special_tokens(UpperCAmelCase , UpperCAmelCase )
assert encoded_sentence == [tokenizer.cls_token_id] + text + [tokenizer.sep_token_id]
assert encoded_pair == [tokenizer.cls_token_id] + text + [tokenizer.sep_token_id] + text_a + [
tokenizer.sep_token_id
]
| 50
|
from __future__ import annotations
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> list[tuple[int, int]]:
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = position
lowerCamelCase__ : Optional[Any] = [
(y + 1, x + 2),
(y - 1, x + 2),
(y + 1, x - 2),
(y - 1, x - 2),
(y + 2, x + 1),
(y + 2, x - 1),
(y - 2, x + 1),
(y - 2, x - 1),
]
lowerCamelCase__ : Dict = []
for position in positions:
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = position
if 0 <= y_test < n and 0 <= x_test < n:
permissible_positions.append(_UpperCAmelCase )
return permissible_positions
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> bool:
return not any(elem == 0 for row in board for elem in row )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> bool:
if is_complete(_UpperCAmelCase ):
return True
for position in get_valid_pos(_UpperCAmelCase , len(_UpperCAmelCase ) ):
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = position
if board[y][x] == 0:
lowerCamelCase__ : List[Any] = curr + 1
if open_knight_tour_helper(_UpperCAmelCase , _UpperCAmelCase , curr + 1 ):
return True
lowerCamelCase__ : Optional[Any] = 0
return False
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> list[list[int]]:
lowerCamelCase__ : Any = [[0 for i in range(_UpperCAmelCase )] for j in range(_UpperCAmelCase )]
for i in range(_UpperCAmelCase ):
for j in range(_UpperCAmelCase ):
lowerCamelCase__ : Optional[int] = 1
if open_knight_tour_helper(_UpperCAmelCase , (i, j) , 1 ):
return board
lowerCamelCase__ : Optional[Any] = 0
lowerCamelCase__ : Any = F"""Open Kight Tour cannot be performed on a board of size {n}"""
raise ValueError(_UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
import math
from typing import Dict, Iterable, List, Optional, Tuple, Union
import numpy as np
from ...image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict
from ...image_transforms import normalize, rescale, resize, to_channel_dimension_format
from ...image_utils import (
IMAGENET_STANDARD_MEAN,
IMAGENET_STANDARD_STD,
ChannelDimension,
ImageInput,
PILImageResampling,
get_image_size,
is_torch_available,
is_torch_tensor,
make_list_of_images,
to_numpy_array,
valid_images,
)
from ...utils import TensorType, is_vision_available, logging
if is_torch_available():
import torch
if is_vision_available():
import PIL
_UpperCAmelCase : int = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Tuple[int, int]:
def constraint_to_multiple_of(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=0 , _UpperCAmelCase=None ):
lowerCamelCase__ : int = round(val / multiple ) * multiple
if max_val is not None and x > max_val:
lowerCamelCase__ : Tuple = math.floor(val / multiple ) * multiple
if x < min_val:
lowerCamelCase__ : Tuple = math.ceil(val / multiple ) * multiple
return x
lowerCamelCase__ : Any = (output_size, output_size) if isinstance(_UpperCAmelCase , _UpperCAmelCase ) else output_size
lowerCamelCase__ , lowerCamelCase__ : Dict = get_image_size(_UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Optional[int] = output_size
# determine new height and width
lowerCamelCase__ : Dict = output_height / input_height
lowerCamelCase__ : Optional[int] = output_width / input_width
if keep_aspect_ratio:
# scale as little as possible
if abs(1 - scale_width ) < abs(1 - scale_height ):
# fit width
lowerCamelCase__ : Union[str, Any] = scale_width
else:
# fit height
lowerCamelCase__ : Tuple = scale_height
lowerCamelCase__ : Optional[int] = constraint_to_multiple_of(scale_height * input_height , multiple=_UpperCAmelCase )
lowerCamelCase__ : Dict = constraint_to_multiple_of(scale_width * input_width , multiple=_UpperCAmelCase )
return (new_height, new_width)
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = ["""pixel_values"""]
def __init__( self : Optional[int] , UpperCAmelCase : bool = True , UpperCAmelCase : Dict[str, int] = None , UpperCAmelCase : PILImageResampling = PILImageResampling.BILINEAR , UpperCAmelCase : bool = False , UpperCAmelCase : int = 1 , UpperCAmelCase : bool = True , UpperCAmelCase : Union[int, float] = 1 / 255 , UpperCAmelCase : bool = True , UpperCAmelCase : Optional[Union[float, List[float]]] = None , UpperCAmelCase : Optional[Union[float, List[float]]] = None , **UpperCAmelCase : Optional[Any] , ) -> None:
super().__init__(**UpperCAmelCase )
lowerCamelCase__ : Optional[int] = size if size is not None else {'height': 384, 'width': 384}
lowerCamelCase__ : str = get_size_dict(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = do_resize
lowerCamelCase__ : List[str] = size
lowerCamelCase__ : Optional[int] = keep_aspect_ratio
lowerCamelCase__ : Union[str, Any] = ensure_multiple_of
lowerCamelCase__ : int = resample
lowerCamelCase__ : str = do_rescale
lowerCamelCase__ : List[Any] = rescale_factor
lowerCamelCase__ : List[Any] = do_normalize
lowerCamelCase__ : str = image_mean if image_mean is not None else IMAGENET_STANDARD_MEAN
lowerCamelCase__ : Tuple = image_std if image_std is not None else IMAGENET_STANDARD_STD
def A_ ( self : List[str] , UpperCAmelCase : np.ndarray , UpperCAmelCase : Dict[str, int] , UpperCAmelCase : bool = False , UpperCAmelCase : int = 1 , UpperCAmelCase : PILImageResampling = PILImageResampling.BICUBIC , UpperCAmelCase : Optional[Union[str, ChannelDimension]] = None , **UpperCAmelCase : str , ) -> np.ndarray:
lowerCamelCase__ : Optional[int] = get_size_dict(UpperCAmelCase )
if "height" not in size or "width" not in size:
raise ValueError(F"""The size dictionary must contain the keys 'height' and 'width'. Got {size.keys()}""" )
lowerCamelCase__ : Union[str, Any] = get_resize_output_image_size(
UpperCAmelCase , output_size=(size['height'], size['width']) , keep_aspect_ratio=UpperCAmelCase , multiple=UpperCAmelCase , )
return resize(UpperCAmelCase , size=UpperCAmelCase , resample=UpperCAmelCase , data_format=UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : List[str] , UpperCAmelCase : np.ndarray , UpperCAmelCase : Union[int, float] , UpperCAmelCase : Optional[Union[str, ChannelDimension]] = None , **UpperCAmelCase : int , ) -> str:
return rescale(UpperCAmelCase , scale=UpperCAmelCase , data_format=UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Optional[int] , UpperCAmelCase : np.ndarray , UpperCAmelCase : Union[float, List[float]] , UpperCAmelCase : Union[float, List[float]] , UpperCAmelCase : Optional[Union[str, ChannelDimension]] = None , **UpperCAmelCase : str , ) -> np.ndarray:
return normalize(UpperCAmelCase , mean=UpperCAmelCase , std=UpperCAmelCase , data_format=UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Union[str, Any] , UpperCAmelCase : ImageInput , UpperCAmelCase : bool = None , UpperCAmelCase : int = None , UpperCAmelCase : bool = None , UpperCAmelCase : int = None , UpperCAmelCase : PILImageResampling = None , UpperCAmelCase : bool = None , UpperCAmelCase : float = None , UpperCAmelCase : bool = None , UpperCAmelCase : Optional[Union[float, List[float]]] = None , UpperCAmelCase : Optional[Union[float, List[float]]] = None , UpperCAmelCase : Optional[Union[str, TensorType]] = None , UpperCAmelCase : ChannelDimension = ChannelDimension.FIRST , **UpperCAmelCase : List[str] , ) -> PIL.Image.Image:
lowerCamelCase__ : Optional[int] = do_resize if do_resize is not None else self.do_resize
lowerCamelCase__ : str = size if size is not None else self.size
lowerCamelCase__ : int = get_size_dict(UpperCAmelCase )
lowerCamelCase__ : List[str] = keep_aspect_ratio if keep_aspect_ratio is not None else self.keep_aspect_ratio
lowerCamelCase__ : Union[str, Any] = ensure_multiple_of if ensure_multiple_of is not None else self.ensure_multiple_of
lowerCamelCase__ : Any = resample if resample is not None else self.resample
lowerCamelCase__ : Optional[Any] = do_rescale if do_rescale is not None else self.do_rescale
lowerCamelCase__ : List[Any] = rescale_factor if rescale_factor is not None else self.rescale_factor
lowerCamelCase__ : Dict = do_normalize if do_normalize is not None else self.do_normalize
lowerCamelCase__ : Dict = image_mean if image_mean is not None else self.image_mean
lowerCamelCase__ : List[Any] = image_std if image_std is not None else self.image_std
lowerCamelCase__ : Optional[Any] = make_list_of_images(UpperCAmelCase )
if not valid_images(UpperCAmelCase ):
raise ValueError(
'Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, '
'torch.Tensor, tf.Tensor or jax.ndarray.' )
if do_resize and size is None or resample is None:
raise ValueError('Size and resample must be specified if do_resize is True.' )
if do_rescale and rescale_factor is None:
raise ValueError('Rescale factor must be specified if do_rescale is True.' )
if do_normalize and (image_mean is None or image_std is None):
raise ValueError('Image mean and std must be specified if do_normalize is True.' )
# All transformations expect numpy arrays.
lowerCamelCase__ : Dict = [to_numpy_array(UpperCAmelCase ) for image in images]
if do_resize:
lowerCamelCase__ : Tuple = [self.resize(image=UpperCAmelCase , size=UpperCAmelCase , resample=UpperCAmelCase ) for image in images]
if do_rescale:
lowerCamelCase__ : Tuple = [self.rescale(image=UpperCAmelCase , scale=UpperCAmelCase ) for image in images]
if do_normalize:
lowerCamelCase__ : List[Any] = [self.normalize(image=UpperCAmelCase , mean=UpperCAmelCase , std=UpperCAmelCase ) for image in images]
lowerCamelCase__ : List[str] = [to_channel_dimension_format(UpperCAmelCase , UpperCAmelCase ) for image in images]
lowerCamelCase__ : Tuple = {'pixel_values': images}
return BatchFeature(data=UpperCAmelCase , tensor_type=UpperCAmelCase )
def A_ ( self : Dict , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[Tuple] = None ) -> List[Any]:
lowerCamelCase__ : Dict = outputs.logits
# Resize logits and compute semantic segmentation maps
if target_sizes is not None:
if len(UpperCAmelCase ) != len(UpperCAmelCase ):
raise ValueError(
'Make sure that you pass in as many target sizes as the batch dimension of the logits' )
if is_torch_tensor(UpperCAmelCase ):
lowerCamelCase__ : Dict = target_sizes.numpy()
lowerCamelCase__ : int = []
for idx in range(len(UpperCAmelCase ) ):
lowerCamelCase__ : Optional[int] = torch.nn.functional.interpolate(
logits[idx].unsqueeze(dim=0 ) , size=target_sizes[idx] , mode='bilinear' , align_corners=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = resized_logits[0].argmax(dim=0 )
semantic_segmentation.append(UpperCAmelCase )
else:
lowerCamelCase__ : Optional[Any] = logits.argmax(dim=1 )
lowerCamelCase__ : Dict = [semantic_segmentation[i] for i in range(semantic_segmentation.shape[0] )]
return semantic_segmentation
| 50
|
import argparse
import json
from collections import OrderedDict
import torch
from huggingface_hub import cached_download, hf_hub_url
from transformers import AutoImageProcessor, CvtConfig, CvtForImageClassification
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> int:
lowerCamelCase__ : Optional[int] = []
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.weight""",
F"""stage{idx}.patch_embed.proj.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.bias""",
F"""stage{idx}.patch_embed.proj.bias""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.weight""",
F"""stage{idx}.patch_embed.norm.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.bias""",
F"""stage{idx}.patch_embed.norm.bias""",
) )
return embed
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Tuple = []
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj.bias""",
) )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.weight""", F"""stage{idx}.blocks.{cnt}.norm1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.bias""", F"""stage{idx}.blocks.{cnt}.norm1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.weight""", F"""stage{idx}.blocks.{cnt}.norm2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.bias""", F"""stage{idx}.blocks.{cnt}.norm2.bias""") )
return attention_weights
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = []
token.append((F"""cvt.encoder.stages.{idx}.cls_token""", 'stage2.cls_token') )
return token
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : str = []
head.append(('layernorm.weight', 'norm.weight') )
head.append(('layernorm.bias', 'norm.bias') )
head.append(('classifier.weight', 'head.weight') )
head.append(('classifier.bias', 'head.bias') )
return head
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
lowerCamelCase__ : Tuple = 'imagenet-1k-id2label.json'
lowerCamelCase__ : Union[str, Any] = 1000
lowerCamelCase__ : Optional[Any] = 'huggingface/label-files'
lowerCamelCase__ : Any = num_labels
lowerCamelCase__ : Dict = json.load(open(cached_download(hf_hub_url(_UpperCAmelCase , _UpperCAmelCase , repo_type='dataset' ) ) , 'r' ) )
lowerCamelCase__ : int = {int(_UpperCAmelCase ): v for k, v in idalabel.items()}
lowerCamelCase__ : Tuple = idalabel
lowerCamelCase__ : List[Any] = {v: k for k, v in idalabel.items()}
lowerCamelCase__ : List[str] = CvtConfig(num_labels=_UpperCAmelCase , idalabel=_UpperCAmelCase , labelaid=_UpperCAmelCase )
# For depth size 13 (13 = 1+2+10)
if cvt_model.rsplit('/' , 1 )[-1][4:6] == "13":
lowerCamelCase__ : List[Any] = [1, 2, 10]
# For depth size 21 (21 = 1+4+16)
elif cvt_model.rsplit('/' , 1 )[-1][4:6] == "21":
lowerCamelCase__ : Dict = [1, 4, 16]
# For wide cvt (similar to wide-resnet) depth size 24 (w24 = 2 + 2 20)
else:
lowerCamelCase__ : Optional[Any] = [2, 2, 20]
lowerCamelCase__ : Optional[int] = [3, 12, 16]
lowerCamelCase__ : str = [192, 768, 1024]
lowerCamelCase__ : Any = CvtForImageClassification(_UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = AutoImageProcessor.from_pretrained('facebook/convnext-base-224-22k-1k' )
lowerCamelCase__ : Tuple = image_size
lowerCamelCase__ : List[str] = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
lowerCamelCase__ : Optional[int] = OrderedDict()
lowerCamelCase__ : Tuple = []
for idx in range(len(config.depth ) ):
if config.cls_token[idx]:
lowerCamelCase__ : Optional[Any] = list_of_state_dict + cls_token(_UpperCAmelCase )
lowerCamelCase__ : str = list_of_state_dict + embeddings(_UpperCAmelCase )
for cnt in range(config.depth[idx] ):
lowerCamelCase__ : str = list_of_state_dict + attention(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : int = list_of_state_dict + final()
for gg in list_of_state_dict:
print(_UpperCAmelCase )
for i in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : str = original_weights[list_of_state_dict[i][1]]
model.load_state_dict(_UpperCAmelCase )
model.save_pretrained(_UpperCAmelCase )
image_processor.save_pretrained(_UpperCAmelCase )
# Download the weights from zoo: https://1drv.ms/u/s!AhIXJn_J-blW9RzF3rMW7SsLHa8h?e=blQ0Al
if __name__ == "__main__":
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
parser.add_argument(
"""--cvt_model""",
default="""cvt-w24""",
type=str,
help="""Name of the cvt model you'd like to convert.""",
)
parser.add_argument(
"""--image_size""",
default=3_84,
type=int,
help="""Input Image Size""",
)
parser.add_argument(
"""--cvt_file_name""",
default=R"""cvtmodels\CvT-w24-384x384-IN-22k.pth""",
type=str,
help="""Input Image Size""",
)
parser.add_argument(
"""--pytorch_dump_folder_path""", default=None, type=str, help="""Path to the output PyTorch model directory."""
)
_UpperCAmelCase : List[str] = parser.parse_args()
convert_cvt_checkpoint(args.cvt_model, args.image_size, args.cvt_file_name, args.pytorch_dump_folder_path)
| 50
| 1
|
from __future__ import annotations
import numpy as np
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[Any]:
return np.maximum(0 , _UpperCAmelCase )
if __name__ == "__main__":
print(np.array(relu([-1, 0, 5]))) # --> [0, 0, 5]
| 50
|
# Copyright 2021 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.
import argparse
import os
from accelerate.test_utils import execute_subprocess_async
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase=None ) -> Tuple:
if subparsers is not None:
lowerCamelCase__ : Any = subparsers.add_parser('test' )
else:
lowerCamelCase__ : int = argparse.ArgumentParser('Accelerate test command' )
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 SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = os.path.sep.join(__file__.split(os.path.sep )[:-2] + ['test_utils', 'scripts', 'test_script.py'] )
if args.config_file is None:
lowerCamelCase__ : List[str] = script_name
else:
lowerCamelCase__ : List[Any] = F"""--config_file={args.config_file} {script_name}"""
lowerCamelCase__ : str = ['accelerate-launch'] + test_args.split()
lowerCamelCase__ : Dict = execute_subprocess_async(_UpperCAmelCase , env=os.environ.copy() )
if result.returncode == 0:
print('Test is a success! You are ready for your distributed training!' )
def SCREAMING_SNAKE_CASE ( ) -> Any:
lowerCamelCase__ : Any = test_command_parser()
lowerCamelCase__ : List[Any] = parser.parse_args()
test_command(_UpperCAmelCase )
if __name__ == "__main__":
main()
| 50
| 1
|
from typing import List, Optional, Tuple, Union
import torch
from ...models import UNetaDModel
from ...schedulers import KarrasVeScheduler
from ...utils import randn_tensor
from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = 42
def __init__( self : Any , UpperCAmelCase : UNetaDModel , UpperCAmelCase : KarrasVeScheduler ) -> List[str]:
super().__init__()
self.register_modules(unet=UpperCAmelCase , scheduler=UpperCAmelCase )
@torch.no_grad()
def __call__( self : List[Any] , UpperCAmelCase : int = 1 , UpperCAmelCase : int = 50 , UpperCAmelCase : Optional[Union[torch.Generator, List[torch.Generator]]] = None , UpperCAmelCase : Optional[str] = "pil" , UpperCAmelCase : bool = True , **UpperCAmelCase : Tuple , ) -> Union[Tuple, ImagePipelineOutput]:
lowerCamelCase__ : Any = self.unet.config.sample_size
lowerCamelCase__ : Any = (batch_size, 3, img_size, img_size)
lowerCamelCase__ : Dict = self.unet
# sample x_0 ~ N(0, sigma_0^2 * I)
lowerCamelCase__ : Dict = randn_tensor(UpperCAmelCase , generator=UpperCAmelCase , device=self.device ) * self.scheduler.init_noise_sigma
self.scheduler.set_timesteps(UpperCAmelCase )
for t in self.progress_bar(self.scheduler.timesteps ):
# here sigma_t == t_i from the paper
lowerCamelCase__ : List[Any] = self.scheduler.schedule[t]
lowerCamelCase__ : Any = self.scheduler.schedule[t - 1] if t > 0 else 0
# 1. Select temporarily increased noise level sigma_hat
# 2. Add new noise to move from sample_i to sample_hat
lowerCamelCase__ , lowerCamelCase__ : Any = self.scheduler.add_noise_to_input(UpperCAmelCase , UpperCAmelCase , generator=UpperCAmelCase )
# 3. Predict the noise residual given the noise magnitude `sigma_hat`
# The model inputs and output are adjusted by following eq. (213) in [1].
lowerCamelCase__ : Optional[Any] = (sigma_hat / 2) * model((sample_hat + 1) / 2 , sigma_hat / 2 ).sample
# 4. Evaluate dx/dt at sigma_hat
# 5. Take Euler step from sigma to sigma_prev
lowerCamelCase__ : Dict = self.scheduler.step(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
if sigma_prev != 0:
# 6. Apply 2nd order correction
# The model inputs and output are adjusted by following eq. (213) in [1].
lowerCamelCase__ : Optional[Any] = (sigma_prev / 2) * model((step_output.prev_sample + 1) / 2 , sigma_prev / 2 ).sample
lowerCamelCase__ : Optional[int] = self.scheduler.step_correct(
UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , step_output.prev_sample , step_output['derivative'] , )
lowerCamelCase__ : Union[str, Any] = step_output.prev_sample
lowerCamelCase__ : Optional[int] = (sample / 2 + 0.5).clamp(0 , 1 )
lowerCamelCase__ : Union[str, Any] = sample.cpu().permute(0 , 2 , 3 , 1 ).numpy()
if output_type == "pil":
lowerCamelCase__ : List[Any] = self.numpy_to_pil(UpperCAmelCase )
if not return_dict:
return (image,)
return ImagePipelineOutput(images=UpperCAmelCase )
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 100_0000 ) -> int:
lowerCamelCase__ : int = limit + 1
lowerCamelCase__ : Optional[Any] = [0] * limit
for first_term in range(1 , _UpperCAmelCase ):
for n in range(_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Optional[Any] = first_term + n / first_term
if common_difference % 4: # d must be divisble by 4
continue
else:
common_difference /= 4
if (
first_term > common_difference
and first_term < 4 * common_difference
): # since x,y,z are positive integers
frequency[n] += 1 # so z>0 and a>d ,also 4d<a
lowerCamelCase__ : List[str] = sum(1 for x in frequency[1:limit] if x == 10 )
return count
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase=False ) -> Union[str, Any]:
if isinstance(_UpperCAmelCase , _UpperCAmelCase ) and isinstance(_UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Tuple = len(set_a.intersection(_UpperCAmelCase ) )
if alternative_union:
lowerCamelCase__ : Union[str, Any] = len(_UpperCAmelCase ) + len(_UpperCAmelCase )
else:
lowerCamelCase__ : Any = len(set_a.union(_UpperCAmelCase ) )
return intersection / union
if isinstance(_UpperCAmelCase , (list, tuple) ) and isinstance(_UpperCAmelCase , (list, tuple) ):
lowerCamelCase__ : int = [element for element in set_a if element in set_b]
if alternative_union:
lowerCamelCase__ : int = len(_UpperCAmelCase ) + len(_UpperCAmelCase )
return len(_UpperCAmelCase ) / union
else:
lowerCamelCase__ : Dict = set_a + [element for element in set_b if element not in set_a]
return len(_UpperCAmelCase ) / len(_UpperCAmelCase )
return len(_UpperCAmelCase ) / len(_UpperCAmelCase )
return None
if __name__ == "__main__":
_UpperCAmelCase : Optional[Any] = {"""a""", """b""", """c""", """d""", """e"""}
_UpperCAmelCase : Tuple = {"""c""", """d""", """e""", """f""", """h""", """i"""}
print(jaccard_similarity(set_a, set_b))
| 50
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_base import BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import PaddingStrategy, logging
from .tokenization_realm import RealmTokenizer
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
_UpperCAmelCase : Any = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : int = {
"""vocab_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/vocab.txt"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/vocab.txt"""
),
"""google/realm-orqa-nq-openqa""": """https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-nq-reader""": """https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-openqa""": """https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/vocab.txt""",
"""google/realm-orqa-wq-reader""": """https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""google/realm-cc-news-pretrained-embedder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/tokenizer.jsont"""
),
"""google/realm-cc-news-pretrained-encoder""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-scorer""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/tokenizer.json"""
),
"""google/realm-cc-news-pretrained-openqa""": (
"""https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-openqa""": (
"""https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-nq-reader""": (
"""https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-openqa""": (
"""https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/tokenizer.json"""
),
"""google/realm-orqa-wq-reader""": (
"""https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : int = {
"""google/realm-cc-news-pretrained-embedder""": 5_12,
"""google/realm-cc-news-pretrained-encoder""": 5_12,
"""google/realm-cc-news-pretrained-scorer""": 5_12,
"""google/realm-cc-news-pretrained-openqa""": 5_12,
"""google/realm-orqa-nq-openqa""": 5_12,
"""google/realm-orqa-nq-reader""": 5_12,
"""google/realm-orqa-wq-openqa""": 5_12,
"""google/realm-orqa-wq-reader""": 5_12,
}
_UpperCAmelCase : Any = {
"""google/realm-cc-news-pretrained-embedder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-encoder""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-scorer""": {"""do_lower_case""": True},
"""google/realm-cc-news-pretrained-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-nq-reader""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-openqa""": {"""do_lower_case""": True},
"""google/realm-orqa-wq-reader""": {"""do_lower_case""": True},
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = PRETRAINED_INIT_CONFIGURATION
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = RealmTokenizer
def __init__( self : Optional[int] , UpperCAmelCase : Tuple=None , UpperCAmelCase : Any=None , UpperCAmelCase : List[Any]=True , UpperCAmelCase : Optional[Any]="[UNK]" , UpperCAmelCase : Any="[SEP]" , UpperCAmelCase : Tuple="[PAD]" , UpperCAmelCase : List[Any]="[CLS]" , UpperCAmelCase : Union[str, Any]="[MASK]" , UpperCAmelCase : Optional[Any]=True , UpperCAmelCase : Any=None , **UpperCAmelCase : Optional[int] , ) -> str:
super().__init__(
UpperCAmelCase , tokenizer_file=UpperCAmelCase , do_lower_case=UpperCAmelCase , unk_token=UpperCAmelCase , sep_token=UpperCAmelCase , pad_token=UpperCAmelCase , cls_token=UpperCAmelCase , mask_token=UpperCAmelCase , tokenize_chinese_chars=UpperCAmelCase , strip_accents=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get('lowercase' , UpperCAmelCase ) != do_lower_case
or normalizer_state.get('strip_accents' , UpperCAmelCase ) != strip_accents
or normalizer_state.get('handle_chinese_chars' , UpperCAmelCase ) != tokenize_chinese_chars
):
lowerCamelCase__ : Optional[int] = getattr(UpperCAmelCase , normalizer_state.pop('type' ) )
lowerCamelCase__ : Optional[Any] = do_lower_case
lowerCamelCase__ : str = strip_accents
lowerCamelCase__ : Optional[Any] = tokenize_chinese_chars
lowerCamelCase__ : int = normalizer_class(**UpperCAmelCase )
lowerCamelCase__ : str = do_lower_case
def A_ ( self : Optional[int] , UpperCAmelCase : int , **UpperCAmelCase : int ) -> List[Any]:
lowerCamelCase__ : List[Any] = PaddingStrategy.MAX_LENGTH
lowerCamelCase__ : Optional[int] = text
lowerCamelCase__ : Dict = kwargs.pop('text_pair' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = kwargs.pop('return_tensors' , UpperCAmelCase )
lowerCamelCase__ : List[Any] = {
'input_ids': [],
'attention_mask': [],
'token_type_ids': [],
}
for idx, candidate_text in enumerate(UpperCAmelCase ):
if batch_text_pair is not None:
lowerCamelCase__ : Tuple = batch_text_pair[idx]
else:
lowerCamelCase__ : Dict = None
lowerCamelCase__ : Optional[int] = super().__call__(UpperCAmelCase , UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Any = encoded_candidates.get('input_ids' )
lowerCamelCase__ : Union[str, Any] = encoded_candidates.get('attention_mask' )
lowerCamelCase__ : Tuple = encoded_candidates.get('token_type_ids' )
if encoded_input_ids is not None:
output_data["input_ids"].append(UpperCAmelCase )
if encoded_attention_mask is not None:
output_data["attention_mask"].append(UpperCAmelCase )
if encoded_token_type_ids is not None:
output_data["token_type_ids"].append(UpperCAmelCase )
lowerCamelCase__ : int = {key: item for key, item in output_data.items() if len(UpperCAmelCase ) != 0}
return BatchEncoding(UpperCAmelCase , tensor_type=UpperCAmelCase )
def A_ ( self : int , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Optional[int]=None ) -> List[str]:
lowerCamelCase__ : Tuple = [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 A_ ( self : Tuple , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : List[Any] = [self.sep_token_id]
lowerCamelCase__ : Optional[int] = [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 A_ ( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
lowerCamelCase__ : int = self._tokenizer.model.save(UpperCAmelCase , name=UpperCAmelCase )
return tuple(UpperCAmelCase )
| 50
| 1
|
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
_UpperCAmelCase : Dict = logging.get_logger(__name__)
_UpperCAmelCase : Dict = {
"""google/mobilenet_v2_1.4_224""": """https://huggingface.co/google/mobilenet_v2_1.4_224/resolve/main/config.json""",
"""google/mobilenet_v2_1.0_224""": """https://huggingface.co/google/mobilenet_v2_1.0_224/resolve/main/config.json""",
"""google/mobilenet_v2_0.75_160""": """https://huggingface.co/google/mobilenet_v2_0.75_160/resolve/main/config.json""",
"""google/mobilenet_v2_0.35_96""": """https://huggingface.co/google/mobilenet_v2_0.35_96/resolve/main/config.json""",
# See all MobileNetV2 models at https://huggingface.co/models?filter=mobilenet_v2
}
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """mobilenet_v2"""
def __init__( self : Optional[int] , UpperCAmelCase : Optional[Any]=3 , UpperCAmelCase : Dict=224 , UpperCAmelCase : str=1.0 , UpperCAmelCase : Union[str, Any]=8 , UpperCAmelCase : List[Any]=8 , UpperCAmelCase : Union[str, Any]=6 , UpperCAmelCase : str=32 , UpperCAmelCase : int=True , UpperCAmelCase : Any=True , UpperCAmelCase : Union[str, Any]="relu6" , UpperCAmelCase : int=True , UpperCAmelCase : List[str]=0.8 , UpperCAmelCase : int=0.0_2 , UpperCAmelCase : Union[str, Any]=0.0_0_1 , UpperCAmelCase : Optional[int]=255 , **UpperCAmelCase : Optional[int] , ) -> Optional[int]:
super().__init__(**UpperCAmelCase )
if depth_multiplier <= 0:
raise ValueError('depth_multiplier must be greater than zero.' )
lowerCamelCase__ : Union[str, Any] = num_channels
lowerCamelCase__ : Union[str, Any] = image_size
lowerCamelCase__ : int = depth_multiplier
lowerCamelCase__ : Optional[int] = depth_divisible_by
lowerCamelCase__ : str = min_depth
lowerCamelCase__ : Optional[Any] = expand_ratio
lowerCamelCase__ : str = output_stride
lowerCamelCase__ : Optional[Any] = first_layer_is_expansion
lowerCamelCase__ : Optional[Any] = finegrained_output
lowerCamelCase__ : Optional[int] = hidden_act
lowerCamelCase__ : Optional[int] = tf_padding
lowerCamelCase__ : Dict = classifier_dropout_prob
lowerCamelCase__ : Union[str, Any] = initializer_range
lowerCamelCase__ : Union[str, Any] = layer_norm_eps
lowerCamelCase__ : Union[str, Any] = semantic_loss_ignore_index
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = version.parse("""1.11""" )
@property
def A_ ( self : str ) -> Mapping[str, Mapping[int, str]]:
return OrderedDict([('pixel_values', {0: 'batch'})] )
@property
def A_ ( self : List[str] ) -> Mapping[str, Mapping[int, str]]:
if self.task == "image-classification":
return OrderedDict([('logits', {0: 'batch'})] )
else:
return OrderedDict([('last_hidden_state', {0: 'batch'}), ('pooler_output', {0: 'batch'})] )
@property
def A_ ( self : str ) -> float:
return 1e-4
| 50
|
import logging
import os
from dataclasses import dataclass
from typing import List, Optional, Union
import tqdm
from filelock import FileLock
from transformers import (
BartTokenizer,
BartTokenizerFast,
DataProcessor,
PreTrainedTokenizer,
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
is_tf_available,
is_torch_available,
)
_UpperCAmelCase : Optional[Any] = logging.getLogger(__name__)
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
@dataclass(frozen=__UpperCamelCase )
class lowerCAmelCase :
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = None
if is_torch_available():
import torch
from torch.utils.data import Dataset
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = 42
def __init__( self : int , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = None , UpperCAmelCase : List[str]=False , UpperCAmelCase : bool = False , ) -> List[str]:
lowerCamelCase__ : int = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = os.path.join(
UpperCAmelCase , 'cached_{}_{}_{}_{}'.format(
'dev' if evaluate else 'train' , tokenizer.__class__.__name__ , str(UpperCAmelCase ) , UpperCAmelCase , ) , )
lowerCamelCase__ : int = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = label_list[2], label_list[1]
lowerCamelCase__ : List[str] = label_list
# Make sure only the first process in distributed training processes the dataset,
# and the others will use the cache.
lowerCamelCase__ : str = cached_features_file + '.lock'
with FileLock(UpperCAmelCase ):
if os.path.exists(UpperCAmelCase ) and not overwrite_cache:
logger.info(F"""Loading features from cached file {cached_features_file}""" )
lowerCamelCase__ : int = torch.load(UpperCAmelCase )
else:
logger.info(F"""Creating features from dataset file at {data_dir}""" )
lowerCamelCase__ : str = (
processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
)
logger.info('Training examples: %s' , len(UpperCAmelCase ) )
lowerCamelCase__ : Dict = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
logger.info('Saving features into cached file %s' , UpperCAmelCase )
torch.save(self.features , UpperCAmelCase )
def __len__( self : Optional[int] ) -> Optional[Any]:
return len(self.features )
def __getitem__( self : Tuple , UpperCAmelCase : Dict ) -> InputFeatures:
return self.features[i]
def A_ ( self : int ) -> int:
return self.label_list
if is_tf_available():
import tensorflow as tf
class lowerCAmelCase :
UpperCAmelCase__ = 42
def __init__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : PreTrainedTokenizer , UpperCAmelCase : str , UpperCAmelCase : Optional[int] = 128 , UpperCAmelCase : Any=False , UpperCAmelCase : bool = False , ) -> Union[str, Any]:
lowerCamelCase__ : Any = hans_processors[task]()
lowerCamelCase__ : Optional[Any] = processor.get_labels()
if tokenizer.__class__ in (
RobertaTokenizer,
RobertaTokenizerFast,
XLMRobertaTokenizer,
BartTokenizer,
BartTokenizerFast,
):
# HACK(label indices are swapped in RoBERTa pretrained model)
lowerCamelCase__ , lowerCamelCase__ : str = label_list[2], label_list[1]
lowerCamelCase__ : Optional[int] = label_list
lowerCamelCase__ : int = processor.get_dev_examples(UpperCAmelCase ) if evaluate else processor.get_train_examples(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = hans_convert_examples_to_features(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
def gen():
for ex_index, ex in tqdm.tqdm(enumerate(self.features ) , desc='convert examples to features' ):
if ex_index % 10000 == 0:
logger.info('Writing example %d of %d' % (ex_index, len(UpperCAmelCase )) )
yield (
{
"example_id": 0,
"input_ids": ex.input_ids,
"attention_mask": ex.attention_mask,
"token_type_ids": ex.token_type_ids,
},
ex.label,
)
lowerCamelCase__ : Optional[int] = tf.data.Dataset.from_generator(
UpperCAmelCase , (
{
'example_id': tf.intaa,
'input_ids': tf.intaa,
'attention_mask': tf.intaa,
'token_type_ids': tf.intaa,
},
tf.intaa,
) , (
{
'example_id': tf.TensorShape([] ),
'input_ids': tf.TensorShape([None, None] ),
'attention_mask': tf.TensorShape([None, None] ),
'token_type_ids': tf.TensorShape([None, None] ),
},
tf.TensorShape([] ),
) , )
def A_ ( self : Any ) -> Any:
return self.dataset
def __len__( self : Tuple ) -> int:
return len(self.features )
def __getitem__( self : List[str] , UpperCAmelCase : Any ) -> InputFeatures:
return self.features[i]
def A_ ( self : Dict ) -> str:
return self.label_list
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : int , UpperCAmelCase : List[Any] ) -> int:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_train_set.txt' ) ) , 'train' )
def A_ ( self : Any , UpperCAmelCase : int ) -> List[Any]:
return self._create_examples(self._read_tsv(os.path.join(UpperCAmelCase , 'heuristics_evaluation_set.txt' ) ) , 'dev' )
def A_ ( self : Any ) -> List[Any]:
return ["contradiction", "entailment", "neutral"]
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : List[str] ) -> List[str]:
lowerCamelCase__ : List[str] = []
for i, line in enumerate(UpperCAmelCase ):
if i == 0:
continue
lowerCamelCase__ : Tuple = '%s-%s' % (set_type, line[0])
lowerCamelCase__ : str = line[5]
lowerCamelCase__ : Dict = line[6]
lowerCamelCase__ : int = line[7][2:] if line[7].startswith('ex' ) else line[7]
lowerCamelCase__ : Dict = line[0]
examples.append(InputExample(guid=UpperCAmelCase , text_a=UpperCAmelCase , text_b=UpperCAmelCase , label=UpperCAmelCase , pairID=UpperCAmelCase ) )
return examples
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ) -> Optional[int]:
lowerCamelCase__ : int = {label: i for i, label in enumerate(_UpperCAmelCase )}
lowerCamelCase__ : List[Any] = []
for ex_index, example in tqdm.tqdm(enumerate(_UpperCAmelCase ) , desc='convert examples to features' ):
if ex_index % 1_0000 == 0:
logger.info('Writing example %d' % (ex_index) )
lowerCamelCase__ : List[Any] = tokenizer(
example.text_a , example.text_b , add_special_tokens=_UpperCAmelCase , max_length=_UpperCAmelCase , padding='max_length' , truncation=_UpperCAmelCase , return_overflowing_tokens=_UpperCAmelCase , )
lowerCamelCase__ : List[str] = label_map[example.label] if example.label in label_map else 0
lowerCamelCase__ : Optional[int] = int(example.pairID )
features.append(InputFeatures(**_UpperCAmelCase , label=_UpperCAmelCase , pairID=_UpperCAmelCase ) )
for i, example in enumerate(examples[:5] ):
logger.info('*** Example ***' )
logger.info(F"""guid: {example}""" )
logger.info(F"""features: {features[i]}""" )
return features
_UpperCAmelCase : str = {
"""hans""": 3,
}
_UpperCAmelCase : List[Any] = {
"""hans""": HansProcessor,
}
| 50
| 1
|
import tempfile
import unittest
from transformers import TaConfig, is_torch_available
from transformers.testing_utils import (
require_sentencepiece,
require_tokenizers,
require_torch,
slow,
torch_device,
)
from ...generation.test_utils import GenerationTesterMixin
from ...test_modeling_common import ModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import AutoTokenizer, UMTaForConditionalGeneration, UMTaForQuestionAnswering, UMTaModel
class lowerCAmelCase :
def __init__( self : Tuple , UpperCAmelCase : Dict , UpperCAmelCase : Union[str, Any]=99 , UpperCAmelCase : str=13 , UpperCAmelCase : List[str]=7 , UpperCAmelCase : str=9 , UpperCAmelCase : Optional[int]=True , UpperCAmelCase : str=True , UpperCAmelCase : Any=False , UpperCAmelCase : Union[str, Any]=32 , UpperCAmelCase : List[str]=5 , UpperCAmelCase : Tuple=4 , UpperCAmelCase : Union[str, Any]=37 , UpperCAmelCase : int=8 , UpperCAmelCase : List[str]=0.1 , UpperCAmelCase : Any=0.0_0_2 , UpperCAmelCase : Optional[Any]=1 , UpperCAmelCase : List[Any]=0 , UpperCAmelCase : Union[str, Any]=0 , UpperCAmelCase : Tuple=None , UpperCAmelCase : Optional[Any]=None , ) -> Union[str, Any]:
lowerCamelCase__ : int = parent
lowerCamelCase__ : Any = batch_size
lowerCamelCase__ : Optional[int] = encoder_seq_length
lowerCamelCase__ : int = decoder_seq_length
# For common tests
lowerCamelCase__ : List[str] = self.decoder_seq_length
lowerCamelCase__ : Optional[int] = is_training
lowerCamelCase__ : List[Any] = use_attention_mask
lowerCamelCase__ : Optional[Any] = use_labels
lowerCamelCase__ : Union[str, Any] = vocab_size
lowerCamelCase__ : Union[str, Any] = hidden_size
lowerCamelCase__ : Optional[Any] = num_hidden_layers
lowerCamelCase__ : Any = num_attention_heads
lowerCamelCase__ : str = d_ff
lowerCamelCase__ : Optional[Any] = relative_attention_num_buckets
lowerCamelCase__ : Any = dropout_rate
lowerCamelCase__ : Any = initializer_factor
lowerCamelCase__ : Union[str, Any] = eos_token_id
lowerCamelCase__ : List[str] = pad_token_id
lowerCamelCase__ : List[str] = decoder_start_token_id
lowerCamelCase__ : List[Any] = None
lowerCamelCase__ : Optional[Any] = decoder_layers
def A_ ( self : List[Any] ) -> int:
return TaConfig.from_pretrained('google/umt5-base' )
def A_ ( self : List[Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[int] , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : Tuple=None , UpperCAmelCase : List[str]=None , UpperCAmelCase : str=None , UpperCAmelCase : Optional[Any]=None , UpperCAmelCase : Optional[Any]=None , ) -> List[str]:
if attention_mask is None:
lowerCamelCase__ : Optional[Any] = input_ids.ne(config.pad_token_id )
if decoder_attention_mask is None:
lowerCamelCase__ : Optional[Any] = decoder_input_ids.ne(config.pad_token_id )
if head_mask is None:
lowerCamelCase__ : int = torch.ones(config.num_hidden_layers , config.num_attention_heads , device=UpperCAmelCase )
if decoder_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
if cross_attn_head_mask is None:
lowerCamelCase__ : Dict = torch.ones(
config.num_decoder_layers , config.num_attention_heads , device=UpperCAmelCase )
return {
"input_ids": input_ids,
"decoder_input_ids": decoder_input_ids,
"attention_mask": attention_mask,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
"cross_attn_head_mask": cross_attn_head_mask,
}
def A_ ( self : str ) -> List[str]:
lowerCamelCase__ : Any = ids_tensor([self.batch_size, self.encoder_seq_length] , self.vocab_size )
lowerCamelCase__ : Optional[Any] = ids_tensor([self.batch_size, self.decoder_seq_length] , self.vocab_size )
# we need to clamp the input ids here to avoid having pad token in between
# this is because for NllbMoe the position_ids are prepared such that
# all pad tokens have pos id = 2 and rest are between 2..seq_length
# and the seq_length here is seq_length - num_pad_tokens
# but when using past, there is no way of knowing if the past input ids had
# pad tokens in them, which results in incorrect seq_lenth and which in turn results in
# position_ids being off by num_pad_tokens in past input
lowerCamelCase__ : List[str] = input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Union[str, Any] = decoder_input_ids.clamp(self.pad_token_id + 1 )
lowerCamelCase__ : Dict = self.get_config()
lowerCamelCase__ : Tuple = config.num_attention_heads
lowerCamelCase__ : Any = self.prepare_inputs_dict(UpperCAmelCase , UpperCAmelCase , UpperCAmelCase )
return config, input_dict
def A_ ( self : Tuple ) -> Union[str, Any]:
lowerCamelCase__ , lowerCamelCase__ : Dict = self.prepare_config_and_inputs()
return config, inputs_dict
def A_ ( self : Optional[int] ) -> List[str]:
return TaConfig(
vocab_size=166 , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Union[str, Any] ) -> Dict:
return TaConfig(
vocab_size=self.vocab_size , d_model=self.hidden_size , d_ff=self.d_ff , d_kv=self.hidden_size // self.num_attention_heads , num_layers=self.num_hidden_layers , num_decoder_layers=self.decoder_layers , num_heads=self.num_attention_heads , relative_attention_num_buckets=self.relative_attention_num_buckets , dropout_rate=self.dropout_rate , initializer_factor=self.initializer_factor , eos_token_id=self.eos_token_id , bos_token_id=self.pad_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.decoder_start_token_id , )
def A_ ( self : Any , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[str] , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Any , UpperCAmelCase : Dict , ) -> str:
lowerCamelCase__ : Dict = UMTaModel(config=UpperCAmelCase )
model.to(UpperCAmelCase )
model.eval()
lowerCamelCase__ : Optional[int] = model(
input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase , decoder_attention_mask=UpperCAmelCase , )
lowerCamelCase__ : Any = model(input_ids=UpperCAmelCase , decoder_input_ids=UpperCAmelCase )
lowerCamelCase__ : Dict = result.last_hidden_state
lowerCamelCase__ : Any = result.past_key_values
lowerCamelCase__ : List[Any] = result.encoder_last_hidden_state
self.parent.assertEqual(encoder_output.size() , (self.batch_size, self.encoder_seq_length, self.hidden_size) )
self.parent.assertEqual(decoder_output.size() , (self.batch_size, self.decoder_seq_length, self.hidden_size) )
# There should be `num_layers` key value embeddings stored in decoder_past
self.parent.assertEqual(len(UpperCAmelCase ) , config.num_layers )
# There should be a self attn key, a self attn value, a cross attn key and a cross attn value stored in each decoder_past tuple
self.parent.assertEqual(len(decoder_past[0] ) , 4 )
def A_ ( self : Optional[int] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : List[Any] , UpperCAmelCase : int , UpperCAmelCase : List[str] , UpperCAmelCase : Optional[Any] , ) -> Optional[int]:
lowerCamelCase__ : List[Any] = UMTaModel(config=UpperCAmelCase ).get_decoder().to(UpperCAmelCase ).eval()
# first forward pass
lowerCamelCase__ : Tuple = model(UpperCAmelCase , use_cache=UpperCAmelCase )
lowerCamelCase__ : List[Any] = model(UpperCAmelCase )
lowerCamelCase__ : int = model(UpperCAmelCase , use_cache=UpperCAmelCase )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) )
self.parent.assertTrue(len(UpperCAmelCase ) == len(UpperCAmelCase ) + 1 )
lowerCamelCase__ , lowerCamelCase__ : Dict = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
lowerCamelCase__ : Optional[int] = ids_tensor((self.batch_size, 1) , config.vocab_size )
# append to next input_ids and
lowerCamelCase__ : List[Any] = torch.cat([input_ids, next_tokens] , dim=-1 )
lowerCamelCase__ : List[str] = model(UpperCAmelCase )['last_hidden_state']
lowerCamelCase__ : str = model(UpperCAmelCase , past_key_values=UpperCAmelCase )['last_hidden_state']
# select random slice
lowerCamelCase__ : List[str] = ids_tensor((1,) , output_from_past.shape[-1] ).item()
lowerCamelCase__ : Tuple = output_from_no_past[:, -1, random_slice_idx].detach()
lowerCamelCase__ : List[str] = output_from_past[:, 0, random_slice_idx].detach()
# test that outputs are equal for slice
self.parent.assertTrue(torch.allclose(UpperCAmelCase , UpperCAmelCase , atol=1e-3 ) )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : List[str] , ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = UMTaModel(config=UpperCAmelCase ).to(UpperCAmelCase ).half().eval()
lowerCamelCase__ : Optional[int] = model(**UpperCAmelCase )['last_hidden_state']
self.parent.assertFalse(torch.isnan(UpperCAmelCase ).any().item() )
@require_torch
class lowerCAmelCase ( __UpperCamelCase, __UpperCamelCase, __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = (
(UMTaModel, UMTaForConditionalGeneration, UMTaForQuestionAnswering) if is_torch_available() else ()
)
UpperCAmelCase__ = (UMTaForConditionalGeneration,) if is_torch_available() else ()
UpperCAmelCase__ = (
{
"""conversational""": UMTaForConditionalGeneration,
"""feature-extraction""": UMTaModel,
"""summarization""": UMTaForConditionalGeneration,
"""text2text-generation""": UMTaForConditionalGeneration,
"""translation""": UMTaForConditionalGeneration,
"""question-answering""": UMTaForQuestionAnswering,
}
if is_torch_available()
else {}
)
UpperCAmelCase__ = True
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = True
UpperCAmelCase__ = True
# The small UMT5 model needs higher percentages for CPU/MP tests
UpperCAmelCase__ = [0.8, 0.9]
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = UMTaModelTester(self )
@unittest.skip('Test has a segmentation fault on torch 1.8.0' )
def A_ ( self : Tuple ) -> int:
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Tuple = UMTaModel(config_and_inputs[0] ).to(UpperCAmelCase )
with tempfile.TemporaryDirectory() as tmpdirname:
torch.onnx.export(
UpperCAmelCase , (config_and_inputs[1], config_and_inputs[3], config_and_inputs[2]) , F"""{tmpdirname}/t5_test.onnx""" , export_params=UpperCAmelCase , opset_version=9 , input_names=['input_ids', 'decoder_input_ids'] , )
@unittest.skipIf(torch_device == 'cpu' , 'Cant do half precision' )
def A_ ( self : Tuple ) -> Optional[Any]:
lowerCamelCase__ : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model_fpaa_forward(*UpperCAmelCase )
def A_ ( self : List[Any] ) -> str:
lowerCamelCase__ : int = ['encoder_attentions', 'decoder_attentions', 'cross_attentions']
lowerCamelCase__ : str = self.model_tester.prepare_config_and_inputs()
lowerCamelCase__ : Any = config_and_inputs[0]
lowerCamelCase__ : Any = UMTaForConditionalGeneration(UpperCAmelCase ).eval()
model.to(UpperCAmelCase )
lowerCamelCase__ : Tuple = {
'head_mask': torch.zeros(config.num_layers , config.num_heads , device=UpperCAmelCase ),
'decoder_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
'cross_attn_head_mask': torch.zeros(config.num_decoder_layers , config.num_heads , device=UpperCAmelCase ),
}
for attn_name, (name, mask) in zip(UpperCAmelCase , head_masking.items() ):
lowerCamelCase__ : Union[str, Any] = {name: mask}
# Explicitly pass decoder_head_mask as it is required from T5 model when head_mask specified
if name == "head_mask":
lowerCamelCase__ : Union[str, Any] = torch.ones(
config.num_decoder_layers , config.num_heads , device=UpperCAmelCase )
lowerCamelCase__ : Tuple = model.generate(
config_and_inputs[1]['input_ids'] , num_beams=1 , max_length=3 , output_attentions=UpperCAmelCase , return_dict_in_generate=UpperCAmelCase , **UpperCAmelCase , )
# We check the state of decoder_attentions and cross_attentions just from the last step
lowerCamelCase__ : Union[str, Any] = out[attn_name] if attn_name == attention_names[0] else out[attn_name][-1]
self.assertEqual(sum([w.sum().item() for w in attn_weights] ) , 0.0 )
@unittest.skip('Does not work on the tiny model as we keep hitting edge cases.' )
def A_ ( self : Optional[Any] ) -> Optional[Any]:
pass
@require_torch
@require_sentencepiece
@require_tokenizers
class lowerCAmelCase ( unittest.TestCase ):
@slow
@unittest.skip(
'Unless we stop stripping left and right by default for all special tokens, the expected ids obtained here will not match the original ones. Wait for https://github.com/huggingface/transformers/pull/23909 to be merged' )
def A_ ( self : Any ) -> int:
lowerCamelCase__ : Optional[Any] = UMTaForConditionalGeneration.from_pretrained('google/umt5-small' , return_dict=UpperCAmelCase ).to(UpperCAmelCase )
lowerCamelCase__ : List[str] = AutoTokenizer.from_pretrained('google/umt5-small' , use_fast=UpperCAmelCase , legacy=UpperCAmelCase )
lowerCamelCase__ : Dict = [
'Bonjour monsieur <extra_id_0> bien <extra_id_1>.',
'No se como puedo <extra_id_0>.',
'This is the reason why we <extra_id_0> them.',
'The <extra_id_0> walks in <extra_id_1>, seats',
'A <extra_id_0> walks into a bar and orders a <extra_id_1> with <extra_id_2> pinch of <extra_id_3>.',
]
lowerCamelCase__ : Tuple = tokenizer(UpperCAmelCase , return_tensors='pt' , padding=UpperCAmelCase ).input_ids
# fmt: off
lowerCamelCase__ : Any = torch.tensor(
[
[ 38530, 210703, 256299, 1410, 256298, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 826, 321, 671, 25922, 256299, 274, 1, 0,0, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 1460, 339, 312, 19014, 10620, 758, 256299, 2355,274, 1, 0, 0, 0, 0, 0, 0,0, 0],
[ 517, 256299, 14869, 281, 301, 256298, 275, 119983,1, 0, 0, 0, 0, 0, 0, 0,0, 0],
[ 320, 256299, 14869, 281, 2234, 289, 2275, 333,61391, 289, 256298, 543, 256297, 168714, 329, 256296,274, 1],
] )
# fmt: on
torch.testing.assert_allclose(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : Optional[int] = model.generate(input_ids.to(UpperCAmelCase ) )
lowerCamelCase__ : List[Any] = [
'<pad><extra_id_0> et<extra_id_1> [eod] <extra_id_2><extra_id_55>.. [eod] 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 💐 <extra_id_56>ajšietosto<extra_id_56>lleux<extra_id_19><extra_id_6>ajšie</s>',
'<pad><extra_id_0>.<extra_id_1>.,<0x0A>...spech <0x0A><extra_id_20> <extra_id_21></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> are not going to be a part of the world. We are not going to be a part of<extra_id_1> and<extra_id_2><0x0A><extra_id_48>.<extra_id_48></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0> door<extra_id_1>, the door<extra_id_2> 피해[/</s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
'<pad><extra_id_0>nyone who<extra_id_1> drink<extra_id_2> a<extra_id_3> alcohol<extra_id_4> A<extra_id_5> A. This<extra_id_6> I<extra_id_7><extra_id_52><extra_id_53></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>',
]
lowerCamelCase__ : Union[str, Any] = tokenizer.batch_decode(UpperCAmelCase )
self.assertEqual(UpperCAmelCase , UpperCAmelCase )
| 50
|
import unittest
from transformers import BertGenerationTokenizer
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
_UpperCAmelCase : Optional[Any] = """▁"""
_UpperCAmelCase : Optional[Any] = get_tests_dir("""fixtures/test_sentencepiece.model""")
@require_sentencepiece
class lowerCAmelCase ( __UpperCamelCase, unittest.TestCase ):
UpperCAmelCase__ = BertGenerationTokenizer
UpperCAmelCase__ = False
UpperCAmelCase__ = True
def A_ ( self : List[Any] ) -> List[str]:
super().setUp()
lowerCamelCase__ : Dict = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
tokenizer.save_pretrained(self.tmpdirname )
def A_ ( self : Optional[Any] ) -> Dict:
lowerCamelCase__ : List[str] = '<s>'
lowerCamelCase__ : Optional[int] = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCAmelCase ) , UpperCAmelCase )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCAmelCase ) , UpperCAmelCase )
def A_ ( self : List[str] ) -> Optional[int]:
lowerCamelCase__ : Optional[Any] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] , '<unk>' )
self.assertEqual(vocab_keys[1] , '<s>' )
self.assertEqual(vocab_keys[-1] , '<pad>' )
self.assertEqual(len(UpperCAmelCase ) , 1002 )
def A_ ( self : List[Any] ) -> Optional[Any]:
self.assertEqual(self.get_tokenizer().vocab_size , 1000 )
def A_ ( self : Union[str, Any] ) -> List[Any]:
lowerCamelCase__ : Union[str, Any] = BertGenerationTokenizer(UpperCAmelCase , keep_accents=UpperCAmelCase )
lowerCamelCase__ : List[str] = tokenizer.tokenize('This is a test' )
self.assertListEqual(UpperCAmelCase , ['▁This', '▁is', '▁a', '▁t', 'est'] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(UpperCAmelCase ) , [285, 46, 10, 170, 382] , )
lowerCamelCase__ : List[str] = tokenizer.tokenize('I was born in 92000, and this is falsé.' )
self.assertListEqual(
UpperCAmelCase , [
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',
'.',
] , )
lowerCamelCase__ : Optional[int] = tokenizer.convert_tokens_to_ids(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
lowerCamelCase__ : Optional[Any] = tokenizer.convert_ids_to_tokens(UpperCAmelCase )
self.assertListEqual(
UpperCAmelCase , [
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>',
'.',
] , )
@cached_property
def A_ ( self : Dict ) -> Tuple:
return BertGenerationTokenizer.from_pretrained('google/bert_for_seq_generation_L-24_bbc_encoder' )
@slow
def A_ ( self : Optional[int] ) -> List[str]:
lowerCamelCase__ : Union[str, Any] = 'Hello World!'
lowerCamelCase__ : Dict = [18536, 2260, 101]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@slow
def A_ ( self : Optional[Any] ) -> str:
lowerCamelCase__ : List[Any] = (
'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'
)
lowerCamelCase__ : Any = [
871,
419,
358,
946,
991,
2521,
452,
358,
1357,
387,
7751,
3536,
112,
985,
456,
126,
865,
938,
5400,
5734,
458,
1368,
467,
786,
2462,
5246,
1159,
633,
865,
4519,
457,
582,
852,
2557,
427,
916,
508,
405,
34324,
497,
391,
408,
11342,
1244,
385,
100,
938,
985,
456,
574,
362,
12597,
3200,
3129,
1172,
]
self.assertListEqual(UpperCAmelCase , self.big_tokenizer.encode(UpperCAmelCase ) )
@require_torch
@slow
def A_ ( self : int ) -> Optional[Any]:
import torch
from transformers import BertGenerationConfig, BertGenerationEncoder
# Build sequence
lowerCamelCase__ : str = list(self.big_tokenizer.get_vocab().keys() )[:10]
lowerCamelCase__ : int = ' '.join(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.big_tokenizer.encode_plus(UpperCAmelCase , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.big_tokenizer.batch_encode_plus(
[sequence + ' ' + sequence] , return_tensors='pt' , return_token_type_ids=UpperCAmelCase )
lowerCamelCase__ : Tuple = BertGenerationConfig()
lowerCamelCase__ : Optional[Any] = BertGenerationEncoder(UpperCAmelCase )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**UpperCAmelCase )
model(**UpperCAmelCase )
@slow
def A_ ( self : Optional[int] ) -> List[Any]:
# fmt: off
lowerCamelCase__ : Any = {'input_ids': [[39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114], [448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 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], [484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 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]], '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, 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, 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=UpperCAmelCase , model_name='google/bert_for_seq_generation_L-24_bbc_encoder' , revision='c817d1fd1be2ffa69431227a1fe320544943d4db' , )
| 50
| 1
|
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers.testing_utils import require_vision
from transformers.utils import is_vision_available
if is_vision_available():
from PIL import Image
from transformers import AutoProcessor, BlipaProcessor, BlipImageProcessor, GPTaTokenizer, PreTrainedTokenizerFast
@require_vision
class lowerCAmelCase ( unittest.TestCase ):
def A_ ( self : Tuple ) -> str:
lowerCamelCase__ : int = tempfile.mkdtemp()
lowerCamelCase__ : Tuple = BlipImageProcessor()
lowerCamelCase__ : str = GPTaTokenizer.from_pretrained('hf-internal-testing/tiny-random-GPT2Model' )
lowerCamelCase__ : List[Any] = BlipaProcessor(UpperCAmelCase , UpperCAmelCase )
processor.save_pretrained(self.tmpdirname )
def A_ ( self : List[str] , **UpperCAmelCase : str ) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **UpperCAmelCase ).tokenizer
def A_ ( self : Optional[int] , **UpperCAmelCase : List[str] ) -> Tuple:
return AutoProcessor.from_pretrained(self.tmpdirname , **UpperCAmelCase ).image_processor
def A_ ( self : Optional[Any] ) -> Optional[Any]:
shutil.rmtree(self.tmpdirname )
def A_ ( self : Tuple ) -> Any:
lowerCamelCase__ : Optional[Any] = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta )]
lowerCamelCase__ : Optional[Any] = [Image.fromarray(np.moveaxis(UpperCAmelCase , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def A_ ( self : Dict ) -> List[str]:
lowerCamelCase__ : Optional[Any] = BlipaProcessor(tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
lowerCamelCase__ : Optional[Any] = self.get_tokenizer(bos_token='(BOS)' , eos_token='(EOS)' )
lowerCamelCase__ : Union[str, Any] = self.get_image_processor(do_normalize=UpperCAmelCase , padding_value=1.0 )
lowerCamelCase__ : str = BlipaProcessor.from_pretrained(
self.tmpdirname , bos_token='(BOS)' , eos_token='(EOS)' , do_normalize=UpperCAmelCase , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , UpperCAmelCase )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , UpperCAmelCase )
def A_ ( self : Tuple ) -> str:
lowerCamelCase__ : Dict = self.get_image_processor()
lowerCamelCase__ : List[Any] = self.get_tokenizer()
lowerCamelCase__ : Any = BlipaProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = self.prepare_image_inputs()
lowerCamelCase__ : List[Any] = image_processor(UpperCAmelCase , return_tensors='np' )
lowerCamelCase__ : List[Any] = processor(images=UpperCAmelCase , 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 A_ ( self : int ) -> Any:
lowerCamelCase__ : Optional[Any] = self.get_image_processor()
lowerCamelCase__ : List[str] = self.get_tokenizer()
lowerCamelCase__ : int = BlipaProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = 'lower newer'
lowerCamelCase__ : Optional[int] = processor(text=UpperCAmelCase )
lowerCamelCase__ : Tuple = tokenizer(UpperCAmelCase , return_token_type_ids=UpperCAmelCase )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def A_ ( self : Tuple ) -> List[Any]:
lowerCamelCase__ : Any = self.get_image_processor()
lowerCamelCase__ : Dict = self.get_tokenizer()
lowerCamelCase__ : Optional[Any] = BlipaProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = 'lower newer'
lowerCamelCase__ : Optional[Any] = self.prepare_image_inputs()
lowerCamelCase__ : Optional[Any] = processor(text=UpperCAmelCase , images=UpperCAmelCase )
self.assertListEqual(list(inputs.keys() ) , ['pixel_values', 'input_ids', 'attention_mask'] )
# test if it raises when no input is passed
with pytest.raises(UpperCAmelCase ):
processor()
def A_ ( self : Dict ) -> Union[str, Any]:
lowerCamelCase__ : Optional[int] = self.get_image_processor()
lowerCamelCase__ : Optional[int] = self.get_tokenizer()
lowerCamelCase__ : str = BlipaProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : int = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
lowerCamelCase__ : str = processor.batch_decode(UpperCAmelCase )
lowerCamelCase__ : str = tokenizer.batch_decode(UpperCAmelCase )
self.assertListEqual(UpperCAmelCase , UpperCAmelCase )
def A_ ( self : Tuple ) -> str:
lowerCamelCase__ : Optional[int] = self.get_image_processor()
lowerCamelCase__ : List[Any] = self.get_tokenizer()
lowerCamelCase__ : List[str] = BlipaProcessor(tokenizer=UpperCAmelCase , image_processor=UpperCAmelCase )
lowerCamelCase__ : Any = 'lower newer'
lowerCamelCase__ : Optional[Any] = self.prepare_image_inputs()
lowerCamelCase__ : Union[str, Any] = processor(text=UpperCAmelCase , images=UpperCAmelCase )
# For now the processor supports only ['pixel_values', 'input_ids', 'attention_mask']
self.assertListEqual(list(inputs.keys() ) , ['pixel_values', 'input_ids', 'attention_mask'] )
| 50
|
import os
import tempfile
from functools import partial
from unittest import TestCase
from unittest.mock import patch
import numpy as np
import pytest
from datasets.arrow_dataset import Dataset
from datasets.search import ElasticSearchIndex, FaissIndex, MissingIndex
from .utils import require_elasticsearch, require_faiss
_UpperCAmelCase : str = pytest.mark.integration
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : List[Any] ) -> Union[str, Any]:
lowerCamelCase__ : List[Any] = Dataset.from_dict({'filename': ['my_name-train' + '_' + str(UpperCAmelCase ) for x in np.arange(30 ).tolist()]} )
return dset
def A_ ( self : Optional[Any] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
lowerCamelCase__ : List[Any] = dset.map(
lambda UpperCAmelCase , UpperCAmelCase : {"vecs": i * np.ones(5 , dtype=np.floataa )} , with_indices=UpperCAmelCase , keep_in_memory=UpperCAmelCase )
lowerCamelCase__ : Tuple = dset.add_faiss_index('vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
dset.drop_index('vecs' )
def A_ ( self : Union[str, Any] ) -> int:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , batch_size=100 , metric_type=faiss.METRIC_INNER_PRODUCT , )
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = dset.get_nearest_examples('vecs' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : List[str] ) -> Tuple:
import faiss
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' , metric_type=faiss.METRIC_INNER_PRODUCT , )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
dset.save_faiss_index('vecs' , tmp_file.name )
dset.load_faiss_index('vecs2' , tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ , lowerCamelCase__ : str = dset.get_nearest_examples('vecs2' , np.ones(5 , dtype=np.floataa ) )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
dset.add_faiss_index_from_external_arrays(
external_arrays=np.ones((30, 5) ) * np.arange(30 ).reshape(-1 , 1 ) , index_name='vecs' )
dset.drop_index('vecs' )
self.assertRaises(UpperCAmelCase , partial(dset.get_nearest_examples , 'vecs2' , np.ones(5 , dtype=np.floataa ) ) )
def A_ ( self : Dict ) -> Dict:
from elasticsearch import Elasticsearch
lowerCamelCase__ : Dataset = self._create_dummy_dataset()
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : List[Any] = {'acknowledged': True}
mocked_bulk.return_value([(True, None)] * 30 )
lowerCamelCase__ : int = {'hits': {'hits': [{'_score': 1, '_id': 29}]}}
lowerCamelCase__ : List[str] = Elasticsearch()
dset.add_elasticsearch_index('filename' , es_client=UpperCAmelCase )
lowerCamelCase__ , lowerCamelCase__ : Dict = dset.get_nearest_examples('filename' , 'my_name-train_29' )
self.assertEqual(examples['filename'][0] , 'my_name-train_29' )
@require_faiss
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Any ) -> Dict:
import faiss
lowerCamelCase__ : Tuple = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
# add vectors
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsNotNone(index.faiss_index )
self.assertEqual(index.faiss_index.ntotal , 5 )
index.add_vectors(np.zeros((5, 5) , dtype=np.floataa ) )
self.assertEqual(index.faiss_index.ntotal , 10 )
# single query
lowerCamelCase__ : int = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Any = 1
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = index.search(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search , query.reshape(-1 , 1 ) )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
# batched queries
lowerCamelCase__ : str = np.eye(5 , dtype=np.floataa )[::-1]
lowerCamelCase__ , lowerCamelCase__ : Dict = index.search_batch(UpperCAmelCase )
self.assertRaises(UpperCAmelCase , index.search_batch , queries[0] )
lowerCamelCase__ : str = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[str] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([4, 3, 2, 1, 0] , UpperCAmelCase )
def A_ ( self : List[Any] ) -> List[Any]:
import faiss
lowerCamelCase__ : Any = FaissIndex(string_factory='Flat' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
lowerCamelCase__ : Tuple = FaissIndex(string_factory='LSH' )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexLSH )
with self.assertRaises(UpperCAmelCase ):
lowerCamelCase__ : List[str] = FaissIndex(string_factory='Flat' , custom_index=faiss.IndexFlat(5 ) )
def A_ ( self : List[str] ) -> Optional[int]:
import faiss
lowerCamelCase__ : Optional[Any] = faiss.IndexFlat(5 )
lowerCamelCase__ : int = FaissIndex(custom_index=UpperCAmelCase )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
self.assertIsInstance(index.faiss_index , faiss.IndexFlat )
def A_ ( self : Any ) -> Optional[int]:
import faiss
lowerCamelCase__ : Any = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
# Setting delete=False and unlinking manually is not pretty... but it is required on Windows to
# ensure somewhat stable behaviour. If we don't, we get PermissionErrors. This is an age-old issue.
# see https://bugs.python.org/issue14243 and
# https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file/23212515
with tempfile.NamedTemporaryFile(delete=UpperCAmelCase ) as tmp_file:
index.save(tmp_file.name )
lowerCamelCase__ : List[Any] = FaissIndex.load(tmp_file.name )
os.unlink(tmp_file.name )
lowerCamelCase__ : List[str] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Tuple = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(UpperCAmelCase )
self.assertGreater(scores[0] , 0 )
self.assertEqual(indices[0] , 1 )
@require_faiss
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Any:
import faiss
lowerCamelCase__ : Optional[Any] = FaissIndex(metric_type=faiss.METRIC_INNER_PRODUCT )
index.add_vectors(np.eye(5 , dtype=np.floataa ) )
lowerCamelCase__ : Optional[int] = 'index.faiss'
lowerCamelCase__ : Optional[Any] = F"""mock://{index_name}"""
index.save(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Tuple = FaissIndex.load(_UpperCAmelCase , storage_options=mockfs.storage_options )
lowerCamelCase__ : Optional[int] = np.zeros(5 , dtype=np.floataa )
lowerCamelCase__ : Dict = 1
lowerCamelCase__ , lowerCamelCase__ : str = index.search(_UpperCAmelCase )
assert scores[0] > 0
assert indices[0] == 1
@require_elasticsearch
class lowerCAmelCase ( __UpperCamelCase ):
def A_ ( self : Dict ) -> List[Any]:
from elasticsearch import Elasticsearch
with patch('elasticsearch.Elasticsearch.search' ) as mocked_search, patch(
'elasticsearch.client.IndicesClient.create' ) as mocked_index_create, patch('elasticsearch.helpers.streaming_bulk' ) as mocked_bulk:
lowerCamelCase__ : Any = Elasticsearch()
lowerCamelCase__ : Tuple = {'acknowledged': True}
lowerCamelCase__ : Tuple = ElasticSearchIndex(es_client=UpperCAmelCase )
mocked_bulk.return_value([(True, None)] * 3 )
index.add_documents(['foo', 'bar', 'foobar'] )
# single query
lowerCamelCase__ : Optional[int] = 'foo'
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search(UpperCAmelCase )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# single query with timeout
lowerCamelCase__ : Any = 'foo'
lowerCamelCase__ : List[str] = {'hits': {'hits': [{'_score': 1, '_id': 0}]}}
lowerCamelCase__ , lowerCamelCase__ : Tuple = index.search(UpperCAmelCase , request_timeout=30 )
self.assertEqual(scores[0] , 1 )
self.assertEqual(indices[0] , 0 )
# batched queries
lowerCamelCase__ : List[str] = ['foo', 'bar', 'foobar']
lowerCamelCase__ : str = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : str = index.search_batch(UpperCAmelCase )
lowerCamelCase__ : List[str] = [scores[0] for scores in total_scores]
lowerCamelCase__ : List[Any] = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
# batched queries with timeout
lowerCamelCase__ : str = ['foo', 'bar', 'foobar']
lowerCamelCase__ : List[Any] = {'hits': {'hits': [{'_score': 1, '_id': 1}]}}
lowerCamelCase__ , lowerCamelCase__ : Optional[Any] = index.search_batch(UpperCAmelCase , request_timeout=30 )
lowerCamelCase__ : Optional[Any] = [scores[0] for scores in total_scores]
lowerCamelCase__ : Dict = [indices[0] for indices in total_indices]
self.assertGreater(np.min(UpperCAmelCase ) , 0 )
self.assertListEqual([1, 1, 1] , UpperCAmelCase )
| 50
| 1
|
from __future__ import annotations
_UpperCAmelCase : str = list[tuple[int, int]]
_UpperCAmelCase : int = [
[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],
]
_UpperCAmelCase : Tuple = ([-1, 0], [0, -1], [1, 0], [0, 1]) # up, left, down, right
class lowerCAmelCase :
def __init__( self : Optional[int] , UpperCAmelCase : int , UpperCAmelCase : int , UpperCAmelCase : int , UpperCAmelCase : int , UpperCAmelCase : float , UpperCAmelCase : Node | None , ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = pos_x
lowerCamelCase__ : List[Any] = pos_y
lowerCamelCase__ : Union[str, Any] = (pos_y, pos_x)
lowerCamelCase__ : Optional[Any] = goal_x
lowerCamelCase__ : Optional[int] = goal_y
lowerCamelCase__ : str = g_cost
lowerCamelCase__ : int = parent
lowerCamelCase__ : Union[str, Any] = self.calculate_heuristic()
def A_ ( self : Optional[Any] ) -> float:
lowerCamelCase__ : Dict = abs(self.pos_x - self.goal_x )
lowerCamelCase__ : int = abs(self.pos_y - self.goal_y )
return dx + dy
def __lt__( self : int , UpperCAmelCase : int ) -> bool:
return self.f_cost < other.f_cost
class lowerCAmelCase :
def __init__( self : List[str] , UpperCAmelCase : tuple[int, int] , UpperCAmelCase : tuple[int, int] ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = Node(start[1] , start[0] , goal[1] , goal[0] , 0 , UpperCAmelCase )
lowerCamelCase__ : Tuple = Node(goal[1] , goal[0] , goal[1] , goal[0] , 99999 , UpperCAmelCase )
lowerCamelCase__ : List[str] = [self.start]
lowerCamelCase__ : list[Node] = []
lowerCamelCase__ : Tuple = False
def A_ ( self : Any ) -> Path | None:
while self.open_nodes:
# Open Nodes are sorted using __lt__
self.open_nodes.sort()
lowerCamelCase__ : Optional[Any] = self.open_nodes.pop(0 )
if current_node.pos == self.target.pos:
lowerCamelCase__ : Optional[int] = True
return self.retrace_path(UpperCAmelCase )
self.closed_nodes.append(UpperCAmelCase )
lowerCamelCase__ : List[Any] = self.get_successors(UpperCAmelCase )
for child_node in successors:
if child_node in self.closed_nodes:
continue
if child_node not in self.open_nodes:
self.open_nodes.append(UpperCAmelCase )
else:
# retrieve the best current path
lowerCamelCase__ : Optional[Any] = self.open_nodes.pop(self.open_nodes.index(UpperCAmelCase ) )
if child_node.g_cost < better_node.g_cost:
self.open_nodes.append(UpperCAmelCase )
else:
self.open_nodes.append(UpperCAmelCase )
if not self.reached:
return [self.start.pos]
return None
def A_ ( self : int , UpperCAmelCase : Node ) -> list[Node]:
lowerCamelCase__ : List[Any] = []
for action in delta:
lowerCamelCase__ : Union[str, Any] = parent.pos_x + action[1]
lowerCamelCase__ : Any = parent.pos_y + action[0]
if not (0 <= pos_x <= len(grid[0] ) - 1 and 0 <= pos_y <= len(UpperCAmelCase ) - 1):
continue
if grid[pos_y][pos_x] != 0:
continue
successors.append(
Node(
UpperCAmelCase , UpperCAmelCase , self.target.pos_y , self.target.pos_x , parent.g_cost + 1 , UpperCAmelCase , ) )
return successors
def A_ ( self : Tuple , UpperCAmelCase : Node | None ) -> Path:
lowerCamelCase__ : Optional[Any] = node
lowerCamelCase__ : Any = []
while current_node is not None:
path.append((current_node.pos_y, current_node.pos_x) )
lowerCamelCase__ : Optional[int] = current_node.parent
path.reverse()
return path
if __name__ == "__main__":
_UpperCAmelCase : Tuple = (0, 0)
_UpperCAmelCase : Tuple = (len(grid) - 1, len(grid[0]) - 1)
for elem in grid:
print(elem)
print("""------""")
_UpperCAmelCase : str = GreedyBestFirst(init, goal)
_UpperCAmelCase : int = greedy_bf.search()
if path:
for pos_x, pos_y in path:
_UpperCAmelCase : int = 2
for elem in grid:
print(elem)
| 50
|
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> bool:
lowerCamelCase__ : List[str] = len(_UpperCAmelCase )
lowerCamelCase__ : str = [[False] * (required_sum + 1) for _ in range(arr_len + 1 )]
# for each arr value, a sum of zero(0) can be formed by not taking any element
# hence True/1
for i in range(arr_len + 1 ):
lowerCamelCase__ : Tuple = True
# sum is not zero and set is empty then false
for i in range(1 , required_sum + 1 ):
lowerCamelCase__ : Dict = False
for i in range(1 , arr_len + 1 ):
for j in range(1 , required_sum + 1 ):
if arr[i - 1] > j:
lowerCamelCase__ : str = subset[i - 1][j]
if arr[i - 1] <= j:
lowerCamelCase__ : Dict = subset[i - 1][j] or subset[i - 1][j - arr[i - 1]]
return subset[arr_len][required_sum]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
| 1
|
# 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 lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = (
"""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."""
)
UpperCAmelCase__ = """CIDAS/clipseg-rd64-refined"""
UpperCAmelCase__ = """image_segmenter"""
UpperCAmelCase__ = CLIPSegForImageSegmentation
UpperCAmelCase__ = ["""image""", """text"""]
UpperCAmelCase__ = ["""image"""]
def __init__( self : Optional[Any] , *UpperCAmelCase : Any , **UpperCAmelCase : Tuple ) -> Union[str, Any]:
requires_backends(self , ['vision'] )
super().__init__(*UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Tuple , UpperCAmelCase : "Image" , UpperCAmelCase : str ) -> str:
return self.pre_processor(text=[label] , images=[image] , padding=UpperCAmelCase , return_tensors='pt' )
def A_ ( self : Union[str, Any] , UpperCAmelCase : Dict ) -> Union[str, Any]:
with torch.no_grad():
lowerCamelCase__ : Dict = self.model(**UpperCAmelCase ).logits
return logits
def A_ ( self : Dict , UpperCAmelCase : Union[str, Any] ) -> Dict:
lowerCamelCase__ : List[Any] = outputs.cpu().detach().numpy()
lowerCamelCase__ : Any = 0
lowerCamelCase__ : Dict = 1
return Image.fromarray((array * 255).astype(np.uinta ) )
| 50
|
import torch
from transformers import PreTrainedModel, XLMRobertaConfig, XLMRobertaModel
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = """M-CLIP"""
def __init__( self : Optional[Any] , UpperCAmelCase : Union[str, Any]=1024 , UpperCAmelCase : Tuple=768 , **UpperCAmelCase : Optional[int] ) -> Dict:
lowerCamelCase__ : Optional[int] = transformerDimSize
lowerCamelCase__ : Optional[Any] = imageDimSize
super().__init__(**UpperCAmelCase )
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = MCLIPConfig
def __init__( self : List[Any] , UpperCAmelCase : Dict , *UpperCAmelCase : Any , **UpperCAmelCase : Dict ) -> Dict:
super().__init__(UpperCAmelCase , *UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Tuple = XLMRobertaModel(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = torch.nn.Linear(
in_features=config.transformerDimensions , out_features=config.numDims )
def A_ ( self : Optional[int] , UpperCAmelCase : Optional[Any] , UpperCAmelCase : Dict ) -> Tuple:
lowerCamelCase__ : Any = self.transformer(input_ids=UpperCAmelCase , attention_mask=UpperCAmelCase )[0]
lowerCamelCase__ : int = (embs * attention_mask.unsqueeze(2 )).sum(dim=1 ) / attention_mask.sum(dim=1 )[:, None]
return self.LinearTransformation(UpperCAmelCase ), embs
| 50
| 1
|
from decimal import Decimal, getcontext
from math import ceil, factorial
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
if not isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise TypeError('Undefined for non-integers' )
elif precision < 1:
raise ValueError('Undefined for non-natural numbers' )
lowerCamelCase__ : Tuple = precision
lowerCamelCase__ : Optional[int] = ceil(precision / 14 )
lowerCamelCase__ : Optional[Any] = 42_6880 * Decimal(1_0005 ).sqrt()
lowerCamelCase__ : int = 1
lowerCamelCase__ : List[str] = 1359_1409
lowerCamelCase__ : Union[str, Any] = Decimal(_UpperCAmelCase )
for k in range(1 , _UpperCAmelCase ):
lowerCamelCase__ : int = factorial(6 * k ) // (factorial(3 * k ) * factorial(_UpperCAmelCase ) ** 3)
linear_term += 5_4514_0134
exponential_term *= -26_2537_4126_4076_8000
partial_sum += Decimal(multinomial_term * linear_term ) / exponential_term
return str(constant_term / partial_sum )[:-1]
if __name__ == "__main__":
_UpperCAmelCase : List[str] = 50
print(F"""The first {n} digits of pi is: {pi(n)}""")
| 50
|
from itertools import count
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase = 50 ) -> int:
lowerCamelCase__ : Optional[Any] = [1] * min_block_length
for n in count(_UpperCAmelCase ):
fill_count_functions.append(1 )
for block_length in range(_UpperCAmelCase , n + 1 ):
for block_start in range(n - block_length ):
fill_count_functions[n] += fill_count_functions[
n - block_start - block_length - 1
]
fill_count_functions[n] += 1
if fill_count_functions[n] > 100_0000:
break
return n
if __name__ == "__main__":
print(F"""{solution() = }""")
| 50
| 1
|
import copy
from dataclasses import dataclass
from pathlib import Path
from typing import Dict, Optional, Union
@dataclass
class lowerCAmelCase :
UpperCAmelCase__ = None
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = None
UpperCAmelCase__ = None
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = False
UpperCAmelCase__ = True
UpperCAmelCase__ = None
UpperCAmelCase__ = 1
UpperCAmelCase__ = None
UpperCAmelCase__ = False
UpperCAmelCase__ = None
UpperCAmelCase__ = None
def A_ ( self : List[str] ) -> "DownloadConfig":
return self.__class__(**{k: copy.deepcopy(UpperCAmelCase ) for k, v in self.__dict__.items()} )
| 50
|
from __future__ import annotations
from typing import Any
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> None:
create_state_space_tree(_UpperCAmelCase , [] , 0 )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> None:
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__":
_UpperCAmelCase : list[Any] = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["""A""", """B""", """C"""])
generate_all_subsequences(seq)
| 50
| 1
|
from math import pow
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ) -> tuple[int, int]:
if current_sum == needed_sum:
# If the sum of the powers is equal to needed_sum, then we have a solution.
solutions_count += 1
return current_sum, solutions_count
lowerCamelCase__ : Union[str, Any] = int(pow(_UpperCAmelCase , _UpperCAmelCase ) )
if current_sum + i_to_n <= needed_sum:
# If the sum of the powers is less than needed_sum, then continue adding powers.
current_sum += i_to_n
lowerCamelCase__ , lowerCamelCase__ : Union[str, Any] = backtrack(
_UpperCAmelCase , _UpperCAmelCase , current_number + 1 , _UpperCAmelCase , _UpperCAmelCase )
current_sum -= i_to_n
if i_to_n < needed_sum:
# If the power of i is less than needed_sum, then try with the next power.
lowerCamelCase__ , lowerCamelCase__ : str = backtrack(
_UpperCAmelCase , _UpperCAmelCase , current_number + 1 , _UpperCAmelCase , _UpperCAmelCase )
return current_sum, solutions_count
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> int:
if not (1 <= needed_sum <= 1000 and 2 <= power <= 10):
raise ValueError(
'Invalid input\n'
'needed_sum must be between 1 and 1000, power between 2 and 10.' )
return backtrack(_UpperCAmelCase , _UpperCAmelCase , 1 , 0 , 0 )[1] # Return the solutions_count
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from tokenizers import processors
from ...tokenization_utils import AddedToken, BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_nllb import NllbTokenizer
else:
_UpperCAmelCase : int = None
_UpperCAmelCase : Dict = logging.get_logger(__name__)
_UpperCAmelCase : Optional[int] = {"""vocab_file""": """sentencepiece.bpe.model""", """tokenizer_file""": """tokenizer.json"""}
_UpperCAmelCase : List[Any] = {
"""vocab_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/sentencepiece.bpe.model"""
),
},
"""tokenizer_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/tokenizer.json"""
),
},
}
_UpperCAmelCase : List[str] = {
"""facebook/nllb-large-en-ro""": 10_24,
"""facebook/nllb-200-distilled-600M""": 10_24,
}
# fmt: off
_UpperCAmelCase : Optional[int] = ["""ace_Arab""", """ace_Latn""", """acm_Arab""", """acq_Arab""", """aeb_Arab""", """afr_Latn""", """ajp_Arab""", """aka_Latn""", """amh_Ethi""", """apc_Arab""", """arb_Arab""", """ars_Arab""", """ary_Arab""", """arz_Arab""", """asm_Beng""", """ast_Latn""", """awa_Deva""", """ayr_Latn""", """azb_Arab""", """azj_Latn""", """bak_Cyrl""", """bam_Latn""", """ban_Latn""", """bel_Cyrl""", """bem_Latn""", """ben_Beng""", """bho_Deva""", """bjn_Arab""", """bjn_Latn""", """bod_Tibt""", """bos_Latn""", """bug_Latn""", """bul_Cyrl""", """cat_Latn""", """ceb_Latn""", """ces_Latn""", """cjk_Latn""", """ckb_Arab""", """crh_Latn""", """cym_Latn""", """dan_Latn""", """deu_Latn""", """dik_Latn""", """dyu_Latn""", """dzo_Tibt""", """ell_Grek""", """eng_Latn""", """epo_Latn""", """est_Latn""", """eus_Latn""", """ewe_Latn""", """fao_Latn""", """pes_Arab""", """fij_Latn""", """fin_Latn""", """fon_Latn""", """fra_Latn""", """fur_Latn""", """fuv_Latn""", """gla_Latn""", """gle_Latn""", """glg_Latn""", """grn_Latn""", """guj_Gujr""", """hat_Latn""", """hau_Latn""", """heb_Hebr""", """hin_Deva""", """hne_Deva""", """hrv_Latn""", """hun_Latn""", """hye_Armn""", """ibo_Latn""", """ilo_Latn""", """ind_Latn""", """isl_Latn""", """ita_Latn""", """jav_Latn""", """jpn_Jpan""", """kab_Latn""", """kac_Latn""", """kam_Latn""", """kan_Knda""", """kas_Arab""", """kas_Deva""", """kat_Geor""", """knc_Arab""", """knc_Latn""", """kaz_Cyrl""", """kbp_Latn""", """kea_Latn""", """khm_Khmr""", """kik_Latn""", """kin_Latn""", """kir_Cyrl""", """kmb_Latn""", """kon_Latn""", """kor_Hang""", """kmr_Latn""", """lao_Laoo""", """lvs_Latn""", """lij_Latn""", """lim_Latn""", """lin_Latn""", """lit_Latn""", """lmo_Latn""", """ltg_Latn""", """ltz_Latn""", """lua_Latn""", """lug_Latn""", """luo_Latn""", """lus_Latn""", """mag_Deva""", """mai_Deva""", """mal_Mlym""", """mar_Deva""", """min_Latn""", """mkd_Cyrl""", """plt_Latn""", """mlt_Latn""", """mni_Beng""", """khk_Cyrl""", """mos_Latn""", """mri_Latn""", """zsm_Latn""", """mya_Mymr""", """nld_Latn""", """nno_Latn""", """nob_Latn""", """npi_Deva""", """nso_Latn""", """nus_Latn""", """nya_Latn""", """oci_Latn""", """gaz_Latn""", """ory_Orya""", """pag_Latn""", """pan_Guru""", """pap_Latn""", """pol_Latn""", """por_Latn""", """prs_Arab""", """pbt_Arab""", """quy_Latn""", """ron_Latn""", """run_Latn""", """rus_Cyrl""", """sag_Latn""", """san_Deva""", """sat_Beng""", """scn_Latn""", """shn_Mymr""", """sin_Sinh""", """slk_Latn""", """slv_Latn""", """smo_Latn""", """sna_Latn""", """snd_Arab""", """som_Latn""", """sot_Latn""", """spa_Latn""", """als_Latn""", """srd_Latn""", """srp_Cyrl""", """ssw_Latn""", """sun_Latn""", """swe_Latn""", """swh_Latn""", """szl_Latn""", """tam_Taml""", """tat_Cyrl""", """tel_Telu""", """tgk_Cyrl""", """tgl_Latn""", """tha_Thai""", """tir_Ethi""", """taq_Latn""", """taq_Tfng""", """tpi_Latn""", """tsn_Latn""", """tso_Latn""", """tuk_Latn""", """tum_Latn""", """tur_Latn""", """twi_Latn""", """tzm_Tfng""", """uig_Arab""", """ukr_Cyrl""", """umb_Latn""", """urd_Arab""", """uzn_Latn""", """vec_Latn""", """vie_Latn""", """war_Latn""", """wol_Latn""", """xho_Latn""", """ydd_Hebr""", """yor_Latn""", """yue_Hant""", """zho_Hans""", """zho_Hant""", """zul_Latn"""]
class lowerCAmelCase ( __UpperCamelCase ):
UpperCAmelCase__ = VOCAB_FILES_NAMES
UpperCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCAmelCase__ = ["""input_ids""", """attention_mask"""]
UpperCAmelCase__ = NllbTokenizer
UpperCAmelCase__ = []
UpperCAmelCase__ = []
def __init__( self : Tuple , UpperCAmelCase : int=None , UpperCAmelCase : Any=None , UpperCAmelCase : str="<s>" , UpperCAmelCase : Optional[Any]="</s>" , UpperCAmelCase : str="</s>" , UpperCAmelCase : Tuple="<s>" , UpperCAmelCase : Optional[Any]="<unk>" , UpperCAmelCase : List[str]="<pad>" , UpperCAmelCase : Union[str, Any]="<mask>" , UpperCAmelCase : Tuple=None , UpperCAmelCase : int=None , UpperCAmelCase : Dict=None , UpperCAmelCase : Any=False , **UpperCAmelCase : Optional[int] , ) -> Tuple:
# Mask token behave like a normal word, i.e. include the space before it
lowerCamelCase__ : List[Any] = AddedToken(UpperCAmelCase , lstrip=UpperCAmelCase , rstrip=UpperCAmelCase ) if isinstance(UpperCAmelCase , UpperCAmelCase ) else mask_token
lowerCamelCase__ : Union[str, Any] = legacy_behaviour
super().__init__(
vocab_file=UpperCAmelCase , tokenizer_file=UpperCAmelCase , bos_token=UpperCAmelCase , eos_token=UpperCAmelCase , sep_token=UpperCAmelCase , cls_token=UpperCAmelCase , unk_token=UpperCAmelCase , pad_token=UpperCAmelCase , mask_token=UpperCAmelCase , src_lang=UpperCAmelCase , tgt_lang=UpperCAmelCase , additional_special_tokens=UpperCAmelCase , legacy_behaviour=UpperCAmelCase , **UpperCAmelCase , )
lowerCamelCase__ : List[Any] = vocab_file
lowerCamelCase__ : Dict = False if not self.vocab_file else True
lowerCamelCase__ : Optional[Any] = FAIRSEQ_LANGUAGE_CODES.copy()
if additional_special_tokens is not None:
# Only add those special tokens if they are not already there.
_additional_special_tokens.extend(
[t for t in additional_special_tokens if t not in _additional_special_tokens] )
self.add_special_tokens({'additional_special_tokens': _additional_special_tokens} )
lowerCamelCase__ : str = {
lang_code: self.convert_tokens_to_ids(UpperCAmelCase ) for lang_code in FAIRSEQ_LANGUAGE_CODES
}
lowerCamelCase__ : int = src_lang if src_lang is not None else 'eng_Latn'
lowerCamelCase__ : List[Any] = self.convert_tokens_to_ids(self._src_lang )
lowerCamelCase__ : str = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def A_ ( self : int ) -> str:
return self._src_lang
@src_lang.setter
def A_ ( self : List[Any] , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Any = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
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 A_ ( self : Optional[Any] , UpperCAmelCase : List[int] , UpperCAmelCase : Optional[List[int]] = None ) -> List[int]:
lowerCamelCase__ : Dict = [self.sep_token_id]
lowerCamelCase__ : List[str] = [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 + sep + token_ids_a + sep ) * [0]
def A_ ( self : int , UpperCAmelCase : int , UpperCAmelCase : str , UpperCAmelCase : Optional[str] , UpperCAmelCase : Optional[str] , **UpperCAmelCase : List[str] ) -> Dict:
if src_lang is None or tgt_lang is None:
raise ValueError('Translation requires a `src_lang` and a `tgt_lang` for this model' )
lowerCamelCase__ : Optional[int] = src_lang
lowerCamelCase__ : Optional[int] = self(UpperCAmelCase , add_special_tokens=UpperCAmelCase , return_tensors=UpperCAmelCase , **UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.convert_tokens_to_ids(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = tgt_lang_id
return inputs
def A_ ( self : Dict , UpperCAmelCase : List[str] , UpperCAmelCase : str = "eng_Latn" , UpperCAmelCase : Optional[List[str]] = None , UpperCAmelCase : str = "fra_Latn" , **UpperCAmelCase : Dict , ) -> BatchEncoding:
lowerCamelCase__ : Any = src_lang
lowerCamelCase__ : int = tgt_lang
return super().prepare_seqaseq_batch(UpperCAmelCase , UpperCAmelCase , **UpperCAmelCase )
def A_ ( self : Union[str, Any] ) -> Optional[int]:
return self.set_src_lang_special_tokens(self.src_lang )
def A_ ( self : Any ) -> Union[str, Any]:
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def A_ ( self : str , UpperCAmelCase : Optional[Any] ) -> None:
lowerCamelCase__ : int = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : int = []
lowerCamelCase__ : str = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : int = [self.cur_lang_code]
lowerCamelCase__ : Tuple = [self.eos_token_id]
lowerCamelCase__ : Any = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : Optional[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : str = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : int , UpperCAmelCase : str ) -> None:
lowerCamelCase__ : Union[str, Any] = self.convert_tokens_to_ids(UpperCAmelCase )
if self.legacy_behaviour:
lowerCamelCase__ : Dict = []
lowerCamelCase__ : Union[str, Any] = [self.eos_token_id, self.cur_lang_code]
else:
lowerCamelCase__ : Any = [self.cur_lang_code]
lowerCamelCase__ : Optional[Any] = [self.eos_token_id]
lowerCamelCase__ : Union[str, Any] = self.convert_ids_to_tokens(self.prefix_tokens )
lowerCamelCase__ : List[Any] = self.convert_ids_to_tokens(self.suffix_tokens )
lowerCamelCase__ : Optional[int] = processors.TemplateProcessing(
single=prefix_tokens_str + ['$A'] + suffix_tokens_str , pair=prefix_tokens_str + ['$A', '$B'] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def A_ ( self : Union[str, Any] , UpperCAmelCase : str , UpperCAmelCase : Optional[str] = None ) -> Tuple[str]:
if not self.can_save_slow_tokenizer:
raise ValueError(
'Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '
'tokenizer.' )
if not os.path.isdir(UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory.""" )
return
lowerCamelCase__ : int = os.path.join(
UpperCAmelCase , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCAmelCase ):
copyfile(self.vocab_file , UpperCAmelCase )
return (out_vocab_file,)
| 50
| 1
|
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 : Optional[Any] = False
class lowerCAmelCase ( unittest.TestCase ):
pass
@nightly
@require_torch_gpu
class lowerCAmelCase ( unittest.TestCase ):
def A_ ( self : str ) -> Optional[int]:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Any ) -> Optional[Any]:
lowerCamelCase__ : Union[str, Any] = VersatileDiffusionPipeline.from_pretrained('shi-labs/versatile-diffusion' , torch_dtype=torch.floataa )
pipe.to(UpperCAmelCase )
pipe.set_progress_bar_config(disable=UpperCAmelCase )
lowerCamelCase__ : List[Any] = load_image(
'https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/versatile_diffusion/benz.jpg' )
lowerCamelCase__ : int = torch.manual_seed(0 )
lowerCamelCase__ : List[Any] = pipe.dual_guided(
prompt='first prompt' , image=UpperCAmelCase , text_to_image_strength=0.7_5 , generator=UpperCAmelCase , guidance_scale=7.5 , num_inference_steps=2 , output_type='numpy' , ).images
with tempfile.TemporaryDirectory() as tmpdirname:
pipe.save_pretrained(UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = VersatileDiffusionPipeline.from_pretrained(UpperCAmelCase , torch_dtype=torch.floataa )
pipe.to(UpperCAmelCase )
pipe.set_progress_bar_config(disable=UpperCAmelCase )
lowerCamelCase__ : Any = generator.manual_seed(0 )
lowerCamelCase__ : int = pipe.dual_guided(
prompt='first prompt' , image=UpperCAmelCase , text_to_image_strength=0.7_5 , generator=UpperCAmelCase , 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 A_ ( self : str ) -> int:
lowerCamelCase__ : str = VersatileDiffusionPipeline.from_pretrained('shi-labs/versatile-diffusion' , torch_dtype=torch.floataa )
pipe.to(UpperCAmelCase )
pipe.set_progress_bar_config(disable=UpperCAmelCase )
lowerCamelCase__ : Union[str, Any] = 'cyberpunk 2077'
lowerCamelCase__ : Any = load_image(
'https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/versatile_diffusion/benz.jpg' )
lowerCamelCase__ : List[str] = torch.manual_seed(0 )
lowerCamelCase__ : List[str] = pipe.dual_guided(
prompt=UpperCAmelCase , image=UpperCAmelCase , text_to_image_strength=0.7_5 , generator=UpperCAmelCase , guidance_scale=7.5 , num_inference_steps=50 , output_type='numpy' , ).images
lowerCamelCase__ : Optional[int] = image[0, 253:256, 253:256, -1]
assert image.shape == (1, 512, 512, 3)
lowerCamelCase__ : Tuple = np.array([0.1_4_4_8, 0.1_6_1_9, 0.1_7_4_1, 0.1_0_8_6, 0.1_1_4_7, 0.1_1_2_8, 0.1_1_9_9, 0.1_1_6_5, 0.1_0_0_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-1
lowerCamelCase__ : Union[str, Any] = 'A painting of a squirrel eating a burger '
lowerCamelCase__ : Dict = torch.manual_seed(0 )
lowerCamelCase__ : int = pipe.text_to_image(
prompt=UpperCAmelCase , generator=UpperCAmelCase , guidance_scale=7.5 , num_inference_steps=50 , output_type='numpy' ).images
lowerCamelCase__ : Union[str, Any] = image[0, 253:256, 253:256, -1]
assert image.shape == (1, 512, 512, 3)
lowerCamelCase__ : Optional[Any] = np.array([0.3_3_6_7, 0.3_1_6_9, 0.2_6_5_6, 0.3_8_7_0, 0.4_7_9_0, 0.3_7_9_6, 0.4_0_0_9, 0.4_8_7_8, 0.4_7_7_8] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-1
lowerCamelCase__ : Optional[Any] = pipe.image_variation(UpperCAmelCase , generator=UpperCAmelCase , output_type='numpy' ).images
lowerCamelCase__ : Dict = image[0, 253:256, 253:256, -1]
assert image.shape == (1, 512, 512, 3)
lowerCamelCase__ : Optional[int] = np.array([0.3_0_7_6, 0.3_1_2_3, 0.3_2_8_4, 0.3_7_8_2, 0.3_7_7_0, 0.3_8_9_4, 0.4_2_9_7, 0.4_3_3_1, 0.4_4_5_6] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-1
| 50
|
import argparse
import torch
from torch import nn
from transformers import MaMaaaConfig, MaMaaaForConditionalGeneration
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Optional[int] = [
'encoder.version',
'decoder.version',
'model.encoder.version',
'model.decoder.version',
'decoder.output_projection.weight',
'_float_tensor',
'encoder.embed_positions._float_tensor',
'decoder.embed_positions._float_tensor',
]
for k in ignore_keys:
state_dict.pop(_UpperCAmelCase , _UpperCAmelCase )
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> List[Any]:
lowerCamelCase__ , lowerCamelCase__ : List[str] = emb.weight.shape
lowerCamelCase__ : Tuple = nn.Linear(_UpperCAmelCase , _UpperCAmelCase , bias=_UpperCAmelCase )
lowerCamelCase__ : Dict = emb.weight.data
return lin_layer
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> str:
lowerCamelCase__ : Tuple = torch.load(_UpperCAmelCase , map_location='cpu' )
lowerCamelCase__ : List[str] = mam_aaa['args'] or mam_aaa['cfg']['model']
lowerCamelCase__ : Optional[int] = mam_aaa['model']
remove_ignore_keys_(_UpperCAmelCase )
lowerCamelCase__ : str = state_dict['encoder.embed_tokens.weight'].shape[0]
lowerCamelCase__ : Union[str, Any] = MaMaaaConfig(
vocab_size=_UpperCAmelCase , max_position_embeddings=1024 , encoder_layers=args.encoder_layers , decoder_layers=args.decoder_layers , encoder_attention_heads=args.encoder_attention_heads , decoder_attention_heads=args.decoder_attention_heads , encoder_ffn_dim=args.encoder_ffn_embed_dim , decoder_ffn_dim=args.decoder_ffn_embed_dim , d_model=args.encoder_embed_dim , encoder_layerdrop=args.encoder_layerdrop , decoder_layerdrop=args.decoder_layerdrop , dropout=args.dropout , attention_dropout=args.attention_dropout , activation_dropout=args.activation_dropout , activation_function='relu' , )
lowerCamelCase__ : Optional[Any] = state_dict['decoder.embed_tokens.weight']
lowerCamelCase__ : Union[str, Any] = MaMaaaForConditionalGeneration(_UpperCAmelCase )
model.model.load_state_dict(_UpperCAmelCase , strict=_UpperCAmelCase )
lowerCamelCase__ : List[str] = make_linear_from_emb(model.model.shared )
return model
if __name__ == "__main__":
_UpperCAmelCase : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""fairseq_path""", type=str, help="""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.""")
_UpperCAmelCase : str = parser.parse_args()
_UpperCAmelCase : Optional[Any] = convert_fairseq_mamaaa_checkpoint_from_disk(args.fairseq_pathß)
model.save_pretrained(args.pytorch_dump_folder_path)
| 50
| 1
|
import argparse
import json
from collections import OrderedDict
import torch
from huggingface_hub import cached_download, hf_hub_url
from transformers import AutoImageProcessor, CvtConfig, CvtForImageClassification
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> int:
lowerCamelCase__ : Optional[int] = []
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.weight""",
F"""stage{idx}.patch_embed.proj.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.projection.bias""",
F"""stage{idx}.patch_embed.proj.bias""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.weight""",
F"""stage{idx}.patch_embed.norm.weight""",
) )
embed.append(
(
F"""cvt.encoder.stages.{idx}.embedding.convolution_embeddings.normalization.bias""",
F"""stage{idx}.patch_embed.norm.bias""",
) )
return embed
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Tuple = []
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_query.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_q.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_key.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_k.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.convolution.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.conv.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.weight""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.bias""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_mean""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_mean""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.running_var""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.running_var""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.convolution_projection_value.convolution_projection.normalization.num_batches_tracked""",
F"""stage{idx}.blocks.{cnt}.attn.conv_proj_v.bn.num_batches_tracked""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_query.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_q.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_key.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_k.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.attention.projection_value.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj_v.bias""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.weight""",
F"""stage{idx}.blocks.{cnt}.attn.proj.weight""",
) )
attention_weights.append(
(
F"""cvt.encoder.stages.{idx}.layers.{cnt}.attention.output.dense.bias""",
F"""stage{idx}.blocks.{cnt}.attn.proj.bias""",
) )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.intermediate.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.weight""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.output.dense.bias""", F"""stage{idx}.blocks.{cnt}.mlp.fc2.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.weight""", F"""stage{idx}.blocks.{cnt}.norm1.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_before.bias""", F"""stage{idx}.blocks.{cnt}.norm1.bias""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.weight""", F"""stage{idx}.blocks.{cnt}.norm2.weight""") )
attention_weights.append(
(F"""cvt.encoder.stages.{idx}.layers.{cnt}.layernorm_after.bias""", F"""stage{idx}.blocks.{cnt}.norm2.bias""") )
return attention_weights
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
lowerCamelCase__ : Union[str, Any] = []
token.append((F"""cvt.encoder.stages.{idx}.cls_token""", 'stage2.cls_token') )
return token
def SCREAMING_SNAKE_CASE ( ) -> str:
lowerCamelCase__ : str = []
head.append(('layernorm.weight', 'norm.weight') )
head.append(('layernorm.bias', 'norm.bias') )
head.append(('classifier.weight', 'head.weight') )
head.append(('classifier.bias', 'head.bias') )
return head
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> Optional[int]:
lowerCamelCase__ : Tuple = 'imagenet-1k-id2label.json'
lowerCamelCase__ : Union[str, Any] = 1000
lowerCamelCase__ : Optional[Any] = 'huggingface/label-files'
lowerCamelCase__ : Any = num_labels
lowerCamelCase__ : Dict = json.load(open(cached_download(hf_hub_url(_UpperCAmelCase , _UpperCAmelCase , repo_type='dataset' ) ) , 'r' ) )
lowerCamelCase__ : int = {int(_UpperCAmelCase ): v for k, v in idalabel.items()}
lowerCamelCase__ : Tuple = idalabel
lowerCamelCase__ : List[Any] = {v: k for k, v in idalabel.items()}
lowerCamelCase__ : List[str] = CvtConfig(num_labels=_UpperCAmelCase , idalabel=_UpperCAmelCase , labelaid=_UpperCAmelCase )
# For depth size 13 (13 = 1+2+10)
if cvt_model.rsplit('/' , 1 )[-1][4:6] == "13":
lowerCamelCase__ : List[Any] = [1, 2, 10]
# For depth size 21 (21 = 1+4+16)
elif cvt_model.rsplit('/' , 1 )[-1][4:6] == "21":
lowerCamelCase__ : Dict = [1, 4, 16]
# For wide cvt (similar to wide-resnet) depth size 24 (w24 = 2 + 2 20)
else:
lowerCamelCase__ : Optional[Any] = [2, 2, 20]
lowerCamelCase__ : Optional[int] = [3, 12, 16]
lowerCamelCase__ : str = [192, 768, 1024]
lowerCamelCase__ : Any = CvtForImageClassification(_UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = AutoImageProcessor.from_pretrained('facebook/convnext-base-224-22k-1k' )
lowerCamelCase__ : Tuple = image_size
lowerCamelCase__ : List[str] = torch.load(_UpperCAmelCase , map_location=torch.device('cpu' ) )
lowerCamelCase__ : Optional[int] = OrderedDict()
lowerCamelCase__ : Tuple = []
for idx in range(len(config.depth ) ):
if config.cls_token[idx]:
lowerCamelCase__ : Optional[Any] = list_of_state_dict + cls_token(_UpperCAmelCase )
lowerCamelCase__ : str = list_of_state_dict + embeddings(_UpperCAmelCase )
for cnt in range(config.depth[idx] ):
lowerCamelCase__ : str = list_of_state_dict + attention(_UpperCAmelCase , _UpperCAmelCase )
lowerCamelCase__ : int = list_of_state_dict + final()
for gg in list_of_state_dict:
print(_UpperCAmelCase )
for i in range(len(_UpperCAmelCase ) ):
lowerCamelCase__ : str = original_weights[list_of_state_dict[i][1]]
model.load_state_dict(_UpperCAmelCase )
model.save_pretrained(_UpperCAmelCase )
image_processor.save_pretrained(_UpperCAmelCase )
# Download the weights from zoo: https://1drv.ms/u/s!AhIXJn_J-blW9RzF3rMW7SsLHa8h?e=blQ0Al
if __name__ == "__main__":
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
parser.add_argument(
"""--cvt_model""",
default="""cvt-w24""",
type=str,
help="""Name of the cvt model you'd like to convert.""",
)
parser.add_argument(
"""--image_size""",
default=3_84,
type=int,
help="""Input Image Size""",
)
parser.add_argument(
"""--cvt_file_name""",
default=R"""cvtmodels\CvT-w24-384x384-IN-22k.pth""",
type=str,
help="""Input Image Size""",
)
parser.add_argument(
"""--pytorch_dump_folder_path""", default=None, type=str, help="""Path to the output PyTorch model directory."""
)
_UpperCAmelCase : List[str] = parser.parse_args()
convert_cvt_checkpoint(args.cvt_model, args.image_size, args.cvt_file_name, args.pytorch_dump_folder_path)
| 50
|
import argparse
import json
import os
import time
import zipfile
from get_ci_error_statistics import download_artifact, get_artifacts_links
from transformers import logging
_UpperCAmelCase : Optional[int] = logging.get_logger(__name__)
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Dict:
lowerCamelCase__ : str = set()
lowerCamelCase__ : Any = []
def parse_line(_UpperCAmelCase ):
for line in fp:
if isinstance(_UpperCAmelCase , _UpperCAmelCase ):
lowerCamelCase__ : Any = line.decode('UTF-8' )
if "warnings summary (final)" in line:
continue
# This means we are outside the body of a warning
elif not line.startswith(' ' ):
# process a single warning and move it to `selected_warnings`.
if len(_UpperCAmelCase ) > 0:
lowerCamelCase__ : str = '\n'.join(_UpperCAmelCase )
# Only keep the warnings specified in `targets`
if any(F""": {x}: """ in warning for x in targets ):
selected_warnings.add(_UpperCAmelCase )
buffer.clear()
continue
else:
lowerCamelCase__ : List[str] = line.strip()
buffer.append(_UpperCAmelCase )
if from_gh:
for filename in os.listdir(_UpperCAmelCase ):
lowerCamelCase__ : Dict = os.path.join(_UpperCAmelCase , _UpperCAmelCase )
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
else:
try:
with zipfile.ZipFile(_UpperCAmelCase ) as z:
for filename in z.namelist():
if not os.path.isdir(_UpperCAmelCase ):
# read the file
if filename != "warnings.txt":
continue
with z.open(_UpperCAmelCase ) as fp:
parse_line(_UpperCAmelCase )
except Exception:
logger.warning(
F"""{artifact_path} is either an invalid zip file or something else wrong. This file is skipped.""" )
return selected_warnings
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase ) -> Union[str, Any]:
lowerCamelCase__ : Tuple = set()
lowerCamelCase__ : Optional[int] = [os.path.join(_UpperCAmelCase , _UpperCAmelCase ) for p in os.listdir(_UpperCAmelCase ) if (p.endswith('.zip' ) or from_gh)]
for p in paths:
selected_warnings.update(extract_warnings_from_single_artifact(_UpperCAmelCase , _UpperCAmelCase ) )
return selected_warnings
if __name__ == "__main__":
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase ) -> Tuple:
return values.split(',' )
_UpperCAmelCase : List[str] = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""--workflow_run_id""", type=str, required=True, help="""A GitHub Actions workflow run id.""")
parser.add_argument(
"""--output_dir""",
type=str,
required=True,
help="""Where to store the downloaded artifacts and other result files.""",
)
parser.add_argument("""--token""", default=None, type=str, help="""A token that has actions:read permission.""")
# optional parameters
parser.add_argument(
"""--targets""",
default="""DeprecationWarning,UserWarning,FutureWarning""",
type=list_str,
help="""Comma-separated list of target warning(s) which we want to extract.""",
)
parser.add_argument(
"""--from_gh""",
action="""store_true""",
help="""If running from a GitHub action workflow and collecting warnings from its artifacts.""",
)
_UpperCAmelCase : Union[str, Any] = parser.parse_args()
_UpperCAmelCase : Dict = args.from_gh
if from_gh:
# The artifacts have to be downloaded using `actions/download-artifact@v3`
pass
else:
os.makedirs(args.output_dir, exist_ok=True)
# get download links
_UpperCAmelCase : Union[str, Any] = get_artifacts_links(args.workflow_run_id, token=args.token)
with open(os.path.join(args.output_dir, """artifacts.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(artifacts, fp, ensure_ascii=False, indent=4)
# download artifacts
for idx, (name, url) in enumerate(artifacts.items()):
print(name)
print(url)
print("""=""" * 80)
download_artifact(name, url, args.output_dir, args.token)
# Be gentle to GitHub
time.sleep(1)
# extract warnings from artifacts
_UpperCAmelCase : Dict = extract_warnings(args.output_dir, args.targets)
_UpperCAmelCase : Optional[Any] = sorted(selected_warnings)
with open(os.path.join(args.output_dir, """selected_warnings.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(selected_warnings, fp, ensure_ascii=False, indent=4)
| 50
| 1
|
from __future__ import annotations
from math import pow, sqrt
def SCREAMING_SNAKE_CASE ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase ) -> dict[str, float]:
if (resistance, reactance, impedance).count(0 ) != 1:
raise ValueError('One and only one argument must be 0' )
if resistance == 0:
return {"resistance": sqrt(pow(_UpperCAmelCase , 2 ) - pow(_UpperCAmelCase , 2 ) )}
elif reactance == 0:
return {"reactance": sqrt(pow(_UpperCAmelCase , 2 ) - pow(_UpperCAmelCase , 2 ) )}
elif impedance == 0:
return {"impedance": sqrt(pow(_UpperCAmelCase , 2 ) + pow(_UpperCAmelCase , 2 ) )}
else:
raise ValueError('Exactly one argument must be 0' )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 50
|
import flax.linen as nn
import jax
import jax.numpy as jnp
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : Any ) -> Any:
lowerCamelCase__ : str = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : int , UpperCAmelCase : Dict ) -> Optional[Any]:
lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ , lowerCamelCase__ : str = hidden_states.shape
lowerCamelCase__ : Union[str, Any] = jax.image.resize(
UpperCAmelCase , shape=(batch, height * 2, width * 2, channels) , method='nearest' , )
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> int:
lowerCamelCase__ : Tuple = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : str , UpperCAmelCase : Union[str, Any] ) -> Optional[Any]:
# pad = ((0, 0), (0, 1), (0, 1), (0, 0)) # pad height and width dim
# hidden_states = jnp.pad(hidden_states, pad_width=pad)
lowerCamelCase__ : Optional[Any] = self.conv(UpperCAmelCase )
return hidden_states
class lowerCAmelCase ( nn.Module ):
UpperCAmelCase__ = 42
UpperCAmelCase__ = None
UpperCAmelCase__ = 0.0
UpperCAmelCase__ = None
UpperCAmelCase__ = jnp.floataa
def A_ ( self : List[str] ) -> Union[str, Any]:
lowerCamelCase__ : Optional[Any] = self.in_channels if self.out_channels is None else self.out_channels
lowerCamelCase__ : Tuple = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : int = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Union[str, Any] = nn.Dense(UpperCAmelCase , dtype=self.dtype )
lowerCamelCase__ : Union[str, Any] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
lowerCamelCase__ : List[Any] = nn.Dropout(self.dropout_prob )
lowerCamelCase__ : Tuple = nn.Conv(
UpperCAmelCase , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowerCamelCase__ : Optional[Any] = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
lowerCamelCase__ : Union[str, Any] = None
if use_nin_shortcut:
lowerCamelCase__ : Dict = nn.Conv(
UpperCAmelCase , kernel_size=(1, 1) , strides=(1, 1) , padding='VALID' , dtype=self.dtype , )
def __call__( self : Optional[Any] , UpperCAmelCase : str , UpperCAmelCase : Union[str, Any] , UpperCAmelCase : int=True ) -> Optional[int]:
lowerCamelCase__ : Union[str, Any] = hidden_states
lowerCamelCase__ : List[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[Any] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Any = self.conva(UpperCAmelCase )
lowerCamelCase__ : Optional[Any] = self.time_emb_proj(nn.swish(UpperCAmelCase ) )
lowerCamelCase__ : List[str] = jnp.expand_dims(jnp.expand_dims(UpperCAmelCase , 1 ) , 1 )
lowerCamelCase__ : List[str] = hidden_states + temb
lowerCamelCase__ : Optional[Any] = self.norma(UpperCAmelCase )
lowerCamelCase__ : List[str] = nn.swish(UpperCAmelCase )
lowerCamelCase__ : Optional[int] = self.dropout(UpperCAmelCase , UpperCAmelCase )
lowerCamelCase__ : str = self.conva(UpperCAmelCase )
if self.conv_shortcut is not None:
lowerCamelCase__ : Dict = self.conv_shortcut(UpperCAmelCase )
return hidden_states + residual
| 50
| 1
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.