repo
stringlengths 7
54
| path
stringlengths 4
223
| func_name
stringlengths 1
134
| original_string
stringlengths 75
104k
| language
stringclasses 1
value | code
stringlengths 75
104k
| code_tokens
listlengths 20
28.4k
| docstring
stringlengths 1
46.3k
| docstring_tokens
listlengths 1
1.66k
| sha
stringlengths 40
40
| url
stringlengths 87
315
| partition
stringclasses 1
value | summary
stringlengths 4
350
| obf_code
stringlengths 7.85k
764k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.bottom
|
def bottom(self, features):
"""Transforms features to feed into body.
Args:
features: dict of str to Tensor. Typically it is the preprocessed data
batch after Problem's preprocess_example().
Returns:
transformed_features: dict of same key-value pairs as features. The value
Tensors are newly transformed.
"""
if not self._problem_hparams:
log_warn("Without a Problem, T2TModel.bottom is a passthrough.")
return features
transformed_features = collections.OrderedDict()
all_previous_modalities = []
target_modality = _create_target_modality(self._problem_hparams.modality)
# Transform features via its corresponding modality.
for feature_name, modality in sorted(
six.iteritems(self._problem_hparams.modality)):
if feature_name not in features:
tf.logging.warning("Missing feature %s - ignoring." % feature_name)
continue
vocab_size = self._problem_hparams.vocab_size[feature_name]
if vocab_size is not None and hasattr(self._hparams, "vocab_divisor"):
vocab_size += (-vocab_size) % self._hparams.vocab_divisor
modality_name = self._hparams.name.get(
feature_name,
modalities.get_name(modality))(self._hparams, vocab_size)
# Use if-else clauses to preserve behavior of previous changes: namely,
# the variable scope name for the targets feature if there is only one
# target modality; and to reuse variable scopes for only input modalities.
if feature_name in target_modality:
if len(target_modality) > 1:
variable_scope_name = "%s/%s" % (modality_name, feature_name)
else:
variable_scope_name = modality_name
bottom = self._hparams.bottom.get(
feature_name,
modalities.get_targets_bottom(modality))
# TODO(aidangomez): share variables?
with tf.variable_scope(variable_scope_name) as vs:
self._add_variable_scope(variable_scope_name, vs)
log_info("Transforming feature '%s' with %s.targets_bottom",
feature_name,
modality_name)
transformed_features[feature_name] = bottom(features[feature_name],
self._hparams,
vocab_size)
else:
bottom = self._hparams.bottom.get(feature_name,
modalities.get_bottom(modality))
do_reuse = modality_name in all_previous_modalities
with tf.variable_scope(modality_name, reuse=do_reuse) as vs:
self._add_variable_scope(modality_name, vs)
log_info("Transforming feature '%s' with %s.bottom",
feature_name,
modality_name)
transformed_features[feature_name] = bottom(features[feature_name],
self._hparams,
vocab_size)
all_previous_modalities.append(modality_name)
for key in features:
if key not in transformed_features:
# For features without a modality, we pass them along as is
transformed_features[key] = features[key]
else:
# Other features get passed along with the "raw" suffix
transformed_features[key + "_raw"] = features[key]
return transformed_features
|
python
|
def bottom(self, features):
"""Transforms features to feed into body.
Args:
features: dict of str to Tensor. Typically it is the preprocessed data
batch after Problem's preprocess_example().
Returns:
transformed_features: dict of same key-value pairs as features. The value
Tensors are newly transformed.
"""
if not self._problem_hparams:
log_warn("Without a Problem, T2TModel.bottom is a passthrough.")
return features
transformed_features = collections.OrderedDict()
all_previous_modalities = []
target_modality = _create_target_modality(self._problem_hparams.modality)
# Transform features via its corresponding modality.
for feature_name, modality in sorted(
six.iteritems(self._problem_hparams.modality)):
if feature_name not in features:
tf.logging.warning("Missing feature %s - ignoring." % feature_name)
continue
vocab_size = self._problem_hparams.vocab_size[feature_name]
if vocab_size is not None and hasattr(self._hparams, "vocab_divisor"):
vocab_size += (-vocab_size) % self._hparams.vocab_divisor
modality_name = self._hparams.name.get(
feature_name,
modalities.get_name(modality))(self._hparams, vocab_size)
# Use if-else clauses to preserve behavior of previous changes: namely,
# the variable scope name for the targets feature if there is only one
# target modality; and to reuse variable scopes for only input modalities.
if feature_name in target_modality:
if len(target_modality) > 1:
variable_scope_name = "%s/%s" % (modality_name, feature_name)
else:
variable_scope_name = modality_name
bottom = self._hparams.bottom.get(
feature_name,
modalities.get_targets_bottom(modality))
# TODO(aidangomez): share variables?
with tf.variable_scope(variable_scope_name) as vs:
self._add_variable_scope(variable_scope_name, vs)
log_info("Transforming feature '%s' with %s.targets_bottom",
feature_name,
modality_name)
transformed_features[feature_name] = bottom(features[feature_name],
self._hparams,
vocab_size)
else:
bottom = self._hparams.bottom.get(feature_name,
modalities.get_bottom(modality))
do_reuse = modality_name in all_previous_modalities
with tf.variable_scope(modality_name, reuse=do_reuse) as vs:
self._add_variable_scope(modality_name, vs)
log_info("Transforming feature '%s' with %s.bottom",
feature_name,
modality_name)
transformed_features[feature_name] = bottom(features[feature_name],
self._hparams,
vocab_size)
all_previous_modalities.append(modality_name)
for key in features:
if key not in transformed_features:
# For features without a modality, we pass them along as is
transformed_features[key] = features[key]
else:
# Other features get passed along with the "raw" suffix
transformed_features[key + "_raw"] = features[key]
return transformed_features
|
[
"def",
"bottom",
"(",
"self",
",",
"features",
")",
":",
"if",
"not",
"self",
".",
"_problem_hparams",
":",
"log_warn",
"(",
"\"Without a Problem, T2TModel.bottom is a passthrough.\"",
")",
"return",
"features",
"transformed_features",
"=",
"collections",
".",
"OrderedDict",
"(",
")",
"all_previous_modalities",
"=",
"[",
"]",
"target_modality",
"=",
"_create_target_modality",
"(",
"self",
".",
"_problem_hparams",
".",
"modality",
")",
"# Transform features via its corresponding modality.",
"for",
"feature_name",
",",
"modality",
"in",
"sorted",
"(",
"six",
".",
"iteritems",
"(",
"self",
".",
"_problem_hparams",
".",
"modality",
")",
")",
":",
"if",
"feature_name",
"not",
"in",
"features",
":",
"tf",
".",
"logging",
".",
"warning",
"(",
"\"Missing feature %s - ignoring.\"",
"%",
"feature_name",
")",
"continue",
"vocab_size",
"=",
"self",
".",
"_problem_hparams",
".",
"vocab_size",
"[",
"feature_name",
"]",
"if",
"vocab_size",
"is",
"not",
"None",
"and",
"hasattr",
"(",
"self",
".",
"_hparams",
",",
"\"vocab_divisor\"",
")",
":",
"vocab_size",
"+=",
"(",
"-",
"vocab_size",
")",
"%",
"self",
".",
"_hparams",
".",
"vocab_divisor",
"modality_name",
"=",
"self",
".",
"_hparams",
".",
"name",
".",
"get",
"(",
"feature_name",
",",
"modalities",
".",
"get_name",
"(",
"modality",
")",
")",
"(",
"self",
".",
"_hparams",
",",
"vocab_size",
")",
"# Use if-else clauses to preserve behavior of previous changes: namely,",
"# the variable scope name for the targets feature if there is only one",
"# target modality; and to reuse variable scopes for only input modalities.",
"if",
"feature_name",
"in",
"target_modality",
":",
"if",
"len",
"(",
"target_modality",
")",
">",
"1",
":",
"variable_scope_name",
"=",
"\"%s/%s\"",
"%",
"(",
"modality_name",
",",
"feature_name",
")",
"else",
":",
"variable_scope_name",
"=",
"modality_name",
"bottom",
"=",
"self",
".",
"_hparams",
".",
"bottom",
".",
"get",
"(",
"feature_name",
",",
"modalities",
".",
"get_targets_bottom",
"(",
"modality",
")",
")",
"# TODO(aidangomez): share variables?",
"with",
"tf",
".",
"variable_scope",
"(",
"variable_scope_name",
")",
"as",
"vs",
":",
"self",
".",
"_add_variable_scope",
"(",
"variable_scope_name",
",",
"vs",
")",
"log_info",
"(",
"\"Transforming feature '%s' with %s.targets_bottom\"",
",",
"feature_name",
",",
"modality_name",
")",
"transformed_features",
"[",
"feature_name",
"]",
"=",
"bottom",
"(",
"features",
"[",
"feature_name",
"]",
",",
"self",
".",
"_hparams",
",",
"vocab_size",
")",
"else",
":",
"bottom",
"=",
"self",
".",
"_hparams",
".",
"bottom",
".",
"get",
"(",
"feature_name",
",",
"modalities",
".",
"get_bottom",
"(",
"modality",
")",
")",
"do_reuse",
"=",
"modality_name",
"in",
"all_previous_modalities",
"with",
"tf",
".",
"variable_scope",
"(",
"modality_name",
",",
"reuse",
"=",
"do_reuse",
")",
"as",
"vs",
":",
"self",
".",
"_add_variable_scope",
"(",
"modality_name",
",",
"vs",
")",
"log_info",
"(",
"\"Transforming feature '%s' with %s.bottom\"",
",",
"feature_name",
",",
"modality_name",
")",
"transformed_features",
"[",
"feature_name",
"]",
"=",
"bottom",
"(",
"features",
"[",
"feature_name",
"]",
",",
"self",
".",
"_hparams",
",",
"vocab_size",
")",
"all_previous_modalities",
".",
"append",
"(",
"modality_name",
")",
"for",
"key",
"in",
"features",
":",
"if",
"key",
"not",
"in",
"transformed_features",
":",
"# For features without a modality, we pass them along as is",
"transformed_features",
"[",
"key",
"]",
"=",
"features",
"[",
"key",
"]",
"else",
":",
"# Other features get passed along with the \"raw\" suffix",
"transformed_features",
"[",
"key",
"+",
"\"_raw\"",
"]",
"=",
"features",
"[",
"key",
"]",
"return",
"transformed_features"
] |
Transforms features to feed into body.
Args:
features: dict of str to Tensor. Typically it is the preprocessed data
batch after Problem's preprocess_example().
Returns:
transformed_features: dict of same key-value pairs as features. The value
Tensors are newly transformed.
|
[
"Transforms",
"features",
"to",
"feed",
"into",
"body",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L443-L516
|
train
|
Transforms features to feed into body.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(1677 - 1626) + chr(0b110 + 0o56) + chr(0b110001), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110011) + '\065' + chr(276 - 227), 0o10), ehT0Px3KOsy9(chr(2053 - 2005) + '\x6f' + '\x34' + chr(2162 - 2112), 0o10), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(0b110000 + 0o77) + chr(0b110111), 23713 - 23705), ehT0Px3KOsy9('\060' + chr(10338 - 10227) + chr(1239 - 1190) + '\064' + chr(48), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x37' + chr(48), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(961 - 907) + '\x31', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110001) + chr(0b110 + 0o52) + chr(54), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(49) + '\060' + chr(50), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\063' + chr(48) + '\x34', 0o10), ehT0Px3KOsy9('\x30' + chr(111) + '\061' + '\x33' + chr(0b100100 + 0o22), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1110 + 0o141) + chr(0b110011) + chr(52) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + '\062' + chr(0b100111 + 0o15), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b10011 + 0o36) + '\x31' + '\060', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(535 - 487) + chr(8183 - 8072) + chr(0b11000 + 0o33) + chr(0b110000 + 0o3) + '\x35', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1 + 0o156) + chr(53), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b1000 + 0o57) + chr(2519 - 2468), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(49) + chr(51) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(111) + '\063' + chr(433 - 380) + chr(0b10111 + 0o34), 0b1000), ehT0Px3KOsy9(chr(48) + chr(9524 - 9413) + chr(1074 - 1019), 8), ehT0Px3KOsy9(chr(1375 - 1327) + chr(0b1001100 + 0o43) + chr(50) + chr(0b1011 + 0o45) + chr(1591 - 1540), 46299 - 46291), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\061' + chr(0b11010 + 0o33) + chr(2509 - 2456), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b100010 + 0o21) + chr(1091 - 1042) + chr(0b10001 + 0o42), ord("\x08")), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b1101111) + '\065' + chr(412 - 363), 803 - 795), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + chr(48) + chr(350 - 295), 28851 - 28843), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + chr(0b110110) + chr(2820 - 2765), ord("\x08")), ehT0Px3KOsy9('\060' + chr(1312 - 1201) + chr(307 - 258) + chr(1174 - 1124) + chr(0b110001), 37825 - 37817), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(5222 - 5111) + chr(49) + chr(0b101111 + 0o1) + '\062', 8), ehT0Px3KOsy9('\060' + '\157' + '\061' + chr(0b110000 + 0o1) + '\066', 61904 - 61896), ehT0Px3KOsy9(chr(1875 - 1827) + '\x6f' + '\066' + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(0b110111) + chr(1010 - 959), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(1246 - 1195) + chr(0b110001) + chr(53), 0o10), ehT0Px3KOsy9('\060' + chr(4177 - 4066) + '\061' + chr(0b110101) + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\157' + '\062' + chr(1649 - 1601) + chr(547 - 498), 34234 - 34226), ehT0Px3KOsy9(chr(48) + chr(1665 - 1554) + '\x32' + chr(0b110010) + '\x32', 28230 - 28222), ehT0Px3KOsy9(chr(55 - 7) + '\x6f' + '\x35' + '\062', ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(426 - 377) + chr(0b110110) + '\x33', 0o10), ehT0Px3KOsy9(chr(530 - 482) + chr(0b1111 + 0o140) + chr(1712 - 1663) + '\067' + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(597 - 549) + chr(111) + '\062' + chr(0b110010) + chr(0b110011), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(113 - 65) + chr(111) + chr(53) + chr(0b101000 + 0o10), 7350 - 7342)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xcf'), chr(9801 - 9701) + '\x65' + '\x63' + chr(0b10000 + 0o137) + '\x64' + chr(0b1100101))('\x75' + chr(116) + chr(102) + '\x2d' + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def kXxsZxlIQUSQ(oVre8I6UXc3b, EEf4r9nUvta_):
if not xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xbe\x99Ph\x02TT+\x1d\xac.\x85\xa6)y\xae'), chr(100) + '\x65' + chr(6639 - 6540) + '\157' + '\144' + '\145')(chr(0b1110101) + chr(3841 - 3725) + chr(102) + '\055' + '\x38')):
DyguQXdCse_Q(xafqLlk3kkUe(SXOLrMavuUCe(b'\xb6\x80Vo\x0fMEf#\xe4\x0e\x96\xbb*x\xb8h\x17z\xbf5\xd7 \xa4M\xe2\xdb\x7f !C\x8f\xa2\x02\xa1\xfbYS\x1e\xb3\x91\x88Qt\x14PC)7\xa36\xca'), chr(4784 - 4684) + '\145' + chr(0b1100011) + chr(0b1000100 + 0o53) + '\144' + chr(0b1100101))('\x75' + '\164' + chr(0b10101 + 0o121) + '\x2d' + chr(1188 - 1132)))
return EEf4r9nUvta_
kRp2XqtfzLyV = FGhnnwoh1Dd8.OrderedDict()
jtsP69nFE2aJ = []
FfV_wqUqiesq = sg60tMitsGCm(oVre8I6UXc3b._problem_hparams.bYPswhysd3s2)
for (lPuZQT6rFAxL, bYPswhysd3s2) in vUlqIvNSaRMa(xafqLlk3kkUe(sYby0kpfssd4, xafqLlk3kkUe(SXOLrMavuUCe(b'\x88\x9dGu\tLT+1'), '\x64' + chr(0b100000 + 0o105) + chr(0b1011 + 0o130) + chr(4527 - 4416) + chr(7592 - 7492) + chr(0b110110 + 0o57))(chr(0b1100100 + 0o21) + '\164' + chr(0b1010111 + 0o17) + chr(0b11010 + 0o23) + chr(0b111000)))(xafqLlk3kkUe(oVre8I6UXc3b._problem_hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'\x83\xb0rt\x17PH5&\xf7-\xd6'), '\144' + chr(0b1100101) + chr(4400 - 4301) + '\157' + chr(5168 - 5068) + chr(0b111111 + 0o46))('\165' + '\x74' + '\x66' + '\055' + '\070')))):
if lPuZQT6rFAxL not in EEf4r9nUvta_:
xafqLlk3kkUe(IDJ2eXGCBCDu.logging, xafqLlk3kkUe(SXOLrMavuUCe(b'\x96\x88Pi\tVV'), '\x64' + chr(101) + chr(99) + chr(0b1100110 + 0o11) + chr(2706 - 2606) + chr(0b100010 + 0o103))(chr(0b1101111 + 0o6) + '\164' + chr(3554 - 3452) + chr(0b100000 + 0o15) + chr(2732 - 2676)))(xafqLlk3kkUe(SXOLrMavuUCe(b"\xac\x80Qt\tVVf$\xa1?\x90\xa1:q\xfd Hz\xc6'\xea\n\xa5F\xf5\xde?%`"), chr(0b1100100) + chr(101) + '\143' + chr(0b1101000 + 0o7) + '\144' + chr(0b1100101))('\x75' + '\164' + chr(0b1100110) + '\055' + '\070') % lPuZQT6rFAxL)
continue
CeyMIoSyrpkQ = oVre8I6UXc3b._problem_hparams.CeyMIoSyrpkQ[lPuZQT6rFAxL]
if CeyMIoSyrpkQ is not None and lot1PSoAwYhj(xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\x8f\xb9m2#@G\x19\x10\x97\x0f\xb5'), '\x64' + chr(101) + '\143' + chr(2994 - 2883) + '\x64' + '\x65')(chr(117) + chr(342 - 226) + chr(3037 - 2935) + chr(0b101101) + chr(0b111000))), xafqLlk3kkUe(SXOLrMavuUCe(b'\x97\x86Af\x02gU/4\xad-\x8b\xa6'), chr(0b11000 + 0o114) + chr(101) + chr(99) + '\x6f' + chr(0b1100100) + '\x65')('\165' + chr(0b1110100) + '\x66' + chr(45) + chr(0b111000 + 0o0))):
CeyMIoSyrpkQ += -CeyMIoSyrpkQ % oVre8I6UXc3b._hparams.vocab_divisor
RjR6XYrAEBW8 = oVre8I6UXc3b._hparams.name.get(lPuZQT6rFAxL, PuPeNl0CuqOQ.get_name(bYPswhysd3s2))(oVre8I6UXc3b.nPO5Cxv_RSQQ, CeyMIoSyrpkQ)
if lPuZQT6rFAxL in FfV_wqUqiesq:
if c2A0yzQpDQB3(FfV_wqUqiesq) > ehT0Px3KOsy9(chr(0b100101 + 0o13) + '\x6f' + chr(2086 - 2037), 0o10):
BV9WgWDuYX5m = xafqLlk3kkUe(SXOLrMavuUCe(b'\xc4\x9a\r"\x13'), '\x64' + chr(6297 - 6196) + chr(3658 - 3559) + chr(111) + chr(0b1100100) + chr(0b11010 + 0o113))('\165' + chr(0b11101 + 0o127) + '\x66' + '\055' + chr(1375 - 1319)) % (RjR6XYrAEBW8, lPuZQT6rFAxL)
else:
BV9WgWDuYX5m = RjR6XYrAEBW8
kXxsZxlIQUSQ = oVre8I6UXc3b._hparams.bottom.get(lPuZQT6rFAxL, PuPeNl0CuqOQ.get_targets_bottom(bYPswhysd3s2))
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x97\x88Pn\x01Z]#\x1d\xb7=\x8b\xa4-'), '\x64' + chr(0b1101 + 0o130) + chr(7283 - 7184) + chr(8899 - 8788) + chr(993 - 893) + chr(0b110000 + 0o65))(chr(6302 - 6185) + '\164' + '\x66' + chr(1742 - 1697) + chr(2969 - 2913)))(BV9WgWDuYX5m) as qGaVI8v_Oz7A:
xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xbe\x88Fc?NP4+\xa5<\x88\xb1\x17g\xbejK?'), '\144' + '\145' + '\143' + chr(111) + chr(0b1100100) + '\145')(chr(0b1110101) + chr(0b1110100) + chr(102) + '\x2d' + chr(56)))(BV9WgWDuYX5m, qGaVI8v_Oz7A)
iBNBQu99lQEa(xafqLlk3kkUe(SXOLrMavuUCe(b"\xb5\x9bCi\x13^^4/\xad0\x83\xf4.q\xbcqN(\x8e'\xa4H\xb8\x0e\xa7\xc086&\x17\xde\xbeA\xf5\xf3X\x14\x1a\xe7\x92\xb6@h\x14L^+"), chr(100) + chr(7823 - 7722) + '\x63' + '\x6f' + '\144' + chr(2548 - 2447))(chr(0b110111 + 0o76) + '\x74' + '\146' + chr(2002 - 1957) + chr(0b111000)), lPuZQT6rFAxL, RjR6XYrAEBW8)
kRp2XqtfzLyV[lPuZQT6rFAxL] = kXxsZxlIQUSQ(EEf4r9nUvta_[lPuZQT6rFAxL], oVre8I6UXc3b.nPO5Cxv_RSQQ, CeyMIoSyrpkQ)
else:
kXxsZxlIQUSQ = oVre8I6UXc3b._hparams.bottom.get(lPuZQT6rFAxL, PuPeNl0CuqOQ.get_bottom(bYPswhysd3s2))
RDoMxwW19f9W = RjR6XYrAEBW8 in jtsP69nFE2aJ
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x97\x88Pn\x01Z]#\x1d\xb7=\x8b\xa4-'), chr(4351 - 4251) + chr(2421 - 2320) + chr(99) + '\x6f' + '\x64' + chr(101))(chr(0b1110101) + '\164' + '\146' + chr(45) + '\070'))(RjR6XYrAEBW8, reuse=RDoMxwW19f9W) as qGaVI8v_Oz7A:
xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xbe\x88Fc?NP4+\xa5<\x88\xb1\x17g\xbejK?'), chr(9309 - 9209) + chr(101) + chr(99) + chr(0b1100111 + 0o10) + chr(3102 - 3002) + chr(0b111101 + 0o50))('\165' + chr(5689 - 5573) + '\146' + '\055' + chr(0b111000)))(RjR6XYrAEBW8, qGaVI8v_Oz7A)
iBNBQu99lQEa(xafqLlk3kkUe(SXOLrMavuUCe(b"\xb5\x9bCi\x13^^4/\xad0\x83\xf4.q\xbcqN(\x8e'\xa4H\xb8\x0e\xa7\xc086&\x17\xde\xbeA\xe3\xfd^\x07\x10\xfe"), chr(0b1101 + 0o127) + chr(101) + chr(0b1100011) + chr(0b100 + 0o153) + chr(8382 - 8282) + chr(0b1011 + 0o132))('\165' + chr(116) + chr(102) + '\x2d' + chr(0b111000)), lPuZQT6rFAxL, RjR6XYrAEBW8)
kRp2XqtfzLyV[lPuZQT6rFAxL] = kXxsZxlIQUSQ(EEf4r9nUvta_[lPuZQT6rFAxL], oVre8I6UXc3b.nPO5Cxv_RSQQ, CeyMIoSyrpkQ)
xafqLlk3kkUe(jtsP69nFE2aJ, xafqLlk3kkUe(SXOLrMavuUCe(b'\x80\x99Rb\x0e\\'), chr(0b1000 + 0o134) + chr(101) + chr(0b10111 + 0o114) + '\x6f' + chr(0b1100100) + chr(6391 - 6290))('\165' + '\x74' + chr(10183 - 10081) + chr(1810 - 1765) + chr(56)))(RjR6XYrAEBW8)
for K3J4ZwSlE0sT in EEf4r9nUvta_:
if K3J4ZwSlE0sT not in kRp2XqtfzLyV:
kRp2XqtfzLyV[K3J4ZwSlE0sT] = EEf4r9nUvta_[K3J4ZwSlE0sT]
else:
kRp2XqtfzLyV[K3J4ZwSlE0sT + xafqLlk3kkUe(SXOLrMavuUCe(b'\xbe\x9bCp'), '\144' + chr(6439 - 6338) + '\143' + chr(3383 - 3272) + '\x64' + chr(6286 - 6185))('\165' + chr(116) + chr(102) + '\x2d' + chr(0b111000))] = EEf4r9nUvta_[K3J4ZwSlE0sT]
return kRp2XqtfzLyV
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.top
|
def top(self, body_output, features):
"""Computes logits given body output and features.
Args:
body_output: dict of str to Tensor, comprising one key-value pair for each
target. Each value denotes the target's pre-logit activations.
Alternatively, it may be a single Tensor denoting the pre-logits for
that target.
features: dict of str to Tensor. Typically it is the preprocessed data
batch after Problem's preprocess_example().
Returns:
logits: dict of str to Tensor, denoting each logits for each target; or
a single Tensor denoting the logits for that target.
When targets are generated at training time:
logits == {
"self_generated_targets": <generated targets tensor>
"logits": <original logits Tensor or dict>
}
"""
if isinstance(body_output, dict):
logits = {}
for k, v in six.iteritems(body_output):
# TODO(aidangomez): share variables here?
with tf.variable_scope(k) as top_vs:
self._add_variable_scope("top_%s" % k, top_vs)
logits[k] = self._top_single(v, k, features)
return logits
else:
return self._top_single(body_output, "targets", features)
|
python
|
def top(self, body_output, features):
"""Computes logits given body output and features.
Args:
body_output: dict of str to Tensor, comprising one key-value pair for each
target. Each value denotes the target's pre-logit activations.
Alternatively, it may be a single Tensor denoting the pre-logits for
that target.
features: dict of str to Tensor. Typically it is the preprocessed data
batch after Problem's preprocess_example().
Returns:
logits: dict of str to Tensor, denoting each logits for each target; or
a single Tensor denoting the logits for that target.
When targets are generated at training time:
logits == {
"self_generated_targets": <generated targets tensor>
"logits": <original logits Tensor or dict>
}
"""
if isinstance(body_output, dict):
logits = {}
for k, v in six.iteritems(body_output):
# TODO(aidangomez): share variables here?
with tf.variable_scope(k) as top_vs:
self._add_variable_scope("top_%s" % k, top_vs)
logits[k] = self._top_single(v, k, features)
return logits
else:
return self._top_single(body_output, "targets", features)
|
[
"def",
"top",
"(",
"self",
",",
"body_output",
",",
"features",
")",
":",
"if",
"isinstance",
"(",
"body_output",
",",
"dict",
")",
":",
"logits",
"=",
"{",
"}",
"for",
"k",
",",
"v",
"in",
"six",
".",
"iteritems",
"(",
"body_output",
")",
":",
"# TODO(aidangomez): share variables here?",
"with",
"tf",
".",
"variable_scope",
"(",
"k",
")",
"as",
"top_vs",
":",
"self",
".",
"_add_variable_scope",
"(",
"\"top_%s\"",
"%",
"k",
",",
"top_vs",
")",
"logits",
"[",
"k",
"]",
"=",
"self",
".",
"_top_single",
"(",
"v",
",",
"k",
",",
"features",
")",
"return",
"logits",
"else",
":",
"return",
"self",
".",
"_top_single",
"(",
"body_output",
",",
"\"targets\"",
",",
"features",
")"
] |
Computes logits given body output and features.
Args:
body_output: dict of str to Tensor, comprising one key-value pair for each
target. Each value denotes the target's pre-logit activations.
Alternatively, it may be a single Tensor denoting the pre-logits for
that target.
features: dict of str to Tensor. Typically it is the preprocessed data
batch after Problem's preprocess_example().
Returns:
logits: dict of str to Tensor, denoting each logits for each target; or
a single Tensor denoting the logits for that target.
When targets are generated at training time:
logits == {
"self_generated_targets": <generated targets tensor>
"logits": <original logits Tensor or dict>
}
|
[
"Computes",
"logits",
"given",
"body",
"output",
"and",
"features",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L583-L612
|
train
|
Computes logits given body output and features.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b11010 + 0o125) + chr(0b110010) + chr(54) + '\x34', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110001) + '\x35' + '\x37', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(1479 - 1430) + chr(0b101010 + 0o11) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + '\063' + chr(928 - 875) + '\065', 7747 - 7739), ehT0Px3KOsy9(chr(48) + chr(0b101110 + 0o101) + chr(51) + '\060' + chr(0b110100), 0b1000), ehT0Px3KOsy9('\x30' + chr(1353 - 1242) + chr(465 - 414) + chr(54) + chr(0b110001), 0o10), ehT0Px3KOsy9(chr(0b1000 + 0o50) + chr(4434 - 4323) + chr(49) + chr(834 - 779), 29795 - 29787), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062' + '\x37' + '\x30', 0b1000), ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(0b1011011 + 0o24) + chr(54) + '\066', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001) + chr(0b1101 + 0o44) + chr(0b100110 + 0o16), 0b1000), ehT0Px3KOsy9(chr(2153 - 2105) + '\x6f' + chr(54) + chr(50), 0o10), ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(111) + chr(0b100001 + 0o21) + chr(2695 - 2643), 26496 - 26488), ehT0Px3KOsy9(chr(1790 - 1742) + chr(0b1011001 + 0o26) + '\061' + chr(0b101110 + 0o3), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(51) + chr(1979 - 1930) + chr(929 - 881), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(10142 - 10031) + chr(0b110111) + chr(53), 26612 - 26604), ehT0Px3KOsy9(chr(0b100001 + 0o17) + chr(0b1101111) + chr(0b110111) + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b101011 + 0o7) + '\063' + '\067', 55296 - 55288), ehT0Px3KOsy9(chr(48) + chr(0b100111 + 0o110) + '\x31' + '\x30' + '\x30', 0o10), ehT0Px3KOsy9(chr(1509 - 1461) + chr(5154 - 5043) + '\062' + chr(0b110110) + chr(1250 - 1199), 13814 - 13806), ehT0Px3KOsy9(chr(343 - 295) + chr(0b1101111) + chr(51) + chr(2527 - 2472), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\061' + '\x31' + chr(2000 - 1948), 8), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\062' + chr(0b100111 + 0o16) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(54) + '\x37', 34660 - 34652), ehT0Px3KOsy9(chr(48) + '\157' + chr(2107 - 2056) + '\x33' + chr(2788 - 2733), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110010) + chr(0b101111 + 0o7) + '\066', ord("\x08")), ehT0Px3KOsy9(chr(437 - 389) + '\157' + chr(49) + chr(0b110111) + chr(0b1011 + 0o46), 0o10), ehT0Px3KOsy9('\060' + chr(427 - 316) + chr(0b10100 + 0o35) + '\x35' + chr(48), 0o10), ehT0Px3KOsy9(chr(115 - 67) + chr(0b1101111) + chr(0b110001) + chr(912 - 864) + chr(0b110101), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(0b101101 + 0o102) + chr(0b110110) + '\x31', 0b1000), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(0b1100000 + 0o17) + '\x31' + '\061' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(1350 - 1302) + '\x6f' + '\062' + chr(49) + chr(0b101111 + 0o2), 33939 - 33931), ehT0Px3KOsy9('\x30' + chr(111) + chr(573 - 524) + chr(1459 - 1406) + '\061', 32210 - 32202), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(51) + '\065' + chr(0b10111 + 0o37), 41652 - 41644), ehT0Px3KOsy9(chr(789 - 741) + chr(111) + '\062' + chr(0b10110 + 0o37) + chr(0b100000 + 0o23), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b10100 + 0o133) + chr(0b110011) + chr(53) + '\066', 8), ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(0b100000 + 0o117) + '\x34' + chr(0b10110 + 0o33), 31999 - 31991), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b101000 + 0o13) + chr(2663 - 2608) + chr(2134 - 2085), 0o10), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(3281 - 3170) + chr(50) + chr(1757 - 1707) + '\067', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\062' + chr(0b1011 + 0o47) + chr(51), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(419 - 371) + chr(12094 - 11983) + chr(0b110101) + chr(48), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x9a'), chr(0b1000001 + 0o43) + chr(0b1000110 + 0o37) + '\x63' + chr(0b1101111) + '\x64' + chr(761 - 660))('\x75' + chr(0b1011011 + 0o31) + '\146' + chr(45) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def qxrVBjeryNEZ(oVre8I6UXc3b, KLYceYPUcF5e, EEf4r9nUvta_):
if PlSM16l2KDPD(KLYceYPUcF5e, wLqBDw8l0eIm):
wF9nmvjsKjYM = {}
for (OolUPRJhRaJd, cMbll0QYhULo) in xafqLlk3kkUe(sYby0kpfssd4, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdd\x1b:\x9f\xad\x85j\xd5E'), chr(823 - 723) + chr(101) + chr(0b1100011) + chr(0b1001100 + 0o43) + chr(0b1100100) + '\x65')(chr(8816 - 8699) + chr(0b111011 + 0o71) + chr(102) + chr(0b101101) + '\070'))(KLYceYPUcF5e):
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc2\x0e-\x84\xa5\x93c\xddi?Q{y\xad'), '\144' + '\x65' + chr(0b1100011) + chr(0b1000101 + 0o52) + '\x64' + chr(0b100 + 0o141))(chr(117) + chr(116) + chr(0b1100110) + chr(0b1100 + 0o41) + chr(0b10011 + 0o45)))(OolUPRJhRaJd) as qhwHZC31wpme:
xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xeb\x0e;\x89\x9b\x87n\xca_-Pxl\x97N\\\xf00\xd6'), chr(0b111110 + 0o46) + '\145' + '\143' + chr(0b1101111) + chr(3368 - 3268) + chr(0b1100101))('\165' + chr(0b1011101 + 0o27) + chr(102) + chr(385 - 340) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xc0\x00/\xb2\xe1\x82'), chr(8880 - 8780) + chr(0b1100101) + chr(0b1100011) + chr(4171 - 4060) + '\x64' + chr(0b1100101))(chr(11935 - 11818) + chr(3043 - 2927) + chr(102) + '\055' + '\x38') % OolUPRJhRaJd, qhwHZC31wpme)
wF9nmvjsKjYM[OolUPRJhRaJd] = oVre8I6UXc3b._top_single(cMbll0QYhULo, OolUPRJhRaJd, EEf4r9nUvta_)
return wF9nmvjsKjYM
else:
return xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xeb\x1b0\x9d\x9b\x82f\xd6Q W'), chr(0b1100100) + '\145' + chr(99) + '\x6f' + chr(0b1100100) + chr(0b1101 + 0o130))(chr(5821 - 5704) + chr(10426 - 10310) + chr(102) + '\x2d' + chr(2696 - 2640)))(KLYceYPUcF5e, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc0\x0e-\x8a\xa1\x85|'), chr(0b1100100) + chr(4561 - 4460) + '\x63' + chr(0b1100110 + 0o11) + chr(0b1100100) + '\145')(chr(12338 - 12221) + '\x74' + '\146' + chr(45) + chr(0b10001 + 0o47)), EEf4r9nUvta_)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.optimize
|
def optimize(self, loss, num_async_replicas=1, use_tpu=False):
"""Return a training op minimizing loss."""
lr = learning_rate.learning_rate_schedule(self.hparams)
if num_async_replicas > 1:
log_info("Dividing learning rate by num_async_replicas: %d",
num_async_replicas)
lr /= math.sqrt(float(num_async_replicas))
train_op = optimize.optimize(loss, lr, self.hparams, use_tpu=use_tpu)
return train_op
|
python
|
def optimize(self, loss, num_async_replicas=1, use_tpu=False):
"""Return a training op minimizing loss."""
lr = learning_rate.learning_rate_schedule(self.hparams)
if num_async_replicas > 1:
log_info("Dividing learning rate by num_async_replicas: %d",
num_async_replicas)
lr /= math.sqrt(float(num_async_replicas))
train_op = optimize.optimize(loss, lr, self.hparams, use_tpu=use_tpu)
return train_op
|
[
"def",
"optimize",
"(",
"self",
",",
"loss",
",",
"num_async_replicas",
"=",
"1",
",",
"use_tpu",
"=",
"False",
")",
":",
"lr",
"=",
"learning_rate",
".",
"learning_rate_schedule",
"(",
"self",
".",
"hparams",
")",
"if",
"num_async_replicas",
">",
"1",
":",
"log_info",
"(",
"\"Dividing learning rate by num_async_replicas: %d\"",
",",
"num_async_replicas",
")",
"lr",
"/=",
"math",
".",
"sqrt",
"(",
"float",
"(",
"num_async_replicas",
")",
")",
"train_op",
"=",
"optimize",
".",
"optimize",
"(",
"loss",
",",
"lr",
",",
"self",
".",
"hparams",
",",
"use_tpu",
"=",
"use_tpu",
")",
"return",
"train_op"
] |
Return a training op minimizing loss.
|
[
"Return",
"a",
"training",
"op",
"minimizing",
"loss",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L710-L718
|
train
|
Return a training op minimizing loss.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2067 - 2018) + '\x35' + chr(0b1000 + 0o55), 54594 - 54586), ehT0Px3KOsy9('\x30' + chr(7442 - 7331) + chr(55) + '\067', 0b1000), ehT0Px3KOsy9(chr(48) + chr(12135 - 12024) + chr(802 - 751) + chr(0b110101) + '\x35', ord("\x08")), ehT0Px3KOsy9(chr(0b100011 + 0o15) + '\x6f' + '\x33' + '\061' + '\062', 24429 - 24421), ehT0Px3KOsy9(chr(1982 - 1934) + chr(111) + '\x31' + '\x37' + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(111) + chr(1982 - 1931) + '\x32' + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(1542 - 1492) + '\x36' + chr(0b1111 + 0o46), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\064' + '\064', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(87 - 37) + chr(0b11 + 0o60), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b101110 + 0o101) + chr(1691 - 1638) + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(50) + chr(49) + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(0b1011 + 0o45) + '\157' + chr(50) + chr(0b110011 + 0o1) + '\067', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(49) + '\060', 9785 - 9777), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + chr(50) + chr(48), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x34' + '\061', 9492 - 9484), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + chr(0b10010 + 0o42) + chr(1544 - 1489), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b10 + 0o155) + '\061' + chr(475 - 425) + '\061', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(1080 - 969) + '\064' + '\x34', 8), ehT0Px3KOsy9('\x30' + chr(0b1010000 + 0o37) + chr(1250 - 1202), 0b1000), ehT0Px3KOsy9(chr(1603 - 1555) + chr(0b1101111) + '\062' + '\x35' + chr(0b110011), 0o10), ehT0Px3KOsy9(chr(267 - 219) + '\157' + chr(2370 - 2319) + chr(0b110101) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(183 - 135) + chr(0b1101111) + '\065' + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110010) + chr(0b110101) + '\065', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\061' + chr(0b10000 + 0o42) + chr(2273 - 2218), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\062' + chr(0b110110) + chr(1286 - 1234), 18076 - 18068), ehT0Px3KOsy9(chr(0b11101 + 0o23) + '\157' + chr(756 - 707) + '\065' + chr(51), 4045 - 4037), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + '\060' + chr(548 - 500), 38927 - 38919), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + chr(0b110100) + '\x34', 21816 - 21808), ehT0Px3KOsy9(chr(1262 - 1214) + '\157' + '\x33' + chr(53) + chr(0b110000), 2825 - 2817), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(0b1101001 + 0o6) + chr(0b110001) + chr(52) + '\063', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x33' + '\063' + chr(0b110110), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b10100 + 0o36) + chr(0b10100 + 0o42) + chr(2750 - 2696), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110011) + chr(0b110001) + chr(483 - 433), 8), ehT0Px3KOsy9(chr(48) + '\157' + chr(2437 - 2382) + '\x31', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b101011 + 0o6) + '\x32' + chr(0b110001), 8), ehT0Px3KOsy9('\060' + '\157' + chr(678 - 629) + '\x37' + chr(0b101000 + 0o12), 58034 - 58026), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(908 - 857) + chr(619 - 570) + chr(1031 - 981), 8), ehT0Px3KOsy9('\060' + '\157' + chr(50) + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\061' + chr(0b110111) + chr(0b101111 + 0o5), ord("\x08")), ehT0Px3KOsy9('\060' + chr(1180 - 1069) + chr(50) + chr(0b110011), 8)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110101) + chr(0b110000), 12031 - 12023)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'C'), chr(0b1100100) + chr(101) + chr(99) + chr(0b0 + 0o157) + chr(0b1010110 + 0o16) + chr(0b11010 + 0o113))(chr(0b1110101) + '\164' + chr(10221 - 10119) + chr(45) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def M4lwI8bLCQGq(oVre8I6UXc3b, YpO0BcZ6fMsf, FsJHiE9x_3jA=ehT0Px3KOsy9(chr(48) + chr(111) + chr(49), 47435 - 47427), L4eE7kczIJwa=ehT0Px3KOsy9(chr(975 - 927) + chr(0b1101111) + chr(48), 8)):
Zzs55KO_HKfp = QGSIpd_yUNzU.Lz_s7neUzM5V(oVre8I6UXc3b.n4ljua2gi1Pr)
if FsJHiE9x_3jA > ehT0Px3KOsy9(chr(1832 - 1784) + chr(8321 - 8210) + chr(49), 8):
iBNBQu99lQEa(xafqLlk3kkUe(SXOLrMavuUCe(b')V\x9b\xa3\x19\xbc\xdbC3@\xeb\xc6z\xfeH\x896\x04r\xca\x9aN\x8d5\xcdK\xae\xd1\x85<\xbf\xc5\xb3W2%\xb8r\x0c\x7f\x04\\\x8c\xb9G\xf5\x90@'), chr(0b1100100) + chr(101) + '\143' + '\157' + chr(8920 - 8820) + '\145')('\165' + chr(0b1001111 + 0o45) + chr(102) + '\x2d' + '\x38'), FsJHiE9x_3jA)
Zzs55KO_HKfp /= yhiZVkosCjBm.sqrt(kkSX4ccExqw4(FsJHiE9x_3jA))
_sRzZqw7qhHl = M4lwI8bLCQGq.optimize(YpO0BcZ6fMsf, Zzs55KO_HKfp, oVre8I6UXc3b.n4ljua2gi1Pr, use_tpu=L4eE7kczIJwa)
return _sRzZqw7qhHl
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.set_mode
|
def set_mode(self, mode):
"""Set hparams with the given mode."""
log_info("Setting T2TModel mode to '%s'", mode)
hparams = hparams_lib.copy_hparams(self._original_hparams)
hparams.add_hparam("mode", mode)
# When not in training mode, set all forms of dropout to zero.
if mode != tf.estimator.ModeKeys.TRAIN:
for key in hparams.values():
if key.endswith("dropout") or key == "label_smoothing":
log_info("Setting hparams.%s to 0.0", key)
setattr(hparams, key, 0.0)
self._hparams = hparams
|
python
|
def set_mode(self, mode):
"""Set hparams with the given mode."""
log_info("Setting T2TModel mode to '%s'", mode)
hparams = hparams_lib.copy_hparams(self._original_hparams)
hparams.add_hparam("mode", mode)
# When not in training mode, set all forms of dropout to zero.
if mode != tf.estimator.ModeKeys.TRAIN:
for key in hparams.values():
if key.endswith("dropout") or key == "label_smoothing":
log_info("Setting hparams.%s to 0.0", key)
setattr(hparams, key, 0.0)
self._hparams = hparams
|
[
"def",
"set_mode",
"(",
"self",
",",
"mode",
")",
":",
"log_info",
"(",
"\"Setting T2TModel mode to '%s'\"",
",",
"mode",
")",
"hparams",
"=",
"hparams_lib",
".",
"copy_hparams",
"(",
"self",
".",
"_original_hparams",
")",
"hparams",
".",
"add_hparam",
"(",
"\"mode\"",
",",
"mode",
")",
"# When not in training mode, set all forms of dropout to zero.",
"if",
"mode",
"!=",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"TRAIN",
":",
"for",
"key",
"in",
"hparams",
".",
"values",
"(",
")",
":",
"if",
"key",
".",
"endswith",
"(",
"\"dropout\"",
")",
"or",
"key",
"==",
"\"label_smoothing\"",
":",
"log_info",
"(",
"\"Setting hparams.%s to 0.0\"",
",",
"key",
")",
"setattr",
"(",
"hparams",
",",
"key",
",",
"0.0",
")",
"self",
".",
"_hparams",
"=",
"hparams"
] |
Set hparams with the given mode.
|
[
"Set",
"hparams",
"with",
"the",
"given",
"mode",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L720-L731
|
train
|
Set hparams with the given mode.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b100001 + 0o21) + chr(50), 0b1000), ehT0Px3KOsy9(chr(2023 - 1975) + chr(111) + '\x32' + chr(52) + '\060', 0o10), ehT0Px3KOsy9(chr(2105 - 2057) + chr(4389 - 4278) + '\061' + chr(53) + chr(0b101 + 0o57), 51352 - 51344), ehT0Px3KOsy9(chr(1367 - 1319) + '\x6f' + chr(0b110011) + '\064' + '\x30', 6189 - 6181), ehT0Px3KOsy9('\x30' + chr(111) + '\063' + chr(53) + '\062', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b100100 + 0o113) + '\061' + chr(0b101100 + 0o10) + chr(0b110000), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010) + chr(49) + chr(1803 - 1752), 49940 - 49932), ehT0Px3KOsy9(chr(2196 - 2148) + chr(111) + chr(49) + chr(0b111 + 0o52) + chr(2199 - 2151), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\065' + chr(737 - 687), 0b1000), ehT0Px3KOsy9(chr(1936 - 1888) + chr(0b1000010 + 0o55) + chr(54) + '\x30', 50000 - 49992), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + chr(2161 - 2111), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b100001 + 0o21) + '\x34' + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(1696 - 1648) + '\157' + '\061' + '\066' + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(10938 - 10827) + chr(0b100000 + 0o21) + '\x31' + chr(1111 - 1057), 57821 - 57813), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110010) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(50) + chr(48) + chr(2560 - 2509), 44870 - 44862), ehT0Px3KOsy9(chr(48) + chr(8912 - 8801) + chr(49) + chr(1658 - 1605) + '\064', 8), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(11980 - 11869) + chr(0b110011) + chr(0b10101 + 0o37) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b111 + 0o54) + chr(51) + chr(0b11 + 0o55), 2707 - 2699), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(0b1101111) + chr(108 - 59) + chr(0b101101 + 0o4) + chr(55), 63031 - 63023), ehT0Px3KOsy9('\060' + '\157' + '\x31' + '\061', 0b1000), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(10931 - 10820) + chr(51) + chr(51) + '\067', 0b1000), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(0b111110 + 0o61) + '\x32' + chr(1421 - 1367) + chr(0b101 + 0o55), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1100100 + 0o13) + chr(564 - 515) + '\x37' + '\x36', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1010 + 0o145) + chr(0b101110 + 0o5) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + chr(0b110010) + chr(1885 - 1833), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(50) + chr(806 - 758) + chr(49), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101001 + 0o6) + chr(0b1001 + 0o50) + chr(1459 - 1407) + '\060', 8), ehT0Px3KOsy9(chr(48) + chr(111) + chr(1322 - 1273) + chr(49) + chr(1355 - 1304), 3295 - 3287), ehT0Px3KOsy9('\060' + chr(4100 - 3989) + chr(0b110011) + chr(51) + chr(0b111 + 0o52), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b111100 + 0o63) + '\062' + chr(1157 - 1105) + '\064', 0b1000), ehT0Px3KOsy9(chr(1720 - 1672) + chr(0b110011 + 0o74) + chr(0b110010) + chr(467 - 413) + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(0b1101111) + chr(0b10001 + 0o46) + '\065', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\063' + chr(48), 57085 - 57077), ehT0Px3KOsy9('\x30' + chr(0b1001011 + 0o44) + '\x32' + chr(1752 - 1702) + '\x35', 0b1000), ehT0Px3KOsy9(chr(48) + chr(11881 - 11770) + chr(1088 - 1037) + chr(819 - 769) + chr(51), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(50) + chr(55) + '\x34', ord("\x08")), ehT0Px3KOsy9('\060' + chr(11506 - 11395) + '\061' + chr(2161 - 2110) + chr(0b110100), 25915 - 25907), ehT0Px3KOsy9('\x30' + '\157' + '\063' + chr(0b101110 + 0o11) + chr(1845 - 1790), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b11100 + 0o123) + '\x32' + chr(0b11010 + 0o35) + chr(2865 - 2810), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\065' + chr(1851 - 1803), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'<'), '\x64' + chr(0b1010000 + 0o25) + '\143' + chr(0b1011000 + 0o27) + chr(0b111011 + 0o51) + chr(0b1000001 + 0o44))(chr(0b1010101 + 0o40) + '\164' + '\x66' + '\x2d' + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def TtFAsiR8IOG5(oVre8I6UXc3b, holLFgwB7vsP):
iBNBQu99lQEa(xafqLlk3kkUe(SXOLrMavuUCe(b'A\xc5\x18\x94\x83\x0e\xech~\x02N\xee\x18*v\xefL\x8b\xb7+l\xff\xc6\x11<\xb3\xa4\x13l'), chr(3735 - 3635) + '\145' + chr(6373 - 6274) + chr(111) + chr(4259 - 4159) + chr(3435 - 3334))(chr(0b1 + 0o164) + chr(0b1110100) + '\146' + chr(0b101101 + 0o0) + chr(0b111000)), holLFgwB7vsP)
n4ljua2gi1Pr = BMqTy4F2E1fh.copy_hparams(oVre8I6UXc3b._original_hparams)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b's\xc4\x08\xbf\x82\x10\xea:K]'), chr(3406 - 3306) + chr(101) + chr(0b1000100 + 0o37) + '\x6f' + chr(100) + '\x65')(chr(0b1110101) + chr(0b100001 + 0o123) + chr(102) + chr(0b101101) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x7f\xcf\x08\x85'), '\x64' + '\145' + '\x63' + chr(0b101111 + 0o100) + '\x64' + '\x65')(chr(0b11101 + 0o130) + '\164' + chr(0b1100 + 0o132) + chr(45) + chr(0b111000)), holLFgwB7vsP)
if holLFgwB7vsP != xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'F\xf2-\xa9\xa4'), chr(0b1100100) + '\145' + chr(0b100010 + 0o101) + chr(111) + chr(4857 - 4757) + '\145')('\165' + chr(116) + chr(0b1010011 + 0o23) + chr(1114 - 1069) + chr(0b111000))):
for K3J4ZwSlE0sT in xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'A\xf0\x02\xa3\xa4\x15\xbe|b\x01~\xc1'), chr(0b1000001 + 0o43) + chr(6110 - 6009) + chr(99) + chr(111) + chr(5310 - 5210) + chr(0b1100101))('\x75' + chr(116) + chr(102) + chr(1235 - 1190) + chr(0b10111 + 0o41)))():
if xafqLlk3kkUe(K3J4ZwSlE0sT, xafqLlk3kkUe(SXOLrMavuUCe(b'w\xce\x08\x93\x9d\t\xff '), chr(100) + chr(101) + chr(706 - 607) + chr(8527 - 8416) + chr(100) + '\145')(chr(5625 - 5508) + chr(6498 - 6382) + chr(0b1011011 + 0o13) + '\055' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'v\xd2\x03\x90\x85\x15\xff'), '\x64' + chr(101) + chr(99) + '\157' + chr(0b1100100) + chr(101))('\x75' + '\164' + '\x66' + chr(45) + '\070')) or K3J4ZwSlE0sT == xafqLlk3kkUe(SXOLrMavuUCe(b'~\xc1\x0e\x85\x86?\xf8%E_n\xcb\x1e t'), chr(5908 - 5808) + chr(6995 - 6894) + chr(0b1100011) + chr(0b1101111) + '\x64' + '\x65')(chr(117) + chr(0b1110100) + chr(102) + chr(0b1101 + 0o40) + chr(56)):
iBNBQu99lQEa(xafqLlk3kkUe(SXOLrMavuUCe(b'A\xc5\x18\x94\x83\x0e\xechB@{\xd1\x16#`\xadI\x95\xf8;f\xff\x82P,'), chr(0b1100100) + '\145' + chr(0b1000011 + 0o40) + '\157' + chr(710 - 610) + chr(0b1010100 + 0o21))('\165' + chr(116) + chr(0b11001 + 0o115) + chr(255 - 210) + '\070'), K3J4ZwSlE0sT)
t0rOMsrOC7R_(n4ljua2gi1Pr, K3J4ZwSlE0sT, 0.0)
oVre8I6UXc3b.nPO5Cxv_RSQQ = n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.eval_autoregressive
|
def eval_autoregressive(self, features=None, decode_length=50):
"""Autoregressive eval.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
Returns:
logits: `Tensor`
losses: a dictionary: {loss-name (string): floating point `Scalar`}.
Contains a single key "training".
"""
results = self._slow_greedy_infer(features, decode_length=decode_length)
return results["logits"], results["losses"]
|
python
|
def eval_autoregressive(self, features=None, decode_length=50):
"""Autoregressive eval.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
Returns:
logits: `Tensor`
losses: a dictionary: {loss-name (string): floating point `Scalar`}.
Contains a single key "training".
"""
results = self._slow_greedy_infer(features, decode_length=decode_length)
return results["logits"], results["losses"]
|
[
"def",
"eval_autoregressive",
"(",
"self",
",",
"features",
"=",
"None",
",",
"decode_length",
"=",
"50",
")",
":",
"results",
"=",
"self",
".",
"_slow_greedy_infer",
"(",
"features",
",",
"decode_length",
"=",
"decode_length",
")",
"return",
"results",
"[",
"\"logits\"",
"]",
",",
"results",
"[",
"\"losses\"",
"]"
] |
Autoregressive eval.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
Returns:
logits: `Tensor`
losses: a dictionary: {loss-name (string): floating point `Scalar`}.
Contains a single key "training".
|
[
"Autoregressive",
"eval",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L737-L752
|
train
|
Autoregressive eval.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110011) + chr(50) + chr(0b10011 + 0o36), 27239 - 27231), ehT0Px3KOsy9('\x30' + chr(12001 - 11890) + chr(862 - 811) + '\x37' + chr(704 - 654), 50481 - 50473), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(111) + chr(0b110010) + chr(0b11101 + 0o25) + chr(2046 - 1998), 0b1000), ehT0Px3KOsy9(chr(0b101011 + 0o5) + '\x6f' + chr(1573 - 1522) + '\065' + '\060', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(54) + chr(0b110000), 0b1000), ehT0Px3KOsy9(chr(1023 - 975) + chr(111) + chr(0b110110) + '\x31', 0b1000), ehT0Px3KOsy9(chr(0b10 + 0o56) + '\x6f' + '\063' + '\x35' + chr(1507 - 1455), 0o10), ehT0Px3KOsy9(chr(0b1 + 0o57) + '\x6f' + chr(0b110010) + chr(0b110100) + chr(1217 - 1163), 16712 - 16704), ehT0Px3KOsy9(chr(780 - 732) + chr(6293 - 6182) + '\x33' + chr(49) + chr(1922 - 1868), 6213 - 6205), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\066' + chr(53), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(384 - 333) + chr(51) + chr(0b11000 + 0o37), 55311 - 55303), ehT0Px3KOsy9('\x30' + chr(10374 - 10263) + chr(0b110001) + chr(0b11101 + 0o26) + chr(0b110100 + 0o1), 45786 - 45778), ehT0Px3KOsy9('\060' + '\157' + '\x34' + '\063', 32652 - 32644), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001) + chr(2220 - 2165) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(2138 - 2090) + chr(7583 - 7472) + chr(0b10110 + 0o34) + chr(55) + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(1311 - 1263) + chr(0b1101111) + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x31' + '\x31' + '\x36', 0o10), ehT0Px3KOsy9(chr(0b100001 + 0o17) + chr(0b1101111) + '\063' + '\062' + '\x36', 0o10), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\157' + '\063' + chr(0b110011) + chr(0b100010 + 0o24), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(7730 - 7619) + '\061' + chr(0b110010) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(1465 - 1417) + chr(111) + '\x32' + chr(730 - 678) + '\063', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(50) + chr(414 - 360), 58839 - 58831), ehT0Px3KOsy9(chr(1720 - 1672) + chr(0b1101111) + chr(51) + chr(2160 - 2112) + chr(55), 22005 - 21997), ehT0Px3KOsy9('\060' + '\157' + '\062' + '\x35' + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(2048 - 1995), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110 + 0o53) + '\066' + chr(0b110110), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\062' + '\x36' + chr(54), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x33' + chr(53) + chr(54), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(50) + chr(915 - 864) + chr(0b100 + 0o61), 0b1000), ehT0Px3KOsy9(chr(243 - 195) + '\157' + '\063' + chr(0b10000 + 0o46) + chr(54), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\066' + '\066', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1634 - 1584) + chr(53), 0o10), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(111) + chr(0b0 + 0o63) + '\062' + chr(50), 9396 - 9388), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011) + '\063' + '\x33', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b10011 + 0o44) + chr(48), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x33' + '\x37' + chr(1773 - 1720), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\061' + '\061' + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + '\x31', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(52) + chr(0b110011 + 0o2), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49) + '\064', ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(53) + chr(0b110000), 30228 - 30220)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xff'), chr(0b1100100) + '\x65' + '\143' + chr(111) + '\144' + '\x65')(chr(117) + '\164' + '\146' + chr(45) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def i54u9J66Xm15(oVre8I6UXc3b, EEf4r9nUvta_=None, U6Ej34SVvx1Y=ehT0Px3KOsy9(chr(48) + chr(11343 - 11232) + '\x36' + chr(0b1010 + 0o50), ord("\x08"))):
iIGKX2zSEGYP = oVre8I6UXc3b._slow_greedy_infer(EEf4r9nUvta_, decode_length=U6Ej34SVvx1Y)
return (iIGKX2zSEGYP[xafqLlk3kkUe(SXOLrMavuUCe(b'\xbd\x95\xf7\x81DB'), '\x64' + chr(0b1100101) + chr(7485 - 7386) + chr(6260 - 6149) + '\x64' + chr(0b1100101))('\165' + chr(0b1110100) + '\x66' + chr(576 - 531) + chr(56))], iIGKX2zSEGYP[xafqLlk3kkUe(SXOLrMavuUCe(b'\xbd\x95\xe3\x9bUB'), '\144' + '\145' + chr(0b110100 + 0o57) + chr(111) + chr(0b1100100) + '\145')('\x75' + chr(0b1110100) + chr(2597 - 2495) + chr(594 - 549) + chr(0b100100 + 0o24))])
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.infer
|
def infer(self,
features=None,
decode_length=50,
beam_size=1,
top_beams=1,
alpha=0.0,
use_tpu=False):
"""A inference method.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: bool, whether to build the inference graph for TPU.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": decoding log probs from the beam search,
None if using greedy decoding (beam_size=1)
}
if slow greedy decoding is used then the dict will also contain {
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`
}
"""
set_custom_getter_compose(self._custom_getter)
with self._eager_var_store.as_default():
# TODO(rsepassi): Make decoding work with real-valued model outputs
# (i.e. if the target modality is RealModality).
self.prepare_features_for_infer(features)
if not self.has_input and beam_size > 1:
log_warn("Beam searching for a model with no inputs.")
if not self.has_input and self.hparams.sampling_method != "random":
log_warn("Non-random sampling for a model with no inputs.")
self._fill_problem_hparams_features(features)
if self._problem_hparams:
target_modality = self._problem_hparams.modality["targets"]
if target_modality == modalities.ModalityType.CLASS_LABEL:
beam_size = 1 # No use to run beam-search for a single class.
if beam_size == 1:
log_info("Greedy Decoding")
results = self._greedy_infer(features, decode_length, use_tpu)
else:
log_info("Beam Decoding with beam size %d" % beam_size)
results = self._beam_decode(features, decode_length, beam_size,
top_beams, alpha, use_tpu)
return results
|
python
|
def infer(self,
features=None,
decode_length=50,
beam_size=1,
top_beams=1,
alpha=0.0,
use_tpu=False):
"""A inference method.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: bool, whether to build the inference graph for TPU.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": decoding log probs from the beam search,
None if using greedy decoding (beam_size=1)
}
if slow greedy decoding is used then the dict will also contain {
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`
}
"""
set_custom_getter_compose(self._custom_getter)
with self._eager_var_store.as_default():
# TODO(rsepassi): Make decoding work with real-valued model outputs
# (i.e. if the target modality is RealModality).
self.prepare_features_for_infer(features)
if not self.has_input and beam_size > 1:
log_warn("Beam searching for a model with no inputs.")
if not self.has_input and self.hparams.sampling_method != "random":
log_warn("Non-random sampling for a model with no inputs.")
self._fill_problem_hparams_features(features)
if self._problem_hparams:
target_modality = self._problem_hparams.modality["targets"]
if target_modality == modalities.ModalityType.CLASS_LABEL:
beam_size = 1 # No use to run beam-search for a single class.
if beam_size == 1:
log_info("Greedy Decoding")
results = self._greedy_infer(features, decode_length, use_tpu)
else:
log_info("Beam Decoding with beam size %d" % beam_size)
results = self._beam_decode(features, decode_length, beam_size,
top_beams, alpha, use_tpu)
return results
|
[
"def",
"infer",
"(",
"self",
",",
"features",
"=",
"None",
",",
"decode_length",
"=",
"50",
",",
"beam_size",
"=",
"1",
",",
"top_beams",
"=",
"1",
",",
"alpha",
"=",
"0.0",
",",
"use_tpu",
"=",
"False",
")",
":",
"set_custom_getter_compose",
"(",
"self",
".",
"_custom_getter",
")",
"with",
"self",
".",
"_eager_var_store",
".",
"as_default",
"(",
")",
":",
"# TODO(rsepassi): Make decoding work with real-valued model outputs",
"# (i.e. if the target modality is RealModality).",
"self",
".",
"prepare_features_for_infer",
"(",
"features",
")",
"if",
"not",
"self",
".",
"has_input",
"and",
"beam_size",
">",
"1",
":",
"log_warn",
"(",
"\"Beam searching for a model with no inputs.\"",
")",
"if",
"not",
"self",
".",
"has_input",
"and",
"self",
".",
"hparams",
".",
"sampling_method",
"!=",
"\"random\"",
":",
"log_warn",
"(",
"\"Non-random sampling for a model with no inputs.\"",
")",
"self",
".",
"_fill_problem_hparams_features",
"(",
"features",
")",
"if",
"self",
".",
"_problem_hparams",
":",
"target_modality",
"=",
"self",
".",
"_problem_hparams",
".",
"modality",
"[",
"\"targets\"",
"]",
"if",
"target_modality",
"==",
"modalities",
".",
"ModalityType",
".",
"CLASS_LABEL",
":",
"beam_size",
"=",
"1",
"# No use to run beam-search for a single class.",
"if",
"beam_size",
"==",
"1",
":",
"log_info",
"(",
"\"Greedy Decoding\"",
")",
"results",
"=",
"self",
".",
"_greedy_infer",
"(",
"features",
",",
"decode_length",
",",
"use_tpu",
")",
"else",
":",
"log_info",
"(",
"\"Beam Decoding with beam size %d\"",
"%",
"beam_size",
")",
"results",
"=",
"self",
".",
"_beam_decode",
"(",
"features",
",",
"decode_length",
",",
"beam_size",
",",
"top_beams",
",",
"alpha",
",",
"use_tpu",
")",
"return",
"results"
] |
A inference method.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: bool, whether to build the inference graph for TPU.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": decoding log probs from the beam search,
None if using greedy decoding (beam_size=1)
}
if slow greedy decoding is used then the dict will also contain {
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`
}
|
[
"A",
"inference",
"method",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L761-L817
|
train
|
A method that performs the inference of the model.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b1110 + 0o42) + chr(111) + chr(0b110010) + '\065' + chr(53), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + '\x31' + '\065' + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011011 + 0o24) + chr(0b11 + 0o56) + '\x36' + chr(54), 2316 - 2308), ehT0Px3KOsy9(chr(693 - 645) + chr(0b11011 + 0o124) + chr(0b110101) + chr(0b101001 + 0o12), 61344 - 61336), ehT0Px3KOsy9(chr(1756 - 1708) + chr(12014 - 11903) + '\x33' + '\065' + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(225 - 174) + chr(50), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110010) + chr(53) + chr(49), 62082 - 62074), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x32' + '\060' + '\065', 12366 - 12358), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(0b1010001 + 0o36) + chr(1352 - 1302) + chr(55) + chr(0b10100 + 0o42), 0b1000), ehT0Px3KOsy9('\060' + chr(11981 - 11870) + chr(1753 - 1702) + '\x37' + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(0b100001 + 0o17) + chr(640 - 529) + chr(1067 - 1016) + chr(0b101 + 0o60) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110 + 0o54) + chr(0b101100 + 0o4) + '\x36', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b110111) + chr(54), 0b1000), ehT0Px3KOsy9(chr(0b11100 + 0o24) + chr(7361 - 7250) + chr(0b100001 + 0o21) + '\x31' + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110010) + chr(0b11011 + 0o31) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b0 + 0o61) + chr(0b110001) + '\x34', 50393 - 50385), ehT0Px3KOsy9('\060' + '\x6f' + chr(49) + chr(2221 - 2168) + chr(0b110101), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1010111 + 0o30) + chr(0b1 + 0o61) + chr(54) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(497 - 449) + '\157' + chr(50) + '\062' + chr(0b11001 + 0o35), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110001) + chr(2298 - 2247) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(3295 - 3184) + '\x33' + '\063' + '\x37', 15821 - 15813), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\062' + chr(0b110111) + chr(0b10011 + 0o37), 0b1000), ehT0Px3KOsy9(chr(48) + chr(134 - 23) + chr(50) + chr(2129 - 2080) + '\x36', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + '\x36' + chr(55), 8), ehT0Px3KOsy9(chr(1356 - 1308) + chr(0b11101 + 0o122) + chr(0b10110 + 0o35) + chr(0b110110) + chr(53), 0b1000), ehT0Px3KOsy9('\060' + chr(3717 - 3606) + chr(0b110010) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(11066 - 10955) + chr(49) + chr(554 - 502) + chr(0b1110 + 0o42), 48500 - 48492), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b100100 + 0o17) + chr(0b110001) + chr(1314 - 1260), 14871 - 14863), ehT0Px3KOsy9(chr(1831 - 1783) + '\x6f' + '\066' + '\064', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\x32' + '\x32' + chr(0b110110), 8), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x31' + chr(96 - 46) + '\x32', 0o10), ehT0Px3KOsy9('\060' + chr(4111 - 4000) + '\063' + '\061' + chr(0b101001 + 0o13), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b10100 + 0o35) + chr(1156 - 1106) + chr(53), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b110001) + chr(2011 - 1958) + chr(0b101111 + 0o2), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b10101 + 0o36) + chr(50) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b1000 + 0o51) + chr(0b110110) + chr(52), 0b1000), ehT0Px3KOsy9(chr(0b1111 + 0o41) + '\157' + chr(49) + '\060' + chr(0b10101 + 0o36), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101101 + 0o2) + chr(2374 - 2325) + chr(372 - 322), 25931 - 25923), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + chr(262 - 212) + chr(49), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(2478 - 2427) + chr(0b110011) + chr(1187 - 1135), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + '\157' + chr(0b100011 + 0o22) + chr(1630 - 1582), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x83'), '\x64' + chr(0b110100 + 0o61) + chr(3318 - 3219) + chr(0b1101111) + '\x64' + '\x65')('\165' + chr(0b1110100) + chr(102) + chr(0b101011 + 0o2) + chr(0b0 + 0o70)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def IhRMh3nN8G5I(oVre8I6UXc3b, EEf4r9nUvta_=None, U6Ej34SVvx1Y=ehT0Px3KOsy9('\x30' + '\157' + chr(2530 - 2476) + '\x32', 0b1000), PQZjDxhiHJGf=ehT0Px3KOsy9('\x30' + chr(12047 - 11936) + chr(0b110001), ord("\x08")), oC1hU_0mlSje=ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x31', 8), gDUX9w35YHFE=0.0, L4eE7kczIJwa=ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110000), 0b1000)):
C34rEtF0E667(xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf2\x0c\x03X\x9bu\xc4\xda\xd9\x99\x99\xf2[\xfe'), chr(0b1110 + 0o126) + chr(101) + chr(0b110101 + 0o56) + chr(111) + '\144' + chr(10175 - 10074))(chr(10328 - 10211) + chr(0b1110100) + chr(0b1100110) + '\055' + '\x38')))
with xafqLlk3kkUe(oVre8I6UXc3b._eager_var_store, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcc\x1c)O\x8a|\xc8\xf0\xd2\x88'), chr(8901 - 8801) + chr(6174 - 6073) + chr(0b1010000 + 0o23) + chr(111) + '\144' + chr(0b11001 + 0o114))(chr(0b1100111 + 0o16) + chr(0b1110100) + '\146' + chr(45) + '\070'))():
xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdd\x1d\x13[\x8eh\xcc\xda\xd8\x99\x8c\xf2K\xfe\xc4?2\xd1\xf9(\x80\xd3\xe0\xd1\x90a'), '\x64' + chr(0b1100101) + chr(0b1010101 + 0o16) + chr(1562 - 1451) + '\144' + chr(0b1100101))(chr(0b110101 + 0o100) + chr(0b10101 + 0o137) + chr(0b1011100 + 0o12) + chr(45) + chr(56)))(EEf4r9nUvta_)
if not xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5\x0e\x05t\x86t\xd9\xf0\xca'), chr(0b100011 + 0o101) + chr(101) + chr(0b111110 + 0o45) + '\157' + chr(0b1100100) + chr(5196 - 5095))(chr(0b10000 + 0o145) + chr(116) + '\146' + chr(45) + chr(0b1 + 0o67))) and PQZjDxhiHJGf > ehT0Px3KOsy9(chr(48) + chr(3826 - 3715) + chr(1085 - 1036), 8):
DyguQXdCse_Q(xafqLlk3kkUe(SXOLrMavuUCe(b'\xef\n\x17F\xcfi\xcc\xe4\xcc\x9f\x85\xefP\xeb\x81*\x02\xc5\xb6;\xff\xd7\xe1\xd3\x90\x7f@\xb9]\xe2:\xec;\x06\xf3\x1d-z\xa4\xa8\xdeA'), '\x64' + chr(0b100101 + 0o100) + chr(99) + chr(0b111101 + 0o62) + chr(0b1100100) + chr(7042 - 6941))(chr(0b10 + 0o163) + chr(0b11110 + 0o126) + chr(3301 - 3199) + chr(0b101101) + chr(0b11010 + 0o36)))
if not xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5\x0e\x05t\x86t\xd9\xf0\xca'), chr(100) + chr(9412 - 9311) + chr(8430 - 8331) + chr(0b1101111) + chr(0b1000 + 0o134) + '\x65')(chr(0b110011 + 0o102) + '\164' + '\x66' + '\x2d' + chr(0b10001 + 0o47))) and xafqLlk3kkUe(oVre8I6UXc3b.hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf8\x0bGb\x81K\x9e\xed\xdf\x8c\x82\xf6'), '\144' + '\145' + chr(4408 - 4309) + chr(111) + chr(8937 - 8837) + '\145')('\x75' + chr(0b1011100 + 0o30) + chr(102) + '\055' + '\070')) != xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x0e\x18O\x80w'), '\144' + chr(2012 - 1911) + chr(0b100111 + 0o74) + chr(0b1101111) + chr(100) + chr(0b10000 + 0o125))('\x75' + chr(2093 - 1977) + chr(382 - 280) + '\x2d' + chr(56)):
DyguQXdCse_Q(xafqLlk3kkUe(SXOLrMavuUCe(b'\xe3\x00\x18\x06\x9d{\xc7\xe1\xd1\x91\xcd\xf5_\xe1\xd1 \x04\xd9\xf1z\xb9\xd5\xfc\x97\x943\r\xa1P\xf3>\xec"\x00\xa7\x1ccd\xbe\xfc\xc4\x01\x06^\x9bi\x87'), '\x64' + '\x65' + '\x63' + '\157' + chr(0b1100100) + chr(101))(chr(12501 - 12384) + chr(0b1011 + 0o151) + chr(5255 - 5153) + '\x2d' + '\x38'))
xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf2\t\x1fG\x83E\xd9\xf7\xd1\x9e\x81\xe3S\xd3\xc9<\x0c\xc5\xf77\xac\xe5\xe8\xd2\x94g\x15\xbcQ\xe5'), '\x64' + chr(0b1100101) + chr(99) + chr(0b101101 + 0o102) + '\144' + '\x65')(chr(0b11100 + 0o131) + chr(0b1000 + 0o154) + chr(3175 - 3073) + chr(0b0 + 0o55) + chr(0b111000)))(EEf4r9nUvta_)
if xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf2\x1f\x04D\x8dv\xcc\xe8\xe1\x94\x9d\xe7L\xed\xcc?'), chr(100) + chr(0b1011011 + 0o12) + chr(0b1100011) + chr(9536 - 9425) + chr(0b101010 + 0o72) + '\145')('\x75' + chr(0b101010 + 0o112) + chr(102) + chr(0b101101 + 0o0) + '\x38')):
FfV_wqUqiesq = oVre8I6UXc3b._problem_hparams.bYPswhysd3s2[xafqLlk3kkUe(SXOLrMavuUCe(b'\xd9\x0e\x04L\x8an\xda'), chr(4215 - 4115) + chr(0b1011011 + 0o12) + chr(99) + '\157' + '\x64' + chr(0b1100101))('\x75' + '\164' + chr(3235 - 3133) + chr(0b101101) + '\x38')]
if FfV_wqUqiesq == xafqLlk3kkUe(PuPeNl0CuqOQ.ModalityType, xafqLlk3kkUe(SXOLrMavuUCe(b'\xee#7x\xbcE\xe5\xc4\xfc\xb9\xa1'), '\144' + chr(101) + chr(0b1100011) + chr(2597 - 2486) + chr(100) + chr(0b1100101))('\165' + chr(0b100100 + 0o120) + '\x66' + chr(0b101101) + '\070')):
PQZjDxhiHJGf = ehT0Px3KOsy9(chr(116 - 68) + chr(0b1101111) + chr(49), 8)
if PQZjDxhiHJGf == ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49), 8):
iBNBQu99lQEa(xafqLlk3kkUe(SXOLrMavuUCe(b'\xea\x1d\x13N\x8bc\x89\xc1\xdb\x9f\x82\xe2W\xe2\xc6'), chr(100) + chr(101) + chr(0b10001 + 0o122) + chr(0b1101111) + chr(0b1100100) + '\x65')(chr(11278 - 11161) + chr(0b1101011 + 0o11) + chr(7574 - 7472) + chr(0b101101) + '\x38'))
iIGKX2zSEGYP = oVre8I6UXc3b._greedy_infer(EEf4r9nUvta_, U6Ej34SVvx1Y, L4eE7kczIJwa)
else:
iBNBQu99lQEa(xafqLlk3kkUe(SXOLrMavuUCe(b'\xef\n\x17F\xcf^\xcc\xe6\xd1\x98\x84\xe8Y\xac\xd6%\x19\xdf\xb68\xba\xdb\xe3\x97\x86z\x1a\xab\x14\xb36'), chr(0b110011 + 0o61) + '\145' + '\x63' + chr(11168 - 11057) + chr(0b1100010 + 0o2) + '\145')('\165' + chr(0b1110100) + '\x66' + chr(0b101101) + chr(0b111000)) % PQZjDxhiHJGf)
iIGKX2zSEGYP = oVre8I6UXc3b._beam_decode(EEf4r9nUvta_, U6Ej34SVvx1Y, PQZjDxhiHJGf, oC1hU_0mlSje, gDUX9w35YHFE, L4eE7kczIJwa)
return iIGKX2zSEGYP
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel._beam_decode
|
def _beam_decode(self,
features,
decode_length,
beam_size,
top_beams,
alpha,
use_tpu=False):
"""Beam search decoding.
Models should ideally implement a more efficient version of this function.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: A bool, whether to do beam decode on TPU.
Returns:
samples: an integer `Tensor`. Top samples from the beam search
"""
return self._beam_decode_slow(features, decode_length, beam_size, top_beams,
alpha, use_tpu)
|
python
|
def _beam_decode(self,
features,
decode_length,
beam_size,
top_beams,
alpha,
use_tpu=False):
"""Beam search decoding.
Models should ideally implement a more efficient version of this function.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: A bool, whether to do beam decode on TPU.
Returns:
samples: an integer `Tensor`. Top samples from the beam search
"""
return self._beam_decode_slow(features, decode_length, beam_size, top_beams,
alpha, use_tpu)
|
[
"def",
"_beam_decode",
"(",
"self",
",",
"features",
",",
"decode_length",
",",
"beam_size",
",",
"top_beams",
",",
"alpha",
",",
"use_tpu",
"=",
"False",
")",
":",
"return",
"self",
".",
"_beam_decode_slow",
"(",
"features",
",",
"decode_length",
",",
"beam_size",
",",
"top_beams",
",",
"alpha",
",",
"use_tpu",
")"
] |
Beam search decoding.
Models should ideally implement a more efficient version of this function.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: A bool, whether to do beam decode on TPU.
Returns:
samples: an integer `Tensor`. Top samples from the beam search
|
[
"Beam",
"search",
"decoding",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L819-L843
|
train
|
Beam search decoding.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + '\x6f' + chr(328 - 278) + '\x33' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(1367 - 1256) + '\x33' + chr(53), 36808 - 36800), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b0 + 0o61) + chr(0b101010 + 0o6) + chr(52), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\061' + chr(0b10011 + 0o41) + chr(0b110111), 6859 - 6851), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(111) + chr(1633 - 1584) + chr(0b100 + 0o54) + chr(51), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b111100 + 0o63) + '\x33' + chr(50) + chr(677 - 622), ord("\x08")), ehT0Px3KOsy9('\060' + chr(5413 - 5302) + chr(1025 - 975) + '\060' + chr(52), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b10001 + 0o41) + chr(2230 - 2177) + chr(0b101101 + 0o10), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110001) + '\x34', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110011) + '\x34' + chr(0b110001), 43363 - 43355), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110111) + '\x37', 0o10), ehT0Px3KOsy9('\060' + chr(1235 - 1124) + chr(0b101010 + 0o7) + chr(48) + chr(50), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(663 - 613) + chr(0b10010 + 0o41) + chr(2062 - 2007), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + '\063' + '\061' + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(1938 - 1890) + '\157' + chr(50) + chr(53) + '\x34', 9727 - 9719), ehT0Px3KOsy9(chr(48) + '\157' + chr(572 - 517) + chr(50), 56065 - 56057), ehT0Px3KOsy9('\060' + chr(7189 - 7078) + chr(0b110001) + chr(55) + '\066', 25305 - 25297), ehT0Px3KOsy9('\060' + chr(9328 - 9217) + '\066', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + chr(0b110110) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b1111 + 0o45) + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + chr(54) + '\x32', 0o10), ehT0Px3KOsy9('\x30' + chr(1552 - 1441) + chr(0b110011) + chr(1693 - 1642) + chr(1188 - 1136), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + chr(53) + chr(49), 64678 - 64670), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + '\061' + chr(0b110100), 25257 - 25249), ehT0Px3KOsy9('\x30' + '\157' + chr(0b101 + 0o62) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(0b1101111) + chr(0b11011 + 0o30) + '\061' + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(57 - 9) + chr(1951 - 1840) + chr(0b110011) + '\x35' + chr(0b101010 + 0o15), 0b1000), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(3538 - 3427) + '\x32' + chr(0b1110 + 0o47) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(1866 - 1818) + chr(111) + '\061' + '\064' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(2232 - 2181) + chr(0b100111 + 0o12) + '\064', 8), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + chr(50) + chr(53), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(49) + '\065' + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(621 - 572) + '\062' + chr(1084 - 1033), 33254 - 33246), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110001) + '\060' + '\x37', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b0 + 0o61) + chr(0b101011 + 0o14) + chr(738 - 683), 0o10), ehT0Px3KOsy9(chr(1044 - 996) + '\157' + chr(0b11010 + 0o30) + '\067' + chr(0b11010 + 0o34), 0o10), ehT0Px3KOsy9(chr(85 - 37) + chr(10294 - 10183) + chr(0b10 + 0o57) + chr(0b10110 + 0o36) + '\x33', 21984 - 21976), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\063' + '\064' + '\060', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(130 - 79) + chr(48) + chr(52), 31734 - 31726), ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(0b1010111 + 0o30) + chr(450 - 399) + chr(0b100011 + 0o24) + chr(50), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(4321 - 4210) + chr(53) + chr(48), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xb4'), chr(0b1001100 + 0o30) + chr(101) + chr(776 - 677) + chr(0b1101111) + chr(0b1100100) + chr(101))(chr(4206 - 4089) + chr(0b1110100) + '\146' + chr(0b101101) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def hfmKGI32Gc0q(oVre8I6UXc3b, EEf4r9nUvta_, U6Ej34SVvx1Y, PQZjDxhiHJGf, oC1hU_0mlSje, gDUX9w35YHFE, L4eE7kczIJwa=ehT0Px3KOsy9(chr(0b101 + 0o53) + chr(0b1101111) + '\060', 832 - 824)):
return xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b"\xc5\x7f!\x06.\x88\xe0\xc8o\x8e['\xc0\xb9\xc0U\t"), chr(0b1100100) + chr(0b1100101) + '\143' + chr(1494 - 1383) + chr(0b100110 + 0o76) + chr(0b1100101))(chr(117) + chr(116) + chr(0b100000 + 0o106) + chr(1063 - 1018) + chr(56)))(EEf4r9nUvta_, U6Ej34SVvx1Y, PQZjDxhiHJGf, oC1hU_0mlSje, gDUX9w35YHFE, L4eE7kczIJwa)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel._beam_decode_slow
|
def _beam_decode_slow(self, features, decode_length, beam_size, top_beams,
alpha, use_tpu=False):
"""Slow version of Beam search decoding.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: A bool, whether to do slow beam decode on TPU.
Returns:
samples: an integer `Tensor`. Top samples from the beam search.
Raises:
NotImplementedError: If use_tpu is set to true.
"""
batch_size = common_layers.shape_list(features["inputs"])[0]
def symbols_to_logits_fn(ids, i=None):
"""Go from ids to logits."""
ids = tf.expand_dims(tf.expand_dims(ids, axis=2), axis=3)
ids = tf.pad(ids[:, 1:], [[0, 0], [0, 1], [0, 0], [0, 0]])
if "partial_targets" in features:
pt = features["partial_targets"]
pt_length = common_layers.shape_list(pt)[1]
pt = tf.tile(pt, [1, beam_size])
pt = tf.reshape(pt, [batch_size * beam_size, pt_length, 1, 1])
ids = tf.concat([pt, ids], axis=1)
features["targets"] = ids
if i is not None:
features["decode_loop_step"] = i
self._coverage = None
logits, _ = self(features) # pylint: disable=not-callable
# now self._coverage is a coverage tensor for the first datashard.
# it has shape [batch_size] and contains floats between 0 and
# source_length.
if self._problem_hparams:
modality = self._problem_hparams.modality["targets"]
top = self._hparams.top.get("targets", modalities.get_top(modality))
if getattr(top, "pointwise", False):
return tf.squeeze(logits, axis=[1, 2, 3])
# -1 due to the pad above.
current_output_position = common_layers.shape_list(ids)[1] - 1
logits = logits[:, current_output_position, :, :]
return tf.squeeze(logits, axis=[1, 2])
def _clone_examples_for_beam(old_feature, n):
"""Clone each example n times."""
old_shape = common_layers.shape_list(old_feature)
assert len(old_shape) >= 1
# Expand the inputs in to the beam size.
feature = tf.expand_dims(old_feature, 1)
feature = tf.tile(feature, [1, n] + [1] * (len(old_shape) - 1))
new_shape = common_layers.shape_list(feature)
feature = tf.reshape(feature,
[new_shape[0] * new_shape[1]] + new_shape[2:])
return feature
initial_ids = tf.zeros([batch_size], dtype=tf.int32)
# Clone select features multiple times to account for beam size.
old_features = {}
for feature_name in ["inputs", "knowledge"]:
if feature_name not in features:
continue
old_features[feature_name] = features[feature_name]
features[feature_name] = _clone_examples_for_beam(
features[feature_name], beam_size)
vocab_size = self._problem_hparams.vocab_size["targets"]
if vocab_size is not None and hasattr(self._hparams, "vocab_divisor"):
vocab_size += (-vocab_size) % self._hparams.vocab_divisor
# Setting decode length to input length + decode_length
if "partial_targets" not in features:
inputs = features["inputs"]
decode_length = (common_layers.shape_list(inputs)[1] +
features.get("decode_length", decode_length))
ids, scores, _ = beam_search.beam_search(
symbols_to_logits_fn,
initial_ids,
beam_size,
decode_length,
vocab_size,
alpha,
stop_early=(top_beams == 1),
use_tpu=use_tpu)
# Set features back to the unexpanded form to not to confuse the
# Estimator!
features.update(old_features)
# Return `top_beams` decodings (also remove initial id from the beam search)
# TODO(lukaszkaiser): make it work multi-problem.
if top_beams == 1:
samples = ids[:, 0, 1:]
else:
samples = ids[:, :top_beams, 1:]
return {"outputs": samples, "scores": scores}
|
python
|
def _beam_decode_slow(self, features, decode_length, beam_size, top_beams,
alpha, use_tpu=False):
"""Slow version of Beam search decoding.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: A bool, whether to do slow beam decode on TPU.
Returns:
samples: an integer `Tensor`. Top samples from the beam search.
Raises:
NotImplementedError: If use_tpu is set to true.
"""
batch_size = common_layers.shape_list(features["inputs"])[0]
def symbols_to_logits_fn(ids, i=None):
"""Go from ids to logits."""
ids = tf.expand_dims(tf.expand_dims(ids, axis=2), axis=3)
ids = tf.pad(ids[:, 1:], [[0, 0], [0, 1], [0, 0], [0, 0]])
if "partial_targets" in features:
pt = features["partial_targets"]
pt_length = common_layers.shape_list(pt)[1]
pt = tf.tile(pt, [1, beam_size])
pt = tf.reshape(pt, [batch_size * beam_size, pt_length, 1, 1])
ids = tf.concat([pt, ids], axis=1)
features["targets"] = ids
if i is not None:
features["decode_loop_step"] = i
self._coverage = None
logits, _ = self(features) # pylint: disable=not-callable
# now self._coverage is a coverage tensor for the first datashard.
# it has shape [batch_size] and contains floats between 0 and
# source_length.
if self._problem_hparams:
modality = self._problem_hparams.modality["targets"]
top = self._hparams.top.get("targets", modalities.get_top(modality))
if getattr(top, "pointwise", False):
return tf.squeeze(logits, axis=[1, 2, 3])
# -1 due to the pad above.
current_output_position = common_layers.shape_list(ids)[1] - 1
logits = logits[:, current_output_position, :, :]
return tf.squeeze(logits, axis=[1, 2])
def _clone_examples_for_beam(old_feature, n):
"""Clone each example n times."""
old_shape = common_layers.shape_list(old_feature)
assert len(old_shape) >= 1
# Expand the inputs in to the beam size.
feature = tf.expand_dims(old_feature, 1)
feature = tf.tile(feature, [1, n] + [1] * (len(old_shape) - 1))
new_shape = common_layers.shape_list(feature)
feature = tf.reshape(feature,
[new_shape[0] * new_shape[1]] + new_shape[2:])
return feature
initial_ids = tf.zeros([batch_size], dtype=tf.int32)
# Clone select features multiple times to account for beam size.
old_features = {}
for feature_name in ["inputs", "knowledge"]:
if feature_name not in features:
continue
old_features[feature_name] = features[feature_name]
features[feature_name] = _clone_examples_for_beam(
features[feature_name], beam_size)
vocab_size = self._problem_hparams.vocab_size["targets"]
if vocab_size is not None and hasattr(self._hparams, "vocab_divisor"):
vocab_size += (-vocab_size) % self._hparams.vocab_divisor
# Setting decode length to input length + decode_length
if "partial_targets" not in features:
inputs = features["inputs"]
decode_length = (common_layers.shape_list(inputs)[1] +
features.get("decode_length", decode_length))
ids, scores, _ = beam_search.beam_search(
symbols_to_logits_fn,
initial_ids,
beam_size,
decode_length,
vocab_size,
alpha,
stop_early=(top_beams == 1),
use_tpu=use_tpu)
# Set features back to the unexpanded form to not to confuse the
# Estimator!
features.update(old_features)
# Return `top_beams` decodings (also remove initial id from the beam search)
# TODO(lukaszkaiser): make it work multi-problem.
if top_beams == 1:
samples = ids[:, 0, 1:]
else:
samples = ids[:, :top_beams, 1:]
return {"outputs": samples, "scores": scores}
|
[
"def",
"_beam_decode_slow",
"(",
"self",
",",
"features",
",",
"decode_length",
",",
"beam_size",
",",
"top_beams",
",",
"alpha",
",",
"use_tpu",
"=",
"False",
")",
":",
"batch_size",
"=",
"common_layers",
".",
"shape_list",
"(",
"features",
"[",
"\"inputs\"",
"]",
")",
"[",
"0",
"]",
"def",
"symbols_to_logits_fn",
"(",
"ids",
",",
"i",
"=",
"None",
")",
":",
"\"\"\"Go from ids to logits.\"\"\"",
"ids",
"=",
"tf",
".",
"expand_dims",
"(",
"tf",
".",
"expand_dims",
"(",
"ids",
",",
"axis",
"=",
"2",
")",
",",
"axis",
"=",
"3",
")",
"ids",
"=",
"tf",
".",
"pad",
"(",
"ids",
"[",
":",
",",
"1",
":",
"]",
",",
"[",
"[",
"0",
",",
"0",
"]",
",",
"[",
"0",
",",
"1",
"]",
",",
"[",
"0",
",",
"0",
"]",
",",
"[",
"0",
",",
"0",
"]",
"]",
")",
"if",
"\"partial_targets\"",
"in",
"features",
":",
"pt",
"=",
"features",
"[",
"\"partial_targets\"",
"]",
"pt_length",
"=",
"common_layers",
".",
"shape_list",
"(",
"pt",
")",
"[",
"1",
"]",
"pt",
"=",
"tf",
".",
"tile",
"(",
"pt",
",",
"[",
"1",
",",
"beam_size",
"]",
")",
"pt",
"=",
"tf",
".",
"reshape",
"(",
"pt",
",",
"[",
"batch_size",
"*",
"beam_size",
",",
"pt_length",
",",
"1",
",",
"1",
"]",
")",
"ids",
"=",
"tf",
".",
"concat",
"(",
"[",
"pt",
",",
"ids",
"]",
",",
"axis",
"=",
"1",
")",
"features",
"[",
"\"targets\"",
"]",
"=",
"ids",
"if",
"i",
"is",
"not",
"None",
":",
"features",
"[",
"\"decode_loop_step\"",
"]",
"=",
"i",
"self",
".",
"_coverage",
"=",
"None",
"logits",
",",
"_",
"=",
"self",
"(",
"features",
")",
"# pylint: disable=not-callable",
"# now self._coverage is a coverage tensor for the first datashard.",
"# it has shape [batch_size] and contains floats between 0 and",
"# source_length.",
"if",
"self",
".",
"_problem_hparams",
":",
"modality",
"=",
"self",
".",
"_problem_hparams",
".",
"modality",
"[",
"\"targets\"",
"]",
"top",
"=",
"self",
".",
"_hparams",
".",
"top",
".",
"get",
"(",
"\"targets\"",
",",
"modalities",
".",
"get_top",
"(",
"modality",
")",
")",
"if",
"getattr",
"(",
"top",
",",
"\"pointwise\"",
",",
"False",
")",
":",
"return",
"tf",
".",
"squeeze",
"(",
"logits",
",",
"axis",
"=",
"[",
"1",
",",
"2",
",",
"3",
"]",
")",
"# -1 due to the pad above.",
"current_output_position",
"=",
"common_layers",
".",
"shape_list",
"(",
"ids",
")",
"[",
"1",
"]",
"-",
"1",
"logits",
"=",
"logits",
"[",
":",
",",
"current_output_position",
",",
":",
",",
":",
"]",
"return",
"tf",
".",
"squeeze",
"(",
"logits",
",",
"axis",
"=",
"[",
"1",
",",
"2",
"]",
")",
"def",
"_clone_examples_for_beam",
"(",
"old_feature",
",",
"n",
")",
":",
"\"\"\"Clone each example n times.\"\"\"",
"old_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"old_feature",
")",
"assert",
"len",
"(",
"old_shape",
")",
">=",
"1",
"# Expand the inputs in to the beam size.",
"feature",
"=",
"tf",
".",
"expand_dims",
"(",
"old_feature",
",",
"1",
")",
"feature",
"=",
"tf",
".",
"tile",
"(",
"feature",
",",
"[",
"1",
",",
"n",
"]",
"+",
"[",
"1",
"]",
"*",
"(",
"len",
"(",
"old_shape",
")",
"-",
"1",
")",
")",
"new_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"feature",
")",
"feature",
"=",
"tf",
".",
"reshape",
"(",
"feature",
",",
"[",
"new_shape",
"[",
"0",
"]",
"*",
"new_shape",
"[",
"1",
"]",
"]",
"+",
"new_shape",
"[",
"2",
":",
"]",
")",
"return",
"feature",
"initial_ids",
"=",
"tf",
".",
"zeros",
"(",
"[",
"batch_size",
"]",
",",
"dtype",
"=",
"tf",
".",
"int32",
")",
"# Clone select features multiple times to account for beam size.",
"old_features",
"=",
"{",
"}",
"for",
"feature_name",
"in",
"[",
"\"inputs\"",
",",
"\"knowledge\"",
"]",
":",
"if",
"feature_name",
"not",
"in",
"features",
":",
"continue",
"old_features",
"[",
"feature_name",
"]",
"=",
"features",
"[",
"feature_name",
"]",
"features",
"[",
"feature_name",
"]",
"=",
"_clone_examples_for_beam",
"(",
"features",
"[",
"feature_name",
"]",
",",
"beam_size",
")",
"vocab_size",
"=",
"self",
".",
"_problem_hparams",
".",
"vocab_size",
"[",
"\"targets\"",
"]",
"if",
"vocab_size",
"is",
"not",
"None",
"and",
"hasattr",
"(",
"self",
".",
"_hparams",
",",
"\"vocab_divisor\"",
")",
":",
"vocab_size",
"+=",
"(",
"-",
"vocab_size",
")",
"%",
"self",
".",
"_hparams",
".",
"vocab_divisor",
"# Setting decode length to input length + decode_length",
"if",
"\"partial_targets\"",
"not",
"in",
"features",
":",
"inputs",
"=",
"features",
"[",
"\"inputs\"",
"]",
"decode_length",
"=",
"(",
"common_layers",
".",
"shape_list",
"(",
"inputs",
")",
"[",
"1",
"]",
"+",
"features",
".",
"get",
"(",
"\"decode_length\"",
",",
"decode_length",
")",
")",
"ids",
",",
"scores",
",",
"_",
"=",
"beam_search",
".",
"beam_search",
"(",
"symbols_to_logits_fn",
",",
"initial_ids",
",",
"beam_size",
",",
"decode_length",
",",
"vocab_size",
",",
"alpha",
",",
"stop_early",
"=",
"(",
"top_beams",
"==",
"1",
")",
",",
"use_tpu",
"=",
"use_tpu",
")",
"# Set features back to the unexpanded form to not to confuse the",
"# Estimator!",
"features",
".",
"update",
"(",
"old_features",
")",
"# Return `top_beams` decodings (also remove initial id from the beam search)",
"# TODO(lukaszkaiser): make it work multi-problem.",
"if",
"top_beams",
"==",
"1",
":",
"samples",
"=",
"ids",
"[",
":",
",",
"0",
",",
"1",
":",
"]",
"else",
":",
"samples",
"=",
"ids",
"[",
":",
",",
":",
"top_beams",
",",
"1",
":",
"]",
"return",
"{",
"\"outputs\"",
":",
"samples",
",",
"\"scores\"",
":",
"scores",
"}"
] |
Slow version of Beam search decoding.
Quadratic time in decode_length.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
beam_size: number of beams.
top_beams: an integer. How many of the beams to return.
alpha: Float that controls the length penalty. larger the alpha, stronger
the preference for longer translations.
use_tpu: A bool, whether to do slow beam decode on TPU.
Returns:
samples: an integer `Tensor`. Top samples from the beam search.
Raises:
NotImplementedError: If use_tpu is set to true.
|
[
"Slow",
"version",
"of",
"Beam",
"search",
"decoding",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L845-L951
|
train
|
Slow version of Beam search decoding.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b101111 + 0o3) + chr(48), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + chr(0b100110 + 0o16) + chr(1120 - 1072), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\063' + chr(54) + chr(85 - 36), 11646 - 11638), ehT0Px3KOsy9('\x30' + chr(111) + chr(51) + '\064' + '\x35', 51251 - 51243), ehT0Px3KOsy9(chr(48) + chr(6206 - 6095) + '\062' + chr(0b101101 + 0o6) + chr(0b101111 + 0o1), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110110) + chr(1456 - 1401), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b1010 + 0o51) + chr(948 - 897) + chr(48), 0b1000), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b1101111) + '\x32' + chr(0b110110) + '\060', 30 - 22), ehT0Px3KOsy9('\060' + '\x6f' + chr(50) + chr(0b1011 + 0o51) + chr(91 - 36), 0o10), ehT0Px3KOsy9(chr(0b10000 + 0o40) + '\157' + chr(0b110011) + '\x37', 0b1000), ehT0Px3KOsy9(chr(0b11 + 0o55) + '\x6f' + chr(1601 - 1552) + chr(0b110001), 38648 - 38640), ehT0Px3KOsy9(chr(2006 - 1958) + '\x6f' + '\x32' + chr(0b110000) + '\x33', 57831 - 57823), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110001) + chr(2019 - 1965) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(2473 - 2423) + '\060' + '\065', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x31' + '\063', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(2105 - 2054) + chr(54), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(0b1111 + 0o44) + chr(54) + chr(0b110110), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1576 - 1525) + chr(0b110001) + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + chr(2391 - 2280) + chr(0b111 + 0o52) + chr(0b110011) + '\x37', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001) + chr(49) + chr(48), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b0 + 0o157) + chr(1427 - 1376) + chr(50) + chr(0b1 + 0o60), 0o10), ehT0Px3KOsy9(chr(1112 - 1064) + '\157' + chr(274 - 226), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1000111 + 0o50) + '\x33' + chr(0b11110 + 0o27) + chr(50), 53104 - 53096), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110101) + chr(0b100101 + 0o22), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(1806 - 1755) + chr(0b110101) + chr(1185 - 1133), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b111010 + 0o65) + '\x32' + '\060', 8), ehT0Px3KOsy9('\060' + chr(0b111001 + 0o66) + '\x34' + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b11100 + 0o24) + chr(9532 - 9421) + '\x31' + chr(52) + chr(54), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\063' + chr(1817 - 1765) + '\063', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b101 + 0o152) + chr(0b110010) + chr(0b110100) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(49) + '\x37' + chr(0b11001 + 0o33), 25887 - 25879), ehT0Px3KOsy9(chr(48) + chr(111) + '\x32' + chr(0b110001 + 0o0) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(0b1101111) + chr(0b110011) + '\x37', 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010) + chr(418 - 363) + '\061', 17547 - 17539), ehT0Px3KOsy9('\060' + chr(3882 - 3771) + '\x31' + chr(48) + '\063', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(830 - 781) + chr(462 - 407) + chr(51), 6809 - 6801), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(48) + chr(50), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(0b1111 + 0o44) + chr(0b110100) + chr(0b10010 + 0o43), 8), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(3342 - 3231) + '\062' + chr(0b110010) + '\061', 22305 - 22297), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110001) + chr(129 - 80) + chr(1492 - 1441), 23775 - 23767)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110101) + chr(0b110000), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'>'), chr(0b101011 + 0o71) + chr(0b1001101 + 0o30) + chr(99) + chr(111) + chr(5105 - 5005) + chr(0b10001 + 0o124))('\x75' + chr(0b1101000 + 0o14) + '\x66' + chr(0b10 + 0o53) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def yKaf2PtmF5kg(oVre8I6UXc3b, EEf4r9nUvta_, U6Ej34SVvx1Y, PQZjDxhiHJGf, oC1hU_0mlSje, gDUX9w35YHFE, L4eE7kczIJwa=ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x30', 8)):
ix9dZyeAmUxY = jSKPaHwSAfVv.shape_list(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'y\x8e\x81\x90\xecD'), chr(0b1000011 + 0o41) + '\x65' + chr(0b1010111 + 0o14) + chr(0b1101111) + chr(789 - 689) + chr(0b101110 + 0o67))(chr(0b1110101) + '\164' + chr(0b1100110) + '\055' + chr(0b111000))])[ehT0Px3KOsy9('\060' + '\157' + chr(0b110000), 8)]
def P8dwKNPITUWX(zdjj2pRemk_P, WVxHKyX45z_L=None):
zdjj2pRemk_P = IDJ2eXGCBCDu.expand_dims(IDJ2eXGCBCDu.expand_dims(zdjj2pRemk_P, axis=ehT0Px3KOsy9(chr(0b110000) + chr(0b10100 + 0o133) + chr(0b11101 + 0o25), ord("\x08"))), axis=ehT0Px3KOsy9(chr(609 - 561) + '\157' + chr(0b11010 + 0o31), 0b1000))
zdjj2pRemk_P = IDJ2eXGCBCDu.pad(zdjj2pRemk_P[:, ehT0Px3KOsy9('\060' + chr(9272 - 9161) + chr(1903 - 1854), 0o10):], [[ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x30', 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x30', 8)], [ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b11011 + 0o25), 8), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(49), 8)], [ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110000), 8), ehT0Px3KOsy9(chr(647 - 599) + '\157' + '\060', 8)], [ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(111) + chr(0b110000), 8), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1101111) + chr(520 - 472), 8)]])
if xafqLlk3kkUe(SXOLrMavuUCe(b'`\x81\x83\x91\xf1V\xb4\xef\xd3~{\xe4\x11\xf1x'), chr(100) + '\x65' + '\143' + chr(0b1101111) + chr(2777 - 2677) + chr(5936 - 5835))(chr(0b1110101) + chr(116) + chr(102) + '\x2d' + chr(1455 - 1399)) in EEf4r9nUvta_:
SZHrkL2hsxgj = EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'`\x81\x83\x91\xf1V\xb4\xef\xd3~{\xe4\x11\xf1x'), '\x64' + chr(8768 - 8667) + '\x63' + chr(4608 - 4497) + '\x64' + chr(6624 - 6523))('\165' + '\x74' + '\x66' + chr(627 - 582) + '\070')]
nF0c0xkCAtzZ = jSKPaHwSAfVv.shape_list(SZHrkL2hsxgj)[ehT0Px3KOsy9('\060' + '\x6f' + '\061', 8)]
SZHrkL2hsxgj = IDJ2eXGCBCDu.tile(SZHrkL2hsxgj, [ehT0Px3KOsy9(chr(442 - 394) + '\157' + chr(0b110001), 8), PQZjDxhiHJGf])
SZHrkL2hsxgj = IDJ2eXGCBCDu.reshape(SZHrkL2hsxgj, [ix9dZyeAmUxY * PQZjDxhiHJGf, nF0c0xkCAtzZ, ehT0Px3KOsy9('\x30' + chr(0b111000 + 0o67) + chr(49), 8), ehT0Px3KOsy9('\x30' + chr(8263 - 8152) + chr(0b110001), 8)])
zdjj2pRemk_P = IDJ2eXGCBCDu.concat([SZHrkL2hsxgj, zdjj2pRemk_P], axis=ehT0Px3KOsy9('\060' + chr(0b11 + 0o154) + chr(49), 8))
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'd\x81\x83\x82\xfdC\xab'), chr(0b1100100) + chr(0b1001001 + 0o34) + '\143' + '\157' + chr(0b1100100) + chr(5982 - 5881))('\x75' + chr(0b1001001 + 0o53) + chr(0b1001010 + 0o34) + chr(1746 - 1701) + chr(0b100001 + 0o27))] = zdjj2pRemk_P
if WVxHKyX45z_L is not None:
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b't\x85\x92\x8a\xfcR\x87\xdc\xc8py\xdc\x07\xf1nn'), chr(0b1100100) + chr(101) + '\x63' + chr(9399 - 9288) + chr(100) + chr(0b1100101))(chr(2671 - 2554) + '\164' + chr(102) + chr(0b101101) + chr(0b111000))] = WVxHKyX45z_L
oVre8I6UXc3b.lySXM3CyEydI = None
(wF9nmvjsKjYM, VNGQdHSFPrso) = oVre8I6UXc3b(EEf4r9nUvta_)
if xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'O\x90\x83\x8a\xfa[\xbd\xdd\xf8wy\xe2\x06\xe4fm'), '\x64' + '\145' + chr(2729 - 2630) + '\x6f' + chr(100) + chr(7761 - 7660))(chr(9179 - 9062) + '\x74' + chr(102) + chr(45) + '\070')):
bYPswhysd3s2 = oVre8I6UXc3b._problem_hparams.bYPswhysd3s2[xafqLlk3kkUe(SXOLrMavuUCe(b'd\x81\x83\x82\xfdC\xab'), '\x64' + '\145' + '\x63' + chr(0b1101111) + chr(100) + chr(0b11100 + 0o111))('\165' + chr(0b1110100) + chr(8010 - 7908) + chr(0b100111 + 0o6) + '\070')]
qxrVBjeryNEZ = oVre8I6UXc3b._hparams.top.get(xafqLlk3kkUe(SXOLrMavuUCe(b'd\x81\x83\x82\xfdC\xab'), chr(6484 - 6384) + '\145' + '\x63' + '\x6f' + '\144' + '\145')(chr(0b1110101) + '\x74' + '\x66' + '\x2d' + chr(0b11 + 0o65)), PuPeNl0CuqOQ.get_top(bYPswhysd3s2))
if xafqLlk3kkUe(qxrVBjeryNEZ, xafqLlk3kkUe(SXOLrMavuUCe(b'`\x8f\x98\x8b\xec@\xb1\xc3\xc2'), chr(0b1011 + 0o131) + chr(0b1100101) + chr(0b1100011) + chr(11662 - 11551) + chr(4272 - 4172) + '\145')('\x75' + '\x74' + '\x66' + '\x2d' + chr(1306 - 1250)), ehT0Px3KOsy9(chr(1670 - 1622) + chr(111) + chr(252 - 204), 8)):
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'c\x91\x84\x80\xfdM\xbd'), chr(0b1100011 + 0o1) + chr(101) + chr(0b1100001 + 0o2) + '\x6f' + '\x64' + chr(0b1100101))('\165' + chr(182 - 66) + chr(0b101110 + 0o70) + chr(0b110 + 0o47) + '\x38'))(wF9nmvjsKjYM, axis=[ehT0Px3KOsy9(chr(1417 - 1369) + '\x6f' + chr(0b100100 + 0o15), 8), ehT0Px3KOsy9(chr(1506 - 1458) + '\157' + chr(0b100101 + 0o15), 8), ehT0Px3KOsy9(chr(0b1110 + 0o42) + '\x6f' + '\x33', 8)])
Ne_DQ1Ggwi__ = jSKPaHwSAfVv.shape_list(zdjj2pRemk_P)[ehT0Px3KOsy9(chr(1477 - 1429) + chr(1735 - 1624) + '\061', 8)] - ehT0Px3KOsy9(chr(0b10011 + 0o35) + '\157' + chr(0b10001 + 0o40), 8)
wF9nmvjsKjYM = wF9nmvjsKjYM[:, Ne_DQ1Ggwi__, :, :]
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'c\x91\x84\x80\xfdM\xbd'), '\x64' + chr(5015 - 4914) + chr(0b1100011) + '\157' + chr(0b100011 + 0o101) + chr(0b1100101))('\x75' + chr(0b1000101 + 0o57) + '\x66' + chr(45) + chr(0b1000 + 0o60)))(wF9nmvjsKjYM, axis=[ehT0Px3KOsy9('\x30' + chr(111) + '\x31', 8), ehT0Px3KOsy9(chr(48) + chr(7670 - 7559) + chr(0b100101 + 0o15), 8)])
def Ew5wbUDrG4Dr(nGSRjlNO7cWT, m1NkCryOw9Bx):
Y8ZxZNJ8zK0n = jSKPaHwSAfVv.shape_list(nGSRjlNO7cWT)
assert c2A0yzQpDQB3(Y8ZxZNJ8zK0n) >= ehT0Px3KOsy9(chr(1622 - 1574) + chr(4463 - 4352) + '\061', 8)
fVxZREPfp9Oo = IDJ2eXGCBCDu.expand_dims(nGSRjlNO7cWT, ehT0Px3KOsy9(chr(48) + chr(0b1101101 + 0o2) + chr(2045 - 1996), 8))
fVxZREPfp9Oo = IDJ2eXGCBCDu.tile(fVxZREPfp9Oo, [ehT0Px3KOsy9('\x30' + chr(0b110011 + 0o74) + '\x31', 8), m1NkCryOw9Bx] + [ehT0Px3KOsy9(chr(48) + '\x6f' + '\061', 8)] * (c2A0yzQpDQB3(Y8ZxZNJ8zK0n) - ehT0Px3KOsy9(chr(48) + chr(111) + '\x31', 8)))
P7dVzv6_yXeE = jSKPaHwSAfVv.shape_list(fVxZREPfp9Oo)
fVxZREPfp9Oo = IDJ2eXGCBCDu.reshape(fVxZREPfp9Oo, [P7dVzv6_yXeE[ehT0Px3KOsy9(chr(0b110000) + chr(10482 - 10371) + chr(0b11110 + 0o22), 8)] * P7dVzv6_yXeE[ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b100 + 0o55), 8)]] + P7dVzv6_yXeE[ehT0Px3KOsy9(chr(1018 - 970) + chr(0b1101111) + '\x32', 8):])
return fVxZREPfp9Oo
WPnuojygthCW = IDJ2eXGCBCDu.zeros([ix9dZyeAmUxY], dtype=IDJ2eXGCBCDu.int32)
q3FGcy7nraMW = {}
for lPuZQT6rFAxL in [xafqLlk3kkUe(SXOLrMavuUCe(b'y\x8e\x81\x90\xecD'), '\x64' + '\x65' + '\x63' + chr(0b10 + 0o155) + chr(0b1100100) + chr(0b110110 + 0o57))(chr(117) + chr(116) + '\146' + '\055' + chr(343 - 287)), xafqLlk3kkUe(SXOLrMavuUCe(b'{\x8e\x9e\x92\xf4R\xbc\xd7\xc2'), chr(0b1100100) + chr(6326 - 6225) + chr(99) + '\157' + '\x64' + chr(9190 - 9089))('\165' + chr(0b1110100) + chr(5725 - 5623) + chr(0b10101 + 0o30) + '\x38')]:
if lPuZQT6rFAxL not in EEf4r9nUvta_:
continue
q3FGcy7nraMW[lPuZQT6rFAxL] = EEf4r9nUvta_[lPuZQT6rFAxL]
EEf4r9nUvta_[lPuZQT6rFAxL] = Ew5wbUDrG4Dr(EEf4r9nUvta_[lPuZQT6rFAxL], PQZjDxhiHJGf)
CeyMIoSyrpkQ = oVre8I6UXc3b._problem_hparams.CeyMIoSyrpkQ[xafqLlk3kkUe(SXOLrMavuUCe(b'd\x81\x83\x82\xfdC\xab'), chr(525 - 425) + chr(4015 - 3914) + chr(5350 - 5251) + chr(111) + '\144' + '\x65')('\165' + chr(116) + '\146' + '\x2d' + chr(0b111000))]
if CeyMIoSyrpkQ is not None and lot1PSoAwYhj(xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'~\xb0\xbe\xd0\xdbO\xae\xef\xf5LX\xd2'), chr(0b1100100) + chr(6998 - 6897) + chr(0b1100010 + 0o1) + chr(3278 - 3167) + chr(0b1000011 + 0o41) + chr(7795 - 7694))(chr(11900 - 11783) + chr(0b100111 + 0o115) + chr(0b110010 + 0o64) + chr(0b101101) + '\x38')), xafqLlk3kkUe(SXOLrMavuUCe(b'f\x8f\x92\x84\xfah\xbc\xd9\xd1vz\xec\x06'), chr(0b100001 + 0o103) + '\145' + chr(0b1100011) + chr(0b1100001 + 0o16) + '\x64' + chr(0b1010010 + 0o23))(chr(3598 - 3481) + chr(116) + chr(9718 - 9616) + '\055' + chr(0b111000))):
CeyMIoSyrpkQ += -CeyMIoSyrpkQ % oVre8I6UXc3b._hparams.vocab_divisor
if xafqLlk3kkUe(SXOLrMavuUCe(b'`\x81\x83\x91\xf1V\xb4\xef\xd3~{\xe4\x11\xf1x'), chr(0b1100100) + '\145' + '\143' + chr(111) + '\144' + chr(101))('\x75' + '\164' + chr(2340 - 2238) + chr(0b10100 + 0o31) + chr(1009 - 953)) not in EEf4r9nUvta_:
vXoupepMtCXU = EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'y\x8e\x81\x90\xecD'), '\x64' + '\145' + chr(99) + chr(0b10 + 0o155) + chr(8063 - 7963) + chr(101))(chr(117) + chr(116) + chr(0b1010110 + 0o20) + '\x2d' + '\070')]
U6Ej34SVvx1Y = jSKPaHwSAfVv.shape_list(vXoupepMtCXU)[ehT0Px3KOsy9(chr(48) + '\x6f' + '\x31', 8)] + EEf4r9nUvta_.get(xafqLlk3kkUe(SXOLrMavuUCe(b't\x85\x92\x8a\xfcR\x87\xdc\xc2qn\xf7\x1c'), chr(100) + chr(0b1100101) + '\x63' + chr(1315 - 1204) + chr(0b1100100) + chr(0b1100101))(chr(0b101011 + 0o112) + '\x74' + chr(102) + chr(0b100110 + 0o7) + chr(1464 - 1408)), U6Ej34SVvx1Y)
(zdjj2pRemk_P, b8rpGniBNUPr, VNGQdHSFPrso) = M4QqcqvVKFSA.beam_search(P8dwKNPITUWX, WPnuojygthCW, PQZjDxhiHJGf, U6Ej34SVvx1Y, CeyMIoSyrpkQ, gDUX9w35YHFE, stop_early=oC1hU_0mlSje == ehT0Px3KOsy9(chr(0b11100 + 0o24) + chr(0b1101111) + '\x31', 8), use_tpu=L4eE7kczIJwa)
xafqLlk3kkUe(EEf4r9nUvta_, xafqLlk3kkUe(SXOLrMavuUCe(b'J\x94\xb0\xa0\xf1y\x92\xde\xde+l\xb3'), chr(2836 - 2736) + chr(0b0 + 0o145) + '\143' + chr(6516 - 6405) + chr(5822 - 5722) + chr(101))(chr(0b1110101) + '\x74' + '\146' + chr(45) + chr(0b111000)))(q3FGcy7nraMW)
if oC1hU_0mlSje == ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(4458 - 4347) + chr(49), 8):
db1_IZvznkcy = zdjj2pRemk_P[:, ehT0Px3KOsy9('\x30' + chr(10898 - 10787) + chr(0b101001 + 0o7), 8), ehT0Px3KOsy9(chr(1874 - 1826) + chr(111) + '\x31', 8):]
else:
db1_IZvznkcy = zdjj2pRemk_P[:, :oC1hU_0mlSje, ehT0Px3KOsy9(chr(0b110000) + chr(0b10110 + 0o131) + chr(1503 - 1454), 8):]
return {xafqLlk3kkUe(SXOLrMavuUCe(b'\x7f\x95\x85\x95\xedC\xab'), '\x64' + '\145' + chr(0b100110 + 0o75) + chr(0b1011110 + 0o21) + '\144' + chr(1911 - 1810))('\x75' + chr(0b1110100) + '\x66' + '\x2d' + chr(0b111000)): db1_IZvznkcy, xafqLlk3kkUe(SXOLrMavuUCe(b'c\x83\x9e\x97\xfdD'), '\144' + chr(0b1100101) + '\143' + '\x6f' + '\x64' + chr(0b1100101))(chr(0b1110101) + '\164' + chr(102) + chr(0b1011 + 0o42) + chr(0b10 + 0o66)): b8rpGniBNUPr}
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel._greedy_infer
|
def _greedy_infer(self, features, decode_length, use_tpu=False):
"""A greedy inference method.
Models should ideally implement a more efficient version of this function.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
use_tpu: A bool, whether to build the inference graph for TPU.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": None
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`}
}
"""
if use_tpu:
return self._slow_greedy_infer_tpu(features, decode_length)
return self._slow_greedy_infer(features, decode_length)
|
python
|
def _greedy_infer(self, features, decode_length, use_tpu=False):
"""A greedy inference method.
Models should ideally implement a more efficient version of this function.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
use_tpu: A bool, whether to build the inference graph for TPU.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": None
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`}
}
"""
if use_tpu:
return self._slow_greedy_infer_tpu(features, decode_length)
return self._slow_greedy_infer(features, decode_length)
|
[
"def",
"_greedy_infer",
"(",
"self",
",",
"features",
",",
"decode_length",
",",
"use_tpu",
"=",
"False",
")",
":",
"if",
"use_tpu",
":",
"return",
"self",
".",
"_slow_greedy_infer_tpu",
"(",
"features",
",",
"decode_length",
")",
"return",
"self",
".",
"_slow_greedy_infer",
"(",
"features",
",",
"decode_length",
")"
] |
A greedy inference method.
Models should ideally implement a more efficient version of this function.
Args:
features: an map of string to `Tensor`
decode_length: an integer. How many additional timesteps to decode.
use_tpu: A bool, whether to build the inference graph for TPU.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": None
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`}
}
|
[
"A",
"greedy",
"inference",
"method",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L953-L975
|
train
|
A greedy inference method.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\063' + '\065' + chr(622 - 571), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110001) + '\066' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110000 + 0o1) + chr(0b110001), 12670 - 12662), ehT0Px3KOsy9('\x30' + chr(4343 - 4232) + chr(49) + '\x36' + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b10000 + 0o40) + '\x6f' + chr(1925 - 1874) + chr(53) + chr(2582 - 2528), 0b1000), ehT0Px3KOsy9(chr(582 - 534) + '\157' + '\061' + chr(0b100101 + 0o21) + chr(51), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110011) + '\063' + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(0b1010110 + 0o31) + chr(51) + chr(54) + chr(1133 - 1081), 0b1000), ehT0Px3KOsy9(chr(391 - 343) + chr(111) + chr(0b11110 + 0o25) + chr(499 - 450) + chr(0b100101 + 0o15), 0o10), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(10682 - 10571) + '\063' + chr(0b1100 + 0o50) + '\x37', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + chr(0b110011) + chr(0b110101), 0o10), ehT0Px3KOsy9('\x30' + chr(0b10011 + 0o134) + chr(0b100101 + 0o14) + '\x33' + chr(213 - 161), 0b1000), ehT0Px3KOsy9(chr(0b1001 + 0o47) + chr(0b101111 + 0o100) + chr(0b11110 + 0o25) + chr(48) + '\x31', 52700 - 52692), ehT0Px3KOsy9('\x30' + chr(0b1011100 + 0o23) + chr(1441 - 1392) + '\066', 40236 - 40228), ehT0Px3KOsy9(chr(0b110000) + chr(4595 - 4484) + chr(0b110 + 0o55) + chr(2054 - 2003) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b11101 + 0o23) + '\x6f' + chr(1367 - 1317) + chr(0b110111) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\061' + chr(0b110011) + chr(1760 - 1710), 59905 - 59897), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(0b1101111) + chr(49) + '\060', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1010100 + 0o33) + '\062' + chr(0b110111) + chr(0b101010 + 0o13), 8), ehT0Px3KOsy9('\060' + '\x6f' + chr(50) + '\063' + '\x37', 0o10), ehT0Px3KOsy9(chr(844 - 796) + chr(5660 - 5549) + '\062' + chr(48) + chr(0b10 + 0o62), ord("\x08")), ehT0Px3KOsy9(chr(1698 - 1650) + chr(0b1101111) + chr(936 - 885) + chr(0b110000) + '\064', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(1506 - 1455) + chr(0b1101 + 0o47), 49277 - 49269), ehT0Px3KOsy9('\x30' + chr(111) + chr(1438 - 1388) + '\064' + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(332 - 284) + chr(0b1101111) + chr(0b110011) + chr(0b110101) + chr(51), 8), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b100101 + 0o20) + chr(49), 62828 - 62820), ehT0Px3KOsy9(chr(0b100 + 0o54) + '\157' + '\x34' + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(1432 - 1384) + '\157' + '\060', ord("\x08")), ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(111) + chr(0b110101) + chr(0b101001 + 0o10), 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(50) + '\x30' + chr(0b100111 + 0o17), 48116 - 48108), ehT0Px3KOsy9(chr(342 - 294) + '\x6f' + chr(49) + chr(1015 - 964) + chr(2055 - 2001), 0b1000), ehT0Px3KOsy9(chr(1390 - 1342) + chr(0b1101111) + chr(0b0 + 0o63) + chr(0b110110) + chr(167 - 113), ord("\x08")), ehT0Px3KOsy9(chr(1690 - 1642) + chr(768 - 657) + chr(1922 - 1872) + chr(49) + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(55) + '\062', 4544 - 4536), ehT0Px3KOsy9(chr(0b110000) + chr(0b10001 + 0o136) + '\x32' + chr(0b100011 + 0o24) + chr(0b11110 + 0o30), 14931 - 14923), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b101001 + 0o11) + chr(49) + '\062', 8), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1449 - 1399) + '\064' + chr(875 - 823), 8), ehT0Px3KOsy9(chr(0b11101 + 0o23) + '\157' + chr(0b110010) + '\065' + chr(0b10 + 0o63), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(651 - 601) + chr(55) + '\x35', 8), ehT0Px3KOsy9(chr(0b10110 + 0o32) + '\x6f' + chr(50) + '\063' + chr(49), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(1829 - 1781) + '\x6f' + chr(0b110101) + chr(900 - 852), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5'), chr(100) + chr(101) + '\143' + chr(111) + chr(0b1100100) + chr(3900 - 3799))('\165' + chr(0b1001 + 0o153) + chr(102) + chr(0b1000 + 0o45) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def txHpm4iC4mly(oVre8I6UXc3b, EEf4r9nUvta_, U6Ej34SVvx1Y, L4eE7kczIJwa=ehT0Px3KOsy9(chr(0b110000) + chr(0b10010 + 0o135) + '\x30', 8)):
if L4eE7kczIJwa:
return xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb4\x02e\x1a\x99,\xbd\x86\xaf\x82\xb9%\x9eg\xb7\xdb?z\xe2\xeb\xf4\xb1'), '\144' + '\x65' + chr(9592 - 9493) + chr(111) + '\144' + chr(6179 - 6078))('\165' + chr(0b11011 + 0o131) + chr(102) + '\055' + chr(0b111000)))(EEf4r9nUvta_, U6Ej34SVvx1Y)
return xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb4\x02e\x1a\x99,\xbd\x86\xaf\x82\xb9%\x9eg\xb7\xdb?z'), chr(9707 - 9607) + chr(0b110110 + 0o57) + chr(3031 - 2932) + chr(0b100010 + 0o115) + chr(0b1100100) + '\145')('\165' + chr(6189 - 6073) + chr(102) + '\055' + chr(3008 - 2952)))(EEf4r9nUvta_, U6Ej34SVvx1Y)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel._slow_greedy_infer_tpu
|
def _slow_greedy_infer_tpu(self, features, decode_length):
"""A slow greedy inference method on TPU.
Quadratic time in decode_length.
Args:
features: An map of string to `Tensor`.
decode_length: An integer, how many additional timesteps to decode.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": None
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`}
}
"""
if not features:
features = {}
inputs_old = None
if "inputs" in features and len(features["inputs"].shape) < 4:
inputs_old = features["inputs"]
features["inputs"] = tf.expand_dims(features["inputs"], 2)
if not self.has_input:
# Prepare partial targets.
# In either features["inputs"] or features["targets"].
# We force the outputs to begin with these sequences.
partial_targets = features.get("inputs")
if partial_targets is None:
partial_targets = features["targets"]
features["partial_targets"] = tf.to_int64(partial_targets)
# Save the targets in a var and reassign it after the tf.while loop to avoid
# having targets being in a 'while' frame. This ensures targets when used
# in metric functions stays in the same frame as other vars.
targets_old = features.get("targets", None)
target_modality = self._problem_hparams.modality["targets"]
def infer_step(i, recent_output, recent_logits, unused_loss):
"""Inference step."""
if not tf.executing_eagerly():
recent_output.set_shape([None, None, None, 1])
padded = tf.pad(recent_output, [[0, 0], [0, 1], [0, 0], [0, 0]])
features["targets"] = padded
# This is inefficient in that it generates samples at all timesteps,
# not just the last one, except if target_modality is pointwise.
features["decode_loop_step"] = i
samples, logits, losses = self.sample(features)
# Concatenate the already-generated recent_output with last timestep
# of the newly-generated samples.z
top = self._hparams.top.get("targets",
modalities.get_top(target_modality))
if getattr(top, "pointwise", False):
cur_sample = samples[:, -1, :, :]
else:
cur_sample = samples[:, i, :, :]
samples = tf.transpose(recent_output, perm=[1, 0, 2, 3])
samples = inplace_ops.alias_inplace_update(samples, i,
tf.to_int64(cur_sample))
samples = tf.transpose(samples, perm=[1, 0, 2, 3])
if not tf.executing_eagerly():
samples.set_shape([None, None, None, 1])
# Assuming we have one shard for logits.
recent_logits = tf.transpose(recent_logits, perm=[1, 0, 2, 3, 4])
recent_logits = inplace_ops.alias_inplace_update(
recent_logits, i, tf.squeeze(logits[:, -1:], axis=1))
logits = tf.transpose(recent_logits, perm=[1, 0, 2, 3, 4])
loss = sum([l for l in losses.values() if l is not None])
return i + 1, samples, logits, loss
# Create an initial output tensor. This will be passed
# to the infer_step, which adds one timestep at every iteration.
if "partial_targets" in features:
initial_output = tf.to_int64(features["partial_targets"])
while len(initial_output.get_shape().as_list()) < 4:
initial_output = tf.expand_dims(initial_output, 2)
batch_size = common_layers.shape_list(initial_output)[0]
else:
batch_size = common_layers.shape_list(features["inputs"])[0]
initial_output = tf.zeros((batch_size, 0, 1, 1), dtype=tf.int64)
# Hack: foldl complains when the output shape is less specified than the
# input shape, so we confuse it about the input shape.
initial_output = tf.slice(initial_output, [0, 0, 0, 0],
common_layers.shape_list(initial_output))
target_modality = self._problem_hparams.modality["targets"]
if target_modality == modalities.ModalityType.CLASS_LABEL:
decode_length = 1
else:
if "partial_targets" in features:
prefix_length = common_layers.shape_list(features["partial_targets"])[1]
else:
prefix_length = common_layers.shape_list(features["inputs"])[1]
decode_length = prefix_length + decode_length
# Initial values of result, logits and loss.
result = tf.concat(
[initial_output,
tf.zeros([batch_size, decode_length, 1, 1], tf.int64)],
axis=1)
# tensor padded to [batch_size, decode_length, 1, 1, vocab_size]
vocab_size = self._problem_hparams.vocab_size["targets"]
if vocab_size is not None and hasattr(self._hparams, "vocab_divisor"):
vocab_size += (-vocab_size) % self._hparams.vocab_divisor
logits = tf.zeros((batch_size, decode_length, 1, 1, vocab_size))
if not tf.executing_eagerly():
logits.set_shape([None, None, None, None, None])
loss = 0.0
def while_exit_cond(i, result, logits, loss): # pylint: disable=unused-argument
"""Exit the loop either if reach decode_length or EOS."""
not_overflow = i < decode_length
if self._problem_hparams.stop_at_eos:
def fn_not_eos():
# Check if the last predicted element is a EOS
return tf.reduce_any(
tf.not_equal(
tf.squeeze(result[:, -1, :, :]), text_encoder.EOS_ID))
not_eos = tf.cond(
# We only check for early stopping if there is at least 1 element (
# otherwise not_eos will crash).
tf.not_equal(i, 0),
fn_not_eos,
lambda: True,
)
return tf.cond(
tf.equal(batch_size, 1),
# If batch_size == 1, we check EOS for early stopping.
lambda: tf.logical_and(not_overflow, not_eos),
# Else, just wait for max length
lambda: not_overflow)
return not_overflow
_, result, logits, loss = tf.while_loop(
while_exit_cond,
infer_step, [tf.constant(0), result, logits, loss],
shape_invariants=[
tf.TensorShape([]),
tf.TensorShape([batch_size, decode_length, 1, 1]),
tf.TensorShape([batch_size, decode_length, 1, 1, vocab_size]),
tf.TensorShape([]),
],
back_prop=False,
parallel_iterations=1)
if inputs_old is not None: # Restore to not confuse Estimator.
features["inputs"] = inputs_old
# Reassign targets back to the previous value.
if targets_old is not None:
features["targets"] = targets_old
losses = {"training": loss}
if "partial_targets" in features:
partial_target_length = common_layers.shape_list(
features["partial_targets"])[1]
result = tf.slice(result, [0, partial_target_length, 0, 0],
[-1, -1, -1, -1])
return {
"outputs": result,
"scores": None,
"logits": logits,
"losses": losses,
}
|
python
|
def _slow_greedy_infer_tpu(self, features, decode_length):
"""A slow greedy inference method on TPU.
Quadratic time in decode_length.
Args:
features: An map of string to `Tensor`.
decode_length: An integer, how many additional timesteps to decode.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": None
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`}
}
"""
if not features:
features = {}
inputs_old = None
if "inputs" in features and len(features["inputs"].shape) < 4:
inputs_old = features["inputs"]
features["inputs"] = tf.expand_dims(features["inputs"], 2)
if not self.has_input:
# Prepare partial targets.
# In either features["inputs"] or features["targets"].
# We force the outputs to begin with these sequences.
partial_targets = features.get("inputs")
if partial_targets is None:
partial_targets = features["targets"]
features["partial_targets"] = tf.to_int64(partial_targets)
# Save the targets in a var and reassign it after the tf.while loop to avoid
# having targets being in a 'while' frame. This ensures targets when used
# in metric functions stays in the same frame as other vars.
targets_old = features.get("targets", None)
target_modality = self._problem_hparams.modality["targets"]
def infer_step(i, recent_output, recent_logits, unused_loss):
"""Inference step."""
if not tf.executing_eagerly():
recent_output.set_shape([None, None, None, 1])
padded = tf.pad(recent_output, [[0, 0], [0, 1], [0, 0], [0, 0]])
features["targets"] = padded
# This is inefficient in that it generates samples at all timesteps,
# not just the last one, except if target_modality is pointwise.
features["decode_loop_step"] = i
samples, logits, losses = self.sample(features)
# Concatenate the already-generated recent_output with last timestep
# of the newly-generated samples.z
top = self._hparams.top.get("targets",
modalities.get_top(target_modality))
if getattr(top, "pointwise", False):
cur_sample = samples[:, -1, :, :]
else:
cur_sample = samples[:, i, :, :]
samples = tf.transpose(recent_output, perm=[1, 0, 2, 3])
samples = inplace_ops.alias_inplace_update(samples, i,
tf.to_int64(cur_sample))
samples = tf.transpose(samples, perm=[1, 0, 2, 3])
if not tf.executing_eagerly():
samples.set_shape([None, None, None, 1])
# Assuming we have one shard for logits.
recent_logits = tf.transpose(recent_logits, perm=[1, 0, 2, 3, 4])
recent_logits = inplace_ops.alias_inplace_update(
recent_logits, i, tf.squeeze(logits[:, -1:], axis=1))
logits = tf.transpose(recent_logits, perm=[1, 0, 2, 3, 4])
loss = sum([l for l in losses.values() if l is not None])
return i + 1, samples, logits, loss
# Create an initial output tensor. This will be passed
# to the infer_step, which adds one timestep at every iteration.
if "partial_targets" in features:
initial_output = tf.to_int64(features["partial_targets"])
while len(initial_output.get_shape().as_list()) < 4:
initial_output = tf.expand_dims(initial_output, 2)
batch_size = common_layers.shape_list(initial_output)[0]
else:
batch_size = common_layers.shape_list(features["inputs"])[0]
initial_output = tf.zeros((batch_size, 0, 1, 1), dtype=tf.int64)
# Hack: foldl complains when the output shape is less specified than the
# input shape, so we confuse it about the input shape.
initial_output = tf.slice(initial_output, [0, 0, 0, 0],
common_layers.shape_list(initial_output))
target_modality = self._problem_hparams.modality["targets"]
if target_modality == modalities.ModalityType.CLASS_LABEL:
decode_length = 1
else:
if "partial_targets" in features:
prefix_length = common_layers.shape_list(features["partial_targets"])[1]
else:
prefix_length = common_layers.shape_list(features["inputs"])[1]
decode_length = prefix_length + decode_length
# Initial values of result, logits and loss.
result = tf.concat(
[initial_output,
tf.zeros([batch_size, decode_length, 1, 1], tf.int64)],
axis=1)
# tensor padded to [batch_size, decode_length, 1, 1, vocab_size]
vocab_size = self._problem_hparams.vocab_size["targets"]
if vocab_size is not None and hasattr(self._hparams, "vocab_divisor"):
vocab_size += (-vocab_size) % self._hparams.vocab_divisor
logits = tf.zeros((batch_size, decode_length, 1, 1, vocab_size))
if not tf.executing_eagerly():
logits.set_shape([None, None, None, None, None])
loss = 0.0
def while_exit_cond(i, result, logits, loss): # pylint: disable=unused-argument
"""Exit the loop either if reach decode_length or EOS."""
not_overflow = i < decode_length
if self._problem_hparams.stop_at_eos:
def fn_not_eos():
# Check if the last predicted element is a EOS
return tf.reduce_any(
tf.not_equal(
tf.squeeze(result[:, -1, :, :]), text_encoder.EOS_ID))
not_eos = tf.cond(
# We only check for early stopping if there is at least 1 element (
# otherwise not_eos will crash).
tf.not_equal(i, 0),
fn_not_eos,
lambda: True,
)
return tf.cond(
tf.equal(batch_size, 1),
# If batch_size == 1, we check EOS for early stopping.
lambda: tf.logical_and(not_overflow, not_eos),
# Else, just wait for max length
lambda: not_overflow)
return not_overflow
_, result, logits, loss = tf.while_loop(
while_exit_cond,
infer_step, [tf.constant(0), result, logits, loss],
shape_invariants=[
tf.TensorShape([]),
tf.TensorShape([batch_size, decode_length, 1, 1]),
tf.TensorShape([batch_size, decode_length, 1, 1, vocab_size]),
tf.TensorShape([]),
],
back_prop=False,
parallel_iterations=1)
if inputs_old is not None: # Restore to not confuse Estimator.
features["inputs"] = inputs_old
# Reassign targets back to the previous value.
if targets_old is not None:
features["targets"] = targets_old
losses = {"training": loss}
if "partial_targets" in features:
partial_target_length = common_layers.shape_list(
features["partial_targets"])[1]
result = tf.slice(result, [0, partial_target_length, 0, 0],
[-1, -1, -1, -1])
return {
"outputs": result,
"scores": None,
"logits": logits,
"losses": losses,
}
|
[
"def",
"_slow_greedy_infer_tpu",
"(",
"self",
",",
"features",
",",
"decode_length",
")",
":",
"if",
"not",
"features",
":",
"features",
"=",
"{",
"}",
"inputs_old",
"=",
"None",
"if",
"\"inputs\"",
"in",
"features",
"and",
"len",
"(",
"features",
"[",
"\"inputs\"",
"]",
".",
"shape",
")",
"<",
"4",
":",
"inputs_old",
"=",
"features",
"[",
"\"inputs\"",
"]",
"features",
"[",
"\"inputs\"",
"]",
"=",
"tf",
".",
"expand_dims",
"(",
"features",
"[",
"\"inputs\"",
"]",
",",
"2",
")",
"if",
"not",
"self",
".",
"has_input",
":",
"# Prepare partial targets.",
"# In either features[\"inputs\"] or features[\"targets\"].",
"# We force the outputs to begin with these sequences.",
"partial_targets",
"=",
"features",
".",
"get",
"(",
"\"inputs\"",
")",
"if",
"partial_targets",
"is",
"None",
":",
"partial_targets",
"=",
"features",
"[",
"\"targets\"",
"]",
"features",
"[",
"\"partial_targets\"",
"]",
"=",
"tf",
".",
"to_int64",
"(",
"partial_targets",
")",
"# Save the targets in a var and reassign it after the tf.while loop to avoid",
"# having targets being in a 'while' frame. This ensures targets when used",
"# in metric functions stays in the same frame as other vars.",
"targets_old",
"=",
"features",
".",
"get",
"(",
"\"targets\"",
",",
"None",
")",
"target_modality",
"=",
"self",
".",
"_problem_hparams",
".",
"modality",
"[",
"\"targets\"",
"]",
"def",
"infer_step",
"(",
"i",
",",
"recent_output",
",",
"recent_logits",
",",
"unused_loss",
")",
":",
"\"\"\"Inference step.\"\"\"",
"if",
"not",
"tf",
".",
"executing_eagerly",
"(",
")",
":",
"recent_output",
".",
"set_shape",
"(",
"[",
"None",
",",
"None",
",",
"None",
",",
"1",
"]",
")",
"padded",
"=",
"tf",
".",
"pad",
"(",
"recent_output",
",",
"[",
"[",
"0",
",",
"0",
"]",
",",
"[",
"0",
",",
"1",
"]",
",",
"[",
"0",
",",
"0",
"]",
",",
"[",
"0",
",",
"0",
"]",
"]",
")",
"features",
"[",
"\"targets\"",
"]",
"=",
"padded",
"# This is inefficient in that it generates samples at all timesteps,",
"# not just the last one, except if target_modality is pointwise.",
"features",
"[",
"\"decode_loop_step\"",
"]",
"=",
"i",
"samples",
",",
"logits",
",",
"losses",
"=",
"self",
".",
"sample",
"(",
"features",
")",
"# Concatenate the already-generated recent_output with last timestep",
"# of the newly-generated samples.z",
"top",
"=",
"self",
".",
"_hparams",
".",
"top",
".",
"get",
"(",
"\"targets\"",
",",
"modalities",
".",
"get_top",
"(",
"target_modality",
")",
")",
"if",
"getattr",
"(",
"top",
",",
"\"pointwise\"",
",",
"False",
")",
":",
"cur_sample",
"=",
"samples",
"[",
":",
",",
"-",
"1",
",",
":",
",",
":",
"]",
"else",
":",
"cur_sample",
"=",
"samples",
"[",
":",
",",
"i",
",",
":",
",",
":",
"]",
"samples",
"=",
"tf",
".",
"transpose",
"(",
"recent_output",
",",
"perm",
"=",
"[",
"1",
",",
"0",
",",
"2",
",",
"3",
"]",
")",
"samples",
"=",
"inplace_ops",
".",
"alias_inplace_update",
"(",
"samples",
",",
"i",
",",
"tf",
".",
"to_int64",
"(",
"cur_sample",
")",
")",
"samples",
"=",
"tf",
".",
"transpose",
"(",
"samples",
",",
"perm",
"=",
"[",
"1",
",",
"0",
",",
"2",
",",
"3",
"]",
")",
"if",
"not",
"tf",
".",
"executing_eagerly",
"(",
")",
":",
"samples",
".",
"set_shape",
"(",
"[",
"None",
",",
"None",
",",
"None",
",",
"1",
"]",
")",
"# Assuming we have one shard for logits.",
"recent_logits",
"=",
"tf",
".",
"transpose",
"(",
"recent_logits",
",",
"perm",
"=",
"[",
"1",
",",
"0",
",",
"2",
",",
"3",
",",
"4",
"]",
")",
"recent_logits",
"=",
"inplace_ops",
".",
"alias_inplace_update",
"(",
"recent_logits",
",",
"i",
",",
"tf",
".",
"squeeze",
"(",
"logits",
"[",
":",
",",
"-",
"1",
":",
"]",
",",
"axis",
"=",
"1",
")",
")",
"logits",
"=",
"tf",
".",
"transpose",
"(",
"recent_logits",
",",
"perm",
"=",
"[",
"1",
",",
"0",
",",
"2",
",",
"3",
",",
"4",
"]",
")",
"loss",
"=",
"sum",
"(",
"[",
"l",
"for",
"l",
"in",
"losses",
".",
"values",
"(",
")",
"if",
"l",
"is",
"not",
"None",
"]",
")",
"return",
"i",
"+",
"1",
",",
"samples",
",",
"logits",
",",
"loss",
"# Create an initial output tensor. This will be passed",
"# to the infer_step, which adds one timestep at every iteration.",
"if",
"\"partial_targets\"",
"in",
"features",
":",
"initial_output",
"=",
"tf",
".",
"to_int64",
"(",
"features",
"[",
"\"partial_targets\"",
"]",
")",
"while",
"len",
"(",
"initial_output",
".",
"get_shape",
"(",
")",
".",
"as_list",
"(",
")",
")",
"<",
"4",
":",
"initial_output",
"=",
"tf",
".",
"expand_dims",
"(",
"initial_output",
",",
"2",
")",
"batch_size",
"=",
"common_layers",
".",
"shape_list",
"(",
"initial_output",
")",
"[",
"0",
"]",
"else",
":",
"batch_size",
"=",
"common_layers",
".",
"shape_list",
"(",
"features",
"[",
"\"inputs\"",
"]",
")",
"[",
"0",
"]",
"initial_output",
"=",
"tf",
".",
"zeros",
"(",
"(",
"batch_size",
",",
"0",
",",
"1",
",",
"1",
")",
",",
"dtype",
"=",
"tf",
".",
"int64",
")",
"# Hack: foldl complains when the output shape is less specified than the",
"# input shape, so we confuse it about the input shape.",
"initial_output",
"=",
"tf",
".",
"slice",
"(",
"initial_output",
",",
"[",
"0",
",",
"0",
",",
"0",
",",
"0",
"]",
",",
"common_layers",
".",
"shape_list",
"(",
"initial_output",
")",
")",
"target_modality",
"=",
"self",
".",
"_problem_hparams",
".",
"modality",
"[",
"\"targets\"",
"]",
"if",
"target_modality",
"==",
"modalities",
".",
"ModalityType",
".",
"CLASS_LABEL",
":",
"decode_length",
"=",
"1",
"else",
":",
"if",
"\"partial_targets\"",
"in",
"features",
":",
"prefix_length",
"=",
"common_layers",
".",
"shape_list",
"(",
"features",
"[",
"\"partial_targets\"",
"]",
")",
"[",
"1",
"]",
"else",
":",
"prefix_length",
"=",
"common_layers",
".",
"shape_list",
"(",
"features",
"[",
"\"inputs\"",
"]",
")",
"[",
"1",
"]",
"decode_length",
"=",
"prefix_length",
"+",
"decode_length",
"# Initial values of result, logits and loss.",
"result",
"=",
"tf",
".",
"concat",
"(",
"[",
"initial_output",
",",
"tf",
".",
"zeros",
"(",
"[",
"batch_size",
",",
"decode_length",
",",
"1",
",",
"1",
"]",
",",
"tf",
".",
"int64",
")",
"]",
",",
"axis",
"=",
"1",
")",
"# tensor padded to [batch_size, decode_length, 1, 1, vocab_size]",
"vocab_size",
"=",
"self",
".",
"_problem_hparams",
".",
"vocab_size",
"[",
"\"targets\"",
"]",
"if",
"vocab_size",
"is",
"not",
"None",
"and",
"hasattr",
"(",
"self",
".",
"_hparams",
",",
"\"vocab_divisor\"",
")",
":",
"vocab_size",
"+=",
"(",
"-",
"vocab_size",
")",
"%",
"self",
".",
"_hparams",
".",
"vocab_divisor",
"logits",
"=",
"tf",
".",
"zeros",
"(",
"(",
"batch_size",
",",
"decode_length",
",",
"1",
",",
"1",
",",
"vocab_size",
")",
")",
"if",
"not",
"tf",
".",
"executing_eagerly",
"(",
")",
":",
"logits",
".",
"set_shape",
"(",
"[",
"None",
",",
"None",
",",
"None",
",",
"None",
",",
"None",
"]",
")",
"loss",
"=",
"0.0",
"def",
"while_exit_cond",
"(",
"i",
",",
"result",
",",
"logits",
",",
"loss",
")",
":",
"# pylint: disable=unused-argument",
"\"\"\"Exit the loop either if reach decode_length or EOS.\"\"\"",
"not_overflow",
"=",
"i",
"<",
"decode_length",
"if",
"self",
".",
"_problem_hparams",
".",
"stop_at_eos",
":",
"def",
"fn_not_eos",
"(",
")",
":",
"# Check if the last predicted element is a EOS",
"return",
"tf",
".",
"reduce_any",
"(",
"tf",
".",
"not_equal",
"(",
"tf",
".",
"squeeze",
"(",
"result",
"[",
":",
",",
"-",
"1",
",",
":",
",",
":",
"]",
")",
",",
"text_encoder",
".",
"EOS_ID",
")",
")",
"not_eos",
"=",
"tf",
".",
"cond",
"(",
"# We only check for early stopping if there is at least 1 element (",
"# otherwise not_eos will crash).",
"tf",
".",
"not_equal",
"(",
"i",
",",
"0",
")",
",",
"fn_not_eos",
",",
"lambda",
":",
"True",
",",
")",
"return",
"tf",
".",
"cond",
"(",
"tf",
".",
"equal",
"(",
"batch_size",
",",
"1",
")",
",",
"# If batch_size == 1, we check EOS for early stopping.",
"lambda",
":",
"tf",
".",
"logical_and",
"(",
"not_overflow",
",",
"not_eos",
")",
",",
"# Else, just wait for max length",
"lambda",
":",
"not_overflow",
")",
"return",
"not_overflow",
"_",
",",
"result",
",",
"logits",
",",
"loss",
"=",
"tf",
".",
"while_loop",
"(",
"while_exit_cond",
",",
"infer_step",
",",
"[",
"tf",
".",
"constant",
"(",
"0",
")",
",",
"result",
",",
"logits",
",",
"loss",
"]",
",",
"shape_invariants",
"=",
"[",
"tf",
".",
"TensorShape",
"(",
"[",
"]",
")",
",",
"tf",
".",
"TensorShape",
"(",
"[",
"batch_size",
",",
"decode_length",
",",
"1",
",",
"1",
"]",
")",
",",
"tf",
".",
"TensorShape",
"(",
"[",
"batch_size",
",",
"decode_length",
",",
"1",
",",
"1",
",",
"vocab_size",
"]",
")",
",",
"tf",
".",
"TensorShape",
"(",
"[",
"]",
")",
",",
"]",
",",
"back_prop",
"=",
"False",
",",
"parallel_iterations",
"=",
"1",
")",
"if",
"inputs_old",
"is",
"not",
"None",
":",
"# Restore to not confuse Estimator.",
"features",
"[",
"\"inputs\"",
"]",
"=",
"inputs_old",
"# Reassign targets back to the previous value.",
"if",
"targets_old",
"is",
"not",
"None",
":",
"features",
"[",
"\"targets\"",
"]",
"=",
"targets_old",
"losses",
"=",
"{",
"\"training\"",
":",
"loss",
"}",
"if",
"\"partial_targets\"",
"in",
"features",
":",
"partial_target_length",
"=",
"common_layers",
".",
"shape_list",
"(",
"features",
"[",
"\"partial_targets\"",
"]",
")",
"[",
"1",
"]",
"result",
"=",
"tf",
".",
"slice",
"(",
"result",
",",
"[",
"0",
",",
"partial_target_length",
",",
"0",
",",
"0",
"]",
",",
"[",
"-",
"1",
",",
"-",
"1",
",",
"-",
"1",
",",
"-",
"1",
"]",
")",
"return",
"{",
"\"outputs\"",
":",
"result",
",",
"\"scores\"",
":",
"None",
",",
"\"logits\"",
":",
"logits",
",",
"\"losses\"",
":",
"losses",
",",
"}"
] |
A slow greedy inference method on TPU.
Quadratic time in decode_length.
Args:
features: An map of string to `Tensor`.
decode_length: An integer, how many additional timesteps to decode.
Returns:
A dict of decoding results {
"outputs": integer `Tensor` of decoded ids of shape
[batch_size, <= decode_length] if beam_size == 1 or
[batch_size, top_beams, <= decode_length]
"scores": None
"logits": `Tensor` of shape [batch_size, time, 1, 1, vocab_size].
"losses": a dictionary: {loss-name (string): floating point `Scalar`}
}
|
[
"A",
"slow",
"greedy",
"inference",
"method",
"on",
"TPU",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L977-L1143
|
train
|
A slow greedy inference method on TPU.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110001) + chr(51) + '\063', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(1564 - 1514) + '\061' + chr(55), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(11519 - 11408) + chr(49) + '\x35' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(1859 - 1811) + '\x6f' + '\x33' + chr(0b110111) + chr(0b100110 + 0o12), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(1323 - 1212) + chr(0b110001) + '\066', 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(822 - 772) + chr(0b1011 + 0o45) + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(1367 - 1319) + chr(0b1001101 + 0o42) + '\x31' + chr(0b110110) + '\x36', 17964 - 17956), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(244 - 193) + '\063' + chr(234 - 186), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(51) + chr(0b110000) + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(0b1101111) + '\x31' + chr(55) + chr(0b110100), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + '\x31', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062' + chr(51) + chr(209 - 160), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(2469 - 2358) + '\063' + chr(48) + chr(1254 - 1205), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(49) + '\067' + '\x36', 53906 - 53898), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011 + 0o0) + chr(0b1001 + 0o50), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110011) + '\060' + '\x35', 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + chr(0b101101 + 0o5) + '\x37', 54114 - 54106), ehT0Px3KOsy9(chr(48) + chr(111) + chr(50) + chr(1255 - 1206) + chr(0b110001 + 0o5), 28761 - 28753), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001) + '\x30', 2330 - 2322), ehT0Px3KOsy9('\x30' + chr(11726 - 11615) + chr(0b110011) + chr(0b11010 + 0o31), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(0b100011 + 0o17) + chr(0b11011 + 0o27) + '\060', 0b1000), ehT0Px3KOsy9(chr(1141 - 1093) + '\x6f' + chr(0b1 + 0o60) + '\x37' + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101001 + 0o6) + chr(0b110 + 0o53) + chr(51) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\063' + chr(0b111 + 0o56) + chr(0b110111), 34561 - 34553), ehT0Px3KOsy9(chr(383 - 335) + chr(0b1010010 + 0o35) + chr(0b110010) + '\x30' + chr(0b110111), 8), ehT0Px3KOsy9('\x30' + chr(4264 - 4153) + '\061' + chr(0b11010 + 0o30) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(593 - 545) + chr(6297 - 6186) + '\x32' + chr(0b11 + 0o62), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(5378 - 5267) + chr(0b100111 + 0o13) + chr(0b10010 + 0o45) + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(48) + chr(10630 - 10519) + '\x30', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101100 + 0o3) + chr(0b110011) + chr(54) + chr(50), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b11001 + 0o32) + '\060' + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110011) + '\x33' + chr(50), 2287 - 2279), ehT0Px3KOsy9(chr(1875 - 1827) + chr(0b1101111) + chr(50) + '\x30', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33' + chr(49) + chr(48), 50852 - 50844), ehT0Px3KOsy9(chr(0b110000) + chr(4962 - 4851) + '\x32' + chr(0b110100) + chr(1523 - 1468), 0b1000), ehT0Px3KOsy9(chr(0b10001 + 0o37) + '\x6f' + chr(0b110011) + chr(0b110111) + chr(0b101 + 0o60), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(51) + chr(50) + chr(98 - 43), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\066' + chr(48), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b10101 + 0o132) + chr(49) + '\x30' + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1692 - 1640) + chr(0b100000 + 0o22), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(108 - 60) + chr(0b111100 + 0o63) + '\x35' + '\060', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'u'), chr(0b111001 + 0o53) + chr(101) + chr(1186 - 1087) + chr(0b1000001 + 0o56) + '\144' + chr(0b1010001 + 0o24))(chr(0b1101011 + 0o12) + chr(116) + chr(102) + '\x2d' + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def YF74WpZ18BcK(oVre8I6UXc3b, EEf4r9nUvta_, U6Ej34SVvx1Y):
if not EEf4r9nUvta_:
EEf4r9nUvta_ = {}
G9nVVSWl2V2F = None
if xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), chr(0b1000000 + 0o44) + chr(0b1100101) + chr(0b1100011) + chr(0b1010 + 0o145) + chr(0b1100100) + chr(0b1100101))(chr(0b1110101) + '\164' + '\x66' + chr(0b11001 + 0o24) + chr(0b111000)) in EEf4r9nUvta_ and c2A0yzQpDQB3(xafqLlk3kkUe(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), '\x64' + chr(0b1010000 + 0o25) + '\x63' + chr(2227 - 2116) + '\144' + '\x65')(chr(0b1110101) + chr(5577 - 5461) + chr(102) + chr(0b11001 + 0o24) + chr(3076 - 3020))], xafqLlk3kkUe(SXOLrMavuUCe(b'5\xee\x1f\xa1h\xcb\x11\xb4\x91\x0cb\xfb'), chr(894 - 794) + '\145' + '\x63' + chr(1595 - 1484) + chr(0b101111 + 0o65) + chr(0b1100101))(chr(117) + chr(12202 - 12086) + chr(2730 - 2628) + chr(45) + chr(324 - 268)))) < ehT0Px3KOsy9(chr(806 - 758) + chr(0b1101111) + chr(2058 - 2006), 0b1000):
G9nVVSWl2V2F = EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), chr(0b11001 + 0o113) + '\145' + chr(6268 - 6169) + chr(10743 - 10632) + chr(1256 - 1156) + '\x65')(chr(117) + chr(0b1101101 + 0o7) + '\146' + '\x2d' + chr(2897 - 2841))]
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), '\144' + '\x65' + '\143' + chr(3538 - 3427) + chr(0b1100100) + '\145')(chr(0b110 + 0o157) + chr(0b1110100) + chr(9814 - 9712) + chr(0b101101) + chr(2275 - 2219))] = IDJ2eXGCBCDu.expand_dims(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), chr(0b1100100) + chr(101) + '\143' + '\x6f' + '\144' + '\145')(chr(0b10011 + 0o142) + chr(0b1010110 + 0o36) + chr(0b100110 + 0o100) + '\x2d' + '\070')], ehT0Px3KOsy9('\x30' + chr(111) + chr(833 - 783), 4476 - 4468))
if not xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'3\xee\x19\xa7g\xe9\x06\xad\xb1'), chr(0b1100100) + chr(3829 - 3728) + chr(0b101110 + 0o65) + chr(0b1011101 + 0o22) + chr(0b11101 + 0o107) + '\145')(chr(117) + chr(116) + chr(0b1100110) + chr(45) + '\070')):
AojJ_vUpav2O = EEf4r9nUvta_.get(xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), chr(0b1100100) + '\x65' + chr(8509 - 8410) + chr(0b101111 + 0o100) + '\x64' + chr(0b1100101))(chr(0b110010 + 0o103) + chr(0b1000000 + 0o64) + chr(102) + '\x2d' + chr(56)))
if AojJ_vUpav2O is None:
AojJ_vUpav2O = EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'/\xee\x18\x9fk\xf3\x05'), '\144' + '\145' + '\143' + chr(111) + chr(7549 - 7449) + chr(0b1100101))(chr(0b1 + 0o164) + chr(116) + chr(0b1100110) + '\055' + chr(0b111000))]
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'+\xee\x18\x8cg\xe6\x1a\x87\xb1\x1ds\xfe\xa4\nj'), '\144' + chr(101) + chr(99) + chr(111) + '\x64' + '\145')('\165' + chr(0b110011 + 0o101) + chr(3853 - 3751) + chr(0b101101) + '\x38')] = IDJ2eXGCBCDu.to_int64(AojJ_vUpav2O)
VAnVTyqvOaPL = EEf4r9nUvta_.get(xafqLlk3kkUe(SXOLrMavuUCe(b'/\xee\x18\x9fk\xf3\x05'), '\x64' + '\145' + '\x63' + chr(0b1101111) + '\x64' + chr(101))('\165' + chr(0b1110100) + '\x66' + chr(0b101001 + 0o4) + '\x38'), None)
FfV_wqUqiesq = oVre8I6UXc3b._problem_hparams.bYPswhysd3s2[xafqLlk3kkUe(SXOLrMavuUCe(b'/\xee\x18\x9fk\xf3\x05'), '\x64' + '\145' + chr(0b1100011) + '\x6f' + chr(0b11110 + 0o106) + chr(101))(chr(0b1011 + 0o152) + chr(3114 - 2998) + '\146' + '\055' + '\x38')]
def WInTwI4hMByi(WVxHKyX45z_L, hlItJECfkP_5, jCJ2auWHy9tB, lYwnyBxWZqd8):
if not xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'>\xf7\x0f\x9b{\xf3\x1f\xb6\xa2#d\xf8\xa6\x1bk\xa9.'), '\x64' + chr(101) + chr(4617 - 4518) + chr(0b1101111) + '\x64' + chr(8189 - 8088))('\x75' + chr(116) + chr(102) + '\055' + chr(2478 - 2422)))():
xafqLlk3kkUe(hlItJECfkP_5, xafqLlk3kkUe(SXOLrMavuUCe(b'(\xea\x1e\xa7}\xef\x17\xa8\xa0'), '\x64' + chr(4356 - 4255) + chr(0b1100011) + '\157' + chr(0b111100 + 0o50) + chr(459 - 358))(chr(8377 - 8260) + chr(12954 - 12838) + chr(1510 - 1408) + chr(0b1000 + 0o45) + '\x38'))([None, None, None, ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(6741 - 6630) + chr(0b110001), 8)])
Jr6qMmXilxlt = IDJ2eXGCBCDu.pad(hlItJECfkP_5, [[ehT0Px3KOsy9('\060' + chr(111) + chr(0b110000), 8), ehT0Px3KOsy9('\x30' + chr(111) + chr(48), 8)], [ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x30', 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(49), 8)], [ehT0Px3KOsy9('\x30' + chr(0b1000001 + 0o56) + '\x30', 8), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(111) + '\x30', 8)], [ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\060', 8), ehT0Px3KOsy9(chr(0b10000 + 0o40) + '\x6f' + chr(285 - 237), 8)]])
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'/\xee\x18\x9fk\xf3\x05'), chr(100) + chr(0b1100101) + chr(0b1001111 + 0o24) + '\157' + chr(646 - 546) + '\145')(chr(0b1110101) + chr(0b1110100) + chr(102) + '\x2d' + '\x38')] = Jr6qMmXilxlt
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'?\xea\t\x97j\xe2)\xb4\xaa\x13q\xc6\xb2\n|\xb5'), chr(100) + '\145' + chr(99) + chr(1757 - 1646) + '\x64' + '\x65')(chr(0b1000000 + 0o65) + chr(0b1110100) + chr(0b1000011 + 0o43) + '\x2d' + chr(0b101 + 0o63))] = WVxHKyX45z_L
(db1_IZvznkcy, wF9nmvjsKjYM, eJKWkHA7qzlZ) = oVre8I6UXc3b.sample(EEf4r9nUvta_)
qxrVBjeryNEZ = oVre8I6UXc3b._hparams.top.get(xafqLlk3kkUe(SXOLrMavuUCe(b'/\xee\x18\x9fk\xf3\x05'), '\144' + '\x65' + '\x63' + chr(9338 - 9227) + '\144' + chr(101))('\x75' + '\x74' + chr(0b1011101 + 0o11) + chr(0b1100 + 0o41) + chr(0b101101 + 0o13)), PuPeNl0CuqOQ.get_top(FfV_wqUqiesq))
if xafqLlk3kkUe(qxrVBjeryNEZ, xafqLlk3kkUe(SXOLrMavuUCe(b'+\xe0\x03\x96z\xf0\x1f\xab\xa0'), chr(8865 - 8765) + chr(0b1100101) + chr(0b101110 + 0o65) + chr(0b1101111) + chr(100) + chr(0b1100101))(chr(5717 - 5600) + '\164' + chr(829 - 727) + chr(0b101101) + chr(56)), ehT0Px3KOsy9(chr(0b11010 + 0o26) + '\157' + chr(0b11111 + 0o21), 8)):
x9og61xsXiTP = db1_IZvznkcy[:, -ehT0Px3KOsy9(chr(0b11100 + 0o24) + '\x6f' + '\061', 8), :, :]
else:
x9og61xsXiTP = db1_IZvznkcy[:, WVxHKyX45z_L, :, :]
db1_IZvznkcy = IDJ2eXGCBCDu.transpose(hlItJECfkP_5, perm=[ehT0Px3KOsy9(chr(48) + '\x6f' + '\x31', 8), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1758 - 1710), 8), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110010), 8), ehT0Px3KOsy9(chr(48) + chr(0b1101100 + 0o3) + chr(0b1111 + 0o44), 0o10)])
db1_IZvznkcy = GanXbkgpxGLx.alias_inplace_update(db1_IZvznkcy, WVxHKyX45z_L, IDJ2eXGCBCDu.to_int64(x9og61xsXiTP))
db1_IZvznkcy = IDJ2eXGCBCDu.transpose(db1_IZvznkcy, perm=[ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001), 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\060', 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062', 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33', 8)])
if not xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'>\xf7\x0f\x9b{\xf3\x1f\xb6\xa2#d\xf8\xa6\x1bk\xa9.'), chr(100) + chr(101) + chr(99) + '\157' + '\144' + '\x65')('\165' + chr(0b1110100) + '\x66' + '\055' + chr(0b111000)))():
xafqLlk3kkUe(db1_IZvznkcy, xafqLlk3kkUe(SXOLrMavuUCe(b'(\xea\x1e\xa7}\xef\x17\xa8\xa0'), '\x64' + chr(101) + chr(0b10110 + 0o115) + chr(111) + chr(0b1011011 + 0o11) + '\x65')(chr(0b111101 + 0o70) + chr(116) + '\146' + '\x2d' + chr(0b111000)))([None, None, None, ehT0Px3KOsy9('\060' + chr(2225 - 2114) + chr(1219 - 1170), 8)])
jCJ2auWHy9tB = IDJ2eXGCBCDu.transpose(jCJ2auWHy9tB, perm=[ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(111) + '\x31', 8), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x30', 8), ehT0Px3KOsy9('\060' + chr(5231 - 5120) + chr(0b100011 + 0o17), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b100011 + 0o114) + '\x33', 8), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110100), 8)])
jCJ2auWHy9tB = GanXbkgpxGLx.alias_inplace_update(jCJ2auWHy9tB, WVxHKyX45z_L, IDJ2eXGCBCDu.squeeze(wF9nmvjsKjYM[:, -ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x31', 8):], axis=ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b10011 + 0o36), 8)))
wF9nmvjsKjYM = IDJ2eXGCBCDu.transpose(jCJ2auWHy9tB, perm=[ehT0Px3KOsy9('\060' + '\157' + '\061', 8), ehT0Px3KOsy9(chr(48) + chr(111) + chr(903 - 855), 8), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(111) + chr(0b110010), 8), ehT0Px3KOsy9(chr(446 - 398) + chr(10346 - 10235) + chr(51), 8), ehT0Px3KOsy9('\x30' + chr(0b10111 + 0o130) + chr(52), 8)])
YpO0BcZ6fMsf = xkxBmo49x2An([aLoH_Mt0dzwO for aLoH_Mt0dzwO in eJKWkHA7qzlZ.SPnCNu54H1db() if aLoH_Mt0dzwO is not None])
return (WVxHKyX45z_L + ehT0Px3KOsy9('\x30' + chr(111) + chr(49), 8), db1_IZvznkcy, wF9nmvjsKjYM, YpO0BcZ6fMsf)
if xafqLlk3kkUe(SXOLrMavuUCe(b'+\xee\x18\x8cg\xe6\x1a\x87\xb1\x1ds\xfe\xa4\nj'), '\144' + chr(101) + chr(0b11010 + 0o111) + '\157' + chr(100) + chr(0b101101 + 0o70))(chr(0b1110101) + chr(116) + chr(0b1100110) + '\055' + chr(0b111000)) in EEf4r9nUvta_:
ESoMhcTQCSlw = IDJ2eXGCBCDu.to_int64(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'+\xee\x18\x8cg\xe6\x1a\x87\xb1\x1ds\xfe\xa4\nj'), chr(1459 - 1359) + '\145' + chr(0b101001 + 0o72) + '\157' + '\x64' + '\x65')(chr(0b1001010 + 0o53) + chr(116) + '\146' + '\x2d' + chr(56))])
while c2A0yzQpDQB3(xafqLlk3kkUe(ESoMhcTQCSlw.get_shape(), xafqLlk3kkUe(SXOLrMavuUCe(b':\xfc5\x94g\xf4\x02'), '\x64' + chr(8919 - 8818) + chr(0b1010100 + 0o17) + chr(0b1101111) + chr(3472 - 3372) + '\145')(chr(0b111010 + 0o73) + chr(116) + chr(0b110000 + 0o66) + chr(0b1010 + 0o43) + chr(2209 - 2153)))()) < ehT0Px3KOsy9(chr(987 - 939) + '\157' + chr(52), 8):
ESoMhcTQCSlw = IDJ2eXGCBCDu.expand_dims(ESoMhcTQCSlw, ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b10011 + 0o37), 8))
ix9dZyeAmUxY = jSKPaHwSAfVv.shape_list(ESoMhcTQCSlw)[ehT0Px3KOsy9(chr(48) + chr(0b1100011 + 0o14) + chr(0b110000 + 0o0), 8)]
else:
ix9dZyeAmUxY = jSKPaHwSAfVv.shape_list(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), '\x64' + '\145' + chr(0b1001101 + 0o26) + chr(3502 - 3391) + chr(9361 - 9261) + chr(5998 - 5897))(chr(0b111001 + 0o74) + chr(0b1011 + 0o151) + '\146' + chr(0b11010 + 0o23) + chr(0b1100 + 0o54))])[ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(48), 8)]
ESoMhcTQCSlw = IDJ2eXGCBCDu.zeros((ix9dZyeAmUxY, ehT0Px3KOsy9(chr(48) + '\157' + '\060', 8), ehT0Px3KOsy9('\x30' + chr(0b1101000 + 0o7) + chr(0b110001), 8), ehT0Px3KOsy9(chr(1904 - 1856) + chr(0b100100 + 0o113) + chr(0b100001 + 0o20), 8)), dtype=IDJ2eXGCBCDu.int64)
ESoMhcTQCSlw = IDJ2eXGCBCDu.slice(ESoMhcTQCSlw, [ehT0Px3KOsy9(chr(1271 - 1223) + chr(1706 - 1595) + '\x30', 8), ehT0Px3KOsy9(chr(76 - 28) + '\157' + chr(48), 8), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(524 - 413) + chr(0b100011 + 0o15), 8), ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(111) + chr(0b110000), 8)], jSKPaHwSAfVv.shape_list(ESoMhcTQCSlw))
FfV_wqUqiesq = oVre8I6UXc3b._problem_hparams.bYPswhysd3s2[xafqLlk3kkUe(SXOLrMavuUCe(b'/\xee\x18\x9fk\xf3\x05'), '\x64' + '\145' + chr(0b11010 + 0o111) + chr(4949 - 4838) + chr(100) + chr(0b1100101))('\165' + chr(0b110111 + 0o75) + '\146' + '\x2d' + '\x38')]
if FfV_wqUqiesq == xafqLlk3kkUe(PuPeNl0CuqOQ.ModalityType, xafqLlk3kkUe(SXOLrMavuUCe(b'\x18\xc3+\xab]\xd8:\x99\x879M'), '\144' + chr(8609 - 8508) + chr(99) + chr(0b1101111) + chr(0b100010 + 0o102) + chr(101))(chr(0b1110101) + '\164' + chr(7704 - 7602) + chr(45) + chr(0b111000))):
U6Ej34SVvx1Y = ehT0Px3KOsy9(chr(48) + chr(0b100111 + 0o110) + chr(0b1111 + 0o42), 8)
else:
if xafqLlk3kkUe(SXOLrMavuUCe(b'+\xee\x18\x8cg\xe6\x1a\x87\xb1\x1ds\xfe\xa4\nj'), '\144' + chr(0b1100101) + '\143' + '\157' + '\x64' + '\x65')(chr(117) + chr(6839 - 6723) + '\146' + '\055' + '\070') in EEf4r9nUvta_:
KI4eMsPc0Vvz = jSKPaHwSAfVv.shape_list(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'+\xee\x18\x8cg\xe6\x1a\x87\xb1\x1ds\xfe\xa4\nj'), chr(0b111110 + 0o46) + chr(0b1100101) + chr(0b1100011) + '\157' + '\144' + chr(1797 - 1696))('\165' + '\164' + chr(0b1011111 + 0o7) + '\055' + chr(56))])[ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31', 8)]
else:
KI4eMsPc0Vvz = jSKPaHwSAfVv.shape_list(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), chr(0b10110 + 0o116) + '\145' + '\x63' + chr(0b1101111) + '\x64' + chr(0b100 + 0o141))(chr(3799 - 3682) + '\x74' + chr(0b1100011 + 0o3) + chr(0b1011 + 0o42) + '\070')])[ehT0Px3KOsy9(chr(48) + chr(7461 - 7350) + chr(1905 - 1856), 8)]
U6Ej34SVvx1Y = KI4eMsPc0Vvz + U6Ej34SVvx1Y
ShZmEKfTkAOZ = IDJ2eXGCBCDu.concat([ESoMhcTQCSlw, IDJ2eXGCBCDu.zeros([ix9dZyeAmUxY, U6Ej34SVvx1Y, ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(0b1101111) + chr(0b101100 + 0o5), 8), ehT0Px3KOsy9('\x30' + chr(1741 - 1630) + chr(0b1000 + 0o51), 8)], IDJ2eXGCBCDu.int64)], axis=ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49), 8))
CeyMIoSyrpkQ = oVre8I6UXc3b._problem_hparams.CeyMIoSyrpkQ[xafqLlk3kkUe(SXOLrMavuUCe(b'/\xee\x18\x9fk\xf3\x05'), chr(0b1100100) + '\145' + chr(99) + '\157' + chr(3741 - 3641) + '\145')(chr(2383 - 2266) + chr(116) + chr(250 - 148) + '\x2d' + chr(0b100000 + 0o30))]
if CeyMIoSyrpkQ is not None and lot1PSoAwYhj(xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xdf%\xcdM\xff\x00\x87\x97/P\xc8'), '\x64' + '\145' + chr(99) + chr(111) + chr(0b1000011 + 0o41) + chr(0b1100101))('\165' + chr(12696 - 12580) + '\x66' + chr(581 - 536) + chr(56))), xafqLlk3kkUe(SXOLrMavuUCe(b'-\xe0\t\x99l\xd8\x12\xb1\xb3\x15r\xf6\xb3'), '\144' + '\145' + chr(0b1100011) + chr(4958 - 4847) + '\144' + chr(0b1010011 + 0o22))('\165' + chr(0b1000101 + 0o57) + chr(102) + chr(45) + chr(56))):
CeyMIoSyrpkQ += -CeyMIoSyrpkQ % oVre8I6UXc3b._hparams.vocab_divisor
wF9nmvjsKjYM = IDJ2eXGCBCDu.zeros((ix9dZyeAmUxY, U6Ej34SVvx1Y, ehT0Px3KOsy9('\060' + '\157' + chr(0b110001), 8), ehT0Px3KOsy9(chr(0b101011 + 0o5) + '\x6f' + chr(0b11000 + 0o31), 8), CeyMIoSyrpkQ))
if not xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'>\xf7\x0f\x9b{\xf3\x1f\xb6\xa2#d\xf8\xa6\x1bk\xa9.'), chr(100) + chr(101) + '\x63' + chr(5006 - 4895) + chr(9002 - 8902) + chr(101))('\x75' + chr(0b100101 + 0o117) + chr(7160 - 7058) + chr(1159 - 1114) + chr(56)))():
xafqLlk3kkUe(wF9nmvjsKjYM, xafqLlk3kkUe(SXOLrMavuUCe(b'(\xea\x1e\xa7}\xef\x17\xa8\xa0'), '\144' + '\145' + '\143' + chr(6481 - 6370) + chr(9299 - 9199) + chr(0b100100 + 0o101))(chr(0b100001 + 0o124) + chr(116) + '\146' + chr(45) + '\x38'))([None, None, None, None, None])
YpO0BcZ6fMsf = 0.0
def QRAZ92UyPH1W(WVxHKyX45z_L, ShZmEKfTkAOZ, wF9nmvjsKjYM, YpO0BcZ6fMsf):
hqq34OblagiW = WVxHKyX45z_L < U6Ej34SVvx1Y
if xafqLlk3kkUe(oVre8I6UXc3b._problem_hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'(\xfb\x05\x88Q\xe6\x02\x87\xa0\x13r'), '\144' + '\145' + chr(0b1100011) + chr(111) + '\x64' + chr(8876 - 8775))(chr(0b1110101) + '\164' + '\146' + chr(748 - 703) + '\x38')):
def qCkbVx6I0YWc():
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b')\xea\x0e\x8dm\xe2)\xb9\xab\x05'), chr(100) + chr(7031 - 6930) + chr(0b1100011) + chr(12085 - 11974) + chr(0b1100100) + chr(0b1100101))('\x75' + chr(0b1110100) + '\146' + chr(0b101101) + chr(0b100 + 0o64)))(xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xe0\x1e\xa7k\xf6\x03\xb9\xa9'), chr(0b1010100 + 0o20) + '\x65' + '\143' + chr(1289 - 1178) + chr(100) + chr(0b1100101))(chr(11023 - 10906) + chr(116) + chr(102) + chr(1955 - 1910) + chr(56)))(xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'(\xfe\x1f\x9dk\xfd\x13'), chr(316 - 216) + '\145' + chr(0b1010100 + 0o17) + chr(0b1100100 + 0o13) + chr(0b1010010 + 0o22) + chr(0b10011 + 0o122))(chr(0b1110101) + '\x74' + '\x66' + chr(922 - 877) + '\x38'))(ShZmEKfTkAOZ[:, -ehT0Px3KOsy9(chr(48) + chr(0b1000100 + 0o53) + chr(0b1001 + 0o50), 8), :, :]), xafqLlk3kkUe(nCRDzZ_Is9fz, xafqLlk3kkUe(SXOLrMavuUCe(b'\x1e\xc09\xa7G\xc3'), chr(3604 - 3504) + chr(101) + chr(9326 - 9227) + chr(4472 - 4361) + chr(0b1100100) + chr(174 - 73))(chr(117) + chr(116) + chr(0b101111 + 0o67) + '\055' + chr(56)))))
qxiXp1QpYkkW = IDJ2eXGCBCDu.cond(IDJ2eXGCBCDu.not_equal(WVxHKyX45z_L, ehT0Px3KOsy9(chr(0b110000) + chr(11195 - 11084) + '\x30', 8)), qCkbVx6I0YWc, lambda : ehT0Px3KOsy9('\x30' + '\157' + chr(0b110001), 8))
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'8\xe0\x04\x9c'), chr(1743 - 1643) + chr(0b1100101) + chr(7694 - 7595) + chr(0b1101111) + chr(0b110010 + 0o62) + chr(0b1100101))('\x75' + chr(0b1110100) + '\146' + chr(45) + chr(0b10011 + 0o45)))(xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'>\xfe\x1f\x99b'), chr(0b1100100) + chr(101) + chr(4820 - 4721) + '\157' + '\144' + '\x65')(chr(197 - 80) + chr(0b1011001 + 0o33) + '\146' + chr(0b101101) + chr(0b1110 + 0o52)))(ix9dZyeAmUxY, ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\x6f' + chr(0b110001), 8)), lambda : xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'7\xe0\r\x91m\xe6\x1a\x87\xa4\x12e'), chr(3384 - 3284) + '\x65' + chr(0b1100011) + chr(111) + '\x64' + '\145')(chr(0b1110101) + chr(0b1001010 + 0o52) + '\x66' + chr(195 - 150) + chr(0b111 + 0o61)))(hqq34OblagiW, qxiXp1QpYkkW), lambda : hqq34OblagiW)
return hqq34OblagiW
(VNGQdHSFPrso, ShZmEKfTkAOZ, wF9nmvjsKjYM, YpO0BcZ6fMsf) = IDJ2eXGCBCDu.while_loop(QRAZ92UyPH1W, WInTwI4hMByi, [IDJ2eXGCBCDu.constant(ehT0Px3KOsy9('\060' + '\157' + chr(0b110000), 8)), ShZmEKfTkAOZ, wF9nmvjsKjYM, YpO0BcZ6fMsf], shape_invariants=[IDJ2eXGCBCDu.TensorShape([]), IDJ2eXGCBCDu.TensorShape([ix9dZyeAmUxY, U6Ej34SVvx1Y, ehT0Px3KOsy9(chr(0b10010 + 0o36) + '\157' + '\x31', 8), ehT0Px3KOsy9('\060' + '\157' + chr(1843 - 1794), 8)]), IDJ2eXGCBCDu.TensorShape([ix9dZyeAmUxY, U6Ej34SVvx1Y, ehT0Px3KOsy9('\060' + '\157' + '\061', 8), ehT0Px3KOsy9(chr(48) + chr(271 - 160) + '\x31', 8), CeyMIoSyrpkQ]), IDJ2eXGCBCDu.TensorShape([])], back_prop=ehT0Px3KOsy9(chr(126 - 78) + chr(0b1101111) + chr(0b11010 + 0o26), 8), parallel_iterations=ehT0Px3KOsy9(chr(494 - 446) + '\x6f' + chr(0b110001), 8))
if G9nVVSWl2V2F is not None:
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'2\xe1\x1a\x8dz\xf4'), '\144' + '\145' + '\x63' + '\157' + chr(6739 - 6639) + '\x65')(chr(117) + '\x74' + chr(102) + '\055' + chr(2671 - 2615))] = G9nVVSWl2V2F
if VAnVTyqvOaPL is not None:
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'/\xee\x18\x9fk\xf3\x05'), chr(0b1100100) + chr(101) + chr(0b1100011) + '\x6f' + chr(0b1100100) + chr(1159 - 1058))(chr(117) + chr(0b1110100) + chr(2725 - 2623) + '\x2d' + chr(56))] = VAnVTyqvOaPL
eJKWkHA7qzlZ = {xafqLlk3kkUe(SXOLrMavuUCe(b'/\xfd\x0b\x91`\xee\x18\xbf'), '\x64' + chr(101) + '\143' + chr(111) + '\x64' + '\145')(chr(0b1110101) + chr(0b1000000 + 0o64) + chr(0b1100110) + '\055' + chr(2414 - 2358)): YpO0BcZ6fMsf}
if xafqLlk3kkUe(SXOLrMavuUCe(b'+\xee\x18\x8cg\xe6\x1a\x87\xb1\x1ds\xfe\xa4\nj'), chr(0b1011111 + 0o5) + chr(101) + '\143' + '\157' + chr(0b101100 + 0o70) + '\x65')(chr(117) + chr(0b110000 + 0o104) + chr(102) + chr(787 - 742) + chr(0b1011 + 0o55)) in EEf4r9nUvta_:
SmHQWqnGhb4U = jSKPaHwSAfVv.shape_list(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'+\xee\x18\x8cg\xe6\x1a\x87\xb1\x1ds\xfe\xa4\nj'), chr(100) + chr(0b1100101) + chr(645 - 546) + chr(111) + '\144' + chr(101))(chr(8893 - 8776) + chr(0b111000 + 0o74) + chr(8344 - 8242) + '\055' + '\070')])[ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x31', 8)]
ShZmEKfTkAOZ = IDJ2eXGCBCDu.slice(ShZmEKfTkAOZ, [ehT0Px3KOsy9('\x30' + chr(3458 - 3347) + chr(0b11010 + 0o26), 8), SmHQWqnGhb4U, ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(0b1101111) + '\060', 8), ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(0b1101111) + chr(0b100000 + 0o20), 8)], [-ehT0Px3KOsy9(chr(1643 - 1595) + chr(111) + chr(49), 8), -ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1926 - 1877), 8), -ehT0Px3KOsy9(chr(0b110000) + chr(9735 - 9624) + chr(0b1011 + 0o46), 8), -ehT0Px3KOsy9('\x30' + chr(9424 - 9313) + chr(0b110001), 8)])
return {xafqLlk3kkUe(SXOLrMavuUCe(b'4\xfa\x1e\x88{\xf3\x05'), chr(0b1100100) + '\145' + '\x63' + chr(3664 - 3553) + chr(0b1100100) + '\145')(chr(8657 - 8540) + '\164' + chr(102) + chr(1014 - 969) + chr(0b110101 + 0o3)): ShZmEKfTkAOZ, xafqLlk3kkUe(SXOLrMavuUCe(b'(\xec\x05\x8ak\xf4'), chr(0b1100100) + chr(0b111011 + 0o52) + chr(0b1100011) + chr(111) + chr(0b10 + 0o142) + '\145')('\165' + chr(2265 - 2149) + '\146' + chr(0b10110 + 0o27) + '\x38'): None, xafqLlk3kkUe(SXOLrMavuUCe(b'7\xe0\r\x91z\xf4'), '\x64' + chr(0b1000 + 0o135) + chr(99) + '\x6f' + chr(0b1100100) + chr(0b111000 + 0o55))(chr(0b100100 + 0o121) + chr(116) + '\x66' + '\055' + chr(0b111000)): wF9nmvjsKjYM, xafqLlk3kkUe(SXOLrMavuUCe(b'7\xe0\x19\x8bk\xf4'), '\x64' + chr(101) + chr(0b110011 + 0o60) + chr(111) + '\x64' + chr(101))(chr(117) + chr(0b100111 + 0o115) + chr(102) + chr(0b101101) + '\070'): eJKWkHA7qzlZ}
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.sample
|
def sample(self, features):
"""Run the model and extract samples.
Args:
features: an map of string to `Tensor`.
Returns:
samples: an integer `Tensor`.
logits: a list of `Tensor`s, one per datashard.
losses: a dictionary: {loss-name (string): floating point `Scalar`}.
"""
logits, losses = self(features) # pylint: disable=not-callable
if self._target_modality_is_real:
return logits, logits, losses # Raw numbers returned from real modality.
if self.hparams.sampling_method == "argmax":
samples = tf.argmax(logits, axis=-1)
else:
assert self.hparams.sampling_method == "random"
def multinomial_squeeze(logits, temperature=1.0):
logits_shape = common_layers.shape_list(logits)
reshaped_logits = (
tf.reshape(logits, [-1, logits_shape[-1]]) / temperature)
choices = tf.multinomial(reshaped_logits, 1)
choices = tf.reshape(choices, logits_shape[:-1])
return choices
samples = multinomial_squeeze(logits, self.hparams.sampling_temp)
return samples, logits, losses
|
python
|
def sample(self, features):
"""Run the model and extract samples.
Args:
features: an map of string to `Tensor`.
Returns:
samples: an integer `Tensor`.
logits: a list of `Tensor`s, one per datashard.
losses: a dictionary: {loss-name (string): floating point `Scalar`}.
"""
logits, losses = self(features) # pylint: disable=not-callable
if self._target_modality_is_real:
return logits, logits, losses # Raw numbers returned from real modality.
if self.hparams.sampling_method == "argmax":
samples = tf.argmax(logits, axis=-1)
else:
assert self.hparams.sampling_method == "random"
def multinomial_squeeze(logits, temperature=1.0):
logits_shape = common_layers.shape_list(logits)
reshaped_logits = (
tf.reshape(logits, [-1, logits_shape[-1]]) / temperature)
choices = tf.multinomial(reshaped_logits, 1)
choices = tf.reshape(choices, logits_shape[:-1])
return choices
samples = multinomial_squeeze(logits, self.hparams.sampling_temp)
return samples, logits, losses
|
[
"def",
"sample",
"(",
"self",
",",
"features",
")",
":",
"logits",
",",
"losses",
"=",
"self",
"(",
"features",
")",
"# pylint: disable=not-callable",
"if",
"self",
".",
"_target_modality_is_real",
":",
"return",
"logits",
",",
"logits",
",",
"losses",
"# Raw numbers returned from real modality.",
"if",
"self",
".",
"hparams",
".",
"sampling_method",
"==",
"\"argmax\"",
":",
"samples",
"=",
"tf",
".",
"argmax",
"(",
"logits",
",",
"axis",
"=",
"-",
"1",
")",
"else",
":",
"assert",
"self",
".",
"hparams",
".",
"sampling_method",
"==",
"\"random\"",
"def",
"multinomial_squeeze",
"(",
"logits",
",",
"temperature",
"=",
"1.0",
")",
":",
"logits_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"logits",
")",
"reshaped_logits",
"=",
"(",
"tf",
".",
"reshape",
"(",
"logits",
",",
"[",
"-",
"1",
",",
"logits_shape",
"[",
"-",
"1",
"]",
"]",
")",
"/",
"temperature",
")",
"choices",
"=",
"tf",
".",
"multinomial",
"(",
"reshaped_logits",
",",
"1",
")",
"choices",
"=",
"tf",
".",
"reshape",
"(",
"choices",
",",
"logits_shape",
"[",
":",
"-",
"1",
"]",
")",
"return",
"choices",
"samples",
"=",
"multinomial_squeeze",
"(",
"logits",
",",
"self",
".",
"hparams",
".",
"sampling_temp",
")",
"return",
"samples",
",",
"logits",
",",
"losses"
] |
Run the model and extract samples.
Args:
features: an map of string to `Tensor`.
Returns:
samples: an integer `Tensor`.
logits: a list of `Tensor`s, one per datashard.
losses: a dictionary: {loss-name (string): floating point `Scalar`}.
|
[
"Run",
"the",
"model",
"and",
"extract",
"samples",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L1326-L1355
|
train
|
Run the model and extract samples.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b100000 + 0o22) + '\x36', 0o10), ehT0Px3KOsy9(chr(1589 - 1541) + '\157' + '\x32' + chr(0b10011 + 0o44) + '\x30', 0o10), ehT0Px3KOsy9(chr(1207 - 1159) + chr(111) + chr(753 - 703) + chr(0b1011 + 0o54) + '\x37', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x33' + chr(0b1101 + 0o46) + chr(0b101011 + 0o14), 0o10), ehT0Px3KOsy9(chr(48) + chr(9914 - 9803) + chr(0b110010) + chr(0b11001 + 0o34) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(1934 - 1886) + chr(8742 - 8631) + chr(1246 - 1196) + '\x30' + chr(0b110111), 13909 - 13901), ehT0Px3KOsy9(chr(1430 - 1382) + chr(0b111100 + 0o63) + '\062' + chr(48) + chr(51), ord("\x08")), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(5821 - 5710) + chr(518 - 469) + chr(0b1011 + 0o52) + '\x36', 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\063' + chr(50) + chr(0b110001), 44702 - 44694), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x32' + chr(0b11001 + 0o32) + chr(50), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(51) + '\x35', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(6217 - 6106) + chr(0b0 + 0o63) + '\066' + chr(51), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x33' + '\067' + '\063', 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110001) + chr(0b110011) + '\x35', 55131 - 55123), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110010) + chr(371 - 323), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(2124 - 2074) + chr(0b110011) + '\x34', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(2347 - 2296) + '\x32' + chr(50), 0b1000), ehT0Px3KOsy9('\x30' + chr(11698 - 11587) + chr(575 - 525), 8), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(10859 - 10748) + chr(51) + chr(0b110011) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110010) + chr(49) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(51) + '\x30' + chr(51), 0b1000), ehT0Px3KOsy9(chr(1861 - 1813) + chr(111) + chr(1056 - 1006) + '\063' + chr(0b110110), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b11111 + 0o24) + chr(54) + chr(54), 41747 - 41739), ehT0Px3KOsy9('\060' + '\157' + chr(51) + chr(0b10001 + 0o44), 8), ehT0Px3KOsy9('\060' + chr(111) + chr(1171 - 1120) + chr(0b11100 + 0o33) + chr(50), 53809 - 53801), ehT0Px3KOsy9(chr(0b110000) + chr(0b110 + 0o151) + chr(0b110010) + chr(386 - 337) + '\067', 52639 - 52631), ehT0Px3KOsy9('\x30' + chr(1808 - 1697) + '\062' + chr(0b110000) + '\x32', 12831 - 12823), ehT0Px3KOsy9(chr(48) + '\x6f' + '\062' + chr(0b110111) + '\064', 0b1000), ehT0Px3KOsy9(chr(1550 - 1502) + '\x6f' + chr(0b110010) + chr(952 - 897) + chr(1744 - 1696), 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33' + chr(2388 - 2336), 39539 - 39531), ehT0Px3KOsy9(chr(0b11110 + 0o22) + chr(10376 - 10265) + chr(0b110011) + chr(0b11111 + 0o22), 0o10), ehT0Px3KOsy9(chr(742 - 694) + chr(0b1101111) + chr(0b110 + 0o53) + chr(1899 - 1846) + '\x35', 7165 - 7157), ehT0Px3KOsy9('\x30' + chr(0b1011110 + 0o21) + '\x32' + chr(0b110010) + chr(0b11001 + 0o32), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1171 - 1117) + '\066', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(50) + chr(1330 - 1276) + '\060', 0b1000), ehT0Px3KOsy9(chr(1761 - 1713) + '\x6f' + chr(0b110 + 0o53) + chr(49) + '\x33', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101101 + 0o2) + chr(1610 - 1561) + chr(52) + chr(0b1011 + 0o53), 0b1000), ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(111) + chr(0b110011) + chr(53) + chr(0b11000 + 0o33), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x35' + chr(1413 - 1358), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(8458 - 8347) + chr(0b10101 + 0o40) + chr(0b10 + 0o56), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'8'), chr(0b10100 + 0o120) + '\145' + chr(6881 - 6782) + '\x6f' + chr(0b1100100) + chr(101))('\x75' + chr(0b1010101 + 0o37) + chr(0b1010001 + 0o25) + chr(45) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def aBu4gMMQp6Jg(oVre8I6UXc3b, EEf4r9nUvta_):
(wF9nmvjsKjYM, eJKWkHA7qzlZ) = oVre8I6UXc3b(EEf4r9nUvta_)
if xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b'I\xcc\xde\xad\xcc\xe2\xf71\xbc\xf5y\x92\x974d{\xc3\xbbq\x83\xfa\xa6\xff\x9a'), chr(4100 - 4000) + chr(0b1100101) + chr(6264 - 6165) + '\x6f' + chr(0b1100100) + chr(0b1100101))(chr(117) + chr(116) + '\146' + chr(45) + chr(1510 - 1454))):
return (wF9nmvjsKjYM, wF9nmvjsKjYM, eJKWkHA7qzlZ)
if xafqLlk3kkUe(oVre8I6UXc3b.hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'C\xdc\x8e\x96\xc5\xd6\xb4\x06\xb0\xear\x83'), chr(100) + '\x65' + '\x63' + chr(111) + '\144' + chr(101))(chr(0b1110101) + '\x74' + chr(0b110110 + 0o60) + '\x2d' + chr(56))) == xafqLlk3kkUe(SXOLrMavuUCe(b'w\xca\xd8\xb2\xca\xff'), chr(0b1100100) + chr(101) + chr(0b1100011) + chr(0b1101111) + '\x64' + chr(0b1100101))(chr(117) + chr(9227 - 9111) + '\x66' + chr(0b10101 + 0o30) + chr(1335 - 1279)):
db1_IZvznkcy = IDJ2eXGCBCDu.argmax(wF9nmvjsKjYM, axis=-ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(496 - 447), 0b1000))
else:
assert xafqLlk3kkUe(oVre8I6UXc3b.hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'C\xdc\x8e\x96\xc5\xd6\xb4\x06\xb0\xear\x83'), chr(100) + '\x65' + chr(1439 - 1340) + chr(0b1101111) + chr(0b11000 + 0o114) + chr(101))(chr(0b111 + 0o156) + '\164' + '\x66' + chr(0b101101) + chr(0b111000))) == xafqLlk3kkUe(SXOLrMavuUCe(b'd\xd9\xd1\xbb\xc4\xea'), '\x64' + '\x65' + chr(0b1 + 0o142) + chr(0b101000 + 0o107) + chr(0b100000 + 0o104) + chr(0b1010010 + 0o23))('\x75' + chr(9100 - 8984) + chr(0b1010011 + 0o23) + chr(0b101101) + '\070')
def FznHmx5xAhvg(wF9nmvjsKjYM, uICaXvjWrxGa=1.0):
Isx8k9uq36YR = jSKPaHwSAfVv.shape_list(wF9nmvjsKjYM)
Wpf5bhZujfAz = IDJ2eXGCBCDu.reshape(wF9nmvjsKjYM, [-ehT0Px3KOsy9(chr(0b110000) + chr(0b10 + 0o155) + '\061', 8), Isx8k9uq36YR[-ehT0Px3KOsy9(chr(0b101101 + 0o3) + '\x6f' + '\061', 8)]]) / uICaXvjWrxGa
XPnoMuK4S7nS = IDJ2eXGCBCDu.multinomial(Wpf5bhZujfAz, ehT0Px3KOsy9(chr(830 - 782) + '\x6f' + '\061', 8))
XPnoMuK4S7nS = IDJ2eXGCBCDu.reshape(XPnoMuK4S7nS, Isx8k9uq36YR[:-ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x31', 8)])
return XPnoMuK4S7nS
db1_IZvznkcy = FznHmx5xAhvg(wF9nmvjsKjYM, oVre8I6UXc3b.hparams.Ep30xVZP6Jij)
return (db1_IZvznkcy, wF9nmvjsKjYM, eJKWkHA7qzlZ)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.estimator_model_fn
|
def estimator_model_fn(cls,
hparams,
features,
labels,
mode,
config=None,
params=None,
decode_hparams=None,
use_tpu=False):
"""Model fn for Estimator.
Args:
hparams: HParams, model hyperparameters
features: dict<str name, Tensor feature>
labels: Tensor
mode: tf.estimator.ModeKeys
config: RunConfig, possibly with data_parallelism attribute
params: dict, may include batch_size, use_tpu
decode_hparams: HParams, used when mode == PREDICT.
use_tpu: A bool, whether to build the inference graph for TPU.
Returns:
TPUEstimatorSpec if use tpu else EstimatorSpec
"""
if mode == tf.estimator.ModeKeys.TRAIN:
create_dummy_vars()
hparams = hparams_lib.copy_hparams(hparams)
# Instantiate model
data_parallelism = None
if not use_tpu and config:
data_parallelism = config.data_parallelism
reuse = tf.get_variable_scope().reuse
model = cls(
hparams,
mode,
data_parallelism=data_parallelism,
decode_hparams=decode_hparams,
_reuse=reuse)
# PREDICT mode
if mode == tf.estimator.ModeKeys.PREDICT:
if use_tpu:
inputs = features.get("inputs")
if inputs is None:
inputs = features["targets"]
shape = inputs.get_shape().as_list()
if shape[0] is None:
shape[0] = decode_hparams.batch_size or hparams.batch_size
if shape[1] is None:
shape[1] = hparams.max_input_seq_length or hparams.max_length
inputs.set_shape(shape)
return model.estimator_spec_predict(features, use_tpu=use_tpu)
# TRAIN and EVAL modes
if hparams.eval_run_autoregressive and mode == tf.estimator.ModeKeys.EVAL:
logits, losses_dict = model.eval_autoregressive(features)
else:
logits, losses_dict = model(features) # pylint: disable=not-callable
# Support model-generated labels by overriding features["targets"] with
# logits["self_generated_targets"].
if isinstance(logits, dict) and "self_generated_targets" in logits:
# Overwrite 'features["targets"]' and 'labels'
# by logits["self_generated_targets"].
tf.logging.info("Replacing targets with model-provided targets.")
features["targets"] = labels = logits.pop("self_generated_targets")
assert list(logits.keys()) == ["logits"], (
# See "Returns" in the "top" method docstring for the expected
# "logits" format when targets are generated at training time.
"Expect only key 'logits' when there is 'self_generated_targets'. "
"Found {}".format(logits.keys())
)
# Recover the original logits tensor from the logits dict.
logits = logits["logits"] # Can be a tf.Tensor or a dict.
# Set known shapes
if common_layers.is_xla_compiled():
if isinstance(logits, dict):
for k, v in sorted(six.iteritems(logits)):
if "scalar/" in k:
continue
shape = v.get_shape().as_list()
if shape[0] is None:
shape[0] = params["batch_size"]
if shape[1] is None:
shape[1] = hparams.max_length
v.set_shape(shape)
else:
shape = logits.get_shape().as_list()
if shape[0] is None:
shape[0] = params["batch_size"]
if shape[1] is None:
shape[1] = hparams.max_length
logits.set_shape(shape)
assert "training" in losses_dict
# Attack mode
if mode == "attack":
return logits
# Summarize losses
model._summarize_losses(losses_dict) # pylint: disable=protected-access
# Accumulate losses
loss = sum(losses_dict[key] for key in sorted(losses_dict.keys()))
# EVAL mode
if mode == tf.estimator.ModeKeys.EVAL:
return model.estimator_spec_eval(features, logits, labels, loss,
losses_dict)
# TRAIN mode
assert mode == tf.estimator.ModeKeys.TRAIN
num_async_replicas = 1
if config and not use_tpu:
num_async_replicas = config.t2t_device_info["num_async_replicas"]
return model.estimator_spec_train(
loss, num_async_replicas=num_async_replicas, use_tpu=use_tpu)
|
python
|
def estimator_model_fn(cls,
hparams,
features,
labels,
mode,
config=None,
params=None,
decode_hparams=None,
use_tpu=False):
"""Model fn for Estimator.
Args:
hparams: HParams, model hyperparameters
features: dict<str name, Tensor feature>
labels: Tensor
mode: tf.estimator.ModeKeys
config: RunConfig, possibly with data_parallelism attribute
params: dict, may include batch_size, use_tpu
decode_hparams: HParams, used when mode == PREDICT.
use_tpu: A bool, whether to build the inference graph for TPU.
Returns:
TPUEstimatorSpec if use tpu else EstimatorSpec
"""
if mode == tf.estimator.ModeKeys.TRAIN:
create_dummy_vars()
hparams = hparams_lib.copy_hparams(hparams)
# Instantiate model
data_parallelism = None
if not use_tpu and config:
data_parallelism = config.data_parallelism
reuse = tf.get_variable_scope().reuse
model = cls(
hparams,
mode,
data_parallelism=data_parallelism,
decode_hparams=decode_hparams,
_reuse=reuse)
# PREDICT mode
if mode == tf.estimator.ModeKeys.PREDICT:
if use_tpu:
inputs = features.get("inputs")
if inputs is None:
inputs = features["targets"]
shape = inputs.get_shape().as_list()
if shape[0] is None:
shape[0] = decode_hparams.batch_size or hparams.batch_size
if shape[1] is None:
shape[1] = hparams.max_input_seq_length or hparams.max_length
inputs.set_shape(shape)
return model.estimator_spec_predict(features, use_tpu=use_tpu)
# TRAIN and EVAL modes
if hparams.eval_run_autoregressive and mode == tf.estimator.ModeKeys.EVAL:
logits, losses_dict = model.eval_autoregressive(features)
else:
logits, losses_dict = model(features) # pylint: disable=not-callable
# Support model-generated labels by overriding features["targets"] with
# logits["self_generated_targets"].
if isinstance(logits, dict) and "self_generated_targets" in logits:
# Overwrite 'features["targets"]' and 'labels'
# by logits["self_generated_targets"].
tf.logging.info("Replacing targets with model-provided targets.")
features["targets"] = labels = logits.pop("self_generated_targets")
assert list(logits.keys()) == ["logits"], (
# See "Returns" in the "top" method docstring for the expected
# "logits" format when targets are generated at training time.
"Expect only key 'logits' when there is 'self_generated_targets'. "
"Found {}".format(logits.keys())
)
# Recover the original logits tensor from the logits dict.
logits = logits["logits"] # Can be a tf.Tensor or a dict.
# Set known shapes
if common_layers.is_xla_compiled():
if isinstance(logits, dict):
for k, v in sorted(six.iteritems(logits)):
if "scalar/" in k:
continue
shape = v.get_shape().as_list()
if shape[0] is None:
shape[0] = params["batch_size"]
if shape[1] is None:
shape[1] = hparams.max_length
v.set_shape(shape)
else:
shape = logits.get_shape().as_list()
if shape[0] is None:
shape[0] = params["batch_size"]
if shape[1] is None:
shape[1] = hparams.max_length
logits.set_shape(shape)
assert "training" in losses_dict
# Attack mode
if mode == "attack":
return logits
# Summarize losses
model._summarize_losses(losses_dict) # pylint: disable=protected-access
# Accumulate losses
loss = sum(losses_dict[key] for key in sorted(losses_dict.keys()))
# EVAL mode
if mode == tf.estimator.ModeKeys.EVAL:
return model.estimator_spec_eval(features, logits, labels, loss,
losses_dict)
# TRAIN mode
assert mode == tf.estimator.ModeKeys.TRAIN
num_async_replicas = 1
if config and not use_tpu:
num_async_replicas = config.t2t_device_info["num_async_replicas"]
return model.estimator_spec_train(
loss, num_async_replicas=num_async_replicas, use_tpu=use_tpu)
|
[
"def",
"estimator_model_fn",
"(",
"cls",
",",
"hparams",
",",
"features",
",",
"labels",
",",
"mode",
",",
"config",
"=",
"None",
",",
"params",
"=",
"None",
",",
"decode_hparams",
"=",
"None",
",",
"use_tpu",
"=",
"False",
")",
":",
"if",
"mode",
"==",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"TRAIN",
":",
"create_dummy_vars",
"(",
")",
"hparams",
"=",
"hparams_lib",
".",
"copy_hparams",
"(",
"hparams",
")",
"# Instantiate model",
"data_parallelism",
"=",
"None",
"if",
"not",
"use_tpu",
"and",
"config",
":",
"data_parallelism",
"=",
"config",
".",
"data_parallelism",
"reuse",
"=",
"tf",
".",
"get_variable_scope",
"(",
")",
".",
"reuse",
"model",
"=",
"cls",
"(",
"hparams",
",",
"mode",
",",
"data_parallelism",
"=",
"data_parallelism",
",",
"decode_hparams",
"=",
"decode_hparams",
",",
"_reuse",
"=",
"reuse",
")",
"# PREDICT mode",
"if",
"mode",
"==",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"PREDICT",
":",
"if",
"use_tpu",
":",
"inputs",
"=",
"features",
".",
"get",
"(",
"\"inputs\"",
")",
"if",
"inputs",
"is",
"None",
":",
"inputs",
"=",
"features",
"[",
"\"targets\"",
"]",
"shape",
"=",
"inputs",
".",
"get_shape",
"(",
")",
".",
"as_list",
"(",
")",
"if",
"shape",
"[",
"0",
"]",
"is",
"None",
":",
"shape",
"[",
"0",
"]",
"=",
"decode_hparams",
".",
"batch_size",
"or",
"hparams",
".",
"batch_size",
"if",
"shape",
"[",
"1",
"]",
"is",
"None",
":",
"shape",
"[",
"1",
"]",
"=",
"hparams",
".",
"max_input_seq_length",
"or",
"hparams",
".",
"max_length",
"inputs",
".",
"set_shape",
"(",
"shape",
")",
"return",
"model",
".",
"estimator_spec_predict",
"(",
"features",
",",
"use_tpu",
"=",
"use_tpu",
")",
"# TRAIN and EVAL modes",
"if",
"hparams",
".",
"eval_run_autoregressive",
"and",
"mode",
"==",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"EVAL",
":",
"logits",
",",
"losses_dict",
"=",
"model",
".",
"eval_autoregressive",
"(",
"features",
")",
"else",
":",
"logits",
",",
"losses_dict",
"=",
"model",
"(",
"features",
")",
"# pylint: disable=not-callable",
"# Support model-generated labels by overriding features[\"targets\"] with",
"# logits[\"self_generated_targets\"].",
"if",
"isinstance",
"(",
"logits",
",",
"dict",
")",
"and",
"\"self_generated_targets\"",
"in",
"logits",
":",
"# Overwrite 'features[\"targets\"]' and 'labels'",
"# by logits[\"self_generated_targets\"].",
"tf",
".",
"logging",
".",
"info",
"(",
"\"Replacing targets with model-provided targets.\"",
")",
"features",
"[",
"\"targets\"",
"]",
"=",
"labels",
"=",
"logits",
".",
"pop",
"(",
"\"self_generated_targets\"",
")",
"assert",
"list",
"(",
"logits",
".",
"keys",
"(",
")",
")",
"==",
"[",
"\"logits\"",
"]",
",",
"(",
"# See \"Returns\" in the \"top\" method docstring for the expected",
"# \"logits\" format when targets are generated at training time.",
"\"Expect only key 'logits' when there is 'self_generated_targets'. \"",
"\"Found {}\"",
".",
"format",
"(",
"logits",
".",
"keys",
"(",
")",
")",
")",
"# Recover the original logits tensor from the logits dict.",
"logits",
"=",
"logits",
"[",
"\"logits\"",
"]",
"# Can be a tf.Tensor or a dict.",
"# Set known shapes",
"if",
"common_layers",
".",
"is_xla_compiled",
"(",
")",
":",
"if",
"isinstance",
"(",
"logits",
",",
"dict",
")",
":",
"for",
"k",
",",
"v",
"in",
"sorted",
"(",
"six",
".",
"iteritems",
"(",
"logits",
")",
")",
":",
"if",
"\"scalar/\"",
"in",
"k",
":",
"continue",
"shape",
"=",
"v",
".",
"get_shape",
"(",
")",
".",
"as_list",
"(",
")",
"if",
"shape",
"[",
"0",
"]",
"is",
"None",
":",
"shape",
"[",
"0",
"]",
"=",
"params",
"[",
"\"batch_size\"",
"]",
"if",
"shape",
"[",
"1",
"]",
"is",
"None",
":",
"shape",
"[",
"1",
"]",
"=",
"hparams",
".",
"max_length",
"v",
".",
"set_shape",
"(",
"shape",
")",
"else",
":",
"shape",
"=",
"logits",
".",
"get_shape",
"(",
")",
".",
"as_list",
"(",
")",
"if",
"shape",
"[",
"0",
"]",
"is",
"None",
":",
"shape",
"[",
"0",
"]",
"=",
"params",
"[",
"\"batch_size\"",
"]",
"if",
"shape",
"[",
"1",
"]",
"is",
"None",
":",
"shape",
"[",
"1",
"]",
"=",
"hparams",
".",
"max_length",
"logits",
".",
"set_shape",
"(",
"shape",
")",
"assert",
"\"training\"",
"in",
"losses_dict",
"# Attack mode",
"if",
"mode",
"==",
"\"attack\"",
":",
"return",
"logits",
"# Summarize losses",
"model",
".",
"_summarize_losses",
"(",
"losses_dict",
")",
"# pylint: disable=protected-access",
"# Accumulate losses",
"loss",
"=",
"sum",
"(",
"losses_dict",
"[",
"key",
"]",
"for",
"key",
"in",
"sorted",
"(",
"losses_dict",
".",
"keys",
"(",
")",
")",
")",
"# EVAL mode",
"if",
"mode",
"==",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"EVAL",
":",
"return",
"model",
".",
"estimator_spec_eval",
"(",
"features",
",",
"logits",
",",
"labels",
",",
"loss",
",",
"losses_dict",
")",
"# TRAIN mode",
"assert",
"mode",
"==",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"TRAIN",
"num_async_replicas",
"=",
"1",
"if",
"config",
"and",
"not",
"use_tpu",
":",
"num_async_replicas",
"=",
"config",
".",
"t2t_device_info",
"[",
"\"num_async_replicas\"",
"]",
"return",
"model",
".",
"estimator_spec_train",
"(",
"loss",
",",
"num_async_replicas",
"=",
"num_async_replicas",
",",
"use_tpu",
"=",
"use_tpu",
")"
] |
Model fn for Estimator.
Args:
hparams: HParams, model hyperparameters
features: dict<str name, Tensor feature>
labels: Tensor
mode: tf.estimator.ModeKeys
config: RunConfig, possibly with data_parallelism attribute
params: dict, may include batch_size, use_tpu
decode_hparams: HParams, used when mode == PREDICT.
use_tpu: A bool, whether to build the inference graph for TPU.
Returns:
TPUEstimatorSpec if use tpu else EstimatorSpec
|
[
"Model",
"fn",
"for",
"Estimator",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L1418-L1538
|
train
|
Model fn for training and evaluation.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b1010 + 0o46) + '\157' + '\x32', 49547 - 49539), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(52) + '\x37', 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(1030 - 981) + chr(0b110000) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x31' + '\x34' + chr(2230 - 2177), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(9049 - 8938) + chr(49) + chr(0b1110 + 0o47), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + '\067' + chr(51), 23406 - 23398), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(6434 - 6323) + chr(50) + chr(0b110100) + chr(0b110111), 34386 - 34378), ehT0Px3KOsy9(chr(1351 - 1303) + chr(4778 - 4667) + chr(0b110001) + chr(0b110001) + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110011) + chr(52), 32831 - 32823), ehT0Px3KOsy9('\060' + chr(0b1011 + 0o144) + '\x33' + chr(1145 - 1091) + chr(1917 - 1863), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(12139 - 12028) + '\062' + chr(0b110000) + '\x30', 51212 - 51204), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(392 - 342) + '\064' + chr(0b110000 + 0o6), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\062' + chr(1155 - 1100) + '\061', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010) + chr(52) + '\x37', 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110111) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(50) + chr(0b110010) + chr(0b11 + 0o60), 25413 - 25405), ehT0Px3KOsy9('\060' + chr(0b110110 + 0o71) + '\062' + chr(0b110110) + '\067', ord("\x08")), ehT0Px3KOsy9('\060' + chr(4266 - 4155) + chr(0b10100 + 0o37) + chr(1805 - 1756) + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(1053 - 1005) + chr(111) + chr(415 - 360) + chr(0b11100 + 0o33), 8), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b101010 + 0o7) + '\060' + chr(985 - 931), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(5169 - 5058) + chr(310 - 261) + '\x30' + '\063', 0o10), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b1101111) + chr(776 - 727) + chr(48), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b100 + 0o153) + '\063' + chr(0b110101) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\063' + chr(0b10110 + 0o32) + chr(0b110000), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(851 - 801) + chr(50) + chr(0b110011), 8), ehT0Px3KOsy9(chr(1659 - 1611) + chr(111) + chr(0b100100 + 0o15) + '\x33' + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(0b110100) + '\x33', 0o10), ehT0Px3KOsy9(chr(2198 - 2150) + chr(111) + chr(0b10010 + 0o40) + '\061' + '\060', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b10101 + 0o132) + chr(0b101101 + 0o5) + chr(0b11000 + 0o35) + '\x30', 46080 - 46072), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(1176 - 1065) + '\x32' + '\063' + chr(2241 - 2191), 0b1000), ehT0Px3KOsy9(chr(782 - 734) + chr(111) + chr(0b110111) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b110001 + 0o76) + chr(447 - 398) + chr(2862 - 2807) + chr(1721 - 1673), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(2407 - 2352) + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111 + 0o0) + chr(0b110000 + 0o1) + '\060' + '\x30', 40033 - 40025), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b11111 + 0o22) + chr(0b110111) + chr(55), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001) + '\064' + chr(229 - 178), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(11197 - 11086) + '\x31' + '\x30' + chr(0b110110), 8), ehT0Px3KOsy9(chr(0b11010 + 0o26) + chr(154 - 43) + chr(49) + chr(0b10 + 0o64) + '\066', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110010) + '\x33' + chr(1735 - 1684), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x35' + chr(48), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b','), chr(0b1100100) + chr(0b1100101) + chr(0b1100011) + '\x6f' + chr(100) + chr(5866 - 5765))('\x75' + '\x74' + '\146' + chr(0b11010 + 0o23) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def wC9130fUBCqT(NSstowUUZlxS, n4ljua2gi1Pr, EEf4r9nUvta_, uXMK81tmdpTM, holLFgwB7vsP, jAj7S20Ct06o=None, nEbJZ4wfte2w=None, LrQSWg3uwmK8=None, L4eE7kczIJwa=ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(1822 - 1774), ord("\x08"))):
if holLFgwB7vsP == xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'Vu\x03,C'), chr(2814 - 2714) + '\145' + '\x63' + '\157' + chr(0b1100100) + chr(0b101001 + 0o74))(chr(117) + '\164' + '\146' + chr(790 - 745) + chr(1817 - 1761))):
gT2F0ANO6vb6()
n4ljua2gi1Pr = BMqTy4F2E1fh.copy_hparams(n4ljua2gi1Pr)
iP4ZSqFqAOG_ = None
if not L4eE7kczIJwa and jAj7S20Ct06o:
iP4ZSqFqAOG_ = jAj7S20Ct06o.iP4ZSqFqAOG_
pmC5wdSFgdFj = IDJ2eXGCBCDu.get_variable_scope().reuse
FK0vqzZ5gPN6 = NSstowUUZlxS(n4ljua2gi1Pr, holLFgwB7vsP, data_parallelism=iP4ZSqFqAOG_, decode_hparams=LrQSWg3uwmK8, _reuse=pmC5wdSFgdFj)
if holLFgwB7vsP == xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'Ru\x07!Dx\xf3'), chr(0b1011111 + 0o5) + '\x65' + '\143' + chr(0b0 + 0o157) + '\144' + chr(0b1100101))(chr(0b1001100 + 0o51) + chr(0b1101000 + 0o14) + chr(0b1000011 + 0o43) + '\055' + chr(56))):
if L4eE7kczIJwa:
vXoupepMtCXU = EEf4r9nUvta_.get(xafqLlk3kkUe(SXOLrMavuUCe(b'kI2\x10yH'), chr(0b1010010 + 0o22) + '\145' + '\143' + chr(0b1101111) + '\144' + chr(101))(chr(6641 - 6524) + chr(0b1110100) + chr(0b11010 + 0o114) + chr(757 - 712) + chr(0b100011 + 0o25)))
if vXoupepMtCXU is None:
vXoupepMtCXU = EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'vF0\x02hO\xd4'), chr(0b110110 + 0o56) + chr(606 - 505) + chr(835 - 736) + chr(0b1101111) + '\144' + '\145')(chr(0b111000 + 0o75) + chr(0b101011 + 0o111) + '\146' + '\x2d' + chr(0b1010 + 0o56))]
nauYfLglTpcb = vXoupepMtCXU.get_shape().as_list()
if nauYfLglTpcb[ehT0Px3KOsy9('\x30' + '\157' + chr(0b10010 + 0o36), 8)] is None:
nauYfLglTpcb[ehT0Px3KOsy9('\x30' + chr(111) + '\060', 8)] = LrQSWg3uwmK8.ix9dZyeAmUxY or n4ljua2gi1Pr.ix9dZyeAmUxY
if nauYfLglTpcb[ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b100110 + 0o13), ord("\x08"))] is None:
nauYfLglTpcb[ehT0Px3KOsy9(chr(0b100000 + 0o20) + chr(0b1101111) + '\061', 8)] = n4ljua2gi1Pr.xa50HGLsAIaS or n4ljua2gi1Pr._o7pVXAdOCRy
xafqLlk3kkUe(vXoupepMtCXU, xafqLlk3kkUe(SXOLrMavuUCe(b'qB6:~S\xc6=y'), chr(3349 - 3249) + chr(2274 - 2173) + '\143' + '\x6f' + chr(0b1100100) + chr(101))(chr(8790 - 8673) + chr(0b100000 + 0o124) + chr(0b1100110) + chr(0b101101) + chr(0b10 + 0o66)))(nauYfLglTpcb)
return xafqLlk3kkUe(FK0vqzZ5gPN6, xafqLlk3kkUe(SXOLrMavuUCe(b'gT6\x0c`Z\xd3"n\xa9\xf3\xd6\xdc\xc4`\xe3\xc5\xaf[/ |'), chr(0b1100100) + chr(0b100001 + 0o104) + chr(0b110010 + 0o61) + chr(9970 - 9859) + '\x64' + chr(0b1100101))('\x75' + chr(11701 - 11585) + '\x66' + chr(45) + chr(0b100000 + 0o30)))(EEf4r9nUvta_, use_tpu=L4eE7kczIJwa)
if xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'gQ#\tRI\xd2#C\x97\xf5\xd2\xd6\xd5Z\xf4\xc5\xafL5*~\x8a'), chr(0b1100100) + chr(463 - 362) + '\143' + '\x6f' + chr(0b1100100) + '\145')(chr(0b1110101) + chr(116) + chr(2832 - 2730) + chr(1817 - 1772) + chr(0b11111 + 0o31))) and holLFgwB7vsP == xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'Gq\x03)'), chr(0b1100100) + chr(0b1100101) + '\143' + '\x6f' + chr(0b1100100) + chr(0b10111 + 0o116))(chr(117) + '\164' + chr(7433 - 7331) + chr(713 - 668) + chr(0b110000 + 0o10))):
(wF9nmvjsKjYM, hovGDa4hwxi3) = FK0vqzZ5gPN6.eval_autoregressive(EEf4r9nUvta_)
else:
(wF9nmvjsKjYM, hovGDa4hwxi3) = FK0vqzZ5gPN6(EEf4r9nUvta_)
if PlSM16l2KDPD(wF9nmvjsKjYM, wLqBDw8l0eIm) and xafqLlk3kkUe(SXOLrMavuUCe(b'qB.\x03R\\\xc2#y\x84\xe1\xd2\xdc\xc3`\xe7\xd6\xb8X#7{'), chr(0b11110 + 0o106) + chr(0b1010010 + 0o23) + chr(1492 - 1393) + '\x6f' + chr(0b1100100) + chr(0b1001101 + 0o30))(chr(12869 - 12752) + chr(116) + '\146' + chr(0b1011 + 0o42) + chr(755 - 699)) in wF9nmvjsKjYM:
xafqLlk3kkUe(IDJ2eXGCBCDu.logging, xafqLlk3kkUe(SXOLrMavuUCe(b'Q\x10\n\x1dxX\xc0zv\x9a\xda\xcd'), chr(0b1100100) + chr(0b1100101) + '\143' + chr(0b1101111) + '\144' + '\x65')('\x75' + chr(843 - 727) + chr(0b1011010 + 0o14) + chr(0b1101 + 0o40) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b"PB2\tlX\xce#{\xd6\xf4\xc7\xcb\xc0Z\xe7\xc4\xeaH/7`\xcf\xad\xfc9\x86\xe4\x85t\xe8Q\x0eP$\x1a\x07f\x02\xfap@'\x11~\x15"), '\144' + chr(101) + chr(0b11001 + 0o112) + '\x6f' + '\144' + '\x65')(chr(7230 - 7113) + chr(4115 - 3999) + '\146' + chr(0b1111 + 0o36) + chr(2521 - 2465)))
EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'vF0\x02hO\xd4'), chr(0b1100100) + '\x65' + '\x63' + chr(0b1101111) + chr(0b1001111 + 0o25) + chr(2494 - 2393))(chr(117) + '\x74' + chr(9561 - 9459) + chr(0b1110 + 0o37) + chr(56))] = uXMK81tmdpTM = wF9nmvjsKjYM.pop(xafqLlk3kkUe(SXOLrMavuUCe(b'qB.\x03R\\\xc2#y\x84\xe1\xd2\xdc\xc3`\xe7\xd6\xb8X#7{'), chr(100) + chr(101) + '\143' + '\157' + chr(100) + '\x65')(chr(0b110 + 0o157) + chr(0b11011 + 0o131) + '\146' + chr(0b101101) + chr(2913 - 2857)))
assert YyaZ4tpXu4lf(xafqLlk3kkUe(wF9nmvjsKjYM, xafqLlk3kkUe(SXOLrMavuUCe(b'iB;\x16'), '\x64' + '\x65' + '\x63' + chr(0b111101 + 0o62) + chr(0b1100100) + chr(0b1000001 + 0o44))(chr(117) + chr(0b101 + 0o157) + chr(0b1100110) + chr(1565 - 1520) + chr(0b10111 + 0o41)))()) == [xafqLlk3kkUe(SXOLrMavuUCe(b'nH%\x0cyH'), '\144' + chr(101) + '\143' + chr(0b1010010 + 0o35) + '\x64' + chr(101))(chr(0b10100 + 0o141) + chr(7536 - 7420) + chr(0b1100110) + '\055' + '\070')], xafqLlk3kkUe(xafqLlk3kkUe(SXOLrMavuUCe(b'G_2\x00nO\x87"r\x9a\xf9\x86\xd2\xc2F\xb3\x90\xa6P!*|\x9c\xe7\xb3*\x8b\xed\xc6$\xeeV\x1dK%_\n5V\xbcqB.\x03R\\\xc2#y\x84\xe1\xd2\xdc\xc3`\xe7\xd6\xb8X#7{\xc8\xee\xb3\x1b\x8c\xfd\xc6`\xbaE\x05'), chr(3699 - 3599) + '\x65' + chr(0b1100011) + '\157' + chr(8105 - 8005) + '\x65')('\x75' + chr(116) + chr(102) + chr(45) + chr(0b101000 + 0o20)), xafqLlk3kkUe(SXOLrMavuUCe(b'T\x130\nEZ\xf4~L\x86\xe5\xcc'), chr(0b10101 + 0o117) + '\145' + '\143' + chr(11992 - 11881) + chr(100) + '\145')(chr(13466 - 13349) + '\x74' + chr(0b1100110) + chr(1996 - 1951) + chr(2783 - 2727)))(xafqLlk3kkUe(wF9nmvjsKjYM, xafqLlk3kkUe(SXOLrMavuUCe(b'iB;\x16'), chr(0b1011100 + 0o10) + chr(0b100100 + 0o101) + '\x63' + chr(111) + chr(0b111110 + 0o46) + chr(101))('\165' + chr(116) + chr(0b1000000 + 0o46) + chr(0b100 + 0o51) + chr(160 - 104)))())
wF9nmvjsKjYM = wF9nmvjsKjYM[xafqLlk3kkUe(SXOLrMavuUCe(b'nH%\x0cyH'), chr(100) + chr(0b1100101) + chr(99) + '\x6f' + chr(100) + '\x65')(chr(0b1110101) + chr(0b1110100) + chr(102) + chr(0b101101) + chr(56))]
if xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'kT\x1d\x1daZ\xf8.s\x9b\xf0\xcf\xd5\xc2['), chr(0b1100100) + chr(101) + chr(946 - 847) + chr(0b1101111) + '\x64' + '\x65')(chr(0b1101000 + 0o15) + chr(656 - 540) + '\x66' + chr(0b101101) + chr(2792 - 2736)))():
if PlSM16l2KDPD(wF9nmvjsKjYM, wLqBDw8l0eIm):
for (OolUPRJhRaJd, cMbll0QYhULo) in vUlqIvNSaRMa(xafqLlk3kkUe(sYby0kpfssd4, xafqLlk3kkUe(SXOLrMavuUCe(b"kS'\x17dO\xc2 o"), '\144' + chr(8907 - 8806) + '\x63' + chr(0b1101111) + chr(0b110101 + 0o57) + chr(101))(chr(3905 - 3788) + chr(13083 - 12967) + chr(0b1100110) + chr(0b101101) + '\070'))(wF9nmvjsKjYM)):
if xafqLlk3kkUe(SXOLrMavuUCe(b'qD#\tlI\x88'), chr(100) + chr(0b1100101) + '\x63' + '\x6f' + '\x64' + '\x65')(chr(117) + chr(9075 - 8959) + '\x66' + chr(0b101101) + chr(56)) in OolUPRJhRaJd:
continue
nauYfLglTpcb = cMbll0QYhULo.get_shape().as_list()
if nauYfLglTpcb[ehT0Px3KOsy9(chr(1923 - 1875) + chr(589 - 478) + chr(1367 - 1319), 8)] is None:
nauYfLglTpcb[ehT0Px3KOsy9('\060' + chr(0b111001 + 0o66) + chr(0b110000), 8)] = nEbJZ4wfte2w[xafqLlk3kkUe(SXOLrMavuUCe(b'`F6\x06ed\xd4$f\x93'), chr(100) + chr(0b1011111 + 0o6) + chr(1025 - 926) + '\x6f' + chr(5063 - 4963) + chr(0b1100101))(chr(117) + '\x74' + chr(0b1100010 + 0o4) + chr(746 - 701) + chr(0b101000 + 0o20))]
if nauYfLglTpcb[ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001), 8)] is None:
nauYfLglTpcb[ehT0Px3KOsy9(chr(136 - 88) + chr(0b1000011 + 0o54) + chr(49), 8)] = n4ljua2gi1Pr._o7pVXAdOCRy
xafqLlk3kkUe(cMbll0QYhULo, xafqLlk3kkUe(SXOLrMavuUCe(b'qB6:~S\xc6=y'), chr(990 - 890) + '\145' + chr(2449 - 2350) + chr(1976 - 1865) + chr(100) + '\x65')('\x75' + chr(0b1101010 + 0o12) + chr(0b1100110) + '\x2d' + '\070'))(nauYfLglTpcb)
else:
nauYfLglTpcb = wF9nmvjsKjYM.get_shape().as_list()
if nauYfLglTpcb[ehT0Px3KOsy9('\060' + '\157' + chr(0b100000 + 0o20), 8)] is None:
nauYfLglTpcb[ehT0Px3KOsy9('\x30' + '\x6f' + chr(48), 8)] = nEbJZ4wfte2w[xafqLlk3kkUe(SXOLrMavuUCe(b'`F6\x06ed\xd4$f\x93'), '\x64' + chr(101) + chr(99) + '\157' + chr(9415 - 9315) + '\145')(chr(0b1110101) + chr(10724 - 10608) + '\x66' + chr(0b100100 + 0o11) + '\070')]
if nauYfLglTpcb[ehT0Px3KOsy9('\x30' + chr(111) + chr(0b11110 + 0o23), 8)] is None:
nauYfLglTpcb[ehT0Px3KOsy9(chr(0b10 + 0o56) + '\157' + chr(0b11101 + 0o24), 8)] = n4ljua2gi1Pr._o7pVXAdOCRy
xafqLlk3kkUe(wF9nmvjsKjYM, xafqLlk3kkUe(SXOLrMavuUCe(b'qB6:~S\xc6=y'), chr(0b1010100 + 0o20) + '\x65' + chr(0b110100 + 0o57) + chr(0b1101111) + chr(6717 - 6617) + chr(0b1100101))(chr(0b1110101) + '\x74' + chr(8069 - 7967) + chr(45) + chr(645 - 589)))(nauYfLglTpcb)
assert xafqLlk3kkUe(SXOLrMavuUCe(b'vU#\x0ccR\xc9*'), chr(0b11010 + 0o112) + '\x65' + chr(99) + '\157' + chr(100) + chr(0b1010 + 0o133))(chr(0b110010 + 0o103) + chr(0b1110100) + chr(0b100 + 0o142) + '\055' + chr(56)) in hovGDa4hwxi3
if holLFgwB7vsP == xafqLlk3kkUe(SXOLrMavuUCe(b'cS6\x04nP'), '\144' + chr(0b1100101) + '\143' + chr(0b100111 + 0o110) + chr(0b1100 + 0o130) + chr(0b1100101))(chr(0b1110101) + chr(0b110010 + 0o102) + chr(9116 - 9014) + '\055' + '\x38'):
return wF9nmvjsKjYM
xafqLlk3kkUe(FK0vqzZ5gPN6, xafqLlk3kkUe(SXOLrMavuUCe(b']T7\x08`Z\xd5$f\x93\xdf\xca\xd6\xd4L\xf6\xc4'), chr(100) + chr(837 - 736) + chr(99) + chr(9314 - 9203) + chr(0b1100100) + chr(101))(chr(0b1100100 + 0o21) + chr(5619 - 5503) + '\146' + chr(0b10001 + 0o34) + chr(0b111000)))(hovGDa4hwxi3)
YpO0BcZ6fMsf = xkxBmo49x2An((hovGDa4hwxi3[K3J4ZwSlE0sT] for K3J4ZwSlE0sT in vUlqIvNSaRMa(hovGDa4hwxi3.keys())))
if holLFgwB7vsP == xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'Gq\x03)'), '\144' + '\x65' + chr(3705 - 3606) + '\157' + '\144' + chr(0b1100101))(chr(0b110 + 0o157) + chr(0b1010101 + 0o37) + chr(102) + '\055' + '\070')):
return xafqLlk3kkUe(FK0vqzZ5gPN6, xafqLlk3kkUe(SXOLrMavuUCe(b'gT6\x0c`Z\xd3"n\xa9\xf3\xd6\xdc\xc4`\xf6\xc1\xabS'), chr(0b1100100) + '\145' + chr(0b1100011) + chr(6538 - 6427) + chr(100) + chr(0b1100101))(chr(0b1000001 + 0o64) + chr(116) + chr(0b1000 + 0o136) + chr(45) + '\070'))(EEf4r9nUvta_, wF9nmvjsKjYM, uXMK81tmdpTM, YpO0BcZ6fMsf, hovGDa4hwxi3)
assert holLFgwB7vsP == xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'Vu\x03,C'), '\144' + chr(0b1011000 + 0o15) + chr(7629 - 7530) + chr(111) + chr(100) + chr(7485 - 7384))(chr(0b1100 + 0o151) + chr(116) + chr(9388 - 9286) + chr(0b11101 + 0o20) + '\070'))
FsJHiE9x_3jA = ehT0Px3KOsy9('\x30' + '\x6f' + '\061', 8)
if jAj7S20Ct06o and (not L4eE7kczIJwa):
FsJHiE9x_3jA = jAj7S20Ct06o.VnXdb2ihoHeL[xafqLlk3kkUe(SXOLrMavuUCe(b'lR/:lH\xde#\x7f\xa9\xf2\xc3\xc9\xcbV\xf0\xd6\xb9'), chr(4585 - 4485) + '\145' + chr(0b100000 + 0o103) + '\157' + '\x64' + '\x65')(chr(0b10110 + 0o137) + chr(0b1110100) + '\146' + '\x2d' + chr(0b101010 + 0o16))]
return xafqLlk3kkUe(FK0vqzZ5gPN6, xafqLlk3kkUe(SXOLrMavuUCe(b'gT6\x0c`Z\xd3"n\xa9\xf3\xd6\xdc\xc4`\xe7\xc5\xabV('), '\144' + chr(0b1100101) + chr(0b1100011) + chr(0b1010101 + 0o32) + chr(1999 - 1899) + '\x65')(chr(117) + chr(0b1000101 + 0o57) + chr(0b1100110) + chr(0b1 + 0o54) + chr(56)))(YpO0BcZ6fMsf, num_async_replicas=FsJHiE9x_3jA, use_tpu=L4eE7kczIJwa)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.estimator_spec_train
|
def estimator_spec_train(self, loss, num_async_replicas=1, use_tpu=False):
"""Constructs `tf.estimator.EstimatorSpec` for TRAIN (training) mode."""
train_op = self.optimize(loss, num_async_replicas=num_async_replicas,
use_tpu=use_tpu)
if use_tpu:
if self._hparams.warm_start_from:
def scaffold_fn():
self.initialize_from_ckpt(self._hparams.warm_start_from)
return tf.train.Scaffold()
else:
scaffold_fn = None
# Note: important to call this before remove_summaries()
if self.hparams.tpu_enable_host_call:
host_call = self.create_train_host_call()
else:
host_call = None
remove_summaries()
return tf.contrib.tpu.TPUEstimatorSpec(
tf.estimator.ModeKeys.TRAIN,
loss=loss,
train_op=train_op,
host_call=host_call,
scaffold_fn=scaffold_fn)
else:
if self._hparams.warm_start_from:
self.initialize_from_ckpt(self._hparams.warm_start_from)
# When loading weights from a pre-trained model, you want to be able to
# load separate weights into the encoder and decoder.
if self._hparams.warm_start_from_second:
self.initialize_from_ckpt(self._hparams.warm_start_from_second)
return tf.estimator.EstimatorSpec(
tf.estimator.ModeKeys.TRAIN, loss=loss, train_op=train_op)
|
python
|
def estimator_spec_train(self, loss, num_async_replicas=1, use_tpu=False):
"""Constructs `tf.estimator.EstimatorSpec` for TRAIN (training) mode."""
train_op = self.optimize(loss, num_async_replicas=num_async_replicas,
use_tpu=use_tpu)
if use_tpu:
if self._hparams.warm_start_from:
def scaffold_fn():
self.initialize_from_ckpt(self._hparams.warm_start_from)
return tf.train.Scaffold()
else:
scaffold_fn = None
# Note: important to call this before remove_summaries()
if self.hparams.tpu_enable_host_call:
host_call = self.create_train_host_call()
else:
host_call = None
remove_summaries()
return tf.contrib.tpu.TPUEstimatorSpec(
tf.estimator.ModeKeys.TRAIN,
loss=loss,
train_op=train_op,
host_call=host_call,
scaffold_fn=scaffold_fn)
else:
if self._hparams.warm_start_from:
self.initialize_from_ckpt(self._hparams.warm_start_from)
# When loading weights from a pre-trained model, you want to be able to
# load separate weights into the encoder and decoder.
if self._hparams.warm_start_from_second:
self.initialize_from_ckpt(self._hparams.warm_start_from_second)
return tf.estimator.EstimatorSpec(
tf.estimator.ModeKeys.TRAIN, loss=loss, train_op=train_op)
|
[
"def",
"estimator_spec_train",
"(",
"self",
",",
"loss",
",",
"num_async_replicas",
"=",
"1",
",",
"use_tpu",
"=",
"False",
")",
":",
"train_op",
"=",
"self",
".",
"optimize",
"(",
"loss",
",",
"num_async_replicas",
"=",
"num_async_replicas",
",",
"use_tpu",
"=",
"use_tpu",
")",
"if",
"use_tpu",
":",
"if",
"self",
".",
"_hparams",
".",
"warm_start_from",
":",
"def",
"scaffold_fn",
"(",
")",
":",
"self",
".",
"initialize_from_ckpt",
"(",
"self",
".",
"_hparams",
".",
"warm_start_from",
")",
"return",
"tf",
".",
"train",
".",
"Scaffold",
"(",
")",
"else",
":",
"scaffold_fn",
"=",
"None",
"# Note: important to call this before remove_summaries()",
"if",
"self",
".",
"hparams",
".",
"tpu_enable_host_call",
":",
"host_call",
"=",
"self",
".",
"create_train_host_call",
"(",
")",
"else",
":",
"host_call",
"=",
"None",
"remove_summaries",
"(",
")",
"return",
"tf",
".",
"contrib",
".",
"tpu",
".",
"TPUEstimatorSpec",
"(",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"TRAIN",
",",
"loss",
"=",
"loss",
",",
"train_op",
"=",
"train_op",
",",
"host_call",
"=",
"host_call",
",",
"scaffold_fn",
"=",
"scaffold_fn",
")",
"else",
":",
"if",
"self",
".",
"_hparams",
".",
"warm_start_from",
":",
"self",
".",
"initialize_from_ckpt",
"(",
"self",
".",
"_hparams",
".",
"warm_start_from",
")",
"# When loading weights from a pre-trained model, you want to be able to",
"# load separate weights into the encoder and decoder.",
"if",
"self",
".",
"_hparams",
".",
"warm_start_from_second",
":",
"self",
".",
"initialize_from_ckpt",
"(",
"self",
".",
"_hparams",
".",
"warm_start_from_second",
")",
"return",
"tf",
".",
"estimator",
".",
"EstimatorSpec",
"(",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"TRAIN",
",",
"loss",
"=",
"loss",
",",
"train_op",
"=",
"train_op",
")"
] |
Constructs `tf.estimator.EstimatorSpec` for TRAIN (training) mode.
|
[
"Constructs",
"tf",
".",
"estimator",
".",
"EstimatorSpec",
"for",
"TRAIN",
"(",
"training",
")",
"mode",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L1549-L1586
|
train
|
Constructs an estimator. EstimatorSpec for training mode.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(111) + chr(1532 - 1483) + chr(48) + '\060', 46451 - 46443), ehT0Px3KOsy9(chr(0b110000) + chr(7474 - 7363) + chr(51) + chr(0b110001 + 0o3) + chr(0b10 + 0o62), 33833 - 33825), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(316 - 266) + '\060' + '\065', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(896 - 846) + chr(786 - 735) + '\062', ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + '\062' + '\x37' + chr(102 - 52), 0o10), ehT0Px3KOsy9(chr(860 - 812) + chr(0b1101111) + chr(49) + '\x36', 31917 - 31909), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1488 - 1439) + chr(0b110110) + chr(2204 - 2156), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b100 + 0o153) + chr(963 - 910) + chr(0b101000 + 0o12), 58401 - 58393), ehT0Px3KOsy9(chr(1973 - 1925) + chr(0b1100001 + 0o16) + '\x35' + chr(50), 8), ehT0Px3KOsy9('\060' + chr(0b1001000 + 0o47) + chr(0b1011 + 0o47) + chr(0b10100 + 0o36) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\061' + '\065', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\061' + chr(55) + '\x33', 21427 - 21419), ehT0Px3KOsy9('\060' + '\157' + chr(288 - 238) + '\063' + chr(53), 1708 - 1700), ehT0Px3KOsy9('\060' + chr(111) + chr(97 - 46) + chr(0b110011) + chr(0b110010), 26864 - 26856), ehT0Px3KOsy9('\060' + chr(2512 - 2401) + '\064' + chr(0b1101 + 0o47), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110001) + chr(0b101101 + 0o7) + chr(1524 - 1469), 0o10), ehT0Px3KOsy9('\x30' + chr(11453 - 11342) + chr(0b1110 + 0o43) + chr(51) + chr(0b110110), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(50) + chr(2082 - 2031) + chr(0b110011), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + '\066' + chr(52), 680 - 672), ehT0Px3KOsy9('\060' + chr(0b11011 + 0o124) + '\x31' + chr(0b110 + 0o55) + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(530 - 479) + chr(1839 - 1787) + chr(1190 - 1139), 11096 - 11088), ehT0Px3KOsy9(chr(0b100 + 0o54) + '\157' + '\062' + '\063' + '\x31', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(1006 - 955) + chr(1371 - 1316) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(2102 - 2054) + chr(0b1101111) + chr(55), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + chr(0b110000) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(51) + chr(1072 - 1018) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(4102 - 3991) + '\062' + chr(0b110000) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(212 - 163) + chr(0b110101) + '\065', 0b1000), ehT0Px3KOsy9('\x30' + chr(3544 - 3433) + chr(541 - 491) + chr(51) + chr(0b110001), 8), ehT0Px3KOsy9(chr(736 - 688) + '\157' + chr(50) + '\x34' + '\x30', 65500 - 65492), ehT0Px3KOsy9('\x30' + '\x6f' + '\066' + '\063', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1939 - 1888) + '\066' + '\x30', 8), ehT0Px3KOsy9(chr(48) + chr(5695 - 5584) + chr(1280 - 1230), 0o10), ehT0Px3KOsy9(chr(0b101000 + 0o10) + '\157' + '\063' + '\063' + '\x36', 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + chr(0b101001 + 0o16) + chr(0b110011), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(49) + chr(0b110011) + chr(1566 - 1515), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001 + 0o0) + chr(0b110000) + chr(156 - 106), 62050 - 62042), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(4243 - 4132) + chr(49) + chr(2953 - 2898) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(0b10 + 0o56) + '\157' + '\x31' + chr(1399 - 1351) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b10011 + 0o134) + chr(0b11110 + 0o25) + '\x36', 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b1000 + 0o50) + '\157' + chr(0b101010 + 0o13) + chr(0b111 + 0o51), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'n'), chr(9923 - 9823) + '\145' + chr(9368 - 9269) + chr(111) + '\144' + chr(0b1100101))(chr(0b1110101) + chr(0b1110100) + chr(6141 - 6039) + chr(0b101101) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def RzLnIXYdB2vf(oVre8I6UXc3b, YpO0BcZ6fMsf, FsJHiE9x_3jA=ehT0Px3KOsy9('\x30' + chr(0b100001 + 0o116) + '\x31', 0b1000), L4eE7kczIJwa=ehT0Px3KOsy9('\x30' + '\157' + chr(0b11 + 0o55), 0b1000)):
_sRzZqw7qhHl = oVre8I6UXc3b.optimize(YpO0BcZ6fMsf, num_async_replicas=FsJHiE9x_3jA, use_tpu=L4eE7kczIJwa)
if L4eE7kczIJwa:
if xafqLlk3kkUe(oVre8I6UXc3b._hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'7\x85\xf7\xf7\xe5}\x04KX\xe4R\x1d\xef\xb5I'), '\x64' + chr(101) + '\x63' + chr(0b110010 + 0o75) + chr(0b1100100) + chr(3723 - 3622))('\x75' + chr(9203 - 9087) + chr(0b1100110) + chr(45) + chr(1589 - 1533))):
def UpF8hjySm5v5():
xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b')\x8a\xec\xee\xd3o\x1cCP\xf5R\x1d\xef\xb5I8=\xf0D.'), '\x64' + '\145' + '\143' + chr(111) + chr(4241 - 4141) + '\145')(chr(117) + chr(0b10111 + 0o135) + chr(0b1100110) + '\x2d' + chr(0b111000)))(xafqLlk3kkUe(oVre8I6UXc3b._hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'7\x85\xf7\xf7\xe5}\x04KX\xe4R\x1d\xef\xb5I'), chr(100) + '\145' + '\143' + chr(0b1 + 0o156) + chr(7591 - 7491) + '\x65')(chr(117) + '\164' + chr(0b1100110) + chr(0b101101) + chr(56))))
return xafqLlk3kkUe(IDJ2eXGCBCDu.train, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\x87\xe4\xfc\xdca\x1cN'), '\144' + '\x65' + chr(3663 - 3564) + chr(0b1101111) + '\144' + chr(0b1100101))(chr(0b1110101) + '\x74' + chr(102) + chr(974 - 929) + '\070'))()
else:
UpF8hjySm5v5 = None
if xafqLlk3kkUe(oVre8I6UXc3b.hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'4\x94\xf0\xc5\xdf`\x11HF\xf5R\x13\xf2\xa9P8=\xfaX6'), chr(100) + chr(0b1001100 + 0o31) + chr(0b1100011) + '\x6f' + chr(0b1101 + 0o127) + chr(0b110100 + 0o61))(chr(0b1110101) + chr(13356 - 13240) + chr(102) + chr(0b101101) + '\070')):
WzD9vnp6mKX7 = oVre8I6UXc3b.create_train_host_call()
else:
WzD9vnp6mKX7 = None
VDFLnc5FoR6Y()
return xafqLlk3kkUe(IDJ2eXGCBCDu.contrib.tpu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x14\xb4\xd0\xdf\xc9z\x19GK\xe4b\t\xce\xaaA\x04'), '\144' + '\x65' + chr(99) + '\x6f' + '\x64' + '\x65')('\x75' + '\x74' + chr(4448 - 4346) + chr(0b100100 + 0o11) + chr(1980 - 1924)))(xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'\x14\xb6\xc4\xd3\xf4'), '\144' + chr(101) + '\x63' + chr(0b1101111) + chr(0b1011001 + 0o13) + chr(101))(chr(249 - 132) + chr(0b1110100) + chr(0b101100 + 0o72) + chr(0b101101) + chr(0b10010 + 0o46))), loss=YpO0BcZ6fMsf, train_op=_sRzZqw7qhHl, host_call=WzD9vnp6mKX7, scaffold_fn=UpF8hjySm5v5)
else:
if xafqLlk3kkUe(oVre8I6UXc3b._hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'7\x85\xf7\xf7\xe5}\x04KX\xe4R\x1d\xef\xb5I'), chr(100) + chr(9500 - 9399) + chr(0b1100011) + '\x6f' + '\x64' + chr(101))(chr(0b1110101) + '\164' + chr(2415 - 2313) + chr(0b0 + 0o55) + chr(56))):
xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b')\x8a\xec\xee\xd3o\x1cCP\xf5R\x1d\xef\xb5I8=\xf0D.'), chr(0b1100100) + chr(0b111100 + 0o51) + chr(99) + chr(2998 - 2887) + '\x64' + chr(0b0 + 0o145))('\165' + chr(9682 - 9566) + '\146' + chr(1053 - 1008) + '\070'))(xafqLlk3kkUe(oVre8I6UXc3b._hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'7\x85\xf7\xf7\xe5}\x04KX\xe4R\x1d\xef\xb5I'), chr(0b1100100) + '\145' + chr(99) + chr(9503 - 9392) + '\x64' + chr(0b1100101))(chr(0b110 + 0o157) + chr(0b1100100 + 0o20) + chr(0b11110 + 0o110) + chr(1061 - 1016) + '\x38')))
if xafqLlk3kkUe(oVre8I6UXc3b._hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'7\x85\xf7\xf7\xe5}\x04KX\xe4R\x1d\xef\xb5I8-\xfeW5\x8d\xfd'), chr(100) + chr(101) + chr(0b1100011) + chr(0b1011100 + 0o23) + '\x64' + chr(0b1000101 + 0o40))('\x75' + '\x74' + chr(102) + chr(45) + chr(0b10011 + 0o45))):
xafqLlk3kkUe(oVre8I6UXc3b, xafqLlk3kkUe(SXOLrMavuUCe(b')\x8a\xec\xee\xd3o\x1cCP\xf5R\x1d\xef\xb5I8=\xf0D.'), chr(5803 - 5703) + chr(0b111101 + 0o50) + chr(99) + chr(111) + chr(100) + chr(0b1100101))(chr(10352 - 10235) + chr(8364 - 8248) + chr(102) + '\055' + '\070'))(xafqLlk3kkUe(oVre8I6UXc3b._hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'7\x85\xf7\xf7\xe5}\x04KX\xe4R\x1d\xef\xb5I8-\xfeW5\x8d\xfd'), '\144' + chr(6233 - 6132) + '\x63' + chr(111) + chr(0b110101 + 0o57) + chr(0b1100101))(chr(6666 - 6549) + chr(9739 - 9623) + '\x66' + chr(0b101101) + chr(812 - 756))))
return xafqLlk3kkUe(IDJ2eXGCBCDu.estimator, xafqLlk3kkUe(SXOLrMavuUCe(b'\x05\x97\xf1\xf3\xd7o\x04EX\xc3}\x1e\xfe'), chr(0b1100100) + chr(0b1100101) + chr(0b1001011 + 0o30) + '\x6f' + chr(0b110100 + 0o60) + '\x65')(chr(1119 - 1002) + '\164' + chr(102) + chr(0b101 + 0o50) + '\x38'))(xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'\x14\xb6\xc4\xd3\xf4'), chr(6733 - 6633) + chr(0b1100101) + chr(99) + '\x6f' + chr(5058 - 4958) + '\x65')(chr(0b1110101) + chr(0b1011000 + 0o34) + chr(0b1000011 + 0o43) + chr(45) + '\070')), loss=YpO0BcZ6fMsf, train_op=_sRzZqw7qhHl)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.estimator_spec_eval
|
def estimator_spec_eval(self, features, logits, labels, loss, losses_dict):
"""Constructs `tf.estimator.EstimatorSpec` for EVAL (evaluation) mode."""
del losses_dict
hparams = self.hparams
if not hasattr(hparams, "problem"):
raise NotImplementedError(_no_problem_err("estimator_spec_eval"))
problem = hparams.problem
if common_layers.is_xla_compiled():
# Note: important to call this before remove_summaries()
if self.hparams.tpu_enable_host_call:
host_call = self.create_eval_host_call()
else:
host_call = None
remove_summaries()
eval_metrics_fn = create_tpu_eval_metrics_fn(problem, hparams)
batch_size = [feature.shape.as_list()[0] for _, feature
in features.items() if feature.shape.ndims][0]
# Add batch dimension to all features since tpu requires the batch
# dimension on all tensors.
for name, feature in features.items():
if not feature.shape.as_list():
# All features must have a batch dimension
feature = tf.tile(tf.expand_dims(feature, 0), [batch_size])
features[name] = feature
eval_metrics_fn_args = dict(
logits=logits, # possibly a dict
labels=labels,
features=features, # dict
)
eval_metrics_fn_flat_args = _flatten_dict(eval_metrics_fn_args)
return tf.contrib.tpu.TPUEstimatorSpec(
tf.estimator.ModeKeys.EVAL,
eval_metrics=(eval_metrics_fn, eval_metrics_fn_flat_args),
host_call=host_call,
loss=loss)
else:
task_list = [problem]
if hasattr(problem, "task_list"):
task_list = problem.task_list
eval_metrics_fns = metrics.create_evaluation_metrics(task_list, hparams)
eval_metrics = {}
for metric_name, metric_fn in six.iteritems(eval_metrics_fns):
if isinstance(logits, dict):
# the key is located in the center of metric_name: "metrics-%s/%s/%s"
k = metric_name.split("/")[1]
if k in logits:
eval_metrics[metric_name] = metric_fn(logits[k], features,
features[k])
else:
# We do not make it an error because we sometimes run models that
# predict only parts of the targets defined by the Problem class.
# For example, an autoencoder or pure-video model can run on a gym
# problem even if another model is also predicting other things,
# like actions or rewards.
tf.logging.warning("No key %s in logits for evaluation." % k)
else:
eval_metrics[metric_name] = metric_fn(logits, features,
features["targets"])
if isinstance(logits, dict):
predictions = logits
else:
predictions = {"predictions": logits}
evaluation_hooks = []
# Create a SummarySaverHook
eval_dir = os.path.join(
self.hparams.model_dir,
self.hparams.get("eval_dir_name", "eval"))
eval_summary_hook = tf.train.SummarySaverHook(
save_steps=1,
output_dir=eval_dir,
summary_op=tf.summary.merge_all())
evaluation_hooks.append(eval_summary_hook)
evaluation_hooks += problem.eval_hooks(features, logits, hparams)
return tf.estimator.EstimatorSpec(
tf.estimator.ModeKeys.EVAL,
predictions=predictions,
eval_metric_ops=eval_metrics,
evaluation_hooks=evaluation_hooks,
loss=loss)
|
python
|
def estimator_spec_eval(self, features, logits, labels, loss, losses_dict):
"""Constructs `tf.estimator.EstimatorSpec` for EVAL (evaluation) mode."""
del losses_dict
hparams = self.hparams
if not hasattr(hparams, "problem"):
raise NotImplementedError(_no_problem_err("estimator_spec_eval"))
problem = hparams.problem
if common_layers.is_xla_compiled():
# Note: important to call this before remove_summaries()
if self.hparams.tpu_enable_host_call:
host_call = self.create_eval_host_call()
else:
host_call = None
remove_summaries()
eval_metrics_fn = create_tpu_eval_metrics_fn(problem, hparams)
batch_size = [feature.shape.as_list()[0] for _, feature
in features.items() if feature.shape.ndims][0]
# Add batch dimension to all features since tpu requires the batch
# dimension on all tensors.
for name, feature in features.items():
if not feature.shape.as_list():
# All features must have a batch dimension
feature = tf.tile(tf.expand_dims(feature, 0), [batch_size])
features[name] = feature
eval_metrics_fn_args = dict(
logits=logits, # possibly a dict
labels=labels,
features=features, # dict
)
eval_metrics_fn_flat_args = _flatten_dict(eval_metrics_fn_args)
return tf.contrib.tpu.TPUEstimatorSpec(
tf.estimator.ModeKeys.EVAL,
eval_metrics=(eval_metrics_fn, eval_metrics_fn_flat_args),
host_call=host_call,
loss=loss)
else:
task_list = [problem]
if hasattr(problem, "task_list"):
task_list = problem.task_list
eval_metrics_fns = metrics.create_evaluation_metrics(task_list, hparams)
eval_metrics = {}
for metric_name, metric_fn in six.iteritems(eval_metrics_fns):
if isinstance(logits, dict):
# the key is located in the center of metric_name: "metrics-%s/%s/%s"
k = metric_name.split("/")[1]
if k in logits:
eval_metrics[metric_name] = metric_fn(logits[k], features,
features[k])
else:
# We do not make it an error because we sometimes run models that
# predict only parts of the targets defined by the Problem class.
# For example, an autoencoder or pure-video model can run on a gym
# problem even if another model is also predicting other things,
# like actions or rewards.
tf.logging.warning("No key %s in logits for evaluation." % k)
else:
eval_metrics[metric_name] = metric_fn(logits, features,
features["targets"])
if isinstance(logits, dict):
predictions = logits
else:
predictions = {"predictions": logits}
evaluation_hooks = []
# Create a SummarySaverHook
eval_dir = os.path.join(
self.hparams.model_dir,
self.hparams.get("eval_dir_name", "eval"))
eval_summary_hook = tf.train.SummarySaverHook(
save_steps=1,
output_dir=eval_dir,
summary_op=tf.summary.merge_all())
evaluation_hooks.append(eval_summary_hook)
evaluation_hooks += problem.eval_hooks(features, logits, hparams)
return tf.estimator.EstimatorSpec(
tf.estimator.ModeKeys.EVAL,
predictions=predictions,
eval_metric_ops=eval_metrics,
evaluation_hooks=evaluation_hooks,
loss=loss)
|
[
"def",
"estimator_spec_eval",
"(",
"self",
",",
"features",
",",
"logits",
",",
"labels",
",",
"loss",
",",
"losses_dict",
")",
":",
"del",
"losses_dict",
"hparams",
"=",
"self",
".",
"hparams",
"if",
"not",
"hasattr",
"(",
"hparams",
",",
"\"problem\"",
")",
":",
"raise",
"NotImplementedError",
"(",
"_no_problem_err",
"(",
"\"estimator_spec_eval\"",
")",
")",
"problem",
"=",
"hparams",
".",
"problem",
"if",
"common_layers",
".",
"is_xla_compiled",
"(",
")",
":",
"# Note: important to call this before remove_summaries()",
"if",
"self",
".",
"hparams",
".",
"tpu_enable_host_call",
":",
"host_call",
"=",
"self",
".",
"create_eval_host_call",
"(",
")",
"else",
":",
"host_call",
"=",
"None",
"remove_summaries",
"(",
")",
"eval_metrics_fn",
"=",
"create_tpu_eval_metrics_fn",
"(",
"problem",
",",
"hparams",
")",
"batch_size",
"=",
"[",
"feature",
".",
"shape",
".",
"as_list",
"(",
")",
"[",
"0",
"]",
"for",
"_",
",",
"feature",
"in",
"features",
".",
"items",
"(",
")",
"if",
"feature",
".",
"shape",
".",
"ndims",
"]",
"[",
"0",
"]",
"# Add batch dimension to all features since tpu requires the batch",
"# dimension on all tensors.",
"for",
"name",
",",
"feature",
"in",
"features",
".",
"items",
"(",
")",
":",
"if",
"not",
"feature",
".",
"shape",
".",
"as_list",
"(",
")",
":",
"# All features must have a batch dimension",
"feature",
"=",
"tf",
".",
"tile",
"(",
"tf",
".",
"expand_dims",
"(",
"feature",
",",
"0",
")",
",",
"[",
"batch_size",
"]",
")",
"features",
"[",
"name",
"]",
"=",
"feature",
"eval_metrics_fn_args",
"=",
"dict",
"(",
"logits",
"=",
"logits",
",",
"# possibly a dict",
"labels",
"=",
"labels",
",",
"features",
"=",
"features",
",",
"# dict",
")",
"eval_metrics_fn_flat_args",
"=",
"_flatten_dict",
"(",
"eval_metrics_fn_args",
")",
"return",
"tf",
".",
"contrib",
".",
"tpu",
".",
"TPUEstimatorSpec",
"(",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"EVAL",
",",
"eval_metrics",
"=",
"(",
"eval_metrics_fn",
",",
"eval_metrics_fn_flat_args",
")",
",",
"host_call",
"=",
"host_call",
",",
"loss",
"=",
"loss",
")",
"else",
":",
"task_list",
"=",
"[",
"problem",
"]",
"if",
"hasattr",
"(",
"problem",
",",
"\"task_list\"",
")",
":",
"task_list",
"=",
"problem",
".",
"task_list",
"eval_metrics_fns",
"=",
"metrics",
".",
"create_evaluation_metrics",
"(",
"task_list",
",",
"hparams",
")",
"eval_metrics",
"=",
"{",
"}",
"for",
"metric_name",
",",
"metric_fn",
"in",
"six",
".",
"iteritems",
"(",
"eval_metrics_fns",
")",
":",
"if",
"isinstance",
"(",
"logits",
",",
"dict",
")",
":",
"# the key is located in the center of metric_name: \"metrics-%s/%s/%s\"",
"k",
"=",
"metric_name",
".",
"split",
"(",
"\"/\"",
")",
"[",
"1",
"]",
"if",
"k",
"in",
"logits",
":",
"eval_metrics",
"[",
"metric_name",
"]",
"=",
"metric_fn",
"(",
"logits",
"[",
"k",
"]",
",",
"features",
",",
"features",
"[",
"k",
"]",
")",
"else",
":",
"# We do not make it an error because we sometimes run models that",
"# predict only parts of the targets defined by the Problem class.",
"# For example, an autoencoder or pure-video model can run on a gym",
"# problem even if another model is also predicting other things,",
"# like actions or rewards.",
"tf",
".",
"logging",
".",
"warning",
"(",
"\"No key %s in logits for evaluation.\"",
"%",
"k",
")",
"else",
":",
"eval_metrics",
"[",
"metric_name",
"]",
"=",
"metric_fn",
"(",
"logits",
",",
"features",
",",
"features",
"[",
"\"targets\"",
"]",
")",
"if",
"isinstance",
"(",
"logits",
",",
"dict",
")",
":",
"predictions",
"=",
"logits",
"else",
":",
"predictions",
"=",
"{",
"\"predictions\"",
":",
"logits",
"}",
"evaluation_hooks",
"=",
"[",
"]",
"# Create a SummarySaverHook",
"eval_dir",
"=",
"os",
".",
"path",
".",
"join",
"(",
"self",
".",
"hparams",
".",
"model_dir",
",",
"self",
".",
"hparams",
".",
"get",
"(",
"\"eval_dir_name\"",
",",
"\"eval\"",
")",
")",
"eval_summary_hook",
"=",
"tf",
".",
"train",
".",
"SummarySaverHook",
"(",
"save_steps",
"=",
"1",
",",
"output_dir",
"=",
"eval_dir",
",",
"summary_op",
"=",
"tf",
".",
"summary",
".",
"merge_all",
"(",
")",
")",
"evaluation_hooks",
".",
"append",
"(",
"eval_summary_hook",
")",
"evaluation_hooks",
"+=",
"problem",
".",
"eval_hooks",
"(",
"features",
",",
"logits",
",",
"hparams",
")",
"return",
"tf",
".",
"estimator",
".",
"EstimatorSpec",
"(",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"EVAL",
",",
"predictions",
"=",
"predictions",
",",
"eval_metric_ops",
"=",
"eval_metrics",
",",
"evaluation_hooks",
"=",
"evaluation_hooks",
",",
"loss",
"=",
"loss",
")"
] |
Constructs `tf.estimator.EstimatorSpec` for EVAL (evaluation) mode.
|
[
"Constructs",
"tf",
".",
"estimator",
".",
"EstimatorSpec",
"for",
"EVAL",
"(",
"evaluation",
")",
"mode",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L1588-L1679
|
train
|
Constructs an estimator. EstimatorSpec for EVAL mode.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(3692 - 3581) + chr(49) + '\065', 0b1000), ehT0Px3KOsy9(chr(0b1001 + 0o47) + chr(111) + chr(694 - 643) + chr(1237 - 1185) + '\x36', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1000111 + 0o50) + chr(0b110010) + chr(680 - 626) + chr(610 - 555), 0o10), ehT0Px3KOsy9('\060' + '\157' + '\x31' + chr(54) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(111) + chr(51) + '\x31' + chr(2388 - 2333), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\x33' + '\x34' + '\x32', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(7387 - 7276) + '\061' + chr(0b111 + 0o53) + chr(0b110101), 52806 - 52798), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b100111 + 0o12) + chr(0b110000) + chr(0b100010 + 0o17), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010 + 0o0) + chr(0b110011) + '\x36', 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(1460 - 1411) + chr(0b100111 + 0o16) + chr(0b10000 + 0o43), ord("\x08")), ehT0Px3KOsy9(chr(544 - 496) + '\157' + chr(228 - 179) + '\061' + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1015 - 960) + chr(2111 - 2059), 6105 - 6097), ehT0Px3KOsy9('\060' + chr(0b1000001 + 0o56) + '\063' + chr(0b110010), 42478 - 42470), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(2297 - 2247) + '\x30' + chr(0b11110 + 0o22), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\x33' + '\x30' + '\x31', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b10110 + 0o131) + chr(0b100010 + 0o21) + chr(0b110111) + chr(48), 0o10), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(8636 - 8525) + chr(0b110010) + '\x34' + chr(51), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + chr(0b110100) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(48) + chr(9119 - 9008) + chr(53) + chr(1431 - 1382), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011111 + 0o20) + chr(51) + chr(0b110110) + '\x35', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x31' + '\x30' + '\x33', 0o10), ehT0Px3KOsy9(chr(48) + chr(2075 - 1964) + chr(1064 - 1015) + chr(0b110 + 0o56), 0b1000), ehT0Px3KOsy9('\x30' + chr(9290 - 9179) + '\062' + chr(52) + chr(0b10000 + 0o47), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101001 + 0o6) + '\062' + chr(0b110110) + chr(0b110 + 0o52), 0b1000), ehT0Px3KOsy9('\x30' + chr(1662 - 1551) + chr(49) + chr(54) + chr(49), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\063' + chr(2702 - 2647) + '\066', ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110101) + chr(52), 0b1000), ehT0Px3KOsy9(chr(1990 - 1942) + '\157' + chr(1081 - 1032) + '\x37' + chr(49), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b11001 + 0o32) + chr(0b110110), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + '\061' + '\x33', 48927 - 48919), ehT0Px3KOsy9('\060' + '\x6f' + '\062' + '\067' + chr(2664 - 2611), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1047 - 994) + chr(2170 - 2115), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(0b10111 + 0o32) + '\x36' + '\063', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b11100 + 0o123) + '\x31' + chr(55) + '\x34', 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(49) + '\060' + chr(653 - 600), 0b1000), ehT0Px3KOsy9(chr(0b100001 + 0o17) + '\157' + chr(220 - 169) + '\x35' + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(1092 - 1044) + '\157' + '\065' + '\066', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51) + '\064' + chr(0b110001), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + chr(51) + '\062', 18282 - 18274), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(111) + chr(922 - 872) + chr(1741 - 1688) + '\067', 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b101001 + 0o7) + '\x6f' + chr(1839 - 1786) + chr(48), 17992 - 17984)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'#'), chr(0b1100100) + '\x65' + '\143' + chr(0b1101111) + chr(0b1100100) + chr(0b111100 + 0o51))(chr(0b1110101) + '\164' + '\x66' + chr(45) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def GgMCm9qp6iiJ(oVre8I6UXc3b, EEf4r9nUvta_, wF9nmvjsKjYM, uXMK81tmdpTM, YpO0BcZ6fMsf, hovGDa4hwxi3):
del hovGDa4hwxi3
n4ljua2gi1Pr = oVre8I6UXc3b.n4ljua2gi1Pr
if not lot1PSoAwYhj(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'}\x94\xd8\x12\xa3\x878'), chr(0b10000 + 0o124) + '\x65' + '\x63' + chr(0b111100 + 0o63) + '\144' + chr(0b1000 + 0o135))('\165' + chr(0b1110100) + chr(1381 - 1279) + chr(0b101 + 0o50) + '\070')):
raise _zJ24Vce7wp0(hRoMtT9RG_6C(xafqLlk3kkUe(SXOLrMavuUCe(b'h\x95\xc3\x19\xa2\x83!.\xec4>L\x7f\xe3\x9e\x9d\xd5\xc5E'), chr(0b1100100) + chr(0b101111 + 0o66) + chr(99) + chr(0b1101111) + chr(3312 - 3212) + chr(0b1100101))(chr(117) + chr(0b1010101 + 0o37) + chr(102) + chr(180 - 135) + chr(1068 - 1012))))
sO7e1A_Mor6Q = n4ljua2gi1Pr.sO7e1A_Mor6Q
if xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'd\x95\xe8\x08\xa3\x83\n"\xf1\x06=Uv\xe5\xa5'), chr(0b10 + 0o142) + chr(0b1011000 + 0o15) + chr(99) + '\x6f' + chr(7380 - 7280) + '\x65')('\x75' + '\x74' + chr(102) + '\055' + chr(0b101010 + 0o16)))():
if xafqLlk3kkUe(oVre8I6UXc3b.hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'y\x96\xc2/\xaa\x8c4#\xf2\x0e\x12Tu\xf3\xb5\xa7\xc0\xc5E\x99'), chr(0b101011 + 0o71) + chr(0b0 + 0o145) + '\143' + chr(0b1101111) + '\x64' + '\x65')(chr(10334 - 10217) + chr(0b1110100) + '\x66' + chr(45) + chr(0b1010 + 0o56))):
WzD9vnp6mKX7 = oVre8I6UXc3b.create_eval_host_call()
else:
WzD9vnp6mKX7 = None
VDFLnc5FoR6Y()
phFtXr_1skn0 = pzG_7F0eMX6Y(sO7e1A_Mor6Q, n4ljua2gi1Pr)
ix9dZyeAmUxY = [fVxZREPfp9Oo.shape.as_list()[ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(577 - 529), 0o10)] for (VNGQdHSFPrso, fVxZREPfp9Oo) in EEf4r9nUvta_.NzveIZ3IlSH9() if fVxZREPfp9Oo.shape.ndims][ehT0Px3KOsy9(chr(573 - 525) + '\157' + '\060', 8)]
for (AIvJRzLdDfgF, fVxZREPfp9Oo) in xafqLlk3kkUe(EEf4r9nUvta_, xafqLlk3kkUe(SXOLrMavuUCe(b'C\x9c\xc1\x15\x86\xb8f\x08\xf28\x05\x05'), chr(100) + chr(0b1100101) + chr(0b1010001 + 0o22) + '\157' + '\x64' + chr(101))('\165' + chr(116) + chr(2241 - 2139) + chr(45) + chr(56)))():
if not xafqLlk3kkUe(fVxZREPfp9Oo.shape, xafqLlk3kkUe(SXOLrMavuUCe(b'l\x95\xe8\x1c\xa6\x91!'), chr(0b1100100) + chr(101) + chr(0b1111 + 0o124) + chr(0b1101111) + chr(0b101100 + 0o70) + chr(101))(chr(0b10111 + 0o136) + '\164' + chr(102) + '\x2d' + chr(1450 - 1394)))():
fVxZREPfp9Oo = IDJ2eXGCBCDu.tile(IDJ2eXGCBCDu.expand_dims(fVxZREPfp9Oo, ehT0Px3KOsy9('\060' + chr(111) + '\060', 8)), [ix9dZyeAmUxY])
EEf4r9nUvta_[AIvJRzLdDfgF] = fVxZREPfp9Oo
zLJ_IUkGWF3z = wLqBDw8l0eIm(logits=wF9nmvjsKjYM, labels=uXMK81tmdpTM, features=EEf4r9nUvta_)
bUlCQBA5Vsp2 = vLyZzUZiuT3H(zLJ_IUkGWF3z)
return xafqLlk3kkUe(IDJ2eXGCBCDu.contrib.tpu, xafqLlk3kkUe(SXOLrMavuUCe(b'Y\xb6\xe25\xbc\x96<,\xff\x1f"NI\xf0\xa4\x9b'), chr(8404 - 8304) + '\x65' + chr(7846 - 7747) + chr(0b10110 + 0o131) + chr(0b10001 + 0o123) + chr(101))(chr(117) + chr(116) + '\x66' + chr(45) + '\x38'))(xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'H\xb0\xf6<'), '\144' + '\x65' + '\x63' + chr(11174 - 11063) + chr(100) + chr(101))(chr(12111 - 11994) + chr(0b1110100) + chr(102) + chr(0b1 + 0o54) + '\x38')), eval_metrics=(phFtXr_1skn0, bUlCQBA5Vsp2), host_call=WzD9vnp6mKX7, loss=YpO0BcZ6fMsf)
else:
MXEimWkpnApK = [sO7e1A_Mor6Q]
if lot1PSoAwYhj(sO7e1A_Mor6Q, xafqLlk3kkUe(SXOLrMavuUCe(b'y\x87\xc4\x1b\x90\x8e<2\xea'), chr(0b11 + 0o141) + '\x65' + chr(0b1011011 + 0o10) + chr(0b1101111) + chr(5763 - 5663) + chr(0b1100101))('\x75' + chr(10162 - 10046) + '\146' + chr(0b100010 + 0o13) + '\x38')):
MXEimWkpnApK = sO7e1A_Mor6Q.task_list
ozon0V7aeb2i = yYegMqDoSfs5.create_evaluation_metrics(MXEimWkpnApK, n4ljua2gi1Pr)
gEY30c7K0x8W = {}
for (Fk10FZM6EP2K, sncLXYohINcs) in xafqLlk3kkUe(sYby0kpfssd4, xafqLlk3kkUe(SXOLrMavuUCe(b'd\x92\xd2\x02\xa6\x960,\xed'), chr(7339 - 7239) + chr(0b1100101) + '\143' + chr(1777 - 1666) + '\x64' + chr(0b1100000 + 0o5))(chr(0b1110101) + '\x74' + chr(8619 - 8517) + chr(0b11001 + 0o24) + chr(1650 - 1594)))(ozon0V7aeb2i):
if PlSM16l2KDPD(wF9nmvjsKjYM, wLqBDw8l0eIm):
OolUPRJhRaJd = Fk10FZM6EP2K.split(xafqLlk3kkUe(SXOLrMavuUCe(b'"'), chr(4334 - 4234) + chr(987 - 886) + '\x63' + chr(0b1001011 + 0o44) + chr(0b1100100) + chr(0b1010111 + 0o16))(chr(0b111010 + 0o73) + chr(0b1110100) + '\146' + chr(45) + chr(1606 - 1550)))[ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49), 0o10)]
if OolUPRJhRaJd in wF9nmvjsKjYM:
gEY30c7K0x8W[Fk10FZM6EP2K] = sncLXYohINcs(wF9nmvjsKjYM[OolUPRJhRaJd], EEf4r9nUvta_, EEf4r9nUvta_[OolUPRJhRaJd])
else:
xafqLlk3kkUe(IDJ2eXGCBCDu.logging, xafqLlk3kkUe(SXOLrMavuUCe(b'z\x87\xc5\x1e\xa6\x8c2'), '\144' + chr(7204 - 7103) + chr(0b1100011) + chr(0b1001 + 0o146) + '\144' + '\x65')('\x75' + '\x74' + chr(0b110100 + 0o62) + chr(0b101100 + 0o1) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'C\x89\x97\x1b\xaa\x9bud\xedK$R:\xec\xae\x9f\xca\xd0Z\xd5%c\xd5\x90\x14\x88M\x15k\xea\xc9F\x1c\x12k'), chr(7147 - 7047) + chr(0b1100101) + chr(5542 - 5443) + chr(111) + chr(0b1100100) + chr(101))(chr(117) + chr(116) + '\x66' + '\055' + '\070') % OolUPRJhRaJd)
else:
gEY30c7K0x8W[Fk10FZM6EP2K] = sncLXYohINcs(wF9nmvjsKjYM, EEf4r9nUvta_, EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'y\x87\xc5\x17\xaa\x96&'), chr(6190 - 6090) + chr(1633 - 1532) + '\x63' + chr(0b1100010 + 0o15) + chr(0b1001101 + 0o27) + '\145')(chr(0b1110101) + chr(116) + '\x66' + '\x2d' + '\x38')])
if PlSM16l2KDPD(wF9nmvjsKjYM, wLqBDw8l0eIm):
qIQi_VFCIFZL = wF9nmvjsKjYM
else:
qIQi_VFCIFZL = {xafqLlk3kkUe(SXOLrMavuUCe(b'}\x94\xd2\x14\xa6\x81!(\xf1\x05>'), chr(8059 - 7959) + '\145' + '\x63' + chr(0b1010111 + 0o30) + chr(0b101001 + 0o73) + chr(0b11111 + 0o106))('\x75' + '\x74' + '\x66' + chr(0b101101) + '\070'): wF9nmvjsKjYM}
vreNpghB8eUB = []
K9Eeuulp_V8T = oqhJDdMJfuwx.path.join(oVre8I6UXc3b.hparams.model_dir, oVre8I6UXc3b.hparams.get(xafqLlk3kkUe(SXOLrMavuUCe(b'h\x90\xd6\x1c\x90\x86<3\xc1\x05,Q\x7f'), chr(100) + '\145' + chr(0b11010 + 0o111) + '\157' + '\x64' + '\145')(chr(12042 - 11925) + chr(116) + '\x66' + chr(570 - 525) + chr(0b111000)), xafqLlk3kkUe(SXOLrMavuUCe(b'h\x90\xd6\x1c'), chr(0b1100100) + chr(3968 - 3867) + chr(99) + chr(111) + chr(3133 - 3033) + '\x65')(chr(0b1110101) + '\x74' + chr(102) + chr(197 - 152) + chr(0b111000))))
jKzlyoeEJ297 = IDJ2eXGCBCDu.train.SummarySaverHook(save_steps=ehT0Px3KOsy9(chr(1328 - 1280) + chr(0b1101111) + chr(0b100100 + 0o15), 8), output_dir=K9Eeuulp_V8T, summary_op=IDJ2eXGCBCDu.summary.merge_all())
xafqLlk3kkUe(vreNpghB8eUB, xafqLlk3kkUe(SXOLrMavuUCe(b'l\x96\xc7\x15\xa1\x86'), chr(100) + chr(0b1011110 + 0o7) + '\x63' + chr(0b11001 + 0o126) + chr(8595 - 8495) + chr(0b10010 + 0o123))(chr(0b100111 + 0o116) + '\164' + chr(102) + chr(239 - 194) + chr(0b10101 + 0o43)))(jKzlyoeEJ297)
vreNpghB8eUB += sO7e1A_Mor6Q.eval_hooks(EEf4r9nUvta_, wF9nmvjsKjYM, n4ljua2gi1Pr)
return xafqLlk3kkUe(IDJ2eXGCBCDu.estimator, xafqLlk3kkUe(SXOLrMavuUCe(b'H\x95\xc3\x19\xa2\x83!.\xec8=Yy'), chr(100) + chr(6900 - 6799) + chr(99) + chr(1395 - 1284) + '\x64' + chr(8247 - 8146))(chr(10484 - 10367) + chr(11107 - 10991) + '\x66' + chr(45) + chr(0b111000)))(xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'H\xb0\xf6<'), chr(4298 - 4198) + '\x65' + '\143' + '\157' + chr(0b1100100) + chr(3785 - 3684))(chr(0b11001 + 0o134) + '\x74' + chr(102) + chr(0b101101) + chr(56))), predictions=qIQi_VFCIFZL, eval_metric_ops=gEY30c7K0x8W, evaluation_hooks=vreNpghB8eUB, loss=YpO0BcZ6fMsf)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.estimator_spec_predict
|
def estimator_spec_predict(self, features, use_tpu=False):
"""Constructs `tf.estimator.EstimatorSpec` for PREDICT (inference) mode."""
decode_hparams = self._decode_hparams
top_beams = decode_hparams.beam_size if decode_hparams.return_beams else 1
infer_out = self.infer(
features,
beam_size=decode_hparams.beam_size,
top_beams=top_beams,
alpha=decode_hparams.alpha,
decode_length=decode_hparams.extra_length,
use_tpu=use_tpu)
if isinstance(infer_out, dict):
outputs = infer_out["outputs"]
scores = infer_out["scores"]
else:
outputs = infer_out
scores = None
inputs = features.get("inputs")
if inputs is None:
inputs = features["targets"]
predictions = {
"outputs": outputs,
"scores": scores,
"inputs": inputs,
"targets": features.get("infer_targets"),
}
# Pass through remaining features
for name, feature in features.items():
if name not in list(predictions.keys()) + ["infer_targets"]:
if name == "decode_loop_step":
continue
if not feature.shape.as_list():
# All features must have a batch dimension
batch_size = common_layers.shape_list(outputs)[0]
feature = tf.tile(tf.expand_dims(feature, 0), [batch_size])
predictions[name] = feature
_del_dict_non_tensors(predictions)
export_out = {"outputs": predictions["outputs"]}
if "scores" in predictions:
export_out["scores"] = predictions["scores"]
# Necessary to rejoin examples in the correct order with the Cloud ML Engine
# batch prediction API.
if "batch_prediction_key" in predictions:
export_out["batch_prediction_key"] = predictions["batch_prediction_key"]
export_outputs = {
tf.saved_model.signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY:
tf.estimator.export.PredictOutput(export_out)
}
if use_tpu:
# Note: important to call this before remove_summaries()
if self.hparams.tpu_enable_host_call:
host_call = self.create_eval_host_call()
else:
host_call = None
remove_summaries()
return tf.contrib.tpu.TPUEstimatorSpec(
tf.estimator.ModeKeys.PREDICT,
predictions=predictions,
host_call=host_call,
export_outputs=export_outputs)
else:
return tf.estimator.EstimatorSpec(
tf.estimator.ModeKeys.PREDICT,
predictions=predictions,
export_outputs=export_outputs)
|
python
|
def estimator_spec_predict(self, features, use_tpu=False):
"""Constructs `tf.estimator.EstimatorSpec` for PREDICT (inference) mode."""
decode_hparams = self._decode_hparams
top_beams = decode_hparams.beam_size if decode_hparams.return_beams else 1
infer_out = self.infer(
features,
beam_size=decode_hparams.beam_size,
top_beams=top_beams,
alpha=decode_hparams.alpha,
decode_length=decode_hparams.extra_length,
use_tpu=use_tpu)
if isinstance(infer_out, dict):
outputs = infer_out["outputs"]
scores = infer_out["scores"]
else:
outputs = infer_out
scores = None
inputs = features.get("inputs")
if inputs is None:
inputs = features["targets"]
predictions = {
"outputs": outputs,
"scores": scores,
"inputs": inputs,
"targets": features.get("infer_targets"),
}
# Pass through remaining features
for name, feature in features.items():
if name not in list(predictions.keys()) + ["infer_targets"]:
if name == "decode_loop_step":
continue
if not feature.shape.as_list():
# All features must have a batch dimension
batch_size = common_layers.shape_list(outputs)[0]
feature = tf.tile(tf.expand_dims(feature, 0), [batch_size])
predictions[name] = feature
_del_dict_non_tensors(predictions)
export_out = {"outputs": predictions["outputs"]}
if "scores" in predictions:
export_out["scores"] = predictions["scores"]
# Necessary to rejoin examples in the correct order with the Cloud ML Engine
# batch prediction API.
if "batch_prediction_key" in predictions:
export_out["batch_prediction_key"] = predictions["batch_prediction_key"]
export_outputs = {
tf.saved_model.signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY:
tf.estimator.export.PredictOutput(export_out)
}
if use_tpu:
# Note: important to call this before remove_summaries()
if self.hparams.tpu_enable_host_call:
host_call = self.create_eval_host_call()
else:
host_call = None
remove_summaries()
return tf.contrib.tpu.TPUEstimatorSpec(
tf.estimator.ModeKeys.PREDICT,
predictions=predictions,
host_call=host_call,
export_outputs=export_outputs)
else:
return tf.estimator.EstimatorSpec(
tf.estimator.ModeKeys.PREDICT,
predictions=predictions,
export_outputs=export_outputs)
|
[
"def",
"estimator_spec_predict",
"(",
"self",
",",
"features",
",",
"use_tpu",
"=",
"False",
")",
":",
"decode_hparams",
"=",
"self",
".",
"_decode_hparams",
"top_beams",
"=",
"decode_hparams",
".",
"beam_size",
"if",
"decode_hparams",
".",
"return_beams",
"else",
"1",
"infer_out",
"=",
"self",
".",
"infer",
"(",
"features",
",",
"beam_size",
"=",
"decode_hparams",
".",
"beam_size",
",",
"top_beams",
"=",
"top_beams",
",",
"alpha",
"=",
"decode_hparams",
".",
"alpha",
",",
"decode_length",
"=",
"decode_hparams",
".",
"extra_length",
",",
"use_tpu",
"=",
"use_tpu",
")",
"if",
"isinstance",
"(",
"infer_out",
",",
"dict",
")",
":",
"outputs",
"=",
"infer_out",
"[",
"\"outputs\"",
"]",
"scores",
"=",
"infer_out",
"[",
"\"scores\"",
"]",
"else",
":",
"outputs",
"=",
"infer_out",
"scores",
"=",
"None",
"inputs",
"=",
"features",
".",
"get",
"(",
"\"inputs\"",
")",
"if",
"inputs",
"is",
"None",
":",
"inputs",
"=",
"features",
"[",
"\"targets\"",
"]",
"predictions",
"=",
"{",
"\"outputs\"",
":",
"outputs",
",",
"\"scores\"",
":",
"scores",
",",
"\"inputs\"",
":",
"inputs",
",",
"\"targets\"",
":",
"features",
".",
"get",
"(",
"\"infer_targets\"",
")",
",",
"}",
"# Pass through remaining features",
"for",
"name",
",",
"feature",
"in",
"features",
".",
"items",
"(",
")",
":",
"if",
"name",
"not",
"in",
"list",
"(",
"predictions",
".",
"keys",
"(",
")",
")",
"+",
"[",
"\"infer_targets\"",
"]",
":",
"if",
"name",
"==",
"\"decode_loop_step\"",
":",
"continue",
"if",
"not",
"feature",
".",
"shape",
".",
"as_list",
"(",
")",
":",
"# All features must have a batch dimension",
"batch_size",
"=",
"common_layers",
".",
"shape_list",
"(",
"outputs",
")",
"[",
"0",
"]",
"feature",
"=",
"tf",
".",
"tile",
"(",
"tf",
".",
"expand_dims",
"(",
"feature",
",",
"0",
")",
",",
"[",
"batch_size",
"]",
")",
"predictions",
"[",
"name",
"]",
"=",
"feature",
"_del_dict_non_tensors",
"(",
"predictions",
")",
"export_out",
"=",
"{",
"\"outputs\"",
":",
"predictions",
"[",
"\"outputs\"",
"]",
"}",
"if",
"\"scores\"",
"in",
"predictions",
":",
"export_out",
"[",
"\"scores\"",
"]",
"=",
"predictions",
"[",
"\"scores\"",
"]",
"# Necessary to rejoin examples in the correct order with the Cloud ML Engine",
"# batch prediction API.",
"if",
"\"batch_prediction_key\"",
"in",
"predictions",
":",
"export_out",
"[",
"\"batch_prediction_key\"",
"]",
"=",
"predictions",
"[",
"\"batch_prediction_key\"",
"]",
"export_outputs",
"=",
"{",
"tf",
".",
"saved_model",
".",
"signature_constants",
".",
"DEFAULT_SERVING_SIGNATURE_DEF_KEY",
":",
"tf",
".",
"estimator",
".",
"export",
".",
"PredictOutput",
"(",
"export_out",
")",
"}",
"if",
"use_tpu",
":",
"# Note: important to call this before remove_summaries()",
"if",
"self",
".",
"hparams",
".",
"tpu_enable_host_call",
":",
"host_call",
"=",
"self",
".",
"create_eval_host_call",
"(",
")",
"else",
":",
"host_call",
"=",
"None",
"remove_summaries",
"(",
")",
"return",
"tf",
".",
"contrib",
".",
"tpu",
".",
"TPUEstimatorSpec",
"(",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"PREDICT",
",",
"predictions",
"=",
"predictions",
",",
"host_call",
"=",
"host_call",
",",
"export_outputs",
"=",
"export_outputs",
")",
"else",
":",
"return",
"tf",
".",
"estimator",
".",
"EstimatorSpec",
"(",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"PREDICT",
",",
"predictions",
"=",
"predictions",
",",
"export_outputs",
"=",
"export_outputs",
")"
] |
Constructs `tf.estimator.EstimatorSpec` for PREDICT (inference) mode.
|
[
"Constructs",
"tf",
".",
"estimator",
".",
"EstimatorSpec",
"for",
"PREDICT",
"(",
"inference",
")",
"mode",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L1681-L1754
|
train
|
Constructs tf. estimator. EstimatorSpec for PREDICT mode.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010 + 0o1) + chr(0b110111) + chr(0b1011 + 0o50), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + chr(54) + chr(0b110110), 2782 - 2774), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(51) + chr(0b101110 + 0o11) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(8272 - 8161) + chr(0b110001) + chr(1393 - 1343) + chr(0b10111 + 0o36), 42724 - 42716), ehT0Px3KOsy9(chr(2280 - 2232) + chr(0b1101111) + chr(2004 - 1954) + chr(890 - 836) + '\065', 0b1000), ehT0Px3KOsy9(chr(48) + chr(167 - 56) + '\063' + '\x31' + chr(0b110100), 0o10), ehT0Px3KOsy9('\x30' + chr(6130 - 6019) + chr(49) + '\061' + chr(54), 0b1000), ehT0Px3KOsy9('\060' + chr(668 - 557) + chr(0b110001) + chr(0b110101) + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1001100 + 0o43) + '\x33' + chr(53) + chr(409 - 357), 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + chr(0b101010 + 0o11) + chr(1323 - 1275), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(12068 - 11957) + chr(50) + '\063' + '\065', 18395 - 18387), ehT0Px3KOsy9('\x30' + '\x6f' + chr(756 - 706) + chr(0b110011) + chr(298 - 244), ord("\x08")), ehT0Px3KOsy9(chr(1907 - 1859) + '\x6f' + '\063' + chr(0b110001), 0o10), ehT0Px3KOsy9('\060' + chr(11190 - 11079) + chr(50) + chr(50) + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(0b1011001 + 0o26) + chr(0b1100 + 0o45) + '\063' + '\062', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1000001 + 0o56) + chr(51) + chr(55) + chr(0b10100 + 0o42), 0o10), ehT0Px3KOsy9('\x30' + chr(0b10100 + 0o133) + chr(0b110010) + chr(2444 - 2392) + chr(54), 0o10), ehT0Px3KOsy9('\x30' + chr(705 - 594) + '\064' + chr(1721 - 1671), ord("\x08")), ehT0Px3KOsy9(chr(765 - 717) + chr(0b1101111) + '\061' + chr(0b110011) + chr(0b110001 + 0o2), 0o10), ehT0Px3KOsy9('\x30' + chr(8841 - 8730) + chr(0b11000 + 0o31) + chr(2618 - 2564) + chr(0b101000 + 0o17), 0b1000), ehT0Px3KOsy9(chr(1009 - 961) + chr(5756 - 5645) + chr(0b101001 + 0o11) + chr(943 - 895) + chr(0b1 + 0o57), 24023 - 24015), ehT0Px3KOsy9('\x30' + '\157' + chr(50) + '\x33' + chr(52), 0o10), ehT0Px3KOsy9('\x30' + chr(7578 - 7467) + chr(0b110010) + chr(1180 - 1131) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(1632 - 1521) + '\x32' + '\063' + chr(0b10000 + 0o40), ord("\x08")), ehT0Px3KOsy9(chr(0b1 + 0o57) + '\x6f' + chr(50) + '\066' + '\061', ord("\x08")), ehT0Px3KOsy9(chr(0b11010 + 0o26) + '\157' + chr(0b1011 + 0o50) + '\061' + '\x30', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110010) + '\061', 0b1000), ehT0Px3KOsy9(chr(1281 - 1233) + chr(11435 - 11324) + '\x32' + chr(1023 - 975) + chr(0b10100 + 0o34), 8), ehT0Px3KOsy9('\060' + '\x6f' + chr(805 - 756) + '\065' + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(2482 - 2430) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(0b101111 + 0o1) + '\x6f' + chr(0b110001) + '\x35' + '\x32', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + '\060', 0b1000), ehT0Px3KOsy9(chr(1315 - 1267) + '\157' + '\x35' + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(870 - 822) + chr(0b100100 + 0o113) + chr(0b110011) + chr(0b110111) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010 + 0o145) + chr(0b101011 + 0o10) + chr(54) + chr(0b110100), 63101 - 63093), ehT0Px3KOsy9(chr(48) + chr(2242 - 2131) + chr(0b11101 + 0o26) + chr(610 - 555) + chr(0b110110), 8), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(0b1101111) + chr(0b100110 + 0o14) + chr(53) + '\x34', 22986 - 22978), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(54) + chr(53), 8139 - 8131), ehT0Px3KOsy9('\x30' + chr(2143 - 2032) + '\063' + chr(0b101101 + 0o6) + chr(1919 - 1865), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1000111 + 0o50) + '\061' + chr(50) + chr(1853 - 1802), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x35' + chr(0b110000), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xd5'), chr(0b1100100) + '\145' + chr(99) + chr(8248 - 8137) + '\x64' + '\145')(chr(0b111111 + 0o66) + chr(0b1010101 + 0o37) + chr(8659 - 8557) + chr(856 - 811) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def vU5K7Q4bydHe(oVre8I6UXc3b, EEf4r9nUvta_, L4eE7kczIJwa=ehT0Px3KOsy9(chr(1928 - 1880) + '\x6f' + chr(48), ord("\x08"))):
LrQSWg3uwmK8 = oVre8I6UXc3b.Doynr1RKrgIF
oC1hU_0mlSje = LrQSWg3uwmK8.PQZjDxhiHJGf if LrQSWg3uwmK8.return_beams else ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(0b1101111) + '\x31', 0o10)
ngRpBOdRUjER = oVre8I6UXc3b.IhRMh3nN8G5I(EEf4r9nUvta_, beam_size=LrQSWg3uwmK8.PQZjDxhiHJGf, top_beams=oC1hU_0mlSje, alpha=LrQSWg3uwmK8.gDUX9w35YHFE, decode_length=LrQSWg3uwmK8.extra_length, use_tpu=L4eE7kczIJwa)
if PlSM16l2KDPD(ngRpBOdRUjER, wLqBDw8l0eIm):
Dx_DllZ8uCko = ngRpBOdRUjER[xafqLlk3kkUe(SXOLrMavuUCe(b'\x94\x83\x8c%\xc0\xb8='), '\x64' + '\145' + chr(7819 - 7720) + chr(0b1011 + 0o144) + chr(0b1100100) + chr(8636 - 8535))(chr(117) + chr(0b1110100) + chr(4571 - 4469) + chr(0b11101 + 0o20) + chr(0b111000))]
b8rpGniBNUPr = ngRpBOdRUjER[xafqLlk3kkUe(SXOLrMavuUCe(b"\x88\x95\x97'\xd0\xbf"), chr(0b111011 + 0o51) + '\x65' + chr(0b1010100 + 0o17) + '\x6f' + chr(100) + '\x65')(chr(3283 - 3166) + chr(1104 - 988) + chr(0b1000100 + 0o42) + chr(0b1110 + 0o37) + chr(725 - 669))]
else:
Dx_DllZ8uCko = ngRpBOdRUjER
b8rpGniBNUPr = None
vXoupepMtCXU = EEf4r9nUvta_.get(xafqLlk3kkUe(SXOLrMavuUCe(b'\x92\x98\x88 \xc1\xbf'), chr(1718 - 1618) + chr(0b1100101) + chr(99) + chr(0b1101111) + '\x64' + chr(0b101101 + 0o70))('\x75' + chr(0b1010100 + 0o40) + chr(5127 - 5025) + '\x2d' + chr(0b10100 + 0o44)))
if vXoupepMtCXU is None:
vXoupepMtCXU = EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b'\x8f\x97\x8a2\xd0\xb8='), chr(100) + chr(101) + chr(6684 - 6585) + chr(111) + chr(0b1100100) + chr(0b1100101))(chr(117) + '\164' + '\x66' + '\x2d' + '\070')]
qIQi_VFCIFZL = {xafqLlk3kkUe(SXOLrMavuUCe(b'\x94\x83\x8c%\xc0\xb8='), '\x64' + '\145' + chr(4149 - 4050) + chr(111) + '\144' + chr(6829 - 6728))('\x75' + chr(116) + chr(0b100000 + 0o106) + chr(1225 - 1180) + chr(1706 - 1650)): Dx_DllZ8uCko, xafqLlk3kkUe(SXOLrMavuUCe(b"\x88\x95\x97'\xd0\xbf"), chr(0b1000100 + 0o40) + chr(5629 - 5528) + chr(5617 - 5518) + chr(0b1101111) + chr(9544 - 9444) + chr(6249 - 6148))(chr(0b1110101) + chr(0b1110100) + chr(102) + '\055' + chr(102 - 46)): b8rpGniBNUPr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x92\x98\x88 \xc1\xbf'), chr(5893 - 5793) + chr(4239 - 4138) + chr(99) + chr(111) + '\144' + chr(0b1100101))(chr(0b1110101) + chr(116) + chr(8630 - 8528) + '\x2d' + chr(56)): vXoupepMtCXU, xafqLlk3kkUe(SXOLrMavuUCe(b'\x8f\x97\x8a2\xd0\xb8='), '\x64' + chr(0b1000001 + 0o44) + chr(99) + '\157' + chr(100) + chr(9540 - 9439))('\x75' + chr(0b1110100) + '\x66' + '\055' + chr(0b10001 + 0o47)): EEf4r9nUvta_.get(xafqLlk3kkUe(SXOLrMavuUCe(b'\x92\x98\x9e0\xc7\x93:\r\x9e\xbf\xf8\xeaj'), '\144' + chr(9568 - 9467) + chr(0b1000001 + 0o42) + chr(0b111001 + 0o66) + chr(1521 - 1421) + chr(0b110 + 0o137))(chr(0b1001010 + 0o53) + chr(0b1100011 + 0o21) + chr(0b1000010 + 0o44) + chr(0b101101 + 0o0) + chr(738 - 682)))}
for (AIvJRzLdDfgF, fVxZREPfp9Oo) in xafqLlk3kkUe(EEf4r9nUvta_, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb5\x8c\x8e0\xfc\x96}%\x80\x8b\xd5\xa7'), chr(100) + chr(10012 - 9911) + chr(2754 - 2655) + chr(0b1101111) + chr(100) + '\x65')(chr(0b110110 + 0o77) + '\x74' + chr(0b101011 + 0o73) + chr(45) + '\x38'))():
if AIvJRzLdDfgF not in YyaZ4tpXu4lf(xafqLlk3kkUe(qIQi_VFCIFZL, xafqLlk3kkUe(SXOLrMavuUCe(b'\x90\x93\x81&'), chr(0b1100100) + chr(101) + chr(0b100110 + 0o75) + chr(10831 - 10720) + chr(0b101110 + 0o66) + '\x65')('\165' + chr(116) + '\146' + '\055' + '\070'))()) + [xafqLlk3kkUe(SXOLrMavuUCe(b'\x92\x98\x9e0\xc7\x93:\r\x9e\xbf\xf8\xeaj'), chr(4177 - 4077) + '\145' + chr(889 - 790) + chr(0b1101111) + chr(0b110010 + 0o62) + chr(4475 - 4374))(chr(0b1001111 + 0o46) + chr(0b1110100) + '\x66' + chr(710 - 665) + chr(0b1011 + 0o55))]:
if AIvJRzLdDfgF == xafqLlk3kkUe(SXOLrMavuUCe(b'\x9f\x93\x9b:\xd1\xa9\x11\x00\x83\xb7\xed\xc1j\xe3?\x8e'), chr(100) + chr(0b1100101) + '\x63' + '\x6f' + chr(0b1100100) + '\x65')(chr(117) + chr(116) + chr(102) + chr(1791 - 1746) + chr(0b11011 + 0o35)):
continue
if not xafqLlk3kkUe(fVxZREPfp9Oo.shape, xafqLlk3kkUe(SXOLrMavuUCe(b'\x9a\x85\xa79\xdc\xbf:'), chr(0b1 + 0o143) + '\x65' + chr(9213 - 9114) + chr(0b1101111) + '\x64' + chr(101))(chr(4894 - 4777) + chr(0b1110000 + 0o4) + chr(102) + chr(1402 - 1357) + chr(85 - 29)))():
ix9dZyeAmUxY = jSKPaHwSAfVv.shape_list(Dx_DllZ8uCko)[ehT0Px3KOsy9('\x30' + '\x6f' + chr(48), 8)]
fVxZREPfp9Oo = IDJ2eXGCBCDu.tile(IDJ2eXGCBCDu.expand_dims(fVxZREPfp9Oo, ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b101001 + 0o7), 8)), [ix9dZyeAmUxY])
qIQi_VFCIFZL[AIvJRzLdDfgF] = fVxZREPfp9Oo
rQZi5JQNSt50(qIQi_VFCIFZL)
Oze97nGFUYKG = {xafqLlk3kkUe(SXOLrMavuUCe(b'\x94\x83\x8c%\xc0\xb8='), '\144' + chr(3516 - 3415) + chr(5155 - 5056) + chr(0b1011010 + 0o25) + chr(100) + chr(0b100 + 0o141))(chr(0b1011101 + 0o30) + chr(0b1000010 + 0o62) + chr(0b111001 + 0o55) + chr(0b101101) + chr(0b101101 + 0o13)): qIQi_VFCIFZL[xafqLlk3kkUe(SXOLrMavuUCe(b'\x94\x83\x8c%\xc0\xb8='), chr(0b100011 + 0o101) + chr(0b11100 + 0o111) + chr(0b1100011) + chr(111) + chr(0b1100100) + chr(0b1100101))(chr(0b1110101) + chr(0b1101111 + 0o5) + chr(0b1100110) + chr(0b10010 + 0o33) + chr(0b111000))]}
if xafqLlk3kkUe(SXOLrMavuUCe(b"\x88\x95\x97'\xd0\xbf"), '\144' + chr(8144 - 8043) + chr(99) + '\157' + chr(4282 - 4182) + chr(0b1011011 + 0o12))('\x75' + '\x74' + '\x66' + chr(0b101101) + chr(0b111000)) in qIQi_VFCIFZL:
Oze97nGFUYKG[xafqLlk3kkUe(SXOLrMavuUCe(b"\x88\x95\x97'\xd0\xbf"), '\144' + '\x65' + '\143' + chr(0b1101111) + chr(0b1100100) + chr(0b1100101))(chr(117) + '\x74' + chr(0b1100110) + chr(45) + '\x38')] = qIQi_VFCIFZL[xafqLlk3kkUe(SXOLrMavuUCe(b"\x88\x95\x97'\xd0\xbf"), '\x64' + chr(0b1011101 + 0o10) + '\143' + chr(0b1101111) + '\144' + '\145')('\x75' + chr(2908 - 2792) + chr(0b1100110) + chr(90 - 45) + chr(3080 - 3024))]
if xafqLlk3kkUe(SXOLrMavuUCe(b'\x99\x97\x8c6\xdd\x93>\x1e\x89\xbc\xf4\xfdm\xfe5\x90\xf9I>\x0e'), chr(100) + chr(0b100100 + 0o101) + chr(6604 - 6505) + chr(0b1101111) + chr(0b10101 + 0o117) + chr(101))('\165' + '\164' + '\146' + chr(0b101101) + chr(1768 - 1712)) in qIQi_VFCIFZL:
Oze97nGFUYKG[xafqLlk3kkUe(SXOLrMavuUCe(b'\x99\x97\x8c6\xdd\x93>\x1e\x89\xbc\xf4\xfdm\xfe5\x90\xf9I>\x0e'), chr(0b1100100) + chr(7774 - 7673) + chr(0b10110 + 0o115) + chr(111) + chr(3907 - 3807) + '\145')('\x75' + chr(1903 - 1787) + chr(0b1000100 + 0o42) + chr(45) + chr(2222 - 2166))] = qIQi_VFCIFZL[xafqLlk3kkUe(SXOLrMavuUCe(b'\x99\x97\x8c6\xdd\x93>\x1e\x89\xbc\xf4\xfdm\xfe5\x90\xf9I>\x0e'), chr(2040 - 1940) + chr(101) + '\x63' + chr(0b111110 + 0o61) + chr(0b1100100) + chr(2477 - 2376))(chr(7007 - 6890) + chr(0b1110100) + '\x66' + '\055' + '\070')]
ZYYcacBFQ24K = {IDJ2eXGCBCDu.saved_model.signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY: IDJ2eXGCBCDu.estimator.export.PredictOutput(Oze97nGFUYKG)}
if L4eE7kczIJwa:
if xafqLlk3kkUe(oVre8I6UXc3b.hparams, xafqLlk3kkUe(SXOLrMavuUCe(b'\x8f\x86\x8d\n\xd0\xa2/\x0e\x80\xbd\xc2\xf6v\xe4.\xa1\xc5C7\x1b'), chr(100) + chr(10197 - 10096) + '\143' + chr(2179 - 2068) + '\x64' + chr(4229 - 4128))('\165' + '\164' + chr(4085 - 3983) + chr(45) + '\070')):
WzD9vnp6mKX7 = oVre8I6UXc3b.create_eval_host_call()
else:
WzD9vnp6mKX7 = None
VDFLnc5FoR6Y()
return xafqLlk3kkUe(IDJ2eXGCBCDu.contrib.tpu, xafqLlk3kkUe(SXOLrMavuUCe(b"\xaf\xa6\xad\x10\xc6\xb8'\x01\x8d\xac\xf2\xecJ\xe7?\x9d"), chr(0b1001100 + 0o30) + chr(0b1000101 + 0o40) + '\143' + chr(7127 - 7016) + chr(0b101111 + 0o65) + '\145')(chr(117) + chr(0b1010111 + 0o35) + '\146' + '\055' + chr(56)))(xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'\xab\xa4\xbd\x11\xfc\x8f\x1a'), chr(100) + chr(0b101000 + 0o75) + chr(470 - 371) + chr(0b0 + 0o157) + '\144' + chr(0b110010 + 0o63))('\165' + '\164' + '\x66' + chr(45) + chr(0b11011 + 0o35))), predictions=qIQi_VFCIFZL, host_call=WzD9vnp6mKX7, export_outputs=ZYYcacBFQ24K)
else:
return xafqLlk3kkUe(IDJ2eXGCBCDu.estimator, xafqLlk3kkUe(SXOLrMavuUCe(b'\xbe\x85\x8c<\xd8\xad:\x03\x9e\x8b\xed\xfbz'), '\x64' + chr(101) + chr(99) + '\x6f' + chr(4126 - 4026) + chr(0b101111 + 0o66))('\165' + chr(116) + chr(0b1100110) + '\055' + chr(0b111000)))(xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'\xab\xa4\xbd\x11\xfc\x8f\x1a'), chr(0b1100100) + chr(0b1100101) + chr(0b1100011) + '\157' + chr(100) + chr(0b1100101))('\165' + chr(0b1000101 + 0o57) + chr(0b1100110) + '\x2d' + chr(0b111000))), predictions=qIQi_VFCIFZL, export_outputs=ZYYcacBFQ24K)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel._summarize_losses
|
def _summarize_losses(self, losses_dict):
"""Adds `tf.summary`s to all terms in the losses dictionary."""
if common_layers.should_generate_summaries():
with tf.name_scope("losses"):
for loss_name, loss_val in sorted(losses_dict.items()):
tf.summary.scalar(loss_name, loss_val)
|
python
|
def _summarize_losses(self, losses_dict):
"""Adds `tf.summary`s to all terms in the losses dictionary."""
if common_layers.should_generate_summaries():
with tf.name_scope("losses"):
for loss_name, loss_val in sorted(losses_dict.items()):
tf.summary.scalar(loss_name, loss_val)
|
[
"def",
"_summarize_losses",
"(",
"self",
",",
"losses_dict",
")",
":",
"if",
"common_layers",
".",
"should_generate_summaries",
"(",
")",
":",
"with",
"tf",
".",
"name_scope",
"(",
"\"losses\"",
")",
":",
"for",
"loss_name",
",",
"loss_val",
"in",
"sorted",
"(",
"losses_dict",
".",
"items",
"(",
")",
")",
":",
"tf",
".",
"summary",
".",
"scalar",
"(",
"loss_name",
",",
"loss_val",
")"
] |
Adds `tf.summary`s to all terms in the losses dictionary.
|
[
"Adds",
"tf",
".",
"summary",
"s",
"to",
"all",
"terms",
"in",
"the",
"losses",
"dictionary",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L1771-L1776
|
train
|
Adds tf. summary s to all terms in the losses dictionary.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + '\x6f' + chr(2750 - 2695) + chr(303 - 250), 0o10), ehT0Px3KOsy9(chr(1896 - 1848) + chr(4221 - 4110) + chr(49) + chr(0b110001) + '\x36', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(50) + chr(50) + '\x31', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + chr(55) + chr(49), 24215 - 24207), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + chr(0b110010) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(0b1 + 0o57) + '\157' + chr(0b110001) + chr(1602 - 1552) + chr(0b110101), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b10010 + 0o135) + chr(0b110101) + chr(460 - 411), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(10100 - 9989) + chr(668 - 618) + chr(1341 - 1287) + '\x32', 44475 - 44467), ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\x6f' + '\067' + '\x31', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(50) + chr(584 - 529) + chr(48), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b11001 + 0o126) + '\061' + chr(0b100111 + 0o17) + '\062', 0b1000), ehT0Px3KOsy9(chr(900 - 852) + chr(0b100101 + 0o112) + chr(0b110011) + '\061' + chr(48), 0b1000), ehT0Px3KOsy9(chr(1161 - 1113) + '\x6f' + chr(54) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(0b1101111) + chr(50) + '\x33' + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(1826 - 1778) + chr(0b1100101 + 0o12) + chr(0b110001) + '\x33' + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b100000 + 0o20) + '\x6f' + '\x33' + chr(0b110111) + chr(506 - 456), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(50) + chr(0b1010 + 0o47) + chr(1567 - 1516), 48456 - 48448), ehT0Px3KOsy9('\060' + '\x6f' + '\063' + chr(50) + '\x32', 0o10), ehT0Px3KOsy9(chr(48) + chr(5700 - 5589) + chr(51) + chr(54) + chr(0b110011), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1715 - 1664) + chr(0b110101) + chr(52), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(0b100010 + 0o17) + chr(0b110001) + '\067', 0o10), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(0b1101111) + '\064' + chr(0b110001), 0o10), ehT0Px3KOsy9('\060' + chr(4120 - 4009) + '\062' + chr(52) + chr(0b110010), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1710 - 1659) + chr(0b101110 + 0o7), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110011) + chr(0b110011) + chr(1983 - 1935), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1401 - 1350) + '\061' + chr(0b10101 + 0o34), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b10111 + 0o130) + chr(55) + '\064', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b11 + 0o154) + chr(1367 - 1314), 57338 - 57330), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x34' + chr(0b111 + 0o57), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110001) + '\067' + chr(639 - 586), 16038 - 16030), ehT0Px3KOsy9(chr(0b101011 + 0o5) + '\x6f' + chr(1240 - 1186) + chr(0b110101), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1001111 + 0o40) + '\064' + '\x34', 40227 - 40219), ehT0Px3KOsy9(chr(834 - 786) + chr(0b110000 + 0o77) + chr(0b101101 + 0o5) + '\x33' + '\x34', 0o10), ehT0Px3KOsy9(chr(646 - 598) + '\x6f' + '\x32' + chr(2094 - 2046) + chr(0b100 + 0o56), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(210 - 161) + chr(48) + chr(51), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(990 - 941) + '\065' + chr(0b110101 + 0o2), 48494 - 48486), ehT0Px3KOsy9('\x30' + chr(0b1000101 + 0o52) + chr(60 - 11) + chr(1324 - 1274) + '\x30', 17884 - 17876), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1764 - 1715) + chr(0b110110), 28109 - 28101), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\061' + chr(0b110100) + '\062', 0o10), ehT0Px3KOsy9('\060' + '\157' + '\x33' + '\x34' + '\061', ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b111001 + 0o66) + chr(53) + chr(0b100101 + 0o13), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x13'), chr(9911 - 9811) + chr(101) + chr(0b1011111 + 0o4) + chr(10671 - 10560) + '\x64' + chr(0b1 + 0o144))('\x75' + chr(13363 - 13247) + chr(0b1100110) + chr(0b100110 + 0o7) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def u1mJzUtVjldY(oVre8I6UXc3b, hovGDa4hwxi3):
if xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'N&\xfe\x8c\xf81v\xd5\\\xd6\x17\xbaU\xea>\xa5\xf8\xa7\x9e\x81.S\xcbx\xab'), '\x64' + '\x65' + chr(0b1100011) + chr(0b1101111) + chr(7343 - 7243) + '\x65')('\x75' + chr(0b1110100) + chr(8107 - 8005) + chr(0b10101 + 0o30) + '\x38'))():
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'S/\xfc\x9c\xcb&J\xddI\xdd'), '\144' + chr(1314 - 1213) + chr(0b101011 + 0o70) + chr(111) + '\144' + chr(0b1100101))('\165' + chr(1378 - 1262) + chr(102) + chr(0b10110 + 0o27) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'Q!\xe2\x8a\xf1&'), chr(5425 - 5325) + chr(0b1010010 + 0o23) + chr(99) + chr(4762 - 4651) + '\144' + '\x65')(chr(0b1110101) + '\164' + chr(102) + chr(0b100001 + 0o14) + chr(1094 - 1038))):
for (qK9CnPKXuUIr, TosWBOlGf8bW) in vUlqIvNSaRMa(xafqLlk3kkUe(hovGDa4hwxi3, xafqLlk3kkUe(SXOLrMavuUCe(b's4\xe7\x9c\xdd\x0f\x1a\xfbU\xeb:\xf1'), chr(0b1000 + 0o134) + '\x65' + chr(0b0 + 0o143) + '\x6f' + chr(100) + '\x65')('\165' + chr(116) + '\146' + chr(45) + chr(0b111000)))()):
xafqLlk3kkUe(IDJ2eXGCBCDu.summary, xafqLlk3kkUe(SXOLrMavuUCe(b"N-\xf0\x95\xf5'"), '\144' + '\x65' + '\143' + chr(11447 - 11336) + chr(100) + chr(101))(chr(13097 - 12980) + chr(116) + '\x66' + chr(1809 - 1764) + chr(2676 - 2620)))(qK9CnPKXuUIr, TosWBOlGf8bW)
|
tensorflow/tensor2tensor
|
tensor2tensor/utils/t2t_model.py
|
T2TModel.maybe_scheduled_sampling
|
def maybe_scheduled_sampling(self, features, logits, losses):
"""Scheduled sampling.
Performs forward inference again with "targets" feature replaced with values
sampled from the model.
This is the identity unless self.hparams.scheduled_sampling_prob > 0
(default).
**WARNING**: This is not a faithful implementation of scheduled sampling.
This implementation samples tokens for timestep t condtioned on gold tokens
1...t-1. A proper implementation must condition on a mix of gold and
sampled tokens. Doing so is not efficient for models such like Transformer.
Args:
features: {str: Tensor}. Features sharded along batch dimension.
logits: Tensor. Logits for each shard of data.
losses: 0-D Tensor or (num: 0-D Tensor, denom: 0-D Tensor). Loss Tensor
Returns:
new_logits: Tensor.
new_losses: {str: loss} where loss is one of (i) a 0-D Tensor or
(ii) a (num: 0-D Tensor, denom: 0-D Tensor) pair to be used in a
weighted average.
"""
hparams = self.hparams
problem_hparams = self._problem_hparams
# Only do scheduled sampling if requested.
if hparams.scheduled_sampling_prob == 0.0:
return (logits, losses)
# Only do scheduled sampling on language tasks.
modality = problem_hparams.modality["targets"]
if modality != modalities.ModalityType.SYMBOL:
assert hparams.scheduled_sampling_prob == 0, (
"Scheduled sampling only applies to ModalityType.SYMBOL. Set "
"hparams.scheduled_sampling_prob == 0.0.")
return (logits, losses)
# Only do scheduled sampling when training.
is_training = (hparams.mode == tf.estimator.ModeKeys.TRAIN)
if not is_training:
tf.logging.info("Running in %s mode. Not using scheduled sampling.",
hparams.mode)
return (logits, losses)
# Pad vocabulary if vocab size must be evenly divisible by vocab_divisor.
vocab_size = problem_hparams.vocab_size["targets"]
assert vocab_size is not None
assert hparams.vocab_divisor == 1
def sample(x):
"""Multinomial sampling from a n-dimensional tensor."""
samples = tf.multinomial(tf.reshape(x, [-1, vocab_size]), 1)
reshaped_samples = tf.reshape(samples, common_layers.shape_list(x)[:-1])
return tf.to_int32(reshaped_samples)
def mix_gold_sampled(gold_targets, sampled_targets, mixin_prob):
"""Interleave sampled and gold tokens randomly."""
return tf.where(
tf.less(
tf.random_uniform(common_layers.shape_list(sampled_targets)),
mixin_prob),
sampled_targets,
gold_targets)
def sampled_results(features, logits, mixin_prob):
"""Generate scheduled sampling results."""
sampled_targets = sample(logits)
new_targets = mix_gold_sampled(features["targets"],
sampled_targets,
mixin_prob)
new_targets = tf.stop_gradient(new_targets) # Treat new_targets as given.
new_features = copy.copy(features)
new_features["targets"] = new_targets
with tf.variable_scope(tf.get_variable_scope(), reuse=True):
# Compute bottom() for new_targets.
#
# TODO(duckworthd): Only apply bottom to 'new_targets'.
new_transformed_features = self.bottom(new_features)
# Compute body.
with tf.variable_scope("body"):
new_body_outputs, new_losses = self._normalize_body_output(
self.body(new_transformed_features))
assert "training" not in new_losses
# Compute top.
new_logits = self.top(new_body_outputs, new_features)
# Compute loss. Use original features (== labels).
if (hparams.mode != tf.estimator.ModeKeys.PREDICT and
hparams.mode != "attack"):
new_losses["training"] = self.loss(new_logits, features)
else:
new_losses["training"] = 0.0
return new_logits, new_losses
tf.logging.info("Using scheduled sampling.")
assert hparams.scheduled_sampling_prob == 1.0, (
"hparams.scheduled_sampling_prob must be 0 or 1.")
# Gradually increase over a warmup period. Lower numbers mean more gold
# tokens.
mixin_prob = (
hparams.scheduled_sampling_gold_mixin_prob *
common_layers.inverse_exp_decay(
hparams.scheduled_sampling_warmup_steps,
min_value=0.001)
)
# Apply scheduled sampling over N passes. The logits from the (n-1)-th pass
# will be mixed with gold tokens for conditioning in the n-th pass.
scheduled_sampling_num_passes = getattr(
hparams, "scheduled_sampling_num_passes", 1)
assert scheduled_sampling_num_passes > 0, (
"hparams.scheduled_sampling_num_passes must be > 0 if "
"hparams.scheduled_sampling_prob > 0.0")
new_logits = logits
new_losses = losses
for _ in range(scheduled_sampling_num_passes):
new_logits, new_losses = sampled_results(features, new_logits, mixin_prob)
return new_logits, new_losses
|
python
|
def maybe_scheduled_sampling(self, features, logits, losses):
"""Scheduled sampling.
Performs forward inference again with "targets" feature replaced with values
sampled from the model.
This is the identity unless self.hparams.scheduled_sampling_prob > 0
(default).
**WARNING**: This is not a faithful implementation of scheduled sampling.
This implementation samples tokens for timestep t condtioned on gold tokens
1...t-1. A proper implementation must condition on a mix of gold and
sampled tokens. Doing so is not efficient for models such like Transformer.
Args:
features: {str: Tensor}. Features sharded along batch dimension.
logits: Tensor. Logits for each shard of data.
losses: 0-D Tensor or (num: 0-D Tensor, denom: 0-D Tensor). Loss Tensor
Returns:
new_logits: Tensor.
new_losses: {str: loss} where loss is one of (i) a 0-D Tensor or
(ii) a (num: 0-D Tensor, denom: 0-D Tensor) pair to be used in a
weighted average.
"""
hparams = self.hparams
problem_hparams = self._problem_hparams
# Only do scheduled sampling if requested.
if hparams.scheduled_sampling_prob == 0.0:
return (logits, losses)
# Only do scheduled sampling on language tasks.
modality = problem_hparams.modality["targets"]
if modality != modalities.ModalityType.SYMBOL:
assert hparams.scheduled_sampling_prob == 0, (
"Scheduled sampling only applies to ModalityType.SYMBOL. Set "
"hparams.scheduled_sampling_prob == 0.0.")
return (logits, losses)
# Only do scheduled sampling when training.
is_training = (hparams.mode == tf.estimator.ModeKeys.TRAIN)
if not is_training:
tf.logging.info("Running in %s mode. Not using scheduled sampling.",
hparams.mode)
return (logits, losses)
# Pad vocabulary if vocab size must be evenly divisible by vocab_divisor.
vocab_size = problem_hparams.vocab_size["targets"]
assert vocab_size is not None
assert hparams.vocab_divisor == 1
def sample(x):
"""Multinomial sampling from a n-dimensional tensor."""
samples = tf.multinomial(tf.reshape(x, [-1, vocab_size]), 1)
reshaped_samples = tf.reshape(samples, common_layers.shape_list(x)[:-1])
return tf.to_int32(reshaped_samples)
def mix_gold_sampled(gold_targets, sampled_targets, mixin_prob):
"""Interleave sampled and gold tokens randomly."""
return tf.where(
tf.less(
tf.random_uniform(common_layers.shape_list(sampled_targets)),
mixin_prob),
sampled_targets,
gold_targets)
def sampled_results(features, logits, mixin_prob):
"""Generate scheduled sampling results."""
sampled_targets = sample(logits)
new_targets = mix_gold_sampled(features["targets"],
sampled_targets,
mixin_prob)
new_targets = tf.stop_gradient(new_targets) # Treat new_targets as given.
new_features = copy.copy(features)
new_features["targets"] = new_targets
with tf.variable_scope(tf.get_variable_scope(), reuse=True):
# Compute bottom() for new_targets.
#
# TODO(duckworthd): Only apply bottom to 'new_targets'.
new_transformed_features = self.bottom(new_features)
# Compute body.
with tf.variable_scope("body"):
new_body_outputs, new_losses = self._normalize_body_output(
self.body(new_transformed_features))
assert "training" not in new_losses
# Compute top.
new_logits = self.top(new_body_outputs, new_features)
# Compute loss. Use original features (== labels).
if (hparams.mode != tf.estimator.ModeKeys.PREDICT and
hparams.mode != "attack"):
new_losses["training"] = self.loss(new_logits, features)
else:
new_losses["training"] = 0.0
return new_logits, new_losses
tf.logging.info("Using scheduled sampling.")
assert hparams.scheduled_sampling_prob == 1.0, (
"hparams.scheduled_sampling_prob must be 0 or 1.")
# Gradually increase over a warmup period. Lower numbers mean more gold
# tokens.
mixin_prob = (
hparams.scheduled_sampling_gold_mixin_prob *
common_layers.inverse_exp_decay(
hparams.scheduled_sampling_warmup_steps,
min_value=0.001)
)
# Apply scheduled sampling over N passes. The logits from the (n-1)-th pass
# will be mixed with gold tokens for conditioning in the n-th pass.
scheduled_sampling_num_passes = getattr(
hparams, "scheduled_sampling_num_passes", 1)
assert scheduled_sampling_num_passes > 0, (
"hparams.scheduled_sampling_num_passes must be > 0 if "
"hparams.scheduled_sampling_prob > 0.0")
new_logits = logits
new_losses = losses
for _ in range(scheduled_sampling_num_passes):
new_logits, new_losses = sampled_results(features, new_logits, mixin_prob)
return new_logits, new_losses
|
[
"def",
"maybe_scheduled_sampling",
"(",
"self",
",",
"features",
",",
"logits",
",",
"losses",
")",
":",
"hparams",
"=",
"self",
".",
"hparams",
"problem_hparams",
"=",
"self",
".",
"_problem_hparams",
"# Only do scheduled sampling if requested.",
"if",
"hparams",
".",
"scheduled_sampling_prob",
"==",
"0.0",
":",
"return",
"(",
"logits",
",",
"losses",
")",
"# Only do scheduled sampling on language tasks.",
"modality",
"=",
"problem_hparams",
".",
"modality",
"[",
"\"targets\"",
"]",
"if",
"modality",
"!=",
"modalities",
".",
"ModalityType",
".",
"SYMBOL",
":",
"assert",
"hparams",
".",
"scheduled_sampling_prob",
"==",
"0",
",",
"(",
"\"Scheduled sampling only applies to ModalityType.SYMBOL. Set \"",
"\"hparams.scheduled_sampling_prob == 0.0.\"",
")",
"return",
"(",
"logits",
",",
"losses",
")",
"# Only do scheduled sampling when training.",
"is_training",
"=",
"(",
"hparams",
".",
"mode",
"==",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"TRAIN",
")",
"if",
"not",
"is_training",
":",
"tf",
".",
"logging",
".",
"info",
"(",
"\"Running in %s mode. Not using scheduled sampling.\"",
",",
"hparams",
".",
"mode",
")",
"return",
"(",
"logits",
",",
"losses",
")",
"# Pad vocabulary if vocab size must be evenly divisible by vocab_divisor.",
"vocab_size",
"=",
"problem_hparams",
".",
"vocab_size",
"[",
"\"targets\"",
"]",
"assert",
"vocab_size",
"is",
"not",
"None",
"assert",
"hparams",
".",
"vocab_divisor",
"==",
"1",
"def",
"sample",
"(",
"x",
")",
":",
"\"\"\"Multinomial sampling from a n-dimensional tensor.\"\"\"",
"samples",
"=",
"tf",
".",
"multinomial",
"(",
"tf",
".",
"reshape",
"(",
"x",
",",
"[",
"-",
"1",
",",
"vocab_size",
"]",
")",
",",
"1",
")",
"reshaped_samples",
"=",
"tf",
".",
"reshape",
"(",
"samples",
",",
"common_layers",
".",
"shape_list",
"(",
"x",
")",
"[",
":",
"-",
"1",
"]",
")",
"return",
"tf",
".",
"to_int32",
"(",
"reshaped_samples",
")",
"def",
"mix_gold_sampled",
"(",
"gold_targets",
",",
"sampled_targets",
",",
"mixin_prob",
")",
":",
"\"\"\"Interleave sampled and gold tokens randomly.\"\"\"",
"return",
"tf",
".",
"where",
"(",
"tf",
".",
"less",
"(",
"tf",
".",
"random_uniform",
"(",
"common_layers",
".",
"shape_list",
"(",
"sampled_targets",
")",
")",
",",
"mixin_prob",
")",
",",
"sampled_targets",
",",
"gold_targets",
")",
"def",
"sampled_results",
"(",
"features",
",",
"logits",
",",
"mixin_prob",
")",
":",
"\"\"\"Generate scheduled sampling results.\"\"\"",
"sampled_targets",
"=",
"sample",
"(",
"logits",
")",
"new_targets",
"=",
"mix_gold_sampled",
"(",
"features",
"[",
"\"targets\"",
"]",
",",
"sampled_targets",
",",
"mixin_prob",
")",
"new_targets",
"=",
"tf",
".",
"stop_gradient",
"(",
"new_targets",
")",
"# Treat new_targets as given.",
"new_features",
"=",
"copy",
".",
"copy",
"(",
"features",
")",
"new_features",
"[",
"\"targets\"",
"]",
"=",
"new_targets",
"with",
"tf",
".",
"variable_scope",
"(",
"tf",
".",
"get_variable_scope",
"(",
")",
",",
"reuse",
"=",
"True",
")",
":",
"# Compute bottom() for new_targets.",
"#",
"# TODO(duckworthd): Only apply bottom to 'new_targets'.",
"new_transformed_features",
"=",
"self",
".",
"bottom",
"(",
"new_features",
")",
"# Compute body.",
"with",
"tf",
".",
"variable_scope",
"(",
"\"body\"",
")",
":",
"new_body_outputs",
",",
"new_losses",
"=",
"self",
".",
"_normalize_body_output",
"(",
"self",
".",
"body",
"(",
"new_transformed_features",
")",
")",
"assert",
"\"training\"",
"not",
"in",
"new_losses",
"# Compute top.",
"new_logits",
"=",
"self",
".",
"top",
"(",
"new_body_outputs",
",",
"new_features",
")",
"# Compute loss. Use original features (== labels).",
"if",
"(",
"hparams",
".",
"mode",
"!=",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"PREDICT",
"and",
"hparams",
".",
"mode",
"!=",
"\"attack\"",
")",
":",
"new_losses",
"[",
"\"training\"",
"]",
"=",
"self",
".",
"loss",
"(",
"new_logits",
",",
"features",
")",
"else",
":",
"new_losses",
"[",
"\"training\"",
"]",
"=",
"0.0",
"return",
"new_logits",
",",
"new_losses",
"tf",
".",
"logging",
".",
"info",
"(",
"\"Using scheduled sampling.\"",
")",
"assert",
"hparams",
".",
"scheduled_sampling_prob",
"==",
"1.0",
",",
"(",
"\"hparams.scheduled_sampling_prob must be 0 or 1.\"",
")",
"# Gradually increase over a warmup period. Lower numbers mean more gold",
"# tokens.",
"mixin_prob",
"=",
"(",
"hparams",
".",
"scheduled_sampling_gold_mixin_prob",
"*",
"common_layers",
".",
"inverse_exp_decay",
"(",
"hparams",
".",
"scheduled_sampling_warmup_steps",
",",
"min_value",
"=",
"0.001",
")",
")",
"# Apply scheduled sampling over N passes. The logits from the (n-1)-th pass",
"# will be mixed with gold tokens for conditioning in the n-th pass.",
"scheduled_sampling_num_passes",
"=",
"getattr",
"(",
"hparams",
",",
"\"scheduled_sampling_num_passes\"",
",",
"1",
")",
"assert",
"scheduled_sampling_num_passes",
">",
"0",
",",
"(",
"\"hparams.scheduled_sampling_num_passes must be > 0 if \"",
"\"hparams.scheduled_sampling_prob > 0.0\"",
")",
"new_logits",
"=",
"logits",
"new_losses",
"=",
"losses",
"for",
"_",
"in",
"range",
"(",
"scheduled_sampling_num_passes",
")",
":",
"new_logits",
",",
"new_losses",
"=",
"sampled_results",
"(",
"features",
",",
"new_logits",
",",
"mixin_prob",
")",
"return",
"new_logits",
",",
"new_losses"
] |
Scheduled sampling.
Performs forward inference again with "targets" feature replaced with values
sampled from the model.
This is the identity unless self.hparams.scheduled_sampling_prob > 0
(default).
**WARNING**: This is not a faithful implementation of scheduled sampling.
This implementation samples tokens for timestep t condtioned on gold tokens
1...t-1. A proper implementation must condition on a mix of gold and
sampled tokens. Doing so is not efficient for models such like Transformer.
Args:
features: {str: Tensor}. Features sharded along batch dimension.
logits: Tensor. Logits for each shard of data.
losses: 0-D Tensor or (num: 0-D Tensor, denom: 0-D Tensor). Loss Tensor
Returns:
new_logits: Tensor.
new_losses: {str: loss} where loss is one of (i) a 0-D Tensor or
(ii) a (num: 0-D Tensor, denom: 0-D Tensor) pair to be used in a
weighted average.
|
[
"Scheduled",
"sampling",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/utils/t2t_model.py#L1778-L1901
|
train
|
This function performs forward inference again with targets feature replaced with values
sampled from the model.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b1001011 + 0o44) + chr(0b1010 + 0o54) + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(7108 - 6997) + chr(0b100011 + 0o20) + chr(51) + '\062', 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b10001 + 0o40) + chr(0b100101 + 0o22) + '\x37', 0o10), ehT0Px3KOsy9(chr(1637 - 1589) + chr(111) + '\x31' + chr(0b101010 + 0o6) + chr(0b11111 + 0o24), 0b1000), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(4363 - 4252) + chr(51) + '\x34' + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(1101 - 1053) + chr(111) + '\x32' + '\061' + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\x31' + '\x32' + chr(2201 - 2149), 0o10), ehT0Px3KOsy9('\060' + chr(11929 - 11818) + chr(1555 - 1504) + chr(0b110010) + chr(0b100100 + 0o22), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x31' + chr(0b11001 + 0o34), 2996 - 2988), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(133 - 78) + chr(284 - 233), 0o10), ehT0Px3KOsy9(chr(48) + chr(11050 - 10939) + chr(0b110010) + '\062' + chr(1416 - 1362), 0o10), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(111) + chr(0b110001) + chr(51) + chr(2345 - 2292), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062' + chr(48) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(48) + chr(6984 - 6873) + '\063' + '\x36' + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33' + chr(0b110010) + chr(2778 - 2724), 8), ehT0Px3KOsy9(chr(494 - 446) + '\157' + chr(0b11001 + 0o32) + chr(0b110010), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(966 - 914) + chr(52), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + chr(0b110110) + '\x30', 49598 - 49590), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(49) + '\x30' + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(1111 - 1063) + '\x6f' + chr(0b101100 + 0o5) + '\x37' + chr(0b11111 + 0o23), ord("\x08")), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\157' + chr(0b110011) + '\x37' + chr(743 - 688), 38170 - 38162), ehT0Px3KOsy9('\060' + '\x6f' + chr(49), 46742 - 46734), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b100101 + 0o16) + chr(50) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(3377 - 3266) + '\x31' + chr(256 - 202) + '\066', 0b1000), ehT0Px3KOsy9(chr(0b1001 + 0o47) + '\x6f' + chr(1842 - 1793) + chr(0b110001) + '\061', 25630 - 25622), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1000111 + 0o50) + '\062' + chr(0b110001) + '\x31', 35723 - 35715), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + '\x31' + chr(0b101101 + 0o5), ord("\x08")), ehT0Px3KOsy9(chr(760 - 712) + '\157' + chr(50) + chr(0b101011 + 0o14) + chr(0b111 + 0o60), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(436 - 385) + chr(55) + chr(913 - 862), 0o10), ehT0Px3KOsy9(chr(2190 - 2142) + '\x6f' + '\x37' + chr(52), 0b1000), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(506 - 395) + chr(2142 - 2093) + chr(50) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(0b1101111) + chr(49) + '\064' + '\x30', 0o10), ehT0Px3KOsy9(chr(128 - 80) + '\157' + '\061' + '\063' + '\065', 8), ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(4761 - 4650) + chr(1625 - 1573), 62729 - 62721), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + '\061' + '\x35', 47960 - 47952), ehT0Px3KOsy9('\x30' + chr(0b11 + 0o154) + chr(1725 - 1676) + '\065' + chr(806 - 753), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b100110 + 0o111) + chr(51) + chr(0b110 + 0o61) + chr(54), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(54), 0b1000), ehT0Px3KOsy9(chr(783 - 735) + '\x6f' + '\062' + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1933 - 1880) + chr(0b100101 + 0o22), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(0b1000100 + 0o53) + chr(0b110101) + chr(735 - 687), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x1b'), chr(0b1100100) + chr(3050 - 2949) + chr(0b1100011) + chr(111) + chr(0b1010011 + 0o21) + '\x65')(chr(9007 - 8890) + chr(116) + '\146' + '\x2d' + chr(885 - 829)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def J4djJLafZoiT(oVre8I6UXc3b, EEf4r9nUvta_, wF9nmvjsKjYM, eJKWkHA7qzlZ):
n4ljua2gi1Pr = oVre8I6UXc3b.n4ljua2gi1Pr
sXqesioLf7Ji = oVre8I6UXc3b._problem_hparams
if xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'g\xebJp\xb3\xce\x13\xe4N?\xda-'), chr(0b100 + 0o140) + chr(0b1100101) + '\x63' + chr(8058 - 7947) + '\144' + chr(0b110111 + 0o56))(chr(0b1110000 + 0o5) + chr(12729 - 12613) + chr(0b1001001 + 0o35) + chr(0b101101) + '\070')) == 0.0:
return (wF9nmvjsKjYM, eJKWkHA7qzlZ)
bYPswhysd3s2 = sXqesioLf7Ji.bYPswhysd3s2[xafqLlk3kkUe(SXOLrMavuUCe(b"A\xbb\r$\x85\xf8'"), '\144' + chr(8769 - 8668) + chr(0b1100011) + chr(0b1101111) + chr(100) + chr(0b1010001 + 0o24))('\x75' + '\x74' + chr(102) + '\x2d' + chr(0b1001 + 0o57))]
if bYPswhysd3s2 != xafqLlk3kkUe(PuPeNl0CuqOQ.ModalityType, xafqLlk3kkUe(SXOLrMavuUCe(b'f\x832\x01\xaf\xc0'), chr(9075 - 8975) + chr(0b1100101) + chr(0b110011 + 0o60) + '\157' + '\x64' + chr(101))('\x75' + chr(5518 - 5402) + '\146' + '\x2d' + chr(2416 - 2360))):
assert xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'g\xebJp\xb3\xce\x13\xe4N?\xda-'), chr(0b111111 + 0o45) + chr(8928 - 8827) + '\143' + chr(111) + chr(0b1100100) + chr(0b111000 + 0o55))(chr(117) + chr(0b1011111 + 0o25) + chr(0b100110 + 0o100) + chr(0b101101 + 0o0) + chr(648 - 592))) == ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110000), ord("\x08")), xafqLlk3kkUe(SXOLrMavuUCe(b'f\xb9\x17&\x84\xf98\xb3i[\xe5<\xe9\x84\xbasJ\xd7e\x15\x91m\xa8V(\xf9\xfa\xd3\x92Y&@)k\xed \x91bpC\\\xae\x06\x17\x99\xfc1\xf8^"\xdb\x1f\xcb\xb8\xf8:w\xd51Z\x97q\xb0\x04(\xe4\xf9\x91\x88_=\x059q\xa1\x08\x9aYbNX\xaa\x13*\x8e\xeb\x0b\xa6\x7f\x14\xf4}\xb9\xc9\xf6*\n\x80k'), chr(100) + '\x65' + '\x63' + chr(0b1101111) + chr(0b1100100) + '\x65')('\x75' + chr(116) + '\x66' + chr(0b101101) + '\x38')
return (wF9nmvjsKjYM, eJKWkHA7qzlZ)
XQJVi3cQFN5l = n4ljua2gi1Pr.mode == IDJ2eXGCBCDu.estimator.ModeKeys.TRAIN
if not XQJVi3cQFN5l:
xafqLlk3kkUe(IDJ2eXGCBCDu.logging, xafqLlk3kkUe(SXOLrMavuUCe(b'f\xed7;\x95\xef3\xe1g\x17\xcc6'), '\x64' + chr(101) + '\143' + chr(1890 - 1779) + chr(8736 - 8636) + chr(101))(chr(0b1110101) + chr(0b1 + 0o163) + chr(102) + '\x2d' + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'g\xaf\x11-\x89\xe23\xf6d\x15\xb6x\xf7\xd4\xbbu@\xd5kZ\xb1n\xa5V<\xfa\xe3\xd1\x9c\x1c&\x035a\xa9\x18\x92cu\x0fF\xbb\x123\x8c\xe5:\xb1#'), chr(0b110001 + 0o63) + chr(0b110001 + 0o64) + chr(5853 - 5754) + chr(1426 - 1315) + chr(5012 - 4912) + chr(101))(chr(13527 - 13410) + chr(0b1110100) + chr(0b1010101 + 0o21) + chr(425 - 380) + '\070'), xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'X\xb5\x1b&'), chr(100) + chr(0b1100101) + '\x63' + chr(0b1000011 + 0o54) + '\x64' + chr(101))(chr(8779 - 8662) + '\164' + chr(0b1100100 + 0o2) + '\055' + chr(56))))
return (wF9nmvjsKjYM, eJKWkHA7qzlZ)
CeyMIoSyrpkQ = sXqesioLf7Ji.CeyMIoSyrpkQ[xafqLlk3kkUe(SXOLrMavuUCe(b"A\xbb\r$\x85\xf8'"), '\x64' + chr(0b1100101) + '\143' + '\x6f' + chr(100) + '\145')(chr(0b1110101) + '\164' + chr(0b10100 + 0o122) + chr(1578 - 1533) + chr(56))]
assert CeyMIoSyrpkQ is not None
assert xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'C\xb5\x1c"\x82\xd30\xbf{\x12\xe52\xf6'), '\144' + '\x65' + '\x63' + chr(0b1101111) + '\x64' + chr(0b100010 + 0o103))(chr(0b1110101) + '\164' + chr(0b1011001 + 0o15) + chr(0b101101) + chr(2091 - 2035))) == ehT0Px3KOsy9(chr(625 - 577) + chr(0b1101111) + chr(0b11 + 0o56), 8)
def aBu4gMMQp6Jg(OeWW0F1dBPRQ):
db1_IZvznkcy = IDJ2eXGCBCDu.multinomial(IDJ2eXGCBCDu.reshape(OeWW0F1dBPRQ, [-ehT0Px3KOsy9(chr(2024 - 1976) + chr(0b1101100 + 0o3) + chr(0b11 + 0o56), 8), CeyMIoSyrpkQ]), ehT0Px3KOsy9('\060' + '\x6f' + chr(49), 8))
EU_KDAywSzFO = IDJ2eXGCBCDu.reshape(db1_IZvznkcy, jSKPaHwSAfVv.shape_list(OeWW0F1dBPRQ)[:-ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x31', 8)])
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'A\xb5 *\x8e\xf8g\xe4'), chr(0b1100100) + chr(101) + chr(0b1011011 + 0o10) + chr(111) + '\x64' + chr(0b10100 + 0o121))('\x75' + '\164' + '\146' + '\x2d' + chr(56)))(EU_KDAywSzFO)
def Jt7aA5x3z7MA(CmeMG4LbscU2, ef2f32THep9j, reWOsQOQs3bO):
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'Q\x889\x02\xa3\xb9m\xaf\\9\xfb\x02'), chr(100) + '\x65' + chr(99) + chr(111) + chr(4054 - 3954) + chr(9354 - 9253))(chr(0b1110101) + '\x74' + chr(0b100 + 0o142) + chr(1614 - 1569) + '\x38'))(xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'Y\xbf\x0c0'), chr(0b10111 + 0o115) + chr(0b1100101) + chr(8404 - 8305) + '\157' + chr(0b1100100) + chr(0b1001100 + 0o31))('\165' + chr(116) + '\146' + chr(45) + chr(0b111000)))(xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b"G\xbb\x11'\x8f\xe1\x0b\xa3c\x12\xf02\xf6\x99"), '\x64' + chr(0b1010000 + 0o25) + chr(0b110011 + 0o60) + '\157' + chr(100) + '\x65')(chr(0b1010101 + 0o40) + chr(0b1011 + 0o151) + chr(0b1100110) + chr(0b10000 + 0o35) + chr(0b111000)))(xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'F\xb2\x1e3\x85\xd38\xbf~\x0f'), '\x64' + chr(0b1100101) + '\143' + chr(111) + chr(1840 - 1740) + '\x65')('\165' + chr(0b101110 + 0o106) + '\146' + chr(0b1101 + 0o40) + chr(1212 - 1156)))(ef2f32THep9j)), reWOsQOQs3bO), ef2f32THep9j, CmeMG4LbscU2)
def z3_SAzcosBci(EEf4r9nUvta_, wF9nmvjsKjYM, reWOsQOQs3bO):
ef2f32THep9j = aBu4gMMQp6Jg(wF9nmvjsKjYM)
EsPsBO7RYJsx = Jt7aA5x3z7MA(EEf4r9nUvta_[xafqLlk3kkUe(SXOLrMavuUCe(b"A\xbb\r$\x85\xf8'"), chr(2411 - 2311) + chr(401 - 300) + chr(3347 - 3248) + chr(111) + chr(0b1100100) + '\x65')(chr(117) + chr(116) + chr(0b1100011 + 0o3) + '\x2d' + chr(2459 - 2403))], ef2f32THep9j, reWOsQOQs3bO)
EsPsBO7RYJsx = IDJ2eXGCBCDu.stop_gradient(EsPsBO7RYJsx)
uLZlp2gT9WJ8 = igThHS4jwVsa.igThHS4jwVsa(EEf4r9nUvta_)
uLZlp2gT9WJ8[xafqLlk3kkUe(SXOLrMavuUCe(b"A\xbb\r$\x85\xf8'"), '\x64' + chr(3924 - 3823) + chr(99) + chr(2580 - 2469) + '\x64' + chr(0b11001 + 0o114))(chr(0b1110101) + chr(0b1110100) + chr(102) + chr(0b11000 + 0o25) + chr(0b101100 + 0o14))] = EsPsBO7RYJsx
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'C\xbb\r*\x81\xee8\xb3R\x08\xf52\xf4\x91'), '\144' + chr(4050 - 3949) + chr(99) + chr(6856 - 6745) + '\x64' + '\x65')('\165' + chr(0b11011 + 0o131) + '\146' + '\055' + '\x38'))(xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'R\xbf\x0b\x1c\x96\xed&\xbfl\x19\xfa8\xdb\x87\xb5uT\xd5'), chr(100) + chr(101) + chr(0b1011011 + 0o10) + chr(11655 - 11544) + chr(2963 - 2863) + '\145')(chr(0b1110101) + chr(951 - 835) + chr(2219 - 2117) + chr(0b1100 + 0o41) + chr(0b100101 + 0o23)))(), reuse=ehT0Px3KOsy9(chr(218 - 170) + chr(111) + chr(2057 - 2008), 8)):
Ao8QxyHMiIHT = oVre8I6UXc3b.kXxsZxlIQUSQ(uLZlp2gT9WJ8)
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'C\xbb\r*\x81\xee8\xb3R\x08\xf52\xf4\x91'), chr(0b1100100) + '\x65' + '\143' + chr(10219 - 10108) + '\144' + chr(10025 - 9924))(chr(0b1110101) + '\x74' + chr(3939 - 3837) + chr(1046 - 1001) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'W\xb5\x1b:'), chr(0b1100100) + chr(6167 - 6066) + chr(0b11001 + 0o112) + chr(111) + chr(0b1001001 + 0o33) + '\145')(chr(0b1110101) + chr(7116 - 7000) + '\x66' + chr(45) + chr(0b111000))):
(Qj31fKV87QoX, Cn2sMYhV3kba) = oVre8I6UXc3b._normalize_body_output(oVre8I6UXc3b.TD8C81EGml3n(Ao8QxyHMiIHT))
assert xafqLlk3kkUe(SXOLrMavuUCe(b'A\xa8\x1e*\x8e\xe5:\xb1'), '\144' + chr(6955 - 6854) + '\143' + chr(0b1100000 + 0o17) + chr(100) + chr(101))(chr(0b1100111 + 0o16) + chr(116) + chr(0b10010 + 0o124) + chr(0b101101) + chr(3084 - 3028)) not in Cn2sMYhV3kba
DaGpVPXdxaey = oVre8I6UXc3b.qxrVBjeryNEZ(Qj31fKV87QoX, uLZlp2gT9WJ8)
if xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'X\xb5\x1b&'), '\144' + '\x65' + '\x63' + '\x6f' + chr(1286 - 1186) + chr(8707 - 8606))(chr(0b1100011 + 0o22) + chr(116) + chr(2663 - 2561) + '\x2d' + '\x38')) != xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'e\x88:\x07\xa9\xcf\x00'), chr(6248 - 6148) + '\x65' + chr(0b1101 + 0o126) + chr(111) + chr(100) + chr(101))(chr(0b1000 + 0o155) + chr(0b1110100) + '\x66' + chr(0b1101 + 0o40) + chr(0b101011 + 0o15))) and xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'X\xb5\x1b&'), '\x64' + chr(1516 - 1415) + chr(0b1010101 + 0o16) + '\x6f' + '\x64' + chr(0b1100101))(chr(117) + chr(0b1110100) + chr(5465 - 5363) + chr(1059 - 1014) + '\x38')) != xafqLlk3kkUe(SXOLrMavuUCe(b'T\xae\x0b"\x83\xe7'), chr(0b1011 + 0o131) + chr(0b1100101) + chr(6604 - 6505) + chr(3794 - 3683) + chr(100) + chr(101))(chr(0b1110101) + chr(116) + chr(3656 - 3554) + '\x2d' + chr(56)):
Cn2sMYhV3kba[xafqLlk3kkUe(SXOLrMavuUCe(b'A\xa8\x1e*\x8e\xe5:\xb1'), chr(8270 - 8170) + chr(0b1100101) + '\143' + '\157' + '\x64' + chr(101))(chr(0b1100 + 0o151) + chr(2445 - 2329) + chr(0b1100110) + chr(0b10011 + 0o32) + chr(56))] = oVre8I6UXc3b.YpO0BcZ6fMsf(DaGpVPXdxaey, EEf4r9nUvta_)
else:
Cn2sMYhV3kba[xafqLlk3kkUe(SXOLrMavuUCe(b'A\xa8\x1e*\x8e\xe5:\xb1'), '\144' + chr(0b1100101) + chr(99) + chr(7853 - 7742) + chr(6934 - 6834) + chr(0b1100101))('\165' + '\164' + chr(0b1100110) + chr(45) + chr(0b111000))] = 0.0
return (DaGpVPXdxaey, Cn2sMYhV3kba)
xafqLlk3kkUe(IDJ2eXGCBCDu.logging, xafqLlk3kkUe(SXOLrMavuUCe(b'f\xed7;\x95\xef3\xe1g\x17\xcc6'), chr(7655 - 7555) + chr(0b101011 + 0o72) + '\143' + chr(111) + '\x64' + chr(5675 - 5574))('\165' + '\164' + chr(2945 - 2843) + chr(0b101100 + 0o1) + chr(0b1110 + 0o52)))(xafqLlk3kkUe(SXOLrMavuUCe(b"`\xa9\x16-\x87\xac'\xb5e\x1e\xf2(\xe8\x91\xb2:W\xd1(\n\x93h\xbf\x11g"), chr(100) + chr(0b1100101) + '\x63' + '\157' + chr(100) + chr(3111 - 3010))(chr(117) + '\x74' + chr(0b11000 + 0o116) + '\055' + chr(1882 - 1826)))
assert xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'g\xebJp\xb3\xce\x13\xe4N?\xda-'), chr(0b1100100) + chr(0b1100101) + '\143' + chr(111) + chr(0b1010100 + 0o20) + chr(0b0 + 0o145))('\x75' + chr(0b110000 + 0o104) + '\x66' + chr(45) + chr(0b111000))) == 1.0, xafqLlk3kkUe(SXOLrMavuUCe(b"]\xaa\x1e1\x81\xe1'\xf8~\x18\xfe8\xe0\x81\xba\x7f@\xef6\x1b\x92q\xbd\x1f'\xee\xd5\xcf\x89S7@0q\xbe\x19\xdedt\x0f\x05\xfa\x101\xc0\xbdz"), chr(0b11101 + 0o107) + '\x65' + '\143' + '\157' + chr(0b1100100) + '\x65')('\x75' + chr(0b1110100) + chr(0b1100110) + chr(1499 - 1454) + chr(56))
reWOsQOQs3bO = n4ljua2gi1Pr.scheduled_sampling_gold_mixin_prob * jSKPaHwSAfVv.inverse_exp_decay(n4ljua2gi1Pr.fS20WIxYM77_, min_value=0.001)
y3HdFQBfHEXu = xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'F\xb9\x17&\x84\xf98\xb3i$\xe5<\xe9\x84\xbasJ\xd7\x1a\x14\x8al\x8e\x06(\xfa\xf9\xda\x88'), chr(8332 - 8232) + chr(101) + chr(5037 - 4938) + chr(111) + chr(100) + chr(0b1100101))(chr(0b1110101) + '\164' + chr(102) + '\x2d' + '\070'), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(535 - 486), 8))
assert y3HdFQBfHEXu > ehT0Px3KOsy9(chr(0b0 + 0o60) + '\157' + chr(48), 8), xafqLlk3kkUe(SXOLrMavuUCe(b"]\xaa\x1e1\x81\xe1'\xf8~\x18\xfe8\xe0\x81\xba\x7f@\xef6\x1b\x92q\xbd\x1f'\xee\xd5\xd1\x8eQ\n\x10<w\xbe\x08\x8d&|ZF\xae_!\x85\xacj\xf6=[\xff;\xa4\x9c\xa6{V\xd1(\t\xd1r\xb2\x1e,\xed\xff\xd3\x9eX\n\x13<i\xbd\x01\x97hvpE\xa8\x10!\xc0\xb2t\xe6#K"), chr(2712 - 2612) + '\x65' + chr(99) + chr(11908 - 11797) + '\x64' + chr(0b1100101))(chr(0b1110101) + chr(0b1110100) + chr(102) + chr(45) + '\x38')
DaGpVPXdxaey = wF9nmvjsKjYM
Cn2sMYhV3kba = eJKWkHA7qzlZ
for VNGQdHSFPrso in vQr8gNKaIaWE(y3HdFQBfHEXu):
(DaGpVPXdxaey, Cn2sMYhV3kba) = z3_SAzcosBci(EEf4r9nUvta_, DaGpVPXdxaey, reWOsQOQs3bO)
return (DaGpVPXdxaey, Cn2sMYhV3kba)
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_prepare_decoder
|
def attention_lm_moe_prepare_decoder(targets, hparams):
"""Prepare one shard of the model for the decoder.
Args:
targets: a Tensor.
hparams: run hyperparameters
Returns:
decoder_input: a Tensor, bottom of decoder stack
decoder_self_attention_bias: a Tensor, containing large negative values
to implement masked attention and possibly biases for diagonal alignments
pad_remover (expert_utils.PadRemover): an util object to remove padding
"""
targets_pad_mask = common_attention.embedding_to_padding(targets)
with tf.name_scope("pad_remover"):
# Because of the shift_right, the <eos> token will be considered as
# padding. In practice, it doesn't really matter, due to the triangular
# mask, this token should never be attended.
pad_remover = expert_utils.PadRemover(targets_pad_mask)
if hparams.prepend_mode == "prepend_inputs_full_attention":
decoder_self_attention_bias = (
common_attention.attention_bias_prepend_inputs_full_attention(
targets_pad_mask))
else:
decoder_self_attention_bias = (
common_attention.attention_bias_lower_triangle(tf.shape(targets)[1]))
decoder_input = common_layers.shift_right_3d(targets)
if hparams.pos == "timing":
decoder_input = common_attention.add_timing_signal_1d(decoder_input)
return (decoder_input, decoder_self_attention_bias, pad_remover)
|
python
|
def attention_lm_moe_prepare_decoder(targets, hparams):
"""Prepare one shard of the model for the decoder.
Args:
targets: a Tensor.
hparams: run hyperparameters
Returns:
decoder_input: a Tensor, bottom of decoder stack
decoder_self_attention_bias: a Tensor, containing large negative values
to implement masked attention and possibly biases for diagonal alignments
pad_remover (expert_utils.PadRemover): an util object to remove padding
"""
targets_pad_mask = common_attention.embedding_to_padding(targets)
with tf.name_scope("pad_remover"):
# Because of the shift_right, the <eos> token will be considered as
# padding. In practice, it doesn't really matter, due to the triangular
# mask, this token should never be attended.
pad_remover = expert_utils.PadRemover(targets_pad_mask)
if hparams.prepend_mode == "prepend_inputs_full_attention":
decoder_self_attention_bias = (
common_attention.attention_bias_prepend_inputs_full_attention(
targets_pad_mask))
else:
decoder_self_attention_bias = (
common_attention.attention_bias_lower_triangle(tf.shape(targets)[1]))
decoder_input = common_layers.shift_right_3d(targets)
if hparams.pos == "timing":
decoder_input = common_attention.add_timing_signal_1d(decoder_input)
return (decoder_input, decoder_self_attention_bias, pad_remover)
|
[
"def",
"attention_lm_moe_prepare_decoder",
"(",
"targets",
",",
"hparams",
")",
":",
"targets_pad_mask",
"=",
"common_attention",
".",
"embedding_to_padding",
"(",
"targets",
")",
"with",
"tf",
".",
"name_scope",
"(",
"\"pad_remover\"",
")",
":",
"# Because of the shift_right, the <eos> token will be considered as",
"# padding. In practice, it doesn't really matter, due to the triangular",
"# mask, this token should never be attended.",
"pad_remover",
"=",
"expert_utils",
".",
"PadRemover",
"(",
"targets_pad_mask",
")",
"if",
"hparams",
".",
"prepend_mode",
"==",
"\"prepend_inputs_full_attention\"",
":",
"decoder_self_attention_bias",
"=",
"(",
"common_attention",
".",
"attention_bias_prepend_inputs_full_attention",
"(",
"targets_pad_mask",
")",
")",
"else",
":",
"decoder_self_attention_bias",
"=",
"(",
"common_attention",
".",
"attention_bias_lower_triangle",
"(",
"tf",
".",
"shape",
"(",
"targets",
")",
"[",
"1",
"]",
")",
")",
"decoder_input",
"=",
"common_layers",
".",
"shift_right_3d",
"(",
"targets",
")",
"if",
"hparams",
".",
"pos",
"==",
"\"timing\"",
":",
"decoder_input",
"=",
"common_attention",
".",
"add_timing_signal_1d",
"(",
"decoder_input",
")",
"return",
"(",
"decoder_input",
",",
"decoder_self_attention_bias",
",",
"pad_remover",
")"
] |
Prepare one shard of the model for the decoder.
Args:
targets: a Tensor.
hparams: run hyperparameters
Returns:
decoder_input: a Tensor, bottom of decoder stack
decoder_self_attention_bias: a Tensor, containing large negative values
to implement masked attention and possibly biases for diagonal alignments
pad_remover (expert_utils.PadRemover): an util object to remove padding
|
[
"Prepare",
"one",
"shard",
"of",
"the",
"model",
"for",
"the",
"decoder",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L334-L364
|
train
|
Prepare one shard of the model for the decoder.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b100 + 0o153) + chr(2324 - 2273) + '\062' + chr(0b110000), ord("\x08")), ehT0Px3KOsy9('\060' + chr(6632 - 6521) + '\x31' + '\060' + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(0b101000 + 0o10) + '\157' + chr(0b11100 + 0o26) + chr(0b1100 + 0o44) + chr(694 - 640), ord("\x08")), ehT0Px3KOsy9(chr(0b101011 + 0o5) + '\157' + '\061' + '\062' + chr(678 - 626), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001 + 0o2) + '\066' + '\x33', 9175 - 9167), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(4595 - 4484) + chr(49) + '\060' + '\x36', 6156 - 6148), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\061' + chr(55) + chr(49), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x36' + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(909 - 858) + chr(0b100111 + 0o14), 41064 - 41056), ehT0Px3KOsy9(chr(48) + chr(111) + chr(1754 - 1703) + chr(53) + chr(601 - 553), 0o10), ehT0Px3KOsy9(chr(1855 - 1807) + chr(0b1101111) + '\x32' + chr(0b1011 + 0o52) + '\063', 46821 - 46813), ehT0Px3KOsy9('\060' + '\x6f' + chr(159 - 109) + chr(2034 - 1983) + chr(0b110010), 40945 - 40937), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010100 + 0o33) + chr(50) + chr(53) + chr(48), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(50) + chr(894 - 842), 4224 - 4216), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(53) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x31' + chr(2156 - 2102) + chr(0b100111 + 0o16), 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(50) + '\x37', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1010010 + 0o35) + chr(51) + '\x30' + chr(2555 - 2503), ord("\x08")), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(0b1101111) + chr(0b100001 + 0o20) + chr(53) + '\x34', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1073 - 1024) + '\x33', 16287 - 16279), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010111 + 0o30) + '\x32' + chr(51) + chr(0b110001), 28803 - 28795), ehT0Px3KOsy9('\x30' + '\x6f' + chr(2476 - 2426) + chr(0b110101) + '\x31', 0o10), ehT0Px3KOsy9(chr(0b11000 + 0o30) + '\x6f' + chr(0b101010 + 0o11) + chr(52) + chr(2676 - 2624), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1000010 + 0o55) + chr(0b100010 + 0o17) + chr(52) + chr(51), 0b1000), ehT0Px3KOsy9(chr(1937 - 1889) + chr(0b1101111) + chr(0b110011) + chr(0b110100) + chr(1801 - 1748), ord("\x08")), ehT0Px3KOsy9(chr(0b101100 + 0o4) + '\x6f' + '\x32' + chr(0b110 + 0o53) + chr(0b1100 + 0o44), ord("\x08")), ehT0Px3KOsy9('\060' + chr(2066 - 1955) + chr(50) + chr(0b100110 + 0o21) + chr(55), 0o10), ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(0b111001 + 0o66) + '\x35' + chr(1008 - 959), 0o10), ehT0Px3KOsy9('\060' + chr(0b10 + 0o155) + chr(50) + chr(0b110001) + '\x36', 0b1000), ehT0Px3KOsy9(chr(1966 - 1918) + chr(111) + chr(0b110100) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b10111 + 0o130) + chr(0b111 + 0o56) + chr(0b110101), 36232 - 36224), ehT0Px3KOsy9(chr(0b101111 + 0o1) + '\157' + chr(0b110010) + '\062' + chr(551 - 502), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + chr(365 - 313) + chr(51), 8), ehT0Px3KOsy9(chr(814 - 766) + '\x6f' + '\x31' + chr(0b110010) + chr(54), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b111 + 0o54) + chr(0b1 + 0o60) + chr(0b1001 + 0o52), 0o10), ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(0b11001 + 0o126) + chr(771 - 720) + chr(0b110000) + '\x34', 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110010) + chr(1647 - 1593) + chr(1670 - 1621), ord("\x08")), ehT0Px3KOsy9(chr(50 - 2) + '\x6f' + chr(51) + chr(1793 - 1743), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\062' + chr(0b1011 + 0o54) + '\x37', 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + chr(48) + '\x33', 8)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(8055 - 7944) + '\x35' + '\060', 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe'), chr(100) + chr(0b1100101) + chr(99) + chr(111) + '\144' + '\145')('\x75' + '\x74' + chr(102) + chr(45) + chr(0b10001 + 0o47)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def dJ1eMLj1f5p6(xIEmRseySp3z, n4ljua2gi1Pr):
FGNRR6U_Delb = WOnrfm4dlYcf.embedding_to_padding(xIEmRseySp3z)
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xbe"\xeb1\xac5\x1aXk\x8d'), '\144' + chr(0b0 + 0o145) + '\x63' + '\157' + '\144' + chr(4764 - 4663))('\165' + '\164' + chr(0b1000011 + 0o43) + '\x2d' + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xa0"\xe2\x0b\x81#\x14Xm\x8d\xd9'), chr(0b1000011 + 0o41) + chr(5162 - 5061) + chr(0b1100011) + '\157' + chr(0b1100100) + chr(0b1100101))(chr(5944 - 5827) + '\x74' + chr(102) + '\x2d' + chr(0b111000))):
bLDzE_zU4vXa = mpdtyez0NuRm.PadRemover(FGNRR6U_Delb)
if xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b"\xa11\xd9c\x9f'3^i\xa9\xc5\xa8"), chr(0b1100100) + '\145' + '\143' + chr(6051 - 5940) + chr(6716 - 6616) + chr(0b1100101))(chr(117) + chr(0b1110100) + '\x66' + chr(1943 - 1898) + '\070')) == xafqLlk3kkUe(SXOLrMavuUCe(b'\xa01\xe3$\x96(\x1dhr\x86\xdb\xef\xdcgw\x13b\xa8\x00T\xf8\xdd\x90\x06\x8b\xfc\xd6F\xe0'), '\144' + '\x65' + '\143' + chr(0b1101111) + '\x64' + chr(0b100110 + 0o77))(chr(0b10100 + 0o141) + chr(116) + chr(0b1100110) + chr(45) + chr(0b111000)):
Z0c2rFCFDCFc = WOnrfm4dlYcf.attention_bias_prepend_inputs_full_attention(FGNRR6U_Delb)
else:
Z0c2rFCFDCFc = WOnrfm4dlYcf.attention_bias_lower_triangle(IDJ2eXGCBCDu.nauYfLglTpcb(xIEmRseySp3z)[ehT0Px3KOsy9(chr(667 - 619) + chr(8405 - 8294) + chr(0b100000 + 0o21), ord("\x08"))])
t5Jz9byuSQ65 = jSKPaHwSAfVv.shift_right_3d(xIEmRseySp3z)
if xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x9e\x1b\xe2d\x927 }\x7f\xdc\xc7\xd1'), chr(2876 - 2776) + '\145' + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + '\145')(chr(0b1110101) + chr(0b1000111 + 0o55) + '\146' + '\x2d' + '\070')) == xafqLlk3kkUe(SXOLrMavuUCe(b'\xa4*\xeb=\x9d!'), '\144' + '\x65' + '\143' + chr(8646 - 8535) + chr(0b1010000 + 0o24) + chr(4023 - 3922))(chr(117) + chr(601 - 485) + '\146' + chr(0b101101) + '\x38'):
t5Jz9byuSQ65 = WOnrfm4dlYcf.add_timing_signal_1d(t5Jz9byuSQ65)
return (t5Jz9byuSQ65, Z0c2rFCFDCFc, bLDzE_zU4vXa)
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
get_batch_coordinate
|
def get_batch_coordinate(x, axis=0):
"""Return a flat int32 tensor of shape [1, batch_size*length, 1]."""
# Compute the batch coordinate before flattening all batches
batch_coordinate = tf.expand_dims(
common_attention.coordinate_tensor(tf.shape(x)[:-1], axis=axis), axis=-1)
return batch_coordinate
|
python
|
def get_batch_coordinate(x, axis=0):
"""Return a flat int32 tensor of shape [1, batch_size*length, 1]."""
# Compute the batch coordinate before flattening all batches
batch_coordinate = tf.expand_dims(
common_attention.coordinate_tensor(tf.shape(x)[:-1], axis=axis), axis=-1)
return batch_coordinate
|
[
"def",
"get_batch_coordinate",
"(",
"x",
",",
"axis",
"=",
"0",
")",
":",
"# Compute the batch coordinate before flattening all batches",
"batch_coordinate",
"=",
"tf",
".",
"expand_dims",
"(",
"common_attention",
".",
"coordinate_tensor",
"(",
"tf",
".",
"shape",
"(",
"x",
")",
"[",
":",
"-",
"1",
"]",
",",
"axis",
"=",
"axis",
")",
",",
"axis",
"=",
"-",
"1",
")",
"return",
"batch_coordinate"
] |
Return a flat int32 tensor of shape [1, batch_size*length, 1].
|
[
"Return",
"a",
"flat",
"int32",
"tensor",
"of",
"shape",
"[",
"1",
"batch_size",
"*",
"length",
"1",
"]",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L368-L373
|
train
|
Return a flat int32 tensor of shape [ 1 batch_size length 1 ).
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(1261 - 1213) + '\157' + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + chr(2005 - 1894) + chr(0b110010) + '\x30' + '\065', 0b1000), ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(0b1101111) + chr(51) + chr(55) + chr(0b11111 + 0o25), 0o10), ehT0Px3KOsy9(chr(2130 - 2082) + '\157' + '\x36' + '\065', 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + chr(0b111 + 0o57), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(1767 - 1718) + chr(54) + chr(51), 50658 - 50650), ehT0Px3KOsy9(chr(48) + chr(0b1100111 + 0o10) + chr(0b110010) + '\065' + chr(228 - 177), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1001010 + 0o45) + chr(53) + '\x37', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(224 - 173) + '\x33' + '\062', 0b1000), ehT0Px3KOsy9(chr(929 - 881) + '\x6f' + chr(0b110011) + '\064' + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110111) + '\064', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x37' + chr(0b110100), 8), ehT0Px3KOsy9('\x30' + chr(10683 - 10572) + chr(0b110010) + '\x37' + '\067', 37188 - 37180), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b1000000 + 0o57) + chr(0b100 + 0o61) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(896 - 845) + '\062' + chr(48), 21403 - 21395), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(2419 - 2368) + chr(2416 - 2361) + chr(2422 - 2370), 8), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b101111 + 0o2) + '\067' + chr(1845 - 1792), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(49) + chr(1737 - 1682) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(442 - 394) + '\157' + chr(1605 - 1556) + chr(2601 - 2550) + '\065', 0b1000), ehT0Px3KOsy9(chr(48) + chr(3868 - 3757) + chr(0b111 + 0o54) + chr(0b1011 + 0o45) + chr(54), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\063' + '\066' + chr(542 - 491), 44557 - 44549), ehT0Px3KOsy9(chr(0b1 + 0o57) + '\157' + '\x32' + chr(0b110110) + chr(0b101111 + 0o4), 63026 - 63018), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51) + chr(1990 - 1935) + '\x32', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(51) + '\x36' + chr(0b101000 + 0o16), 32609 - 32601), ehT0Px3KOsy9('\x30' + chr(3990 - 3879) + '\x33' + chr(53) + chr(55), 0o10), ehT0Px3KOsy9(chr(0b10011 + 0o35) + '\x6f' + chr(0b1110 + 0o44) + chr(2451 - 2401), 0b1000), ehT0Px3KOsy9(chr(1295 - 1247) + chr(111) + chr(1491 - 1439) + '\x31', 52501 - 52493), ehT0Px3KOsy9(chr(883 - 835) + '\157' + chr(0b11110 + 0o24) + chr(2178 - 2124) + chr(1057 - 1005), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x36' + '\x30', 6899 - 6891), ehT0Px3KOsy9(chr(48) + '\x6f' + '\061' + chr(0b1010 + 0o53) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100001 + 0o16) + chr(50) + '\065' + chr(677 - 622), ord("\x08")), ehT0Px3KOsy9(chr(2208 - 2160) + chr(0b1100011 + 0o14) + chr(1764 - 1715) + chr(0b10101 + 0o36) + chr(54), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + chr(0b110110) + chr(0b11011 + 0o31), 36503 - 36495), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b0 + 0o61) + chr(53) + chr(50), 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(51) + chr(0b110101 + 0o2) + '\x37', 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\x32' + chr(0b1010 + 0o47) + chr(641 - 586), 0o10), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\157' + '\x33' + chr(2474 - 2422) + chr(0b110010), 45827 - 45819), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(0b1101111) + chr(1025 - 975) + chr(0b110010) + '\x31', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(127 - 77) + chr(55) + chr(0b10100 + 0o36), 0o10), ehT0Px3KOsy9('\060' + chr(0b111100 + 0o63) + chr(0b1101 + 0o44) + chr(0b100011 + 0o15) + chr(0b110100), 61551 - 61543)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(111) + chr(347 - 294) + '\060', 32891 - 32883)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'+'), '\x64' + '\145' + chr(99) + chr(8307 - 8196) + chr(7041 - 6941) + '\x65')('\x75' + chr(1237 - 1121) + chr(102) + '\055' + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def X86yhumeavQI(OeWW0F1dBPRQ, cRTh61qyvi24=ehT0Px3KOsy9('\060' + chr(111) + '\060', 0b1000)):
E6EllSXMFF9I = IDJ2eXGCBCDu.expand_dims(WOnrfm4dlYcf.coordinate_tensor(IDJ2eXGCBCDu.nauYfLglTpcb(OeWW0F1dBPRQ)[:-ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(49), ord("\x08"))], axis=cRTh61qyvi24), axis=-ehT0Px3KOsy9(chr(2129 - 2081) + '\x6f' + chr(0b101101 + 0o4), 8))
return E6EllSXMFF9I
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
expand_batch_coordinates
|
def expand_batch_coordinates(bc, length_factor):
"""Duplicate elements of bc by length_factor.
Args:
bc (tf.Tensor): int32 tensor of shape [1, length, 1]
length_factor (int):
Returns:
tf.Tensor: of shape [1, length*length_factor, 1] where every elements has
been duplicated length_factor times.
"""
assert bc.get_shape().as_list() == [1, None, 1]
# bc has shape [1, length, 1]
bc *= tf.constant([[1] * length_factor])
# bc has shape [1, length, length_factor]
bc = tf.reshape(bc, [1, -1, 1])
# bc has shape [1, length*length_factor]
return bc
|
python
|
def expand_batch_coordinates(bc, length_factor):
"""Duplicate elements of bc by length_factor.
Args:
bc (tf.Tensor): int32 tensor of shape [1, length, 1]
length_factor (int):
Returns:
tf.Tensor: of shape [1, length*length_factor, 1] where every elements has
been duplicated length_factor times.
"""
assert bc.get_shape().as_list() == [1, None, 1]
# bc has shape [1, length, 1]
bc *= tf.constant([[1] * length_factor])
# bc has shape [1, length, length_factor]
bc = tf.reshape(bc, [1, -1, 1])
# bc has shape [1, length*length_factor]
return bc
|
[
"def",
"expand_batch_coordinates",
"(",
"bc",
",",
"length_factor",
")",
":",
"assert",
"bc",
".",
"get_shape",
"(",
")",
".",
"as_list",
"(",
")",
"==",
"[",
"1",
",",
"None",
",",
"1",
"]",
"# bc has shape [1, length, 1]",
"bc",
"*=",
"tf",
".",
"constant",
"(",
"[",
"[",
"1",
"]",
"*",
"length_factor",
"]",
")",
"# bc has shape [1, length, length_factor]",
"bc",
"=",
"tf",
".",
"reshape",
"(",
"bc",
",",
"[",
"1",
",",
"-",
"1",
",",
"1",
"]",
")",
"# bc has shape [1, length*length_factor]",
"return",
"bc"
] |
Duplicate elements of bc by length_factor.
Args:
bc (tf.Tensor): int32 tensor of shape [1, length, 1]
length_factor (int):
Returns:
tf.Tensor: of shape [1, length*length_factor, 1] where every elements has
been duplicated length_factor times.
|
[
"Duplicate",
"elements",
"of",
"bc",
"by",
"length_factor",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L377-L394
|
train
|
Duplicate elements of bc by length_factor.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(10577 - 10466) + '\063' + chr(0b110100) + '\x37', 0o10), ehT0Px3KOsy9(chr(0b110000 + 0o0) + '\157' + chr(0b110001) + '\x36' + chr(454 - 402), 53289 - 53281), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110001) + '\061', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x32' + chr(0b10001 + 0o46), 62008 - 62000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110010) + chr(0b110001 + 0o4) + chr(126 - 74), 20986 - 20978), ehT0Px3KOsy9(chr(48) + chr(0b1010111 + 0o30) + chr(51) + '\063' + '\x35', 0o10), ehT0Px3KOsy9(chr(48) + chr(6347 - 6236) + chr(0b10010 + 0o45) + chr(55), 47977 - 47969), ehT0Px3KOsy9(chr(0b1000 + 0o50) + '\157' + chr(0b110101) + '\x30', 47553 - 47545), ehT0Px3KOsy9(chr(685 - 637) + chr(111) + '\061' + chr(0b1011 + 0o54) + chr(48), 4528 - 4520), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b100001 + 0o20) + chr(0b110100) + '\x37', 0o10), ehT0Px3KOsy9('\060' + chr(6234 - 6123) + '\061' + '\x35' + chr(1838 - 1787), 0b1000), ehT0Px3KOsy9(chr(1165 - 1117) + chr(111) + chr(0b110011) + chr(124 - 71) + '\061', 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(0b110001) + chr(0b10 + 0o62), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b111 + 0o150) + chr(2245 - 2192) + '\x35', ord("\x08")), ehT0Px3KOsy9(chr(832 - 784) + chr(5321 - 5210) + '\x33' + '\064' + chr(0b10001 + 0o41), 53707 - 53699), ehT0Px3KOsy9(chr(0b11111 + 0o21) + chr(0b1101010 + 0o5) + '\064' + chr(0b100100 + 0o23), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(6485 - 6374) + chr(0b110011) + '\061' + '\x31', 0o10), ehT0Px3KOsy9(chr(1711 - 1663) + chr(597 - 486) + chr(0b110111) + chr(1246 - 1191), 8), ehT0Px3KOsy9('\060' + '\157' + chr(0b1 + 0o60) + chr(284 - 235) + chr(0b110001), 0o10), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(0b1101111) + chr(1347 - 1293) + chr(0b110010), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + chr(52) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b100011 + 0o16) + chr(0b101 + 0o55) + chr(0b1111 + 0o47), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b1001 + 0o52) + chr(2143 - 2091) + chr(0b11000 + 0o33), 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\063' + chr(48) + chr(358 - 306), 2326 - 2318), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(6165 - 6054) + chr(0b100011 + 0o16) + chr(50), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010) + chr(0b11111 + 0o22) + chr(0b110100 + 0o2), 0o10), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(1503 - 1392) + '\x34' + '\x37', 8), ehT0Px3KOsy9('\x30' + '\x6f' + '\x31' + chr(0b110000) + chr(55), 0o10), ehT0Px3KOsy9(chr(981 - 933) + chr(0b100110 + 0o111) + chr(0b110010) + chr(0b11001 + 0o36) + chr(0b110100 + 0o2), ord("\x08")), ehT0Px3KOsy9(chr(1414 - 1366) + chr(0b111 + 0o150) + chr(0b110000 + 0o2) + '\x32' + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(0b11101 + 0o23) + chr(8804 - 8693) + '\061' + '\x35' + '\x33', 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110001) + '\062', 8), ehT0Px3KOsy9('\060' + chr(9685 - 9574) + '\062' + chr(0b110101) + '\063', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(2131 - 2082) + '\x31' + chr(911 - 860), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(2254 - 2205) + '\x30' + '\x35', 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\066' + chr(52), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\x37' + chr(0b10 + 0o62), ord("\x08")), ehT0Px3KOsy9(chr(761 - 713) + chr(111) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(0b1000 + 0o50) + chr(111) + '\x31' + chr(364 - 310), 0o10), ehT0Px3KOsy9(chr(1851 - 1803) + '\157' + chr(0b110010) + chr(0b10000 + 0o43) + chr(0b100 + 0o55), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(123 - 75) + '\157' + chr(1582 - 1529) + chr(247 - 199), 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xc9'), chr(0b1001110 + 0o26) + '\x65' + '\143' + '\x6f' + '\x64' + chr(0b100110 + 0o77))(chr(0b1011011 + 0o32) + chr(0b1110100) + chr(4438 - 4336) + '\055' + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def i4lCvMPvMKb7(qDf1g1be8s4I, FXBPbXkod7y2):
assert xafqLlk3kkUe(qDf1g1be8s4I.get_shape(), xafqLlk3kkUe(SXOLrMavuUCe(b'\x86\x07V{\xc5\xaeK'), chr(100) + chr(0b101101 + 0o70) + '\143' + chr(0b1101111) + chr(100) + '\145')(chr(9975 - 9858) + chr(0b1100010 + 0o22) + '\x66' + chr(0b1111 + 0o36) + '\070'))() == [ehT0Px3KOsy9(chr(749 - 701) + '\157' + chr(0b101111 + 0o2), 0b1000), None, ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b101000 + 0o11), 8)]
qDf1g1be8s4I *= IDJ2eXGCBCDu.constant([[ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\061', 8)] * FXBPbXkod7y2])
qDf1g1be8s4I = IDJ2eXGCBCDu.reshape(qDf1g1be8s4I, [ehT0Px3KOsy9(chr(48) + '\x6f' + '\061', 8), -ehT0Px3KOsy9(chr(1021 - 973) + '\x6f' + '\x31', 8), ehT0Px3KOsy9(chr(48) + chr(7988 - 7877) + chr(0b110001), 8)])
return qDf1g1be8s4I
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
remove_pad
|
def remove_pad(x, pad_remover, mode):
"""Remove padding by concatenating all dimension into one.
Args:
x (tf.Tensor): input of shape [batch_size, length, depth]
pad_remover (obj): a PadRemover object
mode (ModeKeys): infer, train or eval. If inference, the padding remover is
not applied
Returns:
tf.Tensor of shape [1,length_nonpad,depth] where
length_nonpad <= batch_size*length
"""
# Concatenate all tokens (without padding)
x = expert_utils.flatten_all_but_last(x)
# Remove padding for training and eval
if mode != ModeKeys.PREDICT:
# This is a hack to allows inference when the <go> token
# is detected as padding and removed. This works for now because there is
# no padding at inference.
x = pad_remover.remove(x)
x = tf.expand_dims(x, axis=0) # Now batch_size=1
return x
|
python
|
def remove_pad(x, pad_remover, mode):
"""Remove padding by concatenating all dimension into one.
Args:
x (tf.Tensor): input of shape [batch_size, length, depth]
pad_remover (obj): a PadRemover object
mode (ModeKeys): infer, train or eval. If inference, the padding remover is
not applied
Returns:
tf.Tensor of shape [1,length_nonpad,depth] where
length_nonpad <= batch_size*length
"""
# Concatenate all tokens (without padding)
x = expert_utils.flatten_all_but_last(x)
# Remove padding for training and eval
if mode != ModeKeys.PREDICT:
# This is a hack to allows inference when the <go> token
# is detected as padding and removed. This works for now because there is
# no padding at inference.
x = pad_remover.remove(x)
x = tf.expand_dims(x, axis=0) # Now batch_size=1
return x
|
[
"def",
"remove_pad",
"(",
"x",
",",
"pad_remover",
",",
"mode",
")",
":",
"# Concatenate all tokens (without padding)",
"x",
"=",
"expert_utils",
".",
"flatten_all_but_last",
"(",
"x",
")",
"# Remove padding for training and eval",
"if",
"mode",
"!=",
"ModeKeys",
".",
"PREDICT",
":",
"# This is a hack to allows inference when the <go> token",
"# is detected as padding and removed. This works for now because there is",
"# no padding at inference.",
"x",
"=",
"pad_remover",
".",
"remove",
"(",
"x",
")",
"x",
"=",
"tf",
".",
"expand_dims",
"(",
"x",
",",
"axis",
"=",
"0",
")",
"# Now batch_size=1",
"return",
"x"
] |
Remove padding by concatenating all dimension into one.
Args:
x (tf.Tensor): input of shape [batch_size, length, depth]
pad_remover (obj): a PadRemover object
mode (ModeKeys): infer, train or eval. If inference, the padding remover is
not applied
Returns:
tf.Tensor of shape [1,length_nonpad,depth] where
length_nonpad <= batch_size*length
|
[
"Remove",
"padding",
"by",
"concatenating",
"all",
"dimension",
"into",
"one",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L398-L422
|
train
|
Remove padding from the last batch of tokens.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + '\x6f' + chr(2007 - 1958) + chr(2339 - 2288) + chr(0b11111 + 0o27), 29359 - 29351), ehT0Px3KOsy9(chr(0b101001 + 0o7) + '\x6f' + chr(2151 - 2102) + chr(0b110 + 0o56) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110011) + chr(0b10001 + 0o46) + chr(1834 - 1785), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(8229 - 8118) + chr(0b110001) + chr(0b110010), 818 - 810), ehT0Px3KOsy9('\060' + chr(0b1000110 + 0o51) + chr(0b110001) + chr(53) + '\064', 0b1000), ehT0Px3KOsy9(chr(807 - 759) + '\157' + chr(0b110001) + chr(51) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011) + '\x31' + chr(1509 - 1461), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b101001 + 0o11) + '\063' + chr(1670 - 1615), 9982 - 9974), ehT0Px3KOsy9('\x30' + chr(111) + '\062' + chr(53) + chr(1332 - 1284), 64556 - 64548), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(156 - 107) + chr(49), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(0b11101 + 0o31) + '\x32', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b10001 + 0o43) + '\x34', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(50) + chr(0b110010) + chr(54), 51256 - 51248), ehT0Px3KOsy9(chr(0b1111 + 0o41) + '\157' + '\067' + chr(0b110101), 9843 - 9835), ehT0Px3KOsy9(chr(1378 - 1330) + '\x6f' + '\064' + chr(1321 - 1270), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b101100 + 0o103) + '\061' + '\x30' + '\060', 63924 - 63916), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110111) + chr(49), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1000101 + 0o52) + '\x31' + chr(53) + chr(0b11010 + 0o27), ord("\x08")), ehT0Px3KOsy9('\060' + chr(3605 - 3494) + chr(0b101010 + 0o7) + chr(151 - 99) + chr(54), 0o10), ehT0Px3KOsy9('\x30' + chr(0b111111 + 0o60) + '\063' + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b11010 + 0o27) + chr(2737 - 2684), 0o10), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(111) + chr(51) + chr(0b101000 + 0o16) + chr(0b1001 + 0o52), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(2106 - 2056) + '\063' + chr(54), 0o10), ehT0Px3KOsy9(chr(48) + chr(6477 - 6366) + chr(0b110001) + chr(654 - 605) + '\x31', 0b1000), ehT0Px3KOsy9(chr(214 - 166) + chr(0b1101111) + chr(0b101 + 0o56) + chr(1235 - 1187) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(4542 - 4431) + chr(0b110010) + '\060' + chr(921 - 867), 16734 - 16726), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(49) + '\064' + chr(55), 10067 - 10059), ehT0Px3KOsy9('\x30' + '\x6f' + '\x32' + chr(50) + chr(0b11001 + 0o35), 8), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(0b1011000 + 0o27) + '\x32' + chr(1291 - 1240) + '\x35', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x31' + '\x30' + chr(0b0 + 0o67), 0o10), ehT0Px3KOsy9(chr(702 - 654) + '\x6f' + '\061' + chr(0b1011 + 0o46) + chr(55), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\064' + chr(865 - 811), ord("\x08")), ehT0Px3KOsy9('\060' + chr(614 - 503) + '\x31' + chr(0b11001 + 0o33) + chr(1456 - 1401), 8), ehT0Px3KOsy9(chr(48) + chr(0b1011000 + 0o27) + chr(1382 - 1333) + chr(0b111 + 0o55) + chr(0b10111 + 0o36), 8), ehT0Px3KOsy9(chr(0b101 + 0o53) + chr(0b1101111) + '\063' + chr(682 - 631) + chr(389 - 341), 0b1000), ehT0Px3KOsy9(chr(663 - 615) + '\x6f' + '\x31' + chr(0b110001), 8), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(0b1101111) + chr(50) + chr(0b100001 + 0o25) + chr(1715 - 1663), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(895 - 846) + '\x31' + '\062', 0o10), ehT0Px3KOsy9(chr(48) + chr(10048 - 9937) + '\061' + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(54) + chr(1804 - 1753), 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(4710 - 4599) + '\065' + chr(0b1010 + 0o46), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'p'), chr(0b1100100) + chr(0b100111 + 0o76) + '\x63' + '\157' + chr(100) + chr(0b1100101))(chr(0b1110101) + '\x74' + '\x66' + chr(0b100000 + 0o15) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def vXjaekrEI1UA(OeWW0F1dBPRQ, bLDzE_zU4vXa, holLFgwB7vsP):
OeWW0F1dBPRQ = mpdtyez0NuRm.flatten_all_but_last(OeWW0F1dBPRQ)
if holLFgwB7vsP != xafqLlk3kkUe(rwXm4rEGJno8, xafqLlk3kkUe(SXOLrMavuUCe(b'\x0e7\xbcN%\x1b\x9c'), chr(0b1011110 + 0o6) + chr(101) + chr(0b1100011) + '\157' + chr(3949 - 3849) + chr(4138 - 4037))(chr(0b1110101) + chr(116) + '\146' + chr(0b10101 + 0o30) + '\x38')):
OeWW0F1dBPRQ = bLDzE_zU4vXa.remove(OeWW0F1dBPRQ)
OeWW0F1dBPRQ = IDJ2eXGCBCDu.expand_dims(OeWW0F1dBPRQ, axis=ehT0Px3KOsy9('\x30' + chr(111) + '\x30', 0o10))
return OeWW0F1dBPRQ
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_base
|
def attention_lm_moe_base():
"""Set of hyperparameters.
suitable for 1 gpu.
on lm1b_32k:
~229M params
0.9 steps/sec on [GeForce GTX TITAN X]
Returns:
a hparams object
"""
hparams = common_hparams.basic_params1()
hparams.hidden_size = 1024
hparams.batch_size = 8192
hparams.max_length = 256
hparams.dropout = 0.0
hparams.clip_grad_norm = 0. # i.e. no gradient clipping
hparams.optimizer_adam_epsilon = 1e-9
hparams.learning_rate_decay_scheme = "noam"
hparams.learning_rate = 0.1
hparams.learning_rate_warmup_steps = 2000
hparams.initializer_gain = 1.0
hparams.num_hidden_layers = 4
hparams.initializer = "uniform_unit_scaling"
hparams.weight_decay = 0.0
hparams.optimizer_adam_beta1 = 0.9
hparams.optimizer_adam_beta2 = 0.98
hparams.num_sampled_classes = 0
hparams.label_smoothing = 0.0
hparams.shared_embedding_and_softmax_weights = False
hparams.add_hparam("filter_size", 2048) # Add new ones like this.
hparams.moe_num_experts = 32
# attention-related flags
hparams.add_hparam("num_heads", 8)
hparams.add_hparam("attention_key_channels", 0)
hparams.add_hparam("attention_value_channels", 0)
# All hyperparameters ending in "dropout" are automatically set to 0.0
# when not in training mode.
hparams.add_hparam("attention_dropout", 0.0)
hparams.add_hparam("relu_dropout", 0.0)
hparams.add_hparam("pos", "timing") # timing, none
hparams.add_hparam("moe_layers", "2") # comma separated list of layer numbers
# moe params. local attention moe.
# If attention_layers is set, the num_hidden_layers parameter will be ignored
# and each caracter of the string will correspond to one attention
# layer type
hparams.add_hparam("attention_layers", "")
hparams.add_hparam("attention_type", AttentionType.MULTIHEAD)
hparams.add_hparam("attention_local", False)
hparams.add_hparam("attention_moe_k", 2)
hparams.add_hparam("attention_num_head", 1)
hparams.add_hparam("attention_num_experts", 16)
hparams.add_hparam("attention_split_batch", False)
hparams.add_hparam("attention_red_factor", 3)
hparams.add_hparam("attention_block_length", 128)
hparams.add_hparam("attention_reduction_type", "conv")
# Non linearity for the attention reduction. Either "none", or "silu" (
# Sigmoid Linear-Unit described in https://arxiv.org/abs/1710.05941)
hparams.add_hparam("attention_nonlinearity", "none")
# If attention_exp_factor is set, each input to local_expert_attention (of
# dimensionality hidden size) is projected into attention_exp_factor smaller
# inputs, each of dimensionality attention_exp_inputdim. (otherwise
# attention_exp_inputdim is ignored)
hparams.add_hparam("attention_exp_factor", 0)
hparams.add_hparam("attention_exp_inputdim", 128)
# Key, query and value dimensions for the attention
hparams.add_hparam("attention_kq_size", 128)
hparams.add_hparam("attention_v_size", 256)
# Loss coef for load balancing
hparams.add_hparam("attention_load_balance", 2e-2)
# Locality-sensitive hashing params
hparams.add_hparam("lsh_num_hyperplanes", 4)
hparams.add_hparam("lsh_use_map_fn", False)
hparams.add_hparam("use_sepconv", False)
hparams.add_hparam("diet_experts", False)
hparams.add_hparam("memory_efficient_ffn", False)
# if True, we learn a non-autoregressive model from "inputs" to "targets".
# if False, we learn an autoregressive model to generate "targets"
hparams.add_hparam("use_inputs", False)
return hparams
|
python
|
def attention_lm_moe_base():
"""Set of hyperparameters.
suitable for 1 gpu.
on lm1b_32k:
~229M params
0.9 steps/sec on [GeForce GTX TITAN X]
Returns:
a hparams object
"""
hparams = common_hparams.basic_params1()
hparams.hidden_size = 1024
hparams.batch_size = 8192
hparams.max_length = 256
hparams.dropout = 0.0
hparams.clip_grad_norm = 0. # i.e. no gradient clipping
hparams.optimizer_adam_epsilon = 1e-9
hparams.learning_rate_decay_scheme = "noam"
hparams.learning_rate = 0.1
hparams.learning_rate_warmup_steps = 2000
hparams.initializer_gain = 1.0
hparams.num_hidden_layers = 4
hparams.initializer = "uniform_unit_scaling"
hparams.weight_decay = 0.0
hparams.optimizer_adam_beta1 = 0.9
hparams.optimizer_adam_beta2 = 0.98
hparams.num_sampled_classes = 0
hparams.label_smoothing = 0.0
hparams.shared_embedding_and_softmax_weights = False
hparams.add_hparam("filter_size", 2048) # Add new ones like this.
hparams.moe_num_experts = 32
# attention-related flags
hparams.add_hparam("num_heads", 8)
hparams.add_hparam("attention_key_channels", 0)
hparams.add_hparam("attention_value_channels", 0)
# All hyperparameters ending in "dropout" are automatically set to 0.0
# when not in training mode.
hparams.add_hparam("attention_dropout", 0.0)
hparams.add_hparam("relu_dropout", 0.0)
hparams.add_hparam("pos", "timing") # timing, none
hparams.add_hparam("moe_layers", "2") # comma separated list of layer numbers
# moe params. local attention moe.
# If attention_layers is set, the num_hidden_layers parameter will be ignored
# and each caracter of the string will correspond to one attention
# layer type
hparams.add_hparam("attention_layers", "")
hparams.add_hparam("attention_type", AttentionType.MULTIHEAD)
hparams.add_hparam("attention_local", False)
hparams.add_hparam("attention_moe_k", 2)
hparams.add_hparam("attention_num_head", 1)
hparams.add_hparam("attention_num_experts", 16)
hparams.add_hparam("attention_split_batch", False)
hparams.add_hparam("attention_red_factor", 3)
hparams.add_hparam("attention_block_length", 128)
hparams.add_hparam("attention_reduction_type", "conv")
# Non linearity for the attention reduction. Either "none", or "silu" (
# Sigmoid Linear-Unit described in https://arxiv.org/abs/1710.05941)
hparams.add_hparam("attention_nonlinearity", "none")
# If attention_exp_factor is set, each input to local_expert_attention (of
# dimensionality hidden size) is projected into attention_exp_factor smaller
# inputs, each of dimensionality attention_exp_inputdim. (otherwise
# attention_exp_inputdim is ignored)
hparams.add_hparam("attention_exp_factor", 0)
hparams.add_hparam("attention_exp_inputdim", 128)
# Key, query and value dimensions for the attention
hparams.add_hparam("attention_kq_size", 128)
hparams.add_hparam("attention_v_size", 256)
# Loss coef for load balancing
hparams.add_hparam("attention_load_balance", 2e-2)
# Locality-sensitive hashing params
hparams.add_hparam("lsh_num_hyperplanes", 4)
hparams.add_hparam("lsh_use_map_fn", False)
hparams.add_hparam("use_sepconv", False)
hparams.add_hparam("diet_experts", False)
hparams.add_hparam("memory_efficient_ffn", False)
# if True, we learn a non-autoregressive model from "inputs" to "targets".
# if False, we learn an autoregressive model to generate "targets"
hparams.add_hparam("use_inputs", False)
return hparams
|
[
"def",
"attention_lm_moe_base",
"(",
")",
":",
"hparams",
"=",
"common_hparams",
".",
"basic_params1",
"(",
")",
"hparams",
".",
"hidden_size",
"=",
"1024",
"hparams",
".",
"batch_size",
"=",
"8192",
"hparams",
".",
"max_length",
"=",
"256",
"hparams",
".",
"dropout",
"=",
"0.0",
"hparams",
".",
"clip_grad_norm",
"=",
"0.",
"# i.e. no gradient clipping",
"hparams",
".",
"optimizer_adam_epsilon",
"=",
"1e-9",
"hparams",
".",
"learning_rate_decay_scheme",
"=",
"\"noam\"",
"hparams",
".",
"learning_rate",
"=",
"0.1",
"hparams",
".",
"learning_rate_warmup_steps",
"=",
"2000",
"hparams",
".",
"initializer_gain",
"=",
"1.0",
"hparams",
".",
"num_hidden_layers",
"=",
"4",
"hparams",
".",
"initializer",
"=",
"\"uniform_unit_scaling\"",
"hparams",
".",
"weight_decay",
"=",
"0.0",
"hparams",
".",
"optimizer_adam_beta1",
"=",
"0.9",
"hparams",
".",
"optimizer_adam_beta2",
"=",
"0.98",
"hparams",
".",
"num_sampled_classes",
"=",
"0",
"hparams",
".",
"label_smoothing",
"=",
"0.0",
"hparams",
".",
"shared_embedding_and_softmax_weights",
"=",
"False",
"hparams",
".",
"add_hparam",
"(",
"\"filter_size\"",
",",
"2048",
")",
"# Add new ones like this.",
"hparams",
".",
"moe_num_experts",
"=",
"32",
"# attention-related flags",
"hparams",
".",
"add_hparam",
"(",
"\"num_heads\"",
",",
"8",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_key_channels\"",
",",
"0",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_value_channels\"",
",",
"0",
")",
"# All hyperparameters ending in \"dropout\" are automatically set to 0.0",
"# when not in training mode.",
"hparams",
".",
"add_hparam",
"(",
"\"attention_dropout\"",
",",
"0.0",
")",
"hparams",
".",
"add_hparam",
"(",
"\"relu_dropout\"",
",",
"0.0",
")",
"hparams",
".",
"add_hparam",
"(",
"\"pos\"",
",",
"\"timing\"",
")",
"# timing, none",
"hparams",
".",
"add_hparam",
"(",
"\"moe_layers\"",
",",
"\"2\"",
")",
"# comma separated list of layer numbers",
"# moe params. local attention moe.",
"# If attention_layers is set, the num_hidden_layers parameter will be ignored",
"# and each caracter of the string will correspond to one attention",
"# layer type",
"hparams",
".",
"add_hparam",
"(",
"\"attention_layers\"",
",",
"\"\"",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_type\"",
",",
"AttentionType",
".",
"MULTIHEAD",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_local\"",
",",
"False",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_moe_k\"",
",",
"2",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_num_head\"",
",",
"1",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_num_experts\"",
",",
"16",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_split_batch\"",
",",
"False",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_red_factor\"",
",",
"3",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_block_length\"",
",",
"128",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_reduction_type\"",
",",
"\"conv\"",
")",
"# Non linearity for the attention reduction. Either \"none\", or \"silu\" (",
"# Sigmoid Linear-Unit described in https://arxiv.org/abs/1710.05941)",
"hparams",
".",
"add_hparam",
"(",
"\"attention_nonlinearity\"",
",",
"\"none\"",
")",
"# If attention_exp_factor is set, each input to local_expert_attention (of",
"# dimensionality hidden size) is projected into attention_exp_factor smaller",
"# inputs, each of dimensionality attention_exp_inputdim. (otherwise",
"# attention_exp_inputdim is ignored)",
"hparams",
".",
"add_hparam",
"(",
"\"attention_exp_factor\"",
",",
"0",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_exp_inputdim\"",
",",
"128",
")",
"# Key, query and value dimensions for the attention",
"hparams",
".",
"add_hparam",
"(",
"\"attention_kq_size\"",
",",
"128",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_v_size\"",
",",
"256",
")",
"# Loss coef for load balancing",
"hparams",
".",
"add_hparam",
"(",
"\"attention_load_balance\"",
",",
"2e-2",
")",
"# Locality-sensitive hashing params",
"hparams",
".",
"add_hparam",
"(",
"\"lsh_num_hyperplanes\"",
",",
"4",
")",
"hparams",
".",
"add_hparam",
"(",
"\"lsh_use_map_fn\"",
",",
"False",
")",
"hparams",
".",
"add_hparam",
"(",
"\"use_sepconv\"",
",",
"False",
")",
"hparams",
".",
"add_hparam",
"(",
"\"diet_experts\"",
",",
"False",
")",
"hparams",
".",
"add_hparam",
"(",
"\"memory_efficient_ffn\"",
",",
"False",
")",
"# if True, we learn a non-autoregressive model from \"inputs\" to \"targets\".",
"# if False, we learn an autoregressive model to generate \"targets\"",
"hparams",
".",
"add_hparam",
"(",
"\"use_inputs\"",
",",
"False",
")",
"return",
"hparams"
] |
Set of hyperparameters.
suitable for 1 gpu.
on lm1b_32k:
~229M params
0.9 steps/sec on [GeForce GTX TITAN X]
Returns:
a hparams object
|
[
"Set",
"of",
"hyperparameters",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L435-L515
|
train
|
Set of hyperparameters suitable for 1 gpu.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + '\x6f' + chr(50) + chr(50) + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b101110 + 0o101) + chr(0b110011) + '\063' + '\065', 24104 - 24096), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + '\x36' + chr(476 - 427), 52719 - 52711), ehT0Px3KOsy9(chr(48) + '\157' + chr(1213 - 1164) + '\x37' + chr(0b110100), 25320 - 25312), ehT0Px3KOsy9('\060' + chr(111) + chr(0b11000 + 0o32) + '\062' + chr(52), 25012 - 25004), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(51) + '\060' + chr(927 - 878), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101 + 0o142) + chr(0b110010 + 0o1) + '\061' + chr(0b101100 + 0o7), 0b1000), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(650 - 539) + chr(1627 - 1576) + chr(0b110101) + '\062', 35530 - 35522), ehT0Px3KOsy9('\060' + chr(111) + chr(50) + chr(1622 - 1574), 53373 - 53365), ehT0Px3KOsy9(chr(2256 - 2208) + chr(0b1101010 + 0o5) + chr(52) + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(5598 - 5487) + chr(51) + '\x35', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + chr(0b101110 + 0o7) + chr(52), 27649 - 27641), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x31' + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(177 - 129) + chr(5345 - 5234) + chr(0b110001) + '\x37' + '\063', 38103 - 38095), ehT0Px3KOsy9(chr(48) + chr(0b1000010 + 0o55) + '\x31' + chr(0b110000), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(51) + '\060' + chr(0b100110 + 0o15), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\061' + chr(0b100101 + 0o16) + chr(50), 0o10), ehT0Px3KOsy9(chr(309 - 261) + chr(0b101111 + 0o100) + chr(51) + chr(48) + chr(49), 8), ehT0Px3KOsy9('\060' + chr(111) + chr(0b101001 + 0o10) + '\x30' + chr(0b101011 + 0o14), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110010) + chr(55) + chr(0b110001), 0o10), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(8246 - 8135) + chr(0b10011 + 0o37) + chr(0b110010) + chr(51), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(54), 0o10), ehT0Px3KOsy9(chr(1506 - 1458) + chr(0b1101111) + '\063' + chr(1677 - 1625) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(0b101 + 0o56) + chr(0b110101), 8), ehT0Px3KOsy9(chr(717 - 669) + chr(111) + chr(54) + '\063', 43787 - 43779), ehT0Px3KOsy9('\x30' + chr(5843 - 5732) + chr(0b10000 + 0o41) + '\x32' + '\x36', 0b1000), ehT0Px3KOsy9(chr(48) + chr(10675 - 10564) + chr(0b101011 + 0o10) + chr(0b110111) + chr(0b100111 + 0o11), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b101110 + 0o5) + chr(0b110010) + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(8415 - 8304) + chr(55) + '\066', 0b1000), ehT0Px3KOsy9(chr(1216 - 1168) + '\x6f' + chr(0b110011) + '\066' + '\x32', 7367 - 7359), ehT0Px3KOsy9(chr(624 - 576) + '\x6f' + chr(0b110100) + chr(1520 - 1469), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100111 + 0o10) + '\x33' + '\066' + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(0b100010 + 0o16) + '\x6f' + '\x31' + chr(53) + chr(0b101110 + 0o3), 1888 - 1880), ehT0Px3KOsy9(chr(0b110000) + chr(0b1000101 + 0o52) + '\x31' + '\x31' + chr(0b111 + 0o55), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(1904 - 1852) + chr(0b110110), 8), ehT0Px3KOsy9(chr(0b110000) + chr(11237 - 11126) + '\x33' + '\x31' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(1725 - 1677) + chr(111) + '\x31' + chr(0b0 + 0o63), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + '\x33' + '\063' + '\x33', 3382 - 3374), ehT0Px3KOsy9(chr(82 - 34) + chr(0b1101111) + '\066' + chr(0b110001), 32184 - 32176), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1477 - 1426) + '\x30' + chr(54), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(8451 - 8340) + '\x35' + chr(0b1011 + 0o45), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xbc'), '\144' + '\145' + chr(99) + '\x6f' + chr(100) + chr(5283 - 5182))(chr(0b101010 + 0o113) + '\x74' + '\146' + chr(45) + chr(308 - 252)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def kKmBjRbC6Zbc():
n4ljua2gi1Pr = vLnG3ZpOXWXZ.basic_params1()
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9('\x30' + chr(111) + '\x32' + '\x30' + chr(0b110000) + chr(48), 25098 - 25090)
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(50) + chr(0b110000) + chr(0b101001 + 0o7) + chr(48) + '\060', 0o10)
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x34' + chr(0b110000) + chr(0b101100 + 0o4), 51387 - 51379)
n4ljua2gi1Pr.ag0mwEgWzjYv = 0.0
n4ljua2gi1Pr.SdNSZNVkVjLh = 0.0
n4ljua2gi1Pr.o17O_bIptWdl = 1e-09
n4ljua2gi1Pr.v3ZnJE9Hdub1 = xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xb2\xd0\x11'), chr(100) + chr(0b1100101) + '\143' + chr(111) + '\x64' + '\x65')(chr(0b1110101) + chr(1865 - 1749) + '\x66' + chr(0b10101 + 0o30) + chr(56))
n4ljua2gi1Pr.QGSIpd_yUNzU = 0.1
n4ljua2gi1Pr.fHyhoyGmdvM9 = ehT0Px3KOsy9(chr(1754 - 1706) + chr(0b1001100 + 0o43) + chr(0b10010 + 0o41) + chr(2485 - 2430) + chr(0b110010) + '\x30', 48254 - 48246)
n4ljua2gi1Pr.S1SbCBXLapw8 = 1.0
n4ljua2gi1Pr.jZh5_pLUoOoZ = ehT0Px3KOsy9(chr(0b110000 + 0o0) + '\x6f' + chr(52), 0b1000)
n4ljua2gi1Pr.kwfuYzkY5C57 = xafqLlk3kkUe(SXOLrMavuUCe(b'\xe7\xb3\xd8\x1a\xfb\xb3\xa6\xb5eHt\x18S\x08k\xa26\xa8)\xde'), chr(100) + chr(0b1001110 + 0o27) + chr(0b101101 + 0o66) + chr(111) + chr(8500 - 8400) + chr(101))(chr(9043 - 8926) + '\x74' + '\x66' + chr(0b100101 + 0o10) + '\x38')
n4ljua2gi1Pr.eB4rJl6fUxw9 = 0.0
n4ljua2gi1Pr.GcOjyd7zcDH8 = 0.9
n4ljua2gi1Pr.CBOVKNT0M9cG = 0.98
n4ljua2gi1Pr.Syf38YGTPvuw = ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(0b1101111) + chr(48), 0o10)
n4ljua2gi1Pr.FSjUgdaczzRk = 0.0
n4ljua2gi1Pr.qVamxim0L2I1 = ehT0Px3KOsy9(chr(1978 - 1930) + chr(0b1001110 + 0o41) + chr(48), 8)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(0b1100100) + '\145' + chr(0b101011 + 0o70) + chr(4227 - 4116) + '\144' + chr(1432 - 1331))('\x75' + chr(0b1110100) + chr(0b1100110) + chr(1731 - 1686) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf4\xb4\xdd\x08\xf1\xb3\x94\x99y\\x'), chr(100) + '\x65' + '\143' + chr(111) + '\x64' + chr(0b1100101))('\165' + '\x74' + '\x66' + '\055' + chr(0b111000)), ehT0Px3KOsy9('\060' + chr(111) + chr(2189 - 2137) + chr(0b1001 + 0o47) + '\060' + chr(0b11011 + 0o25), ord("\x08")))
n4ljua2gi1Pr.r99iQzD4Y8i3 = ehT0Px3KOsy9('\060' + '\157' + chr(2044 - 1992) + chr(48), 0b1000)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\x64' + '\145' + chr(0b1100011) + chr(0b1101111) + chr(100) + chr(0b1100101))(chr(117) + chr(0b1010101 + 0o37) + chr(0b1100110) + chr(1816 - 1771) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xa8\xdc#\xfc\xa4\xaa\x8ec'), chr(0b1100100) + chr(9788 - 9687) + '\143' + chr(12257 - 12146) + '\x64' + chr(0b110110 + 0o57))(chr(8133 - 8016) + '\164' + chr(0b1100110) + '\x2d' + chr(0b100 + 0o64)), ehT0Px3KOsy9('\060' + '\x6f' + chr(496 - 447) + chr(225 - 177), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\144' + '\145' + chr(0b1100011) + chr(0b1101111) + chr(970 - 870) + chr(101))('\x75' + '\164' + '\146' + '\055' + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yv\tu$k\xab;\xaf)\xdc\xffu'), '\x64' + chr(1484 - 1383) + chr(8199 - 8100) + chr(0b11000 + 0o127) + '\x64' + '\145')(chr(0b1110101) + '\x74' + chr(0b1100110) + '\055' + chr(0b100001 + 0o27)), ehT0Px3KOsy9('\060' + chr(0b11110 + 0o121) + chr(882 - 834), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(4391 - 4291) + chr(7345 - 7244) + chr(7456 - 7357) + '\x6f' + chr(100) + chr(0b1100101))(chr(0b111010 + 0o73) + chr(116) + chr(0b101110 + 0o70) + chr(0b101101) + chr(0b101000 + 0o20)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yk\r`\x0em\x9c9\xa9&\xd7\xfdc\x8e\xae'), chr(0b11110 + 0o106) + '\145' + '\x63' + chr(0b111010 + 0o65) + chr(9936 - 9836) + chr(7632 - 7531))(chr(117) + '\x74' + chr(3141 - 3039) + '\x2d' + '\x38'), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(48), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\x64' + chr(0b11010 + 0o113) + chr(99) + '\x6f' + '\144' + '\145')(chr(117) + chr(116) + chr(0b1001110 + 0o30) + '\x2d' + chr(1178 - 1122)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yy\x1ec\x0bg\xb6.'), chr(0b1100100) + chr(0b1100101) + '\x63' + chr(5740 - 5629) + chr(0b1100100) + chr(4217 - 4116))('\165' + chr(9481 - 9365) + chr(102) + '\x2d' + chr(56)), 0.0)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(100) + '\x65' + chr(99) + chr(0b1101100 + 0o3) + chr(0b1100100) + '\x65')(chr(117) + chr(116) + '\146' + '\x2d' + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xe0\xb8\xdd\t\xcb\xa5\xb9\x85`Ih\x18'), chr(100) + chr(101) + chr(0b111 + 0o134) + '\157' + chr(100) + '\145')('\x75' + '\x74' + chr(102) + chr(475 - 430) + '\070'), 0.0)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(101 - 1) + chr(101) + chr(99) + chr(0b1000100 + 0o53) + chr(5342 - 5242) + '\x65')(chr(0b1110101) + chr(116) + '\x66' + chr(0b100101 + 0o10) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xe2\xb2\xc2'), '\144' + chr(1270 - 1169) + '\143' + '\157' + chr(7162 - 7062) + '\x65')(chr(0b1101100 + 0o11) + '\x74' + chr(10265 - 10163) + chr(1935 - 1890) + chr(0b11110 + 0o32)), xafqLlk3kkUe(SXOLrMavuUCe(b'\xe6\xb4\xdc\x15\xfa\xa6'), '\x64' + '\145' + chr(0b1000001 + 0o42) + chr(111) + chr(2921 - 2821) + chr(101))(chr(0b100110 + 0o117) + chr(0b110011 + 0o101) + '\x66' + '\055' + chr(0b111000)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(7714 - 7614) + '\145' + '\x63' + chr(0b1101111) + chr(100) + '\x65')('\x75' + '\164' + '\x66' + chr(0b11000 + 0o25) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xff\xb2\xd4#\xf8\xa0\xb2\x8fbU'), '\x64' + '\x65' + chr(9721 - 9622) + chr(0b1101111) + '\x64' + '\x65')(chr(117) + chr(126 - 10) + chr(0b1011001 + 0o15) + chr(1252 - 1207) + chr(56)), xafqLlk3kkUe(SXOLrMavuUCe(b'\xa0'), chr(0b1100100) + chr(2045 - 1944) + chr(0b1100011) + chr(111) + '\144' + '\x65')(chr(0b1110101) + chr(116) + chr(8917 - 8815) + '\x2d' + chr(56)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(100) + chr(101) + '\143' + chr(111) + chr(0b11 + 0o141) + '\145')(chr(3210 - 3093) + chr(0b1110100) + '\x66' + chr(0b100100 + 0o11) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yq\ru\x1ez\xb0'), '\x64' + '\x65' + chr(0b1100011) + chr(0b1101111) + '\144' + chr(0b1000100 + 0o41))(chr(0b110 + 0o157) + chr(0b11011 + 0o131) + chr(0b1111 + 0o127) + chr(1232 - 1187) + chr(0b11111 + 0o31)), xafqLlk3kkUe(SXOLrMavuUCe(b''), chr(100) + chr(101) + chr(894 - 795) + chr(9416 - 9305) + chr(0b1100100) + chr(0b1100101))(chr(11810 - 11693) + '\x74' + chr(0b1100110) + chr(255 - 210) + '\070'))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\144' + '\x65' + '\x63' + '\157' + '\144' + '\145')(chr(117) + chr(116) + chr(0b1100110) + chr(45) + chr(2191 - 2135)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yi\x15|\x1e'), '\144' + chr(101) + '\143' + chr(0b10000 + 0o137) + '\x64' + '\145')(chr(0b1110101) + '\164' + chr(581 - 479) + chr(0b101101) + chr(56)), xafqLlk3kkUe(mAAZP2Lf9x3V, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x88\xfd(\xdd\x89\x8e\xabT'), '\144' + chr(0b1100101) + chr(0b1100011) + chr(5488 - 5377) + '\x64' + '\x65')(chr(117) + chr(0b1110100) + chr(0b1000111 + 0o37) + chr(0b11111 + 0o16) + chr(0b110011 + 0o5))))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(100) + chr(101) + chr(5503 - 5404) + '\x6f' + chr(0b1100100) + chr(0b1100101))(chr(0b1101110 + 0o7) + '\164' + '\x66' + '\x2d' + chr(91 - 35)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yq\x03o\x1ad'), chr(0b10110 + 0o116) + chr(0b1100101) + chr(99) + chr(111) + chr(0b1100100) + chr(1659 - 1558))(chr(117) + '\164' + chr(5764 - 5662) + chr(1105 - 1060) + chr(56)), ehT0Px3KOsy9(chr(154 - 106) + '\x6f' + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(0b101111 + 0o65) + chr(4361 - 4260) + chr(3617 - 3518) + '\x6f' + chr(0b1100100) + chr(0b1100101))('\x75' + chr(0b1110001 + 0o3) + '\146' + '\055' + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yp\x03i$c'), chr(0b1100100) + '\145' + '\x63' + '\x6f' + chr(0b1100100) + '\145')('\165' + chr(11951 - 11835) + '\x66' + chr(45) + chr(0b1101 + 0o53)), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\062', 0b1000))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\x64' + chr(0b110001 + 0o64) + '\x63' + chr(111) + '\x64' + chr(0b1100101))(chr(0b1110101) + '\x74' + chr(102) + chr(45) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~ys\x19a$`\xa6;\xa5'), chr(0b1001101 + 0o27) + '\145' + chr(3724 - 3625) + '\x6f' + chr(183 - 83) + chr(0b100011 + 0o102))('\x75' + '\164' + chr(0b1100110) + chr(45) + '\x38'), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b10010 + 0o37), 0b1000))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\x64' + chr(0b1100101) + chr(0b1010011 + 0o20) + '\157' + chr(100) + '\145')('\x75' + chr(0b1010 + 0o152) + chr(102) + chr(338 - 293) + chr(2108 - 2052)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~ys\x19a$m\xbb*\xa45\xcd\xe0'), chr(0b1100100) + chr(0b1011010 + 0o13) + chr(0b1011110 + 0o5) + chr(0b1100001 + 0o16) + chr(100) + chr(101))('\x75' + '\x74' + chr(102) + chr(45) + chr(0b111000)), ehT0Px3KOsy9(chr(48) + chr(0b1001011 + 0o44) + chr(0b101111 + 0o3) + '\x30', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(100) + '\x65' + '\x63' + chr(5729 - 5618) + chr(100) + '\145')('\x75' + chr(4885 - 4769) + '\x66' + '\x2d' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yn\x1c`\x12|\x9c8\xa03\xda\xfb'), chr(0b1011010 + 0o12) + chr(1618 - 1517) + chr(0b1100011) + chr(5444 - 5333) + chr(0b1100100) + chr(2235 - 2134))(chr(117) + chr(0b1110100) + chr(4145 - 4043) + chr(45) + chr(56)), ehT0Px3KOsy9(chr(2296 - 2248) + chr(0b1101111) + chr(48), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(0b1100100) + '\145' + chr(0b10010 + 0o121) + '\x6f' + chr(0b1100100) + '\145')(chr(0b1110101) + chr(5600 - 5484) + chr(102) + '\x2d' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yo\th$n\xa29\xb5(\xcb'), chr(100) + '\145' + chr(99) + '\x6f' + chr(0b110001 + 0o63) + '\x65')(chr(0b100100 + 0o121) + chr(116) + chr(102) + '\x2d' + '\070'), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110011), ord("\x08")))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\x64' + chr(0b1000 + 0o135) + chr(0b1100011) + chr(0b1101111) + '\144' + chr(4454 - 4353))(chr(3030 - 2913) + chr(2367 - 2251) + chr(2161 - 2059) + '\x2d' + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~y\x7f\x00c\x18c\x9c6\xa4)\xde\xe7n'), chr(787 - 687) + chr(8567 - 8466) + '\143' + '\157' + chr(100) + '\145')('\165' + chr(116) + '\146' + chr(45) + chr(56)), ehT0Px3KOsy9(chr(1147 - 1099) + chr(111) + chr(0b101110 + 0o4) + chr(0b110000) + chr(48), 63110 - 63102))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\x64' + chr(8040 - 7939) + chr(0b101000 + 0o73) + chr(0b1011001 + 0o26) + chr(0b1100100) + chr(101))(chr(0b1100111 + 0o16) + chr(4804 - 4688) + chr(102) + chr(1574 - 1529) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yo\th\x0ek\xb73\xae)\xe6\xe7\x7f\x92\xb8'), chr(100) + chr(101) + chr(0b1100000 + 0o3) + '\x6f' + chr(5033 - 4933) + '\x65')(chr(117) + chr(0b1110100) + chr(102) + chr(0b11001 + 0o24) + '\070'), xafqLlk3kkUe(SXOLrMavuUCe(b'\xf1\xb2\xdf\n'), chr(0b1100100) + chr(7519 - 7418) + '\x63' + '\x6f' + chr(8504 - 8404) + chr(2848 - 2747))(chr(128 - 11) + '\x74' + chr(0b11111 + 0o107) + chr(0b101101) + chr(56)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(4632 - 4532) + '\x65' + chr(0b1011001 + 0o12) + '\x6f' + '\x64' + chr(7064 - 6963))(chr(498 - 381) + chr(116) + chr(0b1100110) + '\055' + chr(2918 - 2862)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~ys\x03b\x17a\xad?\xa05\xd0\xe7\x7f'), chr(100) + chr(101) + '\x63' + chr(11976 - 11865) + '\144' + chr(0b101100 + 0o71))(chr(117) + chr(0b1110011 + 0o1) + '\146' + chr(0b1100 + 0o41) + '\070'), xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xb2\xdf\x19'), chr(0b1110 + 0o126) + '\x65' + chr(99) + '\157' + chr(0b1100100) + chr(4815 - 4714))('\x75' + chr(13132 - 13016) + chr(0b1100110) + chr(45) + chr(760 - 704)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(0b100010 + 0o102) + chr(4905 - 4804) + chr(99) + chr(3304 - 3193) + '\x64' + '\145')('\x75' + chr(0b1110100) + '\146' + chr(0b101101) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yx\x14|$n\xa29\xb5(\xcb'), '\144' + '\x65' + chr(99) + '\157' + chr(0b1100100) + chr(0b1100101))(chr(0b101000 + 0o115) + '\164' + chr(102) + '\055' + chr(0b110 + 0o62)), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(6485 - 6374) + chr(370 - 322), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(5746 - 5646) + '\145' + chr(0b1100011) + '\x6f' + '\x64' + chr(101))('\165' + chr(116) + chr(4608 - 4506) + '\055' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yx\x14|$a\xad*\xb43\xdd\xfak'), chr(0b1010011 + 0o21) + '\145' + chr(0b111111 + 0o44) + chr(0b1101111) + '\x64' + chr(0b1100101))(chr(0b1110101) + '\164' + chr(0b1000111 + 0o37) + '\055' + chr(0b111000)), ehT0Px3KOsy9('\x30' + '\157' + chr(2049 - 1999) + '\x30' + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(4583 - 4483) + chr(6246 - 6145) + chr(0b1100011) + chr(0b10111 + 0o130) + '\144' + chr(0b1100101))('\x75' + '\x74' + '\x66' + chr(45) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yv\x1dS\x08a\xb9?'), chr(0b1100100) + '\x65' + chr(2955 - 2856) + '\x6f' + '\x64' + '\x65')('\165' + chr(116) + '\x66' + '\055' + chr(56)), ehT0Px3KOsy9(chr(0b10 + 0o56) + '\157' + '\062' + chr(48) + chr(0b10 + 0o56), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\x64' + chr(4114 - 4013) + chr(0b1100011) + chr(0b111010 + 0o65) + '\144' + chr(0b101011 + 0o72))(chr(117) + chr(116) + chr(0b100001 + 0o105) + '\055' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yk3\x7f\x12r\xa6'), '\x64' + chr(101) + chr(99) + chr(111) + chr(2552 - 2452) + '\145')('\165' + '\164' + '\146' + chr(0b11111 + 0o16) + chr(56)), ehT0Px3KOsy9(chr(48) + '\157' + chr(52) + chr(48) + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(100) + chr(101) + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + '\145')(chr(0b1110101) + chr(116) + chr(102) + '\x2d' + chr(768 - 712)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xa9\xc5\x19\xfa\xb5\xa2\x85~yq\x03m\x1fW\xa1;\xad&\xd7\xf0c'), chr(0b100 + 0o140) + '\x65' + chr(2767 - 2668) + '\157' + '\144' + chr(101))(chr(0b111101 + 0o70) + chr(9271 - 9155) + chr(1683 - 1581) + '\x2d' + '\x38'), 0.02)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(100) + '\x65' + chr(0b1100011) + chr(3027 - 2916) + chr(2674 - 2574) + '\x65')(chr(0b100 + 0o161) + '\x74' + '\146' + chr(0b101101) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe\xae\xd9#\xfa\xb4\xa6\xb5x_m\t~\x0bd\xa24\xa44'), chr(202 - 102) + chr(101) + '\143' + chr(0b1101111) + chr(100) + '\x65')(chr(117) + '\x74' + chr(102) + chr(45) + chr(0b111000)), ehT0Px3KOsy9(chr(1101 - 1053) + chr(11655 - 11544) + chr(66 - 14), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(2930 - 2830) + chr(0b1100101) + chr(99) + chr(10496 - 10385) + chr(0b1100100) + chr(0b111001 + 0o54))(chr(0b1110101) + '\x74' + '\146' + chr(356 - 311) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe\xae\xd9#\xe1\xb2\xae\xb5}Gm3j\x15'), chr(3022 - 2922) + '\145' + chr(2746 - 2647) + chr(0b11001 + 0o126) + '\144' + chr(0b11001 + 0o114))('\x75' + chr(0b1110100) + chr(0b1011011 + 0o13) + chr(45) + '\x38'), ehT0Px3KOsy9('\x30' + chr(0b1101000 + 0o7) + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(0b1100100) + chr(0b1100101) + '\x63' + '\157' + chr(8581 - 8481) + '\145')(chr(0b1110101) + chr(0b101011 + 0o111) + chr(0b1100110) + chr(99 - 54) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xe7\xae\xd4#\xe7\xa4\xbb\x89\x7fHk'), chr(0b1001011 + 0o31) + chr(4023 - 3922) + '\143' + chr(10930 - 10819) + chr(0b1100100) + chr(101))('\x75' + '\164' + '\x66' + chr(0b100 + 0o51) + '\x38'), ehT0Px3KOsy9(chr(0b1010 + 0o46) + chr(0b100 + 0o153) + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(5447 - 5347) + '\145' + chr(99) + chr(11672 - 11561) + '\x64' + chr(0b1100101))('\x75' + chr(13027 - 12911) + chr(0b1000000 + 0o46) + chr(1780 - 1735) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf6\xb4\xd4\x08\xcb\xa4\xb3\x9auTi\x1f'), chr(0b100011 + 0o101) + chr(0b1100101) + chr(0b10 + 0o141) + chr(0b100110 + 0o111) + chr(100) + '\145')('\x75' + chr(116) + chr(0b1000010 + 0o44) + chr(45) + '\x38'), ehT0Px3KOsy9(chr(1824 - 1776) + chr(0b1000001 + 0o56) + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), '\x64' + chr(101) + chr(0b101110 + 0o65) + '\x6f' + chr(0b1001 + 0o133) + chr(0b1100101))(chr(0b1110101) + chr(337 - 221) + chr(0b1100110) + '\x2d' + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xff\xb8\xdc\x13\xe6\xb8\x94\x8fv@t\x0fe\x1ef\xb7\x05\xa7!\xd7'), '\x64' + chr(5995 - 5894) + chr(0b1100011) + chr(0b1000111 + 0o50) + chr(0b11100 + 0o110) + '\145')('\x75' + chr(0b1110100) + '\146' + '\x2d' + chr(0b1111 + 0o51)), ehT0Px3KOsy9(chr(48) + chr(0b10111 + 0o130) + chr(0b110000), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\xb9\xd5#\xfc\xb1\xaa\x98qK'), chr(0b1100100) + chr(101) + chr(0b1100011) + chr(0b1001111 + 0o40) + chr(100) + '\x65')(chr(117) + '\164' + chr(524 - 422) + chr(0b101101) + chr(892 - 836)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xe7\xae\xd4#\xfd\xaf\xbb\x9fdU'), '\x64' + chr(0b1100101) + '\x63' + chr(0b1101111) + chr(0b1100100) + '\145')('\165' + chr(116) + '\x66' + chr(45) + chr(0b111000)), ehT0Px3KOsy9(chr(469 - 421) + '\x6f' + chr(0b11010 + 0o26), 8))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_base_long_seq
|
def attention_lm_moe_base_long_seq():
"""Hyper parameters specifics for long sequence generation."""
hparams = attention_lm_moe_base()
hparams.max_length = 0 # max_length == batch_size
hparams.eval_drop_long_sequences = True
hparams.min_length_bucket = 256 # Avoid cyclic problems for big batches
hparams.use_sepconv = True
return hparams
|
python
|
def attention_lm_moe_base_long_seq():
"""Hyper parameters specifics for long sequence generation."""
hparams = attention_lm_moe_base()
hparams.max_length = 0 # max_length == batch_size
hparams.eval_drop_long_sequences = True
hparams.min_length_bucket = 256 # Avoid cyclic problems for big batches
hparams.use_sepconv = True
return hparams
|
[
"def",
"attention_lm_moe_base_long_seq",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_base",
"(",
")",
"hparams",
".",
"max_length",
"=",
"0",
"# max_length == batch_size",
"hparams",
".",
"eval_drop_long_sequences",
"=",
"True",
"hparams",
".",
"min_length_bucket",
"=",
"256",
"# Avoid cyclic problems for big batches",
"hparams",
".",
"use_sepconv",
"=",
"True",
"return",
"hparams"
] |
Hyper parameters specifics for long sequence generation.
|
[
"Hyper",
"parameters",
"specifics",
"for",
"long",
"sequence",
"generation",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L519-L528
|
train
|
Hyper parameters specifics for long sequence generation.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(0b110001 + 0o76) + chr(51) + chr(0b110 + 0o55) + chr(0b110000), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(11986 - 11875) + chr(0b110001) + chr(52) + '\060', 36018 - 36010), ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(111) + '\062' + '\x32', 0o10), ehT0Px3KOsy9(chr(0b10100 + 0o34) + '\157' + chr(0b10011 + 0o43) + chr(375 - 324), ord("\x08")), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(5113 - 5002) + chr(54), 60308 - 60300), ehT0Px3KOsy9(chr(0b101 + 0o53) + chr(111) + chr(2105 - 2056) + '\x35' + chr(54), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011 + 0o144) + chr(1416 - 1365) + chr(0b11011 + 0o30) + chr(401 - 347), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(1035 - 984) + chr(0b101011 + 0o10) + chr(0b110110), 8), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(6637 - 6526) + '\063' + chr(732 - 682) + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1001011 + 0o44) + chr(51) + chr(0b10 + 0o63) + chr(52), 2067 - 2059), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011100 + 0o23) + chr(49) + chr(52) + chr(48), 8), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(111) + chr(0b110001) + chr(54) + '\061', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(1254 - 1205) + chr(53) + chr(0b101001 + 0o14), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b100001 + 0o116) + chr(2292 - 2241) + chr(48) + '\x35', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110010) + chr(0b11001 + 0o33) + '\x33', 42122 - 42114), ehT0Px3KOsy9('\060' + chr(0b1101000 + 0o7) + '\063' + '\x35' + chr(1967 - 1919), 0b1000), ehT0Px3KOsy9('\x30' + chr(7697 - 7586) + chr(0b100101 + 0o15) + '\060' + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\157' + chr(0b110001) + chr(0b100101 + 0o21) + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(1165 - 1117) + '\x6f' + chr(0b110001) + chr(48) + chr(0b10100 + 0o35), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b10000 + 0o44) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(11566 - 11455) + chr(0b110111) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110 + 0o57), 0o10), ehT0Px3KOsy9('\060' + chr(1778 - 1667) + '\067' + '\064', 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(0b101011 + 0o12) + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(4355 - 4244) + chr(50) + chr(0b110110) + chr(727 - 679), 34976 - 34968), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(0b0 + 0o157) + chr(0b100101 + 0o16) + '\x37' + chr(682 - 631), 31733 - 31725), ehT0Px3KOsy9(chr(93 - 45) + chr(0b111111 + 0o60) + '\x31' + '\060' + chr(2200 - 2149), 15873 - 15865), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + chr(1804 - 1753) + '\x37', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b1111 + 0o43) + chr(1562 - 1512) + chr(0b110001), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(50) + chr(0b110111) + chr(0b101011 + 0o12), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + '\x35' + chr(493 - 440), 43200 - 43192), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110010) + '\x32' + chr(1383 - 1333), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101110 + 0o1) + chr(50) + chr(0b10100 + 0o37) + chr(0b100010 + 0o17), ord("\x08")), ehT0Px3KOsy9(chr(1896 - 1848) + chr(461 - 350) + chr(911 - 861) + chr(0b110001) + '\x34', 0b1000), ehT0Px3KOsy9(chr(1866 - 1818) + chr(12269 - 12158) + chr(0b110001) + chr(0b1000 + 0o57) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b100100 + 0o14) + '\157' + '\x31' + chr(0b110101) + chr(0b110001), 59307 - 59299), ehT0Px3KOsy9('\060' + '\157' + '\x31' + '\060' + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x37' + chr(2214 - 2159), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x33' + '\x31' + chr(50), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062' + chr(0b110111) + '\066', 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1001111 + 0o40) + chr(53) + chr(0b110000), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xf6'), chr(100) + chr(0b1010000 + 0o25) + '\x63' + chr(111) + '\144' + chr(0b1100101))(chr(0b1110101) + chr(116) + '\x66' + chr(0b10110 + 0o27) + chr(2504 - 2448)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def Oj8_r_0ydVTy():
n4ljua2gi1Pr = kKmBjRbC6Zbc()
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110000), 0o10)
n4ljua2gi1Pr.n5sZSNr92T7V = ehT0Px3KOsy9('\x30' + chr(9252 - 9141) + chr(0b110001), ord("\x08"))
n4ljua2gi1Pr.lhJm4Z32JlM2 = ehT0Px3KOsy9('\060' + chr(111) + chr(1020 - 968) + chr(656 - 608) + chr(48), 0o10)
n4ljua2gi1Pr.cASdquKJVXhY = ehT0Px3KOsy9('\x30' + chr(111) + chr(49), 8)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_base_ae
|
def attention_lm_moe_base_ae():
"""Base model with attention expert."""
hparams = attention_lm_moe_base_long_seq()
hparams.attention_type = AttentionType.LOCAL_EXPERTS
hparams.learning_rate = 0.05
hparams.learning_rate_warmup_steps = 10000
# According to noam, ("n", "da") seems better for harder-to-learn models
# hparams.layer_preprocess_sequence = "n"
# hparams.layer_postprocess_sequence = "da"
return hparams
|
python
|
def attention_lm_moe_base_ae():
"""Base model with attention expert."""
hparams = attention_lm_moe_base_long_seq()
hparams.attention_type = AttentionType.LOCAL_EXPERTS
hparams.learning_rate = 0.05
hparams.learning_rate_warmup_steps = 10000
# According to noam, ("n", "da") seems better for harder-to-learn models
# hparams.layer_preprocess_sequence = "n"
# hparams.layer_postprocess_sequence = "da"
return hparams
|
[
"def",
"attention_lm_moe_base_ae",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_base_long_seq",
"(",
")",
"hparams",
".",
"attention_type",
"=",
"AttentionType",
".",
"LOCAL_EXPERTS",
"hparams",
".",
"learning_rate",
"=",
"0.05",
"hparams",
".",
"learning_rate_warmup_steps",
"=",
"10000",
"# According to noam, (\"n\", \"da\") seems better for harder-to-learn models",
"# hparams.layer_preprocess_sequence = \"n\"",
"# hparams.layer_postprocess_sequence = \"da\"",
"return",
"hparams"
] |
Base model with attention expert.
|
[
"Base",
"model",
"with",
"attention",
"expert",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L532-L542
|
train
|
Base model with attention expert.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(111) + chr(0b101111 + 0o3) + chr(2059 - 2010) + '\x30', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(50) + '\062', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(49) + '\x33' + chr(0b110001), 16365 - 16357), ehT0Px3KOsy9('\060' + chr(0b1101101 + 0o2) + chr(52) + chr(1211 - 1159), 9179 - 9171), ehT0Px3KOsy9(chr(1697 - 1649) + '\x6f' + chr(49) + chr(49) + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(51) + chr(0b110000) + '\061', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(2544 - 2489) + chr(1871 - 1819), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1000001 + 0o56) + chr(674 - 625) + chr(50) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b1110 + 0o42) + chr(0b1011010 + 0o25) + chr(50) + '\067' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\061' + chr(0b11110 + 0o24) + chr(0b101 + 0o60), 0o10), ehT0Px3KOsy9('\060' + chr(4162 - 4051) + chr(0b110001) + chr(0b110100) + '\063', 0o10), ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(7176 - 7065) + chr(251 - 200) + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1101111) + chr(50) + chr(1689 - 1639) + chr(0b10110 + 0o35), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(497 - 448) + chr(1953 - 1902) + chr(2289 - 2241), 17392 - 17384), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(0b1001100 + 0o43) + chr(0b1010 + 0o50) + chr(0b110011) + chr(0b100101 + 0o16), 0b1000), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(111) + '\061' + chr(2068 - 2018) + chr(0b110001), 8), ehT0Px3KOsy9('\x30' + chr(0b110101 + 0o72) + chr(0b110011) + chr(0b1101 + 0o43) + chr(0b110001), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b101101 + 0o102) + '\064' + '\066', 2698 - 2690), ehT0Px3KOsy9('\060' + '\157' + chr(0b110001) + chr(0b110000) + chr(1496 - 1443), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1001000 + 0o47) + chr(0b1000 + 0o51) + chr(53) + chr(0b110 + 0o60), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(50) + chr(0b110001) + chr(417 - 364), 53280 - 53272), ehT0Px3KOsy9('\x30' + '\157' + chr(51) + '\x35' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(1448 - 1400) + chr(9692 - 9581) + chr(2118 - 2068) + chr(49) + '\061', 0o10), ehT0Px3KOsy9(chr(0b1101 + 0o43) + '\x6f' + '\x32' + '\064' + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(792 - 741) + chr(0b100000 + 0o20), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110001) + chr(0b110011) + chr(776 - 724), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b101 + 0o54) + chr(0b110110) + chr(0b110110), 34390 - 34382), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1348 - 1296) + chr(52), 8), ehT0Px3KOsy9(chr(48) + chr(0b1100100 + 0o13) + chr(50) + chr(0b100111 + 0o17) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(0b1010 + 0o46) + '\157' + '\x36' + chr(48), 49984 - 49976), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011) + chr(0b110101) + '\060', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b101001 + 0o10) + chr(0b1001 + 0o47) + '\062', 10626 - 10618), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(51) + chr(52) + '\063', ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\061' + '\067' + chr(53), 0o10), ehT0Px3KOsy9(chr(1714 - 1666) + chr(12234 - 12123) + '\061' + chr(0b101000 + 0o13) + chr(1033 - 978), 8614 - 8606), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(50) + chr(48) + chr(2008 - 1958), 0o10), ehT0Px3KOsy9(chr(0b1010 + 0o46) + chr(111) + chr(0b110011) + '\060' + chr(0b101011 + 0o13), 859 - 851), ehT0Px3KOsy9(chr(0b100110 + 0o12) + '\157' + '\x32' + '\066' + '\060', 52660 - 52652), ehT0Px3KOsy9('\x30' + '\157' + chr(1072 - 1022) + chr(0b110000) + chr(0b101001 + 0o14), 0o10), ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(3302 - 3191) + chr(1355 - 1302), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + '\x6f' + '\x35' + '\x30', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xa6'), '\x64' + chr(101) + chr(0b101100 + 0o67) + '\157' + '\x64' + '\145')(chr(0b1110101) + '\164' + chr(0b111111 + 0o47) + chr(362 - 317) + chr(0b11 + 0o65)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def ZoLq4c6Wa9Eh():
n4ljua2gi1Pr = Oj8_r_0ydVTy()
n4ljua2gi1Pr.lZ1GB4L2oMeG = mAAZP2Lf9x3V.LOCAL_EXPERTS
n4ljua2gi1Pr.QGSIpd_yUNzU = 0.05
n4ljua2gi1Pr.fHyhoyGmdvM9 = ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(51) + chr(0b110100) + chr(769 - 719) + chr(48), 0b1000)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_ae_extended
|
def attention_lm_ae_extended():
"""Experiment with the exp_factor params."""
hparams = attention_lm_moe_base_long_seq()
hparams.attention_layers = "eeee"
hparams.attention_local = True
# hparams.factored_logits=1 # Necessary when the number of expert grow bigger
hparams.attention_moe_k = 2
hparams.attention_exp_factor = 4
# hparams.attention_exp_inputdim = 128
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
return hparams
|
python
|
def attention_lm_ae_extended():
"""Experiment with the exp_factor params."""
hparams = attention_lm_moe_base_long_seq()
hparams.attention_layers = "eeee"
hparams.attention_local = True
# hparams.factored_logits=1 # Necessary when the number of expert grow bigger
hparams.attention_moe_k = 2
hparams.attention_exp_factor = 4
# hparams.attention_exp_inputdim = 128
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
return hparams
|
[
"def",
"attention_lm_ae_extended",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_base_long_seq",
"(",
")",
"hparams",
".",
"attention_layers",
"=",
"\"eeee\"",
"hparams",
".",
"attention_local",
"=",
"True",
"# hparams.factored_logits=1 # Necessary when the number of expert grow bigger",
"hparams",
".",
"attention_moe_k",
"=",
"2",
"hparams",
".",
"attention_exp_factor",
"=",
"4",
"# hparams.attention_exp_inputdim = 128",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"n\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"da\"",
"return",
"hparams"
] |
Experiment with the exp_factor params.
|
[
"Experiment",
"with",
"the",
"exp_factor",
"params",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L599-L611
|
train
|
Experiment with the exp_factor params.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(321 - 210) + '\061' + chr(2367 - 2315) + '\062', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(11760 - 11649) + chr(0b110001) + chr(51) + chr(443 - 391), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1457 - 1403) + '\063', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110000 + 0o3) + chr(1752 - 1699), 0b1000), ehT0Px3KOsy9(chr(0b100111 + 0o11) + '\157' + '\x31' + chr(1062 - 1014) + chr(1156 - 1101), ord("\x08")), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(2479 - 2368) + chr(0b110010) + '\067' + chr(0b110111), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b110011) + chr(0b110111) + chr(0b110001), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b10000 + 0o43) + chr(1563 - 1508) + chr(0b10011 + 0o42), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b100100 + 0o15) + chr(55) + '\067', 47592 - 47584), ehT0Px3KOsy9('\060' + chr(0b1100001 + 0o16) + chr(869 - 818) + chr(0b100 + 0o61) + chr(0b100 + 0o57), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1011111 + 0o20) + '\061' + '\x31', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b10 + 0o61) + chr(0b100100 + 0o14) + chr(1692 - 1643), 6521 - 6513), ehT0Px3KOsy9('\x30' + '\x6f' + '\061' + chr(0b10111 + 0o35) + chr(2130 - 2079), 5454 - 5446), ehT0Px3KOsy9(chr(754 - 706) + chr(10079 - 9968) + chr(1112 - 1063) + chr(959 - 907) + '\066', 36129 - 36121), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b1011001 + 0o26) + chr(0b1001 + 0o55) + '\x36', 17614 - 17606), ehT0Px3KOsy9('\x30' + '\x6f' + '\x32' + chr(1919 - 1869) + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1010101 + 0o32) + chr(0b110110) + '\x32', 0b1000), ehT0Px3KOsy9(chr(1484 - 1436) + '\x6f' + '\061' + chr(407 - 354) + chr(0b110000), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + '\x31' + chr(306 - 251) + chr(596 - 541), 8), ehT0Px3KOsy9(chr(396 - 348) + chr(0b101110 + 0o101) + chr(0b10000 + 0o45) + '\x36', 0o10), ehT0Px3KOsy9(chr(0b111 + 0o51) + chr(5166 - 5055) + chr(49) + '\x36' + '\x31', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b10 + 0o57) + '\x30' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b1000000 + 0o57) + '\061' + '\x35' + chr(1603 - 1549), 19371 - 19363), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\x6f' + '\x31' + '\x34' + chr(1474 - 1424), 8), ehT0Px3KOsy9('\x30' + '\157' + chr(107 - 56) + '\062' + '\x36', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(3426 - 3315) + chr(52) + chr(1423 - 1373), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(49) + chr(1939 - 1885) + chr(49), 8), ehT0Px3KOsy9(chr(2030 - 1982) + '\157' + chr(51) + '\x31' + '\x35', 0b1000), ehT0Px3KOsy9(chr(1409 - 1361) + '\157' + '\063' + '\067' + chr(0b1110 + 0o50), 32138 - 32130), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b10010 + 0o40) + '\x35' + chr(1053 - 998), 0b1000), ehT0Px3KOsy9(chr(0b101000 + 0o10) + '\x6f' + '\x33' + '\065' + chr(0b101000 + 0o15), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + chr(664 - 609), 0b1000), ehT0Px3KOsy9(chr(267 - 219) + chr(0b10011 + 0o134) + '\063' + chr(0b10101 + 0o33) + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(1151 - 1103) + '\x6f' + chr(0b1001 + 0o50) + chr(2594 - 2543) + '\064', 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\063' + chr(0b110110) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\063' + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2227 - 2172) + chr(0b110011), 0o10), ehT0Px3KOsy9('\x30' + chr(4873 - 4762) + chr(0b110010) + chr(0b110111), 8), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110001) + chr(55) + '\x37', 8), ehT0Px3KOsy9('\x30' + chr(0b101100 + 0o103) + chr(0b100100 + 0o16) + '\061', 54248 - 54240)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + '\157' + chr(0b100011 + 0o22) + chr(0b110000), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'L'), chr(0b1100100) + chr(1216 - 1115) + chr(99) + chr(0b101000 + 0o107) + '\x64' + chr(0b101110 + 0o67))('\x75' + '\x74' + '\146' + '\055' + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def EhDDH3XToFZI():
n4ljua2gi1Pr = Oj8_r_0ydVTy()
n4ljua2gi1Pr.N6069QsyeXwL = xafqLlk3kkUe(SXOLrMavuUCe(b'\x079Vx'), chr(0b101111 + 0o65) + '\x65' + '\x63' + '\x6f' + '\x64' + chr(0b1100101))('\x75' + '\164' + '\146' + '\055' + chr(0b111000))
n4ljua2gi1Pr.w4zWvvRYqpB5 = ehT0Px3KOsy9('\x30' + chr(0b1011 + 0o144) + chr(1568 - 1519), 0o10)
n4ljua2gi1Pr.TOsbu0AaSNDO = ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110010), 0b1000)
n4ljua2gi1Pr.urAFCjSC7gM2 = ehT0Px3KOsy9(chr(2095 - 2047) + chr(111) + '\x34', 0b1000)
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'\x0c'), '\x64' + '\x65' + chr(9499 - 9400) + '\x6f' + chr(8954 - 8854) + '\145')('\165' + chr(9741 - 9625) + '\x66' + chr(1403 - 1358) + '\070')
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'\x06='), chr(6677 - 6577) + chr(0b1100101) + chr(0b111100 + 0o47) + chr(0b1010011 + 0o34) + chr(1427 - 1327) + chr(4956 - 4855))(chr(0b11101 + 0o130) + chr(12533 - 12417) + chr(102) + chr(0b1000 + 0o45) + chr(0b0 + 0o70))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_base_memeff
|
def attention_lm_moe_base_memeff():
"""Base model with attention expert."""
hparams = attention_lm_moe_base_long_seq()
hparams.use_sepconv = False
hparams.diet_experts = True
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
hparams.layer_prepostprocess_dropout = 0.0
hparams.memory_efficient_ffn = True
hparams.attention_type = AttentionType.MEMORY_EFFICIENT
hparams.num_heads = 8
hparams.factored_logits = True
return hparams
|
python
|
def attention_lm_moe_base_memeff():
"""Base model with attention expert."""
hparams = attention_lm_moe_base_long_seq()
hparams.use_sepconv = False
hparams.diet_experts = True
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
hparams.layer_prepostprocess_dropout = 0.0
hparams.memory_efficient_ffn = True
hparams.attention_type = AttentionType.MEMORY_EFFICIENT
hparams.num_heads = 8
hparams.factored_logits = True
return hparams
|
[
"def",
"attention_lm_moe_base_memeff",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_base_long_seq",
"(",
")",
"hparams",
".",
"use_sepconv",
"=",
"False",
"hparams",
".",
"diet_experts",
"=",
"True",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"n\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"da\"",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.0",
"hparams",
".",
"memory_efficient_ffn",
"=",
"True",
"hparams",
".",
"attention_type",
"=",
"AttentionType",
".",
"MEMORY_EFFICIENT",
"hparams",
".",
"num_heads",
"=",
"8",
"hparams",
".",
"factored_logits",
"=",
"True",
"return",
"hparams"
] |
Base model with attention expert.
|
[
"Base",
"model",
"with",
"attention",
"expert",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L615-L628
|
train
|
Base model with attention expert.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010) + chr(0b101001 + 0o7) + chr(0b100000 + 0o21), ord("\x08")), ehT0Px3KOsy9(chr(1610 - 1562) + chr(10694 - 10583) + '\062' + chr(130 - 75) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b111001 + 0o66) + chr(0b10110 + 0o36) + chr(0b110101), 46155 - 46147), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\157' + '\x35' + '\x37', 52216 - 52208), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(111) + chr(0b110011) + chr(0b10101 + 0o34) + chr(51), 58579 - 58571), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(132 - 77) + chr(1653 - 1601), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1000010 + 0o55) + '\x34' + '\x34', 54318 - 54310), ehT0Px3KOsy9(chr(48) + chr(7709 - 7598) + chr(0b11000 + 0o35) + '\x31', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\062' + chr(0b110111) + '\x34', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(2079 - 2030) + chr(48) + chr(0b110000), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062' + chr(49) + chr(0b1 + 0o63), 5639 - 5631), ehT0Px3KOsy9('\060' + '\157' + chr(0b110 + 0o55) + chr(0b110101) + chr(0b110110), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49) + chr(53) + chr(51), 36848 - 36840), ehT0Px3KOsy9(chr(0b110000) + chr(9389 - 9278) + '\x32' + chr(677 - 626) + chr(93 - 43), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(2960 - 2849) + '\x36' + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b100 + 0o57) + '\x33' + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(3139 - 3028) + '\x31' + chr(0b110110) + chr(0b110010), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + '\063' + chr(923 - 875) + '\065', 57277 - 57269), ehT0Px3KOsy9(chr(2159 - 2111) + chr(9712 - 9601) + '\061' + '\x30' + chr(0b110010), 53989 - 53981), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110001) + chr(0b110000) + chr(927 - 873), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\x36' + '\063', 12204 - 12196), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110001) + chr(55) + '\x36', 58210 - 58202), ehT0Px3KOsy9('\x30' + chr(6571 - 6460) + '\062' + '\x35' + '\064', 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\061' + '\063' + '\060', 18989 - 18981), ehT0Px3KOsy9('\x30' + chr(111) + '\062' + chr(50) + chr(0b101010 + 0o7), 61190 - 61182), ehT0Px3KOsy9(chr(78 - 30) + chr(1300 - 1189) + chr(2350 - 2300) + '\x32' + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(1711 - 1663) + chr(111) + '\062' + chr(52) + chr(1791 - 1738), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\062' + chr(0b110001) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(1003 - 954) + chr(53) + chr(0b11010 + 0o27), 0b1000), ehT0Px3KOsy9('\060' + chr(0b11110 + 0o121) + '\063' + chr(0b10110 + 0o32) + '\062', 28191 - 28183), ehT0Px3KOsy9(chr(48) + chr(0b10011 + 0o134) + chr(0b11011 + 0o26) + chr(0b110110) + '\x35', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(414 - 363) + '\x37', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(11057 - 10946) + chr(1061 - 1012) + '\067' + chr(0b110000), 0o10), ehT0Px3KOsy9('\060' + chr(0b110010 + 0o75) + chr(1473 - 1422) + chr(55) + '\060', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110001 + 0o0) + chr(448 - 395) + '\x37', 4508 - 4500), ehT0Px3KOsy9(chr(293 - 245) + '\x6f' + chr(0b110010) + chr(0b110011) + chr(0b11010 + 0o33), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(8510 - 8399) + chr(119 - 70) + '\066' + chr(55), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b101 + 0o56) + '\x34' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(55) + chr(0b110101), 41709 - 41701), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011010 + 0o25) + '\x31' + '\067' + chr(1201 - 1152), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + '\x6f' + '\x35' + '\x30', 14721 - 14713)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xaf'), '\144' + '\145' + '\x63' + '\x6f' + chr(100) + chr(0b1010100 + 0o21))('\x75' + chr(0b1110100) + chr(102) + '\x2d' + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def x8UvPMTS0YIw():
n4ljua2gi1Pr = Oj8_r_0ydVTy()
n4ljua2gi1Pr.cASdquKJVXhY = ehT0Px3KOsy9('\060' + '\157' + chr(0b110000), 1380 - 1372)
n4ljua2gi1Pr.chv9ul4_iQE1 = ehT0Px3KOsy9(chr(0b110000) + '\157' + '\061', ord("\x08"))
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'\xef'), '\x64' + chr(0b1100101) + '\143' + chr(111) + '\x64' + chr(0b1100101))('\165' + '\164' + chr(0b1100110) + '\x2d' + chr(0b11001 + 0o37))
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'\xe5\xdf'), '\x64' + '\145' + '\143' + chr(0b1001111 + 0o40) + chr(933 - 833) + chr(3547 - 3446))(chr(0b1110101) + chr(116) + chr(0b1100000 + 0o6) + chr(1348 - 1303) + '\x38')
n4ljua2gi1Pr.RW_xSzp18UeS = 0.0
n4ljua2gi1Pr.Cx5KXzfumeJI = ehT0Px3KOsy9(chr(48) + chr(726 - 615) + chr(0b110001), 8)
n4ljua2gi1Pr.lZ1GB4L2oMeG = mAAZP2Lf9x3V.MEMORY_EFFICIENT
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(745 - 697) + '\157' + chr(49) + chr(0b110000), 0o10)
n4ljua2gi1Pr.Sjw4kbDQvNmN = ehT0Px3KOsy9(chr(0b100010 + 0o16) + '\157' + chr(1299 - 1250), 8)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_small
|
def attention_lm_moe_small():
"""Cheap model for single-gpu training.
on lm1b_32k:
~312M params
1.6 steps/sec on [GeForce GTX TITAN X]
After 50K steps on 8 GPUs (synchronous):
eval_log_ppl_per_token = 3.31
Returns:
an hparams object.
"""
hparams = attention_lm_moe_base()
hparams.num_hidden_layers = 4
hparams.hidden_size = 512
hparams.filter_size = 2048
hparams.moe_num_experts = 128
hparams.moe_layers = "2"
return hparams
|
python
|
def attention_lm_moe_small():
"""Cheap model for single-gpu training.
on lm1b_32k:
~312M params
1.6 steps/sec on [GeForce GTX TITAN X]
After 50K steps on 8 GPUs (synchronous):
eval_log_ppl_per_token = 3.31
Returns:
an hparams object.
"""
hparams = attention_lm_moe_base()
hparams.num_hidden_layers = 4
hparams.hidden_size = 512
hparams.filter_size = 2048
hparams.moe_num_experts = 128
hparams.moe_layers = "2"
return hparams
|
[
"def",
"attention_lm_moe_small",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_base",
"(",
")",
"hparams",
".",
"num_hidden_layers",
"=",
"4",
"hparams",
".",
"hidden_size",
"=",
"512",
"hparams",
".",
"filter_size",
"=",
"2048",
"hparams",
".",
"moe_num_experts",
"=",
"128",
"hparams",
".",
"moe_layers",
"=",
"\"2\"",
"return",
"hparams"
] |
Cheap model for single-gpu training.
on lm1b_32k:
~312M params
1.6 steps/sec on [GeForce GTX TITAN X]
After 50K steps on 8 GPUs (synchronous):
eval_log_ppl_per_token = 3.31
Returns:
an hparams object.
|
[
"Cheap",
"model",
"for",
"single",
"-",
"gpu",
"training",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L632-L650
|
train
|
Cheap model for single - gpu training.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(217 - 106) + chr(0b111 + 0o60) + chr(0b11110 + 0o31), 0o10), ehT0Px3KOsy9('\060' + chr(11150 - 11039) + chr(49) + '\066', 37114 - 37106), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + '\066' + chr(49), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110001) + chr(49) + chr(54), 0b1000), ehT0Px3KOsy9('\060' + chr(2491 - 2380) + chr(0b1001 + 0o52) + chr(53) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b10011 + 0o134) + chr(0b110010) + chr(0b100100 + 0o21) + '\066', 0o10), ehT0Px3KOsy9(chr(48) + '\157' + '\063' + chr(2125 - 2070) + chr(48), 40768 - 40760), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(111) + chr(0b110101 + 0o0) + chr(1200 - 1146), 0b1000), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(7445 - 7334) + '\x33' + '\x34' + chr(516 - 465), 0b1000), ehT0Px3KOsy9(chr(48) + chr(10853 - 10742) + '\062' + chr(0b110101) + '\x31', 0b1000), ehT0Px3KOsy9(chr(1402 - 1354) + chr(5665 - 5554) + chr(49) + chr(0b11001 + 0o36), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + '\x31' + chr(0b110010) + chr(52), 55015 - 55007), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b10001 + 0o42) + chr(0b110110) + '\066', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b1010 + 0o51) + chr(49) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(1091 - 1043) + '\x6f' + chr(0b100100 + 0o15) + '\x30' + '\062', 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\x32' + chr(209 - 160) + chr(0b11100 + 0o24), 18820 - 18812), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\061' + chr(50) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(1221 - 1173) + '\x6f' + chr(1028 - 978) + chr(2247 - 2196) + chr(0b10001 + 0o45), 0o10), ehT0Px3KOsy9('\x30' + chr(0b10101 + 0o132) + '\066' + chr(54), 21952 - 21944), ehT0Px3KOsy9(chr(48) + chr(111) + chr(1475 - 1425) + '\x32', 22424 - 22416), ehT0Px3KOsy9(chr(861 - 813) + chr(111) + '\x36' + chr(2477 - 2426), ord("\x08")), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(0b1101111) + '\x31' + '\x35' + chr(53), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(495 - 384) + '\x31' + chr(0b110111) + '\067', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(49) + '\x33' + '\x36', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(1912 - 1801) + chr(50) + '\x31' + chr(586 - 533), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1675 - 1624) + chr(2494 - 2443) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(1566 - 1518) + chr(0b1100010 + 0o15) + '\x33' + '\061' + '\x34', 0o10), ehT0Px3KOsy9(chr(1466 - 1418) + '\x6f' + chr(0b10110 + 0o35) + chr(0b1000 + 0o54) + chr(0b1111 + 0o41), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b10110 + 0o40) + chr(202 - 149), 16953 - 16945), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110001) + chr(0b110110), 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(51) + chr(2085 - 2037) + '\064', 41661 - 41653), ehT0Px3KOsy9(chr(1245 - 1197) + '\x6f' + chr(54), 45127 - 45119), ehT0Px3KOsy9(chr(1756 - 1708) + chr(111) + chr(53) + '\066', 8), ehT0Px3KOsy9('\060' + '\157' + chr(0b110000 + 0o3) + chr(51) + chr(689 - 641), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(6995 - 6884) + chr(0b10000 + 0o42) + chr(1913 - 1859) + chr(55), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(2035 - 1986) + chr(53) + chr(55), 20368 - 20360), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b1110 + 0o51), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(307 - 257), 0b1000), ehT0Px3KOsy9(chr(2151 - 2103) + '\x6f' + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(835 - 787) + '\x6f' + '\x31' + chr(0b110110) + chr(2694 - 2641), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(111) + chr(2516 - 2463) + '\060', 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x11'), chr(0b11110 + 0o106) + chr(0b1010111 + 0o16) + chr(1391 - 1292) + '\x6f' + chr(0b110 + 0o136) + chr(6266 - 6165))(chr(6484 - 6367) + '\x74' + chr(0b1001100 + 0o32) + '\x2d' + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def gDZ9dM4p0W3G():
n4ljua2gi1Pr = kKmBjRbC6Zbc()
n4ljua2gi1Pr.jZh5_pLUoOoZ = ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(111) + '\x34', 60081 - 60073)
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9(chr(48) + '\157' + '\x31' + chr(0b110000) + chr(48) + '\x30', 0b1000)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9('\x30' + '\x6f' + chr(2386 - 2334) + '\060' + chr(0b100010 + 0o16) + '\060', ord("\x08"))
n4ljua2gi1Pr.r99iQzD4Y8i3 = ehT0Px3KOsy9('\x30' + chr(10129 - 10018) + chr(0b11011 + 0o27) + '\060' + chr(48), ord("\x08"))
n4ljua2gi1Pr.zHYy5k5t9tov = xafqLlk3kkUe(SXOLrMavuUCe(b'\r'), '\144' + chr(0b1100101) + '\143' + '\x6f' + chr(0b111011 + 0o51) + chr(4738 - 4637))('\165' + '\x74' + chr(0b1101 + 0o131) + chr(0b101101) + '\x38')
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_attention_moe_tiny
|
def attention_lm_attention_moe_tiny():
"""Cheap model for debugging.
Returns:
an hparams object.
"""
hparams = attention_lm_moe_small()
hparams.moe_layers = ""
hparams.attention_num_experts = 128
hparams.filter_size = 8192
hparams.attention_type = AttentionType.LOCAL_EXPERTS
return hparams
|
python
|
def attention_lm_attention_moe_tiny():
"""Cheap model for debugging.
Returns:
an hparams object.
"""
hparams = attention_lm_moe_small()
hparams.moe_layers = ""
hparams.attention_num_experts = 128
hparams.filter_size = 8192
hparams.attention_type = AttentionType.LOCAL_EXPERTS
return hparams
|
[
"def",
"attention_lm_attention_moe_tiny",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_small",
"(",
")",
"hparams",
".",
"moe_layers",
"=",
"\"\"",
"hparams",
".",
"attention_num_experts",
"=",
"128",
"hparams",
".",
"filter_size",
"=",
"8192",
"hparams",
".",
"attention_type",
"=",
"AttentionType",
".",
"LOCAL_EXPERTS",
"return",
"hparams"
] |
Cheap model for debugging.
Returns:
an hparams object.
|
[
"Cheap",
"model",
"for",
"debugging",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L666-L677
|
train
|
Cheap model for debugging.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(9345 - 9234) + chr(0b100 + 0o60) + chr(1772 - 1721), 10152 - 10144), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(111) + '\x33', 19782 - 19774), ehT0Px3KOsy9('\060' + chr(8510 - 8399) + chr(0b101001 + 0o11) + chr(436 - 387) + chr(235 - 180), 20742 - 20734), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + chr(0b1100 + 0o44) + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1100111 + 0o10) + chr(0b10111 + 0o33) + chr(0b110100) + chr(1447 - 1393), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + chr(48) + chr(1064 - 1013), 0o10), ehT0Px3KOsy9(chr(0b11001 + 0o27) + '\157' + chr(49) + chr(54) + '\x31', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(51) + '\x37', 0b1000), ehT0Px3KOsy9('\060' + chr(0b111011 + 0o64) + chr(0b110011) + chr(1149 - 1101) + chr(0b110010 + 0o4), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\062' + chr(0b110111) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(111) + chr(500 - 449) + '\x36' + chr(0b110110), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b11 + 0o56) + chr(866 - 812), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010) + chr(2059 - 2004) + chr(1661 - 1608), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(49) + '\067' + chr(0b110100 + 0o3), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1100001 + 0o16) + chr(0b1101 + 0o45) + chr(0b110001) + chr(52), 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + '\061' + chr(0b11011 + 0o34) + chr(55), 8), ehT0Px3KOsy9(chr(0b100001 + 0o17) + chr(111) + chr(1366 - 1316) + chr(49) + chr(50), 15042 - 15034), ehT0Px3KOsy9(chr(799 - 751) + chr(0b1101111) + '\x31' + chr(0b110111) + '\x35', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b10100 + 0o37) + chr(53) + chr(1230 - 1179), 0o10), ehT0Px3KOsy9(chr(477 - 429) + chr(5741 - 5630) + chr(0b110001) + '\x35' + chr(2634 - 2582), 0o10), ehT0Px3KOsy9(chr(2110 - 2062) + chr(111) + chr(0b11010 + 0o31) + '\063' + chr(0b110110), 56821 - 56813), ehT0Px3KOsy9(chr(0b110000) + chr(0b10001 + 0o136) + chr(473 - 423) + chr(0b110011) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b101 + 0o53) + '\x6f' + chr(0b11101 + 0o26) + chr(0b110010) + chr(1999 - 1946), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\x32' + '\x37' + '\065', 8), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110011) + chr(0b110000) + chr(51), 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(2302 - 2248) + '\061', 7667 - 7659), ehT0Px3KOsy9(chr(0b110000) + chr(0b10111 + 0o130) + '\061' + chr(654 - 605) + chr(54), 61272 - 61264), ehT0Px3KOsy9(chr(48) + chr(5194 - 5083) + chr(0b110010) + '\x33' + chr(51), 17338 - 17330), ehT0Px3KOsy9('\060' + '\157' + '\x31' + chr(48) + chr(0b110011), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + '\061' + chr(0b110100), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + chr(0b100110 + 0o13) + chr(0b100001 + 0o23), 8), ehT0Px3KOsy9('\x30' + '\157' + '\062' + chr(0b101 + 0o57) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1100111 + 0o10) + chr(0b110001) + chr(53) + chr(51), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b10111 + 0o130) + chr(0b101101 + 0o5) + chr(55), 0o10), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(0b1101111) + chr(0b110001) + chr(0b11011 + 0o26) + '\066', 8), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10010 + 0o37) + '\061', 0b1000), ehT0Px3KOsy9('\060' + chr(1024 - 913) + chr(526 - 471) + chr(1526 - 1477), 19062 - 19054), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x33' + chr(0b11111 + 0o21) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(50) + '\x34' + chr(609 - 559), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\062' + chr(0b100000 + 0o20) + chr(0b110111), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b0 + 0o65) + chr(0b110000), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\r'), chr(0b1001011 + 0o31) + '\145' + '\x63' + chr(4823 - 4712) + chr(5716 - 5616) + chr(101))('\x75' + chr(0b1110100) + '\146' + chr(45) + chr(0b11 + 0o65)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def eYvwUXaHB801():
n4ljua2gi1Pr = gDZ9dM4p0W3G()
n4ljua2gi1Pr.zHYy5k5t9tov = xafqLlk3kkUe(SXOLrMavuUCe(b''), chr(100) + '\145' + chr(111 - 12) + chr(111) + chr(5637 - 5537) + '\145')(chr(0b1011101 + 0o30) + '\164' + chr(102) + '\055' + chr(56))
n4ljua2gi1Pr.RRK2XjaZaWCZ = ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2072 - 2022) + '\x30' + chr(0b100110 + 0o12), 0b1000)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9(chr(1398 - 1350) + '\157' + chr(50) + '\060' + '\x30' + chr(0b10000 + 0o40) + chr(0b110000), ord("\x08"))
n4ljua2gi1Pr.lZ1GB4L2oMeG = mAAZP2Lf9x3V.LOCAL_EXPERTS
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_large
|
def attention_lm_moe_large():
"""Large model for distributed training.
Over 1B parameters, so requires multi-gpu training due to memory
requirements.
on lm1b_32k:
After 45K steps on 8 GPUs (synchronous):
eval_log_ppl_per_token = 3.18
eval_ppl_per_word = exp(1.107893 * eval_log_ppl_per_token) = 33.9
Returns:
an hparams object.
"""
hparams = attention_lm_moe_base()
hparams.num_hidden_layers = 5
hparams.moe_layers = "3"
hparams.hidden_size = 1024
hparams.num_heads = 16
hparams.filter_size = 4096
hparams.moe_hidden_sizes = "4096"
hparams.moe_num_experts = 128
hparams.layer_prepostprocess_dropout = 0.2
return hparams
|
python
|
def attention_lm_moe_large():
"""Large model for distributed training.
Over 1B parameters, so requires multi-gpu training due to memory
requirements.
on lm1b_32k:
After 45K steps on 8 GPUs (synchronous):
eval_log_ppl_per_token = 3.18
eval_ppl_per_word = exp(1.107893 * eval_log_ppl_per_token) = 33.9
Returns:
an hparams object.
"""
hparams = attention_lm_moe_base()
hparams.num_hidden_layers = 5
hparams.moe_layers = "3"
hparams.hidden_size = 1024
hparams.num_heads = 16
hparams.filter_size = 4096
hparams.moe_hidden_sizes = "4096"
hparams.moe_num_experts = 128
hparams.layer_prepostprocess_dropout = 0.2
return hparams
|
[
"def",
"attention_lm_moe_large",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_base",
"(",
")",
"hparams",
".",
"num_hidden_layers",
"=",
"5",
"hparams",
".",
"moe_layers",
"=",
"\"3\"",
"hparams",
".",
"hidden_size",
"=",
"1024",
"hparams",
".",
"num_heads",
"=",
"16",
"hparams",
".",
"filter_size",
"=",
"4096",
"hparams",
".",
"moe_hidden_sizes",
"=",
"\"4096\"",
"hparams",
".",
"moe_num_experts",
"=",
"128",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.2",
"return",
"hparams"
] |
Large model for distributed training.
Over 1B parameters, so requires multi-gpu training due to memory
requirements.
on lm1b_32k:
After 45K steps on 8 GPUs (synchronous):
eval_log_ppl_per_token = 3.18
eval_ppl_per_word = exp(1.107893 * eval_log_ppl_per_token) = 33.9
Returns:
an hparams object.
|
[
"Large",
"model",
"for",
"distributed",
"training",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L699-L722
|
train
|
Large model for distributed training.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b10100 + 0o41) + '\x36', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b1011 + 0o46) + chr(2436 - 2384) + chr(512 - 464), 0o10), ehT0Px3KOsy9(chr(724 - 676) + chr(0b1101111) + chr(0b110010) + chr(163 - 110) + chr(738 - 690), 0b1000), ehT0Px3KOsy9(chr(1619 - 1571) + chr(111) + '\062' + chr(0b110101 + 0o0) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9('\060' + chr(10395 - 10284) + chr(51) + chr(0b1011 + 0o47) + chr(252 - 200), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + '\064' + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b100011 + 0o17) + chr(51) + '\065', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(9728 - 9617) + '\062' + chr(762 - 712), 32674 - 32666), ehT0Px3KOsy9('\x30' + chr(5182 - 5071) + chr(2093 - 2043) + chr(592 - 538) + '\x31', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + chr(0b110 + 0o61) + chr(52), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(50) + '\064' + chr(0b101000 + 0o16), 0b1000), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b1101111) + chr(0b1 + 0o62) + '\060', ord("\x08")), ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(0b110011 + 0o74) + chr(49) + chr(978 - 929) + chr(1008 - 954), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(51) + chr(0b110010) + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(2843 - 2788), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33' + '\063' + '\061', 0o10), ehT0Px3KOsy9(chr(0b110000 + 0o0) + '\x6f' + chr(49) + '\060' + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\063' + '\x31' + '\x37', 5308 - 5300), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(50) + '\x31' + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(50) + chr(0b11 + 0o57) + '\067', 37981 - 37973), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(2154 - 2105) + chr(0b110001) + chr(1360 - 1312), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(1899 - 1849) + chr(0b11100 + 0o26), 8), ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(111) + chr(0b10011 + 0o37) + '\066' + chr(0b11101 + 0o24), 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(53) + chr(1603 - 1554), 63174 - 63166), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b101111 + 0o2) + chr(0b110000) + chr(398 - 343), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x31' + chr(52) + chr(0b11100 + 0o25), 0o10), ehT0Px3KOsy9('\060' + chr(6294 - 6183) + chr(51) + chr(0b110100) + '\067', 0o10), ehT0Px3KOsy9(chr(0b111 + 0o51) + chr(9791 - 9680) + chr(0b110001) + '\x34' + '\064', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b0 + 0o62) + chr(933 - 884) + '\062', 8), ehT0Px3KOsy9(chr(48) + chr(111) + chr(585 - 534) + '\x35' + chr(1796 - 1746), 35598 - 35590), ehT0Px3KOsy9(chr(48) + chr(0b1011100 + 0o23) + chr(863 - 813) + chr(52) + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b100110 + 0o14) + '\x36' + chr(952 - 902), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101100 + 0o3) + '\066' + chr(0b1101 + 0o45), ord("\x08")), ehT0Px3KOsy9(chr(0b101 + 0o53) + chr(0b1101111) + chr(2690 - 2636) + chr(52), 61944 - 61936), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(5689 - 5578) + '\062' + chr(53), 0o10), ehT0Px3KOsy9(chr(0b11011 + 0o25) + '\x6f' + '\063' + chr(2045 - 1995) + '\067', 8), ehT0Px3KOsy9('\060' + '\157' + chr(1691 - 1642) + '\x30' + '\062', ord("\x08")), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(0b1011000 + 0o27) + chr(0b1110 + 0o44) + chr(51), 63510 - 63502), ehT0Px3KOsy9(chr(88 - 40) + chr(4476 - 4365) + '\x32' + '\067' + chr(0b101101 + 0o6), 0b1000), ehT0Px3KOsy9(chr(782 - 734) + chr(0b1100100 + 0o13) + '\x37' + chr(53), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b10100 + 0o41) + '\060', 43308 - 43300)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x00'), chr(967 - 867) + chr(101) + chr(99) + '\157' + chr(2559 - 2459) + '\145')('\165' + chr(116) + '\146' + chr(1266 - 1221) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def cgFL9Y_MXsyr():
n4ljua2gi1Pr = kKmBjRbC6Zbc()
n4ljua2gi1Pr.jZh5_pLUoOoZ = ehT0Px3KOsy9('\060' + chr(240 - 129) + chr(53), 0b1000)
n4ljua2gi1Pr.zHYy5k5t9tov = xafqLlk3kkUe(SXOLrMavuUCe(b'\x1d'), chr(9071 - 8971) + '\x65' + chr(99) + chr(111) + chr(0b1100100) + chr(101))(chr(0b1110101) + '\x74' + chr(0b1100110) + chr(738 - 693) + '\x38')
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(1804 - 1754) + '\x30' + '\x30' + chr(598 - 550), ord("\x08"))
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + chr(48), 0o10)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9(chr(462 - 414) + chr(9839 - 9728) + '\061' + chr(48) + chr(0b10100 + 0o34) + chr(0b110000) + '\060', 60314 - 60306)
n4ljua2gi1Pr.GlalSbqmunyH = xafqLlk3kkUe(SXOLrMavuUCe(b'\x1aP\x91\x9b'), chr(100) + '\145' + '\x63' + '\157' + chr(0b1100100) + '\x65')(chr(117) + '\x74' + chr(102) + chr(0b101101) + chr(2356 - 2300))
n4ljua2gi1Pr.r99iQzD4Y8i3 = ehT0Px3KOsy9(chr(295 - 247) + chr(0b110000 + 0o77) + chr(990 - 940) + chr(48) + '\060', 50097 - 50089)
n4ljua2gi1Pr.RW_xSzp18UeS = 0.2
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_memory_efficient
|
def attention_lm_moe_memory_efficient():
"""Memory-efficient version."""
hparams = attention_lm_moe_large()
hparams.diet_experts = True
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
hparams.layer_prepostprocess_dropout = 0.0
hparams.memory_efficient_ffn = True
hparams.attention_type = AttentionType.MEMORY_EFFICIENT
hparams.num_heads = 8
hparams.factored_logits = True
return hparams
|
python
|
def attention_lm_moe_memory_efficient():
"""Memory-efficient version."""
hparams = attention_lm_moe_large()
hparams.diet_experts = True
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
hparams.layer_prepostprocess_dropout = 0.0
hparams.memory_efficient_ffn = True
hparams.attention_type = AttentionType.MEMORY_EFFICIENT
hparams.num_heads = 8
hparams.factored_logits = True
return hparams
|
[
"def",
"attention_lm_moe_memory_efficient",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_large",
"(",
")",
"hparams",
".",
"diet_experts",
"=",
"True",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"n\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"da\"",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.0",
"hparams",
".",
"memory_efficient_ffn",
"=",
"True",
"hparams",
".",
"attention_type",
"=",
"AttentionType",
".",
"MEMORY_EFFICIENT",
"hparams",
".",
"num_heads",
"=",
"8",
"hparams",
".",
"factored_logits",
"=",
"True",
"return",
"hparams"
] |
Memory-efficient version.
|
[
"Memory",
"-",
"efficient",
"version",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L733-L744
|
train
|
Memory - efficient version.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(202 - 153) + chr(0b110001), 0b1000), ehT0Px3KOsy9('\060' + chr(5958 - 5847) + chr(49) + chr(54) + '\062', 0o10), ehT0Px3KOsy9(chr(1571 - 1523) + chr(0b1101111) + '\063' + chr(1241 - 1190) + '\066', ord("\x08")), ehT0Px3KOsy9(chr(1783 - 1735) + chr(0b1101111) + chr(925 - 872), 0o10), ehT0Px3KOsy9(chr(0b110 + 0o52) + '\157' + chr(0b110001) + '\063' + chr(0b110010 + 0o3), ord("\x08")), ehT0Px3KOsy9(chr(2110 - 2062) + '\157' + '\063' + chr(0b110111) + chr(48), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1011111 + 0o20) + '\066' + '\x31', ord("\x08")), ehT0Px3KOsy9(chr(284 - 236) + chr(0b10 + 0o155) + chr(0b101110 + 0o5) + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(2026 - 1978) + chr(10260 - 10149) + '\064' + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010110 + 0o31) + chr(0b110001) + chr(52) + '\x32', 26157 - 26149), ehT0Px3KOsy9(chr(884 - 836) + chr(0b1101111) + '\x32' + chr(2130 - 2082) + chr(2777 - 2724), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + '\061' + chr(2516 - 2464) + '\x36', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b100101 + 0o112) + '\x32' + chr(0b11111 + 0o22) + chr(0b100111 + 0o17), 0o10), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(2542 - 2431) + chr(277 - 226) + chr(49) + chr(1964 - 1910), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b101110 + 0o5) + chr(0b110110) + chr(0b101000 + 0o16), 0b1000), ehT0Px3KOsy9(chr(505 - 457) + chr(0b1011011 + 0o24) + chr(51) + '\064' + '\064', 0o10), ehT0Px3KOsy9(chr(67 - 19) + chr(111) + '\062' + '\060' + '\x35', 8), ehT0Px3KOsy9(chr(0b110000) + chr(9022 - 8911) + chr(0b110011) + chr(49) + chr(602 - 550), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x32' + chr(51) + chr(50), 0o10), ehT0Px3KOsy9(chr(48) + chr(10604 - 10493) + chr(49) + chr(1226 - 1178) + chr(0b1101 + 0o44), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b10111 + 0o33) + chr(0b1111 + 0o42) + '\x34', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1001 + 0o146) + chr(0b100010 + 0o21) + '\061' + '\063', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(4237 - 4126) + chr(49) + '\x36' + chr(54), 37292 - 37284), ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(0b100001 + 0o116) + '\061' + '\x37' + '\065', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x37' + '\064', 54329 - 54321), ehT0Px3KOsy9('\060' + '\x6f' + '\061' + chr(1105 - 1053) + chr(1758 - 1706), 0b1000), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(4355 - 4244) + chr(50) + chr(0b110101) + chr(0b110100), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1100110 + 0o11) + '\061' + chr(52) + '\065', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b10 + 0o155) + '\061' + chr(0b1010 + 0o53) + chr(0b100000 + 0o20), 40464 - 40456), ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(111) + chr(0b11110 + 0o25) + chr(0b101101 + 0o12) + '\060', 8), ehT0Px3KOsy9(chr(124 - 76) + chr(7393 - 7282) + chr(1359 - 1308) + chr(362 - 308) + chr(1682 - 1627), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(50) + '\065' + chr(54), 0b1000), ehT0Px3KOsy9('\060' + '\157' + '\x32' + chr(1508 - 1453) + chr(0b100001 + 0o24), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b111110 + 0o61) + '\062' + '\063', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(49) + chr(1183 - 1129), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b101100 + 0o103) + '\066' + '\x31', 8), ehT0Px3KOsy9(chr(48) + chr(111) + '\067' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(2100 - 2052) + '\x6f' + chr(1250 - 1199) + chr(0b110001) + chr(53), 57953 - 57945), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110100) + '\064', 0b1000), ehT0Px3KOsy9('\060' + chr(0b110000 + 0o77) + '\x30', 41461 - 41453)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b101101 + 0o10) + chr(1212 - 1164), 1338 - 1330)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b"'"), chr(0b1011 + 0o131) + chr(9794 - 9693) + chr(6011 - 5912) + '\x6f' + '\144' + chr(101))(chr(117) + chr(116) + chr(2427 - 2325) + '\055' + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def Hj1tm8ecSoLA():
n4ljua2gi1Pr = cgFL9Y_MXsyr()
n4ljua2gi1Pr.chv9ul4_iQE1 = ehT0Px3KOsy9('\060' + '\157' + chr(129 - 80), 0b1000)
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'g'), '\144' + chr(0b1100101) + chr(0b101001 + 0o72) + chr(0b1001000 + 0o47) + '\x64' + '\x65')(chr(0b1110101) + chr(1360 - 1244) + '\146' + chr(1993 - 1948) + '\070')
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'm\x13'), chr(100) + chr(101) + '\143' + '\x6f' + chr(0b1100100) + '\x65')(chr(117) + '\x74' + chr(102) + chr(985 - 940) + '\x38')
n4ljua2gi1Pr.RW_xSzp18UeS = 0.0
n4ljua2gi1Pr.Cx5KXzfumeJI = ehT0Px3KOsy9('\x30' + '\x6f' + '\061', 8)
n4ljua2gi1Pr.lZ1GB4L2oMeG = mAAZP2Lf9x3V.MEMORY_EFFICIENT
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(9392 - 9281) + '\x31' + chr(1620 - 1572), 0o10)
n4ljua2gi1Pr.Sjw4kbDQvNmN = ehT0Px3KOsy9(chr(48) + '\x6f' + chr(49), 8)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_24b_diet
|
def attention_lm_moe_24b_diet():
"""Unnecessarily large model with 24B params - because we can."""
hparams = attention_lm_moe_large_diet()
hparams.moe_hidden_sizes = "12288"
hparams.moe_num_experts = 1024
hparams.batch_size = 4096
return hparams
|
python
|
def attention_lm_moe_24b_diet():
"""Unnecessarily large model with 24B params - because we can."""
hparams = attention_lm_moe_large_diet()
hparams.moe_hidden_sizes = "12288"
hparams.moe_num_experts = 1024
hparams.batch_size = 4096
return hparams
|
[
"def",
"attention_lm_moe_24b_diet",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_large_diet",
"(",
")",
"hparams",
".",
"moe_hidden_sizes",
"=",
"\"12288\"",
"hparams",
".",
"moe_num_experts",
"=",
"1024",
"hparams",
".",
"batch_size",
"=",
"4096",
"return",
"hparams"
] |
Unnecessarily large model with 24B params - because we can.
|
[
"Unnecessarily",
"large",
"model",
"with",
"24B",
"params",
"-",
"because",
"we",
"can",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L757-L763
|
train
|
Unnecessarily large model with 24B params - because we can t do it.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(7759 - 7648) + chr(0b110011) + chr(0b10111 + 0o32), 0b1000), ehT0Px3KOsy9('\x30' + chr(754 - 643) + chr(0b110001 + 0o0) + chr(2121 - 2073), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(0b110111) + chr(0b100001 + 0o20), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101100 + 0o3) + '\x31' + chr(0b110100) + '\066', 12329 - 12321), ehT0Px3KOsy9(chr(48) + chr(0b1001010 + 0o45) + '\061' + chr(896 - 847) + chr(723 - 668), 0b1000), ehT0Px3KOsy9(chr(870 - 822) + chr(0b111111 + 0o60) + '\x33' + '\x36' + chr(1514 - 1461), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\061' + '\062' + chr(0b101000 + 0o16), 0o10), ehT0Px3KOsy9(chr(1780 - 1732) + chr(111) + '\063' + chr(0b110 + 0o60) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(262 - 214) + '\x6f' + chr(0b11 + 0o60) + '\x30' + chr(1804 - 1754), 63089 - 63081), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b100000 + 0o26) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110010) + chr(1196 - 1143) + '\x37', 0o10), ehT0Px3KOsy9(chr(2226 - 2178) + chr(0b110111 + 0o70) + chr(1492 - 1443) + chr(0b110110) + chr(0b101011 + 0o6), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\062' + chr(0b110100) + '\062', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + '\x31' + chr(2176 - 2122), 27845 - 27837), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(0b10111 + 0o130) + chr(0b110100) + chr(2427 - 2376), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(1280 - 1169) + '\061' + chr(52) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(52) + chr(0b10010 + 0o40), 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\063' + chr(52) + chr(0b1000 + 0o55), 0o10), ehT0Px3KOsy9(chr(0b11000 + 0o30) + '\x6f' + chr(0b101 + 0o54) + chr(0b110010 + 0o0) + chr(0b10010 + 0o44), 8), ehT0Px3KOsy9('\060' + chr(12312 - 12201) + chr(2482 - 2432) + '\064' + '\x32', 8), ehT0Px3KOsy9(chr(0b10001 + 0o37) + '\157' + chr(1832 - 1783) + '\060', 8), ehT0Px3KOsy9('\x30' + chr(10571 - 10460) + chr(0b10101 + 0o35) + '\x37' + chr(0b101100 + 0o6), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + chr(54) + chr(0b101 + 0o53), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b101100 + 0o5) + chr(48) + chr(0b1101 + 0o50), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110011) + chr(51) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(53) + chr(2265 - 2213), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + '\065' + chr(48), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\x33' + chr(0b110110) + chr(50), 0o10), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(111) + chr(50) + '\065' + '\066', 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\062' + chr(0b101001 + 0o7) + '\067', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(5609 - 5498) + chr(0b110001) + '\x33' + chr(2486 - 2432), ord("\x08")), ehT0Px3KOsy9(chr(2078 - 2030) + chr(1196 - 1085) + chr(50) + '\061' + '\064', 30220 - 30212), ehT0Px3KOsy9(chr(1485 - 1437) + chr(0b110 + 0o151) + chr(0b110011) + '\x36' + chr(1329 - 1277), 15107 - 15099), ehT0Px3KOsy9(chr(479 - 431) + chr(9583 - 9472) + '\066' + chr(2058 - 2009), 0o10), ehT0Px3KOsy9(chr(578 - 530) + chr(0b1101111) + chr(0b11101 + 0o26) + chr(48) + chr(2855 - 2801), 12115 - 12107), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x32' + chr(0b101100 + 0o11) + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b10011 + 0o37) + chr(49) + chr(2063 - 2009), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\061' + '\067' + chr(48), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1001000 + 0o47) + chr(0b10101 + 0o36) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(0b100000 + 0o20) + '\x6f' + chr(50) + chr(964 - 910) + chr(0b110001), 20911 - 20903)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + '\x6f' + chr(2581 - 2528) + chr(0b1 + 0o57), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'7'), chr(100) + chr(4779 - 4678) + '\x63' + '\x6f' + chr(2224 - 2124) + '\x65')(chr(1260 - 1143) + '\x74' + chr(0b1000011 + 0o43) + chr(0b101101) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def UgzbMcgb43hL():
n4ljua2gi1Pr = f1PxCwJWppTY()
n4ljua2gi1Pr.GlalSbqmunyH = xafqLlk3kkUe(SXOLrMavuUCe(b'(:\x8b^w'), chr(100) + '\x65' + chr(0b1100011) + '\157' + chr(0b1100100) + chr(0b1100101))(chr(117) + chr(0b1110100) + chr(7412 - 7310) + chr(1522 - 1477) + chr(0b111000))
n4ljua2gi1Pr.r99iQzD4Y8i3 = ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(0b101001 + 0o7) + chr(48) + '\060', 0b1000)
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(0b110000) + chr(0b1000 + 0o147) + chr(49) + chr(48) + chr(0b110000) + chr(48) + chr(0b110000), 0o10)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_translation
|
def attention_lm_moe_translation():
"""Version to use for seq2seq."""
hparams = attention_lm_moe_base()
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
hparams.learning_rate = 0.4
hparams.prepend_mode = "prepend_inputs_masked_attention"
hparams.max_length = 512
hparams.label_smoothing = 0.1
hparams.layer_prepostprocess_dropout = 0.2
hparams.num_hidden_layers = 6
hparams.moe_layers = "0,1,2,3,4,5"
hparams.shared_embedding_and_softmax_weights = True
return hparams
|
python
|
def attention_lm_moe_translation():
"""Version to use for seq2seq."""
hparams = attention_lm_moe_base()
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
hparams.learning_rate = 0.4
hparams.prepend_mode = "prepend_inputs_masked_attention"
hparams.max_length = 512
hparams.label_smoothing = 0.1
hparams.layer_prepostprocess_dropout = 0.2
hparams.num_hidden_layers = 6
hparams.moe_layers = "0,1,2,3,4,5"
hparams.shared_embedding_and_softmax_weights = True
return hparams
|
[
"def",
"attention_lm_moe_translation",
"(",
")",
":",
"hparams",
"=",
"attention_lm_moe_base",
"(",
")",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"n\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"da\"",
"hparams",
".",
"learning_rate",
"=",
"0.4",
"hparams",
".",
"prepend_mode",
"=",
"\"prepend_inputs_masked_attention\"",
"hparams",
".",
"max_length",
"=",
"512",
"hparams",
".",
"label_smoothing",
"=",
"0.1",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.2",
"hparams",
".",
"num_hidden_layers",
"=",
"6",
"hparams",
".",
"moe_layers",
"=",
"\"0,1,2,3,4,5\"",
"hparams",
".",
"shared_embedding_and_softmax_weights",
"=",
"True",
"return",
"hparams"
] |
Version to use for seq2seq.
|
[
"Version",
"to",
"use",
"for",
"seq2seq",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L767-L780
|
train
|
Version to use for seq2seq.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(1081 - 970) + '\067' + chr(1749 - 1699), ord("\x08")), ehT0Px3KOsy9(chr(1180 - 1132) + chr(10460 - 10349) + '\062' + chr(1207 - 1154) + chr(55), 48025 - 48017), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b100011 + 0o16) + '\x34' + chr(54), 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\x34' + '\061', 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(0b100011 + 0o17) + '\x36', 0o10), ehT0Px3KOsy9(chr(48) + chr(8014 - 7903) + chr(50) + '\062' + chr(1337 - 1284), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2034 - 1983) + chr(0b11110 + 0o26) + chr(0b1111 + 0o42), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\066' + chr(912 - 857), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011 + 0o144) + chr(49) + '\066' + '\063', 47041 - 47033), ehT0Px3KOsy9(chr(872 - 824) + chr(111) + chr(50) + chr(51), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(49) + '\064' + chr(0b110100), 53481 - 53473), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(0b1101111) + '\062' + '\066' + chr(2647 - 2592), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(1766 - 1716) + chr(50) + chr(50), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b100100 + 0o113) + chr(0b11 + 0o57) + '\x37' + chr(54), 0b1000), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(111) + chr(50) + '\x34' + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + chr(0b110011) + chr(48), 0o10), ehT0Px3KOsy9(chr(0b100000 + 0o20) + '\157' + chr(2492 - 2441) + chr(48) + '\066', ord("\x08")), ehT0Px3KOsy9('\060' + chr(10976 - 10865) + chr(939 - 884) + chr(0b1001 + 0o53), 2641 - 2633), ehT0Px3KOsy9('\x30' + chr(0b111011 + 0o64) + chr(50) + chr(0b110110) + '\x30', 19887 - 19879), ehT0Px3KOsy9(chr(239 - 191) + chr(111) + '\062' + chr(48) + chr(55), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(498 - 447) + chr(0b110001) + '\x30', 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\x31' + '\x35' + chr(55), 0b1000), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(6761 - 6650) + '\062' + chr(0b110000) + chr(0b110000), 41927 - 41919), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(0b1001000 + 0o47) + chr(1650 - 1601) + '\x31' + chr(52), 53604 - 53596), ehT0Px3KOsy9(chr(197 - 149) + chr(0b1101111) + chr(0b110010) + chr(0b10101 + 0o37) + chr(2362 - 2311), 16094 - 16086), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(0b1000011 + 0o54) + chr(49) + chr(995 - 943) + chr(1300 - 1249), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1958 - 1908) + chr(0b110001), 58903 - 58895), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b0 + 0o62) + chr(0b110011), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100100 + 0o13) + chr(0b110010) + '\x35' + '\063', 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1469 - 1420) + chr(972 - 918), 0b1000), ehT0Px3KOsy9(chr(557 - 509) + chr(130 - 19) + '\063' + '\x37' + chr(0b110010), 29280 - 29272), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\064' + '\061', 8), ehT0Px3KOsy9('\x30' + chr(0b1010110 + 0o31) + chr(627 - 577) + chr(0b110110) + chr(0b110010), 64391 - 64383), ehT0Px3KOsy9('\x30' + chr(4362 - 4251) + '\x37' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x32' + chr(52) + chr(638 - 588), 8), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b10100 + 0o36) + chr(82 - 32) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(0b10101 + 0o33) + '\x6f' + chr(49) + '\061' + '\x33', 54644 - 54636), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110100) + chr(0b101110 + 0o4), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(401 - 290) + chr(51) + chr(1221 - 1166) + chr(423 - 370), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b101000 + 0o10) + '\x6f' + chr(0b110101) + chr(1541 - 1493), 56142 - 56134)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x14'), '\x64' + chr(4499 - 4398) + chr(0b1010111 + 0o14) + chr(8083 - 7972) + '\144' + chr(101))('\x75' + chr(1470 - 1354) + '\146' + chr(0b101011 + 0o2) + chr(634 - 578)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def hPojHAeHCidc():
n4ljua2gi1Pr = kKmBjRbC6Zbc()
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'T'), chr(0b1010 + 0o132) + '\x65' + chr(99) + '\157' + chr(100) + '\145')(chr(117) + chr(6766 - 6650) + chr(1621 - 1519) + chr(45) + chr(0b1001 + 0o57))
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'^\xce'), chr(100) + chr(101) + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + chr(5582 - 5481))(chr(117) + '\164' + chr(0b1011010 + 0o14) + chr(0b1000 + 0o45) + chr(0b111000))
n4ljua2gi1Pr.QGSIpd_yUNzU = 0.4
n4ljua2gi1Pr.qr_7laJirAn2 = xafqLlk3kkUe(SXOLrMavuUCe(b'J\xdd\x03Qs\xfb\x85h\x1a}\x14\xc2\xe6\xcd\xfd\xf5\xa7O\xdb\xe2\xac0\xe1q8\xc6\rez\xc4`'), chr(0b100111 + 0o75) + chr(0b1011 + 0o132) + chr(99) + '\157' + chr(0b1100100) + chr(6431 - 6330))('\165' + chr(116) + chr(102) + chr(346 - 301) + '\x38')
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9('\060' + '\x6f' + chr(49) + '\060' + '\060' + '\060', 5768 - 5760)
n4ljua2gi1Pr.FSjUgdaczzRk = 0.1
n4ljua2gi1Pr.RW_xSzp18UeS = 0.2
n4ljua2gi1Pr.jZh5_pLUoOoZ = ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\066', 52252 - 52244)
n4ljua2gi1Pr.zHYy5k5t9tov = xafqLlk3kkUe(SXOLrMavuUCe(b'\n\x83W\r$\xb9\xd2\x1bG?Q'), chr(8060 - 7960) + '\x65' + '\143' + chr(0b1000100 + 0o53) + '\x64' + '\145')(chr(117) + chr(116) + chr(0b1100110) + chr(0b101101) + chr(56))
n4ljua2gi1Pr.qVamxim0L2I1 = ehT0Px3KOsy9('\x30' + '\157' + chr(0b110001), 0o10)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/research/attention_lm_moe.py
|
attention_lm_moe_unscramble_base
|
def attention_lm_moe_unscramble_base():
"""Version to use with languagemodel_wiki_scramble1k50."""
hparams = attention_lm_no_moe_small()
hparams.use_inputs = True
hparams.min_length_bucket = 1024
hparams.max_length = 1024
hparams.batch_size = 5000
hparams.layer_prepostprocess_dropout = 0.0
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
return hparams
|
python
|
def attention_lm_moe_unscramble_base():
"""Version to use with languagemodel_wiki_scramble1k50."""
hparams = attention_lm_no_moe_small()
hparams.use_inputs = True
hparams.min_length_bucket = 1024
hparams.max_length = 1024
hparams.batch_size = 5000
hparams.layer_prepostprocess_dropout = 0.0
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
return hparams
|
[
"def",
"attention_lm_moe_unscramble_base",
"(",
")",
":",
"hparams",
"=",
"attention_lm_no_moe_small",
"(",
")",
"hparams",
".",
"use_inputs",
"=",
"True",
"hparams",
".",
"min_length_bucket",
"=",
"1024",
"hparams",
".",
"max_length",
"=",
"1024",
"hparams",
".",
"batch_size",
"=",
"5000",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.0",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"n\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"da\"",
"return",
"hparams"
] |
Version to use with languagemodel_wiki_scramble1k50.
|
[
"Version",
"to",
"use",
"with",
"languagemodel_wiki_scramble1k50",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/research/attention_lm_moe.py#L784-L794
|
train
|
Version to use with languagemodel_wiki_scramble1k50.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + '\157' + chr(52) + chr(1363 - 1313), 19110 - 19102), ehT0Px3KOsy9('\060' + chr(111) + '\x32' + chr(0b110011) + chr(52), 0b1000), ehT0Px3KOsy9(chr(910 - 862) + chr(111) + chr(1619 - 1570) + chr(53) + chr(0b10111 + 0o34), 433 - 425), ehT0Px3KOsy9(chr(0b110000) + chr(0b1001010 + 0o45) + chr(0b1001 + 0o52) + chr(0b110111) + '\060', 0b1000), ehT0Px3KOsy9('\060' + chr(7107 - 6996) + chr(50) + '\064' + chr(51), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(50) + '\061' + '\061', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + chr(0b11110 + 0o24) + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\064' + chr(0b110010), 8), ehT0Px3KOsy9(chr(0b110000) + chr(4766 - 4655) + chr(0b100001 + 0o20) + chr(0b110101) + '\060', 0o10), ehT0Px3KOsy9(chr(314 - 266) + '\157' + '\x34' + '\062', 8), ehT0Px3KOsy9('\060' + '\157' + '\063' + '\x33' + '\063', 49285 - 49277), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + chr(52) + chr(1254 - 1205), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(0b100010 + 0o17) + '\x31', 59010 - 59002), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110011) + '\x32' + chr(0b110000), 14950 - 14942), ehT0Px3KOsy9(chr(1834 - 1786) + chr(0b1101 + 0o142) + '\061' + chr(51) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\061' + chr(0b101101 + 0o6) + chr(54), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110011) + chr(1390 - 1340) + '\x30', 8), ehT0Px3KOsy9('\060' + chr(0b100 + 0o153) + chr(50) + '\x34' + chr(906 - 852), ord("\x08")), ehT0Px3KOsy9(chr(0b1000 + 0o50) + '\x6f' + chr(0b110001) + '\060' + '\x37', ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + '\064' + chr(0b111 + 0o60), 21568 - 21560), ehT0Px3KOsy9(chr(48) + chr(0b1100100 + 0o13) + chr(0b110001) + chr(0b110110) + '\x33', ord("\x08")), ehT0Px3KOsy9('\060' + chr(4822 - 4711) + '\x32' + chr(0b101100 + 0o10) + chr(2149 - 2100), 8), ehT0Px3KOsy9(chr(1075 - 1027) + chr(111) + chr(0b11011 + 0o33), 11301 - 11293), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(6556 - 6445) + chr(0b110001) + chr(0b110101) + chr(0b0 + 0o62), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(978 - 929) + '\065', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(49) + chr(0b110101) + chr(53), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\067' + '\063', ord("\x08")), ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(0b1011001 + 0o26) + chr(1334 - 1279) + '\063', 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x31' + chr(49), 0o10), ehT0Px3KOsy9('\060' + chr(12078 - 11967) + chr(50) + chr(0b1011 + 0o52) + '\x30', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(50) + '\x36' + chr(0b110111), 17866 - 17858), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(0b1101111) + chr(55) + '\x30', 0o10), ehT0Px3KOsy9(chr(1215 - 1167) + chr(0b11000 + 0o127) + chr(718 - 665) + '\063', 0b1000), ehT0Px3KOsy9('\060' + chr(5918 - 5807) + chr(49) + chr(51) + chr(0b110000), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1303 - 1252) + chr(49) + chr(2037 - 1988), 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(617 - 566) + chr(2224 - 2172) + chr(53), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x32' + '\x35', ord("\x08")), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(4948 - 4837) + chr(0b100011 + 0o16) + chr(53) + chr(49), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b10 + 0o60) + chr(0b110011) + chr(0b100101 + 0o22), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(1857 - 1746) + chr(0b101011 + 0o10) + chr(0b110000 + 0o2) + chr(49), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000 + 0o0) + '\157' + '\065' + chr(0b110000), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x0c'), '\x64' + chr(0b100011 + 0o102) + '\x63' + chr(111) + chr(100) + chr(0b1100101))(chr(0b101111 + 0o106) + '\164' + '\146' + chr(45) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def RfIKP2DOXFxL():
n4ljua2gi1Pr = CI4opAsPvodp()
n4ljua2gi1Pr.M4kXVSW2c4p9 = ehT0Px3KOsy9(chr(1947 - 1899) + chr(10306 - 10195) + chr(823 - 774), 0b1000)
n4ljua2gi1Pr.lhJm4Z32JlM2 = ehT0Px3KOsy9(chr(1081 - 1033) + '\x6f' + '\062' + chr(0b10100 + 0o34) + chr(0b110000) + chr(201 - 153), 37444 - 37436)
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9('\060' + chr(111) + '\062' + chr(0b100101 + 0o13) + '\060' + chr(436 - 388), 8)
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(1268 - 1219) + '\x31' + chr(1744 - 1690) + '\x31' + chr(1789 - 1741), 0b1000)
n4ljua2gi1Pr.RW_xSzp18UeS = 0.0
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'L'), chr(0b110111 + 0o55) + chr(0b111101 + 0o50) + '\143' + '\x6f' + '\x64' + '\x65')(chr(0b1110101) + chr(0b101000 + 0o114) + '\146' + '\055' + chr(0b111000))
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'F\xfd'), '\144' + chr(0b1100101) + chr(0b100 + 0o137) + '\x6f' + chr(3950 - 3850) + chr(101))(chr(117) + chr(0b1110100) + chr(0b1100110) + '\x2d' + chr(56))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
audio_bottom
|
def audio_bottom(x, model_hparams, vocab_size):
"""Transform input from data space to model space.
Args:
x: A Tensor with shape [batch, ...]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
body_input: A Tensor with shape [batch, ?, ?,
model_hparams.hidden_size].
"""
del vocab_size # unused arg
inputs = x
with tf.variable_scope("audio_modality"):
# TODO(aidangomez): Will need to sort out a better audio pipeline
def xnet_resblock(x, filters, res_relu, name):
"""Xception block."""
with tf.variable_scope(name):
# Typically audio samples are >100k samples in length and have a width
# of 2 or 4. Mono audio has a single channel while stereo has 2.
y = common_layers.separable_conv_block(
x,
filters, [((1, 1), (3, 3)), ((1, 1), (3, 3))],
first_relu=True,
padding="SAME",
force2d=True,
name="sep_conv_block")
y = common_layers.pool(y, (3, 3), "MAX", "SAME", strides=(2, 2))
return y + common_layers.conv_block(
x,
filters, [((1, 1), (1, 1))],
padding="SAME",
strides=(2, 2),
first_relu=res_relu,
force2d=True,
name="res_conv0")
x = tf.to_float(inputs) / 255.
x.set_shape([None, None, None, 1])
for i in range(model_hparams.audio_compression):
x = xnet_resblock(x, 2**(i + 1), True, "compress_block_%d" % i)
return xnet_resblock(x,
model_hparams.hidden_size,
False,
"compress_block_final")
|
python
|
def audio_bottom(x, model_hparams, vocab_size):
"""Transform input from data space to model space.
Args:
x: A Tensor with shape [batch, ...]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
body_input: A Tensor with shape [batch, ?, ?,
model_hparams.hidden_size].
"""
del vocab_size # unused arg
inputs = x
with tf.variable_scope("audio_modality"):
# TODO(aidangomez): Will need to sort out a better audio pipeline
def xnet_resblock(x, filters, res_relu, name):
"""Xception block."""
with tf.variable_scope(name):
# Typically audio samples are >100k samples in length and have a width
# of 2 or 4. Mono audio has a single channel while stereo has 2.
y = common_layers.separable_conv_block(
x,
filters, [((1, 1), (3, 3)), ((1, 1), (3, 3))],
first_relu=True,
padding="SAME",
force2d=True,
name="sep_conv_block")
y = common_layers.pool(y, (3, 3), "MAX", "SAME", strides=(2, 2))
return y + common_layers.conv_block(
x,
filters, [((1, 1), (1, 1))],
padding="SAME",
strides=(2, 2),
first_relu=res_relu,
force2d=True,
name="res_conv0")
x = tf.to_float(inputs) / 255.
x.set_shape([None, None, None, 1])
for i in range(model_hparams.audio_compression):
x = xnet_resblock(x, 2**(i + 1), True, "compress_block_%d" % i)
return xnet_resblock(x,
model_hparams.hidden_size,
False,
"compress_block_final")
|
[
"def",
"audio_bottom",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"vocab_size",
"# unused arg",
"inputs",
"=",
"x",
"with",
"tf",
".",
"variable_scope",
"(",
"\"audio_modality\"",
")",
":",
"# TODO(aidangomez): Will need to sort out a better audio pipeline",
"def",
"xnet_resblock",
"(",
"x",
",",
"filters",
",",
"res_relu",
",",
"name",
")",
":",
"\"\"\"Xception block.\"\"\"",
"with",
"tf",
".",
"variable_scope",
"(",
"name",
")",
":",
"# Typically audio samples are >100k samples in length and have a width",
"# of 2 or 4. Mono audio has a single channel while stereo has 2.",
"y",
"=",
"common_layers",
".",
"separable_conv_block",
"(",
"x",
",",
"filters",
",",
"[",
"(",
"(",
"1",
",",
"1",
")",
",",
"(",
"3",
",",
"3",
")",
")",
",",
"(",
"(",
"1",
",",
"1",
")",
",",
"(",
"3",
",",
"3",
")",
")",
"]",
",",
"first_relu",
"=",
"True",
",",
"padding",
"=",
"\"SAME\"",
",",
"force2d",
"=",
"True",
",",
"name",
"=",
"\"sep_conv_block\"",
")",
"y",
"=",
"common_layers",
".",
"pool",
"(",
"y",
",",
"(",
"3",
",",
"3",
")",
",",
"\"MAX\"",
",",
"\"SAME\"",
",",
"strides",
"=",
"(",
"2",
",",
"2",
")",
")",
"return",
"y",
"+",
"common_layers",
".",
"conv_block",
"(",
"x",
",",
"filters",
",",
"[",
"(",
"(",
"1",
",",
"1",
")",
",",
"(",
"1",
",",
"1",
")",
")",
"]",
",",
"padding",
"=",
"\"SAME\"",
",",
"strides",
"=",
"(",
"2",
",",
"2",
")",
",",
"first_relu",
"=",
"res_relu",
",",
"force2d",
"=",
"True",
",",
"name",
"=",
"\"res_conv0\"",
")",
"x",
"=",
"tf",
".",
"to_float",
"(",
"inputs",
")",
"/",
"255.",
"x",
".",
"set_shape",
"(",
"[",
"None",
",",
"None",
",",
"None",
",",
"1",
"]",
")",
"for",
"i",
"in",
"range",
"(",
"model_hparams",
".",
"audio_compression",
")",
":",
"x",
"=",
"xnet_resblock",
"(",
"x",
",",
"2",
"**",
"(",
"i",
"+",
"1",
")",
",",
"True",
",",
"\"compress_block_%d\"",
"%",
"i",
")",
"return",
"xnet_resblock",
"(",
"x",
",",
"model_hparams",
".",
"hidden_size",
",",
"False",
",",
"\"compress_block_final\"",
")"
] |
Transform input from data space to model space.
Args:
x: A Tensor with shape [batch, ...]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
body_input: A Tensor with shape [batch, ?, ?,
model_hparams.hidden_size].
|
[
"Transform",
"input",
"from",
"data",
"space",
"to",
"model",
"space",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L128-L173
|
train
|
Bottom transformation for audio.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(1296 - 1248) + '\157' + chr(49) + chr(2793 - 2738), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\x37' + '\065', 0o10), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(0b1101111) + chr(0b110001) + chr(0b110110) + chr(998 - 950), 0o10), ehT0Px3KOsy9(chr(61 - 13) + chr(7815 - 7704) + chr(0b110001) + '\x33' + '\x33', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(7694 - 7583) + '\064' + chr(55), 0o10), ehT0Px3KOsy9(chr(48) + chr(2317 - 2206) + '\063' + chr(52) + chr(1608 - 1558), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1000 + 0o147) + chr(242 - 193) + chr(55), 8), ehT0Px3KOsy9(chr(2164 - 2116) + '\157' + chr(2769 - 2715) + chr(1767 - 1712), 0o10), ehT0Px3KOsy9(chr(0b1000 + 0o50) + chr(0b1101111) + chr(0b101010 + 0o7) + chr(2304 - 2254) + chr(292 - 242), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110011) + chr(0b110110) + chr(53), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(499 - 451) + chr(51), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(49) + chr(1661 - 1606) + chr(2044 - 1992), 0o10), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(0b1101111) + chr(1881 - 1831) + chr(0b110111) + chr(0b110101), 31917 - 31909), ehT0Px3KOsy9(chr(2204 - 2156) + '\x6f' + chr(0b10100 + 0o35) + '\x32' + chr(48), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b101000 + 0o107) + chr(51) + '\x37' + '\061', 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b101110 + 0o4) + '\062' + chr(0b110001 + 0o0), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110 + 0o55) + chr(50) + '\x37', 32066 - 32058), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + chr(0b110010) + '\x33', 47740 - 47732), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(607 - 556) + chr(55), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(0b100001 + 0o20) + '\x32' + '\060', 8), ehT0Px3KOsy9(chr(0b11110 + 0o22) + chr(12146 - 12035) + chr(1026 - 977) + chr(569 - 517) + chr(0b110011), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + '\x37' + chr(54), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101010 + 0o5) + chr(55) + chr(2282 - 2231), 47560 - 47552), ehT0Px3KOsy9(chr(0b110000) + chr(12248 - 12137) + '\x31' + chr(0b101 + 0o55) + chr(49), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(5911 - 5800) + '\062' + chr(55) + chr(0b110011), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(11864 - 11753) + '\x33' + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(9833 - 9722) + chr(53) + chr(49), 0b1000), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(111) + chr(1475 - 1426) + chr(0b110011) + '\061', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1010111 + 0o30) + chr(0b110001) + chr(665 - 611) + chr(1468 - 1414), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x31' + chr(0b110011) + '\060', 43303 - 43295), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(2183 - 2133) + chr(0b110100), 3570 - 3562), ehT0Px3KOsy9(chr(0b110000) + chr(0b111010 + 0o65) + chr(2171 - 2120) + chr(0b110101) + chr(1988 - 1940), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(1201 - 1152) + chr(1783 - 1735) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(932 - 884) + chr(111) + chr(0b110001) + chr(0b110000) + chr(49), 41344 - 41336), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2654 - 2599) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(388 - 340) + '\x6f' + chr(0b100010 + 0o20) + chr(0b110110) + chr(2141 - 2092), 0b1000), ehT0Px3KOsy9(chr(786 - 738) + '\x6f' + '\x33' + chr(1186 - 1138) + '\x30', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110010) + chr(1155 - 1100) + chr(0b10100 + 0o42), ord("\x08")), ehT0Px3KOsy9(chr(0b100000 + 0o20) + chr(0b1011110 + 0o21) + chr(2158 - 2108) + '\x35' + chr(52), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(129 - 74) + '\x31', ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(111) + chr(0b11 + 0o62) + chr(0b110000), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'!'), '\x64' + chr(0b1101 + 0o130) + '\143' + '\157' + chr(3857 - 3757) + chr(101))(chr(0b1110101) + chr(0b11011 + 0o131) + chr(0b1110 + 0o130) + chr(0b101101) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def a1v03mh4yIkx(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del CeyMIoSyrpkQ
vXoupepMtCXU = OeWW0F1dBPRQ
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'y\\\x022F\x80cR\r\x86 \x13\xcd5'), chr(0b1100100) + chr(2321 - 2220) + '\x63' + chr(0b1101111) + chr(7765 - 7665) + '\145')(chr(0b1110101) + '\164' + chr(1670 - 1568) + '\x2d' + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'nH\x142H\xbdbX6\x94/\x15\xc9)'), chr(100) + chr(5098 - 4997) + '\x63' + '\x6f' + chr(0b1100100) + chr(0b1100101))(chr(0b110010 + 0o103) + '\164' + chr(0b1100110) + chr(45) + chr(56))):
def k5DKvlNpTmW6(OeWW0F1dBPRQ, MErh319F3bgE, _g3xQX3jJAGu, AIvJRzLdDfgF):
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'y\\\x022F\x80cR\r\x86 \x13\xcd5'), chr(0b1100100) + '\x65' + chr(0b1100011) + chr(111) + '\144' + chr(101))('\165' + chr(0b1110100) + '\x66' + '\x2d' + '\070'))(AIvJRzLdDfgF):
SqiSOtYOqOJH = jSKPaHwSAfVv.separable_conv_block(OeWW0F1dBPRQ, MErh319F3bgE, [((ehT0Px3KOsy9('\060' + chr(3116 - 3005) + chr(0b110001), 6895 - 6887), ehT0Px3KOsy9(chr(800 - 752) + '\157' + chr(0b110001), 8)), (ehT0Px3KOsy9(chr(1705 - 1657) + '\157' + '\x33', ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110011), 8))), ((ehT0Px3KOsy9(chr(0b110000) + chr(12194 - 12083) + '\061', 8), ehT0Px3KOsy9('\x30' + '\157' + chr(595 - 546), 8)), (ehT0Px3KOsy9('\060' + chr(0b1100 + 0o143) + '\x33', 8), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011), 8)))], first_relu=ehT0Px3KOsy9(chr(467 - 419) + chr(0b1000111 + 0o50) + chr(0b11 + 0o56), 8), padding=xafqLlk3kkUe(SXOLrMavuUCe(b'\\|=\x1e'), chr(4785 - 4685) + chr(0b100001 + 0o104) + chr(0b1 + 0o142) + chr(0b1101111) + chr(3314 - 3214) + '\145')(chr(117) + chr(0b1110100) + chr(0b11110 + 0o110) + chr(0b101101) + chr(56)), force2d=ehT0Px3KOsy9('\x30' + '\157' + chr(49), 8), name=xafqLlk3kkUe(SXOLrMavuUCe(b'|X\x00\x04D\x8daA\r\x97/\x13\xde;'), '\144' + chr(1437 - 1336) + '\x63' + '\x6f' + '\144' + chr(0b111 + 0o136))(chr(0b1100 + 0o151) + chr(116) + chr(3473 - 3371) + chr(1643 - 1598) + chr(56)))
SqiSOtYOqOJH = jSKPaHwSAfVv.pool(SqiSOtYOqOJH, (ehT0Px3KOsy9(chr(374 - 326) + chr(0b1101111) + chr(0b11100 + 0o27), 8), ehT0Px3KOsy9('\x30' + chr(7814 - 7703) + '\x33', 8)), xafqLlk3kkUe(SXOLrMavuUCe(b'B|('), '\144' + chr(0b1100101) + chr(99) + chr(0b1101111) + chr(100) + chr(101))(chr(0b1110101) + chr(5844 - 5728) + '\146' + chr(0b101 + 0o50) + chr(0b111000)), xafqLlk3kkUe(SXOLrMavuUCe(b'\\|=\x1e'), chr(100) + '\x65' + chr(0b1100011) + chr(1442 - 1331) + chr(0b110000 + 0o64) + chr(0b1100101))(chr(2866 - 2749) + '\164' + chr(102) + chr(0b110 + 0o47) + '\070'), strides=(ehT0Px3KOsy9('\x30' + chr(111) + chr(50), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(50), 8)))
return SqiSOtYOqOJH + xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'lR\x1e-x\x80cX1\x9e'), '\x64' + chr(0b10011 + 0o122) + chr(4647 - 4548) + chr(111) + chr(100) + '\145')('\x75' + '\x74' + '\x66' + chr(0b101101) + chr(0b1011 + 0o55)))(OeWW0F1dBPRQ, MErh319F3bgE, [((ehT0Px3KOsy9(chr(48) + chr(7204 - 7093) + '\061', 8), ehT0Px3KOsy9('\x30' + chr(111) + chr(49), 8)), (ehT0Px3KOsy9(chr(48) + '\157' + chr(991 - 942), 8), ehT0Px3KOsy9(chr(48) + chr(9691 - 9580) + chr(0b11 + 0o56), 8)))], padding=xafqLlk3kkUe(SXOLrMavuUCe(b'\\|=\x1e'), chr(100) + chr(101) + '\143' + chr(111) + '\144' + chr(1114 - 1013))(chr(0b1011010 + 0o33) + chr(0b101010 + 0o112) + chr(1403 - 1301) + chr(45) + chr(0b111000)), strides=(ehT0Px3KOsy9('\x30' + chr(6656 - 6545) + chr(546 - 496), 8), ehT0Px3KOsy9('\x30' + chr(111) + '\x32', 8)), first_relu=_g3xQX3jJAGu, force2d=ehT0Px3KOsy9(chr(168 - 120) + chr(9572 - 9461) + '\x31', 8), name=xafqLlk3kkUe(SXOLrMavuUCe(b'}X\x03\x04D\x8daAb'), '\144' + chr(0b1011011 + 0o12) + chr(0b10111 + 0o114) + chr(111) + chr(0b1000110 + 0o36) + '\x65')(chr(0b1110101) + chr(6058 - 5942) + chr(3311 - 3209) + chr(0b101101) + chr(687 - 631)))
OeWW0F1dBPRQ = IDJ2eXGCBCDu.to_float(vXoupepMtCXU) / 255.0
xafqLlk3kkUe(OeWW0F1dBPRQ, xafqLlk3kkUe(SXOLrMavuUCe(b'|X\x04\x04T\x8anG7'), chr(100) + chr(101) + chr(99) + chr(0b1101111) + chr(100) + chr(0b1100101 + 0o0))('\165' + chr(0b11001 + 0o133) + chr(102) + chr(0b1100 + 0o41) + chr(0b111000)))([None, None, None, ehT0Px3KOsy9(chr(770 - 722) + chr(0b1101111) + '\061', 8)])
for WVxHKyX45z_L in vQr8gNKaIaWE(xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'{|Fof\xdb~\x05\x0b\xb2!\x18'), chr(0b1100100) + chr(1887 - 1786) + '\x63' + '\157' + chr(7493 - 7393) + '\x65')('\165' + chr(9151 - 9035) + '\146' + chr(684 - 639) + chr(0b111000)))):
OeWW0F1dBPRQ = k5DKvlNpTmW6(OeWW0F1dBPRQ, ehT0Px3KOsy9(chr(0b1111 + 0o41) + '\x6f' + '\062', 8) ** (WVxHKyX45z_L + ehT0Px3KOsy9(chr(0b100001 + 0o17) + chr(8981 - 8870) + chr(0b10101 + 0o34), 8)), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b0 + 0o61), 8), xafqLlk3kkUe(SXOLrMavuUCe(b'lR\x1d+U\x87|D\r\x97/\x13\xde;\xa6\xb4\xb3'), chr(1491 - 1391) + chr(0b0 + 0o145) + '\143' + '\x6f' + chr(100) + chr(0b1100101))(chr(3027 - 2910) + chr(9736 - 9620) + '\146' + chr(968 - 923) + chr(56)) % WVxHKyX45z_L)
return k5DKvlNpTmW6(OeWW0F1dBPRQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'~G\x1f"\x7f\xac<\\6\x9d\x070'), chr(3232 - 3132) + chr(9555 - 9454) + '\x63' + '\157' + chr(0b1100100) + chr(101))(chr(117) + chr(11240 - 11124) + chr(0b1100110) + chr(45) + chr(56))), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x30', 56232 - 56224), xafqLlk3kkUe(SXOLrMavuUCe(b'lR\x1d+U\x87|D\r\x97/\x13\xde;\xa6\xf7\xbe\xfd~<'), chr(4716 - 4616) + chr(0b1100101) + chr(0b111001 + 0o52) + chr(6743 - 6632) + chr(0b1100100) + chr(0b1100101))(chr(0b1110101) + '\x74' + chr(102) + '\055' + '\x38'))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
image_targets_bottom
|
def image_targets_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for target images."""
pixel_embedding_size = 64
inputs = x
with tf.variable_scope("image_modality"):
if not tf.executing_eagerly():
tf.summary.image(
"targets_bottom",
common_layers.tpu_safe_image_summary(inputs),
max_outputs=1)
inputs_shape = common_layers.shape_list(inputs)
if len(inputs_shape) != 4:
raise ValueError("Assuming images given as int tensors in the format "
"[batch, height, width, channels] (256 values).")
# We embed each of 256=vocab_size possible pixel values.
embedding_var = tf.get_variable(
"pixel_embedding",
[vocab_size, pixel_embedding_size])
hot_inputs = tf.one_hot(tf.to_int32(inputs), vocab_size)
hot_inputs = tf.reshape(hot_inputs, [-1, vocab_size])
embedded = tf.matmul(hot_inputs, embedding_var)
# Let's now merge all channels that were embedded into a single vector.
merged_size = pixel_embedding_size * inputs_shape[3]
embedded = tf.reshape(embedded, inputs_shape[:3] + [merged_size])
merged = tf.layers.dense(
embedded,
model_hparams.hidden_size,
name="merge_pixel_embedded_channels")
return merged
|
python
|
def image_targets_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for target images."""
pixel_embedding_size = 64
inputs = x
with tf.variable_scope("image_modality"):
if not tf.executing_eagerly():
tf.summary.image(
"targets_bottom",
common_layers.tpu_safe_image_summary(inputs),
max_outputs=1)
inputs_shape = common_layers.shape_list(inputs)
if len(inputs_shape) != 4:
raise ValueError("Assuming images given as int tensors in the format "
"[batch, height, width, channels] (256 values).")
# We embed each of 256=vocab_size possible pixel values.
embedding_var = tf.get_variable(
"pixel_embedding",
[vocab_size, pixel_embedding_size])
hot_inputs = tf.one_hot(tf.to_int32(inputs), vocab_size)
hot_inputs = tf.reshape(hot_inputs, [-1, vocab_size])
embedded = tf.matmul(hot_inputs, embedding_var)
# Let's now merge all channels that were embedded into a single vector.
merged_size = pixel_embedding_size * inputs_shape[3]
embedded = tf.reshape(embedded, inputs_shape[:3] + [merged_size])
merged = tf.layers.dense(
embedded,
model_hparams.hidden_size,
name="merge_pixel_embedded_channels")
return merged
|
[
"def",
"image_targets_bottom",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"pixel_embedding_size",
"=",
"64",
"inputs",
"=",
"x",
"with",
"tf",
".",
"variable_scope",
"(",
"\"image_modality\"",
")",
":",
"if",
"not",
"tf",
".",
"executing_eagerly",
"(",
")",
":",
"tf",
".",
"summary",
".",
"image",
"(",
"\"targets_bottom\"",
",",
"common_layers",
".",
"tpu_safe_image_summary",
"(",
"inputs",
")",
",",
"max_outputs",
"=",
"1",
")",
"inputs_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"inputs",
")",
"if",
"len",
"(",
"inputs_shape",
")",
"!=",
"4",
":",
"raise",
"ValueError",
"(",
"\"Assuming images given as int tensors in the format \"",
"\"[batch, height, width, channels] (256 values).\"",
")",
"# We embed each of 256=vocab_size possible pixel values.",
"embedding_var",
"=",
"tf",
".",
"get_variable",
"(",
"\"pixel_embedding\"",
",",
"[",
"vocab_size",
",",
"pixel_embedding_size",
"]",
")",
"hot_inputs",
"=",
"tf",
".",
"one_hot",
"(",
"tf",
".",
"to_int32",
"(",
"inputs",
")",
",",
"vocab_size",
")",
"hot_inputs",
"=",
"tf",
".",
"reshape",
"(",
"hot_inputs",
",",
"[",
"-",
"1",
",",
"vocab_size",
"]",
")",
"embedded",
"=",
"tf",
".",
"matmul",
"(",
"hot_inputs",
",",
"embedding_var",
")",
"# Let's now merge all channels that were embedded into a single vector.",
"merged_size",
"=",
"pixel_embedding_size",
"*",
"inputs_shape",
"[",
"3",
"]",
"embedded",
"=",
"tf",
".",
"reshape",
"(",
"embedded",
",",
"inputs_shape",
"[",
":",
"3",
"]",
"+",
"[",
"merged_size",
"]",
")",
"merged",
"=",
"tf",
".",
"layers",
".",
"dense",
"(",
"embedded",
",",
"model_hparams",
".",
"hidden_size",
",",
"name",
"=",
"\"merge_pixel_embedded_channels\"",
")",
"return",
"merged"
] |
Bottom transformation for target images.
|
[
"Bottom",
"transformation",
"for",
"target",
"images",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L260-L288
|
train
|
Bottom transformation for target images.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b11101 + 0o25) + '\x33' + chr(0b110110), 0b1000), ehT0Px3KOsy9('\060' + chr(0b101000 + 0o107) + '\x33' + chr(2101 - 2049) + '\063', ord("\x08")), ehT0Px3KOsy9(chr(295 - 247) + '\157' + '\x31' + '\x35' + chr(52), ord("\x08")), ehT0Px3KOsy9('\060' + chr(11109 - 10998) + chr(0b110011) + chr(0b110100) + chr(48), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b10011 + 0o134) + chr(0b110101) + chr(1571 - 1523), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + '\x35' + chr(1590 - 1542), 0o10), ehT0Px3KOsy9(chr(1720 - 1672) + chr(9830 - 9719) + chr(0b110001) + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b111011 + 0o64) + chr(51) + chr(0b110000) + chr(2022 - 1968), 48462 - 48454), ehT0Px3KOsy9(chr(0b110000) + chr(8744 - 8633) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(1608 - 1560) + chr(189 - 78) + '\063' + '\x30' + '\065', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\061' + chr(52) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(0b1101111) + chr(0b10101 + 0o36) + chr(0b10111 + 0o33) + chr(0b101111 + 0o5), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + chr(0b110010) + chr(0b110110), 58019 - 58011), ehT0Px3KOsy9(chr(232 - 184) + '\157' + chr(0b110001) + chr(0b101 + 0o62) + '\067', 2250 - 2242), ehT0Px3KOsy9('\060' + chr(111) + '\064' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(1269 - 1221) + chr(0b111111 + 0o60) + chr(0b110001) + chr(50) + '\x36', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(9053 - 8942) + chr(0b10111 + 0o32) + chr(48) + chr(0b100 + 0o60), 0b1000), ehT0Px3KOsy9('\x30' + chr(7941 - 7830) + '\063' + '\x32' + '\x32', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b1101 + 0o46) + chr(0b11101 + 0o25) + chr(0b110111), 4829 - 4821), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b101111 + 0o2) + chr(0b101010 + 0o15) + chr(0b110010), 446 - 438), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110110) + '\063', 0o10), ehT0Px3KOsy9(chr(623 - 575) + chr(0b1101111) + chr(0b110000 + 0o1) + chr(0b1100 + 0o53) + chr(50), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111 + 0o0) + chr(0b110010) + chr(1607 - 1559) + chr(54), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110010) + chr(1837 - 1782) + chr(0b110010 + 0o1), 19878 - 19870), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x32' + chr(0b100100 + 0o15) + chr(0b0 + 0o67), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(1156 - 1106) + chr(581 - 526), 0b1000), ehT0Px3KOsy9(chr(1690 - 1642) + chr(111) + chr(0b110110), 29309 - 29301), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110011) + '\060' + '\062', 48642 - 48634), ehT0Px3KOsy9('\x30' + '\157' + '\x34' + chr(50), 8), ehT0Px3KOsy9('\x30' + chr(0b1100011 + 0o14) + chr(0b110000 + 0o2) + chr(0b101111 + 0o2) + chr(0b110001), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101 + 0o142) + '\x32' + '\x32', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1100 + 0o143) + '\061' + chr(0b100001 + 0o23) + chr(0b100001 + 0o22), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b100110 + 0o111) + '\x33' + chr(0b11100 + 0o31) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100001 + 0o16) + '\063' + '\066' + '\061', 0o10), ehT0Px3KOsy9(chr(0b1110 + 0o42) + '\157' + chr(906 - 857) + chr(0b110000), 8), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + '\064' + chr(0b101110 + 0o4), 0o10), ehT0Px3KOsy9(chr(2082 - 2034) + chr(9956 - 9845) + '\x33' + '\060' + '\065', 8), ehT0Px3KOsy9(chr(750 - 702) + chr(111) + chr(0b110011) + chr(0b1011 + 0o50) + chr(0b110010), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(50) + chr(1116 - 1067), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x32' + chr(0b110010), 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(3697 - 3586) + chr(924 - 871) + '\060', 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xb0'), chr(0b1100100) + chr(101) + chr(99) + '\x6f' + chr(0b10100 + 0o120) + chr(101))(chr(4862 - 4745) + chr(3207 - 3091) + '\146' + chr(45) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def h7VqVNsx2j8T(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
jk6x7fS5a3ck = ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(48) + chr(48), ord("\x08"))
vXoupepMtCXU = OeWW0F1dBPRQ
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe8\x82\x1e\x89I\nd\xa3[\xb6\x03\xbb\xa6\x1a'), chr(0b1101 + 0o127) + chr(8310 - 8209) + chr(99) + chr(0b100010 + 0o115) + chr(0b1100100) + chr(0b1100101))(chr(0b1110101) + chr(0b1100010 + 0o22) + chr(0b1010110 + 0o20) + chr(0b101101) + chr(103 - 47)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf7\x8e\r\x87M7e\xa9`\xa4\x0c\xbd\xa2\x06'), chr(0b1100100) + chr(5948 - 5847) + chr(6140 - 6041) + '\x6f' + '\x64' + chr(0b11010 + 0o113))(chr(0b10011 + 0o142) + chr(0b1110000 + 0o4) + chr(0b1100110) + chr(0b101000 + 0o5) + '\x38')):
if not xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfb\x9b\t\x83]\x1ca\xa8c\x9a\x05\xb5\xb1\x1aP:='), chr(0b101 + 0o137) + '\145' + chr(0b11000 + 0o113) + chr(111) + chr(960 - 860) + chr(101))(chr(0b1111 + 0o146) + chr(116) + '\x66' + chr(0b101101) + '\x38'))():
xafqLlk3kkUe(IDJ2eXGCBCDu.summary, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd7\x87\x01\xa1`?n\x85u\xb7\x0e\xa4'), '\144' + '\145' + '\143' + chr(0b1101111) + chr(0b1001000 + 0o34) + '\x65')(chr(5372 - 5255) + '\x74' + chr(102) + chr(0b101101) + chr(356 - 300)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xea\x82\x1e\x87M\x1c{\x99f\xaa\x14\xa0\xb9\x12'), chr(7271 - 7171) + chr(101) + chr(0b1100011) + chr(4296 - 4185) + chr(0b1100100) + chr(0b1100101))(chr(0b1110101) + chr(116) + chr(0b100100 + 0o102) + chr(485 - 440) + chr(0b111000)), xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'\xea\x93\x19\xbf[\tn\xa3[\xac\r\xb5\xb1\x1a}%1\xbf\xba\x1bA\x03'), chr(100) + chr(0b1100101) + '\143' + chr(1311 - 1200) + '\x64' + chr(0b1010010 + 0o23))(chr(1963 - 1846) + '\164' + chr(102) + chr(45) + chr(912 - 856)))(vXoupepMtCXU), max_outputs=ehT0Px3KOsy9(chr(48) + chr(0b110100 + 0o73) + chr(49), ord("\x08")))
VgP_McURhCb5 = jSKPaHwSAfVv.shape_list(vXoupepMtCXU)
if c2A0yzQpDQB3(VgP_McURhCb5) != ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(0b1101111) + chr(0b110100), 8):
raise q1QCh3W88sgk(xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x90\x1f\x95E\x01f\xa1$\xac\r\xb5\xb1\x1aQv#\xbb\xa1\x1f]Z\xe7\xc8\xaf~h\xb6\x02\xfdw\r\x9b\x9ez\x91\xe5\xb3\x7f2\xea\x8b\t\xc0N\x07z\xabe\xb1@\x8f\xb4\x1eV5,\xfe\xf7\x12V\x13\xe1\xd3\xfb;&\xb5K\xedf\x0b\xc4\xd1k\x8a\xa4\xb4\x7fw\xf2\x901\xc0\x00Z=\xf0$\xb3\x01\xb8\xa3\x1aQ\x7fj'), chr(100) + chr(101) + '\x63' + chr(111) + '\144' + chr(0b1010100 + 0o21))(chr(0b1110101) + chr(116) + chr(907 - 805) + '\x2d' + '\070'))
g9mMUZCHfbgl = IDJ2eXGCBCDu.get_variable(xafqLlk3kkUe(SXOLrMavuUCe(b'\xee\x8a\x14\x85D7m\xabf\xa0\x04\xb0\xbf\x11E'), chr(3564 - 3464) + '\x65' + '\143' + '\x6f' + chr(2551 - 2451) + chr(101))('\x75' + chr(11861 - 11745) + chr(0b1100110) + chr(0b101101 + 0o0) + chr(1780 - 1724)), [CeyMIoSyrpkQ, jk6x7fS5a3ck])
ork2YCRdxUVj = IDJ2eXGCBCDu.Hq3fv4Yp0EhD(IDJ2eXGCBCDu.to_int32(vXoupepMtCXU), CeyMIoSyrpkQ)
ork2YCRdxUVj = IDJ2eXGCBCDu.reshape(ork2YCRdxUVj, [-ehT0Px3KOsy9('\x30' + chr(111) + chr(1786 - 1737), 8), CeyMIoSyrpkQ])
FaFRbyrGgzRu = IDJ2eXGCBCDu.matmul(ork2YCRdxUVj, g9mMUZCHfbgl)
LfXds5rEnpTE = jk6x7fS5a3ck * VgP_McURhCb5[ehT0Px3KOsy9(chr(1377 - 1329) + '\x6f' + chr(1536 - 1485), 9344 - 9336)]
FaFRbyrGgzRu = IDJ2eXGCBCDu.reshape(FaFRbyrGgzRu, VgP_McURhCb5[:ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b1100 + 0o47), 8)] + [LfXds5rEnpTE])
dEDdb0D1WnBU = IDJ2eXGCBCDu.layers.dense(FaFRbyrGgzRu, tq24Tk6UZ6u1.qzoyXN3kdhDL, name=xafqLlk3kkUe(SXOLrMavuUCe(b'\xf3\x86\x1e\x87M7x\xaf|\xa0\x0c\x8b\xb3\x12@3 \xb6\xb2\x1el\x19\xee\xda\xe1yc\xaeQ'), chr(0b110000 + 0o64) + '\x65' + chr(99) + '\x6f' + '\144' + chr(0b1100101))('\165' + '\164' + chr(102) + '\055' + '\070'))
return dEDdb0D1WnBU
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
_image_channel_compress_bottom
|
def _image_channel_compress_bottom(inputs, model_hparams, name="bottom"):
"""Compresses channel-wise input pixels into whole pixel representions.
Perform conversion of RGB pixel values to a real number in the range -1 to
1. This combines pixel channels to form a representation of shape
[img_len, img_len].
Args:
inputs: Tensor representing RGB pixel intensities as integers, of shape
[batch, img_len, img_len, channels].
model_hparams: HParams, model hyperparmeters.
name: string, scope.
Returns:
body_input: Tensor of shape
[batch, img_len, img_len, model_hparams.hidden_size].
"""
num_channels = 3
with tf.variable_scope(name):
inputs = tf.to_float(inputs)
hp = model_hparams
if hp.mode != tf.estimator.ModeKeys.PREDICT:
tf.summary.image(
"inputs",
common_layers.tpu_safe_image_summary(inputs),
max_outputs=2)
inputs = common_layers.convert_rgb_to_symmetric_real(inputs)
# Reshape inputs to apply convolutions across [img_len, img_len*channels].
inputs_shape = common_layers.shape_list(inputs)
inputs = tf.reshape(
inputs, [-1, inputs_shape[1], inputs_shape[2] * inputs_shape[3], 1])
# Compress RGB intensities for each pixel using a convolution.
outputs = tf.layers.conv2d(
inputs,
model_hparams.hidden_size,
kernel_size=(1, num_channels),
padding="VALID",
strides=(1, num_channels),
activation=tf.nn.relu,
name="conv_input")
return outputs
|
python
|
def _image_channel_compress_bottom(inputs, model_hparams, name="bottom"):
"""Compresses channel-wise input pixels into whole pixel representions.
Perform conversion of RGB pixel values to a real number in the range -1 to
1. This combines pixel channels to form a representation of shape
[img_len, img_len].
Args:
inputs: Tensor representing RGB pixel intensities as integers, of shape
[batch, img_len, img_len, channels].
model_hparams: HParams, model hyperparmeters.
name: string, scope.
Returns:
body_input: Tensor of shape
[batch, img_len, img_len, model_hparams.hidden_size].
"""
num_channels = 3
with tf.variable_scope(name):
inputs = tf.to_float(inputs)
hp = model_hparams
if hp.mode != tf.estimator.ModeKeys.PREDICT:
tf.summary.image(
"inputs",
common_layers.tpu_safe_image_summary(inputs),
max_outputs=2)
inputs = common_layers.convert_rgb_to_symmetric_real(inputs)
# Reshape inputs to apply convolutions across [img_len, img_len*channels].
inputs_shape = common_layers.shape_list(inputs)
inputs = tf.reshape(
inputs, [-1, inputs_shape[1], inputs_shape[2] * inputs_shape[3], 1])
# Compress RGB intensities for each pixel using a convolution.
outputs = tf.layers.conv2d(
inputs,
model_hparams.hidden_size,
kernel_size=(1, num_channels),
padding="VALID",
strides=(1, num_channels),
activation=tf.nn.relu,
name="conv_input")
return outputs
|
[
"def",
"_image_channel_compress_bottom",
"(",
"inputs",
",",
"model_hparams",
",",
"name",
"=",
"\"bottom\"",
")",
":",
"num_channels",
"=",
"3",
"with",
"tf",
".",
"variable_scope",
"(",
"name",
")",
":",
"inputs",
"=",
"tf",
".",
"to_float",
"(",
"inputs",
")",
"hp",
"=",
"model_hparams",
"if",
"hp",
".",
"mode",
"!=",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"PREDICT",
":",
"tf",
".",
"summary",
".",
"image",
"(",
"\"inputs\"",
",",
"common_layers",
".",
"tpu_safe_image_summary",
"(",
"inputs",
")",
",",
"max_outputs",
"=",
"2",
")",
"inputs",
"=",
"common_layers",
".",
"convert_rgb_to_symmetric_real",
"(",
"inputs",
")",
"# Reshape inputs to apply convolutions across [img_len, img_len*channels].",
"inputs_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"inputs",
")",
"inputs",
"=",
"tf",
".",
"reshape",
"(",
"inputs",
",",
"[",
"-",
"1",
",",
"inputs_shape",
"[",
"1",
"]",
",",
"inputs_shape",
"[",
"2",
"]",
"*",
"inputs_shape",
"[",
"3",
"]",
",",
"1",
"]",
")",
"# Compress RGB intensities for each pixel using a convolution.",
"outputs",
"=",
"tf",
".",
"layers",
".",
"conv2d",
"(",
"inputs",
",",
"model_hparams",
".",
"hidden_size",
",",
"kernel_size",
"=",
"(",
"1",
",",
"num_channels",
")",
",",
"padding",
"=",
"\"VALID\"",
",",
"strides",
"=",
"(",
"1",
",",
"num_channels",
")",
",",
"activation",
"=",
"tf",
".",
"nn",
".",
"relu",
",",
"name",
"=",
"\"conv_input\"",
")",
"return",
"outputs"
] |
Compresses channel-wise input pixels into whole pixel representions.
Perform conversion of RGB pixel values to a real number in the range -1 to
1. This combines pixel channels to form a representation of shape
[img_len, img_len].
Args:
inputs: Tensor representing RGB pixel intensities as integers, of shape
[batch, img_len, img_len, channels].
model_hparams: HParams, model hyperparmeters.
name: string, scope.
Returns:
body_input: Tensor of shape
[batch, img_len, img_len, model_hparams.hidden_size].
|
[
"Compresses",
"channel",
"-",
"wise",
"input",
"pixels",
"into",
"whole",
"pixel",
"representions",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L291-L333
|
train
|
Bottom image - channel compression.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(1167 - 1119) + chr(0b110011 + 0o74) + '\061' + chr(48) + chr(0b11000 + 0o30), 0o10), ehT0Px3KOsy9('\x30' + chr(678 - 567) + chr(964 - 915) + chr(0b110111) + '\063', 17873 - 17865), ehT0Px3KOsy9(chr(1371 - 1323) + '\157' + chr(0b101 + 0o56) + chr(645 - 597) + chr(48), 0o10), ehT0Px3KOsy9(chr(1263 - 1215) + '\x6f' + chr(49) + chr(0b110100) + chr(925 - 870), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110010) + '\062', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(1195 - 1146) + chr(51), 18317 - 18309), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b0 + 0o62) + chr(50) + '\065', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b100110 + 0o111) + chr(2165 - 2115) + '\065' + chr(55), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(3942 - 3831) + chr(51) + '\x36' + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(49) + chr(190 - 138) + '\x31', 51717 - 51709), ehT0Px3KOsy9('\060' + '\157' + '\x32' + '\061' + chr(0b100001 + 0o26), ord("\x08")), ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(0b1101111) + chr(0b110010) + chr(0b101010 + 0o15) + '\064', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(2426 - 2376) + chr(0b100111 + 0o15), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b111010 + 0o65) + '\063' + chr(51) + chr(54), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + '\064' + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1001001 + 0o46) + chr(2370 - 2321) + '\062' + '\067', 37207 - 37199), ehT0Px3KOsy9(chr(0b101101 + 0o3) + '\x6f' + chr(0b1111 + 0o44) + chr(0b110101) + chr(0b110000), 58745 - 58737), ehT0Px3KOsy9('\x30' + chr(0b10011 + 0o134) + '\x33' + '\062' + chr(0b11111 + 0o26), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\067' + chr(53), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10 + 0o65) + '\x34', 0o10), ehT0Px3KOsy9(chr(0b101000 + 0o10) + '\157' + '\066' + chr(2142 - 2093), 64645 - 64637), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b1001 + 0o51) + chr(2666 - 2613) + chr(0b110001), 0o10), ehT0Px3KOsy9(chr(360 - 312) + chr(0b111010 + 0o65) + '\063' + chr(55) + '\065', 0o10), ehT0Px3KOsy9(chr(48) + '\157' + '\063' + chr(0b11101 + 0o23) + '\x32', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + '\x31' + '\065', 0o10), ehT0Px3KOsy9('\060' + chr(10356 - 10245) + chr(0b11010 + 0o27) + chr(0b110110) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(611 - 563) + chr(4698 - 4587) + chr(0b110010) + chr(51) + chr(0b110100), 32223 - 32215), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + '\065' + chr(48), 0o10), ehT0Px3KOsy9(chr(1893 - 1845) + chr(0b101011 + 0o104) + chr(49) + chr(0b10111 + 0o35) + chr(0b11111 + 0o24), 9501 - 9493), ehT0Px3KOsy9(chr(462 - 414) + chr(111) + chr(55) + '\x33', 0o10), ehT0Px3KOsy9('\060' + chr(0b1001 + 0o146) + chr(2365 - 2312) + '\061', 0b1000), ehT0Px3KOsy9(chr(2213 - 2165) + chr(8309 - 8198) + '\x32' + chr(0b1001 + 0o56) + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(0b100101 + 0o112) + chr(770 - 718) + '\x36', 20673 - 20665), ehT0Px3KOsy9(chr(0b110000) + chr(822 - 711) + chr(54) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(48) + chr(5202 - 5091) + '\x31' + chr(0b110000), 58907 - 58899), ehT0Px3KOsy9('\x30' + chr(9783 - 9672) + chr(49) + chr(55) + chr(800 - 745), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(669 - 619) + chr(55) + '\060', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(50) + chr(0b110010) + chr(54), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + '\x31' + chr(0b1100 + 0o47) + chr(0b1100 + 0o50), 15049 - 15041), ehT0Px3KOsy9(chr(0b110000) + chr(1362 - 1251) + chr(53) + chr(0b110010), 42821 - 42813)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\065' + '\x30', 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'n'), chr(3575 - 3475) + chr(0b1100101) + chr(99) + chr(111) + '\144' + chr(8356 - 8255))(chr(0b1110101) + chr(7692 - 7576) + '\146' + chr(1454 - 1409) + chr(3047 - 2991)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def hcq0d9yI9rqh(vXoupepMtCXU, tq24Tk6UZ6u1, AIvJRzLdDfgF=xafqLlk3kkUe(SXOLrMavuUCe(b'"\x14\xb4\x13}f'), chr(5099 - 4999) + chr(7538 - 7437) + '\143' + chr(0b11111 + 0o120) + chr(0b1100100) + chr(0b1001010 + 0o33))(chr(117) + chr(116) + chr(0b1100110) + chr(0b11111 + 0o16) + chr(0b1111 + 0o51))):
X1ZpHSxyKbHn = ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011), 58064 - 58056)
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'6\x1a\xb2\x0esi\xf9\xca\xaa\x12\xec\xd3d\xbb'), chr(100) + '\145' + chr(0b1100011) + chr(4333 - 4222) + '\144' + '\145')(chr(117) + chr(0b1110100) + '\x66' + '\055' + chr(0b111000)))(AIvJRzLdDfgF):
vXoupepMtCXU = IDJ2eXGCBCDu.to_float(vXoupepMtCXU)
ny6shRSJO9Wm = tq24Tk6UZ6u1
if xafqLlk3kkUe(ny6shRSJO9Wm, xafqLlk3kkUe(SXOLrMavuUCe(b'-\x14\xa4\x02'), chr(0b11110 + 0o106) + chr(3109 - 3008) + chr(0b1100011) + chr(111) + chr(0b1100100) + chr(0b100001 + 0o104))(chr(0b100010 + 0o123) + chr(116) + '\146' + '\x2d' + '\070')) != xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'\x10)\x85#[H\xc1'), '\144' + chr(0b1100101) + chr(0b1100011) + '\157' + chr(100) + chr(4871 - 4770))(chr(0b1110101) + chr(5280 - 5164) + chr(6798 - 6696) + '\055' + '\070')):
xafqLlk3kkUe(IDJ2eXGCBCDu.summary, xafqLlk3kkUe(SXOLrMavuUCe(b'\t\x1f\xad&Z\\\xf3\xec\x84\x13\xe1\xcc'), chr(0b1100100) + '\145' + chr(6598 - 6499) + chr(111) + chr(0b1100100) + chr(101))(chr(117) + chr(0b1000101 + 0o57) + chr(6142 - 6040) + chr(1253 - 1208) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b')\x15\xb0\x12fx'), '\x64' + '\x65' + '\143' + chr(5773 - 5662) + '\x64' + chr(0b1100101))(chr(0b1110101) + chr(116) + chr(0b1100110) + chr(45) + chr(0b111000)), xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'4\x0b\xb58aj\xf3\xca\xaa\x08\xe2\xdds\xbb|$\x9d\xb8P]C\xd0'), '\x64' + '\x65' + '\x63' + '\157' + chr(5776 - 5676) + chr(101))(chr(2730 - 2613) + '\x74' + chr(0b1100110) + chr(0b100000 + 0o15) + chr(56)))(vXoupepMtCXU), max_outputs=ehT0Px3KOsy9('\060' + chr(3871 - 3760) + '\x32', 0b1000))
vXoupepMtCXU = jSKPaHwSAfVv.convert_rgb_to_symmetric_real(vXoupepMtCXU)
VgP_McURhCb5 = jSKPaHwSAfVv.shape_list(vXoupepMtCXU)
vXoupepMtCXU = IDJ2eXGCBCDu.reshape(vXoupepMtCXU, [-ehT0Px3KOsy9(chr(0b110000) + chr(4555 - 4444) + '\061', ord("\x08")), VgP_McURhCb5[ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(111) + '\x31', 8)], VgP_McURhCb5[ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110010), 8)] * VgP_McURhCb5[ehT0Px3KOsy9('\060' + chr(0b111110 + 0o61) + chr(0b110011), 8)], ehT0Px3KOsy9(chr(2059 - 2011) + '\157' + chr(49), 8)])
Dx_DllZ8uCko = IDJ2eXGCBCDu.layers.conv2d(vXoupepMtCXU, tq24Tk6UZ6u1.qzoyXN3kdhDL, kernel_size=(ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b11010 + 0o27), 8), X1ZpHSxyKbHn), padding=xafqLlk3kkUe(SXOLrMavuUCe(b'\x16:\x8c.V'), chr(100) + chr(0b10111 + 0o116) + chr(5482 - 5383) + chr(0b1101111) + '\144' + chr(101))('\x75' + '\x74' + '\x66' + chr(0b100 + 0o51) + '\x38'), strides=(ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b10011 + 0o36), 8), X1ZpHSxyKbHn), activation=IDJ2eXGCBCDu.nn.relu, name=xafqLlk3kkUe(SXOLrMavuUCe(b'#\x14\xae\x11Mb\xfb\xdf\x80\x15'), chr(0b110110 + 0o56) + chr(101) + chr(6341 - 6242) + chr(10848 - 10737) + chr(0b1000011 + 0o41) + chr(2422 - 2321))(chr(6045 - 5928) + '\164' + chr(6361 - 6259) + chr(0b101101) + chr(0b111000)))
return Dx_DllZ8uCko
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
image_channel_embeddings_bottom
|
def image_channel_embeddings_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for image targets."""
del vocab_size # unused arg
inputs = tf.to_int32(x)
io_depth = model_hparams.num_channels
tshape = common_layers.shape_list(inputs)
hidden_size = model_hparams.hidden_size
target_embeddings = cia.get_channel_embeddings(
io_depth, inputs, hidden_size, "input_bottom")
return tf.reshape(target_embeddings,
[tshape[0], tshape[1], tshape[2] * io_depth, hidden_size])
|
python
|
def image_channel_embeddings_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for image targets."""
del vocab_size # unused arg
inputs = tf.to_int32(x)
io_depth = model_hparams.num_channels
tshape = common_layers.shape_list(inputs)
hidden_size = model_hparams.hidden_size
target_embeddings = cia.get_channel_embeddings(
io_depth, inputs, hidden_size, "input_bottom")
return tf.reshape(target_embeddings,
[tshape[0], tshape[1], tshape[2] * io_depth, hidden_size])
|
[
"def",
"image_channel_embeddings_bottom",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"vocab_size",
"# unused arg",
"inputs",
"=",
"tf",
".",
"to_int32",
"(",
"x",
")",
"io_depth",
"=",
"model_hparams",
".",
"num_channels",
"tshape",
"=",
"common_layers",
".",
"shape_list",
"(",
"inputs",
")",
"hidden_size",
"=",
"model_hparams",
".",
"hidden_size",
"target_embeddings",
"=",
"cia",
".",
"get_channel_embeddings",
"(",
"io_depth",
",",
"inputs",
",",
"hidden_size",
",",
"\"input_bottom\"",
")",
"return",
"tf",
".",
"reshape",
"(",
"target_embeddings",
",",
"[",
"tshape",
"[",
"0",
"]",
",",
"tshape",
"[",
"1",
"]",
",",
"tshape",
"[",
"2",
"]",
"*",
"io_depth",
",",
"hidden_size",
"]",
")"
] |
Bottom transformation for image targets.
|
[
"Bottom",
"transformation",
"for",
"image",
"targets",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L346-L356
|
train
|
Bottom transformation for image targets.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b10000 + 0o137) + chr(0b110011) + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(1590 - 1540) + chr(0b110110) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(5530 - 5419) + '\062' + chr(1858 - 1808) + '\061', 0o10), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(0b1001110 + 0o41) + chr(438 - 387) + '\061' + '\061', 0b1000), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(2568 - 2457) + chr(0b110011 + 0o0) + chr(53) + chr(0b1000 + 0o57), 0o10), ehT0Px3KOsy9(chr(48) + chr(12048 - 11937) + chr(0b110010) + chr(0b110101), 56699 - 56691), ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(3401 - 3290) + chr(49) + chr(636 - 585) + '\x36', 52376 - 52368), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1496 - 1445) + '\066' + '\065', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x32' + chr(0b110010) + chr(0b1011 + 0o50), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b110101 + 0o72) + chr(431 - 380) + chr(2240 - 2188) + chr(52), 0b1000), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(111) + chr(0b110011) + chr(0b1000 + 0o55) + chr(0b110111), 8), ehT0Px3KOsy9(chr(1279 - 1231) + '\x6f' + chr(2145 - 2095) + chr(0b110000) + chr(49), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(49) + '\067' + chr(983 - 932), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(51) + chr(49) + '\x36', 0b1000), ehT0Px3KOsy9(chr(914 - 866) + '\x6f' + chr(49) + '\066' + chr(0b110100 + 0o3), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110011) + '\060' + '\062', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\x31' + chr(0b110010) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x33' + chr(2868 - 2813) + '\066', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(2152 - 2103) + chr(54) + '\066', 0o10), ehT0Px3KOsy9(chr(1266 - 1218) + '\157' + '\067' + chr(55), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010) + '\062' + chr(1507 - 1459), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\061' + '\x36' + '\x36', 8), ehT0Px3KOsy9(chr(48) + '\x6f' + '\061' + chr(0b1001 + 0o56) + chr(52), 64275 - 64267), ehT0Px3KOsy9(chr(572 - 524) + '\157' + '\x32' + chr(54), 7549 - 7541), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(50) + chr(50) + chr(52), 42190 - 42182), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(0b1011001 + 0o26) + chr(2183 - 2132) + chr(166 - 112) + chr(0b111 + 0o55), 0b1000), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(0b1101111) + chr(284 - 234) + chr(0b111 + 0o57) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(1408 - 1360) + chr(4023 - 3912) + '\062' + chr(1313 - 1264) + '\066', 38077 - 38069), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(0b1101111) + chr(69 - 19) + chr(0b110000) + chr(0b10 + 0o62), ord("\x08")), ehT0Px3KOsy9('\060' + chr(10896 - 10785) + chr(0b110010) + chr(2160 - 2112) + chr(0b110010), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(51) + chr(1107 - 1053) + chr(0b110111), 28150 - 28142), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x32' + chr(339 - 291) + chr(0b110101), 0o10), ehT0Px3KOsy9('\x30' + chr(0b111111 + 0o60) + '\x33' + '\x34' + '\x30', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1231 - 1182) + '\x31' + '\066', ord("\x08")), ehT0Px3KOsy9(chr(1265 - 1217) + chr(0b1101111) + chr(0b110010) + chr(0b1010 + 0o50) + '\062', 0b1000), ehT0Px3KOsy9(chr(1831 - 1783) + chr(111) + chr(0b110001) + '\x31' + chr(0b110110), 8), ehT0Px3KOsy9(chr(1524 - 1476) + chr(0b1101111) + '\061' + chr(0b110101 + 0o1) + chr(53), 9168 - 9160)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + '\157' + '\065' + chr(616 - 568), 22236 - 22228)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'3'), chr(100) + chr(0b1001010 + 0o33) + chr(0b101001 + 0o72) + chr(11960 - 11849) + chr(0b10011 + 0o121) + chr(101))('\x75' + chr(6856 - 6740) + chr(0b1100110) + chr(0b101101) + chr(2066 - 2010)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def CLe5qtCdo0pJ(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del CeyMIoSyrpkQ
vXoupepMtCXU = IDJ2eXGCBCDu.to_int32(OeWW0F1dBPRQ)
Q4VesPieKI0W = tq24Tk6UZ6u1.X1ZpHSxyKbHn
_67Eltc3Jvt8 = jSKPaHwSAfVv.shape_list(vXoupepMtCXU)
qzoyXN3kdhDL = tq24Tk6UZ6u1.qzoyXN3kdhDL
cmo4EAY8dayI = oIL3U1EOcJgs.get_channel_embeddings(Q4VesPieKI0W, vXoupepMtCXU, qzoyXN3kdhDL, xafqLlk3kkUe(SXOLrMavuUCe(b't\xde\xe1\xbc\x9bJ<\x9a\xe7\x90\x80\xec'), chr(0b1011001 + 0o13) + chr(101) + chr(5733 - 5634) + chr(0b1101111) + chr(0b1111 + 0o125) + chr(0b101010 + 0o73))(chr(9606 - 9489) + '\164' + chr(102) + '\055' + '\x38'))
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'o\xd5\xe2\xa1\x8ee;'), chr(0b1100100) + chr(0b10000 + 0o125) + chr(0b11111 + 0o104) + chr(111) + chr(100) + '\x65')(chr(825 - 708) + chr(116) + chr(102) + chr(0b101101) + '\x38'))(cmo4EAY8dayI, [_67Eltc3Jvt8[ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110000), 8)], _67Eltc3Jvt8[ehT0Px3KOsy9('\060' + '\x6f' + chr(576 - 527), 0o10)], _67Eltc3Jvt8[ehT0Px3KOsy9(chr(48) + '\157' + chr(1752 - 1702), ord("\x08"))] * Q4VesPieKI0W, qzoyXN3kdhDL])
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
speech_recognition_bottom
|
def speech_recognition_bottom(x, model_hparams, vocab_size):
"""Use batchnorm instead of CMVN and shorten the stft with strided convs.
Args:
x: float32 tensor with shape [batch_size, len, 1, freqs * channels]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
float32 tensor with shape [batch_size, shorter_len, 1, hidden_size]
"""
del vocab_size # unused arg
inputs = x
p = model_hparams
num_mel_bins = p.audio_num_mel_bins
num_channels = 3 if p.audio_add_delta_deltas else 1
with tf.variable_scope("speech_recognition_modality"):
if p.audio_preproc_in_bottom:
# Compute filterbanks
with tf.variable_scope("fbanks"):
waveforms = tf.squeeze(inputs, [2, 3])
mel_fbanks = common_audio.compute_mel_filterbank_features(
waveforms,
sample_rate=p.audio_sample_rate,
dither=p.audio_dither,
preemphasis=p.audio_preemphasis,
frame_length=p.audio_frame_length,
frame_step=p.audio_frame_step,
lower_edge_hertz=p.audio_lower_edge_hertz,
upper_edge_hertz=p.audio_upper_edge_hertz,
num_mel_bins=p.audio_num_mel_bins,
apply_mask=True)
if p.audio_add_delta_deltas:
mel_fbanks = common_audio.add_delta_deltas(mel_fbanks)
x = tf.reshape(mel_fbanks,
common_layers.shape_list(mel_fbanks)[:2] +
[num_mel_bins, num_channels])
nonpadding_mask = 1. - common_attention.embedding_to_padding(x)
num_of_nonpadding_elements = tf.reduce_sum(
nonpadding_mask) * num_mel_bins * num_channels
# This replaces CMVN estimation on data
var_epsilon = 1e-09
mean = tf.reduce_sum(
x, axis=[1], keepdims=True) / num_of_nonpadding_elements
variance = (num_of_nonpadding_elements * mean**2. -
2. * mean * tf.reduce_sum(x, axis=[1], keepdims=True) +
tf.reduce_sum(x**2, axis=[1], keepdims=True)
) / num_of_nonpadding_elements
x = (x - mean) * tf.rsqrt(variance + var_epsilon) * tf.expand_dims(
nonpadding_mask, -1)
else:
x = inputs
# The convention is that the models are flattened along the spatial,
# dimensions, thus the speech preprocessor treats frequencies and
# channels as image colors (last axis)
x.set_shape([None, None, num_mel_bins, num_channels])
# TODO(chorowski): how to specify bottom's hparams and avoid hardcoding?
x = tf.pad(x, [[0, 0], [0, 8], [0, 0], [0, 0]])
for _ in range(2):
x = tf.layers.conv2d(
x, 128, (3, 3), (2, 2), use_bias=False)
x = common_layers.layer_norm(x)
x = tf.nn.relu(x)
xshape = common_layers.shape_list(x)
# apply a conv that will remove all frequencies and at the same time
# project the output into desired hidden_size
x = tf.pad(x, [[0, 0], [0, 2], [0, 0], [0, 0]])
x = tf.layers.conv2d(x, p.hidden_size, (3, xshape[2]), use_bias=False)
assert common_layers.shape_list(x)[2] == 1
x = common_layers.layer_norm(x)
x = tf.nn.relu(x)
return x
|
python
|
def speech_recognition_bottom(x, model_hparams, vocab_size):
"""Use batchnorm instead of CMVN and shorten the stft with strided convs.
Args:
x: float32 tensor with shape [batch_size, len, 1, freqs * channels]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
float32 tensor with shape [batch_size, shorter_len, 1, hidden_size]
"""
del vocab_size # unused arg
inputs = x
p = model_hparams
num_mel_bins = p.audio_num_mel_bins
num_channels = 3 if p.audio_add_delta_deltas else 1
with tf.variable_scope("speech_recognition_modality"):
if p.audio_preproc_in_bottom:
# Compute filterbanks
with tf.variable_scope("fbanks"):
waveforms = tf.squeeze(inputs, [2, 3])
mel_fbanks = common_audio.compute_mel_filterbank_features(
waveforms,
sample_rate=p.audio_sample_rate,
dither=p.audio_dither,
preemphasis=p.audio_preemphasis,
frame_length=p.audio_frame_length,
frame_step=p.audio_frame_step,
lower_edge_hertz=p.audio_lower_edge_hertz,
upper_edge_hertz=p.audio_upper_edge_hertz,
num_mel_bins=p.audio_num_mel_bins,
apply_mask=True)
if p.audio_add_delta_deltas:
mel_fbanks = common_audio.add_delta_deltas(mel_fbanks)
x = tf.reshape(mel_fbanks,
common_layers.shape_list(mel_fbanks)[:2] +
[num_mel_bins, num_channels])
nonpadding_mask = 1. - common_attention.embedding_to_padding(x)
num_of_nonpadding_elements = tf.reduce_sum(
nonpadding_mask) * num_mel_bins * num_channels
# This replaces CMVN estimation on data
var_epsilon = 1e-09
mean = tf.reduce_sum(
x, axis=[1], keepdims=True) / num_of_nonpadding_elements
variance = (num_of_nonpadding_elements * mean**2. -
2. * mean * tf.reduce_sum(x, axis=[1], keepdims=True) +
tf.reduce_sum(x**2, axis=[1], keepdims=True)
) / num_of_nonpadding_elements
x = (x - mean) * tf.rsqrt(variance + var_epsilon) * tf.expand_dims(
nonpadding_mask, -1)
else:
x = inputs
# The convention is that the models are flattened along the spatial,
# dimensions, thus the speech preprocessor treats frequencies and
# channels as image colors (last axis)
x.set_shape([None, None, num_mel_bins, num_channels])
# TODO(chorowski): how to specify bottom's hparams and avoid hardcoding?
x = tf.pad(x, [[0, 0], [0, 8], [0, 0], [0, 0]])
for _ in range(2):
x = tf.layers.conv2d(
x, 128, (3, 3), (2, 2), use_bias=False)
x = common_layers.layer_norm(x)
x = tf.nn.relu(x)
xshape = common_layers.shape_list(x)
# apply a conv that will remove all frequencies and at the same time
# project the output into desired hidden_size
x = tf.pad(x, [[0, 0], [0, 2], [0, 0], [0, 0]])
x = tf.layers.conv2d(x, p.hidden_size, (3, xshape[2]), use_bias=False)
assert common_layers.shape_list(x)[2] == 1
x = common_layers.layer_norm(x)
x = tf.nn.relu(x)
return x
|
[
"def",
"speech_recognition_bottom",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"vocab_size",
"# unused arg",
"inputs",
"=",
"x",
"p",
"=",
"model_hparams",
"num_mel_bins",
"=",
"p",
".",
"audio_num_mel_bins",
"num_channels",
"=",
"3",
"if",
"p",
".",
"audio_add_delta_deltas",
"else",
"1",
"with",
"tf",
".",
"variable_scope",
"(",
"\"speech_recognition_modality\"",
")",
":",
"if",
"p",
".",
"audio_preproc_in_bottom",
":",
"# Compute filterbanks",
"with",
"tf",
".",
"variable_scope",
"(",
"\"fbanks\"",
")",
":",
"waveforms",
"=",
"tf",
".",
"squeeze",
"(",
"inputs",
",",
"[",
"2",
",",
"3",
"]",
")",
"mel_fbanks",
"=",
"common_audio",
".",
"compute_mel_filterbank_features",
"(",
"waveforms",
",",
"sample_rate",
"=",
"p",
".",
"audio_sample_rate",
",",
"dither",
"=",
"p",
".",
"audio_dither",
",",
"preemphasis",
"=",
"p",
".",
"audio_preemphasis",
",",
"frame_length",
"=",
"p",
".",
"audio_frame_length",
",",
"frame_step",
"=",
"p",
".",
"audio_frame_step",
",",
"lower_edge_hertz",
"=",
"p",
".",
"audio_lower_edge_hertz",
",",
"upper_edge_hertz",
"=",
"p",
".",
"audio_upper_edge_hertz",
",",
"num_mel_bins",
"=",
"p",
".",
"audio_num_mel_bins",
",",
"apply_mask",
"=",
"True",
")",
"if",
"p",
".",
"audio_add_delta_deltas",
":",
"mel_fbanks",
"=",
"common_audio",
".",
"add_delta_deltas",
"(",
"mel_fbanks",
")",
"x",
"=",
"tf",
".",
"reshape",
"(",
"mel_fbanks",
",",
"common_layers",
".",
"shape_list",
"(",
"mel_fbanks",
")",
"[",
":",
"2",
"]",
"+",
"[",
"num_mel_bins",
",",
"num_channels",
"]",
")",
"nonpadding_mask",
"=",
"1.",
"-",
"common_attention",
".",
"embedding_to_padding",
"(",
"x",
")",
"num_of_nonpadding_elements",
"=",
"tf",
".",
"reduce_sum",
"(",
"nonpadding_mask",
")",
"*",
"num_mel_bins",
"*",
"num_channels",
"# This replaces CMVN estimation on data",
"var_epsilon",
"=",
"1e-09",
"mean",
"=",
"tf",
".",
"reduce_sum",
"(",
"x",
",",
"axis",
"=",
"[",
"1",
"]",
",",
"keepdims",
"=",
"True",
")",
"/",
"num_of_nonpadding_elements",
"variance",
"=",
"(",
"num_of_nonpadding_elements",
"*",
"mean",
"**",
"2.",
"-",
"2.",
"*",
"mean",
"*",
"tf",
".",
"reduce_sum",
"(",
"x",
",",
"axis",
"=",
"[",
"1",
"]",
",",
"keepdims",
"=",
"True",
")",
"+",
"tf",
".",
"reduce_sum",
"(",
"x",
"**",
"2",
",",
"axis",
"=",
"[",
"1",
"]",
",",
"keepdims",
"=",
"True",
")",
")",
"/",
"num_of_nonpadding_elements",
"x",
"=",
"(",
"x",
"-",
"mean",
")",
"*",
"tf",
".",
"rsqrt",
"(",
"variance",
"+",
"var_epsilon",
")",
"*",
"tf",
".",
"expand_dims",
"(",
"nonpadding_mask",
",",
"-",
"1",
")",
"else",
":",
"x",
"=",
"inputs",
"# The convention is that the models are flattened along the spatial,",
"# dimensions, thus the speech preprocessor treats frequencies and",
"# channels as image colors (last axis)",
"x",
".",
"set_shape",
"(",
"[",
"None",
",",
"None",
",",
"num_mel_bins",
",",
"num_channels",
"]",
")",
"# TODO(chorowski): how to specify bottom's hparams and avoid hardcoding?",
"x",
"=",
"tf",
".",
"pad",
"(",
"x",
",",
"[",
"[",
"0",
",",
"0",
"]",
",",
"[",
"0",
",",
"8",
"]",
",",
"[",
"0",
",",
"0",
"]",
",",
"[",
"0",
",",
"0",
"]",
"]",
")",
"for",
"_",
"in",
"range",
"(",
"2",
")",
":",
"x",
"=",
"tf",
".",
"layers",
".",
"conv2d",
"(",
"x",
",",
"128",
",",
"(",
"3",
",",
"3",
")",
",",
"(",
"2",
",",
"2",
")",
",",
"use_bias",
"=",
"False",
")",
"x",
"=",
"common_layers",
".",
"layer_norm",
"(",
"x",
")",
"x",
"=",
"tf",
".",
"nn",
".",
"relu",
"(",
"x",
")",
"xshape",
"=",
"common_layers",
".",
"shape_list",
"(",
"x",
")",
"# apply a conv that will remove all frequencies and at the same time",
"# project the output into desired hidden_size",
"x",
"=",
"tf",
".",
"pad",
"(",
"x",
",",
"[",
"[",
"0",
",",
"0",
"]",
",",
"[",
"0",
",",
"2",
"]",
",",
"[",
"0",
",",
"0",
"]",
",",
"[",
"0",
",",
"0",
"]",
"]",
")",
"x",
"=",
"tf",
".",
"layers",
".",
"conv2d",
"(",
"x",
",",
"p",
".",
"hidden_size",
",",
"(",
"3",
",",
"xshape",
"[",
"2",
"]",
")",
",",
"use_bias",
"=",
"False",
")",
"assert",
"common_layers",
".",
"shape_list",
"(",
"x",
")",
"[",
"2",
"]",
"==",
"1",
"x",
"=",
"common_layers",
".",
"layer_norm",
"(",
"x",
")",
"x",
"=",
"tf",
".",
"nn",
".",
"relu",
"(",
"x",
")",
"return",
"x"
] |
Use batchnorm instead of CMVN and shorten the stft with strided convs.
Args:
x: float32 tensor with shape [batch_size, len, 1, freqs * channels]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
float32 tensor with shape [batch_size, shorter_len, 1, hidden_size]
|
[
"Use",
"batchnorm",
"instead",
"of",
"CMVN",
"and",
"shorten",
"the",
"stft",
"with",
"strided",
"convs",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L373-L452
|
train
|
Bottom of speech recognition.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x36' + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(0b10 + 0o62) + chr(0b10100 + 0o35), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101000 + 0o7) + chr(0b110011) + chr(0b101110 + 0o7) + chr(1345 - 1297), 23959 - 23951), ehT0Px3KOsy9('\x30' + '\x6f' + '\062' + chr(0b110000) + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010111 + 0o30) + '\061' + '\x31' + chr(0b10100 + 0o40), 35167 - 35159), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110101) + '\x34', 0b1000), ehT0Px3KOsy9('\060' + '\157' + '\063' + '\x33' + '\x35', 38772 - 38764), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(1772 - 1722) + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(2264 - 2213) + chr(0b110111), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b100101 + 0o14) + chr(0b110111) + chr(0b11100 + 0o30), 49223 - 49215), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\067' + '\060', 29639 - 29631), ehT0Px3KOsy9(chr(48) + chr(2516 - 2405) + chr(0b101000 + 0o11) + chr(772 - 717) + chr(803 - 749), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b100 + 0o55) + chr(52) + '\x32', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(128 - 78) + '\060' + '\063', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(378 - 327) + chr(51) + '\x34', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110010) + chr(1355 - 1301) + '\065', 34543 - 34535), ehT0Px3KOsy9('\060' + '\157' + chr(1703 - 1651) + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(50) + chr(51) + '\065', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110101) + chr(50), 0b1000), ehT0Px3KOsy9(chr(442 - 394) + chr(0b111000 + 0o67) + chr(51) + chr(0b110000) + chr(50), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(49) + chr(51) + '\x33', 51478 - 51470), ehT0Px3KOsy9('\x30' + chr(3415 - 3304) + chr(0b1011 + 0o47) + chr(0b1110 + 0o47) + '\067', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + chr(52) + chr(55), 0o10), ehT0Px3KOsy9(chr(0b10101 + 0o33) + '\x6f' + chr(206 - 157) + '\060' + '\062', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(218 - 169) + '\x36', 63351 - 63343), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b101101 + 0o4) + chr(51) + chr(55), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + '\x31' + chr(55), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1128 - 1073) + chr(53), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(11847 - 11736) + '\066' + '\x36', 0b1000), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1101111 + 0o0) + '\x33' + chr(1650 - 1598) + chr(0b110100), 2766 - 2758), ehT0Px3KOsy9(chr(0b110000) + chr(0b1001110 + 0o41) + chr(0b1101 + 0o51) + '\x32', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110111) + chr(0b110111), 0o10), ehT0Px3KOsy9('\060' + '\157' + '\x32' + chr(0b100111 + 0o17) + '\062', 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(51) + chr(55) + chr(53), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b1100 + 0o46) + chr(49) + chr(0b110111), 8), ehT0Px3KOsy9('\060' + chr(0b1000000 + 0o57) + chr(360 - 310) + chr(52), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x31' + '\x32' + '\x32', 13375 - 13367), ehT0Px3KOsy9(chr(0b11 + 0o55) + '\157' + chr(0b110010) + chr(190 - 142) + chr(107 - 58), 0o10), ehT0Px3KOsy9(chr(48) + chr(3372 - 3261) + '\061' + '\x35' + '\x37', 0b1000), ehT0Px3KOsy9(chr(644 - 596) + chr(12020 - 11909) + chr(0b100101 + 0o15) + chr(0b110011 + 0o2) + chr(0b100101 + 0o21), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(9473 - 9362) + '\065' + '\060', 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x1f'), chr(7923 - 7823) + chr(0b1100101) + chr(99) + chr(0b1000111 + 0o50) + '\144' + chr(0b1011000 + 0o15))('\165' + chr(0b1110100) + chr(102) + '\055' + chr(0b110100 + 0o4)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def BUwtcxYvLr23(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del CeyMIoSyrpkQ
vXoupepMtCXU = OeWW0F1dBPRQ
UyakMW2IMFEj = tq24Tk6UZ6u1
Df5qJhi6L1Sy = UyakMW2IMFEj.audio_num_mel_bins
X1ZpHSxyKbHn = ehT0Px3KOsy9(chr(330 - 282) + chr(111) + chr(0b110011), 0o10) if UyakMW2IMFEj.audio_add_delta_deltas else ehT0Px3KOsy9(chr(48) + chr(0b111100 + 0o63) + '\x31', 0b1000)
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'G@\x9a\xef-N\xb1p@\x0f[\x11\x12\xe6'), '\x64' + chr(101) + '\x63' + chr(111) + chr(1942 - 1842) + chr(0b1000 + 0o135))(chr(117) + chr(8506 - 8390) + chr(102) + chr(807 - 762) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'BQ\x8d\xe3/D\x82gz\x1fW\x19\x0c\xea\xa8\xdcJ\xf3u\xaf4\xcb\x06.g+\xf6'), chr(0b11110 + 0o106) + chr(101) + chr(0b111100 + 0o47) + chr(0b1101111) + '\144' + chr(5066 - 4965))('\x75' + chr(5104 - 4988) + chr(0b101010 + 0o74) + chr(0b101101) + '\x38')):
if xafqLlk3kkUe(UyakMW2IMFEj, xafqLlk3kkUe(SXOLrMavuUCe(b'PT\x8c\xef#s\xadgz\x0cJ\x11\x01\xdc\xb5\xdbz\xffE\xb6/\xc0\n'), chr(0b1100100) + '\145' + '\x63' + chr(11456 - 11345) + '\x64' + '\x65')('\165' + chr(116) + '\146' + chr(1160 - 1115) + chr(0b10 + 0o66))):
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'G@\x9a\xef-N\xb1p@\x0f[\x11\x12\xe6'), chr(0b11110 + 0o106) + chr(101) + '\x63' + chr(1323 - 1212) + chr(100) + chr(0b11100 + 0o111))('\165' + chr(0b10111 + 0o135) + '\x66' + chr(1535 - 1490) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b"WC\x89\xe8'_"), chr(0b1100100) + chr(101) + chr(99) + '\157' + chr(0b1100100) + chr(0b1001 + 0o134))('\165' + chr(116) + chr(8755 - 8653) + chr(0b101101) + chr(56))):
yP0a4QRs0KoT = IDJ2eXGCBCDu.squeeze(vXoupepMtCXU, [ehT0Px3KOsy9('\060' + chr(7553 - 7442) + chr(0b110010 + 0o0), 45083 - 45075), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110011), 8)])
FHplnhVY104Z = z4qcsyFe_dDD.compute_mel_filterbank_features(yP0a4QRs0KoT, sample_rate=UyakMW2IMFEj.audio_sample_rate, dither=UyakMW2IMFEj.audio_dither, preemphasis=UyakMW2IMFEj.audio_preemphasis, frame_length=UyakMW2IMFEj.audio_frame_length, frame_step=UyakMW2IMFEj.audio_frame_step, lower_edge_hertz=UyakMW2IMFEj.audio_lower_edge_hertz, upper_edge_hertz=UyakMW2IMFEj.audio_upper_edge_hertz, num_mel_bins=UyakMW2IMFEj.audio_num_mel_bins, apply_mask=ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(49), 8))
if xafqLlk3kkUe(UyakMW2IMFEj, xafqLlk3kkUe(SXOLrMavuUCe(b'PT\x8c\xef#s\xbcq{#\\\x1b\x0e\xf7\xbd\xeaA\xf8F\xb6:\xdc'), chr(100) + chr(0b1010 + 0o133) + chr(0b1100011) + chr(2105 - 1994) + chr(0b1100100) + '\145')(chr(4233 - 4116) + chr(0b1110100) + '\x66' + chr(0b10000 + 0o35) + '\x38')):
FHplnhVY104Z = z4qcsyFe_dDD.add_delta_deltas(FHplnhVY104Z)
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reshape(FHplnhVY104Z, jSKPaHwSAfVv.shape_list(FHplnhVY104Z)[:ehT0Px3KOsy9(chr(282 - 234) + '\x6f' + chr(50), 8)] + [Df5qJhi6L1Sy, X1ZpHSxyKbHn])
UyiM64E6iSsw = 1.0 - WOnrfm4dlYcf.embedding_to_padding(OeWW0F1dBPRQ)
zJlHoaNl04cR = IDJ2eXGCBCDu.reduce_sum(UyiM64E6iSsw) * Df5qJhi6L1Sy * X1ZpHSxyKbHn
jffdHMbLiy_2 = 1e-09
aJhItC_Vawlw = IDJ2eXGCBCDu.reduce_sum(OeWW0F1dBPRQ, axis=[ehT0Px3KOsy9(chr(48) + '\x6f' + '\061', 8)], keepdims=ehT0Px3KOsy9(chr(463 - 415) + chr(10092 - 9981) + chr(463 - 414), 8)) / zJlHoaNl04cR
nVKbP5sF7181 = (zJlHoaNl04cR * aJhItC_Vawlw ** 2.0 - 2.0 * aJhItC_Vawlw * IDJ2eXGCBCDu.reduce_sum(OeWW0F1dBPRQ, axis=[ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(111) + '\061', 8)], keepdims=ehT0Px3KOsy9(chr(0b10110 + 0o32) + '\157' + '\061', 8)) + IDJ2eXGCBCDu.reduce_sum(OeWW0F1dBPRQ ** ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b0 + 0o62), 8), axis=[ehT0Px3KOsy9(chr(0b11010 + 0o26) + chr(0b1101111) + chr(1501 - 1452), 8)], keepdims=ehT0Px3KOsy9('\x30' + '\157' + '\061', 8))) / zJlHoaNl04cR
OeWW0F1dBPRQ = (OeWW0F1dBPRQ - aJhItC_Vawlw) * IDJ2eXGCBCDu.rsqrt(nVKbP5sF7181 + jffdHMbLiy_2) * IDJ2eXGCBCDu.expand_dims(UyiM64E6iSsw, -ehT0Px3KOsy9('\x30' + chr(4331 - 4220) + chr(0b110001), 8))
else:
OeWW0F1dBPRQ = vXoupepMtCXU
xafqLlk3kkUe(OeWW0F1dBPRQ, xafqLlk3kkUe(SXOLrMavuUCe(b'BD\x9c\xd9?D\xbcez'), chr(0b1100100) + chr(101) + chr(99) + '\x6f' + chr(100) + '\145')('\x75' + '\x74' + '\x66' + '\x2d' + '\070'))([None, None, Df5qJhi6L1Sy, X1ZpHSxyKbHn])
OeWW0F1dBPRQ = IDJ2eXGCBCDu.pad(OeWW0F1dBPRQ, [[ehT0Px3KOsy9(chr(900 - 852) + chr(0b1101111) + chr(48), 0o10), ehT0Px3KOsy9(chr(1994 - 1946) + chr(0b10111 + 0o130) + '\060', 8)], [ehT0Px3KOsy9(chr(0b101000 + 0o10) + '\x6f' + '\x30', 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + '\060', 0o10)], [ehT0Px3KOsy9('\x30' + chr(111) + chr(2234 - 2186), 8), ehT0Px3KOsy9(chr(1213 - 1165) + chr(0b1101111) + chr(1785 - 1737), 8)], [ehT0Px3KOsy9(chr(48) + chr(4303 - 4192) + '\060', 8), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b100011 + 0o15), 8)]])
for VNGQdHSFPrso in vQr8gNKaIaWE(ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(1549 - 1499), 8)):
OeWW0F1dBPRQ = IDJ2eXGCBCDu.layers.conv2d(OeWW0F1dBPRQ, ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + '\x30' + '\060', 0b1000), (ehT0Px3KOsy9('\060' + '\x6f' + '\063', 8), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b1100011 + 0o14) + '\063', 8)), (ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(1173 - 1123), 8), ehT0Px3KOsy9(chr(0b1000 + 0o50) + chr(0b1101111) + '\x32', 8)), use_bias=ehT0Px3KOsy9(chr(48) + '\x6f' + chr(48), 8))
OeWW0F1dBPRQ = jSKPaHwSAfVv.layer_norm(OeWW0F1dBPRQ)
OeWW0F1dBPRQ = IDJ2eXGCBCDu.nn.relu(OeWW0F1dBPRQ)
XH_WlT4vj_oW = jSKPaHwSAfVv.shape_list(OeWW0F1dBPRQ)
OeWW0F1dBPRQ = IDJ2eXGCBCDu.pad(OeWW0F1dBPRQ, [[ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b1101111) + chr(48), 8), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110000), 8)], [ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x30', 8), ehT0Px3KOsy9('\x30' + chr(111) + '\x32', 8)], [ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(48), 8), ehT0Px3KOsy9('\x30' + '\157' + chr(48), 8)], [ehT0Px3KOsy9('\x30' + '\x6f' + '\060', 8), ehT0Px3KOsy9('\x30' + chr(111) + '\060', 8)]])
OeWW0F1dBPRQ = IDJ2eXGCBCDu.layers.conv2d(OeWW0F1dBPRQ, UyakMW2IMFEj.qzoyXN3kdhDL, (ehT0Px3KOsy9('\x30' + chr(0b1000001 + 0o56) + chr(0b110011), 8), XH_WlT4vj_oW[ehT0Px3KOsy9('\060' + chr(0b1100110 + 0o11) + chr(0b110000 + 0o2), 8)]), use_bias=ehT0Px3KOsy9(chr(1736 - 1688) + chr(1991 - 1880) + chr(0b110000), 8))
assert xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'BI\x89\xf6)s\xb1|l\x08'), chr(0b1100100) + chr(0b101 + 0o140) + chr(99) + '\x6f' + chr(0b1100100) + chr(0b1100101))('\165' + chr(0b1110100) + '\x66' + '\055' + '\070'))(OeWW0F1dBPRQ)[ehT0Px3KOsy9(chr(48) + '\157' + '\x32', 8)] == ehT0Px3KOsy9(chr(1155 - 1107) + chr(111) + chr(0b1101 + 0o44), 8)
OeWW0F1dBPRQ = jSKPaHwSAfVv.layer_norm(OeWW0F1dBPRQ)
OeWW0F1dBPRQ = IDJ2eXGCBCDu.nn.relu(OeWW0F1dBPRQ)
return OeWW0F1dBPRQ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
get_weights
|
def get_weights(model_hparams, vocab_size, hidden_dim=None):
"""Create or get concatenated embedding or softmax variable.
Args:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
hidden_dim: dim of the variable. Defaults to _model_hparams' hidden_size
Returns:
a list of num_shards Tensors.
"""
if hidden_dim is None:
hidden_dim = model_hparams.hidden_size
num_shards = model_hparams.symbol_modality_num_shards
shards = []
for i in range(num_shards):
shard_size = (vocab_size // num_shards) + (
1 if i < vocab_size % num_shards else 0)
var_name = "weights_%d" % i
shards.append(
tf.get_variable(
var_name, [shard_size, hidden_dim],
initializer=tf.random_normal_initializer(0.0, hidden_dim**-0.5)))
if num_shards == 1:
ret = shards[0]
else:
ret = tf.concat(shards, 0)
# Convert ret to tensor.
if not tf.executing_eagerly():
ret = common_layers.convert_gradient_to_tensor(ret)
return ret
|
python
|
def get_weights(model_hparams, vocab_size, hidden_dim=None):
"""Create or get concatenated embedding or softmax variable.
Args:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
hidden_dim: dim of the variable. Defaults to _model_hparams' hidden_size
Returns:
a list of num_shards Tensors.
"""
if hidden_dim is None:
hidden_dim = model_hparams.hidden_size
num_shards = model_hparams.symbol_modality_num_shards
shards = []
for i in range(num_shards):
shard_size = (vocab_size // num_shards) + (
1 if i < vocab_size % num_shards else 0)
var_name = "weights_%d" % i
shards.append(
tf.get_variable(
var_name, [shard_size, hidden_dim],
initializer=tf.random_normal_initializer(0.0, hidden_dim**-0.5)))
if num_shards == 1:
ret = shards[0]
else:
ret = tf.concat(shards, 0)
# Convert ret to tensor.
if not tf.executing_eagerly():
ret = common_layers.convert_gradient_to_tensor(ret)
return ret
|
[
"def",
"get_weights",
"(",
"model_hparams",
",",
"vocab_size",
",",
"hidden_dim",
"=",
"None",
")",
":",
"if",
"hidden_dim",
"is",
"None",
":",
"hidden_dim",
"=",
"model_hparams",
".",
"hidden_size",
"num_shards",
"=",
"model_hparams",
".",
"symbol_modality_num_shards",
"shards",
"=",
"[",
"]",
"for",
"i",
"in",
"range",
"(",
"num_shards",
")",
":",
"shard_size",
"=",
"(",
"vocab_size",
"//",
"num_shards",
")",
"+",
"(",
"1",
"if",
"i",
"<",
"vocab_size",
"%",
"num_shards",
"else",
"0",
")",
"var_name",
"=",
"\"weights_%d\"",
"%",
"i",
"shards",
".",
"append",
"(",
"tf",
".",
"get_variable",
"(",
"var_name",
",",
"[",
"shard_size",
",",
"hidden_dim",
"]",
",",
"initializer",
"=",
"tf",
".",
"random_normal_initializer",
"(",
"0.0",
",",
"hidden_dim",
"**",
"-",
"0.5",
")",
")",
")",
"if",
"num_shards",
"==",
"1",
":",
"ret",
"=",
"shards",
"[",
"0",
"]",
"else",
":",
"ret",
"=",
"tf",
".",
"concat",
"(",
"shards",
",",
"0",
")",
"# Convert ret to tensor.",
"if",
"not",
"tf",
".",
"executing_eagerly",
"(",
")",
":",
"ret",
"=",
"common_layers",
".",
"convert_gradient_to_tensor",
"(",
"ret",
")",
"return",
"ret"
] |
Create or get concatenated embedding or softmax variable.
Args:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
hidden_dim: dim of the variable. Defaults to _model_hparams' hidden_size
Returns:
a list of num_shards Tensors.
|
[
"Create",
"or",
"get",
"concatenated",
"embedding",
"or",
"softmax",
"variable",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L455-L485
|
train
|
Create or get concatenated embedding or softmax variable.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\x6f' + chr(0b1101 + 0o46) + '\x36' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\063' + chr(0b10100 + 0o37) + '\x30', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b10110 + 0o35) + '\066' + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b100 + 0o56) + '\x33', ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + '\061' + chr(51) + '\x35', ord("\x08")), ehT0Px3KOsy9(chr(0b100001 + 0o17) + chr(0b100000 + 0o117) + chr(49) + chr(1134 - 1084) + chr(0b11010 + 0o35), 0b1000), ehT0Px3KOsy9(chr(0b111 + 0o51) + '\157' + chr(0b100011 + 0o16) + chr(0b10000 + 0o41) + chr(0b110101), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + '\x32' + '\066' + chr(454 - 406), 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\x31' + chr(0b110000) + chr(609 - 559), ord("\x08")), ehT0Px3KOsy9(chr(0b11100 + 0o24) + chr(0b1000011 + 0o54) + '\062' + '\061', 39963 - 39955), ehT0Px3KOsy9(chr(48) + chr(695 - 584) + '\061' + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1970 - 1920) + chr(0b11101 + 0o32) + '\x34', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b110100 + 0o73) + chr(0b101111 + 0o2) + chr(49) + chr(453 - 403), 31400 - 31392), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110011) + '\x36' + chr(1058 - 1007), ord("\x08")), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(111) + chr(0b100110 + 0o14) + '\061' + chr(0b110110), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b101 + 0o57) + chr(0b110101), 61076 - 61068), ehT0Px3KOsy9(chr(0b110000) + chr(7428 - 7317) + '\x32' + '\x30' + chr(0b11110 + 0o23), 0o10), ehT0Px3KOsy9(chr(0b11100 + 0o24) + '\157' + chr(49) + chr(54) + chr(1136 - 1082), 0b1000), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(0b111111 + 0o60) + chr(55) + '\063', 45975 - 45967), ehT0Px3KOsy9(chr(638 - 590) + '\157' + chr(0b110011) + '\065' + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x32' + chr(0b110111) + '\066', 7506 - 7498), ehT0Px3KOsy9(chr(48) + chr(285 - 174) + chr(0b110001) + chr(460 - 408) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(0b11100 + 0o24) + chr(111) + chr(0b110101) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x31' + '\062' + chr(0b10010 + 0o43), 32081 - 32073), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + chr(0b110111) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(0b1100011 + 0o14) + '\062' + chr(49) + '\062', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\062' + chr(394 - 340), ord("\x08")), ehT0Px3KOsy9(chr(168 - 120) + chr(0b1001011 + 0o44) + chr(0b1111 + 0o43) + chr(1160 - 1108) + chr(0b100 + 0o54), 11096 - 11088), ehT0Px3KOsy9(chr(0b100100 + 0o14) + '\x6f' + '\x32' + chr(0b1001 + 0o50), 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x33' + chr(0b101 + 0o57) + chr(52), 0o10), ehT0Px3KOsy9(chr(136 - 88) + chr(111) + chr(2234 - 2183) + chr(126 - 77) + '\x35', 0b1000), ehT0Px3KOsy9('\060' + chr(111) + '\x31' + '\063' + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(1764 - 1716) + '\157' + '\066' + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(0b1101111) + '\064' + '\067', 0o10), ehT0Px3KOsy9('\x30' + chr(111) + '\x32' + chr(50) + chr(898 - 843), ord("\x08")), ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(0b1101111) + chr(0b100 + 0o63) + chr(0b11001 + 0o34), 24772 - 24764), ehT0Px3KOsy9('\060' + '\x6f' + chr(54) + chr(54), 888 - 880), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b11 + 0o56) + chr(1199 - 1147) + chr(581 - 528), 0o10), ehT0Px3KOsy9('\x30' + chr(11852 - 11741) + chr(311 - 262) + chr(0b110010) + chr(55), 8), ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(111) + chr(2509 - 2455) + chr(2478 - 2424), 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(62 - 14) + chr(0b1010 + 0o145) + '\x35' + chr(48), 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xdc'), chr(4241 - 4141) + '\145' + '\143' + '\x6f' + '\144' + chr(101))(chr(0b110000 + 0o105) + chr(8706 - 8590) + chr(3580 - 3478) + chr(0b101101) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def WZaPMPa_FY26(tq24Tk6UZ6u1, CeyMIoSyrpkQ, CQFa_cRh0Tor=None):
if CQFa_cRh0Tor is None:
CQFa_cRh0Tor = tq24Tk6UZ6u1.qzoyXN3kdhDL
WJU3qUPk_Uro = tq24Tk6UZ6u1.iBYlnqUAwgIX
Yk4EJarBd6ai = []
for WVxHKyX45z_L in vQr8gNKaIaWE(WJU3qUPk_Uro):
gSOQb9H2BNO9 = CeyMIoSyrpkQ // WJU3qUPk_Uro + (ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110001), 0o10) if WVxHKyX45z_L < CeyMIoSyrpkQ % WJU3qUPk_Uro else ehT0Px3KOsy9('\060' + chr(111) + chr(48), 0o10))
rh85H97CENf3 = xafqLlk3kkUe(SXOLrMavuUCe(b'\x85\xbd\x9et5#>\xefgu'), chr(100) + chr(101) + chr(99) + chr(0b1000010 + 0o55) + chr(403 - 303) + chr(101))(chr(0b1110101) + '\x74' + chr(1700 - 1598) + chr(45) + chr(56)) % WVxHKyX45z_L
xafqLlk3kkUe(Yk4EJarBd6ai, xafqLlk3kkUe(SXOLrMavuUCe(b'\x93\xa8\x87v33'), '\144' + chr(7253 - 7152) + '\143' + chr(0b1010101 + 0o32) + chr(0b11111 + 0o105) + chr(0b11101 + 0o110))('\x75' + chr(0b1100000 + 0o24) + '\x66' + '\055' + '\070'))(xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x95\xbd\x83L+6?\xd9#sc\xd9'), chr(0b1100100) + chr(101) + chr(0b1000110 + 0o35) + chr(111) + '\x64' + chr(1342 - 1241))('\165' + chr(116) + chr(0b1100001 + 0o5) + chr(0b101101) + chr(56)))(rh85H97CENf3, [gSOQb9H2BNO9, CQFa_cRh0Tor], initializer=xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x80\xb9\x99w2:\x12\xde-cb\xdd&\xac\xe7K\xe8\x02R\x8c\xd2\x0fR0\x89'), chr(100) + '\145' + chr(2213 - 2114) + '\157' + chr(100) + chr(0b111111 + 0o46))(chr(0b11000 + 0o135) + '\x74' + chr(0b1001 + 0o135) + chr(45) + chr(0b111000)))(0.0, CQFa_cRh0Tor ** (-0.5))))
if WJU3qUPk_Uro == ehT0Px3KOsy9(chr(1786 - 1738) + chr(2038 - 1927) + chr(49), 8):
VHn4CV4Ymrei = Yk4EJarBd6ai[ehT0Px3KOsy9(chr(683 - 635) + '\x6f' + chr(0b1101 + 0o43), 8)]
else:
VHn4CV4Ymrei = IDJ2eXGCBCDu.concat(Yk4EJarBd6ai, ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2091 - 2043), 8))
if not xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x97\xa0\x92p(#$\xde%Nj\xdd-\x96\xfcI\xf8'), '\x64' + '\x65' + '\x63' + '\157' + '\x64' + chr(287 - 186))(chr(117) + '\x74' + chr(0b1100110) + chr(340 - 295) + '\070'))():
VHn4CV4Ymrei = jSKPaHwSAfVv.convert_gradient_to_tensor(VHn4CV4Ymrei)
return VHn4CV4Ymrei
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
_symbol_bottom_simple
|
def _symbol_bottom_simple(x, model_hparams, vocab_size, name, reuse):
"""Bottom transformation for symbols."""
with tf.variable_scope(name, reuse=reuse):
# Ensure the inputs are 3-D
if len(x.get_shape()) == 4:
x = tf.squeeze(x, axis=3)
while len(x.get_shape()) < 3:
x = tf.expand_dims(x, axis=-1)
var = get_weights(model_hparams, vocab_size)
x = common_layers.dropout_no_scaling(
x, 1.0 - model_hparams.symbol_dropout)
ret = common_layers.gather(var, x)
if model_hparams.multiply_embedding_mode == "sqrt_depth":
ret *= model_hparams.hidden_size**0.5
ret *= tf.expand_dims(
common_layers.cast_like(tf.not_equal(x, 0), ret), -1)
return ret
|
python
|
def _symbol_bottom_simple(x, model_hparams, vocab_size, name, reuse):
"""Bottom transformation for symbols."""
with tf.variable_scope(name, reuse=reuse):
# Ensure the inputs are 3-D
if len(x.get_shape()) == 4:
x = tf.squeeze(x, axis=3)
while len(x.get_shape()) < 3:
x = tf.expand_dims(x, axis=-1)
var = get_weights(model_hparams, vocab_size)
x = common_layers.dropout_no_scaling(
x, 1.0 - model_hparams.symbol_dropout)
ret = common_layers.gather(var, x)
if model_hparams.multiply_embedding_mode == "sqrt_depth":
ret *= model_hparams.hidden_size**0.5
ret *= tf.expand_dims(
common_layers.cast_like(tf.not_equal(x, 0), ret), -1)
return ret
|
[
"def",
"_symbol_bottom_simple",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
",",
"name",
",",
"reuse",
")",
":",
"with",
"tf",
".",
"variable_scope",
"(",
"name",
",",
"reuse",
"=",
"reuse",
")",
":",
"# Ensure the inputs are 3-D",
"if",
"len",
"(",
"x",
".",
"get_shape",
"(",
")",
")",
"==",
"4",
":",
"x",
"=",
"tf",
".",
"squeeze",
"(",
"x",
",",
"axis",
"=",
"3",
")",
"while",
"len",
"(",
"x",
".",
"get_shape",
"(",
")",
")",
"<",
"3",
":",
"x",
"=",
"tf",
".",
"expand_dims",
"(",
"x",
",",
"axis",
"=",
"-",
"1",
")",
"var",
"=",
"get_weights",
"(",
"model_hparams",
",",
"vocab_size",
")",
"x",
"=",
"common_layers",
".",
"dropout_no_scaling",
"(",
"x",
",",
"1.0",
"-",
"model_hparams",
".",
"symbol_dropout",
")",
"ret",
"=",
"common_layers",
".",
"gather",
"(",
"var",
",",
"x",
")",
"if",
"model_hparams",
".",
"multiply_embedding_mode",
"==",
"\"sqrt_depth\"",
":",
"ret",
"*=",
"model_hparams",
".",
"hidden_size",
"**",
"0.5",
"ret",
"*=",
"tf",
".",
"expand_dims",
"(",
"common_layers",
".",
"cast_like",
"(",
"tf",
".",
"not_equal",
"(",
"x",
",",
"0",
")",
",",
"ret",
")",
",",
"-",
"1",
")",
"return",
"ret"
] |
Bottom transformation for symbols.
|
[
"Bottom",
"transformation",
"for",
"symbols",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L488-L505
|
train
|
Bottom transformation for symbols.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b1000 + 0o50) + '\x6f' + chr(0b110011) + chr(0b1 + 0o64) + '\x31', 0o10), ehT0Px3KOsy9('\060' + chr(1598 - 1487) + chr(0b10011 + 0o41), ord("\x08")), ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(111) + '\x33' + '\065' + chr(0b11011 + 0o33), 36951 - 36943), ehT0Px3KOsy9('\060' + chr(0b101100 + 0o103) + chr(50) + '\x37' + '\x30', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(0b110010) + '\062', 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + '\063' + chr(0b10010 + 0o45), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b100010 + 0o21) + '\x34' + '\060', 62974 - 62966), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10000 + 0o42) + '\x30', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1010101 + 0o32) + '\062' + '\064' + chr(54), 25413 - 25405), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(971 - 921) + chr(55) + '\x31', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(55), ord("\x08")), ehT0Px3KOsy9('\060' + chr(4517 - 4406) + chr(0b110001 + 0o2) + '\x31' + chr(0b110001), 23971 - 23963), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(7014 - 6903) + chr(0b10000 + 0o42) + chr(0b110110) + '\x32', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\061' + chr(0b100011 + 0o23) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(842 - 794) + chr(9135 - 9024) + chr(0b1011 + 0o46) + '\067' + '\062', 39560 - 39552), ehT0Px3KOsy9('\x30' + '\x6f' + chr(1550 - 1501) + chr(50), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\064' + chr(49), 8384 - 8376), ehT0Px3KOsy9(chr(0b1 + 0o57) + '\x6f' + '\x30', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(9555 - 9444) + chr(51) + chr(1911 - 1861) + chr(2141 - 2093), 18212 - 18204), ehT0Px3KOsy9(chr(1453 - 1405) + chr(111) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(852 - 804) + '\x6f' + chr(207 - 154) + chr(178 - 130), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(51) + chr(0b110110) + '\062', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110001) + chr(2010 - 1958) + '\064', 35108 - 35100), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(911 - 861) + '\067', 0b1000), ehT0Px3KOsy9(chr(1121 - 1073) + chr(111) + chr(49) + '\066' + chr(2686 - 2633), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(1266 - 1216) + chr(52) + '\x32', 21944 - 21936), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(0b1001 + 0o146) + chr(0b101100 + 0o7) + '\061' + chr(48), 32135 - 32127), ehT0Px3KOsy9(chr(0b110000) + chr(0b10100 + 0o133) + chr(0b110010) + chr(0b101100 + 0o10) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b11101 + 0o23) + chr(0b101 + 0o152) + chr(0b1100 + 0o47) + chr(2363 - 2311) + chr(48), 8), ehT0Px3KOsy9(chr(1856 - 1808) + '\x6f' + chr(0b110011) + chr(1253 - 1201) + chr(0b10110 + 0o36), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\063' + chr(0b110111) + '\x35', 56978 - 56970), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010) + chr(0b11100 + 0o26) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b1101 + 0o43) + '\157' + chr(0b110000 + 0o2) + chr(1349 - 1299) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011) + '\x35' + chr(671 - 623), ord("\x08")), ehT0Px3KOsy9('\060' + chr(1508 - 1397) + chr(55) + chr(55), 0o10), ehT0Px3KOsy9(chr(0b101001 + 0o7) + '\157' + '\x32' + chr(0b110110) + chr(371 - 321), 8), ehT0Px3KOsy9('\x30' + '\x6f' + chr(51) + '\x36', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b101111 + 0o3) + chr(1848 - 1797) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b10011 + 0o134) + chr(1404 - 1353) + '\064' + '\x34', 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\062' + '\060' + chr(55), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(9396 - 9285) + chr(53) + '\060', 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xc7'), chr(0b1011111 + 0o5) + chr(101) + chr(0b1100001 + 0o2) + chr(1560 - 1449) + '\x64' + chr(101))('\x75' + '\x74' + '\x66' + chr(45) + chr(2880 - 2824)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def hUdBpNM71xFW(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ, AIvJRzLdDfgF, pmC5wdSFgdFj):
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x9fe\x9c\xd1\xb3\xbd\x8cu\xf9\xcad\xa6\xc2\x17'), chr(0b1100100) + chr(0b1100101) + chr(99) + '\x6f' + chr(0b1000100 + 0o40) + chr(0b111010 + 0o53))(chr(7702 - 7585) + chr(0b1110100) + chr(6342 - 6240) + '\055' + chr(56)))(AIvJRzLdDfgF, reuse=pmC5wdSFgdFj):
if c2A0yzQpDQB3(xafqLlk3kkUe(OeWW0F1dBPRQ, xafqLlk3kkUe(SXOLrMavuUCe(b'\x8ea\x9a\xe7\xa1\xb7\x81`\xc3'), chr(2236 - 2136) + chr(0b1010101 + 0o20) + '\x63' + chr(0b1101111) + chr(100) + chr(101))(chr(0b1011111 + 0o26) + chr(2177 - 2061) + '\x66' + chr(699 - 654) + chr(0b111000)))()) == ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b100010 + 0o22), 8):
OeWW0F1dBPRQ = IDJ2eXGCBCDu.squeeze(OeWW0F1dBPRQ, axis=ehT0Px3KOsy9('\060' + chr(0b1101111) + '\063', ord("\x08")))
while c2A0yzQpDQB3(xafqLlk3kkUe(OeWW0F1dBPRQ, xafqLlk3kkUe(SXOLrMavuUCe(b'\x8ea\x9a\xe7\xa1\xb7\x81`\xc3'), chr(0b1100100) + '\x65' + chr(0b1100011) + '\157' + chr(1650 - 1550) + '\x65')(chr(117) + chr(5171 - 5055) + chr(102) + '\x2d' + chr(56)))()) < ehT0Px3KOsy9(chr(48) + chr(0b10 + 0o155) + chr(0b110011), 8):
OeWW0F1dBPRQ = IDJ2eXGCBCDu.expand_dims(OeWW0F1dBPRQ, axis=-ehT0Px3KOsy9(chr(48) + chr(111) + chr(1404 - 1355), 37886 - 37878))
l38lb8xQZNsE = WZaPMPa_FY26(tq24Tk6UZ6u1, CeyMIoSyrpkQ)
OeWW0F1dBPRQ = jSKPaHwSAfVv.dropout_no_scaling(OeWW0F1dBPRQ, 1.0 - tq24Tk6UZ6u1.ycYLHKnRG3mu)
VHn4CV4Ymrei = jSKPaHwSAfVv.gather(l38lb8xQZNsE, OeWW0F1dBPRQ)
if xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b"\x981\xbb\xfd\xa2\x97\xad'\xfc\xf0k\x9d"), chr(6654 - 6554) + '\x65' + chr(0b1100011) + '\157' + chr(0b0 + 0o144) + '\x65')(chr(12057 - 11940) + chr(0b1010100 + 0o40) + chr(0b1100110) + chr(246 - 201) + '\070')) == xafqLlk3kkUe(SXOLrMavuUCe(b'\x9au\x9c\xcc\x8d\xbb\x85`\xd2\xd1'), chr(0b1100100) + chr(5687 - 5586) + chr(0b1001000 + 0o33) + chr(0b1101111) + '\144' + '\x65')(chr(0b1110101) + '\x74' + chr(102) + '\055' + chr(0b111000)):
VHn4CV4Ymrei *= tq24Tk6UZ6u1.qzoyXN3kdhDL ** 0.5
VHn4CV4Ymrei *= IDJ2eXGCBCDu.expand_dims(jSKPaHwSAfVv.cast_like(IDJ2eXGCBCDu.not_equal(OeWW0F1dBPRQ, ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10010 + 0o36), 8)), VHn4CV4Ymrei), -ehT0Px3KOsy9(chr(0b110000) + chr(0b11001 + 0o126) + '\061', 8))
return VHn4CV4Ymrei
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
symbol_targets_bottom
|
def symbol_targets_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for target symbols."""
if (model_hparams.shared_embedding_and_softmax_weights or
model_hparams.get("shared_embedding")):
try:
return _symbol_bottom_simple(
x, model_hparams, vocab_size, "shared", reuse=True)
except ValueError:
# perhaps there were no inputs, and this is a new variable.
return _symbol_bottom_simple(
x, model_hparams, vocab_size, "shared", reuse=None)
else:
return _symbol_bottom_simple(
x, model_hparams, vocab_size, "target_emb", reuse=None)
|
python
|
def symbol_targets_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for target symbols."""
if (model_hparams.shared_embedding_and_softmax_weights or
model_hparams.get("shared_embedding")):
try:
return _symbol_bottom_simple(
x, model_hparams, vocab_size, "shared", reuse=True)
except ValueError:
# perhaps there were no inputs, and this is a new variable.
return _symbol_bottom_simple(
x, model_hparams, vocab_size, "shared", reuse=None)
else:
return _symbol_bottom_simple(
x, model_hparams, vocab_size, "target_emb", reuse=None)
|
[
"def",
"symbol_targets_bottom",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"if",
"(",
"model_hparams",
".",
"shared_embedding_and_softmax_weights",
"or",
"model_hparams",
".",
"get",
"(",
"\"shared_embedding\"",
")",
")",
":",
"try",
":",
"return",
"_symbol_bottom_simple",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
",",
"\"shared\"",
",",
"reuse",
"=",
"True",
")",
"except",
"ValueError",
":",
"# perhaps there were no inputs, and this is a new variable.",
"return",
"_symbol_bottom_simple",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
",",
"\"shared\"",
",",
"reuse",
"=",
"None",
")",
"else",
":",
"return",
"_symbol_bottom_simple",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
",",
"\"target_emb\"",
",",
"reuse",
"=",
"None",
")"
] |
Bottom transformation for target symbols.
|
[
"Bottom",
"transformation",
"for",
"target",
"symbols",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L517-L530
|
train
|
Bottom transformation for target symbols.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(756 - 708) + chr(0b1000010 + 0o55) + chr(664 - 614) + '\x36' + chr(49), 64689 - 64681), ehT0Px3KOsy9('\x30' + chr(0b11 + 0o154) + chr(1160 - 1108) + chr(1666 - 1617), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x37' + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b10010 + 0o36) + chr(0b1011100 + 0o23) + chr(0b100111 + 0o13) + chr(0b10001 + 0o40) + chr(55), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b1111 + 0o44) + chr(0b110010) + chr(0b100101 + 0o14), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b10000 + 0o137) + '\x31' + chr(55) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + '\x33' + '\064' + chr(732 - 683), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x32' + chr(0b110101) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110010) + chr(0b110101) + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x32' + '\x33' + chr(48), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + '\x32' + chr(0b11010 + 0o30) + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(0b10001 + 0o37) + '\157' + chr(1971 - 1921) + chr(0b110101) + chr(51), 8), ehT0Px3KOsy9(chr(48) + chr(111) + '\061' + chr(51) + chr(55), 0b1000), ehT0Px3KOsy9(chr(0b10100 + 0o34) + '\157' + chr(2002 - 1949) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b111 + 0o150) + '\065' + chr(51), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b11000 + 0o127) + chr(0b110100) + chr(0b110001), 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\062' + chr(1469 - 1421) + '\x32', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b1011 + 0o47) + chr(54) + '\062', 4886 - 4878), ehT0Px3KOsy9(chr(786 - 738) + chr(0b1101111) + chr(49) + chr(0b110001) + '\061', 31641 - 31633), ehT0Px3KOsy9(chr(2030 - 1982) + chr(0b1101111 + 0o0) + '\063' + chr(54) + '\060', 40663 - 40655), ehT0Px3KOsy9('\060' + chr(111) + chr(0b101101 + 0o4) + chr(0b101110 + 0o3) + chr(0b10010 + 0o36), 0b1000), ehT0Px3KOsy9('\060' + '\157' + '\064' + chr(2432 - 2377), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(879 - 829) + chr(0b1010 + 0o55) + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b11101 + 0o23) + chr(111) + '\067', ord("\x08")), ehT0Px3KOsy9(chr(0b1000 + 0o50) + chr(0b1010000 + 0o37) + '\x32' + chr(55) + chr(0b110111), 52648 - 52640), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b11011 + 0o30) + '\061', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(1650 - 1601) + chr(52) + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(2014 - 1966) + chr(1093 - 982) + '\x32' + chr(55) + '\067', 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\061' + chr(0b110 + 0o60), ord("\x08")), ehT0Px3KOsy9(chr(0b10 + 0o56) + '\157' + '\x35' + chr(0b101001 + 0o11), 23094 - 23086), ehT0Px3KOsy9('\060' + '\x6f' + chr(55) + chr(0b1011 + 0o51), 0o10), ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(0b1101111) + chr(0b110011) + chr(53) + chr(0b110101), 27566 - 27558), ehT0Px3KOsy9('\x30' + chr(111) + '\061' + chr(2084 - 2036) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(452 - 404) + chr(0b110010 + 0o75) + chr(53) + '\x33', 8), ehT0Px3KOsy9(chr(48) + chr(0b1100 + 0o143) + '\x31' + chr(0b110010) + '\067', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(4920 - 4809) + chr(0b110110) + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(11549 - 11438) + chr(50) + chr(0b110000) + chr(1444 - 1389), 58295 - 58287), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x33' + '\067' + chr(308 - 254), 0b1000), ehT0Px3KOsy9(chr(0b1001 + 0o47) + '\x6f' + chr(49) + chr(1014 - 966) + chr(48), 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(2088 - 2040) + chr(53), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(100 - 47) + '\060', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x9f'), chr(0b1100100) + '\x65' + '\x63' + '\157' + chr(5365 - 5265) + '\x65')(chr(117) + chr(0b1111 + 0o145) + chr(102) + chr(45) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def JC4EGLDtEviX(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
if xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc0wY\xe2\xa9\x10\x8c\x9c\xe7\xaa\xde\x9a'), chr(0b1100100) + chr(0b1100101) + chr(99) + '\157' + chr(0b1100100) + '\145')('\x75' + chr(116) + chr(0b10010 + 0o124) + chr(1907 - 1862) + chr(1342 - 1286))) or xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd6DL'), chr(100) + chr(1531 - 1430) + chr(99) + chr(3588 - 3477) + chr(1929 - 1829) + chr(0b100110 + 0o77))('\165' + '\164' + chr(0b1100110) + chr(0b100000 + 0o15) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xc2IY\xfd\xb4\x1d\xbe\xc9\xc6\xfa\xf2\xcf;FEF'), '\144' + '\x65' + chr(780 - 681) + chr(111) + chr(0b1100100) + chr(101))('\x75' + chr(0b1011110 + 0o26) + '\146' + '\055' + chr(56))):
try:
return hUdBpNM71xFW(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc2IY\xfd\xb4\x1d'), chr(2936 - 2836) + chr(0b1100101) + chr(99) + '\x6f' + chr(100) + '\145')(chr(0b1110101) + chr(116) + chr(5181 - 5079) + chr(1506 - 1461) + '\070'), reuse=ehT0Px3KOsy9(chr(2177 - 2129) + '\157' + chr(639 - 590), 0b1000))
except q1QCh3W88sgk:
return hUdBpNM71xFW(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc2IY\xfd\xb4\x1d'), chr(100) + chr(101) + '\x63' + '\x6f' + '\144' + chr(0b1011101 + 0o10))('\x75' + '\164' + chr(7191 - 7089) + '\x2d' + chr(0b110 + 0o62)), reuse=None)
else:
return hUdBpNM71xFW(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5@J\xe8\xb4\r\xbe\xc9\xc6\xfa'), '\144' + chr(6829 - 6728) + chr(0b1100011) + '\x6f' + chr(234 - 134) + '\145')(chr(0b1110101) + '\164' + chr(102) + chr(0b101101) + '\070'), reuse=None)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
video_bitwise_bottom
|
def video_bitwise_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for embedding video bitwise."""
pixel_embedding_size = 64
inputs = x
with tf.variable_scope("video_modality_bitwise", reuse=tf.AUTO_REUSE):
common_layers.summarize_video(inputs, "bottom")
# Embed bitwise.
assert vocab_size == 256
embedded = discretization.int_to_bit_embed(inputs, 8,
pixel_embedding_size)
# Project.
return tf.layers.dense(
embedded,
model_hparams.hidden_size,
name="merge_pixel_embedded_frames")
|
python
|
def video_bitwise_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for embedding video bitwise."""
pixel_embedding_size = 64
inputs = x
with tf.variable_scope("video_modality_bitwise", reuse=tf.AUTO_REUSE):
common_layers.summarize_video(inputs, "bottom")
# Embed bitwise.
assert vocab_size == 256
embedded = discretization.int_to_bit_embed(inputs, 8,
pixel_embedding_size)
# Project.
return tf.layers.dense(
embedded,
model_hparams.hidden_size,
name="merge_pixel_embedded_frames")
|
[
"def",
"video_bitwise_bottom",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"pixel_embedding_size",
"=",
"64",
"inputs",
"=",
"x",
"with",
"tf",
".",
"variable_scope",
"(",
"\"video_modality_bitwise\"",
",",
"reuse",
"=",
"tf",
".",
"AUTO_REUSE",
")",
":",
"common_layers",
".",
"summarize_video",
"(",
"inputs",
",",
"\"bottom\"",
")",
"# Embed bitwise.",
"assert",
"vocab_size",
"==",
"256",
"embedded",
"=",
"discretization",
".",
"int_to_bit_embed",
"(",
"inputs",
",",
"8",
",",
"pixel_embedding_size",
")",
"# Project.",
"return",
"tf",
".",
"layers",
".",
"dense",
"(",
"embedded",
",",
"model_hparams",
".",
"hidden_size",
",",
"name",
"=",
"\"merge_pixel_embedded_frames\"",
")"
] |
Bottom transformation for embedding video bitwise.
|
[
"Bottom",
"transformation",
"for",
"embedding",
"video",
"bitwise",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L552-L566
|
train
|
Bottom transformation for embedding video bitwise.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(9866 - 9755) + '\x31' + '\066' + chr(52), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b100 + 0o56) + chr(1218 - 1169) + chr(883 - 830), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(11591 - 11480) + chr(0b110011) + '\067' + chr(0b110001), 0b1000), ehT0Px3KOsy9('\060' + '\157' + '\x31' + '\066', 0o10), ehT0Px3KOsy9(chr(1544 - 1496) + chr(0b1101111) + '\x32' + chr(0b110000) + chr(51), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(72 - 22) + chr(0b110001) + '\x37', 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(84 - 29) + chr(0b10100 + 0o40), 0o10), ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(111) + chr(51) + '\063' + chr(0b11001 + 0o31), 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + '\061' + chr(55) + chr(51), ord("\x08")), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(111) + '\067', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b100111 + 0o110) + chr(0b110010) + chr(0b110001) + '\062', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(50) + '\x32' + chr(0b101111 + 0o1), 8995 - 8987), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(0b1101110 + 0o1) + chr(0b110001) + chr(0b1001 + 0o52) + chr(0b110000), 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\065' + chr(0b101110 + 0o2), ord("\x08")), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(0b1101111) + chr(51) + '\060', 0o10), ehT0Px3KOsy9(chr(488 - 440) + chr(0b1011110 + 0o21) + '\x37' + '\067', 38116 - 38108), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(111) + '\x34' + chr(0b110001 + 0o2), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(9648 - 9537) + '\x32' + chr(0b110101) + chr(1068 - 1016), 0o10), ehT0Px3KOsy9(chr(886 - 838) + '\157' + '\063' + chr(0b0 + 0o65) + chr(0b10111 + 0o33), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(0b1 + 0o62) + chr(0b110101) + chr(0b100010 + 0o23), 62429 - 62421), ehT0Px3KOsy9('\x30' + chr(9463 - 9352) + chr(2208 - 2154) + chr(54), 49107 - 49099), ehT0Px3KOsy9('\x30' + chr(111) + chr(85 - 34) + chr(48), 8), ehT0Px3KOsy9(chr(48) + chr(111) + '\x33' + chr(0b1110 + 0o47) + chr(0b10110 + 0o37), 8), ehT0Px3KOsy9(chr(48) + chr(0b10 + 0o155) + chr(0b110010) + '\062' + '\063', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110001) + chr(51) + chr(1599 - 1545), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b100011 + 0o22) + '\060', 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110111) + chr(0b110100), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b100000 + 0o117) + chr(52) + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b1100 + 0o44) + '\x6f' + chr(0b100100 + 0o17) + chr(0b10 + 0o57) + '\x30', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + '\063' + '\x34' + chr(2644 - 2590), 0o10), ehT0Px3KOsy9(chr(754 - 706) + chr(4988 - 4877) + chr(0b1111 + 0o43) + chr(1151 - 1097) + chr(1638 - 1586), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b100000 + 0o22) + chr(53) + chr(0b101011 + 0o11), 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110100) + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(7596 - 7485) + chr(49) + '\x35' + chr(2307 - 2253), ord("\x08")), ehT0Px3KOsy9(chr(0b100 + 0o54) + '\x6f' + chr(2042 - 1993) + '\062' + '\066', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1001110 + 0o41) + '\062' + chr(54), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1001100 + 0o43) + '\x36' + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110010) + chr(1642 - 1591) + '\x35', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(51) + '\060' + '\x36', 15087 - 15079), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x33' + chr(0b100110 + 0o12) + '\x32', 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(111) + '\065' + chr(0b11110 + 0o22), 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'Z'), chr(0b1100100) + chr(0b1100101) + '\143' + '\x6f' + '\x64' + chr(0b1100101))(chr(0b1110101) + chr(0b1000101 + 0o57) + chr(0b1011100 + 0o12) + chr(45) + chr(0b101000 + 0o20)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def fJYw_NaESM8R(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
jk6x7fS5a3ck = ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(111) + '\061' + chr(1463 - 1415) + chr(48), ord("\x08"))
vXoupepMtCXU = OeWW0F1dBPRQ
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x02\xec\x8bg\xe2\xedP\xbf$t\xe9\xff(M'), '\x64' + '\145' + '\x63' + '\x6f' + chr(0b1011 + 0o131) + '\145')(chr(0b1110101) + chr(0b1001001 + 0o53) + chr(102) + '\055' + chr(2958 - 2902)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x02\xe4\x9dk\xec\xd0Q\xb5\x1ff\xe6\xf9,QG]J\xd8\x9d\x84E}'), '\144' + chr(0b1100101) + '\143' + chr(157 - 46) + chr(100) + chr(101))(chr(117) + chr(116) + '\x66' + chr(1292 - 1247) + chr(0b111000)), reuse=xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xd8\xadA\xdc\xddy\x8f(B'), chr(100) + chr(0b110101 + 0o60) + chr(99) + chr(0b1101111) + '\x64' + chr(0b1010000 + 0o25))(chr(0b1110101) + '\164' + chr(7472 - 7370) + chr(0b100101 + 0o10) + chr(0b11101 + 0o33)))):
xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'\x07\xf8\x94c\xe2\xfdU\xa0\x1eX\xfc\xf9<Mw'), chr(100) + chr(0b1100101) + chr(0b1100011) + chr(0b1001010 + 0o45) + chr(0b11 + 0o141) + '\x65')(chr(1948 - 1831) + chr(116) + chr(0b1100110) + '\055' + chr(56)))(vXoupepMtCXU, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xe2\x8dz\xec\xe2'), '\144' + chr(101) + '\143' + chr(0b10000 + 0o137) + '\x64' + '\x65')('\165' + chr(116) + chr(102) + chr(0b101101) + chr(0b111000)))
assert CeyMIoSyrpkQ == ehT0Px3KOsy9('\x30' + chr(0b1101100 + 0o3) + chr(0b10111 + 0o35) + chr(0b100110 + 0o12) + chr(0b10111 + 0o31), 0b1000)
FaFRbyrGgzRu = mllfN6mU2TGo.int_to_bit_embed(vXoupepMtCXU, ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(10781 - 10670) + chr(0b10101 + 0o34) + chr(0b110000), 0b1000), jk6x7fS5a3ck)
return xafqLlk3kkUe(IDJ2eXGCBCDu.layers, xafqLlk3kkUe(SXOLrMavuUCe(b'\x10\xe8\x97}\xe6'), chr(0b1100100) + chr(2317 - 2216) + '\143' + chr(111) + chr(6823 - 6723) + '\x65')(chr(0b1110101) + chr(0b1110100) + chr(0b1001000 + 0o36) + '\x2d' + chr(0b11000 + 0o40)))(FaFRbyrGgzRu, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\x05\xf7\x96w\xdb\xc1\x0f\xb1\x1fo\xce\xdc'), chr(4556 - 4456) + chr(0b1100101) + chr(0b10 + 0o141) + chr(7619 - 7508) + chr(100) + chr(4733 - 4632))(chr(0b1110101) + chr(0b1110100) + chr(0b1011010 + 0o14) + chr(1983 - 1938) + chr(0b111000))), name=xafqLlk3kkUe(SXOLrMavuUCe(b'\x19\xe8\x8bi\xe6\xd0L\xb3\x03b\xe6\xcf=EzZG\xc8\x8f\x89i~\x9f\xf23MO'), '\x64' + chr(0b1100101) + chr(0b1100011) + '\157' + chr(0b1100100) + '\x65')(chr(12324 - 12207) + chr(116) + chr(8530 - 8428) + chr(0b11011 + 0o22) + chr(2420 - 2364)))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
video_pixel_noise_bottom
|
def video_pixel_noise_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for video."""
input_noise = getattr(model_hparams, "video_modality_input_noise", 0.25)
inputs = x
if model_hparams.mode == tf.estimator.ModeKeys.TRAIN:
background = tfp.stats.percentile(inputs, 50., axis=[0, 1, 2, 3])
input_shape = common_layers.shape_list(inputs)
input_size = tf.reduce_prod(input_shape[:-1])
input_mask = tf.multinomial(
tf.log([[input_noise, 1.-input_noise]]), input_size)
input_mask = tf.reshape(tf.cast(input_mask, tf.int32),
input_shape[:-1]+[1])
inputs = inputs * input_mask + background * (1 - input_mask)
return video_bottom(inputs, model_hparams, vocab_size)
|
python
|
def video_pixel_noise_bottom(x, model_hparams, vocab_size):
"""Bottom transformation for video."""
input_noise = getattr(model_hparams, "video_modality_input_noise", 0.25)
inputs = x
if model_hparams.mode == tf.estimator.ModeKeys.TRAIN:
background = tfp.stats.percentile(inputs, 50., axis=[0, 1, 2, 3])
input_shape = common_layers.shape_list(inputs)
input_size = tf.reduce_prod(input_shape[:-1])
input_mask = tf.multinomial(
tf.log([[input_noise, 1.-input_noise]]), input_size)
input_mask = tf.reshape(tf.cast(input_mask, tf.int32),
input_shape[:-1]+[1])
inputs = inputs * input_mask + background * (1 - input_mask)
return video_bottom(inputs, model_hparams, vocab_size)
|
[
"def",
"video_pixel_noise_bottom",
"(",
"x",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"input_noise",
"=",
"getattr",
"(",
"model_hparams",
",",
"\"video_modality_input_noise\"",
",",
"0.25",
")",
"inputs",
"=",
"x",
"if",
"model_hparams",
".",
"mode",
"==",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"TRAIN",
":",
"background",
"=",
"tfp",
".",
"stats",
".",
"percentile",
"(",
"inputs",
",",
"50.",
",",
"axis",
"=",
"[",
"0",
",",
"1",
",",
"2",
",",
"3",
"]",
")",
"input_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"inputs",
")",
"input_size",
"=",
"tf",
".",
"reduce_prod",
"(",
"input_shape",
"[",
":",
"-",
"1",
"]",
")",
"input_mask",
"=",
"tf",
".",
"multinomial",
"(",
"tf",
".",
"log",
"(",
"[",
"[",
"input_noise",
",",
"1.",
"-",
"input_noise",
"]",
"]",
")",
",",
"input_size",
")",
"input_mask",
"=",
"tf",
".",
"reshape",
"(",
"tf",
".",
"cast",
"(",
"input_mask",
",",
"tf",
".",
"int32",
")",
",",
"input_shape",
"[",
":",
"-",
"1",
"]",
"+",
"[",
"1",
"]",
")",
"inputs",
"=",
"inputs",
"*",
"input_mask",
"+",
"background",
"*",
"(",
"1",
"-",
"input_mask",
")",
"return",
"video_bottom",
"(",
"inputs",
",",
"model_hparams",
",",
"vocab_size",
")"
] |
Bottom transformation for video.
|
[
"Bottom",
"transformation",
"for",
"video",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L599-L612
|
train
|
Bottom transformation for video.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(1945 - 1897) + chr(111) + chr(50) + chr(0b110101) + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110110) + '\060', 0b1000), ehT0Px3KOsy9('\x30' + chr(3978 - 3867) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(3271 - 3160) + chr(0b101011 + 0o6) + chr(48), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(2369 - 2319) + chr(977 - 924) + chr(0b10111 + 0o31), ord("\x08")), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(111) + chr(51) + chr(2400 - 2350), 0b1000), ehT0Px3KOsy9(chr(614 - 566) + '\x6f' + chr(0b1000 + 0o51) + chr(0b101100 + 0o4) + '\063', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + chr(0b110000) + '\067', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1756 - 1707) + chr(50), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b1000 + 0o51) + chr(0b110011 + 0o2), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + chr(846 - 798) + chr(51), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33' + chr(0b110001) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001) + chr(53) + chr(2524 - 2473), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(49) + chr(52) + chr(0b110011), 4532 - 4524), ehT0Px3KOsy9(chr(48) + '\157' + chr(89 - 37) + chr(0b10001 + 0o41), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + '\060' + chr(0b110010), 6982 - 6974), ehT0Px3KOsy9(chr(0b11100 + 0o24) + chr(0b1101111) + chr(50) + '\x37' + chr(768 - 715), ord("\x08")), ehT0Px3KOsy9(chr(1124 - 1076) + '\x6f' + '\x31' + chr(0b10 + 0o56) + '\x32', 8), ehT0Px3KOsy9('\060' + chr(7981 - 7870) + chr(1986 - 1931), 16095 - 16087), ehT0Px3KOsy9(chr(48) + '\157' + chr(2275 - 2224) + '\x36', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1001 + 0o146) + chr(1935 - 1885) + '\065' + chr(0b101011 + 0o14), 8), ehT0Px3KOsy9('\x30' + chr(10527 - 10416) + '\x32' + '\x35' + '\065', ord("\x08")), ehT0Px3KOsy9(chr(0b10010 + 0o36) + chr(111) + chr(545 - 494) + '\064' + '\x35', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001) + chr(52) + '\067', ord("\x08")), ehT0Px3KOsy9(chr(0b10000 + 0o40) + '\x6f' + chr(51) + chr(55) + chr(0b100 + 0o54), 60195 - 60187), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\061' + chr(0b110110) + chr(0b10100 + 0o40), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\x33' + '\x31' + chr(0b11000 + 0o36), 0b1000), ehT0Px3KOsy9(chr(402 - 354) + '\157' + '\x32' + chr(0b110001) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(2016 - 1968) + '\x6f' + '\x32' + '\067' + '\062', 0o10), ehT0Px3KOsy9(chr(0b1010 + 0o46) + chr(5836 - 5725) + chr(0b110001) + chr(0b110001 + 0o6) + '\x31', 42417 - 42409), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\157' + '\061' + chr(2026 - 1974) + chr(0b10001 + 0o37), 0b1000), ehT0Px3KOsy9(chr(761 - 713) + '\157' + '\x36', 2284 - 2276), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10111 + 0o34) + chr(1385 - 1337) + chr(558 - 505), 0o10), ehT0Px3KOsy9(chr(167 - 119) + chr(0b1101111) + '\x31' + chr(1443 - 1392) + '\064', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(2162 - 2112) + chr(52) + '\067', 0o10), ehT0Px3KOsy9(chr(0b101000 + 0o10) + '\157' + '\x31' + chr(2629 - 2574) + chr(0b110000), 46048 - 46040), ehT0Px3KOsy9(chr(48) + chr(1776 - 1665) + '\x32' + '\064' + '\063', 0b1000), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(0b1101111) + '\x32' + chr(0b100010 + 0o22) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(1950 - 1899) + chr(0b11001 + 0o35) + chr(0b11111 + 0o24), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(49) + '\x34' + chr(0b110101), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b111 + 0o56) + chr(0b101100 + 0o4), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x81'), chr(0b1100100) + chr(101) + chr(1891 - 1792) + chr(3542 - 3431) + chr(100) + '\x65')(chr(0b1110101) + chr(0b111110 + 0o66) + chr(0b11101 + 0o111) + chr(677 - 632) + chr(0b10101 + 0o43)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def mxRCIiyni4wa(OeWW0F1dBPRQ, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
DnfcEvYU1Z8O = xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd9Yfm\xc7E.\xa8nl\xaf\xa5\x1f\x1a}+\xd32rj\xf0\xc3\x8a\x0e\x8b\x11'), '\144' + '\x65' + chr(0b100101 + 0o76) + chr(0b1100010 + 0o15) + '\x64' + '\145')(chr(0b101 + 0o160) + chr(0b101111 + 0o105) + chr(0b101111 + 0o67) + chr(45) + chr(56)), 0.25)
vXoupepMtCXU = OeWW0F1dBPRQ
if xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc2_fm'), chr(5547 - 5447) + chr(0b1100101) + chr(0b1100011) + '\157' + chr(0b111101 + 0o47) + '\145')('\x75' + '\164' + chr(0b100011 + 0o103) + chr(0b100110 + 0o7) + '\070')) == xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfbbCA\xe6'), chr(0b1100100) + '\145' + chr(0b10111 + 0o114) + chr(0b1101111) + chr(0b100001 + 0o103) + chr(1736 - 1635))(chr(0b111111 + 0o66) + chr(0b1110100) + chr(102) + chr(45) + '\070')):
aWgHYEoppQVf = Ys555qziAbad.stats.percentile(vXoupepMtCXU, 50.0, axis=[ehT0Px3KOsy9('\060' + chr(111) + '\x30', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + chr(420 - 309) + '\062', 8), ehT0Px3KOsy9(chr(1325 - 1277) + chr(0b1000011 + 0o54) + chr(593 - 542), 0o10)])
tANyZeuTfu5y = jSKPaHwSAfVv.shape_list(vXoupepMtCXU)
uNAGRcXaNiGu = IDJ2eXGCBCDu.reduce_prod(tANyZeuTfu5y[:-ehT0Px3KOsy9(chr(1166 - 1118) + '\x6f' + chr(0b10100 + 0o35), 8)])
kA61TR8pjraF = IDJ2eXGCBCDu.multinomial(IDJ2eXGCBCDu.log([[DnfcEvYU1Z8O, 1.0 - DnfcEvYU1Z8O]]), uNAGRcXaNiGu)
kA61TR8pjraF = IDJ2eXGCBCDu.reshape(IDJ2eXGCBCDu.cast(kA61TR8pjraF, IDJ2eXGCBCDu.int32), tANyZeuTfu5y[:-ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110001), 8)] + [ehT0Px3KOsy9('\060' + chr(6891 - 6780) + chr(49), 8)])
vXoupepMtCXU = vXoupepMtCXU * kA61TR8pjraF + aWgHYEoppQVf * (ehT0Px3KOsy9(chr(0b110000) + chr(4691 - 4580) + '\061', 8) - kA61TR8pjraF)
return UTQU1__m4eux(vXoupepMtCXU, tq24Tk6UZ6u1, CeyMIoSyrpkQ)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
convert_rgb_to_real
|
def convert_rgb_to_real(prediction, targets):
"""Convert prediction and target from rgb to real."""
prediction = tf.squeeze(prediction, axis=-1)
prediction = common_layers.convert_rgb_to_real(prediction)
targets = common_layers.convert_rgb_to_real(targets)
return prediction, targets
|
python
|
def convert_rgb_to_real(prediction, targets):
"""Convert prediction and target from rgb to real."""
prediction = tf.squeeze(prediction, axis=-1)
prediction = common_layers.convert_rgb_to_real(prediction)
targets = common_layers.convert_rgb_to_real(targets)
return prediction, targets
|
[
"def",
"convert_rgb_to_real",
"(",
"prediction",
",",
"targets",
")",
":",
"prediction",
"=",
"tf",
".",
"squeeze",
"(",
"prediction",
",",
"axis",
"=",
"-",
"1",
")",
"prediction",
"=",
"common_layers",
".",
"convert_rgb_to_real",
"(",
"prediction",
")",
"targets",
"=",
"common_layers",
".",
"convert_rgb_to_real",
"(",
"targets",
")",
"return",
"prediction",
",",
"targets"
] |
Convert prediction and target from rgb to real.
|
[
"Convert",
"prediction",
"and",
"target",
"from",
"rgb",
"to",
"real",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L615-L620
|
train
|
Convert prediction and target from rgb to real.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x33' + chr(1774 - 1723), 14837 - 14829), ehT0Px3KOsy9(chr(0b110 + 0o52) + '\157' + '\x33' + chr(0b10001 + 0o43) + chr(0b11011 + 0o33), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x32' + chr(1701 - 1648) + '\x35', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b10 + 0o57) + '\x35' + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(1283 - 1234) + '\062', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(1605 - 1494) + '\062' + '\067' + chr(0b110100 + 0o3), 37409 - 37401), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1768 - 1718) + chr(2234 - 2183) + chr(1366 - 1316), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b110001 + 0o76) + chr(0b1100 + 0o47) + chr(0b110000) + chr(0b110000), 15353 - 15345), ehT0Px3KOsy9(chr(48) + chr(111) + '\065' + '\062', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(2115 - 2066) + chr(51) + chr(0b10000 + 0o43), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(5309 - 5198) + '\063' + chr(50), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110001) + chr(1520 - 1471) + '\x36', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(50) + '\066' + chr(1389 - 1337), 0o10), ehT0Px3KOsy9(chr(0b1110 + 0o42) + chr(111) + chr(1391 - 1342) + chr(0b110111) + chr(49), 46415 - 46407), ehT0Px3KOsy9(chr(0b110000) + chr(4872 - 4761) + chr(49) + '\061' + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(1320 - 1272) + '\157' + chr(55), 58617 - 58609), ehT0Px3KOsy9(chr(0b0 + 0o60) + chr(111) + '\x33' + chr(1296 - 1241) + '\x35', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b101001 + 0o10) + '\x31' + chr(0b100000 + 0o26), 8), ehT0Px3KOsy9(chr(1543 - 1495) + chr(0b111001 + 0o66) + '\062' + '\065' + '\x31', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1 + 0o156) + chr(1045 - 995) + chr(0b110111), 0o10), ehT0Px3KOsy9('\060' + chr(0b1000101 + 0o52) + '\061' + chr(1699 - 1647) + chr(50), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101110 + 0o1) + chr(1978 - 1927) + chr(48) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(111) + chr(49) + chr(52) + '\066', 0o10), ehT0Px3KOsy9(chr(169 - 121) + chr(111) + '\063' + '\064' + chr(2512 - 2460), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010 + 0o0) + chr(0b110111) + chr(0b11011 + 0o32), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33' + chr(0b10101 + 0o42) + '\063', 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(49) + '\065' + '\x31', 45817 - 45809), ehT0Px3KOsy9(chr(0b11101 + 0o23) + chr(0b100101 + 0o112) + '\061' + '\064' + chr(0b101000 + 0o11), 28232 - 28224), ehT0Px3KOsy9(chr(2237 - 2189) + chr(2475 - 2364) + chr(930 - 880) + chr(0b100100 + 0o17) + '\x31', 39102 - 39094), ehT0Px3KOsy9(chr(48) + chr(8689 - 8578) + chr(50) + chr(50) + chr(54), 33195 - 33187), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1420 - 1366) + '\x32', 52022 - 52014), ehT0Px3KOsy9(chr(1730 - 1682) + chr(0b1101111) + '\061' + '\x37' + '\065', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(279 - 228) + '\063' + chr(51), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b11 + 0o56) + chr(0b10110 + 0o34) + chr(0b11101 + 0o30), 55917 - 55909), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b11110 + 0o25) + chr(55), 53764 - 53756), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110011) + chr(0b110101) + chr(50), 0o10), ehT0Px3KOsy9(chr(0b11100 + 0o24) + '\x6f' + chr(2222 - 2171) + '\064' + chr(963 - 915), ord("\x08")), ehT0Px3KOsy9(chr(396 - 348) + chr(1847 - 1736) + chr(53) + chr(0b1000 + 0o53), 45199 - 45191), ehT0Px3KOsy9(chr(2298 - 2250) + '\x6f' + chr(1210 - 1161) + chr(536 - 482) + chr(0b110111), 41255 - 41247), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110011) + chr(0b110110) + '\x30', 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + '\157' + chr(0b110101) + chr(48), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'5'), chr(0b1100100) + chr(5108 - 5007) + chr(8055 - 7956) + '\157' + '\x64' + chr(1536 - 1435))(chr(0b1001000 + 0o55) + chr(116) + chr(0b1100110) + chr(0b1101 + 0o40) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def vPvjvy22ItpS(ys6Y0jo7ObhM, xIEmRseySp3z):
ys6Y0jo7ObhM = IDJ2eXGCBCDu.squeeze(ys6Y0jo7ObhM, axis=-ehT0Px3KOsy9(chr(0b110000) + chr(0b111111 + 0o60) + chr(0b110001), 23451 - 23443))
ys6Y0jo7ObhM = jSKPaHwSAfVv.convert_rgb_to_real(ys6Y0jo7ObhM)
xIEmRseySp3z = jSKPaHwSAfVv.convert_rgb_to_real(xIEmRseySp3z)
return (ys6Y0jo7ObhM, xIEmRseySp3z)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
ctc_symbol_loss
|
def ctc_symbol_loss(top_out, targets, model_hparams, vocab_size, weight_fn):
"""Compute the CTC loss."""
del model_hparams, vocab_size # unused arg
logits = top_out
with tf.name_scope("ctc_loss", values=[logits, targets]):
# For CTC we assume targets are 1d, [batch, length, 1, 1] here.
targets_shape = targets.get_shape().as_list()
assert len(targets_shape) == 4
assert targets_shape[2] == 1
assert targets_shape[3] == 1
targets = tf.squeeze(targets, axis=[2, 3])
logits = tf.squeeze(logits, axis=[2, 3])
targets_mask = 1 - tf.to_int32(tf.equal(targets, 0))
targets_lengths = tf.reduce_sum(targets_mask, axis=1)
sparse_targets = tf.keras.backend.ctc_label_dense_to_sparse(
targets, targets_lengths)
xent = tf.nn.ctc_loss(
sparse_targets,
logits,
targets_lengths,
time_major=False,
preprocess_collapse_repeated=False,
ctc_merge_repeated=False)
weights = weight_fn(targets)
return tf.reduce_sum(xent), tf.reduce_sum(weights)
|
python
|
def ctc_symbol_loss(top_out, targets, model_hparams, vocab_size, weight_fn):
"""Compute the CTC loss."""
del model_hparams, vocab_size # unused arg
logits = top_out
with tf.name_scope("ctc_loss", values=[logits, targets]):
# For CTC we assume targets are 1d, [batch, length, 1, 1] here.
targets_shape = targets.get_shape().as_list()
assert len(targets_shape) == 4
assert targets_shape[2] == 1
assert targets_shape[3] == 1
targets = tf.squeeze(targets, axis=[2, 3])
logits = tf.squeeze(logits, axis=[2, 3])
targets_mask = 1 - tf.to_int32(tf.equal(targets, 0))
targets_lengths = tf.reduce_sum(targets_mask, axis=1)
sparse_targets = tf.keras.backend.ctc_label_dense_to_sparse(
targets, targets_lengths)
xent = tf.nn.ctc_loss(
sparse_targets,
logits,
targets_lengths,
time_major=False,
preprocess_collapse_repeated=False,
ctc_merge_repeated=False)
weights = weight_fn(targets)
return tf.reduce_sum(xent), tf.reduce_sum(weights)
|
[
"def",
"ctc_symbol_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weight_fn",
")",
":",
"del",
"model_hparams",
",",
"vocab_size",
"# unused arg",
"logits",
"=",
"top_out",
"with",
"tf",
".",
"name_scope",
"(",
"\"ctc_loss\"",
",",
"values",
"=",
"[",
"logits",
",",
"targets",
"]",
")",
":",
"# For CTC we assume targets are 1d, [batch, length, 1, 1] here.",
"targets_shape",
"=",
"targets",
".",
"get_shape",
"(",
")",
".",
"as_list",
"(",
")",
"assert",
"len",
"(",
"targets_shape",
")",
"==",
"4",
"assert",
"targets_shape",
"[",
"2",
"]",
"==",
"1",
"assert",
"targets_shape",
"[",
"3",
"]",
"==",
"1",
"targets",
"=",
"tf",
".",
"squeeze",
"(",
"targets",
",",
"axis",
"=",
"[",
"2",
",",
"3",
"]",
")",
"logits",
"=",
"tf",
".",
"squeeze",
"(",
"logits",
",",
"axis",
"=",
"[",
"2",
",",
"3",
"]",
")",
"targets_mask",
"=",
"1",
"-",
"tf",
".",
"to_int32",
"(",
"tf",
".",
"equal",
"(",
"targets",
",",
"0",
")",
")",
"targets_lengths",
"=",
"tf",
".",
"reduce_sum",
"(",
"targets_mask",
",",
"axis",
"=",
"1",
")",
"sparse_targets",
"=",
"tf",
".",
"keras",
".",
"backend",
".",
"ctc_label_dense_to_sparse",
"(",
"targets",
",",
"targets_lengths",
")",
"xent",
"=",
"tf",
".",
"nn",
".",
"ctc_loss",
"(",
"sparse_targets",
",",
"logits",
",",
"targets_lengths",
",",
"time_major",
"=",
"False",
",",
"preprocess_collapse_repeated",
"=",
"False",
",",
"ctc_merge_repeated",
"=",
"False",
")",
"weights",
"=",
"weight_fn",
"(",
"targets",
")",
"return",
"tf",
".",
"reduce_sum",
"(",
"xent",
")",
",",
"tf",
".",
"reduce_sum",
"(",
"weights",
")"
] |
Compute the CTC loss.
|
[
"Compute",
"the",
"CTC",
"loss",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L638-L662
|
train
|
Compute the CTC loss.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(804 - 756) + chr(111) + chr(51) + chr(876 - 828), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1000111 + 0o50) + chr(51) + '\061', 29445 - 29437), ehT0Px3KOsy9('\060' + chr(2499 - 2388) + '\062' + chr(100 - 48) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b110111 + 0o70) + chr(2523 - 2471) + '\063', 45007 - 44999), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110011) + chr(0b110011) + chr(49), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110110) + '\065', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b101111 + 0o2) + '\066' + chr(0b101110 + 0o6), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b11111 + 0o120) + chr(51) + chr(307 - 255) + chr(0b110101), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b1110 + 0o43) + '\x35' + '\061', 0o10), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(111) + chr(0b11011 + 0o30) + '\061', 8), ehT0Px3KOsy9(chr(850 - 802) + chr(0b1101111) + chr(0b110010) + '\060' + '\x36', 0o10), ehT0Px3KOsy9('\060' + chr(0b1000010 + 0o55) + '\062' + '\066' + chr(51), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b11110 + 0o121) + '\x32' + '\x36' + chr(54), 0b1000), ehT0Px3KOsy9(chr(1227 - 1179) + chr(10051 - 9940) + '\x31' + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b10001 + 0o42) + '\060' + '\061', 27490 - 27482), ehT0Px3KOsy9('\x30' + '\x6f' + chr(50) + chr(0b110101) + chr(0b110100), 60189 - 60181), ehT0Px3KOsy9(chr(48) + chr(8084 - 7973) + '\x33' + chr(443 - 395), 8), ehT0Px3KOsy9(chr(286 - 238) + chr(111) + chr(0b10100 + 0o36) + '\x33' + '\x30', 0b1000), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(9787 - 9676) + '\x33' + chr(0b110001) + chr(53), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b10 + 0o155) + '\063' + chr(50) + '\064', 26471 - 26463), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(365 - 316) + chr(54) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(6680 - 6569) + chr(55) + chr(1962 - 1913), 65393 - 65385), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b101111 + 0o3) + '\065' + '\061', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(51) + '\x34' + chr(0b110111), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(0b10001 + 0o40) + '\x34' + chr(1717 - 1667), ord("\x08")), ehT0Px3KOsy9('\060' + chr(3308 - 3197) + chr(50) + chr(0b110000) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(111) + chr(1240 - 1191) + '\065' + chr(2379 - 2329), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001) + chr(0b110111) + chr(2317 - 2266), 0b1000), ehT0Px3KOsy9(chr(322 - 274) + chr(0b111010 + 0o65) + '\x36' + '\x33', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1100010 + 0o15) + '\061' + '\x34', 0b1000), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\157' + '\x33' + chr(0b100000 + 0o26) + chr(48), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(51) + '\062' + chr(0b110011), 17517 - 17509), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(49) + '\x30' + chr(1479 - 1428), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b100110 + 0o13) + chr(0b100001 + 0o22) + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(790 - 742) + chr(111) + chr(0b11011 + 0o30) + chr(49) + chr(52), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(1994 - 1942) + chr(0b10011 + 0o40), 0b1000), ehT0Px3KOsy9(chr(48) + chr(11347 - 11236) + chr(2119 - 2068) + '\063' + chr(1501 - 1449), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + chr(1251 - 1199) + '\061', 0o10), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b1101111) + chr(51) + chr(0b1001 + 0o54) + chr(0b10100 + 0o41), 0b1000), ehT0Px3KOsy9('\060' + chr(4165 - 4054) + chr(51) + chr(52) + chr(0b110010), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(111) + chr(53) + chr(48), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'6'), '\144' + '\145' + chr(2340 - 2241) + '\x6f' + chr(100) + chr(0b1000000 + 0o45))(chr(117) + chr(116) + chr(102) + chr(45) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def QyBSDyDRERUx(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, TtirqPESlfJn):
del tq24Tk6UZ6u1, CeyMIoSyrpkQ
wF9nmvjsKjYM = juqulShQpQzQ
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'vx\xcdF\x86F\x17\x8a\x19|'), '\x64' + '\x65' + chr(0b10010 + 0o121) + chr(0b1101111) + chr(0b11101 + 0o107) + '\145')(chr(117) + chr(0b1110100) + chr(0b1100110) + chr(0b10101 + 0o30) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'{m\xc3|\xb5Z\x07\x96'), chr(0b1100100) + chr(1188 - 1087) + chr(4697 - 4598) + chr(111) + chr(0b1100100) + chr(0b1100101))(chr(0b1101101 + 0o10) + chr(4344 - 4228) + '\x66' + chr(45) + chr(0b110000 + 0o10)), values=[wF9nmvjsKjYM, xIEmRseySp3z]):
qGCVeFvxIRjf = xIEmRseySp3z.get_shape().as_list()
assert c2A0yzQpDQB3(qGCVeFvxIRjf) == ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110 + 0o56), 17106 - 17098)
assert qGCVeFvxIRjf[ehT0Px3KOsy9(chr(130 - 82) + '\157' + '\x32', 0b1000)] == ehT0Px3KOsy9(chr(577 - 529) + '\x6f' + chr(0b1110 + 0o43), 0o10)
assert qGCVeFvxIRjf[ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(645 - 594), ord("\x08"))] == ehT0Px3KOsy9(chr(1264 - 1216) + chr(111) + chr(49), 8)
xIEmRseySp3z = IDJ2eXGCBCDu.squeeze(xIEmRseySp3z, axis=[ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x32', 8), ehT0Px3KOsy9(chr(0b0 + 0o60) + chr(0b1101111) + chr(0b1 + 0o62), 8)])
wF9nmvjsKjYM = IDJ2eXGCBCDu.squeeze(wF9nmvjsKjYM, axis=[ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x32', 8), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b101001 + 0o12), 8)])
BHw9bGdZny4p = ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(0b1101111) + '\x31', 8) - IDJ2eXGCBCDu.to_int32(IDJ2eXGCBCDu.equal(xIEmRseySp3z, ehT0Px3KOsy9(chr(0b110000) + chr(0b1101011 + 0o4) + chr(776 - 728), 0o10)))
NKpmzeuCPl9M = IDJ2eXGCBCDu.reduce_sum(BHw9bGdZny4p, axis=ehT0Px3KOsy9(chr(894 - 846) + '\157' + chr(0b110001), 8))
aDTXZyvZjQhA = IDJ2eXGCBCDu.keras.backend.ctc_label_dense_to_sparse(xIEmRseySp3z, NKpmzeuCPl9M)
_YHpmhjj_eGR = IDJ2eXGCBCDu.nn.ctc_loss(aDTXZyvZjQhA, wF9nmvjsKjYM, NKpmzeuCPl9M, time_major=ehT0Px3KOsy9(chr(1969 - 1921) + chr(111) + chr(0b11101 + 0o23), 8), preprocess_collapse_repeated=ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b100000 + 0o20), 8), ctc_merge_repeated=ehT0Px3KOsy9(chr(1359 - 1311) + chr(10410 - 10299) + chr(2203 - 2155), 8))
ZurHTci57aXw = TtirqPESlfJn(xIEmRseySp3z)
return (xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'j|\xc4V\xbaP+\x96\x1ct'), chr(100) + '\x65' + chr(6051 - 5952) + chr(111) + chr(0b1001101 + 0o27) + chr(101))(chr(117) + chr(797 - 681) + '\146' + chr(0b101101) + chr(2812 - 2756)))(_YHpmhjj_eGR), xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'j|\xc4V\xbaP+\x96\x1ct'), chr(4191 - 4091) + chr(0b1001 + 0o134) + chr(0b1011 + 0o130) + chr(111) + chr(330 - 230) + chr(0b1010001 + 0o24))(chr(0b11001 + 0o134) + chr(0b1110100) + chr(6782 - 6680) + '\x2d' + chr(0b1110 + 0o52)))(ZurHTci57aXw))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
generic_loss
|
def generic_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Compute loss numerator and denominator for one shard of output."""
del vocab_size # unused arg
logits = top_out
logits = common_attention.maybe_upcast(logits, hparams=model_hparams)
cutoff = getattr(model_hparams, "video_modality_loss_cutoff", 0.0)
return common_layers.padded_cross_entropy(
logits,
targets,
model_hparams.label_smoothing,
cutoff=cutoff,
weights_fn=weights_fn)
|
python
|
def generic_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Compute loss numerator and denominator for one shard of output."""
del vocab_size # unused arg
logits = top_out
logits = common_attention.maybe_upcast(logits, hparams=model_hparams)
cutoff = getattr(model_hparams, "video_modality_loss_cutoff", 0.0)
return common_layers.padded_cross_entropy(
logits,
targets,
model_hparams.label_smoothing,
cutoff=cutoff,
weights_fn=weights_fn)
|
[
"def",
"generic_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weights_fn",
")",
":",
"del",
"vocab_size",
"# unused arg",
"logits",
"=",
"top_out",
"logits",
"=",
"common_attention",
".",
"maybe_upcast",
"(",
"logits",
",",
"hparams",
"=",
"model_hparams",
")",
"cutoff",
"=",
"getattr",
"(",
"model_hparams",
",",
"\"video_modality_loss_cutoff\"",
",",
"0.0",
")",
"return",
"common_layers",
".",
"padded_cross_entropy",
"(",
"logits",
",",
"targets",
",",
"model_hparams",
".",
"label_smoothing",
",",
"cutoff",
"=",
"cutoff",
",",
"weights_fn",
"=",
"weights_fn",
")"
] |
Compute loss numerator and denominator for one shard of output.
|
[
"Compute",
"loss",
"numerator",
"and",
"denominator",
"for",
"one",
"shard",
"of",
"output",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L665-L676
|
train
|
Compute loss numerator and denominator for one shard of output.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(1374 - 1326) + '\x6f' + chr(0b11111 + 0o24) + '\065', 38170 - 38162), ehT0Px3KOsy9('\x30' + chr(0b100111 + 0o110) + '\061' + chr(0b110111) + chr(0b100 + 0o60), 38588 - 38580), ehT0Px3KOsy9('\060' + '\157' + '\x33' + chr(0b11011 + 0o26) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(2257 - 2209) + chr(11162 - 11051) + chr(50) + chr(1067 - 1012) + chr(1953 - 1904), 45554 - 45546), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x32' + chr(642 - 594), 0o10), ehT0Px3KOsy9(chr(48) + chr(5082 - 4971) + '\x36' + '\x33', 60377 - 60369), ehT0Px3KOsy9(chr(2134 - 2086) + chr(8759 - 8648) + chr(51) + chr(53) + '\x31', 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1028 - 978) + '\067' + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(50) + chr(0b110101) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\063' + chr(0b1001 + 0o51) + '\060', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1100 + 0o143) + chr(0b110010) + chr(452 - 399) + chr(0b110011), 8), ehT0Px3KOsy9(chr(48) + chr(8866 - 8755) + chr(286 - 236) + '\x32' + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + chr(8368 - 8257) + chr(52) + '\067', ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b1 + 0o60) + chr(1506 - 1458) + chr(0b10010 + 0o43), 0b1000), ehT0Px3KOsy9(chr(2016 - 1968) + '\157' + chr(1705 - 1652) + chr(53), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(0b1001 + 0o55) + chr(50), 25547 - 25539), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49) + chr(48) + '\x33', 0b1000), ehT0Px3KOsy9(chr(1942 - 1894) + '\x6f' + chr(0b110010) + chr(0b110010) + chr(0b110000), 5416 - 5408), ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(6399 - 6288) + chr(0b110011) + chr(0b11101 + 0o27) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(0b11101 + 0o23) + chr(0b10001 + 0o136) + chr(0b110001) + '\x36' + chr(1318 - 1267), ord("\x08")), ehT0Px3KOsy9(chr(0b11010 + 0o26) + '\157' + chr(0b100101 + 0o15) + chr(52) + chr(0b10011 + 0o44), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b111111 + 0o60) + chr(0b11000 + 0o33) + '\065' + chr(49), 8), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33' + '\064' + chr(492 - 443), 8), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + chr(0b110010) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(315 - 204) + '\061' + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b101100 + 0o10) + chr(55), 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1423 - 1374) + chr(51) + '\x32', 11365 - 11357), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(111) + chr(0b1110 + 0o44) + chr(2283 - 2230) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(5526 - 5415) + chr(55) + '\065', 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\065', ord("\x08")), ehT0Px3KOsy9(chr(898 - 850) + chr(0b110111 + 0o70) + chr(0b11001 + 0o34) + chr(51), 0b1000), ehT0Px3KOsy9(chr(1259 - 1211) + chr(0b1001 + 0o146) + chr(0b111 + 0o56) + chr(49), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b111 + 0o53) + chr(1358 - 1309) + chr(1286 - 1238), 0b1000), ehT0Px3KOsy9(chr(0b10101 + 0o33) + '\x6f' + '\x33' + chr(0b101000 + 0o10), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110001 + 0o0) + '\065' + chr(676 - 625), 19621 - 19613), ehT0Px3KOsy9('\x30' + chr(0b10011 + 0o134) + '\x32' + chr(0b110110) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + '\x31' + '\066' + chr(2452 - 2399), 6831 - 6823), ehT0Px3KOsy9('\060' + chr(111) + chr(0b101001 + 0o11) + chr(0b110000) + chr(0b1001 + 0o55), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + chr(49) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(929 - 881) + chr(111) + chr(0b100110 + 0o20), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b11111 + 0o120) + chr(0b100110 + 0o17) + chr(1932 - 1884), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'3'), chr(495 - 395) + chr(0b1100101) + chr(99) + '\157' + chr(100) + chr(5347 - 5246))('\165' + chr(0b11110 + 0o126) + chr(8962 - 8860) + '\055' + chr(271 - 215)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def satn1Z5kItGQ(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, Pdbc6Q2jZ4RQ):
del CeyMIoSyrpkQ
wF9nmvjsKjYM = juqulShQpQzQ
wF9nmvjsKjYM = WOnrfm4dlYcf.maybe_upcast(wF9nmvjsKjYM, hparams=tq24Tk6UZ6u1)
EjnQGacZaia3 = xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'k\x15\xad\xdc\x7fl\x84\xd1\xcf\xb1\xc7\xf8S<r\x9e,\xe3\x92,\xc4\x9c\x95\xbehA'), '\x64' + '\145' + '\x63' + chr(111) + '\144' + chr(101))('\x75' + chr(0b1 + 0o163) + chr(1345 - 1243) + '\055' + chr(0b111000)), 0.0)
return xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'm\x1d\xad\xdduW\xb6\xdd\xd9\xbf\xd8\xe2x C\x861\xff\x91\n'), '\144' + '\x65' + chr(0b1100011) + '\157' + '\x64' + '\x65')('\x75' + '\164' + chr(2872 - 2770) + '\x2d' + chr(0b111000)))(wF9nmvjsKjYM, xIEmRseySp3z, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'[/\xa3\xecwW\x88\xdd\xd1\xaa\xf9\xfa'), chr(0b1100100) + '\145' + chr(0b1 + 0o142) + chr(603 - 492) + chr(0b1100100) + chr(0b1100101))(chr(0b1001 + 0o154) + chr(8824 - 8708) + chr(0b1011011 + 0o13) + chr(45) + chr(56))), cutoff=EjnQGacZaia3, weights_fn=Pdbc6Q2jZ4RQ)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
multi_label_loss
|
def multi_label_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Average loss over the labels."""
del vocab_size # unused arg
logits = top_out
num_labels = tf.shape(targets)[1]
logits = tf.tile(logits, [1, num_labels, 1, 1, 1])
xent, weights = common_layers.padded_cross_entropy(
logits,
targets,
model_hparams.label_smoothing,
weights_fn=weights_fn,
reduce_sum=False,
)
xent = tf.squeeze(xent, [2, 3])
weights = tf.squeeze(weights, [2, 3])
# average loss over all labels
loss = tf.reduce_sum(xent, axis=1)
weights = tf.reduce_sum(weights, axis=1)
loss /= (weights + 1e-8)
weights = tf.to_float(tf.greater(weights, 0.))
return tf.reduce_sum(loss*weights), tf.reduce_sum(weights)
|
python
|
def multi_label_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Average loss over the labels."""
del vocab_size # unused arg
logits = top_out
num_labels = tf.shape(targets)[1]
logits = tf.tile(logits, [1, num_labels, 1, 1, 1])
xent, weights = common_layers.padded_cross_entropy(
logits,
targets,
model_hparams.label_smoothing,
weights_fn=weights_fn,
reduce_sum=False,
)
xent = tf.squeeze(xent, [2, 3])
weights = tf.squeeze(weights, [2, 3])
# average loss over all labels
loss = tf.reduce_sum(xent, axis=1)
weights = tf.reduce_sum(weights, axis=1)
loss /= (weights + 1e-8)
weights = tf.to_float(tf.greater(weights, 0.))
return tf.reduce_sum(loss*weights), tf.reduce_sum(weights)
|
[
"def",
"multi_label_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weights_fn",
")",
":",
"del",
"vocab_size",
"# unused arg",
"logits",
"=",
"top_out",
"num_labels",
"=",
"tf",
".",
"shape",
"(",
"targets",
")",
"[",
"1",
"]",
"logits",
"=",
"tf",
".",
"tile",
"(",
"logits",
",",
"[",
"1",
",",
"num_labels",
",",
"1",
",",
"1",
",",
"1",
"]",
")",
"xent",
",",
"weights",
"=",
"common_layers",
".",
"padded_cross_entropy",
"(",
"logits",
",",
"targets",
",",
"model_hparams",
".",
"label_smoothing",
",",
"weights_fn",
"=",
"weights_fn",
",",
"reduce_sum",
"=",
"False",
",",
")",
"xent",
"=",
"tf",
".",
"squeeze",
"(",
"xent",
",",
"[",
"2",
",",
"3",
"]",
")",
"weights",
"=",
"tf",
".",
"squeeze",
"(",
"weights",
",",
"[",
"2",
",",
"3",
"]",
")",
"# average loss over all labels",
"loss",
"=",
"tf",
".",
"reduce_sum",
"(",
"xent",
",",
"axis",
"=",
"1",
")",
"weights",
"=",
"tf",
".",
"reduce_sum",
"(",
"weights",
",",
"axis",
"=",
"1",
")",
"loss",
"/=",
"(",
"weights",
"+",
"1e-8",
")",
"weights",
"=",
"tf",
".",
"to_float",
"(",
"tf",
".",
"greater",
"(",
"weights",
",",
"0.",
")",
")",
"return",
"tf",
".",
"reduce_sum",
"(",
"loss",
"*",
"weights",
")",
",",
"tf",
".",
"reduce_sum",
"(",
"weights",
")"
] |
Average loss over the labels.
|
[
"Average",
"loss",
"over",
"the",
"labels",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L689-L711
|
train
|
Average loss over the labels.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b10011 + 0o134) + chr(0b110101) + chr(0b100101 + 0o22), 0b1000), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(111) + chr(51) + '\065' + chr(48), 20715 - 20707), ehT0Px3KOsy9(chr(1366 - 1318) + chr(0b1101111) + '\061' + chr(0b10010 + 0o43) + '\x30', 0o10), ehT0Px3KOsy9(chr(48) + chr(5156 - 5045) + chr(0b110000 + 0o3) + chr(0b110111) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b10100 + 0o37) + chr(468 - 414) + chr(432 - 382), 0o10), ehT0Px3KOsy9('\060' + '\157' + '\063' + chr(0b11000 + 0o33), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(50) + '\x31' + '\062', ord("\x08")), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(111) + chr(0b110110) + chr(417 - 366), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(1475 - 1425) + chr(0b110000), 28471 - 28463), ehT0Px3KOsy9('\x30' + chr(111) + chr(1724 - 1671) + '\066', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101010 + 0o5) + chr(49) + chr(0b110010) + chr(0b110001), 0o10), ehT0Px3KOsy9('\x30' + chr(0b111011 + 0o64) + '\061' + chr(0b11111 + 0o21) + '\x36', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b100 + 0o153) + chr(871 - 820) + chr(230 - 177) + chr(0b110000), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1000011 + 0o54) + chr(0b11100 + 0o27) + '\x32' + '\x30', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + chr(0b110111) + '\063', 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + chr(55) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(53) + chr(0b10111 + 0o32), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b100001 + 0o22) + chr(0b11001 + 0o33) + chr(0b10101 + 0o35), 0o10), ehT0Px3KOsy9(chr(0b10110 + 0o32) + '\157' + chr(0b100011 + 0o17) + chr(0b100000 + 0o21) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(0b1100110 + 0o11) + chr(0b10000 + 0o43) + '\x35' + chr(2405 - 2355), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(49) + chr(0b110100) + chr(53), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(0b100011 + 0o17) + chr(865 - 812), 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\062' + '\060' + chr(2030 - 1976), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(0b11001 + 0o30) + chr(0b1 + 0o60), 10312 - 10304), ehT0Px3KOsy9(chr(48) + chr(0b101111 + 0o100) + '\063' + '\062' + '\x30', 8), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(51) + chr(0b10001 + 0o42) + '\062', 0o10), ehT0Px3KOsy9(chr(162 - 114) + '\157' + chr(2004 - 1954) + chr(0b11 + 0o64) + '\x32', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1100001 + 0o16) + '\061' + '\x30' + chr(0b110110), 8), ehT0Px3KOsy9('\x30' + chr(0b1000011 + 0o54) + chr(49) + chr(51) + chr(50), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b100101 + 0o112) + chr(0b110001) + chr(0b100 + 0o61) + chr(0b1 + 0o60), 43739 - 43731), ehT0Px3KOsy9(chr(0b110000) + chr(5153 - 5042) + chr(0b110110) + chr(0b1 + 0o60), ord("\x08")), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(0b101100 + 0o103) + chr(0b110011) + '\067' + chr(52), 8), ehT0Px3KOsy9(chr(0b110000) + chr(10641 - 10530) + chr(55) + chr(1070 - 1020), 42504 - 42496), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110110) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(398 - 350) + '\157' + chr(0b101000 + 0o12) + chr(1287 - 1232) + chr(777 - 724), ord("\x08")), ehT0Px3KOsy9(chr(348 - 300) + chr(11109 - 10998) + '\063' + chr(55) + chr(1195 - 1143), 8), ehT0Px3KOsy9(chr(48) + chr(0b110 + 0o151) + '\062' + chr(943 - 895) + chr(50), 0b1000), ehT0Px3KOsy9(chr(667 - 619) + chr(3754 - 3643) + chr(0b110010) + chr(0b101100 + 0o5) + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(0b1010 + 0o46) + '\x6f' + chr(49) + chr(0b110101) + '\x31', 8), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + chr(49) + '\066', 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(0b1010010 + 0o35) + chr(53) + '\x30', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x01'), chr(0b1100100) + chr(9018 - 8917) + chr(99) + '\x6f' + '\144' + chr(0b111100 + 0o51))(chr(0b1110101) + chr(0b1110100) + '\146' + chr(45) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def digenOuecR7e(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, Pdbc6Q2jZ4RQ):
del CeyMIoSyrpkQ
wF9nmvjsKjYM = juqulShQpQzQ
cNrGXe_50UpS = IDJ2eXGCBCDu.nauYfLglTpcb(xIEmRseySp3z)[ehT0Px3KOsy9(chr(48) + chr(6213 - 6102) + '\061', 0o10)]
wF9nmvjsKjYM = IDJ2eXGCBCDu.tile(wF9nmvjsKjYM, [ehT0Px3KOsy9(chr(0b11010 + 0o26) + chr(0b1010101 + 0o32) + chr(0b110001), 8), cNrGXe_50UpS, ehT0Px3KOsy9('\x30' + '\x6f' + chr(49), 8), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(234 - 185), 8), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(49), 8)])
(_YHpmhjj_eGR, ZurHTci57aXw) = jSKPaHwSAfVv.padded_cross_entropy(wF9nmvjsKjYM, xIEmRseySp3z, tq24Tk6UZ6u1.FSjUgdaczzRk, weights_fn=Pdbc6Q2jZ4RQ, reduce_sum=ehT0Px3KOsy9('\060' + chr(11672 - 11561) + '\x30', 0o10))
_YHpmhjj_eGR = IDJ2eXGCBCDu.squeeze(_YHpmhjj_eGR, [ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062', 20020 - 20012), ehT0Px3KOsy9('\060' + '\x6f' + chr(51), 0b1000)])
ZurHTci57aXw = IDJ2eXGCBCDu.squeeze(ZurHTci57aXw, [ehT0Px3KOsy9('\060' + chr(3472 - 3361) + chr(0b11010 + 0o30), 8), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33', 8)])
YpO0BcZ6fMsf = IDJ2eXGCBCDu.reduce_sum(_YHpmhjj_eGR, axis=ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110001), 8))
ZurHTci57aXw = IDJ2eXGCBCDu.reduce_sum(ZurHTci57aXw, axis=ehT0Px3KOsy9('\x30' + chr(0b1001001 + 0o46) + chr(1569 - 1520), 8))
YpO0BcZ6fMsf /= ZurHTci57aXw + 1e-08
ZurHTci57aXw = IDJ2eXGCBCDu.to_float(IDJ2eXGCBCDu.greater(ZurHTci57aXw, 0.0))
return (xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b']\x8d\x0c\x89\x91~\xd5@eC'), '\x64' + '\x65' + '\x63' + chr(111) + chr(100) + '\x65')(chr(0b1110101) + chr(0b101111 + 0o105) + '\x66' + '\055' + chr(0b111000)))(YpO0BcZ6fMsf * ZurHTci57aXw), xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b']\x8d\x0c\x89\x91~\xd5@eC'), chr(0b1100100) + chr(0b1100101) + chr(0b1000001 + 0o42) + '\x6f' + '\x64' + chr(0b1100101))('\x75' + chr(0b1100 + 0o150) + chr(0b1000110 + 0o40) + chr(0b101101) + chr(56)))(ZurHTci57aXw))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
one_hot_class_label_loss
|
def one_hot_class_label_loss(top_out,
targets,
model_hparams,
vocab_size,
weights_fn):
"""Apply softmax cross-entropy between outputs and targets.
Args:
top_out: logits Tensor with shape [batch, ?, ?, num_classes]
targets: one-hot encoding Tensor with shape [batch, ?, ?, num_classes]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
weights_fn:
Returns:
loss_scale (cross-entropy), loss_denom
"""
del model_hparams, vocab_size # unused arg
loss_scale = tf.losses.softmax_cross_entropy(
onehot_labels=targets, logits=top_out)
weights = weights_fn(targets)
loss_denom = tf.reduce_sum(weights)
return loss_scale, loss_denom
|
python
|
def one_hot_class_label_loss(top_out,
targets,
model_hparams,
vocab_size,
weights_fn):
"""Apply softmax cross-entropy between outputs and targets.
Args:
top_out: logits Tensor with shape [batch, ?, ?, num_classes]
targets: one-hot encoding Tensor with shape [batch, ?, ?, num_classes]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
weights_fn:
Returns:
loss_scale (cross-entropy), loss_denom
"""
del model_hparams, vocab_size # unused arg
loss_scale = tf.losses.softmax_cross_entropy(
onehot_labels=targets, logits=top_out)
weights = weights_fn(targets)
loss_denom = tf.reduce_sum(weights)
return loss_scale, loss_denom
|
[
"def",
"one_hot_class_label_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weights_fn",
")",
":",
"del",
"model_hparams",
",",
"vocab_size",
"# unused arg",
"loss_scale",
"=",
"tf",
".",
"losses",
".",
"softmax_cross_entropy",
"(",
"onehot_labels",
"=",
"targets",
",",
"logits",
"=",
"top_out",
")",
"weights",
"=",
"weights_fn",
"(",
"targets",
")",
"loss_denom",
"=",
"tf",
".",
"reduce_sum",
"(",
"weights",
")",
"return",
"loss_scale",
",",
"loss_denom"
] |
Apply softmax cross-entropy between outputs and targets.
Args:
top_out: logits Tensor with shape [batch, ?, ?, num_classes]
targets: one-hot encoding Tensor with shape [batch, ?, ?, num_classes]
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
weights_fn:
Returns:
loss_scale (cross-entropy), loss_denom
|
[
"Apply",
"softmax",
"cross",
"-",
"entropy",
"between",
"outputs",
"and",
"targets",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L714-L736
|
train
|
Applies softmax cross - entropy between outputs and targets.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(511 - 463) + chr(0b1101111) + chr(0b110010) + chr(1640 - 1590) + chr(0b1010 + 0o52), 38769 - 38761), ehT0Px3KOsy9('\x30' + '\x6f' + chr(49) + chr(0b110011) + chr(0b110101), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(1290 - 1241) + chr(1471 - 1416) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110011 + 0o0) + chr(54) + chr(49), 32999 - 32991), ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(0b1101111) + chr(50) + chr(0b110111) + chr(201 - 153), 4260 - 4252), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(49) + chr(0b110101) + '\064', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b10011 + 0o36) + chr(0b110100) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(6533 - 6422) + chr(51) + chr(0b110100) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(974 - 926) + '\157' + chr(55) + chr(68 - 20), ord("\x08")), ehT0Px3KOsy9(chr(1272 - 1224) + chr(6006 - 5895) + '\061' + chr(0b1000 + 0o50) + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(50) + '\x33' + chr(54), 8359 - 8351), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(0b1101111) + chr(50) + '\x34' + chr(0b110000), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x32' + chr(54) + chr(0b0 + 0o61), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110010) + chr(2137 - 2086), 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + '\x36', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110011) + chr(0b110000) + '\064', 52322 - 52314), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(0b1100101 + 0o12) + chr(0b10110 + 0o33) + '\x32' + chr(0b110000), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(855 - 804) + '\x31' + '\065', ord("\x08")), ehT0Px3KOsy9(chr(0b11100 + 0o24) + chr(7681 - 7570) + chr(0b101100 + 0o6) + chr(0b110110), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(0b1111 + 0o43) + chr(0b101011 + 0o11) + chr(0b110011), 0o10), ehT0Px3KOsy9(chr(0b101 + 0o53) + '\x6f' + chr(2381 - 2331) + chr(0b110111) + chr(561 - 511), ord("\x08")), ehT0Px3KOsy9(chr(179 - 131) + chr(111) + chr(0b110101 + 0o2) + '\065', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101010 + 0o5) + chr(50) + chr(0b110101) + chr(0b110010 + 0o4), 13308 - 13300), ehT0Px3KOsy9(chr(48) + '\x6f' + '\060', 20955 - 20947), ehT0Px3KOsy9(chr(1329 - 1281) + chr(3739 - 3628) + '\061' + '\062' + chr(0b101101 + 0o4), 29687 - 29679), ehT0Px3KOsy9('\x30' + chr(0b1001011 + 0o44) + '\x32' + chr(55) + chr(51), 0b1000), ehT0Px3KOsy9(chr(1549 - 1501) + '\157' + chr(51) + '\063' + '\066', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1028 - 978) + chr(0b110001) + '\067', ord("\x08")), ehT0Px3KOsy9(chr(2103 - 2055) + chr(4343 - 4232) + '\067' + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(0b1101111) + '\x34' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\x34' + chr(0b110011), 25250 - 25242), ehT0Px3KOsy9(chr(1085 - 1037) + '\157' + '\066' + chr(55), 0o10), ehT0Px3KOsy9(chr(0b100111 + 0o11) + '\157' + chr(1332 - 1282) + chr(0b101000 + 0o12) + '\065', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1001001 + 0o46) + '\065' + '\x33', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1000111 + 0o50) + chr(0b101 + 0o56) + chr(0b11010 + 0o31) + '\x31', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b11010 + 0o27) + '\060' + chr(54), 8), ehT0Px3KOsy9(chr(933 - 885) + chr(0b111010 + 0o65) + chr(0b10011 + 0o37) + chr(54) + chr(0b110110), 34464 - 34456), ehT0Px3KOsy9('\x30' + chr(0b1011001 + 0o26) + chr(732 - 681) + chr(55), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110001) + chr(53) + chr(0b10001 + 0o44), ord("\x08")), ehT0Px3KOsy9(chr(344 - 296) + chr(0b1110 + 0o141) + chr(49) + chr(0b11111 + 0o23) + chr(0b100101 + 0o16), 41434 - 41426)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(111) + '\065' + chr(0b101010 + 0o6), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xba'), '\144' + '\x65' + chr(0b1100011) + chr(0b101011 + 0o104) + '\144' + chr(101))(chr(0b1110101) + chr(116) + chr(0b1001 + 0o135) + chr(0b100011 + 0o12) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def sQFKa_Gsem2q(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, Pdbc6Q2jZ4RQ):
del tq24Tk6UZ6u1, CeyMIoSyrpkQ
Z44FhoIBMgHP = IDJ2eXGCBCDu.losses.softmax_cross_entropy(onehot_labels=xIEmRseySp3z, logits=juqulShQpQzQ)
ZurHTci57aXw = Pdbc6Q2jZ4RQ(xIEmRseySp3z)
FEhwUNri8SNx = IDJ2eXGCBCDu.reduce_sum(ZurHTci57aXw)
return (Z44FhoIBMgHP, FEhwUNri8SNx)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
real_log_poisson_loss
|
def real_log_poisson_loss(top_out,
targets,
model_hparams,
vocab_size,
weights_fn):
"""Poisson loss for real."""
del model_hparams, vocab_size # unused arg
predictions = top_out
if (len(common_layers.shape_list(top_out)) != len(
common_layers.shape_list(targets))):
predictions = tf.squeeze(top_out, axis=[-1])
with tf.name_scope("log_possion"):
weights = weights_fn(targets)
lp_loss = tf.nn.log_poisson_loss(targets, predictions)
return tf.reduce_sum(lp_loss * weights), tf.reduce_sum(weights)
|
python
|
def real_log_poisson_loss(top_out,
targets,
model_hparams,
vocab_size,
weights_fn):
"""Poisson loss for real."""
del model_hparams, vocab_size # unused arg
predictions = top_out
if (len(common_layers.shape_list(top_out)) != len(
common_layers.shape_list(targets))):
predictions = tf.squeeze(top_out, axis=[-1])
with tf.name_scope("log_possion"):
weights = weights_fn(targets)
lp_loss = tf.nn.log_poisson_loss(targets, predictions)
return tf.reduce_sum(lp_loss * weights), tf.reduce_sum(weights)
|
[
"def",
"real_log_poisson_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weights_fn",
")",
":",
"del",
"model_hparams",
",",
"vocab_size",
"# unused arg",
"predictions",
"=",
"top_out",
"if",
"(",
"len",
"(",
"common_layers",
".",
"shape_list",
"(",
"top_out",
")",
")",
"!=",
"len",
"(",
"common_layers",
".",
"shape_list",
"(",
"targets",
")",
")",
")",
":",
"predictions",
"=",
"tf",
".",
"squeeze",
"(",
"top_out",
",",
"axis",
"=",
"[",
"-",
"1",
"]",
")",
"with",
"tf",
".",
"name_scope",
"(",
"\"log_possion\"",
")",
":",
"weights",
"=",
"weights_fn",
"(",
"targets",
")",
"lp_loss",
"=",
"tf",
".",
"nn",
".",
"log_poisson_loss",
"(",
"targets",
",",
"predictions",
")",
"return",
"tf",
".",
"reduce_sum",
"(",
"lp_loss",
"*",
"weights",
")",
",",
"tf",
".",
"reduce_sum",
"(",
"weights",
")"
] |
Poisson loss for real.
|
[
"Poisson",
"loss",
"for",
"real",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L751-L765
|
train
|
Poisson loss for real.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(147 - 99) + chr(7867 - 7756) + chr(0b100100 + 0o15) + '\x30' + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b100011 + 0o114) + '\061' + chr(0b110010) + '\x34', 26944 - 26936), ehT0Px3KOsy9('\x30' + chr(0b1000010 + 0o55) + '\061' + chr(0b0 + 0o64) + chr(0b10010 + 0o45), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2026 - 1977) + chr(2267 - 2216) + chr(0b100 + 0o57), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110100 + 0o2) + chr(0b101000 + 0o14), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b11001 + 0o27), 0o10), ehT0Px3KOsy9(chr(1961 - 1913) + chr(3905 - 3794) + chr(940 - 890) + '\062' + '\065', 21714 - 21706), ehT0Px3KOsy9('\x30' + '\157' + chr(1349 - 1296) + chr(49), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33' + chr(49) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\063' + chr(54) + chr(55), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(49) + chr(1339 - 1289) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(111) + chr(0b111 + 0o54) + chr(1998 - 1945) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x31' + '\063' + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(6534 - 6423) + chr(0b100111 + 0o13) + '\x30' + '\064', 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x31' + '\060' + '\x37', 64626 - 64618), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\067' + '\062', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(2690 - 2638) + chr(55), 13857 - 13849), ehT0Px3KOsy9(chr(48) + '\157' + '\066' + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\x6f' + '\062' + '\066' + chr(0b100 + 0o61), 0b1000), ehT0Px3KOsy9(chr(429 - 381) + '\157' + '\061' + '\x37' + chr(0b1010 + 0o46), 0b1000), ehT0Px3KOsy9('\060' + '\157' + '\x33' + '\060', 24875 - 24867), ehT0Px3KOsy9('\x30' + chr(111) + chr(1630 - 1579) + chr(0b110110) + '\x35', 0o10), ehT0Px3KOsy9(chr(712 - 664) + chr(0b1011110 + 0o21) + chr(922 - 872) + chr(0b110110) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110111) + '\x37', 0o10), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(111) + chr(0b101100 + 0o7) + chr(0b110101) + '\063', 0o10), ehT0Px3KOsy9('\x30' + chr(7705 - 7594) + chr(49) + chr(882 - 834) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(7251 - 7140) + chr(0b100111 + 0o12) + chr(48), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110000 + 0o2) + '\066' + chr(55), 40627 - 40619), ehT0Px3KOsy9('\060' + '\157' + chr(371 - 320) + chr(53) + chr(0b110000 + 0o6), 5565 - 5557), ehT0Px3KOsy9('\060' + '\157' + '\063' + chr(0b110001) + '\066', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(396 - 344) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(6015 - 5904) + '\064' + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(5580 - 5469) + chr(2059 - 2010) + '\x32' + '\x33', 0o10), ehT0Px3KOsy9('\060' + '\157' + '\x32' + chr(0b110010 + 0o5) + chr(0b110010 + 0o3), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(2246 - 2195) + '\064' + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(1894 - 1846) + chr(0b100000 + 0o117) + chr(989 - 940) + chr(51), 0b1000), ehT0Px3KOsy9(chr(48) + chr(3819 - 3708) + chr(0b110011) + chr(0b1010 + 0o51) + chr(0b1010 + 0o53), 0b1000), ehT0Px3KOsy9(chr(554 - 506) + '\x6f' + '\x33' + '\064' + chr(0b101100 + 0o6), ord("\x08")), ehT0Px3KOsy9(chr(301 - 253) + chr(0b1101111) + '\066' + chr(0b110010), 8), ehT0Px3KOsy9(chr(48) + '\157' + chr(50) + chr(1542 - 1491), 33364 - 33356)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1001111 + 0o40) + '\x35' + chr(0b110000), 28121 - 28113)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'k'), chr(3179 - 3079) + chr(3773 - 3672) + chr(6044 - 5945) + chr(0b1101111) + chr(0b100111 + 0o75) + chr(101))(chr(4366 - 4249) + chr(0b1101010 + 0o12) + chr(8638 - 8536) + chr(0b101101 + 0o0) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def wLa7yBGuyt7m(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, Pdbc6Q2jZ4RQ):
del tq24Tk6UZ6u1, CeyMIoSyrpkQ
qIQi_VFCIFZL = juqulShQpQzQ
if c2A0yzQpDQB3(xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'6<\x06+Q_\xf9@\xbf\x83'), chr(0b1100100) + '\145' + chr(99) + chr(0b1101011 + 0o4) + chr(100) + chr(101))('\x75' + '\x74' + chr(102) + '\x2d' + chr(56)))(juqulShQpQzQ)) != c2A0yzQpDQB3(xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'6<\x06+Q_\xf9@\xbf\x83'), chr(9003 - 8903) + chr(0b1100101) + chr(1143 - 1044) + chr(11889 - 11778) + chr(8056 - 7956) + chr(101))(chr(0b1110101) + chr(12495 - 12379) + '\x66' + chr(45) + chr(0b111000)))(xIEmRseySp3z)):
qIQi_VFCIFZL = IDJ2eXGCBCDu.squeeze(juqulShQpQzQ, axis=[-ehT0Px3KOsy9('\060' + chr(3351 - 3240) + '\x31', ord("\x08"))])
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'+5\n>ks\xf6F\xbc\x92'), '\x64' + chr(0b101110 + 0o67) + chr(1475 - 1376) + '\157' + chr(4219 - 4119) + chr(101))('\165' + chr(866 - 750) + '\146' + chr(0b101101) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b');\x00\x04Do\xe6Z\xa5\x98<'), chr(4245 - 4145) + '\x65' + chr(0b1100011) + chr(9109 - 8998) + chr(100) + '\145')('\x75' + '\x74' + chr(0b1100110) + chr(0b101101) + chr(56))):
ZurHTci57aXw = Pdbc6Q2jZ4RQ(xIEmRseySp3z)
Tn4EciR0ZCyh = IDJ2eXGCBCDu.nn.log_poisson_loss(xIEmRseySp3z, qIQi_VFCIFZL)
return (xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'71\x03.We\xcaZ\xb9\x9a'), chr(100) + chr(0b1010011 + 0o22) + '\x63' + chr(111) + '\x64' + chr(101))(chr(4695 - 4578) + chr(7464 - 7348) + chr(0b101001 + 0o75) + chr(0b101101) + chr(1276 - 1220)))(Tn4EciR0ZCyh * ZurHTci57aXw), xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'71\x03.We\xcaZ\xb9\x9a'), '\144' + '\145' + chr(0b1100011) + '\x6f' + chr(100) + chr(0b111001 + 0o54))('\x75' + chr(116) + chr(2095 - 1993) + '\055' + '\x38'))(ZurHTci57aXw))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
sigmoid_class_label_loss
|
def sigmoid_class_label_loss(top_out,
targets,
model_hparams,
vocab_size,
weights_fn):
"""Loss for class label."""
# Expect inputs of size [batch-size, timesteps, 1, num-classes], where the
# last dimension of num-classes represents logits for binary labels
del model_hparams, vocab_size # unused arg
loss_scale = tf.losses.sigmoid_cross_entropy(
multi_class_labels=targets, logits=top_out)
weights = weights_fn(targets)
loss_denom = tf.reduce_sum(weights)
return loss_scale, loss_denom
|
python
|
def sigmoid_class_label_loss(top_out,
targets,
model_hparams,
vocab_size,
weights_fn):
"""Loss for class label."""
# Expect inputs of size [batch-size, timesteps, 1, num-classes], where the
# last dimension of num-classes represents logits for binary labels
del model_hparams, vocab_size # unused arg
loss_scale = tf.losses.sigmoid_cross_entropy(
multi_class_labels=targets, logits=top_out)
weights = weights_fn(targets)
loss_denom = tf.reduce_sum(weights)
return loss_scale, loss_denom
|
[
"def",
"sigmoid_class_label_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weights_fn",
")",
":",
"# Expect inputs of size [batch-size, timesteps, 1, num-classes], where the",
"# last dimension of num-classes represents logits for binary labels",
"del",
"model_hparams",
",",
"vocab_size",
"# unused arg",
"loss_scale",
"=",
"tf",
".",
"losses",
".",
"sigmoid_cross_entropy",
"(",
"multi_class_labels",
"=",
"targets",
",",
"logits",
"=",
"top_out",
")",
"weights",
"=",
"weights_fn",
"(",
"targets",
")",
"loss_denom",
"=",
"tf",
".",
"reduce_sum",
"(",
"weights",
")",
"return",
"loss_scale",
",",
"loss_denom"
] |
Loss for class label.
|
[
"Loss",
"for",
"class",
"label",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L768-L781
|
train
|
Loss for class label.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(0b1101111 + 0o0) + chr(0b110010) + '\x32', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b1111 + 0o42) + '\065' + chr(0b0 + 0o60), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b111 + 0o150) + chr(0b110001) + chr(55) + '\063', 0b1000), ehT0Px3KOsy9(chr(1429 - 1381) + chr(0b1101111) + chr(0b110001) + '\x35' + '\065', 938 - 930), ehT0Px3KOsy9(chr(2069 - 2021) + chr(0b1101111) + chr(0b11111 + 0o24) + chr(51) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + '\062' + chr(0b1000 + 0o56) + '\x31', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110011) + chr(0b110000) + chr(0b110100), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110100), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b0 + 0o157) + chr(50) + '\x31' + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100111 + 0o10) + chr(1266 - 1212), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(9229 - 9118) + chr(1139 - 1089) + '\x32' + chr(0b110111), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + '\060' + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(49) + chr(0b1111 + 0o44) + chr(1524 - 1471), 14277 - 14269), ehT0Px3KOsy9(chr(48) + chr(0b1101011 + 0o4) + chr(0b10111 + 0o32) + chr(683 - 634) + chr(0b10010 + 0o40), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(1159 - 1110) + '\x37' + chr(0b100110 + 0o13), 18570 - 18562), ehT0Px3KOsy9('\060' + chr(10774 - 10663) + '\063' + chr(0b110010) + chr(48), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(1493 - 1382) + chr(824 - 773) + chr(0b10100 + 0o35) + chr(0b10110 + 0o40), ord("\x08")), ehT0Px3KOsy9(chr(1220 - 1172) + chr(111) + '\x32' + chr(0b1011 + 0o50) + chr(50), ord("\x08")), ehT0Px3KOsy9('\060' + chr(325 - 214) + chr(977 - 924) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x33' + chr(0b11111 + 0o22) + chr(0b110011), 51027 - 51019), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110010) + chr(0b110001 + 0o0) + '\063', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(352 - 297) + chr(0b100010 + 0o24), 0o10), ehT0Px3KOsy9('\060' + chr(6238 - 6127) + '\061' + chr(135 - 87) + chr(2000 - 1945), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b101000 + 0o107) + chr(49) + '\065' + chr(0b100111 + 0o16), 8), ehT0Px3KOsy9(chr(826 - 778) + '\x6f' + chr(0b110110), 8), ehT0Px3KOsy9('\060' + chr(0b1000001 + 0o56) + '\065' + '\x36', 54367 - 54359), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + chr(2006 - 1955), 0o10), ehT0Px3KOsy9(chr(684 - 636) + chr(5193 - 5082) + chr(0b100111 + 0o14) + chr(0b110 + 0o52) + chr(0b111 + 0o51), 32386 - 32378), ehT0Px3KOsy9(chr(48) + chr(0b1001110 + 0o41) + chr(51) + chr(53) + '\060', ord("\x08")), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(4133 - 4022) + chr(0b110011) + chr(54) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(49) + chr(0b101111 + 0o6) + chr(0b10111 + 0o31), 8), ehT0Px3KOsy9('\060' + chr(962 - 851) + '\x33' + '\065' + chr(0b10011 + 0o35), 8), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + '\x37' + chr(0b110001), 8), ehT0Px3KOsy9('\x30' + chr(0b101110 + 0o101) + chr(0b1 + 0o60) + chr(0b10011 + 0o44) + '\062', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1011111 + 0o20) + '\x35' + '\x36', 8), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(111) + '\063' + chr(0b110010) + '\060', 8), ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(0b0 + 0o157) + '\062' + chr(53) + chr(0b11101 + 0o31), 7315 - 7307), ehT0Px3KOsy9('\060' + '\x6f' + chr(51) + chr(0b110100) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b10100 + 0o34) + '\157' + chr(54) + '\062', 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\063' + chr(0b110101 + 0o1) + chr(0b110011), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b11111 + 0o26) + chr(2261 - 2213), 15808 - 15800)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'<'), chr(2902 - 2802) + chr(0b1100101) + chr(0b101101 + 0o66) + chr(0b1101111) + '\144' + chr(6599 - 6498))(chr(0b1110101) + '\164' + '\146' + chr(483 - 438) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def m0T3xdcEkCcx(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, Pdbc6Q2jZ4RQ):
del tq24Tk6UZ6u1, CeyMIoSyrpkQ
Z44FhoIBMgHP = IDJ2eXGCBCDu.losses.sigmoid_cross_entropy(multi_class_labels=xIEmRseySp3z, logits=juqulShQpQzQ)
ZurHTci57aXw = Pdbc6Q2jZ4RQ(xIEmRseySp3z)
FEhwUNri8SNx = IDJ2eXGCBCDu.reduce_sum(ZurHTci57aXw)
return (Z44FhoIBMgHP, FEhwUNri8SNx)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
video_loss
|
def video_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Compute loss numerator and denominator for one shard of output."""
del vocab_size # unused arg
logits = top_out
logits = tf.reshape(logits, [-1] + common_layers.shape_list(logits)[2:])
targets = tf.reshape(targets, [-1] + common_layers.shape_list(targets)[2:])
cutoff = getattr(model_hparams, "video_modality_loss_cutoff", 0.01)
return common_layers.padded_cross_entropy(
logits,
targets,
model_hparams.label_smoothing,
cutoff=cutoff,
weights_fn=weights_fn)
|
python
|
def video_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Compute loss numerator and denominator for one shard of output."""
del vocab_size # unused arg
logits = top_out
logits = tf.reshape(logits, [-1] + common_layers.shape_list(logits)[2:])
targets = tf.reshape(targets, [-1] + common_layers.shape_list(targets)[2:])
cutoff = getattr(model_hparams, "video_modality_loss_cutoff", 0.01)
return common_layers.padded_cross_entropy(
logits,
targets,
model_hparams.label_smoothing,
cutoff=cutoff,
weights_fn=weights_fn)
|
[
"def",
"video_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weights_fn",
")",
":",
"del",
"vocab_size",
"# unused arg",
"logits",
"=",
"top_out",
"logits",
"=",
"tf",
".",
"reshape",
"(",
"logits",
",",
"[",
"-",
"1",
"]",
"+",
"common_layers",
".",
"shape_list",
"(",
"logits",
")",
"[",
"2",
":",
"]",
")",
"targets",
"=",
"tf",
".",
"reshape",
"(",
"targets",
",",
"[",
"-",
"1",
"]",
"+",
"common_layers",
".",
"shape_list",
"(",
"targets",
")",
"[",
"2",
":",
"]",
")",
"cutoff",
"=",
"getattr",
"(",
"model_hparams",
",",
"\"video_modality_loss_cutoff\"",
",",
"0.01",
")",
"return",
"common_layers",
".",
"padded_cross_entropy",
"(",
"logits",
",",
"targets",
",",
"model_hparams",
".",
"label_smoothing",
",",
"cutoff",
"=",
"cutoff",
",",
"weights_fn",
"=",
"weights_fn",
")"
] |
Compute loss numerator and denominator for one shard of output.
|
[
"Compute",
"loss",
"numerator",
"and",
"denominator",
"for",
"one",
"shard",
"of",
"output",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L812-L824
|
train
|
Compute loss numerator and denominator for one shard of output.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(111) + '\062' + chr(0b110000) + chr(52), 0b1000), ehT0Px3KOsy9(chr(0b10010 + 0o36) + chr(0b101100 + 0o103) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(456 - 408) + '\x6f' + '\x37' + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x32' + chr(2170 - 2116) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(53), 8), ehT0Px3KOsy9(chr(0b11110 + 0o22) + chr(0b1101111) + '\061' + chr(0b110101) + chr(53), 0b1000), ehT0Px3KOsy9(chr(1196 - 1148) + chr(0b1101110 + 0o1) + chr(54) + chr(48), 0o10), ehT0Px3KOsy9(chr(625 - 577) + chr(0b1010111 + 0o30) + '\x32' + chr(0b11100 + 0o33), 0o10), ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(0b101 + 0o152) + chr(49) + chr(0b100001 + 0o21) + chr(53), 0b1000), ehT0Px3KOsy9(chr(2177 - 2129) + chr(0b1010 + 0o145) + chr(1416 - 1363) + chr(300 - 252), 26856 - 26848), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b1001 + 0o52) + '\066' + '\063', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(1375 - 1264) + chr(0b110001) + '\063' + chr(910 - 855), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110010) + chr(1471 - 1421) + chr(0b110100), 45109 - 45101), ehT0Px3KOsy9('\x30' + chr(0b1011010 + 0o25) + chr(50) + '\x32' + chr(52), 8), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x33' + chr(55) + chr(51), ord("\x08")), ehT0Px3KOsy9(chr(1038 - 990) + '\157' + '\062' + chr(0b101010 + 0o12) + '\x32', 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b1110 + 0o43) + chr(54), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\x32' + '\062', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x33' + chr(2727 - 2673) + chr(55), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(462 - 412) + chr(0b11011 + 0o34) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(0b1101111) + chr(538 - 488) + '\060' + chr(50), 31445 - 31437), ehT0Px3KOsy9('\060' + chr(2179 - 2068) + chr(1588 - 1539) + chr(0b11011 + 0o31) + chr(0b10100 + 0o43), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + chr(0b110000) + chr(0b110110), 61259 - 61251), ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(0b1001011 + 0o44) + '\x32' + chr(0b110000) + '\x31', 0b1000), ehT0Px3KOsy9(chr(283 - 235) + chr(0b1001001 + 0o46) + chr(0b110100) + chr(0b110011), 22646 - 22638), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x32' + chr(48) + chr(0b110101), 1521 - 1513), ehT0Px3KOsy9('\x30' + chr(0b101000 + 0o107) + chr(2390 - 2340) + chr(48) + chr(0b110011), 34405 - 34397), ehT0Px3KOsy9('\060' + chr(0b100101 + 0o112) + '\065' + chr(0b101101 + 0o4), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + chr(48), 27800 - 27792), ehT0Px3KOsy9('\060' + '\157' + chr(0b110101) + chr(0b101010 + 0o10), 3783 - 3775), ehT0Px3KOsy9('\x30' + '\x6f' + '\x32' + '\x35' + chr(1871 - 1819), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33' + chr(0b110001) + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(990 - 942) + '\x6f' + chr(2133 - 2082) + chr(0b110010) + '\061', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\066' + chr(0b1101 + 0o52), 20128 - 20120), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(0b1101111) + '\061' + chr(0b110001) + '\066', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101011 + 0o4) + chr(0b10110 + 0o33) + '\x30' + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(448 - 400) + '\157' + '\x31' + chr(0b101111 + 0o6) + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + '\062' + '\063', 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b101110 + 0o4) + chr(0b1000 + 0o50), 8), ehT0Px3KOsy9('\x30' + '\157' + '\061' + chr(0b11100 + 0o30) + '\067', 8)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(2730 - 2677) + chr(0b10 + 0o56), 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xaa'), '\144' + '\x65' + chr(0b1001 + 0o132) + '\157' + '\x64' + chr(0b1010 + 0o133))('\x75' + '\x74' + chr(0b1100011 + 0o3) + chr(0b1000 + 0o45) + chr(0b11110 + 0o32)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def JIGZcfgBtYav(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, Pdbc6Q2jZ4RQ):
del CeyMIoSyrpkQ
wF9nmvjsKjYM = juqulShQpQzQ
wF9nmvjsKjYM = IDJ2eXGCBCDu.reshape(wF9nmvjsKjYM, [-ehT0Px3KOsy9(chr(0b110000) + chr(0b101110 + 0o101) + chr(665 - 616), 0o10)] + jSKPaHwSAfVv.shape_list(wF9nmvjsKjYM)[ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x32', ord("\x08")):])
xIEmRseySp3z = IDJ2eXGCBCDu.reshape(xIEmRseySp3z, [-ehT0Px3KOsy9('\060' + '\157' + '\061', 8)] + jSKPaHwSAfVv.shape_list(xIEmRseySp3z)[ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50), 8):])
EjnQGacZaia3 = xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf2l_\xd1j2]x1I\x9f6\xe0\xed\xa4\xceaa\x84\xe3\xe1\x122\xeeE\xe3'), '\144' + chr(0b1100101) + chr(0b1100011) + '\157' + '\x64' + chr(101))('\165' + chr(0b1111 + 0o145) + chr(0b1100110) + '\055' + chr(0b101010 + 0o16)), 0.01)
return xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b"\xf4d_\xd0`\tot'G\x80,\xcb\xf1\x95\xd6|}\x87\xc5"), '\144' + chr(101) + chr(0b1100011) + '\157' + '\144' + chr(7444 - 7343))(chr(0b1010001 + 0o44) + chr(7432 - 7316) + chr(0b1100110) + '\x2d' + '\x38'))(wF9nmvjsKjYM, xIEmRseySp3z, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc2VQ\xe1b\tQt/R\xa14'), chr(100) + chr(0b1100101) + chr(99) + '\157' + chr(0b1100100) + chr(10011 - 9910))(chr(0b1110101) + '\164' + chr(0b100111 + 0o77) + chr(0b101101) + chr(0b100101 + 0o23))), cutoff=EjnQGacZaia3, weights_fn=Pdbc6Q2jZ4RQ)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
video_l1_loss
|
def video_l1_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Compute loss numerator and denominator for one shard of output."""
del vocab_size # unused arg
logits = top_out
logits = tf.reshape(logits, [-1] + common_layers.shape_list(logits)[2:-1])
targets = tf.reshape(targets, [-1] + common_layers.shape_list(targets)[2:])
weights = weights_fn(targets)
# Shift targets by 0.5 so later just casting to int gives the prediction.
# So for int targets, say 0 and 7, we actually train to predict 0.5 and 7.5.
# Later (in merics or infer) this is cast to int anyway. Also, we have no
# loss beyond cutoff = 0.2 as these are already correct predictions.
targets = tf.to_float(targets) + 0.5
loss = video_l1_internal_loss(logits, targets, model_hparams)
return tf.reduce_sum(loss * weights), tf.reduce_sum(weights)
|
python
|
def video_l1_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Compute loss numerator and denominator for one shard of output."""
del vocab_size # unused arg
logits = top_out
logits = tf.reshape(logits, [-1] + common_layers.shape_list(logits)[2:-1])
targets = tf.reshape(targets, [-1] + common_layers.shape_list(targets)[2:])
weights = weights_fn(targets)
# Shift targets by 0.5 so later just casting to int gives the prediction.
# So for int targets, say 0 and 7, we actually train to predict 0.5 and 7.5.
# Later (in merics or infer) this is cast to int anyway. Also, we have no
# loss beyond cutoff = 0.2 as these are already correct predictions.
targets = tf.to_float(targets) + 0.5
loss = video_l1_internal_loss(logits, targets, model_hparams)
return tf.reduce_sum(loss * weights), tf.reduce_sum(weights)
|
[
"def",
"video_l1_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weights_fn",
")",
":",
"del",
"vocab_size",
"# unused arg",
"logits",
"=",
"top_out",
"logits",
"=",
"tf",
".",
"reshape",
"(",
"logits",
",",
"[",
"-",
"1",
"]",
"+",
"common_layers",
".",
"shape_list",
"(",
"logits",
")",
"[",
"2",
":",
"-",
"1",
"]",
")",
"targets",
"=",
"tf",
".",
"reshape",
"(",
"targets",
",",
"[",
"-",
"1",
"]",
"+",
"common_layers",
".",
"shape_list",
"(",
"targets",
")",
"[",
"2",
":",
"]",
")",
"weights",
"=",
"weights_fn",
"(",
"targets",
")",
"# Shift targets by 0.5 so later just casting to int gives the prediction.",
"# So for int targets, say 0 and 7, we actually train to predict 0.5 and 7.5.",
"# Later (in merics or infer) this is cast to int anyway. Also, we have no",
"# loss beyond cutoff = 0.2 as these are already correct predictions.",
"targets",
"=",
"tf",
".",
"to_float",
"(",
"targets",
")",
"+",
"0.5",
"loss",
"=",
"video_l1_internal_loss",
"(",
"logits",
",",
"targets",
",",
"model_hparams",
")",
"return",
"tf",
".",
"reduce_sum",
"(",
"loss",
"*",
"weights",
")",
",",
"tf",
".",
"reduce_sum",
"(",
"weights",
")"
] |
Compute loss numerator and denominator for one shard of output.
|
[
"Compute",
"loss",
"numerator",
"and",
"denominator",
"for",
"one",
"shard",
"of",
"output",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L852-L865
|
train
|
Compute loss numerator and denominator for one shard of output.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(6077 - 5966) + chr(51) + '\x37' + chr(0b10001 + 0o44), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x35' + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b1111 + 0o41) + '\x6f' + chr(50) + chr(0b11011 + 0o32) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110111) + '\064', 0b1000), ehT0Px3KOsy9(chr(471 - 423) + chr(0b1101111) + '\063' + chr(53) + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(1543 - 1495) + chr(111) + chr(319 - 270) + chr(55) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(12225 - 12114) + '\063' + chr(53) + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(629 - 580) + chr(52) + chr(1814 - 1765), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101100 + 0o3) + '\062' + '\066' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\x33' + '\066' + chr(48), 50666 - 50658), ehT0Px3KOsy9(chr(1275 - 1227) + '\x6f' + chr(0b1101 + 0o44) + '\061' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b10 + 0o56) + '\157' + chr(0b101110 + 0o4) + '\065' + chr(0b1010 + 0o50), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1186 - 1134), 2485 - 2477), ehT0Px3KOsy9(chr(48) + chr(0b10111 + 0o130) + '\x31' + chr(0b1001 + 0o54) + '\x34', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51) + '\063' + chr(0b10011 + 0o42), 0o10), ehT0Px3KOsy9(chr(780 - 732) + '\x6f' + '\064' + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\x32' + '\x36' + chr(0b10001 + 0o43), 46606 - 46598), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x31' + chr(163 - 114) + chr(275 - 225), 8), ehT0Px3KOsy9('\x30' + '\x6f' + '\063' + chr(2116 - 2063) + chr(262 - 213), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110010) + '\x35' + chr(51), 3291 - 3283), ehT0Px3KOsy9(chr(2122 - 2074) + chr(0b1000011 + 0o54) + '\062' + chr(1039 - 989) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(289 - 178) + chr(0b110111) + chr(2844 - 2789), 0o10), ehT0Px3KOsy9('\060' + chr(0b11 + 0o154) + chr(0b110010) + chr(0b110010 + 0o1) + '\x34', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b111001 + 0o66) + chr(0b101001 + 0o12) + chr(123 - 70) + chr(55), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b10011 + 0o134) + '\062' + chr(52) + chr(2540 - 2486), 0b1000), ehT0Px3KOsy9(chr(0b1010 + 0o46) + '\157' + chr(1042 - 992) + '\066' + chr(1266 - 1214), 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + chr(50) + chr(49), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010) + chr(0b110011) + '\060', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110010) + '\060' + '\063', 1935 - 1927), ehT0Px3KOsy9('\060' + chr(111) + chr(51) + chr(0b110110) + chr(0b110000 + 0o4), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(50) + '\066' + chr(770 - 717), 0b1000), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(0b101100 + 0o103) + chr(0b11101 + 0o25) + chr(0b1 + 0o63) + chr(2299 - 2246), 0o10), ehT0Px3KOsy9(chr(613 - 565) + '\x6f' + '\063' + chr(0b1011 + 0o54) + chr(549 - 495), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110010) + chr(621 - 569) + chr(1170 - 1122), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b111001 + 0o66) + chr(708 - 658) + '\x32' + chr(0b11101 + 0o25), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49) + '\x30' + chr(0b110010), 59215 - 59207), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2583 - 2531) + chr(49), 24213 - 24205), ehT0Px3KOsy9(chr(48) + chr(694 - 583) + chr(1965 - 1915) + chr(0b11011 + 0o25) + chr(0b10111 + 0o40), ord("\x08")), ehT0Px3KOsy9(chr(0b10001 + 0o37) + '\x6f' + chr(49) + chr(54) + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(949 - 901) + chr(4975 - 4864) + '\061' + '\x31' + chr(52), 27337 - 27329)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(7967 - 7856) + chr(53) + chr(0b110000), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xd3'), chr(0b1100100) + '\145' + '\143' + chr(111) + '\x64' + chr(6391 - 6290))(chr(8113 - 7996) + chr(1690 - 1574) + '\146' + chr(45) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def N9rPShunrV5z(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, Pdbc6Q2jZ4RQ):
del CeyMIoSyrpkQ
wF9nmvjsKjYM = juqulShQpQzQ
wF9nmvjsKjYM = IDJ2eXGCBCDu.reshape(wF9nmvjsKjYM, [-ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b10001 + 0o40), 0b1000)] + jSKPaHwSAfVv.shape_list(wF9nmvjsKjYM)[ehT0Px3KOsy9(chr(128 - 80) + chr(0b1101011 + 0o4) + chr(0b1111 + 0o43), 0b1000):-ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b100100 + 0o15), 8)])
xIEmRseySp3z = IDJ2eXGCBCDu.reshape(xIEmRseySp3z, [-ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001), 8)] + jSKPaHwSAfVv.shape_list(xIEmRseySp3z)[ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010), 8):])
ZurHTci57aXw = Pdbc6Q2jZ4RQ(xIEmRseySp3z)
xIEmRseySp3z = IDJ2eXGCBCDu.to_float(xIEmRseySp3z) + 0.5
YpO0BcZ6fMsf = FbxNVHwNCwFN(wF9nmvjsKjYM, xIEmRseySp3z, tq24Tk6UZ6u1)
return (xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x8fI\xc9I\x89\x1b\xb2\x12\xc7\x9d'), '\144' + '\x65' + chr(7992 - 7893) + '\157' + chr(100) + chr(2840 - 2739))('\x75' + '\164' + chr(102) + '\x2d' + chr(0b111000)))(YpO0BcZ6fMsf * ZurHTci57aXw), xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x8fI\xc9I\x89\x1b\xb2\x12\xc7\x9d'), chr(0b1100100) + chr(0b1100101) + chr(0b1011010 + 0o11) + chr(111) + chr(0b100011 + 0o101) + chr(2706 - 2605))(chr(6047 - 5930) + chr(0b1110010 + 0o2) + chr(0b1100110 + 0o0) + '\x2d' + chr(0b10010 + 0o46)))(ZurHTci57aXw))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
video_l2_loss
|
def video_l2_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Compute loss numerator and denominator for one shard of output."""
del vocab_size # unused arg
logits = top_out
logits = tf.reshape(logits, [-1] + common_layers.shape_list(logits)[2:-1])
targets = tf.reshape(targets, [-1] + common_layers.shape_list(targets)[2:])
weights = weights_fn(targets)
# Shift targets by 0.5 so later just casting to int gives the prediction.
# So for int targets, say 0 and 7, we actually train to predict 0.5 and 7.5.
# Later (in merics or infer) this is cast to int anyway. Also, we have no
# loss beyond cutoff = 0.2 as these are already correct predictions.
targets = tf.to_float(targets) + 0.5
loss = video_l2_internal_loss(logits, targets, model_hparams)
return tf.reduce_sum(loss * weights), tf.reduce_sum(weights)
|
python
|
def video_l2_loss(top_out, targets, model_hparams, vocab_size, weights_fn):
"""Compute loss numerator and denominator for one shard of output."""
del vocab_size # unused arg
logits = top_out
logits = tf.reshape(logits, [-1] + common_layers.shape_list(logits)[2:-1])
targets = tf.reshape(targets, [-1] + common_layers.shape_list(targets)[2:])
weights = weights_fn(targets)
# Shift targets by 0.5 so later just casting to int gives the prediction.
# So for int targets, say 0 and 7, we actually train to predict 0.5 and 7.5.
# Later (in merics or infer) this is cast to int anyway. Also, we have no
# loss beyond cutoff = 0.2 as these are already correct predictions.
targets = tf.to_float(targets) + 0.5
loss = video_l2_internal_loss(logits, targets, model_hparams)
return tf.reduce_sum(loss * weights), tf.reduce_sum(weights)
|
[
"def",
"video_l2_loss",
"(",
"top_out",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
",",
"weights_fn",
")",
":",
"del",
"vocab_size",
"# unused arg",
"logits",
"=",
"top_out",
"logits",
"=",
"tf",
".",
"reshape",
"(",
"logits",
",",
"[",
"-",
"1",
"]",
"+",
"common_layers",
".",
"shape_list",
"(",
"logits",
")",
"[",
"2",
":",
"-",
"1",
"]",
")",
"targets",
"=",
"tf",
".",
"reshape",
"(",
"targets",
",",
"[",
"-",
"1",
"]",
"+",
"common_layers",
".",
"shape_list",
"(",
"targets",
")",
"[",
"2",
":",
"]",
")",
"weights",
"=",
"weights_fn",
"(",
"targets",
")",
"# Shift targets by 0.5 so later just casting to int gives the prediction.",
"# So for int targets, say 0 and 7, we actually train to predict 0.5 and 7.5.",
"# Later (in merics or infer) this is cast to int anyway. Also, we have no",
"# loss beyond cutoff = 0.2 as these are already correct predictions.",
"targets",
"=",
"tf",
".",
"to_float",
"(",
"targets",
")",
"+",
"0.5",
"loss",
"=",
"video_l2_internal_loss",
"(",
"logits",
",",
"targets",
",",
"model_hparams",
")",
"return",
"tf",
".",
"reduce_sum",
"(",
"loss",
"*",
"weights",
")",
",",
"tf",
".",
"reduce_sum",
"(",
"weights",
")"
] |
Compute loss numerator and denominator for one shard of output.
|
[
"Compute",
"loss",
"numerator",
"and",
"denominator",
"for",
"one",
"shard",
"of",
"output",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L874-L887
|
train
|
Compute loss numerator and denominator for one shard of output.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110001) + chr(0b100111 + 0o12) + '\061', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x31' + '\x35' + '\x31', 50388 - 50380), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(50) + chr(0b110101 + 0o0) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(377 - 329) + '\157' + chr(0b1001 + 0o50) + '\x30' + '\x34', 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b101101 + 0o5) + '\x35' + chr(280 - 229), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(2156 - 2045) + chr(0b111 + 0o54) + '\x31', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b101 + 0o55) + chr(0b101100 + 0o7) + chr(0b110000), 46017 - 46009), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + chr(0b110100) + chr(712 - 664), 0o10), ehT0Px3KOsy9(chr(0b11111 + 0o21) + chr(11093 - 10982) + '\061' + chr(0b11110 + 0o24) + chr(0b110001 + 0o1), 0b1000), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(3423 - 3312) + chr(53) + chr(0b100010 + 0o23), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1 + 0o156) + '\061' + chr(0b110111) + chr(0b100 + 0o63), 0o10), ehT0Px3KOsy9('\x30' + chr(3738 - 3627) + chr(0b101 + 0o55) + '\064', 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(2310 - 2261) + chr(52) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(0b1 + 0o57) + '\157' + '\x37' + '\066', 15217 - 15209), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51) + chr(2461 - 2411) + chr(0b110000), 50347 - 50339), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(51) + chr(0b10001 + 0o40) + chr(0b1010 + 0o46), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b10010 + 0o40) + chr(281 - 227) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(1307 - 1259) + chr(0b1101111) + chr(0b1011 + 0o51) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110110) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(2268 - 2220) + chr(5581 - 5470) + chr(0b110011) + '\062' + '\x33', 0b1000), ehT0Px3KOsy9(chr(930 - 882) + '\x6f' + '\x35' + chr(54), 5435 - 5427), ehT0Px3KOsy9('\x30' + chr(752 - 641) + chr(51) + chr(0b110100) + '\x32', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(50) + chr(0b110011) + chr(780 - 726), 27311 - 27303), ehT0Px3KOsy9(chr(0b100101 + 0o13) + '\157' + '\062' + '\x36' + '\x35', 8), ehT0Px3KOsy9(chr(0b0 + 0o60) + '\157' + chr(51) + chr(53) + '\x37', 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\063' + chr(48) + '\x37', 36749 - 36741), ehT0Px3KOsy9(chr(963 - 915) + chr(7876 - 7765) + chr(0b11101 + 0o31) + chr(635 - 586), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(4642 - 4531) + '\x32' + chr(51) + chr(0b100011 + 0o21), 0o10), ehT0Px3KOsy9(chr(1836 - 1788) + chr(111) + chr(51) + '\x34' + '\064', 52713 - 52705), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(0b110000) + '\x32', 54886 - 54878), ehT0Px3KOsy9(chr(48) + chr(4473 - 4362) + chr(52) + chr(0b100011 + 0o20), 58804 - 58796), ehT0Px3KOsy9('\x30' + '\x6f' + chr(544 - 490) + chr(0b110100), 0o10), ehT0Px3KOsy9('\x30' + chr(0b101 + 0o152) + chr(0b110011) + '\067' + chr(0b110101), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(50) + chr(208 - 158) + chr(51), 6205 - 6197), ehT0Px3KOsy9(chr(0b110000) + chr(2486 - 2375) + chr(0b110011) + chr(0b110100) + chr(1149 - 1099), 8), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b100011 + 0o16) + chr(2393 - 2338) + '\x30', 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(0b11011 + 0o26) + chr(0b110100) + chr(49), 31878 - 31870), ehT0Px3KOsy9(chr(701 - 653) + chr(0b1100110 + 0o11) + chr(1091 - 1040) + chr(0b110101) + chr(52), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b100 + 0o55) + chr(48) + chr(1663 - 1614), 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + chr(0b1001 + 0o53) + chr(0b110100), 8)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(5897 - 5786) + chr(464 - 411) + '\060', 62079 - 62071)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'g'), '\x64' + chr(0b0 + 0o145) + chr(0b1000000 + 0o43) + chr(0b1101111) + chr(0b11110 + 0o106) + chr(0b1100101))('\165' + chr(9083 - 8967) + chr(0b1111 + 0o127) + '\055' + chr(0b1000 + 0o60)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def uPyZteooulGN(juqulShQpQzQ, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ, Pdbc6Q2jZ4RQ):
del CeyMIoSyrpkQ
wF9nmvjsKjYM = juqulShQpQzQ
wF9nmvjsKjYM = IDJ2eXGCBCDu.reshape(wF9nmvjsKjYM, [-ehT0Px3KOsy9(chr(0b110000) + chr(1180 - 1069) + chr(2096 - 2047), 20216 - 20208)] + jSKPaHwSAfVv.shape_list(wF9nmvjsKjYM)[ehT0Px3KOsy9(chr(0b110000) + chr(8129 - 8018) + '\062', 0b1000):-ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b101001 + 0o10), 8)])
xIEmRseySp3z = IDJ2eXGCBCDu.reshape(xIEmRseySp3z, [-ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(49), 8)] + jSKPaHwSAfVv.shape_list(xIEmRseySp3z)[ehT0Px3KOsy9(chr(1255 - 1207) + '\x6f' + '\x32', 8):])
ZurHTci57aXw = Pdbc6Q2jZ4RQ(xIEmRseySp3z)
xIEmRseySp3z = IDJ2eXGCBCDu.to_float(xIEmRseySp3z) + 0.5
YpO0BcZ6fMsf = F2LU7d09n74P(wF9nmvjsKjYM, xIEmRseySp3z, tq24Tk6UZ6u1)
return (xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b";\x0c\xca1\xc8|\xc7\x93'@"), chr(100) + chr(2947 - 2846) + chr(0b1100011) + '\x6f' + chr(100) + '\x65')('\165' + '\x74' + chr(0b1100110) + chr(45) + chr(56)))(YpO0BcZ6fMsf * ZurHTci57aXw), xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b";\x0c\xca1\xc8|\xc7\x93'@"), chr(4567 - 4467) + chr(0b1100101) + chr(0b1100011) + chr(0b1101011 + 0o4) + chr(3720 - 3620) + chr(101))(chr(2003 - 1886) + '\x74' + chr(0b1100110) + chr(0b110 + 0o47) + '\070'))(ZurHTci57aXw))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
class_label_top
|
def class_label_top(body_output, targets, model_hparams, vocab_size):
"""Transform inputs from model space to target space.
Average over inner dims and a linear layer to logits.
Args:
body_output: A Tensor with shape [batch, ?, ?, body_output_size].
targets:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
a Tensors, each with shape [batch_size, 1, 1, 1, vocab_size]
"""
del targets # unused arg
with tf.variable_scope("class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)):
x = body_output
x = tf.reduce_mean(x, axis=[1, 2], keepdims=True)
res = tf.layers.dense(x, vocab_size)
return tf.expand_dims(res, 3)
|
python
|
def class_label_top(body_output, targets, model_hparams, vocab_size):
"""Transform inputs from model space to target space.
Average over inner dims and a linear layer to logits.
Args:
body_output: A Tensor with shape [batch, ?, ?, body_output_size].
targets:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
a Tensors, each with shape [batch_size, 1, 1, 1, vocab_size]
"""
del targets # unused arg
with tf.variable_scope("class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)):
x = body_output
x = tf.reduce_mean(x, axis=[1, 2], keepdims=True)
res = tf.layers.dense(x, vocab_size)
return tf.expand_dims(res, 3)
|
[
"def",
"class_label_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"with",
"tf",
".",
"variable_scope",
"(",
"\"class_label_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
")",
":",
"x",
"=",
"body_output",
"x",
"=",
"tf",
".",
"reduce_mean",
"(",
"x",
",",
"axis",
"=",
"[",
"1",
",",
"2",
"]",
",",
"keepdims",
"=",
"True",
")",
"res",
"=",
"tf",
".",
"layers",
".",
"dense",
"(",
"x",
",",
"vocab_size",
")",
"return",
"tf",
".",
"expand_dims",
"(",
"res",
",",
"3",
")"
] |
Transform inputs from model space to target space.
Average over inner dims and a linear layer to logits.
Args:
body_output: A Tensor with shape [batch, ?, ?, body_output_size].
targets:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
a Tensors, each with shape [batch_size, 1, 1, 1, vocab_size]
|
[
"Transform",
"inputs",
"from",
"model",
"space",
"to",
"target",
"space",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L927-L947
|
train
|
Transform inputs from model space to target space.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b1111 + 0o42) + '\x34' + '\x32', 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(708 - 658) + chr(0b1011 + 0o52) + '\x32', 0o10), ehT0Px3KOsy9(chr(0b101001 + 0o7) + '\x6f' + chr(0b110001 + 0o2) + chr(0b110000) + chr(53), 0o10), ehT0Px3KOsy9(chr(0b1010 + 0o46) + '\x6f' + '\x32' + '\064' + chr(1541 - 1488), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\067' + '\066', 0b1000), ehT0Px3KOsy9(chr(1197 - 1149) + chr(111) + chr(50) + chr(932 - 884) + '\x30', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110110 + 0o1) + chr(0b101111 + 0o4), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(49) + '\065', 0o10), ehT0Px3KOsy9(chr(48) + chr(7429 - 7318) + chr(1275 - 1224) + chr(51) + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + '\067', ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(390 - 339) + chr(0b11100 + 0o25) + chr(2457 - 2407), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b1000 + 0o51) + '\x32' + '\065', ord("\x08")), ehT0Px3KOsy9(chr(739 - 691) + '\157' + chr(1594 - 1540) + chr(510 - 462), 7110 - 7102), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010 + 0o145) + chr(0b110011) + '\066' + chr(54), 3904 - 3896), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110011) + chr(0b10010 + 0o41) + '\x33', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(0b10111 + 0o36) + '\065', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110100) + chr(51), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010 + 0o0) + chr(51) + chr(48), 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110010) + '\060' + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011) + chr(0b1000 + 0o53) + chr(0b100111 + 0o20), 55002 - 54994), ehT0Px3KOsy9('\x30' + '\157' + '\x31' + '\x33' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(1459 - 1411) + chr(111) + chr(2338 - 2289) + '\x33' + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(521 - 472) + '\064' + chr(0b110000), 0o10), ehT0Px3KOsy9('\x30' + chr(1541 - 1430) + chr(0b110001) + chr(49) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b110000 + 0o0) + '\157' + chr(49) + chr(1774 - 1723) + chr(50), 0b1000), ehT0Px3KOsy9(chr(416 - 368) + chr(111) + chr(50) + chr(0b10010 + 0o42) + '\x36', 0b1000), ehT0Px3KOsy9('\060' + '\157' + '\062' + '\066' + chr(48), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001) + chr(0b100110 + 0o14), 0b1000), ehT0Px3KOsy9(chr(0b0 + 0o60) + chr(3916 - 3805) + chr(0b101001 + 0o10) + chr(929 - 879), 8), ehT0Px3KOsy9(chr(0b111 + 0o51) + '\157' + chr(569 - 520) + chr(523 - 471) + chr(0b1010 + 0o46), 8), ehT0Px3KOsy9(chr(1997 - 1949) + chr(111) + '\x33' + chr(0b110101) + '\067', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + chr(0b100001 + 0o23) + '\067', 0o10), ehT0Px3KOsy9(chr(48) + chr(4889 - 4778) + chr(0b110000 + 0o7), 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001) + '\064' + chr(0b11111 + 0o24), ord("\x08")), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(8401 - 8290) + '\063' + chr(0b101110 + 0o10) + chr(203 - 151), 35095 - 35087), ehT0Px3KOsy9(chr(0b110000) + chr(0b1001010 + 0o45) + chr(0b1101 + 0o44) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(3108 - 2997) + chr(0b1101 + 0o44) + chr(0b100000 + 0o20) + '\x31', 21489 - 21481), ehT0Px3KOsy9('\060' + chr(1462 - 1351) + chr(54), 41909 - 41901), ehT0Px3KOsy9(chr(1507 - 1459) + chr(111) + chr(0b10000 + 0o42) + chr(0b100001 + 0o26) + chr(0b1111 + 0o47), 0b1000), ehT0Px3KOsy9(chr(914 - 866) + chr(0b1101101 + 0o2) + chr(723 - 672) + chr(0b11101 + 0o30) + chr(0b101001 + 0o10), 33724 - 33716)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(7023 - 6912) + chr(0b10010 + 0o43) + '\060', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'L'), chr(100) + '\145' + chr(8636 - 8537) + chr(0b1101111) + chr(0b1100100) + chr(0b1011101 + 0o10))(chr(9697 - 9580) + chr(116) + chr(1326 - 1224) + '\x2d' + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def y_HKuLsmsrGR(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x14\xcb\xb7\xcc_\xe2Wh\x87t\xa9:@\x93'), '\x64' + '\145' + '\x63' + '\157' + '\144' + chr(0b100100 + 0o101))('\165' + chr(116) + chr(0b11110 + 0o110) + chr(0b101101) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x01\xc6\xa4\xd6M\xdfWl\xbab\xa6\n]\x99\xbf\xccO\xf1\xf3\xa6\x06y\x04\x15\x7f\xc2'), chr(2100 - 2000) + chr(101) + chr(0b1001 + 0o132) + chr(111) + chr(0b1111 + 0o125) + chr(0b1100101))(chr(117) + chr(0b110100 + 0o100) + chr(0b1100110) + chr(0b101101) + chr(168 - 112)) % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xd0\xaa\xdcf\xce\x08f\xbco\x8e\x19'), chr(4894 - 4794) + '\x65' + chr(99) + '\157' + '\x64' + chr(0b11001 + 0o114))(chr(3998 - 3881) + chr(116) + '\146' + chr(1203 - 1158) + chr(0b111000))))):
OeWW0F1dBPRQ = KLYceYPUcF5e
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reduce_mean(OeWW0F1dBPRQ, axis=[ehT0Px3KOsy9(chr(905 - 857) + chr(0b101001 + 0o106) + chr(49), 55221 - 55213), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(6795 - 6684) + '\x32', 64802 - 64794)], keepdims=ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x31', 8))
MsbwfslwLjRO = IDJ2eXGCBCDu.layers.dense(OeWW0F1dBPRQ, CeyMIoSyrpkQ)
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x07\xd2\xb5\xc4P\xe4di\xb1j\xb9'), '\x64' + chr(0b1100101) + chr(0b10110 + 0o115) + chr(0b1101111) + chr(0b1100100) + '\145')(chr(117) + '\164' + chr(102) + '\x2d' + chr(56)))(MsbwfslwLjRO, ehT0Px3KOsy9('\060' + chr(0b1101111) + '\063', 0o10))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
image_top
|
def image_top(body_output, targets, model_hparams, vocab_size):
"""Top transformation for images."""
del targets # unused arg
# TODO(lukaszkaiser): is this a universal enough way to get channels?
num_channels = model_hparams.problem.num_channels
with tf.variable_scope("rgb_softmax"):
body_output_shape = common_layers.shape_list(body_output)
reshape_shape = body_output_shape[:3]
reshape_shape.extend([num_channels, vocab_size])
res = tf.layers.dense(body_output, vocab_size * num_channels)
res = tf.reshape(res, reshape_shape)
if not tf.get_variable_scope().reuse:
res_argmax = tf.argmax(res, axis=-1)
tf.summary.image(
"result",
common_layers.tpu_safe_image_summary(res_argmax),
max_outputs=1)
return res
|
python
|
def image_top(body_output, targets, model_hparams, vocab_size):
"""Top transformation for images."""
del targets # unused arg
# TODO(lukaszkaiser): is this a universal enough way to get channels?
num_channels = model_hparams.problem.num_channels
with tf.variable_scope("rgb_softmax"):
body_output_shape = common_layers.shape_list(body_output)
reshape_shape = body_output_shape[:3]
reshape_shape.extend([num_channels, vocab_size])
res = tf.layers.dense(body_output, vocab_size * num_channels)
res = tf.reshape(res, reshape_shape)
if not tf.get_variable_scope().reuse:
res_argmax = tf.argmax(res, axis=-1)
tf.summary.image(
"result",
common_layers.tpu_safe_image_summary(res_argmax),
max_outputs=1)
return res
|
[
"def",
"image_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"# TODO(lukaszkaiser): is this a universal enough way to get channels?",
"num_channels",
"=",
"model_hparams",
".",
"problem",
".",
"num_channels",
"with",
"tf",
".",
"variable_scope",
"(",
"\"rgb_softmax\"",
")",
":",
"body_output_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"body_output",
")",
"reshape_shape",
"=",
"body_output_shape",
"[",
":",
"3",
"]",
"reshape_shape",
".",
"extend",
"(",
"[",
"num_channels",
",",
"vocab_size",
"]",
")",
"res",
"=",
"tf",
".",
"layers",
".",
"dense",
"(",
"body_output",
",",
"vocab_size",
"*",
"num_channels",
")",
"res",
"=",
"tf",
".",
"reshape",
"(",
"res",
",",
"reshape_shape",
")",
"if",
"not",
"tf",
".",
"get_variable_scope",
"(",
")",
".",
"reuse",
":",
"res_argmax",
"=",
"tf",
".",
"argmax",
"(",
"res",
",",
"axis",
"=",
"-",
"1",
")",
"tf",
".",
"summary",
".",
"image",
"(",
"\"result\"",
",",
"common_layers",
".",
"tpu_safe_image_summary",
"(",
"res_argmax",
")",
",",
"max_outputs",
"=",
"1",
")",
"return",
"res"
] |
Top transformation for images.
|
[
"Top",
"transformation",
"for",
"images",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L955-L972
|
train
|
Top transformation for images.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(49) + chr(0b11011 + 0o32) + chr(52), 28640 - 28632), ehT0Px3KOsy9(chr(553 - 505) + chr(2903 - 2792) + chr(2259 - 2208), 0o10), ehT0Px3KOsy9(chr(1739 - 1691) + chr(111) + chr(51) + chr(0b101110 + 0o3), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010) + chr(51) + '\x31', 0b1000), ehT0Px3KOsy9(chr(1429 - 1381) + chr(7016 - 6905) + chr(51) + chr(0b110100) + chr(0b110110), 55199 - 55191), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + '\x34' + chr(304 - 254), 0o10), ehT0Px3KOsy9(chr(0b11101 + 0o23) + '\157' + chr(0b11011 + 0o26) + '\x34' + chr(0b110011), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\062' + chr(0b110001 + 0o0) + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(374 - 326) + chr(0b1001001 + 0o46) + chr(0b10011 + 0o36) + '\064' + chr(0b110101), 0o10), ehT0Px3KOsy9('\x30' + chr(4064 - 3953) + '\061' + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1010001 + 0o36) + chr(0b10001 + 0o40), 28760 - 28752), ehT0Px3KOsy9('\060' + chr(0b100110 + 0o111) + chr(0b110101) + chr(1090 - 1035), 19136 - 19128), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110011) + chr(1582 - 1531) + '\x33', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101 + 0o142) + chr(1895 - 1844) + chr(55) + chr(0b1001 + 0o50), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(246 - 195) + chr(107 - 53) + '\x36', 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110010) + chr(1470 - 1415) + chr(87 - 34), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\063' + chr(50), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\x31' + chr(738 - 687) + chr(48), 58742 - 58734), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011110 + 0o21) + chr(0b111 + 0o53) + '\x36' + chr(0b110101), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(50) + '\063', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\061' + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b110010 + 0o75) + chr(0b110011) + chr(0b101010 + 0o13) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b100111 + 0o110) + chr(0b110001) + chr(0b110100) + chr(52), 38242 - 38234), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101011 + 0o4) + chr(0b110011) + '\x30' + chr(1185 - 1137), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b100 + 0o153) + '\061' + chr(0b101101 + 0o7) + chr(0b110101), 8), ehT0Px3KOsy9(chr(0b1010 + 0o46) + chr(111) + '\061' + chr(2190 - 2139), 0b1000), ehT0Px3KOsy9(chr(369 - 321) + chr(111) + chr(0b10001 + 0o42) + chr(0b1001 + 0o50) + '\x34', 0b1000), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(6741 - 6630) + chr(0b101101 + 0o5) + '\062' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(0b101100 + 0o4) + '\157' + chr(2372 - 2323) + chr(0b1 + 0o64) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(2051 - 2003) + chr(0b111000 + 0o67) + '\063' + '\065' + '\065', 8), ehT0Px3KOsy9(chr(1038 - 990) + chr(8330 - 8219) + chr(0b110001) + chr(0b110011), 8), ehT0Px3KOsy9('\060' + chr(111) + chr(2312 - 2261) + chr(0b101001 + 0o16) + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(0b100110 + 0o12) + '\x6f' + chr(0b110010) + chr(0b10011 + 0o43), 0o10), ehT0Px3KOsy9('\x30' + chr(5165 - 5054) + chr(0b110001) + '\x36', 8), ehT0Px3KOsy9('\060' + chr(0b110110 + 0o71) + chr(0b110101) + chr(0b10111 + 0o34), 47763 - 47755), ehT0Px3KOsy9('\x30' + '\x6f' + chr(51) + '\x33' + chr(2376 - 2326), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b11001 + 0o30) + chr(0b11001 + 0o31) + '\060', 0b1000), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(111) + chr(0b110010) + chr(0b1010 + 0o46) + chr(51), 42544 - 42536), ehT0Px3KOsy9(chr(0b110000) + chr(2427 - 2316) + chr(0b110010) + chr(0b110010) + chr(92 - 41), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(53) + chr(0b110100), 64511 - 64503)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(11154 - 11043) + chr(0b110101) + chr(48), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'B'), chr(0b1100100) + chr(2087 - 1986) + chr(2593 - 2494) + chr(0b11110 + 0o121) + chr(0b1100100) + chr(101))(chr(117) + chr(0b1110100) + '\146' + '\x2d' + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def xnD2VGWgfz3u(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
X1ZpHSxyKbHn = tq24Tk6UZ6u1.problem.X1ZpHSxyKbHn
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x1abk\xf0\x87\x80\x0f\xef:yb@\xab\x9c'), chr(100) + chr(0b111110 + 0o47) + '\x63' + '\x6f' + '\144' + '\145')(chr(0b10111 + 0o136) + chr(12586 - 12470) + chr(102) + chr(0b101101) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x1ed{\xc6\x95\x8d\x05\xfe\x08ky'), chr(0b1001101 + 0o27) + chr(0b1100101) + chr(0b1011100 + 0o7) + chr(0b1101111) + '\144' + chr(6847 - 6746))(chr(117) + chr(0b1110100) + chr(0b100010 + 0o104) + chr(45) + '\070')):
oknqTBFY6iX3 = jSKPaHwSAfVv.shape_list(KLYceYPUcF5e)
TTb6rlTz2h_U = oknqTBFY6iX3[:ehT0Px3KOsy9('\x30' + chr(111) + '\x33', 8)]
xafqLlk3kkUe(TTb6rlTz2h_U, xafqLlk3kkUe(SXOLrMavuUCe(b'\t{m\xfc\x88\x86'), chr(0b1001010 + 0o32) + chr(0b1001001 + 0o34) + chr(0b111 + 0o134) + chr(0b1101111) + chr(100) + chr(0b111101 + 0o50))(chr(0b1111 + 0o146) + chr(0b1110100) + chr(0b1100110) + chr(0b1011 + 0o42) + chr(56)))([X1ZpHSxyKbHn, CeyMIoSyrpkQ])
MsbwfslwLjRO = IDJ2eXGCBCDu.layers.dense(KLYceYPUcF5e, CeyMIoSyrpkQ * X1ZpHSxyKbHn)
MsbwfslwLjRO = IDJ2eXGCBCDu.reshape(MsbwfslwLjRO, TTb6rlTz2h_U)
if not xafqLlk3kkUe(IDJ2eXGCBCDu.get_variable_scope(), xafqLlk3kkUe(SXOLrMavuUCe(b'\x1efl\xea\x83'), chr(4998 - 4898) + chr(0b100100 + 0o101) + chr(0b1001011 + 0o30) + chr(0b1010111 + 0o30) + chr(0b10 + 0o142) + chr(0b1010011 + 0o22))(chr(5779 - 5662) + '\164' + chr(0b1100110) + chr(0b10011 + 0o32) + '\x38')):
mDMFq4zNSnwx = IDJ2eXGCBCDu.argmax(MsbwfslwLjRO, axis=-ehT0Px3KOsy9('\x30' + '\x6f' + chr(49), 8))
xafqLlk3kkUe(IDJ2eXGCBCDu.summary, xafqLlk3kkUe(SXOLrMavuUCe(b'%gt\xd8\xae\xb5\x05\xc9\x14xo_'), chr(100) + chr(7028 - 6927) + chr(0b1100011) + chr(0b111110 + 0o61) + chr(7174 - 7074) + chr(0b1000000 + 0o45))(chr(0b1110101) + chr(116) + chr(0b0 + 0o146) + chr(176 - 131) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x1efj\xec\x8a\x96'), chr(0b1010101 + 0o17) + '\145' + chr(99) + chr(466 - 355) + chr(0b10110 + 0o116) + chr(0b1100101))(chr(2096 - 1979) + chr(0b1001 + 0o153) + '\x66' + '\x2d' + chr(1646 - 1590)), xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'\x18sl\xc6\x95\x83\x05\xef:clN\xbc\x9c\xa9\xceo5\xd8r|\x94'), chr(0b1100100) + chr(101) + chr(0b1100011) + chr(0b1101111) + '\x64' + '\145')(chr(117) + '\164' + chr(102) + '\055' + chr(0b111000)))(mDMFq4zNSnwx), max_outputs=ehT0Px3KOsy9('\060' + chr(5906 - 5795) + '\x31', 8))
return MsbwfslwLjRO
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
image_channel_compress_top
|
def image_channel_compress_top(body_output, targets, model_hparams, vocab_size):
"""Transforms body output to return logits.
Args:
body_output: Tensor of shape [batch, img_len, img_len, depth].
targets:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
Tensor of shape [batch, img_len, img_len, channels, vocab_size].
"""
del targets # unused arg
with tf.variable_scope("image_channel_compress_modality"):
hidden_size = model_hparams.hidden_size
img_len = model_hparams.img_len
channels = 3 # RGB
batch = common_layers.shape_list(body_output)[0]
x = tf.layers.conv2d(
body_output,
hidden_size * channels,
kernel_size=(1, 1),
strides=(1, 1),
padding="VALID",
activation=tf.nn.relu,
name="decompress_conv")
x = tf.reshape(x, [batch, img_len, img_len * channels, hidden_size])
x = common_layers.layer_preprocess(x, model_hparams)
x = tf.layers.dense(x,
vocab_size,
use_bias=True,
activation=None,
name="output_conv")
x = tf.reshape(
x, [batch, img_len, img_len, channels, vocab_size])
return x
|
python
|
def image_channel_compress_top(body_output, targets, model_hparams, vocab_size):
"""Transforms body output to return logits.
Args:
body_output: Tensor of shape [batch, img_len, img_len, depth].
targets:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
Tensor of shape [batch, img_len, img_len, channels, vocab_size].
"""
del targets # unused arg
with tf.variable_scope("image_channel_compress_modality"):
hidden_size = model_hparams.hidden_size
img_len = model_hparams.img_len
channels = 3 # RGB
batch = common_layers.shape_list(body_output)[0]
x = tf.layers.conv2d(
body_output,
hidden_size * channels,
kernel_size=(1, 1),
strides=(1, 1),
padding="VALID",
activation=tf.nn.relu,
name="decompress_conv")
x = tf.reshape(x, [batch, img_len, img_len * channels, hidden_size])
x = common_layers.layer_preprocess(x, model_hparams)
x = tf.layers.dense(x,
vocab_size,
use_bias=True,
activation=None,
name="output_conv")
x = tf.reshape(
x, [batch, img_len, img_len, channels, vocab_size])
return x
|
[
"def",
"image_channel_compress_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"with",
"tf",
".",
"variable_scope",
"(",
"\"image_channel_compress_modality\"",
")",
":",
"hidden_size",
"=",
"model_hparams",
".",
"hidden_size",
"img_len",
"=",
"model_hparams",
".",
"img_len",
"channels",
"=",
"3",
"# RGB",
"batch",
"=",
"common_layers",
".",
"shape_list",
"(",
"body_output",
")",
"[",
"0",
"]",
"x",
"=",
"tf",
".",
"layers",
".",
"conv2d",
"(",
"body_output",
",",
"hidden_size",
"*",
"channels",
",",
"kernel_size",
"=",
"(",
"1",
",",
"1",
")",
",",
"strides",
"=",
"(",
"1",
",",
"1",
")",
",",
"padding",
"=",
"\"VALID\"",
",",
"activation",
"=",
"tf",
".",
"nn",
".",
"relu",
",",
"name",
"=",
"\"decompress_conv\"",
")",
"x",
"=",
"tf",
".",
"reshape",
"(",
"x",
",",
"[",
"batch",
",",
"img_len",
",",
"img_len",
"*",
"channels",
",",
"hidden_size",
"]",
")",
"x",
"=",
"common_layers",
".",
"layer_preprocess",
"(",
"x",
",",
"model_hparams",
")",
"x",
"=",
"tf",
".",
"layers",
".",
"dense",
"(",
"x",
",",
"vocab_size",
",",
"use_bias",
"=",
"True",
",",
"activation",
"=",
"None",
",",
"name",
"=",
"\"output_conv\"",
")",
"x",
"=",
"tf",
".",
"reshape",
"(",
"x",
",",
"[",
"batch",
",",
"img_len",
",",
"img_len",
",",
"channels",
",",
"vocab_size",
"]",
")",
"return",
"x"
] |
Transforms body output to return logits.
Args:
body_output: Tensor of shape [batch, img_len, img_len, depth].
targets:
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
Tensor of shape [batch, img_len, img_len, channels, vocab_size].
|
[
"Transforms",
"body",
"output",
"to",
"return",
"logits",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L975-L1010
|
train
|
Transforms body output to return logits.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x32', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1001101 + 0o42) + chr(1512 - 1461) + chr(50) + chr(0b110011), 0b1000), ehT0Px3KOsy9('\060' + chr(8270 - 8159) + chr(51) + chr(0b10 + 0o60) + chr(0b101111 + 0o1), 61952 - 61944), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + '\061', ord("\x08")), ehT0Px3KOsy9(chr(0b110000 + 0o0) + '\157' + chr(1647 - 1596) + '\x31' + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + '\066' + '\067', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b11100 + 0o27) + '\061', 8), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(0b1010001 + 0o36) + '\062' + chr(51) + '\x37', 0o10), ehT0Px3KOsy9(chr(48) + chr(9014 - 8903) + chr(314 - 263) + '\x35' + chr(2134 - 2084), 0o10), ehT0Px3KOsy9(chr(1728 - 1680) + '\x6f' + chr(1359 - 1310) + chr(0b110100) + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + '\x37' + '\x35', 24410 - 24402), ehT0Px3KOsy9('\060' + chr(0b1001111 + 0o40) + '\063' + chr(2028 - 1976), ord("\x08")), ehT0Px3KOsy9('\060' + chr(1952 - 1841) + chr(1420 - 1370) + chr(54) + '\063', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110010) + chr(0b110111) + chr(0b110 + 0o57), 0b1000), ehT0Px3KOsy9(chr(123 - 75) + '\157' + chr(1256 - 1207) + '\067' + chr(451 - 402), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(977 - 926) + chr(0b1011 + 0o52) + chr(0b110001 + 0o5), 65492 - 65484), ehT0Px3KOsy9(chr(2012 - 1964) + '\x6f' + chr(0b110011) + chr(0b10111 + 0o36) + chr(3009 - 2954), 64518 - 64510), ehT0Px3KOsy9(chr(0b110000) + chr(4566 - 4455) + chr(0b110010) + chr(52), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(8587 - 8476) + chr(51) + '\067' + chr(50), ord("\x08")), ehT0Px3KOsy9('\060' + chr(3591 - 3480) + '\062' + chr(0b101000 + 0o16) + chr(0b101101 + 0o11), 16763 - 16755), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x31' + chr(0b1011 + 0o54) + chr(0b110001 + 0o2), 38444 - 38436), ehT0Px3KOsy9(chr(1659 - 1611) + chr(111) + '\063' + chr(0b110000) + chr(0b110101), 7475 - 7467), ehT0Px3KOsy9('\x30' + '\157' + '\x37' + '\x35', 8), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(0b110100 + 0o73) + '\062' + '\x37' + chr(2245 - 2196), 0b1000), ehT0Px3KOsy9(chr(681 - 633) + chr(0b1101111) + chr(49) + '\067', 0b1000), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(0b1101111) + '\x33' + chr(0b110001 + 0o6) + chr(0b11111 + 0o25), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(1842 - 1791) + chr(0b1000 + 0o57) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + '\x37' + chr(0b110101), 8), ehT0Px3KOsy9(chr(863 - 815) + chr(111) + '\x32' + chr(0b11101 + 0o32) + '\064', 37697 - 37689), ehT0Px3KOsy9(chr(1518 - 1470) + '\157' + chr(0b101001 + 0o13) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\062' + chr(2188 - 2140) + '\062', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b110011) + '\x36' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(0b110011) + chr(52), 8), ehT0Px3KOsy9(chr(0b110000) + chr(2666 - 2555) + chr(2107 - 2057) + chr(55) + chr(52), 8), ehT0Px3KOsy9(chr(1918 - 1870) + '\157' + chr(0b110010) + chr(49), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + chr(51) + '\x32', 56434 - 56426), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b1101111) + chr(0b1101 + 0o44) + '\x31' + chr(50), 0o10), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(0b10 + 0o155) + '\x31' + chr(2611 - 2559) + chr(0b110000), 0o10), ehT0Px3KOsy9('\x30' + chr(3963 - 3852) + chr(0b110001) + '\x33' + chr(49), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + chr(0b1 + 0o61), 52507 - 52499)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(697 - 649) + chr(8089 - 7978) + '\065' + '\x30', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b','), '\x64' + chr(0b11110 + 0o107) + chr(99) + '\x6f' + '\x64' + chr(0b1100101))(chr(0b101111 + 0o106) + chr(952 - 836) + '\146' + chr(0b101101) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def vLZKgij6ya6Z(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b't\xb2\xa2p\xac\x15u\xfa\xb5\x14^s\xc3\xd8'), '\144' + chr(101) + chr(1294 - 1195) + chr(111) + chr(0b1100100) + '\x65')(chr(3025 - 2908) + '\x74' + chr(0b10 + 0o144) + chr(45) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'k\xbe\xb1~\xa8(z\xf7\x8b\tSy\xdf\xe2\x1a\x81\x82d\x88\xd3\x08\xb6b\xd4`\x98\x99\xd1\xd5Q\xfb'), chr(0b1100100) + '\x65' + chr(0b1100011) + '\x6f' + '\144' + chr(101))(chr(117) + chr(9392 - 9276) + '\x66' + '\055' + chr(0b111000))):
qzoyXN3kdhDL = tq24Tk6UZ6u1.qzoyXN3kdhDL
laxD7jy5y7k1 = tq24Tk6UZ6u1.laxD7jy5y7k1
H2MQqAZeamNo = ehT0Px3KOsy9(chr(0b110000) + chr(7941 - 7830) + '\x33', ord("\x08"))
dNwAahu8tvoY = jSKPaHwSAfVv.shape_list(KLYceYPUcF5e)[ehT0Px3KOsy9(chr(0b110000) + chr(646 - 535) + chr(1416 - 1368), 19021 - 19013)]
OeWW0F1dBPRQ = IDJ2eXGCBCDu.layers.conv2d(KLYceYPUcF5e, qzoyXN3kdhDL * H2MQqAZeamNo, kernel_size=(ehT0Px3KOsy9(chr(1612 - 1564) + '\x6f' + '\x31', 8041 - 8033), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\061', 8)), strides=(ehT0Px3KOsy9('\x30' + '\157' + '\061', 8), ehT0Px3KOsy9(chr(2134 - 2086) + chr(4457 - 4346) + '\061', 8)), padding=xafqLlk3kkUe(SXOLrMavuUCe(b'T\x92\x9cP\x89'), '\144' + chr(0b1100101) + chr(99) + '\x6f' + '\x64' + '\145')(chr(117) + chr(116) + chr(9809 - 9707) + '\x2d' + chr(0b111000)), activation=IDJ2eXGCBCDu.nn.relu, name=xafqLlk3kkUe(SXOLrMavuUCe(b'f\xb6\xb3v\xa0\x07k\xfa\x99\x14b\x7f\xdc\xd3\x0f'), chr(0b1100100) + chr(0b1100101) + chr(5475 - 5376) + chr(111) + chr(100) + chr(829 - 728))(chr(0b1110101) + '\x74' + chr(328 - 226) + '\x2d' + chr(1938 - 1882)))
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reshape(OeWW0F1dBPRQ, [dNwAahu8tvoY, laxD7jy5y7k1, laxD7jy5y7k1 * H2MQqAZeamNo, qzoyXN3kdhDL])
OeWW0F1dBPRQ = jSKPaHwSAfVv.layer_preprocess(OeWW0F1dBPRQ, tq24Tk6UZ6u1)
OeWW0F1dBPRQ = IDJ2eXGCBCDu.layers.dense(OeWW0F1dBPRQ, CeyMIoSyrpkQ, use_bias=ehT0Px3KOsy9('\060' + '\x6f' + '\x31', 8), activation=None, name=xafqLlk3kkUe(SXOLrMavuUCe(b'm\xa6\xa4i\xb8\x03F\xfc\x85\tK'), chr(0b10100 + 0o120) + chr(101) + chr(99) + chr(6885 - 6774) + '\x64' + '\145')('\165' + chr(0b1001010 + 0o52) + chr(4059 - 3957) + '\x2d' + '\x38'))
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reshape(OeWW0F1dBPRQ, [dNwAahu8tvoY, laxD7jy5y7k1, laxD7jy5y7k1, H2MQqAZeamNo, CeyMIoSyrpkQ])
return OeWW0F1dBPRQ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
image_channel_embeddings_top
|
def image_channel_embeddings_top(body_output,
targets,
model_hparams,
vocab_size):
"""Top transformation for images."""
del targets # unused arg
with tf.variable_scope("image_channel_embeddings_bottom"):
img_len = model_hparams.img_len
channels = model_hparams.num_channels
x = tf.layers.dense(
body_output, 256, use_bias=True, activation=None, name="output_conv")
x = tf.reshape(x,
[-1, img_len, img_len, channels, vocab_size])
return x
|
python
|
def image_channel_embeddings_top(body_output,
targets,
model_hparams,
vocab_size):
"""Top transformation for images."""
del targets # unused arg
with tf.variable_scope("image_channel_embeddings_bottom"):
img_len = model_hparams.img_len
channels = model_hparams.num_channels
x = tf.layers.dense(
body_output, 256, use_bias=True, activation=None, name="output_conv")
x = tf.reshape(x,
[-1, img_len, img_len, channels, vocab_size])
return x
|
[
"def",
"image_channel_embeddings_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"with",
"tf",
".",
"variable_scope",
"(",
"\"image_channel_embeddings_bottom\"",
")",
":",
"img_len",
"=",
"model_hparams",
".",
"img_len",
"channels",
"=",
"model_hparams",
".",
"num_channels",
"x",
"=",
"tf",
".",
"layers",
".",
"dense",
"(",
"body_output",
",",
"256",
",",
"use_bias",
"=",
"True",
",",
"activation",
"=",
"None",
",",
"name",
"=",
"\"output_conv\"",
")",
"x",
"=",
"tf",
".",
"reshape",
"(",
"x",
",",
"[",
"-",
"1",
",",
"img_len",
",",
"img_len",
",",
"channels",
",",
"vocab_size",
"]",
")",
"return",
"x"
] |
Top transformation for images.
|
[
"Top",
"transformation",
"for",
"images",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1013-L1026
|
train
|
Top transformation for images.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + '\157' + chr(55) + chr(1135 - 1085), 0o10), ehT0Px3KOsy9('\x30' + chr(8456 - 8345) + chr(50) + chr(0b101111 + 0o4) + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + '\x31' + chr(2562 - 2509), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(0b10000 + 0o43) + '\067' + '\x33', 55299 - 55291), ehT0Px3KOsy9(chr(48) + chr(111) + '\062' + chr(0b110011) + chr(0b1111 + 0o47), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110001) + chr(2201 - 2153) + chr(0b10 + 0o64), ord("\x08")), ehT0Px3KOsy9(chr(1421 - 1373) + chr(0b1101111) + chr(0b110010) + chr(0b110011) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110011) + chr(0b110011) + '\066', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010101 + 0o32) + chr(0b1010 + 0o47) + chr(276 - 227), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111 + 0o0) + chr(0b110010) + chr(49) + chr(2285 - 2232), 8), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b11001 + 0o31) + chr(0b110101) + chr(724 - 675), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110001) + '\x30' + '\x31', 37793 - 37785), ehT0Px3KOsy9(chr(0b110000) + chr(11456 - 11345) + chr(49) + '\063' + '\060', ord("\x08")), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(0b11010 + 0o125) + chr(0b110010) + chr(55) + '\061', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(49) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(50) + '\061' + chr(50), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b0 + 0o67) + chr(438 - 385), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(8073 - 7962) + '\063' + '\066' + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b1101 + 0o44) + chr(0b111 + 0o54) + '\061', 17048 - 17040), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + '\065' + chr(0b111 + 0o55), 0b1000), ehT0Px3KOsy9(chr(112 - 64) + chr(0b1000010 + 0o55) + chr(0b110010) + chr(0b11 + 0o64), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(4782 - 4671) + '\063' + chr(0b101110 + 0o6) + '\063', ord("\x08")), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(0b101110 + 0o101) + chr(1488 - 1439) + '\065' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(2111 - 2063) + chr(0b111111 + 0o60) + chr(1852 - 1797) + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(0b0 + 0o60) + chr(0b1100101 + 0o12) + chr(49) + chr(0b110011) + chr(51), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + '\063' + chr(1432 - 1384) + '\x34', 26034 - 26026), ehT0Px3KOsy9(chr(0b110000) + chr(1615 - 1504) + chr(490 - 439) + '\x35' + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(111) + '\062' + chr(53) + chr(51), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1011111 + 0o20) + '\063' + chr(1046 - 995) + chr(0b100101 + 0o15), 1662 - 1654), ehT0Px3KOsy9(chr(0b10100 + 0o34) + '\157' + chr(359 - 309) + chr(0b110110) + '\065', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(4539 - 4428) + chr(1568 - 1517) + chr(0b1000 + 0o52) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(547 - 499) + chr(11056 - 10945) + '\x32' + chr(0b110100) + chr(0b101001 + 0o13), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x33' + chr(1527 - 1479), 0o10), ehT0Px3KOsy9(chr(0b11100 + 0o24) + '\157' + chr(1972 - 1921) + chr(0b110011) + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(111) + chr(0b110111) + '\x35', 8), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\061' + '\x33' + chr(845 - 795), 0o10), ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(4612 - 4501) + chr(0b10011 + 0o40) + '\063' + chr(0b110111), 48839 - 48831), ehT0Px3KOsy9(chr(0b100110 + 0o12) + '\x6f' + chr(51) + chr(1979 - 1927) + chr(54), 45189 - 45181), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110010) + chr(0b101001 + 0o11) + chr(53), 0o10), ehT0Px3KOsy9(chr(0b10110 + 0o32) + '\157' + chr(0b10011 + 0o41), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(354 - 306) + chr(111) + '\x35' + chr(485 - 437), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x14'), chr(0b1100100) + chr(101) + chr(6805 - 6706) + chr(6146 - 6035) + chr(411 - 311) + chr(0b100010 + 0o103))(chr(117) + '\x74' + chr(0b10 + 0o144) + chr(45) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def iE3aUQwOOXi8(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'L\xfe\xff\x92\xff$\xf1\xbbV\xfe\xca.(\xdc'), chr(100) + chr(2691 - 2590) + '\x63' + chr(5998 - 5887) + chr(7571 - 7471) + chr(0b11001 + 0o114))('\165' + chr(0b1110100) + chr(102) + chr(0b101101) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'S\xf2\xec\x9c\xfb\x19\xfe\xb6h\xe3\xc7$4\xe6h\xe7_\x95=\x88~\x8d\x0eL\x04\xa6\x82\xdf\xae\xda\xb8'), chr(0b1100100) + '\145' + chr(9126 - 9027) + chr(1938 - 1827) + '\144' + '\x65')(chr(0b1110101) + chr(12810 - 12694) + '\x66' + '\x2d' + chr(398 - 342))):
laxD7jy5y7k1 = tq24Tk6UZ6u1.laxD7jy5y7k1
H2MQqAZeamNo = tq24Tk6UZ6u1.X1ZpHSxyKbHn
OeWW0F1dBPRQ = IDJ2eXGCBCDu.layers.dense(KLYceYPUcF5e, ehT0Px3KOsy9(chr(183 - 135) + '\157' + '\064' + chr(48) + chr(48), 20749 - 20741), use_bias=ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49), 0o10), activation=None, name=xafqLlk3kkUe(SXOLrMavuUCe(b'U\xea\xf9\x8b\xeb2\xc2\xbdf\xe3\xdf'), '\144' + '\145' + chr(0b1100011) + chr(919 - 808) + chr(0b1100100) + chr(0b1011 + 0o132))('\165' + '\x74' + chr(102) + chr(0b1010 + 0o43) + '\070'))
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reshape(OeWW0F1dBPRQ, [-ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110001), 8), laxD7jy5y7k1, laxD7jy5y7k1, H2MQqAZeamNo, CeyMIoSyrpkQ])
return OeWW0F1dBPRQ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
softmax_average_pooling_class_label_top
|
def softmax_average_pooling_class_label_top(body_output,
targets,
model_hparams,
vocab_size):
"""Loss for class label."""
del targets # unused arg
with tf.variable_scope(
"softmax_average_pooling_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)):
x = body_output
x = tf.reduce_mean(x, axis=1, keepdims=True)
return tf.layers.dense(x, vocab_size)
|
python
|
def softmax_average_pooling_class_label_top(body_output,
targets,
model_hparams,
vocab_size):
"""Loss for class label."""
del targets # unused arg
with tf.variable_scope(
"softmax_average_pooling_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)):
x = body_output
x = tf.reduce_mean(x, axis=1, keepdims=True)
return tf.layers.dense(x, vocab_size)
|
[
"def",
"softmax_average_pooling_class_label_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"with",
"tf",
".",
"variable_scope",
"(",
"\"softmax_average_pooling_onehot_class_label_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
")",
":",
"x",
"=",
"body_output",
"x",
"=",
"tf",
".",
"reduce_mean",
"(",
"x",
",",
"axis",
"=",
"1",
",",
"keepdims",
"=",
"True",
")",
"return",
"tf",
".",
"layers",
".",
"dense",
"(",
"x",
",",
"vocab_size",
")"
] |
Loss for class label.
|
[
"Loss",
"for",
"class",
"label",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1062-L1073
|
train
|
Loss for class label.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(111) + chr(1857 - 1807) + chr(0b110000) + chr(0b110110), 52564 - 52556), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(0b110 + 0o52) + chr(987 - 936), 12400 - 12392), ehT0Px3KOsy9(chr(1457 - 1409) + '\x6f' + chr(0b110010) + '\062' + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x31' + '\067' + chr(51), 0o10), ehT0Px3KOsy9(chr(402 - 354) + chr(7742 - 7631) + '\063' + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(621 - 573) + chr(4700 - 4589) + chr(0b110011) + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b111100 + 0o63) + '\063' + chr(1667 - 1614), 8), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(0b1101111) + chr(0b110001) + chr(0b11101 + 0o23) + '\x35', 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + chr(0b110111) + chr(0b110100), 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\x33' + '\x36', 8), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110010) + chr(49) + chr(50), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110010) + chr(0b110001 + 0o0) + '\062', 8), ehT0Px3KOsy9(chr(1306 - 1258) + chr(0b1010100 + 0o33) + chr(51) + '\064' + chr(53), 34967 - 34959), ehT0Px3KOsy9(chr(0b101111 + 0o1) + '\x6f' + chr(0b111 + 0o53) + chr(1932 - 1881) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(263 - 210) + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b11111 + 0o21) + chr(0b111010 + 0o65) + chr(50) + '\067' + chr(0b11 + 0o56), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(2191 - 2141) + '\x31', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(352 - 301) + '\x34' + '\x36', 0o10), ehT0Px3KOsy9(chr(2239 - 2191) + chr(0b1101111) + chr(182 - 131) + '\067' + '\x36', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(48) + '\065', 25228 - 25220), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(0b110110) + '\067', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(732 - 681) + chr(51) + chr(52), 0b1000), ehT0Px3KOsy9(chr(353 - 305) + chr(0b1010100 + 0o33) + '\x32' + '\062' + '\x34', 0b1000), ehT0Px3KOsy9(chr(2122 - 2074) + '\157' + '\061' + chr(0b0 + 0o61) + '\x33', 0b1000), ehT0Px3KOsy9(chr(1263 - 1215) + '\x6f' + chr(1152 - 1101) + chr(48) + '\064', 0b1000), ehT0Px3KOsy9(chr(2004 - 1956) + chr(0b1101111) + chr(0b11 + 0o56) + chr(50) + '\x33', 0o10), ehT0Px3KOsy9('\060' + '\x6f' + '\x33' + chr(0b10001 + 0o40) + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1100010 + 0o15) + '\x33' + chr(0b101111 + 0o1) + chr(2048 - 1999), 11802 - 11794), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(111) + chr(414 - 365) + chr(0b100011 + 0o20) + '\061', 0o10), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(7325 - 7214) + chr(0b100100 + 0o17) + chr(1314 - 1263) + '\060', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1954 - 1903) + chr(744 - 694) + chr(0b10111 + 0o36), 51642 - 51634), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\062' + chr(955 - 906) + '\065', 0o10), ehT0Px3KOsy9(chr(1497 - 1449) + '\x6f' + '\x32' + chr(0b11001 + 0o33) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(9503 - 9392) + chr(513 - 462) + chr(1583 - 1531) + chr(0b110011), 23042 - 23034), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + '\064' + chr(49), 59207 - 59199), ehT0Px3KOsy9(chr(0b11001 + 0o27) + '\157' + chr(49) + '\060' + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(181 - 133) + '\157' + chr(0b10101 + 0o36) + chr(2155 - 2103) + chr(54), 8), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(49) + chr(48), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b1101 + 0o45) + chr(52) + chr(49), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b100111 + 0o110) + '\063' + '\066' + chr(0b1011 + 0o50), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b11110 + 0o121) + chr(0b110101) + chr(770 - 722), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xa8'), chr(0b1100100) + chr(0b111010 + 0o53) + '\143' + chr(111) + chr(100) + chr(5963 - 5862))(chr(1392 - 1275) + chr(0b1110100) + '\146' + chr(45) + chr(0b11011 + 0o35)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def ESIkx9jSR_C2(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf0\xa2\xe4\x12\x7f\x7fq \xe3m\xe9\xe5\x95\xfe'), chr(0b1100100) + '\x65' + chr(0b1100011) + chr(0b1000000 + 0o57) + '\x64' + chr(0b1100101))(chr(0b1110101) + chr(116) + '\146' + '\x2d' + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xf5\xac\xf0\x0fs|e\x1a\xddh\xef\xf8\x84\xfcH\xe6a\x89\x91\xe9\x1e\xb2\xf3\x14\xab=\xac\xa96\xac\x8a\xee\xcc\x82\x922\xb9d\xc0\x91\xe3\xaf\xc9\x16qy|)\xd5j\xf3\xd5\xc0\xffr\x9cu'), '\144' + chr(101) + chr(99) + '\x6f' + '\x64' + chr(0b1011110 + 0o7))('\x75' + chr(4418 - 4302) + chr(0b101001 + 0o75) + '\x2d' + '\x38') % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf7\xb9\xf9\x02FS..\xd8v\xce\xc6'), chr(0b101000 + 0o74) + chr(1757 - 1656) + chr(99) + '\157' + '\x64' + '\x65')(chr(0b1100000 + 0o25) + '\164' + '\146' + chr(0b101101) + chr(1321 - 1265))))):
OeWW0F1dBPRQ = KLYceYPUcF5e
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reduce_mean(OeWW0F1dBPRQ, axis=ehT0Px3KOsy9(chr(1631 - 1583) + chr(7231 - 7120) + chr(1140 - 1091), 0b1000), keepdims=ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110001), 8))
return xafqLlk3kkUe(IDJ2eXGCBCDu.layers, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe2\xa6\xf8\x08{'), chr(9686 - 9586) + chr(0b111110 + 0o47) + chr(0b1100010 + 0o1) + '\157' + chr(2205 - 2105) + chr(0b1011111 + 0o6))('\x75' + chr(2011 - 1895) + chr(102) + chr(1984 - 1939) + chr(0b111000)))(OeWW0F1dBPRQ, CeyMIoSyrpkQ)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
softmax_last_timestep_class_label_top
|
def softmax_last_timestep_class_label_top(body_output,
targets,
model_hparams,
vocab_size):
"""Loss for class label."""
del targets # unused arg
with tf.variable_scope(
"softmax_last_timestep_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)):
x = body_output
x = tf.expand_dims(x[:, -1], 1) # Pick the last timestep
return tf.layers.dense(x, vocab_size)
|
python
|
def softmax_last_timestep_class_label_top(body_output,
targets,
model_hparams,
vocab_size):
"""Loss for class label."""
del targets # unused arg
with tf.variable_scope(
"softmax_last_timestep_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)):
x = body_output
x = tf.expand_dims(x[:, -1], 1) # Pick the last timestep
return tf.layers.dense(x, vocab_size)
|
[
"def",
"softmax_last_timestep_class_label_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"with",
"tf",
".",
"variable_scope",
"(",
"\"softmax_last_timestep_onehot_class_label_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
")",
":",
"x",
"=",
"body_output",
"x",
"=",
"tf",
".",
"expand_dims",
"(",
"x",
"[",
":",
",",
"-",
"1",
"]",
",",
"1",
")",
"# Pick the last timestep",
"return",
"tf",
".",
"layers",
".",
"dense",
"(",
"x",
",",
"vocab_size",
")"
] |
Loss for class label.
|
[
"Loss",
"for",
"class",
"label",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1076-L1087
|
train
|
Loss for class label.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b10011 + 0o134) + chr(576 - 525) + chr(2731 - 2677) + chr(49), 0o10), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(0b1010101 + 0o32) + chr(1636 - 1585) + chr(0b11011 + 0o33), 0b1000), ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\x6f' + '\x32' + chr(49) + chr(0b100001 + 0o17), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(7268 - 7157) + chr(1319 - 1270) + chr(48), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1110 + 0o141) + '\062' + '\063' + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(1150 - 1098) + chr(53), 0b1000), ehT0Px3KOsy9(chr(1721 - 1673) + '\x6f' + chr(1985 - 1930) + '\x33', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x33' + chr(1537 - 1487) + chr(0b10100 + 0o37), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110011 + 0o0) + '\x31' + chr(0b10011 + 0o36), 39818 - 39810), ehT0Px3KOsy9('\060' + '\157' + '\061' + chr(1204 - 1151) + '\065', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + '\062' + chr(0b10011 + 0o44) + '\x31', 57091 - 57083), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110011) + chr(0b110110), 8), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(0b1011000 + 0o27) + '\x33' + '\x35' + chr(480 - 429), 6588 - 6580), ehT0Px3KOsy9(chr(1775 - 1727) + chr(0b10100 + 0o133) + chr(1682 - 1633) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(331 - 283) + '\x6f' + chr(50) + '\061' + chr(50), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b111100 + 0o63) + chr(50) + chr(51) + chr(0b110000 + 0o6), 20976 - 20968), ehT0Px3KOsy9(chr(1771 - 1723) + chr(111) + '\062' + chr(0b110000) + chr(52), 19315 - 19307), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110011) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(0b110101) + chr(0b110010), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110010) + '\x34', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + '\061' + '\x37' + chr(0b101011 + 0o14), 58340 - 58332), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(0b1101111) + chr(51) + '\067' + chr(51), 38693 - 38685), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b0 + 0o63) + '\060' + chr(0b110 + 0o54), ord("\x08")), ehT0Px3KOsy9(chr(1103 - 1055) + chr(0b1101111) + '\062' + chr(53) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(0b100 + 0o54) + '\x6f' + '\062' + chr(0b11111 + 0o27) + chr(0b100110 + 0o14), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1010011 + 0o34) + chr(0b10011 + 0o37) + '\065' + chr(419 - 371), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1010000 + 0o37) + '\066' + '\x35', 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\x32' + chr(0b110011) + '\x33', 31682 - 31674), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011 + 0o2), 8621 - 8613), ehT0Px3KOsy9('\060' + chr(111) + chr(1411 - 1362) + chr(0b110011) + chr(2247 - 2194), 0b1000), ehT0Px3KOsy9('\060' + chr(11146 - 11035) + chr(0b110011) + chr(0b110111) + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(712 - 664) + '\x6f' + chr(51) + '\x30' + chr(0b11101 + 0o32), ord("\x08")), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(135 - 24) + '\061' + '\060' + chr(52), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\061' + chr(51) + chr(2330 - 2280), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b100 + 0o56) + '\063' + chr(1794 - 1740), 8), ehT0Px3KOsy9(chr(0b110000) + chr(3013 - 2902) + chr(635 - 585) + chr(0b110100) + chr(0b110101), 25183 - 25175), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x36' + '\067', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x32' + '\062' + chr(1427 - 1372), ord("\x08")), ehT0Px3KOsy9(chr(2286 - 2238) + chr(0b1101111) + chr(0b110100) + '\060', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\063' + '\061' + '\063', 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(8310 - 8199) + chr(0b100010 + 0o23) + '\x30', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf'), '\144' + '\x65' + chr(0b100110 + 0o75) + '\x6f' + chr(0b1111 + 0o125) + chr(0b1100101))(chr(12018 - 11901) + chr(0b100000 + 0o124) + chr(0b110010 + 0o64) + '\x2d' + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def zCVG3mWcnQw1(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\x87\x7f\xfaa\xf8GW\xb6\x96\x1e\xdaq\x9bh'), '\x64' + chr(0b1100101) + chr(99) + '\157' + '\x64' + '\145')(chr(0b1110101) + '\164' + '\x66' + '\055' + chr(0b110010 + 0o6)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x82q\xee|\xf4DC\x8c\xa5\x0c\xcaj\xb4y\xe3\xf3\xe1i^q\x0f\xa4\xad\xc4\xd7\xc0Z\xefZ>\x90\xa67)\xc1\xc9V\xf5E\xa7\xaes\xe7l\xf8IR\xa7\xb02\x9cz\xb4(\xee'), chr(100) + '\x65' + chr(0b110000 + 0o63) + '\157' + chr(0b1100100) + chr(0b1100101))('\165' + chr(2993 - 2877) + chr(0b1100110) + chr(411 - 366) + '\x38') % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\x80d\xe7q\xc1k\x08\xb8\xad\x05\xfdR'), chr(100) + chr(8847 - 8746) + '\143' + chr(0b1101111) + chr(0b1000101 + 0o37) + chr(259 - 158))(chr(7983 - 7866) + chr(116) + chr(0b1100110) + '\055' + chr(0b111000))))):
OeWW0F1dBPRQ = KLYceYPUcF5e
OeWW0F1dBPRQ = IDJ2eXGCBCDu.expand_dims(OeWW0F1dBPRQ[:, -ehT0Px3KOsy9('\x30' + '\x6f' + chr(49), 50888 - 50880)], ehT0Px3KOsy9(chr(0b110000) + chr(0b1011111 + 0o20) + '\x31', 8))
return xafqLlk3kkUe(IDJ2eXGCBCDu.layers, xafqLlk3kkUe(SXOLrMavuUCe(b'\x95{\xe6{\xfc'), '\x64' + chr(101) + '\x63' + chr(0b1001000 + 0o47) + chr(7023 - 6923) + chr(4350 - 4249))(chr(0b110111 + 0o76) + chr(0b10110 + 0o136) + '\146' + chr(0b11100 + 0o21) + chr(0b111 + 0o61)))(OeWW0F1dBPRQ, CeyMIoSyrpkQ)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
softmax_max_pooling_class_label_top
|
def softmax_max_pooling_class_label_top(body_output,
targets,
model_hparams,
vocab_size):
"""Loss for class label."""
del targets # unused arg
with tf.variable_scope(
"softmax_max_pooling_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)):
x = body_output
x = tf.reduce_max(x, axis=1, keepdims=True)
return tf.layers.dense(x, vocab_size)
|
python
|
def softmax_max_pooling_class_label_top(body_output,
targets,
model_hparams,
vocab_size):
"""Loss for class label."""
del targets # unused arg
with tf.variable_scope(
"softmax_max_pooling_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)):
x = body_output
x = tf.reduce_max(x, axis=1, keepdims=True)
return tf.layers.dense(x, vocab_size)
|
[
"def",
"softmax_max_pooling_class_label_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"with",
"tf",
".",
"variable_scope",
"(",
"\"softmax_max_pooling_onehot_class_label_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
")",
":",
"x",
"=",
"body_output",
"x",
"=",
"tf",
".",
"reduce_max",
"(",
"x",
",",
"axis",
"=",
"1",
",",
"keepdims",
"=",
"True",
")",
"return",
"tf",
".",
"layers",
".",
"dense",
"(",
"x",
",",
"vocab_size",
")"
] |
Loss for class label.
|
[
"Loss",
"for",
"class",
"label",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1090-L1101
|
train
|
Loss for class label.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10111 + 0o33) + '\x36' + chr(0b1110 + 0o51), 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\x31' + '\060' + chr(0b110010 + 0o2), ord("\x08")), ehT0Px3KOsy9(chr(0b11 + 0o55) + '\157' + chr(54) + chr(51), 0o10), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(4932 - 4821) + chr(0b11100 + 0o25) + chr(48) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(812 - 763) + chr(1366 - 1313) + chr(0b110011), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(1882 - 1833) + chr(51) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1872 - 1821) + chr(0b101001 + 0o11), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\061' + chr(51), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b11001 + 0o126) + '\x31' + chr(0b110100) + chr(0b11111 + 0o30), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b10000 + 0o42) + chr(1782 - 1734) + '\x37', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110011) + chr(50) + chr(0b10100 + 0o41), 52546 - 52538), ehT0Px3KOsy9(chr(837 - 789) + '\x6f' + '\x36' + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(1419 - 1371) + chr(2531 - 2420) + '\061' + '\062' + '\062', 49671 - 49663), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(111) + chr(0b110010) + chr(49) + chr(2083 - 2028), 33583 - 33575), ehT0Px3KOsy9(chr(48) + chr(0b111111 + 0o60) + chr(2266 - 2217) + chr(0b110100), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101 + 0o142) + chr(51) + chr(1235 - 1187), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(903 - 850) + '\x37', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(594 - 544) + '\x34', 0o10), ehT0Px3KOsy9(chr(340 - 292) + chr(111) + chr(0b110011) + chr(0b110011) + chr(0b110111), 38011 - 38003), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100000 + 0o17) + chr(0b10111 + 0o34) + chr(1267 - 1219) + '\062', 62072 - 62064), ehT0Px3KOsy9(chr(1156 - 1108) + chr(0b1100111 + 0o10) + chr(0b11001 + 0o30) + chr(0b110111) + chr(49), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(556 - 506) + '\065' + '\064', 53725 - 53717), ehT0Px3KOsy9('\060' + chr(0b110000 + 0o77) + chr(0b110001) + '\x37' + chr(0b10011 + 0o36), 8), ehT0Px3KOsy9(chr(0b110000) + chr(5013 - 4902) + chr(53) + chr(749 - 699), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1000 + 0o147) + '\x33' + chr(1078 - 1025), 0o10), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(0b1101111) + chr(49) + chr(0b110110) + chr(0b101 + 0o62), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x32' + '\x37' + chr(892 - 837), 50198 - 50190), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1101111) + chr(50) + chr(0b110000) + chr(2046 - 1994), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(50) + chr(1918 - 1867) + '\x30', 17478 - 17470), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062' + '\067' + chr(1036 - 985), 30755 - 30747), ehT0Px3KOsy9(chr(48) + '\157' + chr(50) + chr(1398 - 1346) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110100) + '\x33', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x31' + chr(0b100101 + 0o22) + '\x37', 6195 - 6187), ehT0Px3KOsy9(chr(48) + chr(111) + chr(55) + chr(0b101101 + 0o4), 0o10), ehT0Px3KOsy9('\x30' + chr(0b101001 + 0o106) + '\063' + chr(192 - 143) + chr(0b10000 + 0o44), 59666 - 59658), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(0b10000 + 0o41) + chr(0b110001 + 0o0), 58517 - 58509), ehT0Px3KOsy9(chr(48) + chr(0b1010010 + 0o35) + chr(50) + chr(0b110001) + chr(106 - 57), 0b1000), ehT0Px3KOsy9(chr(0b1001 + 0o47) + chr(0b1101 + 0o142) + chr(54) + chr(429 - 379), 0o10), ehT0Px3KOsy9(chr(969 - 921) + chr(5569 - 5458) + chr(0b110100) + chr(0b11000 + 0o30), 0o10), ehT0Px3KOsy9(chr(48) + chr(3861 - 3750) + chr(0b11 + 0o57) + chr(51) + chr(0b110101), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(2047 - 1999) + chr(0b1100000 + 0o17) + chr(1938 - 1885) + '\060', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x99'), chr(100) + chr(101) + '\x63' + '\x6f' + chr(0b1100100) + chr(0b1100101))(chr(0b1111 + 0o146) + chr(0b11001 + 0o133) + '\146' + chr(45) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def kipLeNz9LwPg(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc1%A/\n;vn8\xf4\xb6^"\xea'), chr(100) + '\145' + '\x63' + chr(3556 - 3445) + chr(0b1100000 + 0o4) + chr(0b10101 + 0o120))(chr(117) + '\x74' + '\146' + chr(45) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xc4+U2\x068bT\n\xe6\xadn"\xe0ctFz\xb8\x9d\x16\xc2\x1cBr\x03\xe0\xe7\xf4\xda\xd6P U\xad\xab\xec^\x7f\xf0\xd8 R*\x02-cTB\xe3\x8a\x146'), chr(3440 - 3340) + chr(0b101011 + 0o72) + chr(0b1000110 + 0o35) + chr(1965 - 1854) + chr(0b1100100) + chr(0b1100101))(chr(13355 - 13238) + chr(0b100110 + 0o116) + chr(102) + chr(45) + chr(0b111000)) % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc6>\\?3\x17)`\x03\xef\x91}'), '\144' + chr(0b1100101) + chr(0b1010011 + 0o20) + '\157' + chr(8583 - 8483) + chr(0b1100101))(chr(0b1110101) + chr(116) + '\146' + chr(0b101101) + chr(0b111000))))):
OeWW0F1dBPRQ = KLYceYPUcF5e
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reduce_max(OeWW0F1dBPRQ, axis=ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31', ord("\x08")), keepdims=ehT0Px3KOsy9('\x30' + chr(10897 - 10786) + chr(1171 - 1122), 8))
return xafqLlk3kkUe(IDJ2eXGCBCDu.layers, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd3!]5\x0e'), chr(100) + chr(3317 - 3216) + '\x63' + chr(0b1101111) + chr(100) + '\145')(chr(0b1110101) + chr(10687 - 10571) + '\x66' + '\x2d' + '\070'))(OeWW0F1dBPRQ, CeyMIoSyrpkQ)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
symbol_top
|
def symbol_top(body_output, targets, model_hparams, vocab_size):
"""Generate logits.
Args:
body_output: A Tensor with shape
[batch, p0, p1, model_hparams.hidden_size].
targets: Unused.
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
logits: A Tensor with shape [batch, p0, p1, ?, vocab_size].
"""
del targets # unused arg
if model_hparams.shared_embedding_and_softmax_weights:
scope_name = "shared"
reuse = tf.AUTO_REUSE
else:
scope_name = "softmax"
reuse = False
with tf.variable_scope(scope_name, reuse=reuse):
body_output_shape = common_layers.shape_list(body_output)
var = get_weights(model_hparams, vocab_size, body_output_shape[-1])
if (model_hparams.factored_logits and
model_hparams.mode == tf.estimator.ModeKeys.TRAIN):
# insert channels dimension
body_output = tf.expand_dims(body_output, 3)
return common_layers.FactoredTensor(body_output, var)
else:
body_output = tf.reshape(body_output, [-1, body_output_shape[-1]])
logits = tf.matmul(body_output, var, transpose_b=True)
return tf.reshape(logits,
body_output_shape[:-1] + [1, vocab_size])
|
python
|
def symbol_top(body_output, targets, model_hparams, vocab_size):
"""Generate logits.
Args:
body_output: A Tensor with shape
[batch, p0, p1, model_hparams.hidden_size].
targets: Unused.
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
logits: A Tensor with shape [batch, p0, p1, ?, vocab_size].
"""
del targets # unused arg
if model_hparams.shared_embedding_and_softmax_weights:
scope_name = "shared"
reuse = tf.AUTO_REUSE
else:
scope_name = "softmax"
reuse = False
with tf.variable_scope(scope_name, reuse=reuse):
body_output_shape = common_layers.shape_list(body_output)
var = get_weights(model_hparams, vocab_size, body_output_shape[-1])
if (model_hparams.factored_logits and
model_hparams.mode == tf.estimator.ModeKeys.TRAIN):
# insert channels dimension
body_output = tf.expand_dims(body_output, 3)
return common_layers.FactoredTensor(body_output, var)
else:
body_output = tf.reshape(body_output, [-1, body_output_shape[-1]])
logits = tf.matmul(body_output, var, transpose_b=True)
return tf.reshape(logits,
body_output_shape[:-1] + [1, vocab_size])
|
[
"def",
"symbol_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"if",
"model_hparams",
".",
"shared_embedding_and_softmax_weights",
":",
"scope_name",
"=",
"\"shared\"",
"reuse",
"=",
"tf",
".",
"AUTO_REUSE",
"else",
":",
"scope_name",
"=",
"\"softmax\"",
"reuse",
"=",
"False",
"with",
"tf",
".",
"variable_scope",
"(",
"scope_name",
",",
"reuse",
"=",
"reuse",
")",
":",
"body_output_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"body_output",
")",
"var",
"=",
"get_weights",
"(",
"model_hparams",
",",
"vocab_size",
",",
"body_output_shape",
"[",
"-",
"1",
"]",
")",
"if",
"(",
"model_hparams",
".",
"factored_logits",
"and",
"model_hparams",
".",
"mode",
"==",
"tf",
".",
"estimator",
".",
"ModeKeys",
".",
"TRAIN",
")",
":",
"# insert channels dimension",
"body_output",
"=",
"tf",
".",
"expand_dims",
"(",
"body_output",
",",
"3",
")",
"return",
"common_layers",
".",
"FactoredTensor",
"(",
"body_output",
",",
"var",
")",
"else",
":",
"body_output",
"=",
"tf",
".",
"reshape",
"(",
"body_output",
",",
"[",
"-",
"1",
",",
"body_output_shape",
"[",
"-",
"1",
"]",
"]",
")",
"logits",
"=",
"tf",
".",
"matmul",
"(",
"body_output",
",",
"var",
",",
"transpose_b",
"=",
"True",
")",
"return",
"tf",
".",
"reshape",
"(",
"logits",
",",
"body_output_shape",
"[",
":",
"-",
"1",
"]",
"+",
"[",
"1",
",",
"vocab_size",
"]",
")"
] |
Generate logits.
Args:
body_output: A Tensor with shape
[batch, p0, p1, model_hparams.hidden_size].
targets: Unused.
model_hparams: HParams, model hyperparmeters.
vocab_size: int, vocabulary size.
Returns:
logits: A Tensor with shape [batch, p0, p1, ?, vocab_size].
|
[
"Generate",
"logits",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1105-L1137
|
train
|
Generate logits.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + '\157' + '\x32' + chr(0b110001) + chr(53), 40951 - 40943), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(53) + chr(0b10110 + 0o41), 18024 - 18016), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(914 - 863) + chr(1685 - 1633) + chr(1420 - 1372), 0b1000), ehT0Px3KOsy9(chr(948 - 900) + chr(111) + chr(50) + chr(0b110010) + chr(54), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + chr(571 - 521) + '\064', 55935 - 55927), ehT0Px3KOsy9(chr(48) + chr(0b100100 + 0o113) + chr(0b1100 + 0o45) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b101011 + 0o5) + '\x6f' + chr(0b11100 + 0o25) + chr(51) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1213 - 1163) + chr(0b110010) + '\066', 8), ehT0Px3KOsy9(chr(48) + chr(0b1000000 + 0o57) + chr(0b110010) + chr(0b1011 + 0o51) + '\x32', 31307 - 31299), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011101 + 0o22) + chr(504 - 455) + chr(0b100100 + 0o14) + chr(0b110000), 0b1000), ehT0Px3KOsy9(chr(2168 - 2120) + chr(111) + chr(50) + '\x35' + '\x34', 0b1000), ehT0Px3KOsy9(chr(0b10010 + 0o36) + chr(111) + '\x37' + '\060', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1001 + 0o146) + '\x33' + '\x30' + chr(0b11110 + 0o23), 0o10), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1101111) + chr(0b110011) + '\x35' + chr(1405 - 1353), 0b1000), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(8530 - 8419) + '\x32' + chr(0b110 + 0o53) + '\x32', 25136 - 25128), ehT0Px3KOsy9(chr(605 - 557) + chr(7480 - 7369) + chr(0b101001 + 0o10) + chr(0b110010) + chr(1488 - 1439), 24842 - 24834), ehT0Px3KOsy9(chr(1629 - 1581) + chr(111) + chr(49) + chr(0b111 + 0o53) + chr(1050 - 995), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(1301 - 1253) + '\x6f' + chr(0b110001) + chr(0b1010 + 0o47) + chr(1634 - 1580), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1100011 + 0o14) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + '\061' + chr(2052 - 2003) + '\066', 8), ehT0Px3KOsy9(chr(0b100100 + 0o14) + '\x6f' + chr(0b110011) + chr(0b110001) + chr(0b110100), 38642 - 38634), ehT0Px3KOsy9(chr(0b1101 + 0o43) + '\157' + '\062' + chr(0b110110) + chr(48), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(2597 - 2546) + chr(0b110001) + chr(51), 0b1000), ehT0Px3KOsy9('\060' + chr(6161 - 6050) + chr(2248 - 2199) + chr(0b101 + 0o62), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x31' + chr(0b110000) + '\x31', 0o10), ehT0Px3KOsy9('\x30' + chr(3394 - 3283) + '\062' + chr(0b10111 + 0o36) + chr(0b101101 + 0o10), 48952 - 48944), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(0b1101111) + '\x32' + chr(0b110001), 57670 - 57662), ehT0Px3KOsy9('\060' + chr(0b0 + 0o157) + chr(50) + chr(998 - 949) + chr(49), 0o10), ehT0Px3KOsy9(chr(0b11101 + 0o23) + chr(0b10010 + 0o135) + '\063' + chr(54) + chr(1517 - 1466), 0b1000), ehT0Px3KOsy9(chr(0b1100 + 0o44) + '\x6f' + chr(2091 - 2039) + chr(0b100110 + 0o12), 57020 - 57012), ehT0Px3KOsy9(chr(1269 - 1221) + chr(6099 - 5988) + '\x31' + '\x36' + chr(0b10 + 0o65), 22762 - 22754), ehT0Px3KOsy9(chr(0b101 + 0o53) + chr(0b10000 + 0o137) + chr(52) + chr(1826 - 1774), 0b1000), ehT0Px3KOsy9('\x30' + chr(4844 - 4733) + '\x32' + chr(49) + chr(0b110010), 8), ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(1489 - 1378) + '\067' + '\063', 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(49) + chr(54) + chr(1844 - 1791), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b101110 + 0o5) + chr(0b110011) + '\x34', 0o10), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(9696 - 9585) + '\x31' + chr(1279 - 1225) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(2056 - 2008) + chr(0b1101111) + chr(0b11001 + 0o34) + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(639 - 591) + chr(985 - 874) + chr(51) + chr(0b11111 + 0o30) + chr(0b110011), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + '\157' + chr(0b110101) + '\060', 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xa3'), chr(0b10100 + 0o120) + chr(0b1100101) + '\143' + chr(0b1000110 + 0o51) + chr(0b1011101 + 0o7) + chr(0b1100101))(chr(0b1110101) + chr(12378 - 12262) + '\x66' + chr(1800 - 1755) + chr(0b11011 + 0o35)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def ZdqK7JN8NC8f(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
if xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfcy\x81\xfb\xeca0\xa6\xeer\xe5\t'), '\144' + chr(101) + '\143' + '\157' + '\x64' + chr(101))(chr(0b1110101) + chr(7124 - 7008) + '\x66' + '\055' + chr(0b111000))):
kP9QkIaIzxt1 = xafqLlk3kkUe(SXOLrMavuUCe(b'\xfeG\x81\xe4\xf1l'), chr(716 - 616) + chr(101) + chr(99) + '\157' + '\144' + '\145')('\x75' + chr(0b1110100) + chr(102) + '\055' + '\x38')
pmC5wdSFgdFj = IDJ2eXGCBCDu.AUTO_REUSE
else:
kP9QkIaIzxt1 = xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe@\x86\xe2\xf9i%'), chr(100) + chr(2887 - 2786) + chr(2905 - 2806) + chr(111) + chr(100) + '\x65')('\x75' + chr(116) + '\146' + chr(0b100000 + 0o15) + chr(1585 - 1529))
pmC5wdSFgdFj = ehT0Px3KOsy9(chr(662 - 614) + chr(111) + '\x30', 59949 - 59941)
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfbN\x92\xff\xf5j1\xf3\xfd3\xcfW\xb1\x89'), chr(0b1100100) + chr(1340 - 1239) + '\143' + chr(11663 - 11552) + chr(100) + chr(0b1100101))(chr(0b100001 + 0o124) + chr(1055 - 939) + '\x66' + chr(45) + '\x38'))(kP9QkIaIzxt1, reuse=pmC5wdSFgdFj):
oknqTBFY6iX3 = jSKPaHwSAfVv.shape_list(KLYceYPUcF5e)
l38lb8xQZNsE = WZaPMPa_FY26(tq24Tk6UZ6u1, CeyMIoSyrpkQ, oknqTBFY6iX3[-ehT0Px3KOsy9(chr(48) + chr(111) + '\061', 0b1000)])
if xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdeE\x97\xa2\xffj\x19\xc7\xd4\x0e\xc1v'), '\144' + chr(101) + chr(0b1100011) + chr(111) + '\x64' + chr(9311 - 9210))('\x75' + '\164' + chr(0b1100110) + '\055' + chr(56))) and xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe0@\x84\xf3'), chr(0b1100100) + chr(101) + chr(0b1100011) + '\157' + chr(7796 - 7696) + '\145')(chr(0b1110101) + chr(0b1001 + 0o153) + chr(621 - 519) + chr(1574 - 1529) + '\x38')) == xafqLlk3kkUe(IDJ2eXGCBCDu.estimator.ModeKeys, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd9}\xa1\xdf\xda'), chr(0b110110 + 0o56) + '\145' + '\143' + chr(0b10101 + 0o132) + chr(0b1100100) + chr(8580 - 8479))(chr(117) + '\x74' + '\x66' + '\055' + chr(0b11010 + 0o36))):
KLYceYPUcF5e = IDJ2eXGCBCDu.expand_dims(KLYceYPUcF5e, ehT0Px3KOsy9(chr(0b110000) + chr(2369 - 2258) + chr(467 - 416), 0o10))
return xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcbN\x83\xe2\xfbz8\xf2\xf6%\xc2K\xae\x9e'), chr(100) + chr(0b110100 + 0o61) + '\x63' + chr(0b1101111) + chr(100) + '\145')(chr(8740 - 8623) + chr(0b10000 + 0o144) + chr(102) + chr(1526 - 1481) + '\x38'))(KLYceYPUcF5e, l38lb8xQZNsE)
else:
KLYceYPUcF5e = IDJ2eXGCBCDu.reshape(KLYceYPUcF5e, [-ehT0Px3KOsy9(chr(83 - 35) + '\x6f' + chr(0b1111 + 0o42), 8), oknqTBFY6iX3[-ehT0Px3KOsy9(chr(0b11 + 0o55) + '\x6f' + '\061', 8)]])
wF9nmvjsKjYM = IDJ2eXGCBCDu.matmul(KLYceYPUcF5e, l38lb8xQZNsE, transpose_b=ehT0Px3KOsy9(chr(0b110000) + chr(0b1100100 + 0o13) + chr(0b1011 + 0o46), 8))
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xffJ\x93\xfe\xf5x8'), chr(100) + chr(2542 - 2441) + chr(0b1100011) + chr(0b1101111) + chr(0b11001 + 0o113) + chr(0b110010 + 0o63))(chr(117) + chr(116) + chr(0b1100110) + chr(45) + chr(0b10 + 0o66)))(wF9nmvjsKjYM, oknqTBFY6iX3[:-ehT0Px3KOsy9(chr(48) + chr(3854 - 3743) + chr(1877 - 1828), 8)] + [ehT0Px3KOsy9('\060' + chr(111) + chr(49), 8), CeyMIoSyrpkQ])
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
video_top
|
def video_top(body_output, targets, model_hparams, vocab_size):
"""Top transformation for video."""
del targets # unused arg
num_channels = model_hparams.problem.num_channels
shape = common_layers.shape_list(body_output)
reshape_shape = shape[:-1] + [num_channels, vocab_size]
res = tf.reshape(body_output, reshape_shape)
# Calculate argmax so as to have a summary with the produced images.
x = tf.argmax(tf.reshape(res, [-1, vocab_size]), axis=-1)
x = tf.reshape(x, shape[:-1] + [num_channels])
common_video.gif_summary("results", x, max_outputs=1)
return res
|
python
|
def video_top(body_output, targets, model_hparams, vocab_size):
"""Top transformation for video."""
del targets # unused arg
num_channels = model_hparams.problem.num_channels
shape = common_layers.shape_list(body_output)
reshape_shape = shape[:-1] + [num_channels, vocab_size]
res = tf.reshape(body_output, reshape_shape)
# Calculate argmax so as to have a summary with the produced images.
x = tf.argmax(tf.reshape(res, [-1, vocab_size]), axis=-1)
x = tf.reshape(x, shape[:-1] + [num_channels])
common_video.gif_summary("results", x, max_outputs=1)
return res
|
[
"def",
"video_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
"# unused arg",
"num_channels",
"=",
"model_hparams",
".",
"problem",
".",
"num_channels",
"shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"body_output",
")",
"reshape_shape",
"=",
"shape",
"[",
":",
"-",
"1",
"]",
"+",
"[",
"num_channels",
",",
"vocab_size",
"]",
"res",
"=",
"tf",
".",
"reshape",
"(",
"body_output",
",",
"reshape_shape",
")",
"# Calculate argmax so as to have a summary with the produced images.",
"x",
"=",
"tf",
".",
"argmax",
"(",
"tf",
".",
"reshape",
"(",
"res",
",",
"[",
"-",
"1",
",",
"vocab_size",
"]",
")",
",",
"axis",
"=",
"-",
"1",
")",
"x",
"=",
"tf",
".",
"reshape",
"(",
"x",
",",
"shape",
"[",
":",
"-",
"1",
"]",
"+",
"[",
"num_channels",
"]",
")",
"common_video",
".",
"gif_summary",
"(",
"\"results\"",
",",
"x",
",",
"max_outputs",
"=",
"1",
")",
"return",
"res"
] |
Top transformation for video.
|
[
"Top",
"transformation",
"for",
"video",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1146-L1157
|
train
|
Top transformation for video.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(6720 - 6609) + chr(50) + chr(2112 - 2061) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(603 - 555) + '\157' + chr(0b110100) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(308 - 259) + chr(0b110110) + '\066', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110011) + chr(0b101000 + 0o12) + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b1101111) + chr(0b100111 + 0o12) + chr(523 - 468) + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(1445 - 1334) + '\x33' + chr(0b100 + 0o60) + chr(1265 - 1215), 56573 - 56565), ehT0Px3KOsy9(chr(1881 - 1833) + chr(111) + '\x33' + '\064', 50125 - 50117), ehT0Px3KOsy9('\060' + chr(2485 - 2374) + '\x33' + chr(444 - 394) + chr(49), 37629 - 37621), ehT0Px3KOsy9(chr(1254 - 1206) + chr(0b1101111) + '\x35' + '\x37', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1745 - 1696) + chr(1857 - 1804) + chr(48), 1576 - 1568), ehT0Px3KOsy9(chr(48) + '\157' + chr(49) + chr(0b100110 + 0o12) + '\063', 31789 - 31781), ehT0Px3KOsy9(chr(660 - 612) + chr(111) + chr(0b110010) + chr(2910 - 2856) + chr(52), 0o10), ehT0Px3KOsy9('\060' + chr(3146 - 3035) + chr(0b110010) + chr(2412 - 2358) + chr(0b11110 + 0o31), 0b1000), ehT0Px3KOsy9(chr(0b11011 + 0o25) + '\x6f' + chr(50) + '\066' + chr(2105 - 2054), 0o10), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(0b1101111) + chr(49) + '\065' + chr(598 - 550), 8), ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(0b1011101 + 0o22) + chr(49) + chr(120 - 72) + chr(0b11100 + 0o31), 25905 - 25897), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\061' + chr(0b101011 + 0o11) + '\061', 0o10), ehT0Px3KOsy9(chr(1904 - 1856) + chr(379 - 268) + chr(0b100100 + 0o17), 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\061' + chr(51) + '\067', 0o10), ehT0Px3KOsy9('\060' + chr(3200 - 3089) + chr(0b110011) + chr(922 - 868) + chr(49), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\061' + chr(52) + '\x37', 0b1000), ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(111) + chr(0b101111 + 0o2) + chr(0b110100), 61289 - 61281), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(3392 - 3281) + chr(50) + chr(0b110011) + '\x32', 0b1000), ehT0Px3KOsy9('\x30' + chr(5592 - 5481) + '\063' + chr(0b101100 + 0o13), ord("\x08")), ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(111) + chr(1047 - 998) + chr(1020 - 969) + chr(48), 12636 - 12628), ehT0Px3KOsy9(chr(1355 - 1307) + chr(0b1101111) + chr(1170 - 1121) + '\x30' + chr(0b110011 + 0o4), ord("\x08")), ehT0Px3KOsy9(chr(0b10 + 0o56) + '\157' + chr(1715 - 1666) + chr(50) + '\065', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b11000 + 0o34) + '\x30', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1238 - 1187) + chr(0b10001 + 0o43) + chr(0b10110 + 0o32), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(1249 - 1198) + '\x36' + '\x33', 0o10), ehT0Px3KOsy9(chr(904 - 856) + chr(111) + chr(2442 - 2392) + '\065' + chr(48), 0b1000), ehT0Px3KOsy9(chr(446 - 398) + chr(0b1101111) + chr(1604 - 1554) + '\063' + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(1166 - 1118) + '\157' + '\x30', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x36' + '\061', ord("\x08")), ehT0Px3KOsy9(chr(1610 - 1562) + chr(0b10110 + 0o131) + chr(0b110010) + '\062' + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(0b1011110 + 0o21) + chr(0b110001) + '\060' + chr(0b100001 + 0o20), 45437 - 45429), ehT0Px3KOsy9('\060' + chr(0b1101100 + 0o3) + '\061' + '\x33' + '\065', 33931 - 33923), ehT0Px3KOsy9('\x30' + chr(6553 - 6442) + '\x32' + chr(0b110010) + chr(0b11101 + 0o24), 22447 - 22439), ehT0Px3KOsy9('\x30' + chr(0b11111 + 0o120) + '\x31' + chr(50) + chr(0b10 + 0o62), 27495 - 27487), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110001) + '\061' + chr(49), 35384 - 35376)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1101101 + 0o2) + chr(0b110101) + '\x30', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7'), '\x64' + '\145' + chr(99) + chr(9847 - 9736) + chr(0b1001010 + 0o32) + '\145')('\165' + chr(0b1110100) + chr(102) + '\055' + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def CV8Z6qeH8WaV(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z
X1ZpHSxyKbHn = tq24Tk6UZ6u1.problem.X1ZpHSxyKbHn
nauYfLglTpcb = jSKPaHwSAfVv.shape_list(KLYceYPUcF5e)
TTb6rlTz2h_U = nauYfLglTpcb[:-ehT0Px3KOsy9(chr(283 - 235) + '\x6f' + '\x31', 0o10)] + [X1ZpHSxyKbHn, CeyMIoSyrpkQ]
MsbwfslwLjRO = IDJ2eXGCBCDu.reshape(KLYceYPUcF5e, TTb6rlTz2h_U)
OeWW0F1dBPRQ = IDJ2eXGCBCDu.argmax(IDJ2eXGCBCDu.reshape(MsbwfslwLjRO, [-ehT0Px3KOsy9(chr(48) + chr(0b10001 + 0o136) + chr(0b110000 + 0o1), 8), CeyMIoSyrpkQ]), axis=-ehT0Px3KOsy9(chr(48) + '\x6f' + '\x31', 8))
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reshape(OeWW0F1dBPRQ, nauYfLglTpcb[:-ehT0Px3KOsy9(chr(0b110000) + chr(0b11010 + 0o125) + '\x31', 8)] + [X1ZpHSxyKbHn])
xafqLlk3kkUe(feDooRjkbHzt, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfeH\x10\x8a\x0e\x97q\xbcN\x1a:'), '\144' + chr(0b1011010 + 0o13) + chr(0b10101 + 0o116) + '\157' + chr(0b1100100) + chr(1720 - 1619))('\x75' + '\164' + chr(0b1100110) + chr(45) + chr(0b1111 + 0o51)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xebD\x05\xa0\x11\x96o'), '\144' + '\x65' + chr(1832 - 1733) + chr(0b1101111) + '\144' + chr(4561 - 4460))('\165' + chr(0b1110100) + chr(102) + chr(0b11100 + 0o21) + chr(1830 - 1774)), OeWW0F1dBPRQ, max_outputs=ehT0Px3KOsy9(chr(48) + chr(0b110100 + 0o73) + chr(2008 - 1959), 8))
return MsbwfslwLjRO
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
video_l1_top
|
def video_l1_top(body_output, targets, model_hparams, vocab_size):
"""Top transformation for video."""
del targets, vocab_size # unused arg
num_channels = model_hparams.problem.num_channels
num_frames = model_hparams.video_num_target_frames
with tf.variable_scope("rgb"):
body_output_shape = common_layers.shape_list(body_output)
res = tf.layers.dense(body_output, num_channels * num_frames, name="cast")
res = tf.reshape(res, body_output_shape[:3] + [num_channels, num_frames])
res = tf.transpose(res, [0, 4, 1, 2, 3]) # Move frames next to batch.
if not tf.get_variable_scope().reuse:
res_argmax = res[:, -1, :, :, :]
tf.summary.image(
"result",
common_layers.tpu_safe_image_summary(res_argmax),
max_outputs=1)
return tf.expand_dims(res, axis=-1)
|
python
|
def video_l1_top(body_output, targets, model_hparams, vocab_size):
"""Top transformation for video."""
del targets, vocab_size # unused arg
num_channels = model_hparams.problem.num_channels
num_frames = model_hparams.video_num_target_frames
with tf.variable_scope("rgb"):
body_output_shape = common_layers.shape_list(body_output)
res = tf.layers.dense(body_output, num_channels * num_frames, name="cast")
res = tf.reshape(res, body_output_shape[:3] + [num_channels, num_frames])
res = tf.transpose(res, [0, 4, 1, 2, 3]) # Move frames next to batch.
if not tf.get_variable_scope().reuse:
res_argmax = res[:, -1, :, :, :]
tf.summary.image(
"result",
common_layers.tpu_safe_image_summary(res_argmax),
max_outputs=1)
return tf.expand_dims(res, axis=-1)
|
[
"def",
"video_l1_top",
"(",
"body_output",
",",
"targets",
",",
"model_hparams",
",",
"vocab_size",
")",
":",
"del",
"targets",
",",
"vocab_size",
"# unused arg",
"num_channels",
"=",
"model_hparams",
".",
"problem",
".",
"num_channels",
"num_frames",
"=",
"model_hparams",
".",
"video_num_target_frames",
"with",
"tf",
".",
"variable_scope",
"(",
"\"rgb\"",
")",
":",
"body_output_shape",
"=",
"common_layers",
".",
"shape_list",
"(",
"body_output",
")",
"res",
"=",
"tf",
".",
"layers",
".",
"dense",
"(",
"body_output",
",",
"num_channels",
"*",
"num_frames",
",",
"name",
"=",
"\"cast\"",
")",
"res",
"=",
"tf",
".",
"reshape",
"(",
"res",
",",
"body_output_shape",
"[",
":",
"3",
"]",
"+",
"[",
"num_channels",
",",
"num_frames",
"]",
")",
"res",
"=",
"tf",
".",
"transpose",
"(",
"res",
",",
"[",
"0",
",",
"4",
",",
"1",
",",
"2",
",",
"3",
"]",
")",
"# Move frames next to batch.",
"if",
"not",
"tf",
".",
"get_variable_scope",
"(",
")",
".",
"reuse",
":",
"res_argmax",
"=",
"res",
"[",
":",
",",
"-",
"1",
",",
":",
",",
":",
",",
":",
"]",
"tf",
".",
"summary",
".",
"image",
"(",
"\"result\"",
",",
"common_layers",
".",
"tpu_safe_image_summary",
"(",
"res_argmax",
")",
",",
"max_outputs",
"=",
"1",
")",
"return",
"tf",
".",
"expand_dims",
"(",
"res",
",",
"axis",
"=",
"-",
"1",
")"
] |
Top transformation for video.
|
[
"Top",
"transformation",
"for",
"video",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1160-L1176
|
train
|
Top transformation for video.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x32' + '\x32' + '\067', 9407 - 9399), ehT0Px3KOsy9(chr(0b110000) + chr(5517 - 5406) + chr(50) + chr(0b110001) + chr(1730 - 1681), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + '\063' + chr(49) + chr(1484 - 1429), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + chr(49) + '\064', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + '\065' + chr(1281 - 1227), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110011) + '\x32' + chr(0b110111), 20153 - 20145), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b11000 + 0o33) + chr(0b1001 + 0o50) + '\x33', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b10110 + 0o33) + '\064', 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b11011 + 0o26) + chr(53) + chr(52), 34702 - 34694), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(49) + chr(0b110011) + chr(1856 - 1808), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + '\x32' + chr(0b110001) + chr(1861 - 1813), 0o10), ehT0Px3KOsy9('\060' + chr(6440 - 6329) + chr(0b110011) + chr(54) + chr(463 - 415), ord("\x08")), ehT0Px3KOsy9(chr(0b100111 + 0o11) + '\157' + chr(1931 - 1880) + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1000110 + 0o51) + '\x32' + chr(55) + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(414 - 366) + '\x6f' + chr(2133 - 2078) + '\x34', 32207 - 32199), ehT0Px3KOsy9('\060' + chr(4542 - 4431) + chr(50) + chr(1548 - 1499), ord("\x08")), ehT0Px3KOsy9(chr(118 - 70) + '\157' + chr(49) + '\063' + '\062', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1449 - 1400) + chr(53) + chr(134 - 80), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\063' + '\x31' + '\065', ord("\x08")), ehT0Px3KOsy9(chr(0b11111 + 0o21) + '\x6f' + chr(1588 - 1539) + chr(0b101000 + 0o14) + '\067', ord("\x08")), ehT0Px3KOsy9(chr(1954 - 1906) + chr(111) + '\062' + chr(65 - 12) + chr(0b110 + 0o61), ord("\x08")), ehT0Px3KOsy9(chr(2164 - 2116) + chr(0b1101111) + '\x33' + '\063', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b11 + 0o56) + '\067' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010101 + 0o32) + chr(1992 - 1938) + chr(0b110100), 41453 - 41445), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + '\062' + '\x36', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(2398 - 2349) + '\x36' + '\066', ord("\x08")), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(111) + chr(0b110011) + chr(0b110001) + chr(0b110101), 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\063' + '\062' + chr(0b110111), 8), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110011) + chr(0b110100) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(0b1000 + 0o50) + '\157' + chr(0b110001) + chr(51) + chr(2071 - 2023), 8), ehT0Px3KOsy9(chr(0b110000) + chr(11644 - 11533) + chr(49) + chr(54) + chr(0b110001), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\061' + '\x34' + '\x32', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110100) + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(49) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(0b1101111) + chr(0b110010) + '\x31', 8), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b101101 + 0o6) + chr(0b100100 + 0o17) + chr(0b110000), 20924 - 20916), ehT0Px3KOsy9(chr(1585 - 1537) + '\x6f' + '\x32' + chr(1736 - 1688) + chr(49), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\062' + '\060' + chr(53), 39592 - 39584), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + chr(0b110001) + chr(2427 - 2374), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + '\x33' + '\x30' + '\x36', ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(53) + chr(0b11100 + 0o24), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xb9'), '\144' + chr(0b1010010 + 0o23) + chr(0b1001 + 0o132) + chr(0b1101111) + chr(100) + chr(101))('\x75' + chr(5452 - 5336) + chr(0b11001 + 0o115) + chr(956 - 911) + chr(3065 - 3009)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def gfSpdw7Luh5y(KLYceYPUcF5e, xIEmRseySp3z, tq24Tk6UZ6u1, CeyMIoSyrpkQ):
del xIEmRseySp3z, CeyMIoSyrpkQ
X1ZpHSxyKbHn = tq24Tk6UZ6u1.problem.X1ZpHSxyKbHn
S89Y3lISBlM4 = tq24Tk6UZ6u1.UxYiT0ZFW2SZ
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe1\xe8\xbd\xe5\xcf\xb5\xa7i3+\xeb\x9fo\xda'), '\144' + chr(0b1000111 + 0o36) + '\x63' + '\x6f' + chr(9853 - 9753) + chr(0b1100101))(chr(0b1110101) + '\164' + '\x66' + chr(0b101101) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xe5\xee\xad'), chr(0b1100100) + chr(0b110101 + 0o60) + '\x63' + '\157' + '\144' + chr(0b1100101))(chr(979 - 862) + chr(11283 - 11167) + chr(102) + chr(45) + '\x38')):
oknqTBFY6iX3 = jSKPaHwSAfVv.shape_list(KLYceYPUcF5e)
MsbwfslwLjRO = IDJ2eXGCBCDu.layers.dense(KLYceYPUcF5e, X1ZpHSxyKbHn * S89Y3lISBlM4, name=xafqLlk3kkUe(SXOLrMavuUCe(b'\xf4\xe8\xbc\xf8'), chr(100) + chr(0b101101 + 0o70) + '\143' + chr(0b1101111) + chr(0b1010100 + 0o20) + chr(101))(chr(1209 - 1092) + '\x74' + chr(329 - 227) + '\055' + chr(0b111000)))
MsbwfslwLjRO = IDJ2eXGCBCDu.reshape(MsbwfslwLjRO, oknqTBFY6iX3[:ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b10101 + 0o36), ord("\x08"))] + [X1ZpHSxyKbHn, S89Y3lISBlM4])
MsbwfslwLjRO = IDJ2eXGCBCDu.transpose(MsbwfslwLjRO, [ehT0Px3KOsy9(chr(0b110000) + '\157' + '\060', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\064', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(6093 - 5982) + '\x31', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(9661 - 9550) + '\x32', 274 - 266), ehT0Px3KOsy9(chr(909 - 861) + chr(0b1101111) + '\x33', 8)])
if not xafqLlk3kkUe(IDJ2eXGCBCDu.get_variable_scope(), xafqLlk3kkUe(SXOLrMavuUCe(b'\xe5\xec\xba\xff\xcb'), '\144' + chr(101) + '\x63' + '\157' + '\144' + '\145')(chr(0b1110101) + chr(116) + chr(102) + chr(0b101101) + chr(56))):
mDMFq4zNSnwx = MsbwfslwLjRO[:, -ehT0Px3KOsy9(chr(48) + chr(0b1010 + 0o145) + chr(0b110001), 8), :, :, :]
xafqLlk3kkUe(IDJ2eXGCBCDu.summary, xafqLlk3kkUe(SXOLrMavuUCe(b'\xde\xed\xa2\xcd\xe6\x80\xadO\x1d*\xe6\x80'), '\x64' + chr(0b1100101) + chr(0b1100011) + '\x6f' + chr(100) + '\x65')(chr(117) + '\x74' + chr(0b10 + 0o144) + '\x2d' + chr(0b100110 + 0o22)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\xe5\xec\xbc\xf9\xc2\xa3'), chr(5870 - 5770) + chr(0b11001 + 0o114) + '\x63' + chr(111) + chr(100) + '\x65')(chr(0b1110101) + chr(3423 - 3307) + chr(0b1100110) + chr(603 - 558) + chr(56)), xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe3\xf9\xba\xd3\xdd\xb6\xadi31\xe5\x91x\xdacb/\x03\xa0\x06\xddb'), '\144' + chr(0b1100101) + chr(99) + chr(111) + '\x64' + chr(0b1100101))('\165' + chr(0b100111 + 0o115) + '\x66' + '\x2d' + chr(56)))(mDMFq4zNSnwx), max_outputs=ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31', 8))
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf2\xf1\xbf\xed\xc0\xb3\x94h\x055\xfb'), chr(348 - 248) + '\145' + chr(0b1011010 + 0o11) + '\157' + chr(3042 - 2942) + chr(0b1100101))('\x75' + '\x74' + chr(0b111110 + 0o50) + '\x2d' + chr(56)))(MsbwfslwLjRO, axis=-ehT0Px3KOsy9(chr(48) + chr(0b1000011 + 0o54) + chr(49), 8))
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
get_bottom
|
def get_bottom(modality_type, value=None):
"""Gets default bottom transformation; if none available, return value."""
if modality_type == ModalityType.AUDIO:
return audio_bottom
elif modality_type == ModalityType.AUDIO_SPECTRAL:
return audio_spectral_bottom
elif modality_type in (ModalityType.CLASS_LABEL,
ModalityType.MULTI_LABEL,
ModalityType.ONE_HOT_CLASS_LABEL,
ModalityType.SIGMOID_CLASS_LABEL,
ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL,
ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL):
return class_label_bottom
elif modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL):
return symbol_bottom
elif modality_type in (ModalityType.GENERIC_L2_LOSS,
ModalityType.IDENTITY,
ModalityType.IDENTITY_SYMBOL,
ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM):
return identity_bottom
elif modality_type == ModalityType.IMAGE:
return image_bottom
elif modality_type in (ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY,
ModalityType.IMAGE_CHANNEL_COMPRESS):
return image_channel_compress_bottom
elif modality_type in (ModalityType.REAL,
ModalityType.REAL_L2_LOSS,
ModalityType.REAL_LOG_POISSON_LOSS):
return real_bottom
elif modality_type == ModalityType.SPEECH_RECOGNITION:
return speech_recognition_bottom
elif modality_type == ModalityType.SYMBOL_ONE_HOT:
return symbol_one_hot_bottom
elif modality_type in (ModalityType.VIDEO,
ModalityType.VIDEO_L1,
ModalityType.VIDEO_L2):
return video_bottom
elif modality_type == ModalityType.VIDEO_BITWISE:
return video_bitwise_bottom
elif modality_type == ModalityType.VIDEO_IDENTITY:
return video_identity_bottom
elif modality_type in (ModalityType.VIDEO_L1_RAW,
ModalityType.VIDEO_L2_RAW):
return video_raw_bottom
elif modality_type == ModalityType.VIDEO_PIXEL_NOISE:
return video_pixel_noise_bottom
return value
|
python
|
def get_bottom(modality_type, value=None):
"""Gets default bottom transformation; if none available, return value."""
if modality_type == ModalityType.AUDIO:
return audio_bottom
elif modality_type == ModalityType.AUDIO_SPECTRAL:
return audio_spectral_bottom
elif modality_type in (ModalityType.CLASS_LABEL,
ModalityType.MULTI_LABEL,
ModalityType.ONE_HOT_CLASS_LABEL,
ModalityType.SIGMOID_CLASS_LABEL,
ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL,
ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL):
return class_label_bottom
elif modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL):
return symbol_bottom
elif modality_type in (ModalityType.GENERIC_L2_LOSS,
ModalityType.IDENTITY,
ModalityType.IDENTITY_SYMBOL,
ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM):
return identity_bottom
elif modality_type == ModalityType.IMAGE:
return image_bottom
elif modality_type in (ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY,
ModalityType.IMAGE_CHANNEL_COMPRESS):
return image_channel_compress_bottom
elif modality_type in (ModalityType.REAL,
ModalityType.REAL_L2_LOSS,
ModalityType.REAL_LOG_POISSON_LOSS):
return real_bottom
elif modality_type == ModalityType.SPEECH_RECOGNITION:
return speech_recognition_bottom
elif modality_type == ModalityType.SYMBOL_ONE_HOT:
return symbol_one_hot_bottom
elif modality_type in (ModalityType.VIDEO,
ModalityType.VIDEO_L1,
ModalityType.VIDEO_L2):
return video_bottom
elif modality_type == ModalityType.VIDEO_BITWISE:
return video_bitwise_bottom
elif modality_type == ModalityType.VIDEO_IDENTITY:
return video_identity_bottom
elif modality_type in (ModalityType.VIDEO_L1_RAW,
ModalityType.VIDEO_L2_RAW):
return video_raw_bottom
elif modality_type == ModalityType.VIDEO_PIXEL_NOISE:
return video_pixel_noise_bottom
return value
|
[
"def",
"get_bottom",
"(",
"modality_type",
",",
"value",
"=",
"None",
")",
":",
"if",
"modality_type",
"==",
"ModalityType",
".",
"AUDIO",
":",
"return",
"audio_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"AUDIO_SPECTRAL",
":",
"return",
"audio_spectral_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CLASS_LABEL",
",",
"ModalityType",
".",
"MULTI_LABEL",
",",
"ModalityType",
".",
"ONE_HOT_CLASS_LABEL",
",",
"ModalityType",
".",
"SIGMOID_CLASS_LABEL",
",",
"ModalityType",
".",
"SIGMOID_MAX_POOLING_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_AVERAGE_POOLING_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_LAST_TIMESTEP_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_MAX_POOLING_CLASS_LABEL",
")",
":",
"return",
"class_label_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CTC_SYMBOL",
",",
"ModalityType",
".",
"SYMBOL",
",",
"ModalityType",
".",
"SYMBOL_WEIGHTS_ALL",
")",
":",
"return",
"symbol_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"GENERIC_L2_LOSS",
",",
"ModalityType",
".",
"IDENTITY",
",",
"ModalityType",
".",
"IDENTITY_SYMBOL",
",",
"ModalityType",
".",
"IMAGE_CHANNEL_EMBEDDINGS_BOTTOM",
")",
":",
"return",
"identity_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE",
":",
"return",
"image_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"IMAGE_CHANNEL_BOTTOM_IDENTITY",
",",
"ModalityType",
".",
"IMAGE_CHANNEL_COMPRESS",
")",
":",
"return",
"image_channel_compress_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"REAL",
",",
"ModalityType",
".",
"REAL_L2_LOSS",
",",
"ModalityType",
".",
"REAL_LOG_POISSON_LOSS",
")",
":",
"return",
"real_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SPEECH_RECOGNITION",
":",
"return",
"speech_recognition_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SYMBOL_ONE_HOT",
":",
"return",
"symbol_one_hot_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"VIDEO",
",",
"ModalityType",
".",
"VIDEO_L1",
",",
"ModalityType",
".",
"VIDEO_L2",
")",
":",
"return",
"video_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_BITWISE",
":",
"return",
"video_bitwise_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_IDENTITY",
":",
"return",
"video_identity_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"VIDEO_L1_RAW",
",",
"ModalityType",
".",
"VIDEO_L2_RAW",
")",
":",
"return",
"video_raw_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_PIXEL_NOISE",
":",
"return",
"video_pixel_noise_bottom",
"return",
"value"
] |
Gets default bottom transformation; if none available, return value.
|
[
"Gets",
"default",
"bottom",
"transformation",
";",
"if",
"none",
"available",
"return",
"value",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1192-L1242
|
train
|
Gets default bottom transformation for given modality type.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + '\157' + chr(0b110010) + chr(53) + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(11569 - 11458) + chr(50) + '\x35' + chr(1669 - 1620), ord("\x08")), ehT0Px3KOsy9(chr(0b101101 + 0o3) + '\x6f' + chr(50) + chr(0b110011) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b110010 + 0o75) + chr(0b110001) + chr(1821 - 1770) + '\x33', 4485 - 4477), ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(111) + chr(0b110011) + chr(0b1000 + 0o57), 44511 - 44503), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101001 + 0o6) + chr(2332 - 2278) + chr(0b1 + 0o66), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(1440 - 1389) + '\x34', 0o10), ehT0Px3KOsy9(chr(0b11110 + 0o22) + chr(0b1101111) + '\063' + chr(0b110011 + 0o4) + chr(0b110011), 60167 - 60159), ehT0Px3KOsy9('\060' + chr(2444 - 2333) + chr(0b110001) + chr(0b110011) + chr(0b110111), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110011) + chr(0b110000) + chr(54), 0b1000), ehT0Px3KOsy9(chr(169 - 121) + chr(0b1000111 + 0o50) + chr(0b110001) + '\064' + chr(55), 0o10), ehT0Px3KOsy9(chr(2183 - 2135) + '\x6f' + chr(0b110001) + chr(53) + chr(0b110001), 5217 - 5209), ehT0Px3KOsy9(chr(48) + chr(111) + '\x33' + chr(49) + '\x35', 52620 - 52612), ehT0Px3KOsy9(chr(1721 - 1673) + chr(0b1101111) + chr(50) + '\064' + '\064', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101101 + 0o2) + '\x33' + '\063' + chr(0b110 + 0o56), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001) + '\063' + '\x36', 47181 - 47173), ehT0Px3KOsy9('\060' + chr(0b0 + 0o157) + '\067' + '\x33', 11687 - 11679), ehT0Px3KOsy9('\060' + chr(0b10010 + 0o135) + chr(49) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(10919 - 10808) + chr(0b100001 + 0o24) + chr(1615 - 1563), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1011111 + 0o20) + chr(716 - 665) + chr(52) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(973 - 925) + chr(111) + '\063' + chr(0b100101 + 0o20) + chr(958 - 903), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1010 + 0o145) + chr(1331 - 1281) + '\063' + '\067', 9485 - 9477), ehT0Px3KOsy9('\060' + '\157' + '\063' + chr(0b110100) + '\067', 0o10), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\x6f' + '\063' + chr(532 - 480) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(1255 - 1205) + '\067' + '\060', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1000011 + 0o54) + '\062' + '\063', 0o10), ehT0Px3KOsy9(chr(0b110000 + 0o0) + '\x6f' + '\x32' + '\x30', 54753 - 54745), ehT0Px3KOsy9(chr(0b110000) + chr(9853 - 9742) + chr(0b110111) + chr(1759 - 1710), 0o10), ehT0Px3KOsy9('\060' + chr(5111 - 5000) + '\061' + '\x35' + '\x35', 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\064' + chr(1570 - 1517), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010) + chr(998 - 944) + chr(0b110101), 0o10), ehT0Px3KOsy9('\060' + chr(5056 - 4945) + chr(0b10001 + 0o42) + chr(50) + chr(54), 0o10), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\x6f' + chr(1816 - 1766) + '\061' + chr(729 - 677), 0o10), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(111) + chr(49) + '\x31' + '\063', 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x36' + chr(0b101 + 0o61), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b101 + 0o57) + chr(48), 0o10), ehT0Px3KOsy9(chr(0b11010 + 0o26) + chr(0b1000000 + 0o57) + '\063' + '\x30' + chr(52), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(0b11110 + 0o30) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(10496 - 10385) + chr(51) + chr(0b110110) + chr(0b101000 + 0o16), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(973 - 862) + chr(0b110001) + chr(51) + chr(2209 - 2157), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(0b1101111) + '\065' + '\060', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x81'), chr(100) + chr(9877 - 9776) + '\143' + '\x6f' + chr(100) + '\x65')('\x75' + chr(116) + chr(0b1100110) + chr(0b1001 + 0o44) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def wgTg0sc4uUA5(N7qTs4FW6YfL, QmmgWUB13VCJ=None):
if N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xee\xfc\xda\x12P'), '\144' + chr(8503 - 8402) + chr(99) + '\157' + '\144' + chr(5474 - 5373))(chr(0b1110101) + chr(0b101101 + 0o107) + chr(338 - 236) + chr(161 - 116) + '\070')):
return a1v03mh4yIkx
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xee\xfc\xda\x12PhO\xab\x1a\x853;\x8c\xe8'), chr(2254 - 2154) + chr(0b1100101) + '\143' + '\x6f' + chr(8166 - 8066) + chr(101))(chr(0b1 + 0o164) + chr(9523 - 9407) + chr(0b101110 + 0o70) + chr(160 - 115) + chr(1634 - 1578))):
return fjzC10DtfeYX
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xec\xe5\xdf\x08LhP\xba\x1d\x83+'), chr(100) + chr(101) + chr(99) + chr(7142 - 7031) + chr(166 - 66) + '\145')('\165' + chr(7565 - 7449) + chr(0b1 + 0o145) + '\x2d' + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe2\xfc\xd2\x0fVhP\xba\x1d\x83+'), chr(0b1100100) + chr(101) + chr(0b1100011) + chr(111) + '\144' + chr(0b1000101 + 0o40))(chr(0b1110101) + '\164' + chr(6686 - 6584) + chr(0b101000 + 0o5) + chr(0b10110 + 0o42))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe0\xe7\xdb\x04WxH\xa4\x1c\x8a&:\x9e\xfb\x90\x1fyM`'), chr(7281 - 7181) + chr(101) + '\143' + '\x6f' + chr(4440 - 4340) + '\x65')(chr(117) + chr(116) + chr(0b1001001 + 0o35) + chr(45) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xe0\xd9\x16P~X\xa4\x1c\x8a&:\x9e\xfb\x90\x1fyM`'), '\144' + chr(0b1100101) + chr(99) + chr(4849 - 4738) + '\144' + chr(101))(chr(532 - 415) + chr(0b1110100) + chr(1915 - 1813) + chr(0b1001 + 0o44) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xe0\xd9\x16P~X\xa4\x12\x87?6\x9d\xeb\x93\x12rFk\xbd\xac\xd3\xa6\xb2\xebL\\x/`\xf9'), chr(5073 - 4973) + '\145' + chr(0b1100011) + '\157' + chr(100) + chr(0b10111 + 0o116))('\165' + chr(116) + '\146' + '\x2d' + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xe6\xd8\x0fRvD\xa4\x1e\x90";\x8c\xe3\x99\x01kGc\xae\xa6\xd1\xa0\xbe\xfb_Qj>z\xf9\x97\xce\x0ez'), chr(100) + chr(0b1100101) + chr(99) + '\x6f' + '\x64' + '\145')('\x75' + chr(0b1001000 + 0o54) + '\x66' + '\055' + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xe6\xd8\x0fRvD\xa4\x13\x874=\x92\xf0\x95\x13~[x\xa7\xbf\xc0\xa4\xad\xf9@Cf!d\xf7\x93\xc0'), '\x64' + '\145' + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + chr(0b1100101))(chr(117) + chr(0b101111 + 0o105) + chr(0b100100 + 0o102) + chr(45) + chr(2485 - 2429))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xe6\xd8\x0fRvD\xa4\x12\x87?6\x9d\xeb\x93\x12rFk\xbd\xac\xd3\xa6\xb2\xebL\\x/`\xf9'), '\x64' + '\x65' + chr(0b110111 + 0o54) + chr(5940 - 5829) + '\144' + chr(0b111 + 0o136))(chr(10978 - 10861) + chr(116) + '\x66' + chr(0b10000 + 0o35) + '\x38'))):
return uk_HoVDtliXI
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xec\xfd\xdd\x04LnQ\xb9\x10\x8a'), '\144' + chr(334 - 233) + '\x63' + '\x6f' + chr(100) + chr(101))(chr(8620 - 8503) + chr(0b111100 + 0o70) + chr(7737 - 7635) + '\x2d' + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xf0\xd3\x19P{'), chr(0b1100100) + '\x65' + '\x63' + chr(0b1101111) + chr(7297 - 7197) + '\145')('\165' + chr(0b1011110 + 0o26) + chr(0b11001 + 0o115) + chr(0b101101) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xf0\xd3\x19P{C\xac\x1a\x8f !\x99\xf7\x83\x1fwD'), chr(0b1010010 + 0o22) + chr(0b110101 + 0o60) + chr(0b1100011) + '\x6f' + chr(0b1100100) + chr(0b1100101))(chr(0b1110101) + chr(0b1001110 + 0o46) + chr(102) + '\055' + chr(2181 - 2125)))):
return Rv03jF3MXvlF
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe8\xec\xd0\x1eM~_\xa4\x13\xf48%\x82\xf7\x8f'), '\x64' + '\x65' + chr(99) + chr(10695 - 10584) + chr(100) + '\x65')('\165' + '\164' + chr(0b1 + 0o145) + '\055' + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe6\xed\xdb\x15K~H\xa2'), '\x64' + chr(101) + chr(99) + chr(3615 - 3504) + chr(100) + chr(101))('\165' + chr(0b10110 + 0o136) + '\146' + chr(45) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe6\xed\xdb\x15K~H\xa2\x00\x95>$\x8f\xeb\x90'), '\x64' + '\x65' + chr(4245 - 4146) + '\x6f' + chr(0b1100100) + '\x65')('\x75' + chr(116) + '\146' + chr(0b101101) + chr(0b100000 + 0o30))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe6\xe4\xdf\x1cZh_\xb3\x1e\x88),\x81\xfb\x99\x13yMh\xa6\xa6\xd1\xa0\xb2\xe7Q_m9j\xf8'), '\144' + chr(9222 - 9121) + '\143' + chr(0b1101111) + '\144' + chr(8839 - 8738))('\165' + chr(0b1110100) + chr(102) + '\055' + '\070'))):
return I4Q6CFwoHJ21
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe6\xe4\xdf\x1cZ'), chr(9506 - 9406) + '\x65' + chr(99) + '\157' + chr(7446 - 7346) + '\x65')(chr(0b1110101) + chr(10561 - 10445) + chr(0b1100110) + chr(995 - 950) + chr(0b111000))):
return nBEWq1NUJOuC
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe6\xe4\xdf\x1cZh_\xb3\x1e\x88),\x81\xfb\x9e\x11o\\c\xaf\xb0\xd6\xa3\xa4\xf6GYm4'), '\144' + '\x65' + chr(0b101110 + 0o65) + chr(0b101001 + 0o106) + '\144' + chr(0b111100 + 0o51))('\165' + '\x74' + '\146' + chr(0b11100 + 0o21) + chr(1154 - 1098))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe6\xe4\xdf\x1cZh_\xb3\x1e\x88),\x81\xfb\x9f\x11vX~\xa7\xbc\xcc'), chr(100) + chr(4279 - 4178) + chr(99) + chr(3243 - 3132) + chr(0b111100 + 0o50) + '\145')(chr(0b100111 + 0o116) + chr(116) + chr(6769 - 6667) + chr(200 - 155) + chr(56)))):
return Bftc64BxcxxT
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd\xec\xdf\x17'), chr(100) + '\x65' + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + chr(101))(chr(117) + '\164' + chr(0b1 + 0o145) + chr(0b11111 + 0o16) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd\xec\xdf\x17@{.\xa4\x13\x894:'), chr(0b101110 + 0o66) + '\145' + '\143' + chr(0b1100000 + 0o17) + chr(100) + chr(101))(chr(0b1110101) + chr(116) + '\146' + chr(45) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd\xec\xdf\x17@{S\xbc\x00\x96( \x9e\xf7\x93\x10dDc\xb1\xbc'), chr(8805 - 8705) + chr(0b1001011 + 0o32) + chr(0b1100011) + chr(0b111111 + 0o60) + chr(2397 - 2297) + chr(296 - 195))(chr(0b1110101) + '\x74' + chr(0b11001 + 0o115) + '\x2d' + chr(253 - 197)))):
return LzVngjm7tWJD
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xf9\xdb\x1e\\\x7fC\xa9\x1a\x85(.\x83\xed\x88\x17tF'), '\144' + chr(0b1100101) + chr(99) + chr(0b1100100 + 0o13) + chr(391 - 291) + '\x65')('\x75' + '\x74' + chr(102) + chr(0b101101) + '\x38')):
return BUwtcxYvLr23
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfc\xf0\xd3\x19P{C\xb4\x11\x838!\x82\xf0'), '\x64' + '\145' + chr(1591 - 1492) + '\x6f' + chr(0b111010 + 0o52) + '\x65')(chr(117) + '\x74' + '\146' + chr(0b101101) + '\070')):
return Lhrai3rloNcV
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf9\xe0\xda\x1eP'), '\x64' + chr(101) + chr(353 - 254) + chr(0b100100 + 0o113) + chr(0b1100100) + chr(0b1100101))('\165' + '\164' + chr(102) + chr(1049 - 1004) + chr(0b10101 + 0o43))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf9\xe0\xda\x1ePhP\xca'), chr(4702 - 4602) + chr(0b10 + 0o143) + chr(1815 - 1716) + chr(0b1101111) + '\144' + chr(0b1110 + 0o127))(chr(4071 - 3954) + '\x74' + chr(3004 - 2902) + '\x2d' + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf9\xe0\xda\x1ePhP\xc9'), '\144' + '\x65' + chr(1598 - 1499) + '\x6f' + '\144' + chr(3143 - 3042))('\x75' + '\164' + '\x66' + chr(45) + '\070'))):
return UTQU1__m4eux
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf9\xe0\xda\x1ePh^\xb2\x0b\x91.:\x88'), chr(0b11101 + 0o107) + '\x65' + '\143' + '\x6f' + '\144' + '\x65')('\x75' + '\x74' + '\146' + chr(0b101100 + 0o1) + chr(0b111000))):
return fJYw_NaESM8R
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf9\xe0\xda\x1ePhU\xbf\x1a\x883 \x99\xfd'), chr(0b111100 + 0o50) + chr(101) + chr(99) + chr(111) + '\144' + chr(101))('\165' + '\164' + chr(0b1001100 + 0o32) + chr(0b101101) + chr(0b111000))):
return SFIit7xawxOd
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf9\xe0\xda\x1ePhP\xca\x00\x94&>'), chr(2368 - 2268) + '\145' + chr(0b1100011) + chr(8055 - 7944) + chr(2594 - 2494) + chr(0b110011 + 0o62))(chr(1173 - 1056) + '\x74' + chr(102) + chr(45) + chr(1455 - 1399))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf9\xe0\xda\x1ePhP\xc9\x00\x94&>'), '\144' + chr(0b11100 + 0o111) + chr(8354 - 8255) + chr(111) + chr(0b1100100) + '\145')('\x75' + chr(0b1110100) + chr(0b1101 + 0o131) + chr(0b101101) + chr(0b111000)))):
return UVkHiIoReFMl
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf9\xe0\xda\x1ePhL\xb2\x07\x83+6\x83\xeb\x95\r~'), chr(0b110111 + 0o55) + chr(0b100110 + 0o77) + '\x63' + '\x6f' + chr(0b1010011 + 0o21) + '\145')('\165' + chr(0b10010 + 0o142) + chr(5923 - 5821) + chr(1415 - 1370) + '\070')):
return mxRCIiyni4wa
return QmmgWUB13VCJ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
get_loss
|
def get_loss(modality_type, value=None):
"""Gets default loss transformation; if none available, return value."""
if modality_type in (ModalityType.AUDIO,
ModalityType.AUDIO_SPECTRAL,
ModalityType.CLASS_LABEL,
ModalityType.IDENTITY,
ModalityType.IDENTITY_SYMBOL,
ModalityType.IMAGE,
ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY,
ModalityType.IMAGE_CHANNEL_COMPRESS,
ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM,
ModalityType.REAL,
ModalityType.SPEECH_RECOGNITION,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL):
return generic_loss
elif modality_type == ModalityType.CTC_SYMBOL:
return ctc_symbol_loss
elif modality_type == ModalityType.GENERIC_L2_LOSS:
return generic_l2_loss
elif modality_type == ModalityType.MULTI_LABEL:
return multi_label_loss
elif modality_type in (ModalityType.ONE_HOT_CLASS_LABEL,
ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL,
ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL):
return one_hot_class_label_loss
elif modality_type == ModalityType.REAL_L2_LOSS:
return real_l2_loss
elif modality_type == ModalityType.REAL_LOG_POISSON_LOSS:
return real_log_poisson_loss
elif modality_type == ModalityType.SIGMOID_CLASS_LABEL:
return sigmoid_class_label_loss
elif modality_type == ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL:
return sigmoid_max_pooling_class_label_loss
elif modality_type == ModalityType.SYMBOL_ONE_HOT:
return symbol_one_hot_loss
elif modality_type in (ModalityType.VIDEO,
ModalityType.VIDEO_BITWISE,
ModalityType.VIDEO_PIXEL_NOISE):
return video_loss
elif modality_type == ModalityType.VIDEO_IDENTITY:
return video_identity_loss
elif modality_type == ModalityType.VIDEO_L1:
return video_l1_loss
elif modality_type == ModalityType.VIDEO_L1_RAW:
return video_l1_raw_loss
elif modality_type == ModalityType.VIDEO_L2:
return video_l2_loss
elif modality_type == ModalityType.VIDEO_L2_RAW:
return video_l2_raw_loss
return value
|
python
|
def get_loss(modality_type, value=None):
"""Gets default loss transformation; if none available, return value."""
if modality_type in (ModalityType.AUDIO,
ModalityType.AUDIO_SPECTRAL,
ModalityType.CLASS_LABEL,
ModalityType.IDENTITY,
ModalityType.IDENTITY_SYMBOL,
ModalityType.IMAGE,
ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY,
ModalityType.IMAGE_CHANNEL_COMPRESS,
ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM,
ModalityType.REAL,
ModalityType.SPEECH_RECOGNITION,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL):
return generic_loss
elif modality_type == ModalityType.CTC_SYMBOL:
return ctc_symbol_loss
elif modality_type == ModalityType.GENERIC_L2_LOSS:
return generic_l2_loss
elif modality_type == ModalityType.MULTI_LABEL:
return multi_label_loss
elif modality_type in (ModalityType.ONE_HOT_CLASS_LABEL,
ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL,
ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL):
return one_hot_class_label_loss
elif modality_type == ModalityType.REAL_L2_LOSS:
return real_l2_loss
elif modality_type == ModalityType.REAL_LOG_POISSON_LOSS:
return real_log_poisson_loss
elif modality_type == ModalityType.SIGMOID_CLASS_LABEL:
return sigmoid_class_label_loss
elif modality_type == ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL:
return sigmoid_max_pooling_class_label_loss
elif modality_type == ModalityType.SYMBOL_ONE_HOT:
return symbol_one_hot_loss
elif modality_type in (ModalityType.VIDEO,
ModalityType.VIDEO_BITWISE,
ModalityType.VIDEO_PIXEL_NOISE):
return video_loss
elif modality_type == ModalityType.VIDEO_IDENTITY:
return video_identity_loss
elif modality_type == ModalityType.VIDEO_L1:
return video_l1_loss
elif modality_type == ModalityType.VIDEO_L1_RAW:
return video_l1_raw_loss
elif modality_type == ModalityType.VIDEO_L2:
return video_l2_loss
elif modality_type == ModalityType.VIDEO_L2_RAW:
return video_l2_raw_loss
return value
|
[
"def",
"get_loss",
"(",
"modality_type",
",",
"value",
"=",
"None",
")",
":",
"if",
"modality_type",
"in",
"(",
"ModalityType",
".",
"AUDIO",
",",
"ModalityType",
".",
"AUDIO_SPECTRAL",
",",
"ModalityType",
".",
"CLASS_LABEL",
",",
"ModalityType",
".",
"IDENTITY",
",",
"ModalityType",
".",
"IDENTITY_SYMBOL",
",",
"ModalityType",
".",
"IMAGE",
",",
"ModalityType",
".",
"IMAGE_CHANNEL_BOTTOM_IDENTITY",
",",
"ModalityType",
".",
"IMAGE_CHANNEL_COMPRESS",
",",
"ModalityType",
".",
"IMAGE_CHANNEL_EMBEDDINGS_BOTTOM",
",",
"ModalityType",
".",
"REAL",
",",
"ModalityType",
".",
"SPEECH_RECOGNITION",
",",
"ModalityType",
".",
"SYMBOL",
",",
"ModalityType",
".",
"SYMBOL_WEIGHTS_ALL",
")",
":",
"return",
"generic_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"CTC_SYMBOL",
":",
"return",
"ctc_symbol_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"GENERIC_L2_LOSS",
":",
"return",
"generic_l2_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"MULTI_LABEL",
":",
"return",
"multi_label_loss",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"ONE_HOT_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_AVERAGE_POOLING_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_LAST_TIMESTEP_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_MAX_POOLING_CLASS_LABEL",
")",
":",
"return",
"one_hot_class_label_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"REAL_L2_LOSS",
":",
"return",
"real_l2_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"REAL_LOG_POISSON_LOSS",
":",
"return",
"real_log_poisson_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SIGMOID_CLASS_LABEL",
":",
"return",
"sigmoid_class_label_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SIGMOID_MAX_POOLING_CLASS_LABEL",
":",
"return",
"sigmoid_max_pooling_class_label_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SYMBOL_ONE_HOT",
":",
"return",
"symbol_one_hot_loss",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"VIDEO",
",",
"ModalityType",
".",
"VIDEO_BITWISE",
",",
"ModalityType",
".",
"VIDEO_PIXEL_NOISE",
")",
":",
"return",
"video_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_IDENTITY",
":",
"return",
"video_identity_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_L1",
":",
"return",
"video_l1_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_L1_RAW",
":",
"return",
"video_l1_raw_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_L2",
":",
"return",
"video_l2_loss",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_L2_RAW",
":",
"return",
"video_l2_raw_loss",
"return",
"value"
] |
Gets default loss transformation; if none available, return value.
|
[
"Gets",
"default",
"loss",
"transformation",
";",
"if",
"none",
"available",
"return",
"value",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1245-L1296
|
train
|
Gets default loss transformation for given modality type.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110001) + chr(0b110100) + '\066', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010) + '\x30' + chr(0b110100), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\x37' + '\061', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(0b110001) + chr(49), 0o10), ehT0Px3KOsy9(chr(48) + chr(1256 - 1145) + chr(494 - 444) + '\x32' + '\x37', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + '\062' + chr(555 - 502), 0o10), ehT0Px3KOsy9(chr(0b1110 + 0o42) + chr(0b1100001 + 0o16) + chr(0b110011) + chr(2148 - 2093) + '\066', 32588 - 32580), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(0b11000 + 0o127) + '\x33' + chr(50) + chr(0b100 + 0o63), ord("\x08")), ehT0Px3KOsy9(chr(0b11010 + 0o26) + chr(111) + chr(0b10101 + 0o34) + '\x36' + '\067', 0o10), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(0b1101111) + '\x32' + '\066' + '\065', 0o10), ehT0Px3KOsy9(chr(48) + chr(8829 - 8718) + chr(0b10000 + 0o43) + chr(0b110100) + chr(0b101111 + 0o1), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(789 - 739) + chr(51) + chr(52), 34514 - 34506), ehT0Px3KOsy9('\x30' + '\x6f' + chr(2690 - 2636), 10891 - 10883), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011) + chr(0b110011) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110011) + chr(0b101101 + 0o3) + '\x32', 0b1000), ehT0Px3KOsy9(chr(1905 - 1857) + chr(0b101111 + 0o100) + chr(528 - 477) + chr(1501 - 1450) + chr(1968 - 1916), ord("\x08")), ehT0Px3KOsy9('\060' + chr(4386 - 4275) + '\063' + chr(51) + chr(2784 - 2730), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\062' + chr(49) + chr(50), 0o10), ehT0Px3KOsy9(chr(328 - 280) + '\157' + chr(0b101110 + 0o3) + chr(49), 56119 - 56111), ehT0Px3KOsy9(chr(0b1000 + 0o50) + chr(8088 - 7977) + '\061' + '\066' + chr(1453 - 1398), 8), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110001) + chr(54) + '\x30', 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33' + chr(0b110001) + chr(2328 - 2276), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100001 + 0o16) + chr(0b11101 + 0o25) + '\x30' + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(2210 - 2162) + chr(111) + '\x33' + '\x35' + chr(0b111 + 0o57), 41791 - 41783), ehT0Px3KOsy9('\x30' + '\157' + chr(0b101110 + 0o5) + chr(0b101001 + 0o11) + chr(207 - 156), 34891 - 34883), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x32' + chr(0b10110 + 0o40), 40850 - 40842), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110010) + '\060' + chr(53), 0o10), ehT0Px3KOsy9(chr(971 - 923) + chr(6174 - 6063) + chr(49) + chr(0b0 + 0o61) + chr(0b101101 + 0o7), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(53) + chr(0b101001 + 0o14), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + '\x36' + chr(49), 0o10), ehT0Px3KOsy9(chr(0b1010 + 0o46) + '\157' + chr(0b10011 + 0o36) + chr(0b1110 + 0o46) + chr(0b110100 + 0o2), 8), ehT0Px3KOsy9(chr(1766 - 1718) + chr(8626 - 8515) + chr(0b110011) + chr(0b110110) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(7443 - 7332) + '\062' + '\x32' + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(668 - 620) + chr(0b1101111) + '\x37' + '\x34', ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(282 - 231) + chr(1725 - 1670) + chr(0b110110), 8), ehT0Px3KOsy9(chr(1851 - 1803) + chr(10040 - 9929) + '\x31' + chr(2449 - 2395) + chr(51), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\064' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x33' + chr(54) + chr(0b101010 + 0o14), ord("\x08")), ehT0Px3KOsy9(chr(0b101101 + 0o3) + '\157' + chr(51) + '\x37' + chr(51), 31180 - 31172), ehT0Px3KOsy9(chr(0b110000) + chr(7973 - 7862) + chr(0b11100 + 0o25) + chr(48) + '\065', 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(254 - 206) + '\x6f' + chr(94 - 41) + '\x30', 16150 - 16142)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'H'), chr(100) + chr(6263 - 6162) + chr(5567 - 5468) + chr(111) + chr(0b100100 + 0o100) + chr(0b111001 + 0o54))('\165' + '\x74' + '\146' + chr(1010 - 965) + chr(0b10110 + 0o42)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def rHg1ZYgTHKgX(N7qTs4FW6YfL, QmmgWUB13VCJ=None):
if N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b"'\xd1}\x00\xd8"), chr(0b1100100) + chr(0b1100101) + chr(0b1100011) + '\157' + chr(0b1100100) + '\145')(chr(117) + chr(0b1110100) + chr(102) + chr(0b101100 + 0o1) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b"'\xd1}\x00\xd8\x8a\xad\x872\xf6\xb4\xceG\x93"), chr(0b1100100) + chr(5565 - 5464) + '\143' + chr(0b1101111) + chr(0b1100100) + '\x65')(chr(0b111111 + 0o66) + chr(0b1110100) + chr(0b1100110) + chr(0b101101) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'%\xc8x\x1a\xc4\x8a\xb2\x965\xf0\xac'), '\x64' + '\x65' + '\x63' + chr(0b10011 + 0o134) + chr(8651 - 8551) + chr(101))(chr(117) + '\x74' + chr(102) + chr(0b1011 + 0o42) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'/\xc0|\x07\xc3\x9c\xaa\x8e'), chr(0b1100100) + chr(3736 - 3635) + chr(0b110011 + 0o60) + '\157' + chr(0b1100100) + '\145')('\x75' + chr(0b1110100) + '\x66' + chr(0b100 + 0o51) + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'/\xc0|\x07\xc3\x9c\xaa\x8e(\xe6\xb9\xd1D\x90\x8e'), chr(0b1100100) + chr(101) + chr(0b1011011 + 0o10) + chr(0b1101111) + '\x64' + chr(101))('\165' + '\164' + '\x66' + chr(45) + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'/\xc9x\x0e\xd2'), chr(100) + chr(0b1100101) + chr(99) + chr(11447 - 11336) + chr(4223 - 4123) + chr(7522 - 7421))(chr(117) + chr(0b1011000 + 0o34) + '\x66' + '\055' + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'/\xc9x\x0e\xd2\x8a\xbd\x9f6\xfb\xae\xd9J\x80\x80\x93\x8a\xdeF:/\x85\xc3\xab\x9dB\xcc\x18t'), '\144' + chr(8532 - 8431) + chr(0b101010 + 0o71) + chr(11347 - 11236) + '\x64' + chr(2814 - 2713))('\x75' + chr(0b1110100) + chr(3167 - 3065) + '\055' + chr(720 - 664))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'/\xc9x\x0e\xd2\x8a\xbd\x9f6\xfb\xae\xd9J\x80\x81\x93\x93\xda[2#\x9f'), chr(3253 - 3153) + chr(1475 - 1374) + chr(0b1100011) + '\157' + '\144' + chr(101))(chr(0b1100100 + 0o21) + '\x74' + chr(0b1100110) + '\055' + chr(0b11001 + 0o37))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'/\xc9x\x0e\xd2\x8a\xbd\x9f6\xfb\xae\xd9J\x80\x87\x91\x9c\xcfM39\x82\xc0\xbd\x8cT\xca\x18y\xbe+'), chr(0b11111 + 0o105) + chr(0b1100101) + chr(5085 - 4986) + chr(2724 - 2613) + '\144' + '\145')('\165' + chr(116) + chr(5526 - 5424) + '\055' + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'4\xc1x\x05'), chr(100) + chr(8123 - 8022) + '\143' + '\x6f' + chr(0b1100100) + '\145')(chr(117) + chr(6368 - 6252) + '\146' + chr(0b100101 + 0o10) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xd4|\x0c\xd4\x9d\xa1\x852\xf6\xaf\xdbH\x96\x96\x95\x91\xc4'), chr(100) + '\145' + chr(99) + chr(111) + chr(0b100101 + 0o77) + '\x65')(chr(0b1110101) + chr(0b1110100) + '\146' + chr(744 - 699) + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xddt\x0b\xd8\x99'), chr(7409 - 7309) + chr(0b1100101) + chr(99) + chr(111) + chr(7673 - 7573) + chr(0b1100101))(chr(0b1100111 + 0o16) + chr(10788 - 10672) + chr(9027 - 8925) + '\x2d' + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xddt\x0b\xd8\x99\xa1\x802\xfc\xa7\xd4R\x8c\x9d\x9d\x92\xc6'), '\x64' + '\x65' + chr(5075 - 4976) + chr(0b111000 + 0o67) + chr(0b1100100) + '\x65')('\x75' + chr(0b11111 + 0o125) + '\146' + chr(96 - 51) + '\x38'))):
return satn1Z5kItGQ
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'%\xd0z\x16\xc4\x8c\xb3\x958\xf9'), chr(0b10101 + 0o117) + '\x65' + '\x63' + chr(0b0 + 0o157) + '\x64' + '\145')(chr(0b1001111 + 0o46) + chr(0b1101111 + 0o5) + '\146' + chr(0b1000 + 0o45) + chr(56))):
return QyBSDyDRERUx
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'!\xc1w\x0c\xc5\x9c\xbd\x88;\x87\xbf\xd0I\x8c\x91'), chr(0b1000100 + 0o40) + chr(9425 - 9324) + '\x63' + chr(0b1101111 + 0o0) + chr(9394 - 9294) + chr(101))(chr(117) + chr(116) + chr(102) + chr(932 - 887) + '\070')):
return D4fZUrWl_2_9
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'+\xd1u\x1d\xde\x8a\xb2\x965\xf0\xac'), '\x64' + chr(101) + chr(0b110101 + 0o56) + chr(0b1101111) + chr(0b1100100) + '\x65')(chr(0b1110101) + '\x74' + '\146' + chr(0b101101) + chr(1269 - 1213))):
return digenOuecR7e
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b')\xca|\x16\xdf\x9a\xaa\x884\xf9\xa1\xcfU\x80\x8e\x9d\x9c\xcfE'), '\144' + chr(0b1001110 + 0o27) + '\x63' + chr(111) + chr(0b1010100 + 0o20) + chr(7745 - 7644))('\165' + chr(116) + chr(102) + '\x2d' + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xcb\x7f\x1d\xda\x94\xa6\x886\xe3\xa5\xceG\x98\x87\x83\x8e\xc5F;9\x82\xc0\xb1\x90Z\xc4\x1f~\xae*\xb5\xd4y\xb2'), chr(0b1100 + 0o130) + chr(0b1100101) + '\x63' + chr(0b1100111 + 0o10) + chr(100) + '\x65')(chr(0b1110101) + chr(0b1110100) + chr(953 - 851) + '\x2d' + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xcb\x7f\x1d\xda\x94\xa6\x88;\xf4\xb3\xc8Y\x8b\x8b\x91\x9b\xd9]2 \x93\xc4\xa2\x92E\xd6\x13a\xb0$\xb1\xda'), chr(0b1100100) + chr(0b1100101) + chr(0b1100011) + chr(0b1101111) + chr(1672 - 1572) + chr(3643 - 3542))(chr(12558 - 12441) + chr(2694 - 2578) + '\x66' + chr(0b101101) + chr(0b11001 + 0o37))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xcb\x7f\x1d\xda\x94\xa6\x88:\xf4\xb8\xc3V\x90\x8d\x90\x97\xc4N(3\x80\xc6\xbd\x80I\xc9\ro\xb4*'), chr(0b1000110 + 0o36) + chr(101) + '\143' + chr(491 - 380) + '\144' + chr(4486 - 4385))(chr(117) + chr(0b1110100) + chr(102) + chr(45) + '\070'))):
return sQFKa_Gsem2q
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'4\xc1x\x05\xc8\x99\xcc\x88;\xfa\xb3\xcf'), chr(0b1000110 + 0o36) + chr(0b1100101) + chr(6885 - 6786) + chr(0b1101111) + chr(0b1100100) + '\145')('\165' + '\x74' + chr(8003 - 7901) + '\055' + '\x38')):
return fTqvcN6mMYpq
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'4\xc1x\x05\xc8\x99\xb1\x90(\xe5\xaf\xd5U\x8c\x8d\x92\x81\xc6F$#'), chr(5195 - 5095) + '\145' + chr(785 - 686) + chr(0b1101111) + chr(0b1000001 + 0o43) + chr(263 - 162))(chr(12530 - 12413) + chr(8906 - 8790) + chr(0b110101 + 0o61) + chr(0b100001 + 0o14) + '\x38')):
return wLa7yBGuyt7m
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xcd~\x04\xd8\x9c\xba\x884\xf9\xa1\xcfU\x80\x8e\x9d\x9c\xcfE'), chr(0b1100100) + chr(0b1100101) + chr(0b1100011) + chr(9149 - 9038) + chr(0b1001 + 0o133) + '\x65')(chr(0b1101110 + 0o7) + '\x74' + chr(102) + chr(0b101101) + chr(0b111000))):
return m0T3xdcEkCcx
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xcd~\x04\xd8\x9c\xba\x88:\xf4\xb8\xc3V\x90\x8d\x90\x97\xc4N(3\x80\xc6\xbd\x80I\xc9\ro\xb4*'), chr(0b1100100) + '\145' + '\x63' + chr(0b1000011 + 0o54) + chr(0b1100100) + '\145')(chr(0b1001100 + 0o51) + '\x74' + chr(102) + chr(0b1 + 0o54) + chr(1982 - 1926))):
return tVVOzaUS0yOu
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'5\xddt\x0b\xd8\x99\xa1\x989\xf0\xbf\xd4I\x8b'), '\144' + chr(0b1000011 + 0o42) + chr(3615 - 3516) + chr(0b1101111) + '\144' + chr(0b111010 + 0o53))('\x75' + chr(10012 - 9896) + chr(102) + chr(45) + '\070')):
return GfUKEMcrk4Tg
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'0\xcd}\x0c\xd8'), chr(1350 - 1250) + chr(5476 - 5375) + '\x63' + '\x6f' + chr(1154 - 1054) + chr(8494 - 8393))(chr(0b1010011 + 0o42) + chr(0b1110100) + chr(3331 - 3229) + chr(0b1000 + 0o45) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'0\xcd}\x0c\xd8\x8a\xbc\x9e#\xe2\xa9\xcfC'), chr(0b110001 + 0o63) + '\x65' + chr(0b1100011) + chr(111) + chr(0b110111 + 0o55) + '\145')('\x75' + chr(0b1001000 + 0o54) + '\x66' + chr(0b101101) + chr(2398 - 2342))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'0\xcd}\x0c\xd8\x8a\xae\x9e/\xf0\xac\xc3H\x90\x8b\x8f\x9b'), chr(0b1100100) + chr(101) + '\x63' + chr(0b1101011 + 0o4) + chr(0b1100 + 0o130) + chr(0b1100101))(chr(11556 - 11439) + chr(10968 - 10852) + '\146' + chr(0b11010 + 0o23) + chr(0b1000 + 0o60)))):
return JIGZcfgBtYav
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'0\xcd}\x0c\xd8\x8a\xb7\x932\xfb\xb4\xd5R\x86'), chr(0b1100100) + '\145' + '\x63' + '\157' + '\144' + chr(0b1100101))(chr(117) + chr(116) + chr(0b1001110 + 0o30) + chr(0b101101) + chr(0b111000))):
return j2z4gnHvpIsk
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'0\xcd}\x0c\xd8\x8a\xb2\xe6'), '\x64' + chr(515 - 414) + '\x63' + chr(111) + '\x64' + chr(101))(chr(10204 - 10087) + chr(116) + chr(0b1100110) + '\055' + chr(1493 - 1437))):
return N9rPShunrV5z
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'0\xcd}\x0c\xd8\x8a\xb2\xe6(\xe7\xa1\xcb'), '\144' + chr(1171 - 1070) + chr(0b1100011) + '\157' + chr(100) + '\x65')(chr(0b1110101) + chr(8488 - 8372) + '\146' + chr(45) + chr(1036 - 980))):
return cMATj0P9XumV
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'0\xcd}\x0c\xd8\x8a\xb2\xe5'), '\x64' + chr(0b1100101) + '\143' + '\x6f' + '\x64' + chr(0b1100101))('\165' + chr(12414 - 12298) + chr(0b1100110) + chr(0b101101) + chr(0b111000))):
return uPyZteooulGN
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'0\xcd}\x0c\xd8\x8a\xb2\xe5(\xe7\xa1\xcb'), chr(0b1010110 + 0o16) + chr(0b1100101) + chr(0b1100011) + chr(11021 - 10910) + chr(100) + chr(0b1010001 + 0o24))('\x75' + chr(6619 - 6503) + '\x66' + chr(0b101101) + chr(56))):
return BKt6ssgpP1bz
return QmmgWUB13VCJ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
get_name
|
def get_name(modality_type, value=None):
"""Gets default name for transformations; if none available, return value."""
# For legacy reasons, modalities vary in their naming scheme. Future plans are
# to remove any need for get_name. We do not recommend using it.
if modality_type == ModalityType.AUDIO:
return lambda model_hparams, vocab_size: "audio_modality"
elif modality_type == ModalityType.AUDIO_SPECTRAL:
return lambda model_hparams, vocab_size: "audio_spectral_modality"
elif modality_type == ModalityType.GENERIC_L2_LOSS:
return lambda model_hparams, vocab_size: "generic_l2_loss_modality"
elif modality_type == ModalityType.IDENTITY:
return lambda model_hparams, vocab_size: "identity_modality"
elif modality_type == ModalityType.IMAGE:
return lambda model_hparams, vocab_size: "image_modality"
elif modality_type == ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY:
return (lambda model_hparams, vocab_size: # pylint: disable=g-long-lambda
"image_channel_bottom_identity_modality")
elif modality_type == ModalityType.IMAGE_CHANNEL_COMPRESS:
return lambda model_hparams, vocab_size: "image_channel_compress_modality"
elif modality_type == ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM:
return lambda model_hparams, vocab_size: "image_channel_embeddings_bottom"
elif modality_type == ModalityType.REAL:
return lambda model_hparams, vocab_size: "real_modality"
elif modality_type == ModalityType.REAL_L2_LOSS:
return lambda model_hparams, vocab_size: "real_l2_loss_modality"
elif modality_type == ModalityType.REAL_LOG_POISSON_LOSS:
return lambda model_hparams, vocab_size: "real_log_poisson_loss_modality"
elif modality_type == ModalityType.SPEECH_RECOGNITION:
return lambda model_hparams, vocab_size: "speech_recognition_modality"
elif modality_type == ModalityType.VIDEO:
return lambda model_hparams, vocab_size: "video_modality"
elif modality_type == ModalityType.VIDEO_BITWISE:
return lambda model_hparams, vocab_size: "video_modality_bitwise"
elif modality_type == ModalityType.VIDEO_IDENTITY:
return lambda model_hparams, vocab_size: "video_modality_identity"
elif modality_type == ModalityType.VIDEO_L1:
return lambda model_hparams, vocab_size: "video_modality_l1"
elif modality_type == ModalityType.VIDEO_L1_RAW:
return lambda model_hparams, vocab_size: "video_modality_l1_raw"
elif modality_type == ModalityType.VIDEO_L2:
return lambda model_hparams, vocab_size: "video_modality_l2"
elif modality_type == ModalityType.VIDEO_L2_RAW:
return lambda model_hparams, vocab_size: "video_modality_l2_raw"
elif modality_type == ModalityType.VIDEO_PIXEL_NOISE:
return lambda model_hparams, vocab_size: "video_modality_pixel_noise"
elif modality_type in (ModalityType.CLASS_LABEL,
ModalityType.MULTI_LABEL,
ModalityType.ONE_HOT_CLASS_LABEL):
def name(model_hparams, vocab_size):
return "class_label_modality_%d_%d" % (vocab_size,
model_hparams.hidden_size)
return name
elif modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.IDENTITY_SYMBOL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL,
ModalityType.SYMBOL_ONE_HOT):
def name(model_hparams, vocab_size):
return "symbol_modality_%d_%d" % (vocab_size, model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SIGMOID_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "sigmoid_class_symbol_modality_%d_%d" % (vocab_size,
model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "sigmoid_max_pooling_class_symbol_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "softmax_average_pooling_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "softmax_last_timestep_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "softmax_max_pooling_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)
return name
return value
|
python
|
def get_name(modality_type, value=None):
"""Gets default name for transformations; if none available, return value."""
# For legacy reasons, modalities vary in their naming scheme. Future plans are
# to remove any need for get_name. We do not recommend using it.
if modality_type == ModalityType.AUDIO:
return lambda model_hparams, vocab_size: "audio_modality"
elif modality_type == ModalityType.AUDIO_SPECTRAL:
return lambda model_hparams, vocab_size: "audio_spectral_modality"
elif modality_type == ModalityType.GENERIC_L2_LOSS:
return lambda model_hparams, vocab_size: "generic_l2_loss_modality"
elif modality_type == ModalityType.IDENTITY:
return lambda model_hparams, vocab_size: "identity_modality"
elif modality_type == ModalityType.IMAGE:
return lambda model_hparams, vocab_size: "image_modality"
elif modality_type == ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY:
return (lambda model_hparams, vocab_size: # pylint: disable=g-long-lambda
"image_channel_bottom_identity_modality")
elif modality_type == ModalityType.IMAGE_CHANNEL_COMPRESS:
return lambda model_hparams, vocab_size: "image_channel_compress_modality"
elif modality_type == ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM:
return lambda model_hparams, vocab_size: "image_channel_embeddings_bottom"
elif modality_type == ModalityType.REAL:
return lambda model_hparams, vocab_size: "real_modality"
elif modality_type == ModalityType.REAL_L2_LOSS:
return lambda model_hparams, vocab_size: "real_l2_loss_modality"
elif modality_type == ModalityType.REAL_LOG_POISSON_LOSS:
return lambda model_hparams, vocab_size: "real_log_poisson_loss_modality"
elif modality_type == ModalityType.SPEECH_RECOGNITION:
return lambda model_hparams, vocab_size: "speech_recognition_modality"
elif modality_type == ModalityType.VIDEO:
return lambda model_hparams, vocab_size: "video_modality"
elif modality_type == ModalityType.VIDEO_BITWISE:
return lambda model_hparams, vocab_size: "video_modality_bitwise"
elif modality_type == ModalityType.VIDEO_IDENTITY:
return lambda model_hparams, vocab_size: "video_modality_identity"
elif modality_type == ModalityType.VIDEO_L1:
return lambda model_hparams, vocab_size: "video_modality_l1"
elif modality_type == ModalityType.VIDEO_L1_RAW:
return lambda model_hparams, vocab_size: "video_modality_l1_raw"
elif modality_type == ModalityType.VIDEO_L2:
return lambda model_hparams, vocab_size: "video_modality_l2"
elif modality_type == ModalityType.VIDEO_L2_RAW:
return lambda model_hparams, vocab_size: "video_modality_l2_raw"
elif modality_type == ModalityType.VIDEO_PIXEL_NOISE:
return lambda model_hparams, vocab_size: "video_modality_pixel_noise"
elif modality_type in (ModalityType.CLASS_LABEL,
ModalityType.MULTI_LABEL,
ModalityType.ONE_HOT_CLASS_LABEL):
def name(model_hparams, vocab_size):
return "class_label_modality_%d_%d" % (vocab_size,
model_hparams.hidden_size)
return name
elif modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.IDENTITY_SYMBOL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL,
ModalityType.SYMBOL_ONE_HOT):
def name(model_hparams, vocab_size):
return "symbol_modality_%d_%d" % (vocab_size, model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SIGMOID_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "sigmoid_class_symbol_modality_%d_%d" % (vocab_size,
model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "sigmoid_max_pooling_class_symbol_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "softmax_average_pooling_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "softmax_last_timestep_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)
return name
elif modality_type == ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL:
def name(model_hparams, vocab_size):
return "softmax_max_pooling_onehot_class_label_modality_%d_%d" % (
vocab_size, model_hparams.hidden_size)
return name
return value
|
[
"def",
"get_name",
"(",
"modality_type",
",",
"value",
"=",
"None",
")",
":",
"# For legacy reasons, modalities vary in their naming scheme. Future plans are",
"# to remove any need for get_name. We do not recommend using it.",
"if",
"modality_type",
"==",
"ModalityType",
".",
"AUDIO",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"audio_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"AUDIO_SPECTRAL",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"audio_spectral_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"GENERIC_L2_LOSS",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"generic_l2_loss_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IDENTITY",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"identity_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"image_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE_CHANNEL_BOTTOM_IDENTITY",
":",
"return",
"(",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"# pylint: disable=g-long-lambda",
"\"image_channel_bottom_identity_modality\"",
")",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE_CHANNEL_COMPRESS",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"image_channel_compress_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE_CHANNEL_EMBEDDINGS_BOTTOM",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"image_channel_embeddings_bottom\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"REAL",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"real_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"REAL_L2_LOSS",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"real_l2_loss_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"REAL_LOG_POISSON_LOSS",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"real_log_poisson_loss_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SPEECH_RECOGNITION",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"speech_recognition_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"video_modality\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_BITWISE",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"video_modality_bitwise\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_IDENTITY",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"video_modality_identity\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_L1",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"video_modality_l1\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_L1_RAW",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"video_modality_l1_raw\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_L2",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"video_modality_l2\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_L2_RAW",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"video_modality_l2_raw\"",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_PIXEL_NOISE",
":",
"return",
"lambda",
"model_hparams",
",",
"vocab_size",
":",
"\"video_modality_pixel_noise\"",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CLASS_LABEL",
",",
"ModalityType",
".",
"MULTI_LABEL",
",",
"ModalityType",
".",
"ONE_HOT_CLASS_LABEL",
")",
":",
"def",
"name",
"(",
"model_hparams",
",",
"vocab_size",
")",
":",
"return",
"\"class_label_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
"return",
"name",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CTC_SYMBOL",
",",
"ModalityType",
".",
"IDENTITY_SYMBOL",
",",
"ModalityType",
".",
"SYMBOL",
",",
"ModalityType",
".",
"SYMBOL_WEIGHTS_ALL",
",",
"ModalityType",
".",
"SYMBOL_ONE_HOT",
")",
":",
"def",
"name",
"(",
"model_hparams",
",",
"vocab_size",
")",
":",
"return",
"\"symbol_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
"return",
"name",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SIGMOID_CLASS_LABEL",
":",
"def",
"name",
"(",
"model_hparams",
",",
"vocab_size",
")",
":",
"return",
"\"sigmoid_class_symbol_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
"return",
"name",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SIGMOID_MAX_POOLING_CLASS_LABEL",
":",
"def",
"name",
"(",
"model_hparams",
",",
"vocab_size",
")",
":",
"return",
"\"sigmoid_max_pooling_class_symbol_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
"return",
"name",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SOFTMAX_AVERAGE_POOLING_CLASS_LABEL",
":",
"def",
"name",
"(",
"model_hparams",
",",
"vocab_size",
")",
":",
"return",
"\"softmax_average_pooling_onehot_class_label_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
"return",
"name",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SOFTMAX_LAST_TIMESTEP_CLASS_LABEL",
":",
"def",
"name",
"(",
"model_hparams",
",",
"vocab_size",
")",
":",
"return",
"\"softmax_last_timestep_onehot_class_label_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
"return",
"name",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SOFTMAX_MAX_POOLING_CLASS_LABEL",
":",
"def",
"name",
"(",
"model_hparams",
",",
"vocab_size",
")",
":",
"return",
"\"softmax_max_pooling_onehot_class_label_modality_%d_%d\"",
"%",
"(",
"vocab_size",
",",
"model_hparams",
".",
"hidden_size",
")",
"return",
"name",
"return",
"value"
] |
Gets default name for transformations; if none available, return value.
|
[
"Gets",
"default",
"name",
"for",
"transformations",
";",
"if",
"none",
"available",
"return",
"value",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1299-L1384
|
train
|
Gets default name for transformations ; if none available return value.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b11110 + 0o22) + chr(0b100110 + 0o111) + chr(0b110010) + chr(49) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1111 + 0o140) + '\x31' + chr(0b110000) + chr(52), 47917 - 47909), ehT0Px3KOsy9('\060' + chr(0b1100000 + 0o17) + '\062' + chr(1345 - 1297) + chr(51), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(1306 - 1254) + '\065', ord("\x08")), ehT0Px3KOsy9(chr(0b10001 + 0o37) + '\x6f' + chr(0b110011) + chr(0b110011) + chr(51), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(53) + chr(51), 28526 - 28518), ehT0Px3KOsy9('\x30' + '\157' + chr(51) + chr(0b10010 + 0o44) + chr(0b110001 + 0o6), 2504 - 2496), ehT0Px3KOsy9('\x30' + '\x6f' + chr(49) + '\x36' + '\x33', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101 + 0o142) + chr(0b110011) + chr(50) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(2183 - 2135) + chr(0b1101111) + chr(815 - 765) + '\062', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110001) + chr(536 - 487) + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + chr(4409 - 4298) + chr(53) + chr(49), 23145 - 23137), ehT0Px3KOsy9(chr(1348 - 1300) + '\157' + chr(1633 - 1580) + '\062', 0b1000), ehT0Px3KOsy9(chr(48) + chr(802 - 691) + '\x31' + chr(50) + '\x31', 0o10), ehT0Px3KOsy9(chr(48) + '\157' + '\061' + chr(55) + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x33' + '\x37' + '\064', ord("\x08")), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(0b1101111) + '\x36' + chr(0b1010 + 0o47), 0o10), ehT0Px3KOsy9('\x30' + chr(5836 - 5725) + '\x32' + chr(0b10010 + 0o40) + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1000001 + 0o56) + chr(578 - 523) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\065' + '\x36', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(11825 - 11714) + chr(0b11000 + 0o32) + '\067', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1001000 + 0o47) + chr(0b1101 + 0o46) + chr(53) + '\x36', 56157 - 56149), ehT0Px3KOsy9(chr(0b110000) + chr(0b111101 + 0o62) + chr(72 - 22) + '\062' + chr(0b110011), 8), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(49) + '\x32' + chr(49), 8), ehT0Px3KOsy9(chr(48) + chr(0b0 + 0o157) + chr(995 - 945) + chr(0b11101 + 0o25) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + chr(0b101100 + 0o11) + chr(0b110011), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + '\061' + chr(0b10110 + 0o41), 562 - 554), ehT0Px3KOsy9(chr(48) + chr(111) + '\061' + '\067' + chr(0b1010 + 0o55), 8), ehT0Px3KOsy9(chr(0b110000) + chr(2826 - 2715) + chr(0b110001 + 0o0) + '\x30' + '\060', 0b1000), ehT0Px3KOsy9(chr(1173 - 1125) + chr(9694 - 9583) + chr(0b110001) + chr(1341 - 1288) + chr(0b1111 + 0o50), 0o10), ehT0Px3KOsy9('\060' + chr(7875 - 7764) + '\062' + '\066' + chr(2341 - 2288), 28043 - 28035), ehT0Px3KOsy9('\060' + '\157' + chr(0b110011) + chr(0b110011) + chr(0b110100), 0b1000), ehT0Px3KOsy9('\060' + chr(0b110100 + 0o73) + chr(1109 - 1060) + chr(427 - 374) + chr(0b101000 + 0o11), 52548 - 52540), ehT0Px3KOsy9('\x30' + '\x6f' + '\062' + '\x33', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(647 - 597) + chr(0b110111) + chr(477 - 423), 56947 - 56939), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + chr(0b110100) + chr(54), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(861 - 810) + '\060' + '\x30', 6509 - 6501), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\062' + '\x36', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b11110 + 0o25) + '\x35' + chr(51), 8), ehT0Px3KOsy9('\060' + '\157' + chr(0b100111 + 0o13) + chr(0b10001 + 0o41) + chr(1079 - 1024), 59573 - 59565)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(1433 - 1385) + '\157' + '\x35' + '\060', 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xa2'), '\144' + chr(0b1100101) + '\143' + '\157' + chr(156 - 56) + chr(2356 - 2255))(chr(0b1110101) + '\x74' + '\146' + chr(45) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def IXqa55sKf4h1(N7qTs4FW6YfL, QmmgWUB13VCJ=None):
if N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcd\x11\xc7l\x94'), chr(0b1100100) + '\x65' + '\143' + chr(0b1101111) + '\x64' + '\145')(chr(3937 - 3820) + chr(854 - 738) + '\x66' + chr(0b101101) + chr(0b1000 + 0o60))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xed1\xe7L\xb4t\x9a\x1c\xb3s%@^('), chr(0b10011 + 0o121) + '\145' + '\x63' + chr(0b1100101 + 0o12) + chr(587 - 487) + chr(0b1100101))(chr(0b1110101) + '\164' + '\146' + chr(0b101101) + chr(770 - 714))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcd\x11\xc7l\x94t\xa4#\x92Q\x1d{k\x1d'), '\144' + chr(0b1011101 + 0o10) + chr(99) + chr(111) + chr(2101 - 2001) + chr(0b1100101))('\165' + chr(4017 - 3901) + chr(102) + chr(0b101101) + chr(56))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xed1\xe7L\xb4t\x84\x03\xb2q=[K= \x91^\xf7XB~\x9a\xea'), '\x64' + '\145' + chr(99) + chr(111) + chr(0b100100 + 0o100) + chr(8377 - 8276))(chr(0b1110101) + '\x74' + '\x66' + '\x2d' + chr(0b0 + 0o70))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcb\x01\xcd`\x89b\xb4,\x9b \x16ee\x02,'), chr(410 - 310) + chr(0b1001000 + 0o35) + '\143' + chr(0b110000 + 0o77) + '\144' + chr(0b1100101))('\165' + '\164' + chr(102) + '\x2d' + '\070')):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xeb!\xed@\xa9B\x94,\xbb \x16EE"\x0c\xa3\\\xfc]O{\x87\xe7('), chr(100) + chr(101) + '\143' + chr(111) + chr(9520 - 9420) + '\145')(chr(0b11111 + 0o126) + chr(0b100 + 0o160) + chr(0b1100110) + chr(0b101101) + chr(0b1110 + 0o52))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5\x00\xc6k\x8fb\xa3*'), chr(100) + '\x65' + chr(99) + '\157' + chr(869 - 769) + chr(0b1100101))('\165' + chr(0b1011000 + 0o34) + chr(0b1100110) + chr(0b11001 + 0o24) + chr(193 - 137))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xe5 \xe6K\xafB\x83\n\x88\x7f&MK=\x16\x88H'), chr(100) + chr(8755 - 8654) + '\143' + '\157' + chr(0b1100100) + chr(0b1100101))(chr(4469 - 4352) + chr(116) + '\x66' + chr(0b101101) + chr(56))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5\t\xc2b\x9e'), chr(0b1100100) + chr(0b101000 + 0o75) + chr(99) + '\x6f' + chr(0b1000000 + 0o44) + chr(0b1100101))(chr(3104 - 2987) + chr(0b111011 + 0o71) + chr(0b1100110) + chr(311 - 266) + chr(0b111000))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xe5)\xe2B\xbet\x9a\x1c\xb3s%@^('), chr(0b1100100) + chr(0b1100101) + '\143' + '\x6f' + '\144' + chr(0b1100101))(chr(117) + chr(116) + '\x66' + '\x2d' + chr(775 - 719))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5\t\xc2b\x9et\xb4;\x96\\\x07lf\x0e=\xb3e\xc7vcH\xa7\xd7\x14\xdb\xbf\x06+\x19'), '\144' + chr(0b1100101) + chr(99) + '\157' + chr(100) + chr(0b1100101))(chr(0b10100 + 0o141) + chr(116) + chr(0b1010111 + 0o17) + chr(0b101101) + chr(0b1000 + 0o60))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b"\xe5)\xe2B\xbet\x94\x1b\xb6|'LF\x0e\x1d\x93E\xe7VCH\x87\xf74\xfb\x9f&\x0b90\xd8\xb3\rr\xd2\xcf\xb4o"), '\144' + chr(8027 - 7926) + chr(0b1100011) + '\x6f' + chr(6565 - 6465) + '\145')(chr(10825 - 10708) + chr(13206 - 13090) + chr(102) + chr(45) + chr(440 - 384))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5\t\xc2b\x9et\xb4;\x96\\\x07lf\x0e<\xb3|\xc3kkD\xbd'), '\144' + chr(0b111001 + 0o54) + chr(0b1010011 + 0o20) + chr(0b1001001 + 0o46) + chr(7578 - 7478) + chr(0b1001100 + 0o31))(chr(0b1110101) + chr(0b1110100) + '\146' + chr(45) + chr(56))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b"\xe5)\xe2B\xbet\x94\x1b\xb6|'LF\x0e\x1c\x93\\\xe3KKd\x9d\xcc<\xfa\x8f.\x13)\x1b\xcc"), chr(100) + chr(5431 - 5330) + chr(0b1011 + 0o130) + chr(111) + chr(0b111101 + 0o47) + chr(0b1100101))('\165' + '\164' + chr(10183 - 10081) + '\x2d' + '\x38')
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5\t\xc2b\x9et\xb4;\x96\\\x07lf\x0e:\xb1s\xd6}j^\xa0\xd4\x02\xca\xa9\x00+\x14 \xf8'), '\144' + chr(0b111111 + 0o46) + chr(0b1001010 + 0o31) + chr(111) + '\144' + chr(0b1100101))(chr(8684 - 8567) + chr(0b1110100) + chr(0b1100110) + '\055' + '\070')):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xe5)\xe2B\xbet\x94\x1b\xb6|\'LF\x0e\x1a\x91S\xf6]J~\x80\xf4"\xca\x89 \x0b4\x00\xd8'), '\x64' + chr(0b111111 + 0o46) + '\x63' + chr(0b1101111) + chr(0b11011 + 0o111) + chr(0b1000000 + 0o45))(chr(6901 - 6784) + chr(0b1110100) + '\146' + chr(0b101101) + '\070')
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xde\x01\xc2i'), chr(0b100 + 0o140) + '\x65' + chr(0b10111 + 0o114) + chr(0b1011111 + 0o20) + '\x64' + chr(0b11101 + 0o110))('\165' + chr(8463 - 8347) + chr(0b1100110) + chr(1358 - 1313) + chr(0b101001 + 0o17))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe!\xe2I\x84F\x98\x17\xb6~ ]S'), chr(0b1100100) + '\x65' + chr(5090 - 4991) + chr(0b1000010 + 0o55) + '\144' + chr(0b1100101))(chr(0b1110101) + chr(0b1110100) + chr(0b1011000 + 0o16) + chr(914 - 869) + chr(1950 - 1894))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xde\x01\xc2i\x84g\xc5,\x9b]\x1az'), chr(7523 - 7423) + '\x65' + chr(0b110111 + 0o54) + chr(5802 - 5691) + chr(248 - 148) + chr(0b1011000 + 0o15))(chr(0b101100 + 0o111) + chr(0b100 + 0o160) + chr(0b1000000 + 0o46) + '\x2d' + '\070')):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe!\xe2I\x84G\xc5,\xbb}:Zu<\x10\x98P\xffPZn'), chr(7336 - 7236) + chr(0b100100 + 0o101) + chr(0b1100011) + '\x6f' + chr(100) + chr(0b1100101))('\x75' + '\164' + chr(0b11010 + 0o114) + chr(45) + '\070')
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xde\x01\xc2i\x84g\xb84\x88B\x06`y\x020\xb2n\xdfv}D'), chr(0b1100100) + '\145' + chr(99) + '\157' + chr(7411 - 7311) + '\145')(chr(0b1101 + 0o150) + '\x74' + chr(0b1100110) + chr(0b101101) + chr(56))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe!\xe2I\x84G\x98\x14\x88b&@Y"\x10\x92n\xffV]d\xb1\xfe>\xf1\x8a#\x164\x16'), chr(0b1100100) + '\x65' + chr(0b1100011) + '\157' + '\144' + chr(9175 - 9074))(chr(12673 - 12556) + chr(116) + chr(102) + '\x2d' + chr(2276 - 2220))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x14\xc6`\x98c\xa8!\x92Q\x06nd\x18+\xb5~\xdd'), '\x64' + chr(101) + chr(99) + chr(0b1100111 + 0o10) + chr(8964 - 8864) + chr(1892 - 1791))(chr(1885 - 1768) + chr(0b110010 + 0o102) + '\146' + chr(0b101101) + '\070')):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xff4\xe6@\xb8C\xa8\x01\xb2q&ND8\x0b\x95^\xfdfCx\x8a\xf2=\xfc\x9f6'), chr(0b1100100) + chr(0b1100101) + chr(0b1100011) + chr(10887 - 10776) + chr(100) + chr(1156 - 1055))(chr(117) + '\164' + chr(0b1001 + 0o135) + '\055' + '\070')
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xda\r\xc7`\x94'), chr(4149 - 4049) + '\x65' + chr(0b1100011) + chr(7655 - 7544) + chr(9900 - 9800) + chr(101))(chr(6746 - 6629) + '\164' + chr(0b1100110) + chr(0b111 + 0o46) + chr(1249 - 1193))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfa-\xe7@\xb4t\x9a\x1c\xb3s%@^('), chr(0b1000000 + 0o44) + '\145' + chr(0b1100 + 0o127) + chr(0b1101111) + '\144' + chr(0b1100101))(chr(0b1110101) + chr(0b1110100) + '\146' + '\055' + chr(0b111000))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xda\r\xc7`\x94t\xb5:\x83E\x00zo'), chr(0b1101 + 0o127) + '\x65' + chr(99) + chr(6032 - 5921) + chr(0b111010 + 0o52) + chr(101))('\x75' + '\x74' + chr(0b1001000 + 0o36) + chr(45) + chr(56))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfa-\xe7@\xb4t\x9a\x1c\xb3s%@^( \x9eX\xe7NGd\x8b'), '\144' + chr(9068 - 8967) + '\x63' + chr(0b1101111) + chr(0b1011111 + 0o5) + chr(101))('\165' + '\x74' + '\x66' + '\055' + '\070')
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xda\r\xc7`\x94t\xbe7\x92\\\x1d`~\x08'), chr(0b1011101 + 0o7) + chr(0b11 + 0o142) + chr(0b1100011) + '\x6f' + chr(100) + '\x65')(chr(0b1011010 + 0o33) + '\x74' + '\146' + chr(45) + chr(1014 - 958))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfa-\xe7@\xb4t\x9a\x1c\xb3s%@^( \x95U\xf6WZ~\x9a\xea'), chr(100) + chr(0b111110 + 0o47) + chr(0b1100011) + chr(4843 - 4732) + '\144' + chr(9143 - 9042))(chr(5832 - 5715) + chr(116) + '\146' + '\055' + chr(0b101011 + 0o15))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xda\r\xc7`\x94t\xbbB'), chr(100) + chr(6892 - 6791) + chr(99) + '\x6f' + chr(0b1100100) + chr(0b1011100 + 0o11))(chr(0b110110 + 0o77) + '\x74' + chr(0b1100110) + chr(841 - 796) + '\070')):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfa-\xe7@\xb4t\x9a\x1c\xb3s%@^( \x90\x00'), chr(0b1100100) + '\x65' + '\143' + chr(0b1010001 + 0o36) + chr(0b1100100) + chr(101))(chr(117) + chr(0b10000 + 0o144) + chr(0b1000001 + 0o45) + '\x2d' + chr(0b100011 + 0o25))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xda\r\xc7`\x94t\xbbB\x88@\x08~'), chr(0b1100100) + chr(3681 - 3580) + '\143' + chr(0b1101100 + 0o3) + chr(0b1100100) + '\x65')(chr(0b111111 + 0o66) + '\x74' + chr(4296 - 4194) + chr(188 - 143) + chr(0b111000))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfa-\xe7@\xb4t\x9a\x1c\xb3s%@^( \x90\x00\xccKO`'), chr(0b111101 + 0o47) + '\x65' + chr(0b1100011) + '\157' + '\144' + chr(0b1100101))(chr(12009 - 11892) + chr(0b1110100) + chr(102) + chr(0b100001 + 0o14) + chr(0b111000))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xda\r\xc7`\x94t\xbbA'), chr(0b1100100) + chr(867 - 766) + chr(0b111 + 0o134) + chr(5138 - 5027) + chr(8089 - 7989) + chr(0b110010 + 0o63))(chr(0b1110101) + chr(5643 - 5527) + '\x66' + chr(45) + chr(0b111000))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfa-\xe7@\xb4t\x9a\x1c\xb3s%@^( \x90\x03'), chr(6418 - 6318) + '\x65' + '\x63' + chr(0b1101111) + chr(0b1100100) + '\x65')(chr(10155 - 10038) + chr(10419 - 10303) + chr(9849 - 9747) + '\x2d' + chr(0b11100 + 0o34))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xda\r\xc7`\x94t\xbbA\x88@\x08~'), chr(0b1001001 + 0o33) + chr(0b1010111 + 0o16) + chr(9415 - 9316) + chr(111) + chr(100) + chr(0b1011000 + 0o15))('\165' + chr(0b1110100) + '\x66' + chr(0b101000 + 0o5) + '\070')):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfa-\xe7@\xb4t\x9a\x1c\xb3s%@^( \x90\x03\xccKO`'), '\144' + chr(0b110100 + 0o61) + '\143' + chr(111) + '\144' + chr(0b1100011 + 0o2))('\165' + chr(116) + chr(9266 - 9164) + chr(77 - 32) + chr(545 - 489))
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xda\r\xc7`\x94t\xa7:\x8fW\x05vd\x1e6\xaft'), chr(0b1100100) + '\145' + '\143' + '\x6f' + chr(0b1100100) + '\145')(chr(4523 - 4406) + '\x74' + chr(0b1001110 + 0o30) + '\x2d' + chr(675 - 619))):
return lambda tq24Tk6UZ6u1, CeyMIoSyrpkQ: xafqLlk3kkUe(SXOLrMavuUCe(b'\xfa-\xe7@\xb4t\x9a\x1c\xb3s%@^( \x8cX\xeb\\BH\x80\xfc8\xe6\x8e'), chr(0b1100100) + '\x65' + '\143' + '\157' + '\x64' + '\145')('\x75' + chr(0b1110100) + chr(0b1100110) + chr(0b101000 + 0o5) + chr(56))
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcf\x08\xc2v\x88t\xbb2\x95W\x05'), chr(8199 - 8099) + chr(8717 - 8616) + chr(0b1100011) + chr(10103 - 9992) + '\x64' + chr(0b1100101))(chr(5552 - 5435) + chr(0b1110100) + chr(0b1100110) + '\x2d' + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc1\x11\xcfq\x92t\xbb2\x95W\x05'), '\144' + chr(0b1100101) + '\x63' + '\157' + chr(100) + chr(0b1111 + 0o126))('\165' + '\x74' + '\146' + chr(0b11011 + 0o22) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc3\n\xc6z\x93d\xa3,\x94^\x08zy\x0e3\xbds\xd6u'), '\x64' + chr(0b1100101) + '\143' + chr(0b1101111) + '\x64' + chr(8889 - 8788))('\x75' + chr(0b1110100) + '\146' + '\x2d' + '\x38'))):
def AIvJRzLdDfgF(tq24Tk6UZ6u1, CeyMIoSyrpkQ):
return xafqLlk3kkUe(SXOLrMavuUCe(b'\xef(\xe2V\xa8t\x9b\x12\xb5w%vG>\x1b\x9d]\xfaMWH\xcb\xf7\x0e\xb0\x8f'), '\144' + chr(101) + '\x63' + '\x6f' + '\144' + chr(4832 - 4731))('\x75' + chr(8159 - 8043) + chr(102) + chr(45) + chr(56)) % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd>\xec\\\x83e\xc4\x18\xb3z\re'), chr(1972 - 1872) + chr(101) + '\x63' + '\x6f' + '\144' + chr(0b1100101))(chr(0b1110101) + chr(0b1011110 + 0o26) + chr(0b11110 + 0o110) + '\055' + chr(1283 - 1227))))
return AIvJRzLdDfgF
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcf\x10\xc0z\x88r\xba1\x98^'), chr(0b1100100) + '\x65' + '\143' + '\x6f' + chr(0b1100100) + chr(0b100110 + 0o77))(chr(0b100011 + 0o122) + '\x74' + chr(0b1011000 + 0o16) + '\055' + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xc5\x00\xc6k\x8fb\xa3*\x88A\x10dh\x1e3'), chr(0b101 + 0o137) + '\x65' + chr(3966 - 3867) + chr(0b110 + 0o151) + chr(0b1100100) + chr(0b1100101))(chr(0b1100100 + 0o21) + chr(0b1 + 0o163) + '\146' + '\055' + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x1d\xceg\x94g'), '\x64' + chr(0b1100101) + chr(0b1000100 + 0o37) + chr(0b11111 + 0o120) + chr(100) + chr(101))('\x75' + '\x74' + chr(9744 - 9642) + chr(1952 - 1907) + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x1d\xceg\x94g\xa8$\x92[\x0ea~\x02 \xbd}\xdf'), chr(0b1001011 + 0o31) + '\x65' + chr(99) + chr(111) + '\144' + chr(0b1100101))(chr(7420 - 7303) + chr(9832 - 9716) + chr(0b1100110) + '\055' + chr(1089 - 1033))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x1d\xceg\x94g\xa8<\x99W\x16ae\x05'), '\144' + chr(0b1001000 + 0o35) + '\143' + chr(111) + chr(0b1100100) + chr(0b1100101))(chr(117) + chr(116) + '\x66' + chr(45) + chr(0b111000)))):
def AIvJRzLdDfgF(tq24Tk6UZ6u1, CeyMIoSyrpkQ):
return xafqLlk3kkUe(SXOLrMavuUCe(b'\xff=\xeeG\xb4G\xa8\x1e\xb8v(EC%\x06\xa3\x14\xf7f\x0bs'), '\144' + chr(0b1000110 + 0o37) + chr(5906 - 5807) + '\x6f' + chr(5044 - 4944) + chr(3788 - 3687))('\165' + chr(0b1110100) + chr(102) + chr(0b1110 + 0o37) + '\070') % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd>\xec\\\x83e\xc4\x18\xb3z\re'), '\x64' + '\145' + '\x63' + chr(1591 - 1480) + '\x64' + chr(101))(chr(0b11101 + 0o130) + '\x74' + chr(0b1100110) + '\055' + chr(0b110 + 0o62))))
return AIvJRzLdDfgF
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\r\xc4h\x94b\xb3,\x94^\x08zy\x0e3\xbds\xd6u'), '\x64' + chr(1468 - 1367) + '\143' + chr(111) + '\x64' + '\x65')(chr(117) + '\164' + '\146' + '\x2d' + chr(0b100000 + 0o30))):
def AIvJRzLdDfgF(tq24Tk6UZ6u1, CeyMIoSyrpkQ):
return xafqLlk3kkUe(SXOLrMavuUCe(b'\xff-\xe4H\xb4B\x93,\xb4~(ZY\x0e\x0c\x85\\\xf1VBH\x83\xfc5\xf4\x87&\x0b90\x90\xb866\xda'), chr(0b101001 + 0o73) + chr(0b11101 + 0o110) + chr(596 - 497) + chr(0b1101111) + chr(0b101111 + 0o65) + chr(101))(chr(117) + chr(0b1000001 + 0o63) + '\146' + chr(0b10011 + 0o32) + chr(713 - 657)) % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd>\xec\\\x83e\xc4\x18\xb3z\re'), chr(0b1100100) + '\145' + '\x63' + chr(9505 - 9394) + '\x64' + chr(7398 - 7297))('\x75' + '\x74' + '\146' + chr(1935 - 1890) + chr(0b11 + 0o65))))
return AIvJRzLdDfgF
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\r\xc4h\x94b\xb3,\x9aS\x11vz\x1e0\xb0x\xdd~qT\xa2\xd2\x02\xc6\xb4\x03>\x02*\xf9'), chr(100) + chr(0b111001 + 0o54) + '\x63' + '\157' + chr(0b1100100) + chr(718 - 617))(chr(9107 - 8990) + chr(116) + chr(102) + chr(918 - 873) + chr(0b111000))):
def AIvJRzLdDfgF(tq24Tk6UZ6u1, CeyMIoSyrpkQ):
return xafqLlk3kkUe(SXOLrMavuUCe(b'\xff-\xe4H\xb4B\x93,\xbas1vZ>\x10\x90X\xfd^qt\x82\xf2"\xe6\xb4<\x06-\r\xda\xb06~\xd1\xc2\xa1z\x82\xe3\xf5\x1b\xa6A\x84\x0e\x93'), chr(100) + '\145' + chr(0b1100011) + '\x6f' + chr(0b1100100) + chr(0b1100011 + 0o2))('\x75' + chr(116) + chr(102) + chr(45) + chr(0b1111 + 0o51)) % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd>\xec\\\x83e\xc4\x18\xb3z\re'), '\144' + '\145' + chr(99) + '\157' + chr(100) + chr(0b1100101))(chr(117) + chr(5337 - 5221) + '\x66' + '\x2d' + '\070')))
return AIvJRzLdDfgF
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x0b\xc5q\x96j\xaf,\x96D\x0c{k\x16:\xa3a\xdcvb^\xa0\xd4\x0e\xd6\xa7\x0e,\x130\xf9\x9d+V\xf2'), '\x64' + chr(101) + chr(99) + chr(0b1101111) + chr(100) + '\x65')(chr(11796 - 11679) + chr(6487 - 6371) + chr(0b1100110) + '\055' + chr(0b111000))):
def AIvJRzLdDfgF(tq24Tk6UZ6u1, CeyMIoSyrpkQ):
return xafqLlk3kkUe(SXOLrMavuUCe(b'\xff+\xe5Q\xb6J\x8f,\xb6d,[K6\x1a\xa3A\xfcVB~\x80\xf4\x0e\xfa\x85*\x17/\x1b\xea\xbf\x05r\xcd\xd5\x9fz\x8a\xf5\xe9(\xdcH\xb4O\x96\x1f\xbef0v\x0f5 \xd9U'), chr(3461 - 3361) + chr(0b1100101) + '\143' + '\157' + '\x64' + '\145')(chr(117) + chr(12122 - 12006) + chr(0b1100110) + chr(795 - 750) + '\070') % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd>\xec\\\x83e\xc4\x18\xb3z\re'), chr(0b1011010 + 0o12) + chr(4323 - 4222) + chr(99) + chr(0b1101111) + '\x64' + chr(0b1010010 + 0o23))(chr(117) + chr(116) + chr(0b11011 + 0o113) + '\055' + chr(56))))
return AIvJRzLdDfgF
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x0b\xc5q\x96j\xaf,\x9bS\x1a}u\x056\xb1t\xc0mkG\xb1\xd0\x1d\xd4\xb8\x1c \x0c.\xf7\x99%'), chr(0b1001010 + 0o32) + chr(0b1100100 + 0o1) + chr(0b1010110 + 0o15) + '\x6f' + chr(3834 - 3734) + '\x65')(chr(3532 - 3415) + chr(0b1110100) + '\146' + chr(0b100010 + 0o13) + '\070')):
def AIvJRzLdDfgF(tq24Tk6UZ6u1, CeyMIoSyrpkQ):
return xafqLlk3kkUe(SXOLrMavuUCe(b'\xff+\xe5Q\xb6J\x8f,\xbbs:]u%\x16\x91T\xe0MKg\xb1\xfc?\xf0\x83 \x0b\x1f\x0c\xd9\xbd\x1a`\xe1\xca\xa1t\x8e\xfb\xd3)\xecA\xbaG\x9e\x07\xaeMlMut\x1b'), chr(0b10001 + 0o123) + chr(0b10001 + 0o124) + chr(8521 - 8422) + chr(111) + chr(0b1100100) + chr(101))(chr(0b1110101) + chr(116) + chr(0b1010101 + 0o21) + chr(0b101101) + chr(0b100111 + 0o21)) % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd>\xec\\\x83e\xc4\x18\xb3z\re'), chr(0b1000110 + 0o36) + '\145' + '\x63' + chr(0b1101111) + chr(7492 - 7392) + '\145')(chr(117) + '\164' + chr(8091 - 7989) + '\x2d' + chr(0b11011 + 0o35))))
return AIvJRzLdDfgF
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdf\x0b\xc5q\x96j\xaf,\x9aS\x11vz\x1e0\xb0x\xdd~qT\xa2\xd2\x02\xc6\xb4\x03>\x02*\xf9'), chr(517 - 417) + chr(0b1100101) + chr(99) + '\157' + chr(100) + chr(0b1100101))('\x75' + '\x74' + chr(0b1011100 + 0o12) + '\x2d' + '\070')):
def AIvJRzLdDfgF(tq24Tk6UZ6u1, CeyMIoSyrpkQ):
return xafqLlk3kkUe(SXOLrMavuUCe(b'\xff+\xe5Q\xb6J\x8f,\xbas1vZ>\x10\x90X\xfd^qx\x80\xf69\xfa\x9f\x10\x1c,\x0e\xc6\xaf6\x7f\xdf\xc4\xa5z\xb4\xfa\xe3 \xe2I\xb2_\x8e,\xf2v\x16\x0cN'), '\x64' + chr(0b100101 + 0o100) + chr(8539 - 8440) + chr(0b111011 + 0o64) + chr(100) + chr(0b1100101))(chr(117) + chr(116) + chr(0b1100110) + chr(0b101101) + chr(0b111000)) % (CeyMIoSyrpkQ, xafqLlk3kkUe(tq24Tk6UZ6u1, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfd>\xec\\\x83e\xc4\x18\xb3z\re'), '\x64' + chr(101) + '\143' + '\x6f' + '\x64' + '\x65')(chr(0b101001 + 0o114) + chr(0b101111 + 0o105) + chr(0b1100110) + chr(0b101101) + chr(0b111000))))
return AIvJRzLdDfgF
return QmmgWUB13VCJ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
get_targets_bottom
|
def get_targets_bottom(modality_type, value=None):
"""Gets default bottom transformation for targets; if none, return value."""
if modality_type == ModalityType.AUDIO:
return make_targets_bottom(audio_bottom)
elif modality_type == ModalityType.AUDIO_SPECTRAL:
return make_targets_bottom(audio_spectral_bottom)
elif modality_type in (ModalityType.CLASS_LABEL,
ModalityType.MULTI_LABEL,
ModalityType.ONE_HOT_CLASS_LABEL,
ModalityType.SIGMOID_CLASS_LABEL,
ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL,
ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL):
return class_label_targets_bottom
elif modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL):
return symbol_targets_bottom
elif modality_type in (ModalityType.GENERIC_L2_LOSS,
ModalityType.IDENTITY_SYMBOL):
return identity_bottom
elif modality_type == ModalityType.IDENTITY:
return make_targets_bottom(identity_bottom)
elif modality_type == ModalityType.IMAGE:
return image_targets_bottom
elif modality_type in (ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY,
ModalityType.IMAGE_CHANNEL_COMPRESS):
return image_channel_compress_targets_bottom
elif modality_type == ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM:
return image_channel_embeddings_bottom
elif modality_type in (ModalityType.REAL,
ModalityType.REAL_L2_LOSS,
ModalityType.REAL_LOG_POISSON_LOSS):
return make_targets_bottom(real_bottom)
elif modality_type == ModalityType.SPEECH_RECOGNITION:
return make_targets_bottom(speech_recognition_bottom)
elif modality_type == ModalityType.SYMBOL_ONE_HOT:
return symbol_one_hot_bottom
elif modality_type in (ModalityType.VIDEO,
ModalityType.VIDEO_L1,
ModalityType.VIDEO_L2):
return video_targets_bottom
elif modality_type == ModalityType.VIDEO_BITWISE:
return video_bitwise_targets_bottom
elif modality_type == ModalityType.VIDEO_IDENTITY:
return video_identity_targets_bottom
elif modality_type in (ModalityType.VIDEO_L1_RAW,
ModalityType.VIDEO_L2_RAW):
return video_raw_targets_bottom
elif modality_type == ModalityType.VIDEO_PIXEL_NOISE:
return make_targets_bottom(video_pixel_noise_bottom)
return value
|
python
|
def get_targets_bottom(modality_type, value=None):
"""Gets default bottom transformation for targets; if none, return value."""
if modality_type == ModalityType.AUDIO:
return make_targets_bottom(audio_bottom)
elif modality_type == ModalityType.AUDIO_SPECTRAL:
return make_targets_bottom(audio_spectral_bottom)
elif modality_type in (ModalityType.CLASS_LABEL,
ModalityType.MULTI_LABEL,
ModalityType.ONE_HOT_CLASS_LABEL,
ModalityType.SIGMOID_CLASS_LABEL,
ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL,
ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL,
ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL):
return class_label_targets_bottom
elif modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL):
return symbol_targets_bottom
elif modality_type in (ModalityType.GENERIC_L2_LOSS,
ModalityType.IDENTITY_SYMBOL):
return identity_bottom
elif modality_type == ModalityType.IDENTITY:
return make_targets_bottom(identity_bottom)
elif modality_type == ModalityType.IMAGE:
return image_targets_bottom
elif modality_type in (ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY,
ModalityType.IMAGE_CHANNEL_COMPRESS):
return image_channel_compress_targets_bottom
elif modality_type == ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM:
return image_channel_embeddings_bottom
elif modality_type in (ModalityType.REAL,
ModalityType.REAL_L2_LOSS,
ModalityType.REAL_LOG_POISSON_LOSS):
return make_targets_bottom(real_bottom)
elif modality_type == ModalityType.SPEECH_RECOGNITION:
return make_targets_bottom(speech_recognition_bottom)
elif modality_type == ModalityType.SYMBOL_ONE_HOT:
return symbol_one_hot_bottom
elif modality_type in (ModalityType.VIDEO,
ModalityType.VIDEO_L1,
ModalityType.VIDEO_L2):
return video_targets_bottom
elif modality_type == ModalityType.VIDEO_BITWISE:
return video_bitwise_targets_bottom
elif modality_type == ModalityType.VIDEO_IDENTITY:
return video_identity_targets_bottom
elif modality_type in (ModalityType.VIDEO_L1_RAW,
ModalityType.VIDEO_L2_RAW):
return video_raw_targets_bottom
elif modality_type == ModalityType.VIDEO_PIXEL_NOISE:
return make_targets_bottom(video_pixel_noise_bottom)
return value
|
[
"def",
"get_targets_bottom",
"(",
"modality_type",
",",
"value",
"=",
"None",
")",
":",
"if",
"modality_type",
"==",
"ModalityType",
".",
"AUDIO",
":",
"return",
"make_targets_bottom",
"(",
"audio_bottom",
")",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"AUDIO_SPECTRAL",
":",
"return",
"make_targets_bottom",
"(",
"audio_spectral_bottom",
")",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CLASS_LABEL",
",",
"ModalityType",
".",
"MULTI_LABEL",
",",
"ModalityType",
".",
"ONE_HOT_CLASS_LABEL",
",",
"ModalityType",
".",
"SIGMOID_CLASS_LABEL",
",",
"ModalityType",
".",
"SIGMOID_MAX_POOLING_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_AVERAGE_POOLING_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_LAST_TIMESTEP_CLASS_LABEL",
",",
"ModalityType",
".",
"SOFTMAX_MAX_POOLING_CLASS_LABEL",
")",
":",
"return",
"class_label_targets_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CTC_SYMBOL",
",",
"ModalityType",
".",
"SYMBOL",
",",
"ModalityType",
".",
"SYMBOL_WEIGHTS_ALL",
")",
":",
"return",
"symbol_targets_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"GENERIC_L2_LOSS",
",",
"ModalityType",
".",
"IDENTITY_SYMBOL",
")",
":",
"return",
"identity_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IDENTITY",
":",
"return",
"make_targets_bottom",
"(",
"identity_bottom",
")",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE",
":",
"return",
"image_targets_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"IMAGE_CHANNEL_BOTTOM_IDENTITY",
",",
"ModalityType",
".",
"IMAGE_CHANNEL_COMPRESS",
")",
":",
"return",
"image_channel_compress_targets_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE_CHANNEL_EMBEDDINGS_BOTTOM",
":",
"return",
"image_channel_embeddings_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"REAL",
",",
"ModalityType",
".",
"REAL_L2_LOSS",
",",
"ModalityType",
".",
"REAL_LOG_POISSON_LOSS",
")",
":",
"return",
"make_targets_bottom",
"(",
"real_bottom",
")",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SPEECH_RECOGNITION",
":",
"return",
"make_targets_bottom",
"(",
"speech_recognition_bottom",
")",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SYMBOL_ONE_HOT",
":",
"return",
"symbol_one_hot_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"VIDEO",
",",
"ModalityType",
".",
"VIDEO_L1",
",",
"ModalityType",
".",
"VIDEO_L2",
")",
":",
"return",
"video_targets_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_BITWISE",
":",
"return",
"video_bitwise_targets_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_IDENTITY",
":",
"return",
"video_identity_targets_bottom",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"VIDEO_L1_RAW",
",",
"ModalityType",
".",
"VIDEO_L2_RAW",
")",
":",
"return",
"video_raw_targets_bottom",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"VIDEO_PIXEL_NOISE",
":",
"return",
"make_targets_bottom",
"(",
"video_pixel_noise_bottom",
")",
"return",
"value"
] |
Gets default bottom transformation for targets; if none, return value.
|
[
"Gets",
"default",
"bottom",
"transformation",
"for",
"targets",
";",
"if",
"none",
"return",
"value",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1387-L1439
|
train
|
Gets default bottom transformation for targets ; if none return value.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b1001 + 0o146) + chr(49) + chr(0b110000) + '\x30', 39692 - 39684), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b11110 + 0o24) + chr(1677 - 1623) + chr(0b110011), 57870 - 57862), ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\x6f' + chr(51) + chr(0b110100) + '\x34', 60530 - 60522), ehT0Px3KOsy9(chr(48) + '\157' + '\062' + chr(0b10001 + 0o41) + chr(0b110000 + 0o7), ord("\x08")), ehT0Px3KOsy9(chr(201 - 153) + chr(1536 - 1425) + chr(0b1011 + 0o46) + chr(689 - 639) + chr(0b100 + 0o63), 0o10), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\x6f' + chr(51) + chr(0b110010) + chr(48), 0o10), ehT0Px3KOsy9(chr(1364 - 1316) + chr(7127 - 7016) + '\061' + chr(79 - 28) + '\x35', 0b1000), ehT0Px3KOsy9(chr(1230 - 1182) + chr(0b1100011 + 0o14) + '\063' + chr(0b1001 + 0o47), 0o10), ehT0Px3KOsy9(chr(294 - 246) + chr(0b1101111) + chr(1927 - 1874) + '\x34', 57587 - 57579), ehT0Px3KOsy9('\x30' + chr(10384 - 10273) + chr(51) + '\065' + '\061', 0o10), ehT0Px3KOsy9('\060' + chr(0b1001110 + 0o41) + chr(50) + chr(1939 - 1886) + '\060', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + chr(0b100010 + 0o23) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(0b1101111) + chr(50) + chr(0b110111) + chr(50), 39733 - 39725), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + chr(0b110001) + '\061', 56679 - 56671), ehT0Px3KOsy9('\x30' + '\157' + chr(1875 - 1826) + chr(54) + '\x34', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + chr(0b10000 + 0o44) + chr(1625 - 1575), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(55), 0o10), ehT0Px3KOsy9('\060' + chr(1318 - 1207) + chr(49) + '\064' + '\x32', 8), ehT0Px3KOsy9(chr(48) + chr(0b11101 + 0o122) + chr(49) + chr(2560 - 2507) + chr(1834 - 1779), 0o10), ehT0Px3KOsy9(chr(485 - 437) + '\157' + chr(0b110001) + chr(52) + chr(1175 - 1120), ord("\x08")), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(5210 - 5099) + chr(50) + chr(1451 - 1396) + '\x33', 30273 - 30265), ehT0Px3KOsy9('\060' + '\157' + chr(1666 - 1617) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(1064 - 1016) + '\x6f' + chr(0b110011) + '\x36' + chr(0b100 + 0o63), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + '\061' + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(2049 - 2001) + '\157' + '\x31' + chr(0b100111 + 0o15), 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + chr(0b100 + 0o55) + chr(51), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + '\x32' + chr(0b110010) + chr(0b101100 + 0o5), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + '\062' + chr(618 - 564), 0b1000), ehT0Px3KOsy9(chr(0b10100 + 0o34) + '\157' + '\x31' + chr(0b100000 + 0o22) + chr(0b110010), 29267 - 29259), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110011) + '\060' + '\x32', 42115 - 42107), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x32' + chr(380 - 329) + '\x34', 53367 - 53359), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(122 - 74) + chr(0b1101111) + chr(0b10100 + 0o37) + chr(1158 - 1110) + '\x34', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(10889 - 10778) + chr(1914 - 1863) + '\x35' + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(2072 - 2024) + chr(111) + chr(2374 - 2323) + '\x33' + chr(50), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(73 - 24) + '\066' + chr(51), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b100001 + 0o20) + chr(0b10100 + 0o41) + chr(55), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b101011 + 0o7) + chr(0b110110) + chr(2593 - 2542), 8), ehT0Px3KOsy9(chr(48) + chr(0b1010000 + 0o37) + chr(54) + '\067', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b101011 + 0o7) + chr(2089 - 2040) + '\065', 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(371 - 323) + chr(9465 - 9354) + '\065' + '\x30', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'n'), '\144' + chr(101) + chr(0b1100011) + chr(0b1101111) + '\x64' + chr(0b101101 + 0o70))(chr(7445 - 7328) + '\x74' + chr(0b1100110) + '\x2d' + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def nBq0m3bO2snh(N7qTs4FW6YfL, QmmgWUB13VCJ=None):
if N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x01\xe6\xa0\xde\x18'), chr(7999 - 7899) + '\x65' + chr(2104 - 2005) + '\x6f' + '\x64' + chr(101))(chr(961 - 844) + '\x74' + chr(4403 - 4301) + '\055' + chr(56))):
return iA8F85yLVO8w(a1v03mh4yIkx)
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x01\xe6\xa0\xde\x18\x8f\x0eHi\xaa\xfc\xf8\xfb\x05'), chr(0b1001001 + 0o33) + chr(101) + '\x63' + '\x6f' + chr(0b11001 + 0o113) + '\145')('\x75' + '\x74' + chr(0b1100110) + chr(0b101101) + '\070')):
return iA8F85yLVO8w(fjzC10DtfeYX)
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x03\xff\xa5\xc4\x04\x8f\x11Yn\xac\xe4'), chr(1026 - 926) + chr(7968 - 7867) + '\x63' + chr(111) + '\144' + chr(0b1100101))(chr(0b1010 + 0o153) + chr(116) + chr(102) + '\x2d' + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\r\xe6\xa8\xc3\x1e\x8f\x11Yn\xac\xe4'), '\144' + chr(2961 - 2860) + chr(0b110000 + 0o63) + '\157' + '\144' + '\145')(chr(0b110111 + 0o76) + '\x74' + '\x66' + chr(0b1100 + 0o41) + chr(3131 - 3075))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b"\x0f\xfd\xa1\xc8\x1f\x9f\tGo\xa5\xe9\xf9\xe9\x168#E'#"), chr(4264 - 4164) + chr(0b1100101) + '\x63' + '\x6f' + chr(100) + chr(0b1010011 + 0o22))(chr(117) + chr(116) + chr(0b110100 + 0o62) + chr(0b101101) + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b"\x13\xfa\xa3\xda\x18\x99\x19Go\xa5\xe9\xf9\xe9\x168#E'#"), chr(3142 - 3042) + chr(101) + chr(0b1100011) + chr(1567 - 1456) + chr(100) + chr(0b1100101))(chr(0b10000 + 0o145) + chr(0b110011 + 0o101) + chr(0b1000001 + 0o45) + chr(666 - 621) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xfa\xa3\xda\x18\x99\x19Ga\xa8\xf0\xf5\xea\x06;.N,(8\xf8@\xb6X_\xd4\xdd\x97\x10\x87\xd0'), '\x64' + chr(101) + chr(0b10000 + 0o123) + chr(0b1101111) + chr(2423 - 2323) + chr(101))(chr(0b1101 + 0o150) + chr(0b1010010 + 0o42) + chr(102) + '\x2d' + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xfc\xa2\xc3\x1a\x91\x05Gm\xbf\xed\xf8\xfb\x0e1=W- +\xf2B\xb0TO\xc7\xd0\x85\x01\x9d\xd0]\x86\xab\x96'), '\x64' + chr(8092 - 7991) + '\x63' + chr(0b1101111) + chr(0b1100100) + chr(0b101 + 0o140))(chr(0b1100000 + 0o25) + '\x74' + '\146' + chr(0b101101) + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xfc\xa2\xc3\x1a\x91\x05G`\xa8\xfb\xfe\xe5\x1d=/B1;"\xebS\xb4GM\xd8\xc2\x89\x1e\x83\xdeY\x88'), chr(3026 - 2926) + chr(0b1100101) + chr(99) + chr(0b1101111) + '\144' + chr(0b1100101))(chr(117) + chr(116) + '\146' + '\x2d' + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xfc\xa2\xc3\x1a\x91\x05Ga\xa8\xf0\xf5\xea\x06;.N,(8\xf8@\xb6X_\xd4\xdd\x97\x10\x87\xd0'), '\x64' + '\145' + '\x63' + '\157' + chr(324 - 224) + chr(1845 - 1744))(chr(0b1110101) + chr(0b1110100) + chr(237 - 135) + '\055' + chr(1591 - 1535)))):
return GNxTAcXheok_
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x03\xe7\xa7\xc8\x04\x89\x10Zc\xa5'), chr(2464 - 2364) + chr(0b1100101) + chr(99) + chr(0b100 + 0o153) + '\144' + chr(0b1100101))('\165' + '\164' + '\x66' + chr(45) + chr(910 - 854))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xea\xa9\xd5\x18\x9c'), chr(7937 - 7837) + chr(101) + chr(0b1010011 + 0o20) + '\x6f' + chr(0b1010001 + 0o23) + chr(101))(chr(4544 - 4427) + chr(6815 - 6699) + chr(0b1100110) + '\x2d' + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xea\xa9\xd5\x18\x9c\x02Oi\xa0\xef\xe2\xee\x1a+#K.'), chr(0b1100100) + '\x65' + chr(99) + '\157' + chr(0b111100 + 0o50) + chr(0b1100101))('\x75' + '\x74' + chr(0b10001 + 0o125) + '\055' + chr(56)))):
return JC4EGLDtEviX
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b"\x07\xf6\xaa\xd2\x05\x99\x1eG`\xdb\xf7\xe6\xf5\x1a'"), chr(100) + '\145' + chr(99) + chr(5664 - 5553) + chr(0b11 + 0o141) + '\x65')(chr(0b1000000 + 0o65) + chr(0b1110100) + chr(0b1100110) + chr(45) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\t\xf7\xa1\xd9\x03\x99\tAs\xba\xf1\xe7\xf8\x068'), chr(0b1100100) + chr(0b1100101) + chr(0b1010010 + 0o21) + '\157' + chr(1458 - 1358) + chr(0b1100101))(chr(4227 - 4110) + chr(6372 - 6256) + chr(0b100011 + 0o103) + chr(0b101101) + chr(2715 - 2659)))):
return I4Q6CFwoHJ21
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\t\xf7\xa1\xd9\x03\x99\tA'), chr(1260 - 1160) + chr(9344 - 9243) + '\x63' + '\157' + '\x64' + chr(0b1010110 + 0o17))(chr(117) + chr(5116 - 5000) + '\146' + '\x2d' + chr(619 - 563))):
return iA8F85yLVO8w(I4Q6CFwoHJ21)
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\t\xfe\xa5\xd0\x12'), chr(0b1100100) + chr(1401 - 1300) + chr(0b1100011) + chr(111) + chr(0b1100100) + chr(3939 - 3838))(chr(0b1110101) + '\x74' + chr(0b1011110 + 0o10) + chr(0b101101) + '\x38')):
return h7VqVNsx2j8T
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\t\xfe\xa5\xd0\x12\x8f\x1ePm\xa7\xe6\xef\xf6\x166-S6 *\xe4E\xb3NB\xdf\xd8\x82\x0b'), chr(100) + '\145' + chr(3823 - 3724) + '\157' + chr(0b1001001 + 0o33) + '\x65')(chr(9522 - 9405) + chr(0b101001 + 0o113) + chr(102) + '\055' + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\t\xfe\xa5\xd0\x12\x8f\x1ePm\xa7\xe6\xef\xf6\x167-J2="\xe8_'), chr(0b11000 + 0o114) + chr(101) + chr(0b1100011) + '\x6f' + '\x64' + '\145')(chr(6828 - 6711) + chr(0b1110100) + chr(0b10010 + 0o124) + chr(644 - 599) + chr(0b111000)))):
return Hl80dG9dKuCn
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b"\t\xfe\xa5\xd0\x12\x8f\x1ePm\xa7\xe6\xef\xf6\x161/E'+#\xf2B\xb0XS\xc9\xde\x82\x06\x8d\xd1"), chr(0b1011000 + 0o14) + '\x65' + '\x63' + chr(6685 - 6574) + chr(0b1001110 + 0o26) + chr(0b1100101))(chr(12863 - 12746) + chr(0b1011000 + 0o34) + chr(102) + '\x2d' + chr(56))):
return CLe5qtCdo0pJ
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x12\xf6\xa5\xdb'), chr(0b1100100) + chr(9870 - 9769) + chr(5070 - 4971) + '\157' + '\x64' + chr(1103 - 1002))(chr(0b10001 + 0o144) + chr(116) + chr(7138 - 7036) + '\055' + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x12\xf6\xa5\xdb\x08\x9coG`\xa6\xfb\xf9'), chr(100) + chr(1318 - 1217) + chr(99) + chr(0b100111 + 0o110) + '\x64' + chr(9801 - 9700))(chr(0b111110 + 0o67) + chr(0b1101000 + 0o14) + chr(0b1100110) + chr(0b101010 + 0o3) + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x12\xf6\xa5\xdb\x08\x9c\x12_s\xb9\xe7\xe3\xe9\x1a;,X. 4\xe8'), chr(8950 - 8850) + chr(101) + '\x63' + chr(111) + chr(0b11100 + 0o110) + '\145')('\165' + '\164' + chr(3339 - 3237) + chr(0b101101) + chr(0b111000)))):
return iA8F85yLVO8w(LzVngjm7tWJD)
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xe3\xa1\xd2\x14\x98\x02Ji\xaa\xe7\xed\xf4\x00 +H,'), '\144' + chr(4124 - 4023) + chr(1917 - 1818) + chr(1999 - 1888) + chr(100) + '\x65')(chr(0b1110101) + chr(116) + '\x66' + chr(0b101101) + chr(0b1 + 0o67))):
return iA8F85yLVO8w(BUwtcxYvLr23)
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x13\xea\xa9\xd5\x18\x9c\x02Wb\xac\xf7\xe2\xf5\x1d'), chr(5765 - 5665) + chr(0b100 + 0o141) + chr(6053 - 5954) + chr(0b1110 + 0o141) + '\144' + '\145')(chr(0b1001010 + 0o53) + chr(0b1110100) + '\146' + chr(0b1100 + 0o41) + chr(1689 - 1633))):
return Lhrai3rloNcV
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xfa\xa0\xd2\x18'), '\x64' + chr(0b1100101) + '\143' + chr(0b101010 + 0o105) + chr(0b1100001 + 0o3) + chr(101))(chr(0b1001101 + 0o50) + chr(10911 - 10795) + chr(0b1100110) + chr(45) + chr(0b10100 + 0o44))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xfa\xa0\xd2\x18\x8f\x11)'), '\144' + chr(0b110110 + 0o57) + chr(0b1001000 + 0o33) + chr(0b1101111) + chr(0b111010 + 0o52) + chr(0b10 + 0o143))('\165' + chr(116) + chr(102) + chr(45) + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xfa\xa0\xd2\x18\x8f\x11*'), chr(2208 - 2108) + chr(0b1100101) + chr(99) + chr(9160 - 9049) + chr(0b10110 + 0o116) + chr(0b111100 + 0o51))(chr(0b1110101) + '\x74' + '\x66' + '\055' + '\x38'))):
return CCPQbtcXn1uV
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xfa\xa0\xd2\x18\x8f\x1fQx\xbe\xe1\xf9\xff'), chr(0b1100100) + '\145' + chr(0b1011100 + 0o7) + chr(0b1101111) + chr(0b1100100) + chr(101))(chr(117) + '\164' + '\146' + chr(0b1111 + 0o36) + '\070')):
return yFQwTrYSUSyG
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xfa\xa0\xd2\x18\x8f\x14\\i\xa7\xfc\xe3\xee\x10'), chr(100) + chr(0b1011011 + 0o12) + chr(3411 - 3312) + chr(0b111101 + 0o62) + chr(0b111110 + 0o46) + chr(7462 - 7361))(chr(0b110001 + 0o104) + '\x74' + chr(102) + '\055' + chr(0b101100 + 0o14))):
return FsqqPqb662Mw
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xfa\xa0\xd2\x18\x8f\x11)s\xbb\xe9\xfd'), chr(0b1100100) + chr(101) + chr(99) + chr(10160 - 10049) + chr(0b1100100) + chr(6855 - 6754))('\x75' + chr(0b110101 + 0o77) + chr(102) + chr(1194 - 1149) + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xfa\xa0\xd2\x18\x8f\x11*s\xbb\xe9\xfd'), chr(7812 - 7712) + '\x65' + '\x63' + chr(111) + chr(7822 - 7722) + chr(0b101101 + 0o70))('\165' + chr(0b1110100) + chr(0b1001110 + 0o30) + chr(45) + chr(283 - 227)))):
return O2OqS4avAZUZ
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\x16\xfa\xa0\xd2\x18\x8f\rQt\xac\xe4\xf5\xf4\x06=1B'), chr(0b1100100) + chr(5680 - 5579) + '\x63' + '\157' + chr(100) + chr(0b1100101))(chr(0b100 + 0o161) + chr(0b1001101 + 0o47) + chr(10028 - 9926) + chr(0b101101) + chr(1497 - 1441))):
return iA8F85yLVO8w(mxRCIiyni4wa)
return QmmgWUB13VCJ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
get_top
|
def get_top(modality_type, value=None):
"""Gets default top transformation; if none available, return value."""
if modality_type in (ModalityType.AUDIO,
ModalityType.AUDIO_SPECTRAL,
ModalityType.GENERIC_L2_LOSS,
ModalityType.IDENTITY,
ModalityType.IDENTITY_SYMBOL,
ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY,
ModalityType.SPEECH_RECOGNITION,
ModalityType.VIDEO_IDENTITY):
return identity_top
elif modality_type in (ModalityType.CLASS_LABEL,
ModalityType.MULTI_LABEL,
ModalityType.ONE_HOT_CLASS_LABEL,
ModalityType.SIGMOID_CLASS_LABEL):
return class_label_top
elif modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL):
return symbol_top
elif modality_type == ModalityType.IMAGE:
return image_top
elif modality_type == ModalityType.IMAGE_CHANNEL_COMPRESS:
return image_channel_compress_top
elif modality_type == ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM:
return image_channel_embeddings_top
elif modality_type in (ModalityType.REAL,
ModalityType.REAL_L2_LOSS,
ModalityType.REAL_LOG_POISSON_LOSS):
return real_top
elif modality_type == ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL:
return sigmoid_max_pooling_class_label_top
elif modality_type == ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL:
return softmax_average_pooling_class_label_top
elif modality_type == ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL:
return softmax_last_timestep_class_label_top
elif modality_type == ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL:
return softmax_max_pooling_class_label_top
elif modality_type == ModalityType.SYMBOL_ONE_HOT:
return symbol_one_hot_top
elif modality_type in (ModalityType.VIDEO,
ModalityType.VIDEO_BITWISE,
ModalityType.VIDEO_PIXEL_NOISE):
return video_top
elif modality_type in (ModalityType.VIDEO_L1,
ModalityType.VIDEO_L2):
return video_l1_top
elif modality_type in (ModalityType.VIDEO_L1_RAW,
ModalityType.VIDEO_L2_RAW):
return video_raw_top
return value
|
python
|
def get_top(modality_type, value=None):
"""Gets default top transformation; if none available, return value."""
if modality_type in (ModalityType.AUDIO,
ModalityType.AUDIO_SPECTRAL,
ModalityType.GENERIC_L2_LOSS,
ModalityType.IDENTITY,
ModalityType.IDENTITY_SYMBOL,
ModalityType.IMAGE_CHANNEL_BOTTOM_IDENTITY,
ModalityType.SPEECH_RECOGNITION,
ModalityType.VIDEO_IDENTITY):
return identity_top
elif modality_type in (ModalityType.CLASS_LABEL,
ModalityType.MULTI_LABEL,
ModalityType.ONE_HOT_CLASS_LABEL,
ModalityType.SIGMOID_CLASS_LABEL):
return class_label_top
elif modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_WEIGHTS_ALL):
return symbol_top
elif modality_type == ModalityType.IMAGE:
return image_top
elif modality_type == ModalityType.IMAGE_CHANNEL_COMPRESS:
return image_channel_compress_top
elif modality_type == ModalityType.IMAGE_CHANNEL_EMBEDDINGS_BOTTOM:
return image_channel_embeddings_top
elif modality_type in (ModalityType.REAL,
ModalityType.REAL_L2_LOSS,
ModalityType.REAL_LOG_POISSON_LOSS):
return real_top
elif modality_type == ModalityType.SIGMOID_MAX_POOLING_CLASS_LABEL:
return sigmoid_max_pooling_class_label_top
elif modality_type == ModalityType.SOFTMAX_AVERAGE_POOLING_CLASS_LABEL:
return softmax_average_pooling_class_label_top
elif modality_type == ModalityType.SOFTMAX_LAST_TIMESTEP_CLASS_LABEL:
return softmax_last_timestep_class_label_top
elif modality_type == ModalityType.SOFTMAX_MAX_POOLING_CLASS_LABEL:
return softmax_max_pooling_class_label_top
elif modality_type == ModalityType.SYMBOL_ONE_HOT:
return symbol_one_hot_top
elif modality_type in (ModalityType.VIDEO,
ModalityType.VIDEO_BITWISE,
ModalityType.VIDEO_PIXEL_NOISE):
return video_top
elif modality_type in (ModalityType.VIDEO_L1,
ModalityType.VIDEO_L2):
return video_l1_top
elif modality_type in (ModalityType.VIDEO_L1_RAW,
ModalityType.VIDEO_L2_RAW):
return video_raw_top
return value
|
[
"def",
"get_top",
"(",
"modality_type",
",",
"value",
"=",
"None",
")",
":",
"if",
"modality_type",
"in",
"(",
"ModalityType",
".",
"AUDIO",
",",
"ModalityType",
".",
"AUDIO_SPECTRAL",
",",
"ModalityType",
".",
"GENERIC_L2_LOSS",
",",
"ModalityType",
".",
"IDENTITY",
",",
"ModalityType",
".",
"IDENTITY_SYMBOL",
",",
"ModalityType",
".",
"IMAGE_CHANNEL_BOTTOM_IDENTITY",
",",
"ModalityType",
".",
"SPEECH_RECOGNITION",
",",
"ModalityType",
".",
"VIDEO_IDENTITY",
")",
":",
"return",
"identity_top",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CLASS_LABEL",
",",
"ModalityType",
".",
"MULTI_LABEL",
",",
"ModalityType",
".",
"ONE_HOT_CLASS_LABEL",
",",
"ModalityType",
".",
"SIGMOID_CLASS_LABEL",
")",
":",
"return",
"class_label_top",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CTC_SYMBOL",
",",
"ModalityType",
".",
"SYMBOL",
",",
"ModalityType",
".",
"SYMBOL_WEIGHTS_ALL",
")",
":",
"return",
"symbol_top",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE",
":",
"return",
"image_top",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE_CHANNEL_COMPRESS",
":",
"return",
"image_channel_compress_top",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"IMAGE_CHANNEL_EMBEDDINGS_BOTTOM",
":",
"return",
"image_channel_embeddings_top",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"REAL",
",",
"ModalityType",
".",
"REAL_L2_LOSS",
",",
"ModalityType",
".",
"REAL_LOG_POISSON_LOSS",
")",
":",
"return",
"real_top",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SIGMOID_MAX_POOLING_CLASS_LABEL",
":",
"return",
"sigmoid_max_pooling_class_label_top",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SOFTMAX_AVERAGE_POOLING_CLASS_LABEL",
":",
"return",
"softmax_average_pooling_class_label_top",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SOFTMAX_LAST_TIMESTEP_CLASS_LABEL",
":",
"return",
"softmax_last_timestep_class_label_top",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SOFTMAX_MAX_POOLING_CLASS_LABEL",
":",
"return",
"softmax_max_pooling_class_label_top",
"elif",
"modality_type",
"==",
"ModalityType",
".",
"SYMBOL_ONE_HOT",
":",
"return",
"symbol_one_hot_top",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"VIDEO",
",",
"ModalityType",
".",
"VIDEO_BITWISE",
",",
"ModalityType",
".",
"VIDEO_PIXEL_NOISE",
")",
":",
"return",
"video_top",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"VIDEO_L1",
",",
"ModalityType",
".",
"VIDEO_L2",
")",
":",
"return",
"video_l1_top",
"elif",
"modality_type",
"in",
"(",
"ModalityType",
".",
"VIDEO_L1_RAW",
",",
"ModalityType",
".",
"VIDEO_L2_RAW",
")",
":",
"return",
"video_raw_top",
"return",
"value"
] |
Gets default top transformation; if none available, return value.
|
[
"Gets",
"default",
"top",
"transformation",
";",
"if",
"none",
"available",
"return",
"value",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1442-L1492
|
train
|
Gets default top transformation for given modality type.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(111) + chr(1231 - 1181) + chr(0b110011) + '\x34', 36141 - 36133), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(51) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(12021 - 11910) + chr(49) + chr(1452 - 1402) + chr(2219 - 2165), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + chr(1779 - 1726) + '\061', 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(51) + chr(53) + chr(1444 - 1390), 0o10), ehT0Px3KOsy9(chr(48) + chr(4944 - 4833) + chr(2029 - 1980) + '\061' + chr(1789 - 1737), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1258 - 1208) + '\x33' + '\063', 0o10), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b1000101 + 0o52) + chr(49) + chr(956 - 904) + chr(0b1011 + 0o52), 0o10), ehT0Px3KOsy9(chr(1372 - 1324) + '\157' + chr(0b100111 + 0o12) + chr(0b100111 + 0o11) + chr(0b110000), 28780 - 28772), ehT0Px3KOsy9('\x30' + '\157' + chr(51) + chr(0b11101 + 0o32) + '\066', 27376 - 27368), ehT0Px3KOsy9(chr(1286 - 1238) + '\x6f' + chr(0b110011) + chr(0b110011) + chr(51), 0o10), ehT0Px3KOsy9('\x30' + chr(9563 - 9452) + chr(53) + chr(278 - 230), 0b1000), ehT0Px3KOsy9(chr(0b11010 + 0o26) + '\x6f' + '\067' + '\x35', 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\063' + chr(0b11011 + 0o31) + chr(0b101011 + 0o5), 52772 - 52764), ehT0Px3KOsy9(chr(48) + chr(11485 - 11374) + chr(0b1 + 0o61) + chr(2445 - 2395) + chr(1451 - 1402), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b10011 + 0o42) + chr(1265 - 1215), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b1010 + 0o50) + chr(53) + chr(0b100010 + 0o25), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110100) + chr(347 - 299), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(1154 - 1103) + '\x31' + chr(50), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + '\063' + chr(1856 - 1801), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b11011 + 0o27) + chr(0b110110) + '\x34', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b100101 + 0o16) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(1144 - 1096) + chr(0b10011 + 0o134) + '\063' + chr(0b110010) + '\x36', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(50) + chr(51) + '\064', 8), ehT0Px3KOsy9('\x30' + chr(8564 - 8453) + chr(0b1101 + 0o45) + chr(48) + chr(0b1010 + 0o47), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b1110 + 0o45) + '\x36' + chr(51), 0b1000), ehT0Px3KOsy9(chr(1943 - 1895) + chr(0b1101111 + 0o0) + chr(0b110010) + chr(0b110100) + '\064', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(285 - 235) + chr(55) + chr(0b100000 + 0o25), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(52 - 3) + '\061' + chr(50), 0b1000), ehT0Px3KOsy9(chr(0b101 + 0o53) + '\157' + chr(0b10011 + 0o37) + '\x32', 0b1000), ehT0Px3KOsy9(chr(2170 - 2122) + '\x6f' + '\x31' + chr(0b100 + 0o62), ord("\x08")), ehT0Px3KOsy9(chr(1257 - 1209) + '\157' + '\x33' + '\x33' + '\065', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110001) + '\x34' + '\x31', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b11010 + 0o31) + chr(0b11 + 0o56), 12384 - 12376), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110010) + '\063', 56379 - 56371), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(0b1101111) + '\x31' + '\x35' + '\060', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(2314 - 2260) + chr(0b110000), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + '\x31' + '\061', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b10010 + 0o41) + chr(51) + chr(0b110001), 5944 - 5936), ehT0Px3KOsy9('\x30' + chr(0b100010 + 0o115) + chr(49) + chr(2658 - 2605) + '\061', 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(1586 - 1533) + chr(48), 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xb2'), chr(6849 - 6749) + chr(8237 - 8136) + '\x63' + chr(0b1101111) + chr(100) + chr(101))(chr(0b1110101) + chr(7284 - 7168) + chr(4376 - 4274) + chr(0b111 + 0o46) + chr(1732 - 1676)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def FMA4xdtyanre(N7qTs4FW6YfL, QmmgWUB13VCJ=None):
if N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xddO\x12\xa0\xa1'), '\144' + chr(0b1100101) + '\143' + chr(2369 - 2258) + chr(100) + '\x65')(chr(10768 - 10651) + chr(0b111001 + 0o73) + chr(7395 - 7293) + chr(1618 - 1573) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xddO\x12\xa0\xa1\x13\xc85\x05\xbd\x8fz|\xac'), chr(0b1100100) + chr(0b1100101) + '\143' + chr(0b111 + 0o150) + '\144' + '\x65')('\165' + chr(0b111011 + 0o71) + '\x66' + '\055' + chr(0b11101 + 0o33))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdb_\x18\xac\xbc\x05\xd8:\x0c\xcc\x84dr\xb3\xc2'), chr(3196 - 3096) + chr(101) + '\143' + chr(111) + chr(100) + chr(0b1000000 + 0o45))('\165' + '\x74' + '\146' + chr(45) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd5^\x13\xa7\xba\x05\xcf<'), chr(0b1100100) + chr(6748 - 6647) + '\x63' + chr(0b1101111) + chr(0b1001000 + 0o34) + chr(7186 - 7085))(chr(0b10101 + 0o140) + '\164' + '\x66' + chr(0b10 + 0o53) + chr(2836 - 2780))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd5^\x13\xa7\xba\x05\xcf<\x1f\xad\x82e\x7f\xaf\xdd'), chr(100) + chr(8440 - 8339) + chr(3079 - 2980) + chr(4685 - 4574) + '\144' + chr(101))(chr(0b1101 + 0o150) + chr(3325 - 3209) + chr(102) + chr(0b1010 + 0o43) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd5W\x17\xae\xab\x13\xd8-\x01\xb0\x95mq\xbf\xd3e\xfbt\x85\x12\xebR\x92\xd9\xcf\xa7\xed\xe9\x13'), '\x64' + '\x65' + chr(99) + chr(9647 - 9536) + chr(0b1100100) + chr(101))('\x75' + '\x74' + chr(0b1100110) + chr(0b101101) + chr(0b1011 + 0o55))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfJ\x13\xac\xad\x04\xc47\x05\xbd\x94os\xa9\xc5c\xe0n'), chr(0b1100100) + chr(0b110 + 0o137) + chr(0b1011001 + 0o12) + '\x6f' + '\144' + chr(8922 - 8821))('\165' + chr(116) + chr(102) + chr(0b101011 + 0o2) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcaS\x12\xac\xa1\x13\xd2!\x05\xb0\x8fai\xb9'), '\x64' + chr(0b1100100 + 0o1) + chr(3475 - 3376) + '\x6f' + chr(2979 - 2879) + chr(0b1100101))(chr(0b111011 + 0o72) + '\x74' + chr(1904 - 1802) + '\x2d' + '\x38'))):
return XNGrNhsgQ7IV
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xdfV\x17\xba\xbd\x13\xd7$\x02\xbb\x97'), chr(0b1100100) + chr(0b110 + 0o137) + chr(3156 - 3057) + '\x6f' + '\x64' + chr(101))(chr(0b1001100 + 0o51) + '\x74' + chr(0b1100110) + chr(0b11010 + 0o23) + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd1O\x1a\xbd\xa7\x13\xd7$\x02\xbb\x97'), chr(100) + '\x65' + '\143' + chr(111) + chr(0b1100100) + chr(5042 - 4941))(chr(4185 - 4068) + chr(116) + chr(4889 - 4787) + chr(0b101101) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd3T\x13\xb6\xa6\x03\xcf:\x03\xb2\x9a{n\xbf\xddk\xede\x86'), chr(0b1100100) + chr(0b1100101) + '\143' + '\157' + chr(5298 - 5198) + chr(328 - 227))('\x75' + chr(5429 - 5313) + chr(1722 - 1620) + '\055' + chr(0b110000 + 0o10))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfS\x11\xa4\xa1\x05\xdf:\x03\xb2\x9a{n\xbf\xddk\xede\x86'), chr(100) + chr(101) + chr(99) + '\x6f' + '\144' + chr(1850 - 1749))('\x75' + chr(0b101001 + 0o113) + chr(0b1100000 + 0o6) + chr(276 - 231) + chr(0b111000)))):
return y_HKuLsmsrGR
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b"\xdfN\x15\xb6\xbd\x15\xd6'\x0f\xb2"), chr(0b1100100) + chr(0b11001 + 0o114) + '\x63' + chr(111) + '\x64' + chr(0b1100101))(chr(117) + chr(0b1110100) + chr(0b11000 + 0o116) + '\055' + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfC\x1b\xab\xa1\x00'), chr(0b1001011 + 0o31) + '\x65' + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + chr(0b1100011 + 0o2))('\165' + chr(116) + chr(0b1100110) + chr(45) + chr(0b1100 + 0o54))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfC\x1b\xab\xa1\x00\xc42\x05\xb7\x9c`i\xb3\xcek\xe3l'), chr(100) + '\x65' + '\x63' + chr(0b110011 + 0o74) + '\144' + chr(0b1000101 + 0o40))(chr(0b1110101) + chr(0b1110100) + chr(102) + chr(0b101101) + '\070'))):
return ZdqK7JN8NC8f
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd5W\x17\xae\xab'), chr(9067 - 8967) + chr(101) + chr(958 - 859) + chr(6936 - 6825) + chr(1909 - 1809) + chr(0b1100101))(chr(0b10111 + 0o136) + chr(0b1110100) + '\x66' + chr(45) + chr(0b111 + 0o61))):
return xnD2VGWgfz3u
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd5W\x17\xae\xab\x13\xd8-\x01\xb0\x95mq\xbf\xd2e\xe2p\x98\x1a\xe7H'), chr(0b1100100) + chr(0b1100101) + chr(99) + chr(3382 - 3271) + chr(100) + '\145')(chr(0b111011 + 0o72) + chr(116) + chr(102) + chr(0b1111 + 0o36) + '\x38')):
return vLZKgij6ya6Z
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xd5W\x17\xae\xab\x13\xd8-\x01\xb0\x95mq\xbf\xd4g\xede\x8e\x1b\xfdU\x91\xcf\xde\xb1\xeb\xe9\x1e]C'), '\x64' + chr(0b1000011 + 0o42) + '\143' + '\157' + chr(363 - 263) + '\x65')('\165' + chr(0b1110100) + chr(0b1100110) + '\x2d' + chr(0b1 + 0o67))):
return iE3aUQwOOXi8
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xce_\x17\xa5'), chr(4775 - 4675) + '\x65' + '\143' + chr(0b1000001 + 0o56) + '\144' + chr(0b1010 + 0o133))(chr(117) + chr(0b1110100) + chr(0b1100110) + chr(45) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xce_\x17\xa5\xb1\x00\xa9:\x0c\xb1\x88{'), chr(0b1 + 0o143) + chr(0b11 + 0o142) + '\x63' + chr(722 - 611) + chr(100) + '\145')('\x75' + chr(116) + chr(0b101100 + 0o72) + chr(0b11101 + 0o20) + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xce_\x17\xa5\xb1\x00\xd4"\x1f\xae\x94an\xb3\xded\xf0l\x85\x0c\xe7'), '\x64' + chr(0b1010101 + 0o20) + '\143' + chr(0b1101111) + '\x64' + chr(101))(chr(117) + chr(1832 - 1716) + chr(0b1100110) + '\055' + chr(0b111000)))):
return c9GhEmQWkIpM
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfS\x11\xa4\xa1\x05\xdf:\r\xbf\x83wm\xaf\xdef\xe6n\x8d\x00\xf7W\x97\xcf\xd2\xac\xe8\xfc\x08WB'), chr(0b1011101 + 0o7) + chr(0b1100101) + chr(1813 - 1714) + chr(5227 - 5116) + chr(0b1100100) + chr(101))(chr(117) + chr(0b1101001 + 0o13) + chr(0b1100110) + '\055' + chr(56))):
return PObzrp0bu9qA
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfU\x10\xbd\xa3\r\xc3:\x01\xa8\x9ez|\xa7\xd4u\xffo\x85\x13\xfdU\x91\xc3\xc2\xbf\xe5\xee\x19MB\x9c#\\_'), chr(0b1010011 + 0o21) + chr(101) + '\x63' + '\157' + '\x64' + chr(0b1100101))(chr(0b1110101) + chr(8168 - 8052) + chr(0b1100110) + '\x2d' + chr(0b100001 + 0o27))):
return ESIkx9jSR_C2
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfU\x10\xbd\xa3\r\xc3:\x0c\xbf\x88|b\xb4\xd8g\xeas\x9e\x1a\xe4D\x95\xd0\xc0\xa0\xf7\xe2\x06SL\x98-'), '\x64' + chr(101) + '\x63' + chr(111) + chr(0b1100100) + chr(5269 - 5168))(chr(0b11110 + 0o127) + '\164' + '\146' + '\x2d' + chr(1867 - 1811))):
return zCVG3mWcnQw1
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfU\x10\xbd\xa3\r\xc3:\r\xbf\x83wm\xaf\xdef\xe6n\x8d\x00\xf7W\x97\xcf\xd2\xac\xe8\xfc\x08WB'), chr(4345 - 4245) + chr(0b1100101) + '\x63' + '\x6f' + '\144' + '\x65')(chr(1306 - 1189) + chr(0b11111 + 0o125) + chr(0b1100110) + '\x2d' + chr(1042 - 986))):
return kipLeNz9LwPg
elif N7qTs4FW6YfL == xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcfC\x1b\xab\xa1\x00\xc4*\x0e\xbb\x84`r\xb4'), chr(100) + '\x65' + '\143' + chr(111) + chr(100) + chr(0b1100101))('\165' + chr(0b1110100) + '\146' + chr(0b110 + 0o47) + chr(56))):
return pPfJoW_C1QlP
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcaS\x12\xac\xa1'), chr(7116 - 7016) + chr(3983 - 3882) + chr(4500 - 4401) + chr(111) + chr(1824 - 1724) + chr(0b1100101))(chr(117) + chr(2055 - 1939) + '\146' + chr(0b10110 + 0o27) + '\x38')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcaS\x12\xac\xa1\x13\xd9,\x14\xa9\x92{x'), chr(0b1001010 + 0o32) + chr(0b1100101) + '\x63' + chr(12128 - 12017) + '\x64' + chr(101))(chr(0b1000111 + 0o56) + chr(116) + chr(0b1100110) + '\055' + chr(2441 - 2385))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcaS\x12\xac\xa1\x13\xcb,\x18\xbb\x97ws\xaf\xd8y\xea'), chr(0b1100100) + '\x65' + chr(0b1100011) + '\157' + '\144' + chr(6420 - 6319))('\x75' + '\164' + chr(0b1100110) + chr(0b10000 + 0o35) + chr(0b111000)))):
return CV8Z6qeH8WaV
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcaS\x12\xac\xa1\x13\xd7T'), chr(4139 - 4039) + '\145' + '\x63' + '\x6f' + '\x64' + '\145')(chr(0b1110101) + chr(0b1110100) + '\x66' + chr(0b100100 + 0o11) + chr(2813 - 2757))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcaS\x12\xac\xa1\x13\xd7W'), chr(100) + chr(101) + chr(0b1100011) + chr(111) + '\x64' + chr(0b1100101))(chr(0b1001000 + 0o55) + '\164' + '\x66' + chr(45) + chr(56)))):
return gfSpdw7Luh5y
elif N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcaS\x12\xac\xa1\x13\xd7T\x1f\xac\x9a\x7f'), chr(0b1100100) + chr(0b1011101 + 0o10) + chr(99) + chr(111) + '\x64' + chr(4647 - 4546))('\165' + chr(5462 - 5346) + '\x66' + chr(0b11100 + 0o21) + '\070')), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\xcaS\x12\xac\xa1\x13\xd7W\x1f\xac\x9a\x7f'), '\144' + '\145' + chr(0b100011 + 0o100) + chr(8343 - 8232) + chr(8036 - 7936) + chr(0b1100101))(chr(0b101111 + 0o106) + chr(9301 - 9185) + chr(6911 - 6809) + '\x2d' + '\x38'))):
return WErnuj6dlB3d
return QmmgWUB13VCJ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/modalities.py
|
get_weights_fn
|
def get_weights_fn(modality_type, value=None):
"""Gets default weights function; if none available, return value."""
if modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.IDENTITY_SYMBOL,
ModalityType.MULTI_LABEL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_ONE_HOT):
return common_layers.weights_nonzero
elif modality_type in ModalityType.get_choices():
return common_layers.weights_all
return value
|
python
|
def get_weights_fn(modality_type, value=None):
"""Gets default weights function; if none available, return value."""
if modality_type in (ModalityType.CTC_SYMBOL,
ModalityType.IDENTITY_SYMBOL,
ModalityType.MULTI_LABEL,
ModalityType.SYMBOL,
ModalityType.SYMBOL_ONE_HOT):
return common_layers.weights_nonzero
elif modality_type in ModalityType.get_choices():
return common_layers.weights_all
return value
|
[
"def",
"get_weights_fn",
"(",
"modality_type",
",",
"value",
"=",
"None",
")",
":",
"if",
"modality_type",
"in",
"(",
"ModalityType",
".",
"CTC_SYMBOL",
",",
"ModalityType",
".",
"IDENTITY_SYMBOL",
",",
"ModalityType",
".",
"MULTI_LABEL",
",",
"ModalityType",
".",
"SYMBOL",
",",
"ModalityType",
".",
"SYMBOL_ONE_HOT",
")",
":",
"return",
"common_layers",
".",
"weights_nonzero",
"elif",
"modality_type",
"in",
"ModalityType",
".",
"get_choices",
"(",
")",
":",
"return",
"common_layers",
".",
"weights_all",
"return",
"value"
] |
Gets default weights function; if none available, return value.
|
[
"Gets",
"default",
"weights",
"function",
";",
"if",
"none",
"available",
"return",
"value",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/modalities.py#L1495-L1505
|
train
|
Gets default weights function ; if none available return value.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(461 - 413) + chr(1041 - 930) + chr(824 - 774) + chr(0b110 + 0o53), 65152 - 65144), ehT0Px3KOsy9('\x30' + chr(10184 - 10073) + '\062' + '\064' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110010) + chr(178 - 127) + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b1000 + 0o50) + chr(0b100010 + 0o115) + '\x32' + chr(1513 - 1463) + chr(0b10 + 0o60), 0o10), ehT0Px3KOsy9('\060' + chr(11782 - 11671) + '\063' + chr(2082 - 2030) + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(0b101111 + 0o1) + '\x6f' + chr(49) + chr(0b10000 + 0o44) + chr(0b110111), 30764 - 30756), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(1475 - 1427) + chr(1507 - 1396) + chr(49) + chr(0b110010) + chr(0b11 + 0o64), 0o10), ehT0Px3KOsy9('\x30' + chr(9181 - 9070) + chr(0b110011) + '\x37' + chr(1899 - 1850), 33579 - 33571), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110010) + chr(51) + chr(0b10111 + 0o36), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x33' + '\x34' + chr(0b100100 + 0o14), 8), ehT0Px3KOsy9(chr(0b100000 + 0o20) + chr(111) + chr(0b1000 + 0o52) + chr(610 - 560) + chr(154 - 106), 0o10), ehT0Px3KOsy9(chr(1267 - 1219) + chr(0b1101111) + chr(0b100101 + 0o16) + chr(0b110101) + chr(0b1001 + 0o54), 0o10), ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(111) + '\x32' + '\x37' + chr(2435 - 2380), 0b1000), ehT0Px3KOsy9('\x30' + chr(12262 - 12151) + '\x33' + chr(0b110010) + '\067', 48316 - 48308), ehT0Px3KOsy9(chr(0b110000) + chr(0b100010 + 0o115) + chr(0b110011) + chr(0b110110) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x33' + '\061' + '\x30', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\061' + chr(48) + chr(0b10110 + 0o35), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + '\x32' + '\062' + '\065', 26730 - 26722), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(1061 - 950) + chr(1036 - 981) + chr(0b10100 + 0o43), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b11010 + 0o27) + chr(0b110110) + chr(49), 0o10), ehT0Px3KOsy9('\x30' + chr(2827 - 2716) + chr(0b110010) + chr(1283 - 1235) + chr(51), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x32' + chr(0b1110 + 0o51) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(2278 - 2230) + '\157' + '\x32' + chr(0b10110 + 0o36) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b101100 + 0o103) + chr(49) + chr(0b101111 + 0o4) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(1320 - 1272) + '\x6f' + '\063' + chr(0b110010) + chr(50), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110001) + chr(0b110000) + '\x33', 8), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b1101111) + '\063' + chr(0b110101) + '\062', 0b1000), ehT0Px3KOsy9('\060' + chr(2182 - 2071) + chr(0b110010) + chr(49) + '\x32', 2713 - 2705), ehT0Px3KOsy9(chr(48) + chr(111) + '\x31' + chr(0b110010) + chr(2081 - 2029), 28941 - 28933), ehT0Px3KOsy9('\x30' + chr(1852 - 1741) + '\061' + chr(0b1000 + 0o55) + '\066', 2234 - 2226), ehT0Px3KOsy9(chr(1253 - 1205) + '\157' + chr(225 - 174) + chr(54) + chr(1845 - 1792), ord("\x08")), ehT0Px3KOsy9(chr(1723 - 1675) + chr(0b1000000 + 0o57) + '\x31' + chr(0b110110) + chr(2302 - 2250), 34199 - 34191), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(10440 - 10329) + chr(0b10010 + 0o40) + chr(1969 - 1915) + chr(0b100000 + 0o25), 26815 - 26807), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(4063 - 3952) + chr(0b11 + 0o57) + '\x36' + '\x32', 8751 - 8743), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\061' + '\x30' + chr(0b1100 + 0o53), 0o10), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(111) + chr(0b110001 + 0o2) + chr(0b110001) + chr(688 - 637), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1100100 + 0o13) + chr(0b1000 + 0o52) + chr(0b110001) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1828 - 1778) + '\x36' + chr(657 - 602), 0o10), ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\157' + chr(0b1000 + 0o52) + chr(0b110110) + chr(0b110000 + 0o3), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(111) + chr(0b110101) + chr(0b110000), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'?'), chr(0b1100100) + chr(0b1 + 0o144) + '\x63' + chr(111) + '\x64' + chr(0b1100101))(chr(10304 - 10187) + chr(116) + '\x66' + chr(45) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def aNKQLXCV9PWL(N7qTs4FW6YfL, QmmgWUB13VCJ=None):
if N7qTs4FW6YfL in (xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'R\xf6\xd9\xcd\xb3>H\x15\xb6\xd1'), '\144' + '\145' + '\x63' + '\157' + chr(100) + chr(0b1100101))('\165' + chr(0b1110100) + chr(102) + chr(45) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'X\xe6\xdf\xdc\xb4.Q\x0e\xa6\xce\xb9\xdd\xaf\xf0\x9b'), chr(100) + '\145' + chr(0b101 + 0o136) + chr(1479 - 1368) + chr(100) + '\x65')(chr(0b1100001 + 0o24) + '\164' + '\x66' + chr(0b101010 + 0o3) + chr(0b111000))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'\\\xf7\xd6\xc6\xa98I\x16\xbb\xd8\xac'), '\x64' + chr(0b1100101) + chr(0b100010 + 0o101) + '\x6f' + chr(0b1100100) + chr(7419 - 7318))('\x75' + chr(0b1110100) + chr(10240 - 10138) + '\055' + chr(56))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'B\xfb\xd7\xd0\xaf+'), chr(0b10101 + 0o117) + '\145' + chr(0b1100011) + '\157' + chr(9565 - 9465) + chr(2107 - 2006))(chr(117) + chr(0b1100100 + 0o20) + chr(0b1100110) + chr(45) + chr(0b110010 + 0o6))), xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'B\xfb\xd7\xd0\xaf+Z\x18\xb7\xd8\xbf\xd8\xa2\xeb'), chr(2243 - 2143) + chr(0b100000 + 0o105) + '\x63' + chr(0b100111 + 0o110) + chr(0b1100100) + '\145')(chr(0b1011011 + 0o32) + chr(116) + chr(0b1000000 + 0o46) + chr(1935 - 1890) + chr(1598 - 1542)))):
return xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'f\xc7\xf3\xf5\x88\x13v\x08\x97\xf2\x8e\xea\x88\xcd\xb8'), chr(0b1100100) + chr(0b1100101) + chr(0b1100011) + chr(153 - 42) + chr(0b1100100) + chr(101))(chr(117) + chr(7477 - 7361) + chr(0b1100110) + '\x2d' + '\070'))
elif N7qTs4FW6YfL in xafqLlk3kkUe(DQ15HPvleepI, xafqLlk3kkUe(SXOLrMavuUCe(b'v\xc7\xee\xcd\x83\x0fj>\x9a\xf8\x93'), chr(3271 - 3171) + chr(101) + '\x63' + chr(111) + chr(100) + '\x65')(chr(117) + chr(0b1110100) + '\x66' + chr(0b11000 + 0o25) + chr(324 - 268)))():
return xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'f\xc7\xf3\xf5\x88\x13v\x08\x98\xf1\x8c'), '\144' + chr(101) + '\x63' + '\157' + chr(0b110010 + 0o62) + chr(0b1100101))(chr(0b1110101) + '\164' + chr(0b101000 + 0o76) + '\055' + chr(0b11001 + 0o37)))
return QmmgWUB13VCJ
|
tensorflow/tensor2tensor
|
tensor2tensor/data_generators/paraphrase_ms_coco.py
|
create_combination
|
def create_combination(list_of_sentences):
"""Generates all possible pair combinations for the input list of sentences.
For example:
input = ["paraphrase1", "paraphrase2", "paraphrase3"]
output = [("paraphrase1", "paraphrase2"),
("paraphrase1", "paraphrase3"),
("paraphrase2", "paraphrase3")]
Args:
list_of_sentences: the list of input sentences.
Returns:
the list of all possible sentence pairs.
"""
num_sentences = len(list_of_sentences) - 1
combinations = []
for i, _ in enumerate(list_of_sentences):
if i == num_sentences:
break
num_pairs = num_sentences - i
populated = num_pairs * [list_of_sentences[i]]
zipped = list(zip(populated, list_of_sentences[i + 1:]))
combinations += zipped
return combinations
|
python
|
def create_combination(list_of_sentences):
"""Generates all possible pair combinations for the input list of sentences.
For example:
input = ["paraphrase1", "paraphrase2", "paraphrase3"]
output = [("paraphrase1", "paraphrase2"),
("paraphrase1", "paraphrase3"),
("paraphrase2", "paraphrase3")]
Args:
list_of_sentences: the list of input sentences.
Returns:
the list of all possible sentence pairs.
"""
num_sentences = len(list_of_sentences) - 1
combinations = []
for i, _ in enumerate(list_of_sentences):
if i == num_sentences:
break
num_pairs = num_sentences - i
populated = num_pairs * [list_of_sentences[i]]
zipped = list(zip(populated, list_of_sentences[i + 1:]))
combinations += zipped
return combinations
|
[
"def",
"create_combination",
"(",
"list_of_sentences",
")",
":",
"num_sentences",
"=",
"len",
"(",
"list_of_sentences",
")",
"-",
"1",
"combinations",
"=",
"[",
"]",
"for",
"i",
",",
"_",
"in",
"enumerate",
"(",
"list_of_sentences",
")",
":",
"if",
"i",
"==",
"num_sentences",
":",
"break",
"num_pairs",
"=",
"num_sentences",
"-",
"i",
"populated",
"=",
"num_pairs",
"*",
"[",
"list_of_sentences",
"[",
"i",
"]",
"]",
"zipped",
"=",
"list",
"(",
"zip",
"(",
"populated",
",",
"list_of_sentences",
"[",
"i",
"+",
"1",
":",
"]",
")",
")",
"combinations",
"+=",
"zipped",
"return",
"combinations"
] |
Generates all possible pair combinations for the input list of sentences.
For example:
input = ["paraphrase1", "paraphrase2", "paraphrase3"]
output = [("paraphrase1", "paraphrase2"),
("paraphrase1", "paraphrase3"),
("paraphrase2", "paraphrase3")]
Args:
list_of_sentences: the list of input sentences.
Returns:
the list of all possible sentence pairs.
|
[
"Generates",
"all",
"possible",
"pair",
"combinations",
"for",
"the",
"input",
"list",
"of",
"sentences",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/data_generators/paraphrase_ms_coco.py#L42-L67
|
train
|
Generates all possible pair combinations for the input list of sentences.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b1101 + 0o46) + chr(712 - 662) + chr(0b110000), 15085 - 15077), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010011 + 0o34) + chr(51) + chr(2031 - 1983) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110011) + '\x36', 22972 - 22964), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(0b110001) + '\x30', 51124 - 51116), ehT0Px3KOsy9('\x30' + '\x6f' + chr(1323 - 1273) + chr(0b110110) + chr(392 - 343), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101010 + 0o5) + chr(51) + chr(55) + chr(48), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\063' + chr(414 - 363) + chr(2030 - 1975), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\066' + '\064', 0o10), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(0b101010 + 0o105) + chr(1211 - 1162) + '\x34' + '\066', 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110 + 0o53) + '\x30' + chr(0b100000 + 0o23), 0o10), ehT0Px3KOsy9(chr(48) + chr(7300 - 7189) + chr(0b100111 + 0o14) + chr(497 - 444) + '\060', ord("\x08")), ehT0Px3KOsy9('\060' + chr(1144 - 1033) + '\x33' + chr(0b10110 + 0o36) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b10010 + 0o40) + chr(2020 - 1972), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(9965 - 9854) + '\065' + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b11001 + 0o27) + '\x6f' + '\062' + '\x31' + chr(0b1101 + 0o52), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(696 - 645) + '\x34' + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(0b10000 + 0o40) + '\157' + chr(0b110001) + chr(55) + '\065', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1100 + 0o143) + chr(0b110001) + chr(0b110001) + chr(49), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x33' + '\066' + chr(0b100000 + 0o22), 5739 - 5731), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1891 - 1841) + chr(0b110010 + 0o1) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b1001 + 0o52) + chr(55), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b1100 + 0o47) + chr(0b110100) + '\x31', ord("\x08")), ehT0Px3KOsy9(chr(1665 - 1617) + '\157' + '\x32' + chr(0b10 + 0o56) + '\x30', 0b1000), ehT0Px3KOsy9(chr(0b100000 + 0o20) + '\x6f' + '\x32' + chr(0b110 + 0o53) + '\066', 12876 - 12868), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b100011 + 0o20) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b100111 + 0o11) + '\157' + '\063' + chr(55) + chr(51), 1544 - 1536), ehT0Px3KOsy9(chr(1959 - 1911) + chr(111) + chr(0b110010) + '\061' + '\x32', 0o10), ehT0Px3KOsy9(chr(232 - 184) + '\x6f' + chr(0b11111 + 0o23) + '\064' + chr(50), 29830 - 29822), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110111) + chr(53), 0o10), ehT0Px3KOsy9(chr(1983 - 1935) + chr(5272 - 5161) + chr(0b11101 + 0o24) + chr(1280 - 1231) + chr(48), 8), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x35' + '\x32', ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010) + chr(2660 - 2608) + '\064', 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(49) + chr(52) + chr(112 - 63), 51491 - 51483), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1208 - 1159) + chr(53) + chr(0b110 + 0o55), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x31' + chr(1839 - 1790) + '\x30', 8), ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(6338 - 6227) + chr(0b11011 + 0o30) + chr(49) + '\060', 11098 - 11090), ehT0Px3KOsy9('\060' + chr(0b100100 + 0o113) + chr(0b110000 + 0o1) + '\x32' + '\062', 0o10), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(0b1101111) + chr(49) + chr(0b110011 + 0o0) + '\x33', 59125 - 59117), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b1010 + 0o54) + chr(1279 - 1229), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1000 + 0o147) + chr(0b110001) + chr(49) + chr(1489 - 1439), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(1313 - 1202) + '\065' + '\x30', 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe'), chr(100) + '\x65' + chr(6135 - 6036) + '\157' + '\144' + '\x65')(chr(0b1110101) + '\x74' + '\x66' + chr(1036 - 991) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def OjpW_eSeCCpV(AgCpA0ru_kOr):
WSQEZ7aaS3nT = c2A0yzQpDQB3(AgCpA0ru_kOr) - ehT0Px3KOsy9(chr(0b10110 + 0o32) + chr(111) + chr(1123 - 1074), 0o10)
yntNrkShk8ke = []
for (WVxHKyX45z_L, VNGQdHSFPrso) in YlkZvXL8qwsX(AgCpA0ru_kOr):
if WVxHKyX45z_L == WSQEZ7aaS3nT:
break
JUsAMfHz04sS = WSQEZ7aaS3nT - WVxHKyX45z_L
m83UPJqCEe29 = JUsAMfHz04sS * [AgCpA0ru_kOr[WVxHKyX45z_L]]
KR_FvVPx_21y = YyaZ4tpXu4lf(pZ0NK2y6HRbn(m83UPJqCEe29, AgCpA0ru_kOr[WVxHKyX45z_L + ehT0Px3KOsy9('\x30' + '\157' + chr(49), 8):]))
yntNrkShk8ke += KR_FvVPx_21y
return yntNrkShk8ke
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
image_transformer2d_base
|
def image_transformer2d_base():
"""Set of hyperparameters."""
hparams = common_hparams.basic_params1()
hparams.hidden_size = 512
hparams.batch_size = 1
hparams.max_length = 256
hparams.dropout = 0.0
hparams.clip_grad_norm = 0. # i.e. no gradient clipping
hparams.optimizer_adam_epsilon = 1e-9
hparams.learning_rate_decay_scheme = "noam"
hparams.learning_rate = 0.1
hparams.learning_rate_warmup_steps = 4000
hparams.initializer_gain = 0.2
hparams.initializer = "uniform_unit_scaling"
hparams.weight_decay = 0.0
hparams.optimizer_adam_beta1 = 0.9
hparams.optimizer_adam_beta2 = 0.98
hparams.label_smoothing = 0.0
hparams.bottom["targets"] = modalities.make_targets_bottom(
modalities.image_channel_embeddings_bottom)
hparams.top["targets"] = modalities.identity_top
hparams.norm_type = "layer"
hparams.layer_prepostprocess_dropout = 0.0
hparams.add_hparam("filter_size", 512) # Add new ones like this.
# attention-related flags
hparams.add_hparam("num_heads", 8)
hparams.add_hparam("attention_key_channels", 0)
hparams.add_hparam("attention_value_channels", 0)
hparams.add_hparam("ffn_layer", "conv_hidden_relu")
# All hyperparameters ending in "dropout" are automatically set to 0.0
# when not in training mode.
hparams.add_hparam("attention_dropout", 0.0)
hparams.add_hparam("relu_dropout", 0.0)
hparams.add_hparam("pos", "timing") # timing, none
hparams.add_hparam("nbr_decoder_problems", 1)
hparams.add_hparam("num_output_layers", 3)
hparams.add_hparam("block_size", 1)
# image size related flags
# assuming that the image has same height and width
hparams.add_hparam("img_len", 32)
hparams.add_hparam("num_channels", 3)
# Local attention params
hparams.add_hparam("local_and_global_att", False)
hparams.add_hparam("block_length", 256)
hparams.add_hparam("block_width", 128)
# Local 2D attention params
hparams.add_hparam("query_shape", (16, 16))
hparams.add_hparam("memory_flange", (16, 32))
hparams.add_hparam("num_encoder_layers", 4)
hparams.add_hparam("num_decoder_layers", 8)
# attention type related params
hparams.add_hparam("enc_attention_type", cia.AttentionType.GLOBAL)
hparams.add_hparam("dec_attention_type", cia.AttentionType.LOCAL_2D)
hparams.add_hparam("block_raster_scan", False)
# multipos attention params
hparams.add_hparam("q_filter_width", 1)
hparams.add_hparam("kv_filter_width", 1)
hparams.add_hparam("unconditional", False) # unconditional generation
# relative embedding hparams
hparams.add_hparam("shared_rel", False)
return hparams
|
python
|
def image_transformer2d_base():
"""Set of hyperparameters."""
hparams = common_hparams.basic_params1()
hparams.hidden_size = 512
hparams.batch_size = 1
hparams.max_length = 256
hparams.dropout = 0.0
hparams.clip_grad_norm = 0. # i.e. no gradient clipping
hparams.optimizer_adam_epsilon = 1e-9
hparams.learning_rate_decay_scheme = "noam"
hparams.learning_rate = 0.1
hparams.learning_rate_warmup_steps = 4000
hparams.initializer_gain = 0.2
hparams.initializer = "uniform_unit_scaling"
hparams.weight_decay = 0.0
hparams.optimizer_adam_beta1 = 0.9
hparams.optimizer_adam_beta2 = 0.98
hparams.label_smoothing = 0.0
hparams.bottom["targets"] = modalities.make_targets_bottom(
modalities.image_channel_embeddings_bottom)
hparams.top["targets"] = modalities.identity_top
hparams.norm_type = "layer"
hparams.layer_prepostprocess_dropout = 0.0
hparams.add_hparam("filter_size", 512) # Add new ones like this.
# attention-related flags
hparams.add_hparam("num_heads", 8)
hparams.add_hparam("attention_key_channels", 0)
hparams.add_hparam("attention_value_channels", 0)
hparams.add_hparam("ffn_layer", "conv_hidden_relu")
# All hyperparameters ending in "dropout" are automatically set to 0.0
# when not in training mode.
hparams.add_hparam("attention_dropout", 0.0)
hparams.add_hparam("relu_dropout", 0.0)
hparams.add_hparam("pos", "timing") # timing, none
hparams.add_hparam("nbr_decoder_problems", 1)
hparams.add_hparam("num_output_layers", 3)
hparams.add_hparam("block_size", 1)
# image size related flags
# assuming that the image has same height and width
hparams.add_hparam("img_len", 32)
hparams.add_hparam("num_channels", 3)
# Local attention params
hparams.add_hparam("local_and_global_att", False)
hparams.add_hparam("block_length", 256)
hparams.add_hparam("block_width", 128)
# Local 2D attention params
hparams.add_hparam("query_shape", (16, 16))
hparams.add_hparam("memory_flange", (16, 32))
hparams.add_hparam("num_encoder_layers", 4)
hparams.add_hparam("num_decoder_layers", 8)
# attention type related params
hparams.add_hparam("enc_attention_type", cia.AttentionType.GLOBAL)
hparams.add_hparam("dec_attention_type", cia.AttentionType.LOCAL_2D)
hparams.add_hparam("block_raster_scan", False)
# multipos attention params
hparams.add_hparam("q_filter_width", 1)
hparams.add_hparam("kv_filter_width", 1)
hparams.add_hparam("unconditional", False) # unconditional generation
# relative embedding hparams
hparams.add_hparam("shared_rel", False)
return hparams
|
[
"def",
"image_transformer2d_base",
"(",
")",
":",
"hparams",
"=",
"common_hparams",
".",
"basic_params1",
"(",
")",
"hparams",
".",
"hidden_size",
"=",
"512",
"hparams",
".",
"batch_size",
"=",
"1",
"hparams",
".",
"max_length",
"=",
"256",
"hparams",
".",
"dropout",
"=",
"0.0",
"hparams",
".",
"clip_grad_norm",
"=",
"0.",
"# i.e. no gradient clipping",
"hparams",
".",
"optimizer_adam_epsilon",
"=",
"1e-9",
"hparams",
".",
"learning_rate_decay_scheme",
"=",
"\"noam\"",
"hparams",
".",
"learning_rate",
"=",
"0.1",
"hparams",
".",
"learning_rate_warmup_steps",
"=",
"4000",
"hparams",
".",
"initializer_gain",
"=",
"0.2",
"hparams",
".",
"initializer",
"=",
"\"uniform_unit_scaling\"",
"hparams",
".",
"weight_decay",
"=",
"0.0",
"hparams",
".",
"optimizer_adam_beta1",
"=",
"0.9",
"hparams",
".",
"optimizer_adam_beta2",
"=",
"0.98",
"hparams",
".",
"label_smoothing",
"=",
"0.0",
"hparams",
".",
"bottom",
"[",
"\"targets\"",
"]",
"=",
"modalities",
".",
"make_targets_bottom",
"(",
"modalities",
".",
"image_channel_embeddings_bottom",
")",
"hparams",
".",
"top",
"[",
"\"targets\"",
"]",
"=",
"modalities",
".",
"identity_top",
"hparams",
".",
"norm_type",
"=",
"\"layer\"",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.0",
"hparams",
".",
"add_hparam",
"(",
"\"filter_size\"",
",",
"512",
")",
"# Add new ones like this.",
"# attention-related flags",
"hparams",
".",
"add_hparam",
"(",
"\"num_heads\"",
",",
"8",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_key_channels\"",
",",
"0",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_value_channels\"",
",",
"0",
")",
"hparams",
".",
"add_hparam",
"(",
"\"ffn_layer\"",
",",
"\"conv_hidden_relu\"",
")",
"# All hyperparameters ending in \"dropout\" are automatically set to 0.0",
"# when not in training mode.",
"hparams",
".",
"add_hparam",
"(",
"\"attention_dropout\"",
",",
"0.0",
")",
"hparams",
".",
"add_hparam",
"(",
"\"relu_dropout\"",
",",
"0.0",
")",
"hparams",
".",
"add_hparam",
"(",
"\"pos\"",
",",
"\"timing\"",
")",
"# timing, none",
"hparams",
".",
"add_hparam",
"(",
"\"nbr_decoder_problems\"",
",",
"1",
")",
"hparams",
".",
"add_hparam",
"(",
"\"num_output_layers\"",
",",
"3",
")",
"hparams",
".",
"add_hparam",
"(",
"\"block_size\"",
",",
"1",
")",
"# image size related flags",
"# assuming that the image has same height and width",
"hparams",
".",
"add_hparam",
"(",
"\"img_len\"",
",",
"32",
")",
"hparams",
".",
"add_hparam",
"(",
"\"num_channels\"",
",",
"3",
")",
"# Local attention params",
"hparams",
".",
"add_hparam",
"(",
"\"local_and_global_att\"",
",",
"False",
")",
"hparams",
".",
"add_hparam",
"(",
"\"block_length\"",
",",
"256",
")",
"hparams",
".",
"add_hparam",
"(",
"\"block_width\"",
",",
"128",
")",
"# Local 2D attention params",
"hparams",
".",
"add_hparam",
"(",
"\"query_shape\"",
",",
"(",
"16",
",",
"16",
")",
")",
"hparams",
".",
"add_hparam",
"(",
"\"memory_flange\"",
",",
"(",
"16",
",",
"32",
")",
")",
"hparams",
".",
"add_hparam",
"(",
"\"num_encoder_layers\"",
",",
"4",
")",
"hparams",
".",
"add_hparam",
"(",
"\"num_decoder_layers\"",
",",
"8",
")",
"# attention type related params",
"hparams",
".",
"add_hparam",
"(",
"\"enc_attention_type\"",
",",
"cia",
".",
"AttentionType",
".",
"GLOBAL",
")",
"hparams",
".",
"add_hparam",
"(",
"\"dec_attention_type\"",
",",
"cia",
".",
"AttentionType",
".",
"LOCAL_2D",
")",
"hparams",
".",
"add_hparam",
"(",
"\"block_raster_scan\"",
",",
"False",
")",
"# multipos attention params",
"hparams",
".",
"add_hparam",
"(",
"\"q_filter_width\"",
",",
"1",
")",
"hparams",
".",
"add_hparam",
"(",
"\"kv_filter_width\"",
",",
"1",
")",
"hparams",
".",
"add_hparam",
"(",
"\"unconditional\"",
",",
"False",
")",
"# unconditional generation",
"# relative embedding hparams",
"hparams",
".",
"add_hparam",
"(",
"\"shared_rel\"",
",",
"False",
")",
"return",
"hparams"
] |
Set of hyperparameters.
|
[
"Set",
"of",
"hyperparameters",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L367-L432
|
train
|
Set of hyperparameters for 2D image transformer.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(0b1101111) + chr(54) + chr(2555 - 2502), 0o10), ehT0Px3KOsy9(chr(0b1101 + 0o43) + '\157' + chr(1597 - 1546) + '\067' + chr(53), 0b1000), ehT0Px3KOsy9(chr(1409 - 1361) + chr(0b1001 + 0o146) + chr(1105 - 1055) + chr(1025 - 974) + '\060', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b10010 + 0o40) + chr(0b11111 + 0o21), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b100 + 0o153) + chr(2280 - 2228) + chr(0b101110 + 0o2), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110011) + '\x37' + chr(0b110101), 8), ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(0b1100111 + 0o10) + '\x37' + chr(0b1100 + 0o44), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1010 + 0o145) + chr(0b110011) + chr(49) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(264 - 213) + chr(48) + chr(253 - 200), 12085 - 12077), ehT0Px3KOsy9(chr(2228 - 2180) + chr(11482 - 11371) + chr(49) + '\060' + '\067', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x33' + '\x36' + '\x31', 7176 - 7168), ehT0Px3KOsy9(chr(0b110000) + chr(11626 - 11515) + '\063' + chr(0b100000 + 0o25) + chr(0b11010 + 0o35), 31825 - 31817), ehT0Px3KOsy9(chr(1465 - 1417) + '\x6f' + chr(0b110011) + '\063' + chr(2267 - 2218), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(6360 - 6249) + chr(51) + '\x34' + chr(50), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + '\x37' + '\061', 0b1000), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\x6f' + '\x33' + chr(1218 - 1166) + '\065', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b10001 + 0o44) + chr(0b110001), 32717 - 32709), ehT0Px3KOsy9(chr(1031 - 983) + chr(111) + '\061' + '\066' + '\x33', 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + '\x31' + chr(53) + '\067', 0o10), ehT0Px3KOsy9('\060' + chr(5844 - 5733) + chr(0b0 + 0o61) + chr(0b100111 + 0o20) + chr(50), 20424 - 20416), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(606 - 557) + '\x35' + '\063', 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(0b1000 + 0o51) + chr(0b110000) + chr(0b1111 + 0o45), 5215 - 5207), ehT0Px3KOsy9(chr(0b10011 + 0o35) + '\x6f' + chr(49) + '\066' + '\x34', 0b1000), ehT0Px3KOsy9(chr(1403 - 1355) + chr(0b1101111) + chr(0b110011) + chr(2528 - 2477) + '\x36', 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + chr(2132 - 2079) + chr(0b1111 + 0o41), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(50) + chr(1329 - 1275) + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(866 - 818) + chr(0b1101111) + '\065' + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(1760 - 1712) + chr(9068 - 8957) + '\063' + '\065' + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(10581 - 10470) + '\061' + chr(51) + chr(1985 - 1930), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b101110 + 0o101) + chr(843 - 794) + chr(0b110010) + chr(54), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x32' + chr(53) + chr(0b0 + 0o67), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(6442 - 6331) + '\063' + chr(0b11110 + 0o23) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(0b1010 + 0o46) + '\x6f' + '\061' + chr(51) + '\060', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1235 - 1184) + '\x34', 0o10), ehT0Px3KOsy9('\x30' + chr(0b101100 + 0o103) + chr(0b110001) + '\x35' + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(0b1010010 + 0o35) + chr(0b0 + 0o63) + chr(0b110011) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(0b1101111) + chr(50) + chr(50) + chr(0b11011 + 0o34), 22766 - 22758), ehT0Px3KOsy9(chr(2035 - 1987) + '\x6f' + chr(0b110010) + chr(217 - 168) + chr(52), 0o10), ehT0Px3KOsy9(chr(1575 - 1527) + chr(111) + chr(408 - 359) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\061' + '\063' + '\x37', 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x35' + chr(0b10111 + 0o31), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x1b'), chr(0b1100100) + chr(159 - 58) + '\143' + chr(9283 - 9172) + chr(100) + chr(0b1100101))(chr(0b1101 + 0o150) + chr(11673 - 11557) + '\146' + '\055' + chr(1650 - 1594)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def Z9uDCoX3BqFw():
n4ljua2gi1Pr = vLnG3ZpOXWXZ.basic_params1()
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9('\x30' + chr(11762 - 11651) + chr(49) + '\060' + '\x30' + chr(0b110000), ord("\x08"))
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(48) + chr(111) + chr(49), 0b1000)
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9('\x30' + chr(2076 - 1965) + chr(52) + chr(2038 - 1990) + chr(0b11011 + 0o25), ord("\x08"))
n4ljua2gi1Pr.ag0mwEgWzjYv = 0.0
n4ljua2gi1Pr.SdNSZNVkVjLh = 0.0
n4ljua2gi1Pr.o17O_bIptWdl = 1e-09
n4ljua2gi1Pr.v3ZnJE9Hdub1 = xafqLlk3kkUe(SXOLrMavuUCe(b'[\x92\xf9}'), chr(0b110000 + 0o64) + chr(0b101011 + 0o72) + chr(9049 - 8950) + chr(111) + chr(1502 - 1402) + chr(101))(chr(117) + chr(0b1110100 + 0o0) + chr(102) + '\055' + chr(0b101110 + 0o12))
n4ljua2gi1Pr.QGSIpd_yUNzU = 0.1
n4ljua2gi1Pr.fHyhoyGmdvM9 = ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\067' + chr(0b11110 + 0o30) + chr(0b101001 + 0o13) + '\060', 0o10)
n4ljua2gi1Pr.S1SbCBXLapw8 = 0.2
n4ljua2gi1Pr.kwfuYzkY5C57 = xafqLlk3kkUe(SXOLrMavuUCe(b'@\x93\xf1vO\x8fU\x90\xb0)\x98\x9b\x86\x91\x9a\x84E\x1a\x01\x1d'), chr(8136 - 8036) + chr(7757 - 7656) + chr(0b1100011) + chr(0b1011001 + 0o26) + chr(4972 - 4872) + chr(101))('\x75' + '\164' + chr(6637 - 6535) + chr(0b100011 + 0o12) + chr(1285 - 1229))
n4ljua2gi1Pr.eB4rJl6fUxw9 = 0.0
n4ljua2gi1Pr.GcOjyd7zcDH8 = 0.9
n4ljua2gi1Pr.CBOVKNT0M9cG = 0.98
n4ljua2gi1Pr.FSjUgdaczzRk = 0.0
n4ljua2gi1Pr.kXxsZxlIQUSQ[xafqLlk3kkUe(SXOLrMavuUCe(b'A\x9c\xeawE\x89K'), chr(0b1100100) + chr(5876 - 5775) + chr(0b11100 + 0o107) + '\157' + chr(0b1011100 + 0o10) + '\x65')(chr(0b1110101) + chr(116) + '\x66' + '\055' + chr(0b111000))] = PuPeNl0CuqOQ.make_targets_bottom(PuPeNl0CuqOQ.image_channel_embeddings_bottom)
n4ljua2gi1Pr.qxrVBjeryNEZ[xafqLlk3kkUe(SXOLrMavuUCe(b'A\x9c\xeawE\x89K'), chr(0b1100100) + chr(4920 - 4819) + chr(0b10110 + 0o115) + '\x6f' + chr(0b1000100 + 0o40) + '\145')('\x75' + '\x74' + chr(0b100110 + 0o100) + chr(1309 - 1264) + chr(1036 - 980))] = PuPeNl0CuqOQ.identity_top
n4ljua2gi1Pr.LE5Fu6Tcl7nw = xafqLlk3kkUe(SXOLrMavuUCe(b'Y\x9c\xe1uR'), '\x64' + chr(101) + '\143' + chr(0b1101111) + chr(0b1100100) + chr(0b1100101))(chr(0b1110101) + chr(0b1110001 + 0o3) + chr(867 - 765) + chr(45) + '\x38')
n4ljua2gi1Pr.RW_xSzp18UeS = 0.0
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), '\144' + chr(0b1100101) + '\x63' + '\x6f' + '\144' + chr(0b1100101))('\x75' + '\164' + chr(0b1100110) + '\055' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'S\x94\xf4dE\x8fg\xbc\xac=\x94'), '\x64' + '\x65' + '\x63' + chr(840 - 729) + chr(100) + chr(0b1010010 + 0o23))(chr(0b110011 + 0o102) + chr(116) + chr(0b1100110) + chr(45) + chr(439 - 383)), ehT0Px3KOsy9('\060' + chr(11168 - 11057) + chr(49) + chr(1148 - 1100) + chr(48) + chr(0b10010 + 0o36), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b101010 + 0o72) + '\145' + '\143' + chr(11411 - 11300) + chr(100) + chr(5538 - 5437))(chr(117) + '\164' + chr(102) + chr(0b101000 + 0o5) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'[\x88\xf5OH\x98Y\xab\xb6'), chr(0b1100001 + 0o3) + chr(3208 - 3107) + chr(0b1100011) + '\x6f' + chr(0b1001 + 0o133) + chr(4579 - 4478))(chr(117) + '\x74' + chr(8744 - 8642) + '\x2d' + chr(0b10001 + 0o47)), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b1001 + 0o50) + chr(1471 - 1423), ord("\x08")))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b1100100) + chr(101) + chr(0b1011110 + 0o5) + chr(11372 - 11261) + chr(100) + chr(7254 - 7153))(chr(7868 - 7751) + chr(0b1010001 + 0o43) + chr(102) + chr(0b101101) + chr(0b1101 + 0o53)))(xafqLlk3kkUe(SXOLrMavuUCe(b'T\x89\xecuN\x89Q\xa0\xab\x18\x9a\x8a\xa0\xbd\x9a\x8dH\x1d\x01\x1f\x077'), '\144' + chr(7405 - 7304) + '\x63' + chr(0b11011 + 0o124) + chr(4013 - 3913) + '\x65')(chr(0b1011000 + 0o35) + chr(0b101010 + 0o112) + '\146' + chr(0b101101) + chr(864 - 808)), ehT0Px3KOsy9('\x30' + '\157' + chr(1539 - 1491), 0o10))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(100) + chr(0b10110 + 0o117) + chr(0b101101 + 0o66) + '\x6f' + chr(0b111 + 0o135) + chr(101))('\165' + chr(0b1110100) + chr(0b1100110) + chr(45) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'T\x89\xecuN\x89Q\xa0\xab\x18\x87\x8e\xb5\x97\x9c\xbaJ\x1b\x0e\x14\x05!\x18\xad'), chr(2854 - 2754) + chr(0b1100101) + chr(0b1100011) + chr(7577 - 7466) + '\144' + '\145')(chr(117) + '\x74' + chr(102) + chr(1218 - 1173) + chr(0b100101 + 0o23)), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110000), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b110000 + 0o64) + chr(0b1100100 + 0o1) + '\143' + chr(8676 - 8565) + chr(100) + '\145')(chr(0b1110101) + chr(0b1110100) + chr(102) + chr(0b101101) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'S\x9b\xf6OL\x9cA\xaa\xb7'), chr(100) + chr(101) + '\143' + chr(0b1010100 + 0o33) + chr(3617 - 3517) + '\x65')('\x75' + chr(0b1110100) + '\146' + chr(45) + chr(0b11001 + 0o37)), xafqLlk3kkUe(SXOLrMavuUCe(b'V\x92\xf6f\x7f\x95Q\xab\xa1"\x9f\xb0\xab\x87\x95\x90'), chr(8502 - 8402) + chr(0b1011001 + 0o14) + chr(0b1100011) + chr(111) + '\144' + '\145')(chr(0b1110101) + '\164' + chr(0b110111 + 0o57) + '\x2d' + chr(0b100011 + 0o25)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b110001 + 0o63) + chr(101) + chr(6438 - 6339) + '\x6f' + chr(100) + chr(5509 - 5408))(chr(792 - 675) + '\164' + chr(1257 - 1155) + chr(45) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'T\x89\xecuN\x89Q\xa0\xab\x18\x95\x9d\xb6\x92\x96\x90]'), '\144' + chr(0b100001 + 0o104) + chr(99) + chr(111) + chr(100) + chr(0b11110 + 0o107))(chr(0b1100101 + 0o20) + '\x74' + chr(0b1100110) + chr(45) + chr(56)), 0.0)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(7813 - 7713) + '\x65' + chr(0b1000110 + 0o35) + chr(4074 - 3963) + '\x64' + chr(0b1100101))(chr(0b100100 + 0o121) + '\164' + '\146' + chr(625 - 580) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'G\x98\xf4e\x7f\x99J\xa0\xb5(\x84\x9b'), chr(4217 - 4117) + chr(2151 - 2050) + '\x63' + chr(3127 - 3016) + chr(0b101 + 0o137) + '\145')(chr(117) + '\164' + chr(0b1100110) + chr(0b101101) + '\x38'), 0.0)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), '\144' + chr(0b10101 + 0o120) + chr(0b1100011) + '\157' + chr(0b100110 + 0o76) + '\x65')('\165' + chr(0b1110100) + chr(102) + '\x2d' + chr(314 - 258)))(xafqLlk3kkUe(SXOLrMavuUCe(b'E\x92\xeb'), '\x64' + chr(101) + '\143' + '\157' + chr(100) + chr(101))(chr(0b1110101) + '\x74' + '\x66' + chr(1017 - 972) + '\x38'), xafqLlk3kkUe(SXOLrMavuUCe(b'A\x94\xf5yN\x9a'), '\144' + chr(101) + '\x63' + chr(8388 - 8277) + chr(100) + chr(101))('\165' + '\x74' + chr(0b1001 + 0o135) + chr(0b101101) + chr(56)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b111011 + 0o51) + chr(0b1011 + 0o132) + chr(0b1100011) + chr(145 - 34) + '\144' + chr(3279 - 3178))(chr(0b1110101) + chr(116) + chr(1371 - 1269) + '\055' + chr(0b110110 + 0o2)))(xafqLlk3kkUe(SXOLrMavuUCe(b'[\x9f\xeaOD\x98[\xa0\xa1"\x83\xb0\xa9\x90\x96\x87E\x16\x02\t'), '\144' + chr(0b1100101) + chr(99) + chr(111) + '\144' + '\x65')('\x75' + chr(116) + '\x66' + '\055' + chr(0b10111 + 0o41)), ehT0Px3KOsy9('\x30' + '\157' + chr(49), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b1100100) + chr(0b1110 + 0o127) + '\x63' + '\x6f' + chr(0b11110 + 0o106) + '\x65')('\165' + '\164' + chr(102) + '\055' + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'[\x88\xf5OO\x88L\xbf\xb03\xae\x83\xb8\x9b\x9c\x97Z'), '\144' + chr(101) + chr(0b110011 + 0o60) + chr(7462 - 7351) + chr(0b1010010 + 0o22) + chr(101))(chr(0b1110101) + chr(7334 - 7218) + chr(102) + '\055' + chr(56)), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110011), ord("\x08")))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(5495 - 5395) + chr(0b111100 + 0o51) + '\x63' + chr(368 - 257) + chr(0b1100100) + '\x65')('\x75' + chr(0b1011001 + 0o33) + chr(6791 - 6689) + '\055' + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'W\x91\xf7sK\xa2K\xa6\xbf"'), chr(0b1000001 + 0o43) + '\145' + '\x63' + chr(0b111010 + 0o65) + chr(0b110000 + 0o64) + chr(5133 - 5032))(chr(0b1110101) + chr(0b1011011 + 0o31) + chr(102) + chr(45) + chr(56)), ehT0Px3KOsy9('\060' + '\x6f' + '\x31', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b10000 + 0o124) + '\145' + chr(99) + '\157' + chr(1620 - 1520) + chr(0b1100101))(chr(117) + chr(116) + '\x66' + chr(0b101 + 0o50) + chr(2285 - 2229)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\\\x90\xffOL\x98V'), '\x64' + '\x65' + chr(0b1100011) + chr(0b1101111) + chr(7132 - 7032) + '\x65')('\165' + '\164' + '\x66' + chr(45) + chr(0b101001 + 0o17)), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(52) + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(100) + chr(0b1010110 + 0o17) + '\x63' + '\157' + '\144' + chr(0b11000 + 0o115))(chr(117) + '\x74' + '\146' + chr(45) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'[\x88\xf5OC\x95Y\xa1\xab"\x9d\x9c'), chr(100) + chr(0b111001 + 0o54) + '\x63' + chr(111) + chr(0b1100100) + '\145')('\x75' + chr(0b1110100) + chr(0b1100110) + chr(1473 - 1428) + '\070'), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110011), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b1110 + 0o126) + '\145' + chr(3296 - 3197) + chr(0b1101111) + '\x64' + '\x65')(chr(117) + chr(0b1110100) + chr(102) + chr(0b100111 + 0o6) + chr(3095 - 3039)))(xafqLlk3kkUe(SXOLrMavuUCe(b'Y\x92\xfbqL\xa2Y\xa1\xa1\x18\x96\x83\xb6\x80\x98\x89v\x12\x1b\x0e'), chr(100) + chr(2064 - 1963) + chr(0b1100011) + '\x6f' + chr(100) + chr(0b1000 + 0o135))('\165' + chr(116) + chr(0b111101 + 0o51) + '\x2d' + '\070'), ehT0Px3KOsy9(chr(261 - 213) + chr(3004 - 2893) + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(100) + chr(0b1010010 + 0o23) + chr(99) + chr(0b1101111) + chr(0b1100000 + 0o4) + '\x65')('\x75' + chr(0b1110100) + '\146' + chr(45) + chr(2426 - 2370)))(xafqLlk3kkUe(SXOLrMavuUCe(b'W\x91\xf7sK\xa2T\xaa\xab \x85\x87'), chr(4787 - 4687) + chr(101) + chr(99) + '\157' + chr(100) + chr(101))(chr(0b1110101) + '\164' + '\146' + '\055' + chr(0b1011 + 0o55)), ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(0b1010101 + 0o32) + '\x34' + chr(0b110000) + chr(0b100011 + 0o15), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), '\x64' + '\145' + chr(99) + chr(0b1101111) + '\144' + '\145')(chr(0b1110101) + chr(116) + chr(2404 - 2302) + '\055' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'W\x91\xf7sK\xa2O\xa6\xa13\x99'), chr(6798 - 6698) + '\145' + chr(99) + chr(111) + '\x64' + chr(0b1000000 + 0o45))(chr(8594 - 8477) + '\164' + '\x66' + chr(0b11101 + 0o20) + '\070'), ehT0Px3KOsy9('\060' + '\157' + '\062' + chr(816 - 768) + '\060', ord("\x08")))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), '\144' + '\145' + chr(99) + chr(0b1101111) + chr(4742 - 4642) + '\145')('\165' + chr(116) + '\146' + chr(0b100 + 0o51) + chr(0b11001 + 0o37)))(xafqLlk3kkUe(SXOLrMavuUCe(b'D\x88\xfdbY\xa2K\xa7\xa47\x94'), chr(4162 - 4062) + chr(0b1100101) + chr(99) + chr(0b1101111) + chr(0b1100100) + '\145')('\165' + '\x74' + chr(0b1100110) + '\x2d' + '\070'), (ehT0Px3KOsy9('\x30' + '\x6f' + '\x32' + chr(901 - 853), 8), ehT0Px3KOsy9('\060' + chr(7348 - 7237) + chr(0b110010) + chr(0b10111 + 0o31), 8)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), '\x64' + chr(101) + chr(0b100100 + 0o77) + chr(8806 - 8695) + '\144' + chr(7181 - 7080))(chr(0b10110 + 0o137) + '\164' + '\146' + '\055' + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'X\x98\xf5\x7fR\x84g\xa9\xa9&\x9f\x88\xbc'), '\x64' + chr(5240 - 5139) + chr(7370 - 7271) + '\157' + '\144' + chr(0b110111 + 0o56))('\x75' + '\164' + chr(0b1100110) + chr(0b101101) + chr(0b10000 + 0o50)), (ehT0Px3KOsy9(chr(1080 - 1032) + chr(0b1101111) + chr(0b100101 + 0o15) + chr(48), 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(2702 - 2650) + '\060', 8)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b1100100) + chr(0b1100101) + '\x63' + chr(0b1101111) + '\144' + '\145')('\x75' + chr(0b1110100) + chr(0b1100110) + chr(541 - 496) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'[\x88\xf5OE\x93[\xa0\xa1"\x83\xb0\xb5\x83\x80\x80[\x00'), chr(9739 - 9639) + '\x65' + chr(1435 - 1336) + '\x6f' + chr(0b111100 + 0o50) + chr(0b111111 + 0o46))(chr(0b1110101) + chr(12912 - 12796) + chr(102) + chr(45) + '\x38'), ehT0Px3KOsy9('\060' + '\157' + chr(52), 0o10))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b100011 + 0o101) + '\145' + chr(713 - 614) + '\157' + chr(6261 - 6161) + '\x65')(chr(117) + chr(228 - 112) + chr(102) + chr(467 - 422) + chr(2821 - 2765)))(xafqLlk3kkUe(SXOLrMavuUCe(b'[\x88\xf5OD\x98[\xa0\xa1"\x83\xb0\xb5\x83\x80\x80[\x00'), chr(0b1100100) + chr(5410 - 5309) + chr(0b101011 + 0o70) + '\x6f' + chr(100) + chr(101))(chr(2696 - 2579) + '\164' + chr(3961 - 3859) + '\055' + chr(2072 - 2016)), ehT0Px3KOsy9('\x30' + chr(111) + '\061' + chr(0b110000), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(100) + chr(5417 - 5316) + '\143' + '\x6f' + chr(4924 - 4824) + chr(101))(chr(0b11001 + 0o134) + chr(0b1110100) + chr(6325 - 6223) + chr(0b1 + 0o54) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'P\x93\xfbOA\x89L\xaa\xab3\x98\x80\xb7\xbd\x8d\x9cY\x16'), chr(0b1000111 + 0o35) + '\x65' + chr(6055 - 5956) + chr(0b1000110 + 0o51) + chr(7348 - 7248) + '\x65')(chr(0b1110101) + chr(6598 - 6482) + chr(0b1100000 + 0o6) + chr(523 - 478) + '\x38'), xafqLlk3kkUe(oIL3U1EOcJgs.AttentionType, xafqLlk3kkUe(SXOLrMavuUCe(b'r\xb1\xd7Ra\xb1'), chr(100) + '\x65' + chr(0b111000 + 0o53) + chr(0b11001 + 0o126) + chr(9749 - 9649) + chr(0b1100101))('\165' + chr(810 - 694) + chr(102) + chr(0b100 + 0o51) + chr(0b10101 + 0o43))))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), '\x64' + '\x65' + chr(99) + '\157' + chr(5830 - 5730) + chr(7642 - 7541))(chr(117) + '\164' + chr(0b1100110) + '\055' + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'Q\x98\xfbOA\x89L\xaa\xab3\x98\x80\xb7\xbd\x8d\x9cY\x16'), chr(100) + chr(6999 - 6898) + chr(0b1100011) + chr(0b1101011 + 0o4) + '\x64' + chr(10171 - 10070))(chr(117) + chr(0b1010100 + 0o40) + '\x66' + '\055' + chr(0b111000)), xafqLlk3kkUe(oIL3U1EOcJgs.AttentionType, xafqLlk3kkUe(SXOLrMavuUCe(b'y\xb2\xdbQl\xa2\n\x8b'), chr(0b101111 + 0o65) + '\145' + chr(99) + chr(0b11001 + 0o126) + chr(0b1000100 + 0o40) + chr(0b1100100 + 0o1))(chr(7413 - 7296) + chr(11702 - 11586) + chr(5845 - 5743) + chr(0b1101 + 0o40) + chr(2294 - 2238))))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b1100100) + chr(0b111100 + 0o51) + chr(7870 - 7771) + chr(111) + '\144' + chr(2671 - 2570))('\x75' + chr(0b101011 + 0o111) + '\x66' + chr(0b101101) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'W\x91\xf7sK\xa2J\xae\xb63\x94\x9d\x86\x91\x9a\x84G'), chr(0b1100100) + '\145' + chr(99) + chr(111) + chr(0b111101 + 0o47) + '\145')(chr(951 - 834) + chr(844 - 728) + chr(3304 - 3202) + chr(691 - 646) + chr(0b111000)), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b100011 + 0o15), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), '\x64' + '\x65' + chr(6288 - 6189) + '\x6f' + chr(100) + chr(101))(chr(13604 - 13487) + '\164' + '\146' + chr(45) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'D\xa2\xfeyL\x89]\xbd\x9a0\x98\x8b\xad\x8a'), chr(0b1100100) + chr(8691 - 8590) + '\143' + '\157' + chr(0b11100 + 0o110) + chr(101))('\x75' + chr(0b1110100) + '\x66' + '\055' + chr(56)), ehT0Px3KOsy9('\060' + chr(0b1100001 + 0o16) + chr(1699 - 1650), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b100110 + 0o76) + chr(0b1100101) + chr(99) + chr(111) + chr(5201 - 5101) + chr(6877 - 6776))('\x75' + chr(0b110011 + 0o101) + chr(9903 - 9801) + chr(45) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'^\x8b\xc7vI\x91L\xaa\xb7\x18\x86\x86\xbd\x96\x91'), chr(5834 - 5734) + '\145' + '\143' + '\x6f' + '\x64' + chr(0b100000 + 0o105))(chr(0b1100100 + 0o21) + chr(0b1110011 + 0o1) + chr(123 - 21) + '\055' + chr(0b111000)), ehT0Px3KOsy9(chr(1069 - 1021) + chr(111) + chr(2259 - 2210), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), '\x64' + chr(2274 - 2173) + chr(0b1000111 + 0o34) + '\157' + chr(0b101100 + 0o70) + '\x65')(chr(0b1110101) + chr(0b101000 + 0o114) + chr(3902 - 3800) + chr(45) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'@\x93\xfb\x7fN\x99Q\xbb\xac(\x9f\x8e\xb5'), chr(0b111010 + 0o52) + chr(1445 - 1344) + chr(5182 - 5083) + chr(111) + chr(100) + '\145')(chr(117) + '\x74' + chr(102) + chr(0b1100 + 0o41) + chr(56)), ehT0Px3KOsy9('\060' + '\x6f' + '\060', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'T\x99\xfcOH\x8dY\xbd\xa4*'), chr(0b1100100) + chr(0b1100101) + '\x63' + chr(0b1101111) + '\144' + chr(9637 - 9536))('\x75' + '\164' + '\x66' + chr(45) + chr(0b11110 + 0o32)))(xafqLlk3kkUe(SXOLrMavuUCe(b'F\x95\xf9bE\x99g\xbd\xa0+'), chr(9990 - 9890) + chr(0b1100101) + '\x63' + '\x6f' + chr(0b1100100) + chr(101))(chr(117) + chr(116) + chr(102) + chr(691 - 646) + '\x38'), ehT0Px3KOsy9('\060' + chr(10647 - 10536) + '\x30', 8))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
imagetransformer2d_base_8l_8_32_big
|
def imagetransformer2d_base_8l_8_32_big():
"""hparams fo 8 layer big 2d model for cifar 10."""
hparams = image_transformer2d_base()
hparams.num_heads = 16
hparams.hidden_size = 1024
hparams.filter_size = 2048
hparams.num_decoder_layers = 8
hparams.batch_size = 1
hparams.layer_prepostprocess_dropout = 0.3
hparams.query_shape = (8, 16)
hparams.memory_flange = (0, 32)
hparams.unconditional = int(False)
return hparams
|
python
|
def imagetransformer2d_base_8l_8_32_big():
"""hparams fo 8 layer big 2d model for cifar 10."""
hparams = image_transformer2d_base()
hparams.num_heads = 16
hparams.hidden_size = 1024
hparams.filter_size = 2048
hparams.num_decoder_layers = 8
hparams.batch_size = 1
hparams.layer_prepostprocess_dropout = 0.3
hparams.query_shape = (8, 16)
hparams.memory_flange = (0, 32)
hparams.unconditional = int(False)
return hparams
|
[
"def",
"imagetransformer2d_base_8l_8_32_big",
"(",
")",
":",
"hparams",
"=",
"image_transformer2d_base",
"(",
")",
"hparams",
".",
"num_heads",
"=",
"16",
"hparams",
".",
"hidden_size",
"=",
"1024",
"hparams",
".",
"filter_size",
"=",
"2048",
"hparams",
".",
"num_decoder_layers",
"=",
"8",
"hparams",
".",
"batch_size",
"=",
"1",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.3",
"hparams",
".",
"query_shape",
"=",
"(",
"8",
",",
"16",
")",
"hparams",
".",
"memory_flange",
"=",
"(",
"0",
",",
"32",
")",
"hparams",
".",
"unconditional",
"=",
"int",
"(",
"False",
")",
"return",
"hparams"
] |
hparams fo 8 layer big 2d model for cifar 10.
|
[
"hparams",
"fo",
"8",
"layer",
"big",
"2d",
"model",
"for",
"cifar",
"10",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L485-L497
|
train
|
Hparams fo 8 layer big 2d model for cifar 10.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(1555 - 1507) + '\x6f' + chr(50) + chr(484 - 429) + '\x34', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(50) + '\x33', 8357 - 8349), ehT0Px3KOsy9(chr(48) + chr(4664 - 4553) + chr(1743 - 1688), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1000011 + 0o54) + chr(50) + '\064', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x33' + '\x30' + '\x30', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b10011 + 0o134) + '\063' + '\064' + '\x31', 11331 - 11323), ehT0Px3KOsy9('\x30' + chr(12291 - 12180) + chr(0b110011) + chr(1545 - 1496) + '\x33', 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\062', 63990 - 63982), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b10 + 0o57) + '\066', ord("\x08")), ehT0Px3KOsy9(chr(0b111 + 0o51) + '\x6f' + chr(0b110110) + '\061', 0b1000), ehT0Px3KOsy9(chr(0b11111 + 0o21) + chr(0b1000001 + 0o56) + chr(1049 - 999) + '\x30' + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(12228 - 12117) + '\x32' + chr(0b10110 + 0o33) + chr(2584 - 2532), 47838 - 47830), ehT0Px3KOsy9('\060' + '\x6f' + chr(1336 - 1285) + chr(0b110011) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(51) + chr(0b110100) + '\064', 36017 - 36009), ehT0Px3KOsy9('\060' + chr(111) + chr(49) + '\060' + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\062' + chr(0b101010 + 0o15) + '\060', 0o10), ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(0b1101111) + chr(0b110011) + '\x37', 0b1000), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(0b1010001 + 0o36) + chr(0b110001) + chr(52), 22371 - 22363), ehT0Px3KOsy9(chr(1146 - 1098) + '\157' + '\061' + '\066' + chr(0b10011 + 0o35), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\063' + chr(0b101111 + 0o4) + chr(48), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1001010 + 0o45) + chr(55) + chr(710 - 660), 23704 - 23696), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b1101111) + '\x36' + '\x35', 0b1000), ehT0Px3KOsy9(chr(48) + chr(1081 - 970) + chr(49) + '\066' + chr(0b1010 + 0o53), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51) + chr(0b110 + 0o53) + '\x37', 38098 - 38090), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b101000 + 0o13) + chr(424 - 371), 54743 - 54735), ehT0Px3KOsy9('\060' + chr(111) + chr(1386 - 1336) + '\062' + chr(0b0 + 0o66), 0b1000), ehT0Px3KOsy9(chr(1275 - 1227) + '\157' + '\x32' + chr(0b100111 + 0o14) + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(50) + chr(0b110010) + '\x37', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + '\062' + '\061' + '\063', 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(2025 - 1976) + '\065' + chr(51), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(768 - 718) + '\061' + '\x37', 0o10), ehT0Px3KOsy9(chr(238 - 190) + chr(4830 - 4719) + chr(2329 - 2280) + chr(0b110010) + chr(54), 0o10), ehT0Px3KOsy9('\060' + chr(3722 - 3611) + '\064' + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b111 + 0o51) + '\157' + '\x33' + '\x32' + chr(0b1111 + 0o46), 0b1000), ehT0Px3KOsy9(chr(1292 - 1244) + chr(0b10111 + 0o130) + chr(53) + '\061', 0b1000), ehT0Px3KOsy9(chr(48) + chr(5740 - 5629) + chr(0b100111 + 0o13) + '\063' + chr(0b10 + 0o61), ord("\x08")), ehT0Px3KOsy9(chr(0b101 + 0o53) + chr(111) + '\063' + '\066' + chr(54), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(49) + chr(2738 - 2683) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b100 + 0o55) + '\065' + chr(714 - 662), 0o10), ehT0Px3KOsy9(chr(1248 - 1200) + '\157' + chr(2050 - 1997) + chr(1996 - 1945), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(0b1101111) + '\065' + chr(0b110000), 29784 - 29776)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x92'), chr(3708 - 3608) + chr(101) + chr(0b1100011) + chr(1220 - 1109) + chr(0b1100100) + '\145')('\x75' + chr(0b101000 + 0o114) + '\146' + chr(86 - 41) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def RAvMms4Y2eXs():
n4ljua2gi1Pr = Z9uDCoX3BqFw()
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(0b10 + 0o56) + '\x6f' + chr(0b1000 + 0o52) + chr(0b101000 + 0o10), 18119 - 18111)
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9('\060' + '\157' + chr(50) + chr(0b1110 + 0o42) + '\x30' + '\x30', 33849 - 33841)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9('\060' + chr(1351 - 1240) + chr(0b100010 + 0o22) + chr(0b10101 + 0o33) + '\060' + chr(580 - 532), 3199 - 3191)
n4ljua2gi1Pr.pRi6YFAYEnH4 = ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(111) + chr(49) + '\x30', ord("\x08"))
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(111) + chr(0b1 + 0o60), ord("\x08"))
n4ljua2gi1Pr.RW_xSzp18UeS = 0.3
n4ljua2gi1Pr.bOgwkN3Z_Ukr = (ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b10111 + 0o32) + '\060', 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\062' + '\060', 8))
n4ljua2gi1Pr.BpQobI3VuWq4 = (ehT0Px3KOsy9('\060' + chr(111) + chr(0b101011 + 0o5), 0b1000), ehT0Px3KOsy9('\060' + '\157' + '\x34' + '\x30', 32902 - 32894))
n4ljua2gi1Pr.IcYfltg0WkcT = ehT0Px3KOsy9(ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b101000 + 0o10), 8))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
imagetransformer_base_10l_8h_big_uncond_dr03_dan_64_2d
|
def imagetransformer_base_10l_8h_big_uncond_dr03_dan_64_2d():
"""big 1d model for unconditional generation on imagenet."""
hparams = image_transformer2d_base()
hparams.unconditional = True
hparams.hidden_size = 512
hparams.batch_size = 1
hparams.img_len = 64
hparams.num_heads = 8
hparams.filter_size = 2048
hparams.batch_size = 1
hparams.max_length = 3075
hparams.max_length = 14000
hparams.layer_preprocess_sequence = "none"
hparams.layer_postprocess_sequence = "dan"
hparams.layer_prepostprocess_dropout = 0.1
hparams.dec_attention_type = cia.AttentionType.LOCAL_2D
hparams.query_shape = (16, 16)
hparams.memory_flange = (8, 8)
return hparams
|
python
|
def imagetransformer_base_10l_8h_big_uncond_dr03_dan_64_2d():
"""big 1d model for unconditional generation on imagenet."""
hparams = image_transformer2d_base()
hparams.unconditional = True
hparams.hidden_size = 512
hparams.batch_size = 1
hparams.img_len = 64
hparams.num_heads = 8
hparams.filter_size = 2048
hparams.batch_size = 1
hparams.max_length = 3075
hparams.max_length = 14000
hparams.layer_preprocess_sequence = "none"
hparams.layer_postprocess_sequence = "dan"
hparams.layer_prepostprocess_dropout = 0.1
hparams.dec_attention_type = cia.AttentionType.LOCAL_2D
hparams.query_shape = (16, 16)
hparams.memory_flange = (8, 8)
return hparams
|
[
"def",
"imagetransformer_base_10l_8h_big_uncond_dr03_dan_64_2d",
"(",
")",
":",
"hparams",
"=",
"image_transformer2d_base",
"(",
")",
"hparams",
".",
"unconditional",
"=",
"True",
"hparams",
".",
"hidden_size",
"=",
"512",
"hparams",
".",
"batch_size",
"=",
"1",
"hparams",
".",
"img_len",
"=",
"64",
"hparams",
".",
"num_heads",
"=",
"8",
"hparams",
".",
"filter_size",
"=",
"2048",
"hparams",
".",
"batch_size",
"=",
"1",
"hparams",
".",
"max_length",
"=",
"3075",
"hparams",
".",
"max_length",
"=",
"14000",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"none\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"dan\"",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.1",
"hparams",
".",
"dec_attention_type",
"=",
"cia",
".",
"AttentionType",
".",
"LOCAL_2D",
"hparams",
".",
"query_shape",
"=",
"(",
"16",
",",
"16",
")",
"hparams",
".",
"memory_flange",
"=",
"(",
"8",
",",
"8",
")",
"return",
"hparams"
] |
big 1d model for unconditional generation on imagenet.
|
[
"big",
"1d",
"model",
"for",
"unconditional",
"generation",
"on",
"imagenet",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L501-L519
|
train
|
big 1d model for unconditional generation on imagenet.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b1100000 + 0o17) + chr(1407 - 1358) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(0b1101111) + chr(51) + '\x30', 0b1000), ehT0Px3KOsy9(chr(48) + chr(9569 - 9458) + '\x33' + '\063', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + '\062' + chr(0b11010 + 0o35) + chr(1759 - 1709), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110001) + chr(0b110110) + '\064', 56141 - 56133), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(0b1101111) + '\063' + chr(2457 - 2407) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + '\064' + chr(0b1110 + 0o44), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(55) + chr(0b110111), 31737 - 31729), ehT0Px3KOsy9('\x30' + '\157' + '\062', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b100011 + 0o114) + chr(0b110011) + chr(0b110011) + '\x31', 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\062' + chr(48) + '\x37', 0o10), ehT0Px3KOsy9(chr(218 - 170) + chr(8793 - 8682) + chr(135 - 85) + chr(0b110110 + 0o0) + chr(835 - 786), 0b1000), ehT0Px3KOsy9(chr(1215 - 1167) + '\x6f' + chr(55) + chr(333 - 279), 55899 - 55891), ehT0Px3KOsy9(chr(0b110000) + chr(1771 - 1660) + chr(51) + chr(0b110001) + chr(0b110000), 45701 - 45693), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\061' + chr(0b110000) + '\x32', 0b1000), ehT0Px3KOsy9('\060' + chr(111) + '\x31' + chr(0b11001 + 0o30) + chr(0b110000 + 0o0), 20020 - 20012), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(0b1101111) + chr(0b11110 + 0o24) + chr(98 - 48) + chr(0b110011), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + chr(0b1001 + 0o50) + chr(0b101110 + 0o4), 28643 - 28635), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(111) + chr(1655 - 1605) + chr(52) + chr(55), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x33' + '\x35' + chr(0b110111), 40796 - 40788), ehT0Px3KOsy9(chr(586 - 538) + chr(0b1010000 + 0o37) + chr(1447 - 1398) + chr(51) + chr(0b110000), 61722 - 61714), ehT0Px3KOsy9(chr(2137 - 2089) + chr(8942 - 8831) + '\x34' + chr(51), 0b1000), ehT0Px3KOsy9(chr(845 - 797) + chr(0b1101111) + chr(0b1101 + 0o46) + chr(0b110101) + chr(52), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b11010 + 0o125) + '\061' + chr(0b110001) + chr(0b110100), 16952 - 16944), ehT0Px3KOsy9(chr(2242 - 2194) + chr(111) + chr(51) + chr(322 - 274) + '\065', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110001) + chr(1873 - 1820) + chr(2198 - 2144), 0o10), ehT0Px3KOsy9(chr(546 - 498) + chr(11723 - 11612) + chr(53), 52865 - 52857), ehT0Px3KOsy9(chr(58 - 10) + chr(0b1001010 + 0o45) + '\061' + chr(2106 - 2053) + chr(0b1101 + 0o51), 8), ehT0Px3KOsy9(chr(0b110000) + chr(11649 - 11538) + '\062' + chr(50) + chr(0b110101 + 0o0), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(6680 - 6569) + chr(0b110010 + 0o1), 0b1000), ehT0Px3KOsy9(chr(0b11101 + 0o23) + chr(111) + chr(994 - 943) + '\x33' + chr(49), 8), ehT0Px3KOsy9(chr(0b110000) + chr(10045 - 9934) + chr(0b11100 + 0o27) + chr(0b100 + 0o61) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(111) + chr(1411 - 1362) + chr(692 - 638) + chr(0b110011), 35260 - 35252), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(49) + chr(0b110001) + chr(0b101011 + 0o14), 22150 - 22142), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110001) + chr(0b110101) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(0b10101 + 0o33) + '\x6f' + '\x35' + chr(0b1011 + 0o50), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b100001 + 0o116) + '\x33' + chr(0b110111) + '\x35', 4608 - 4600), ehT0Px3KOsy9(chr(48) + chr(0b1000101 + 0o52) + '\064' + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\061' + chr(1485 - 1436) + '\x33', 5034 - 5026), ehT0Px3KOsy9(chr(966 - 918) + '\157' + chr(50) + chr(53) + chr(1043 - 988), 17862 - 17854)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b11101 + 0o23) + '\x6f' + '\x35' + chr(1556 - 1508), 42818 - 42810)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'!'), chr(0b1100100) + chr(4320 - 4219) + '\143' + '\x6f' + chr(0b1100100 + 0o0) + '\145')('\x75' + chr(116) + chr(102) + chr(0b101101) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def Kq4yCJI9DWEl():
n4ljua2gi1Pr = Z9uDCoX3BqFw()
n4ljua2gi1Pr.IcYfltg0WkcT = ehT0Px3KOsy9(chr(0b100101 + 0o13) + '\x6f' + chr(49), 59857 - 59849)
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9(chr(0b110000) + chr(10244 - 10133) + chr(192 - 143) + chr(582 - 534) + chr(0b110000) + chr(1789 - 1741), ord("\x08"))
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(1095 - 1047) + chr(0b1101111) + chr(49), 8)
n4ljua2gi1Pr.laxD7jy5y7k1 = ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(0b1101111) + '\x31' + '\x30' + chr(48), 0b1000)
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(1654 - 1606) + chr(0b1101111) + '\061' + chr(0b1010 + 0o46), ord("\x08"))
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9(chr(1889 - 1841) + chr(349 - 238) + '\064' + chr(48) + chr(0b110000) + chr(48), 26056 - 26048)
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(48) + chr(0b1100110 + 0o11) + '\x31', 8)
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9(chr(0b110000) + chr(1920 - 1809) + '\066' + '\060' + chr(2078 - 2030) + '\063', 7857 - 7849)
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9('\060' + chr(0b1101111) + '\063' + chr(0b11100 + 0o27) + chr(0b110010) + chr(261 - 207) + chr(48), 9834 - 9826)
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'awu\xdf'), chr(100) + '\x65' + chr(0b1100011) + chr(9022 - 8911) + chr(0b100101 + 0o77) + chr(101))(chr(5443 - 5326) + '\164' + '\146' + chr(0b101101) + chr(2983 - 2927))
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'kyu'), '\144' + chr(101) + chr(0b1100011) + chr(0b1 + 0o156) + '\144' + '\x65')('\x75' + '\164' + '\146' + chr(1385 - 1340) + chr(1394 - 1338))
n4ljua2gi1Pr.RW_xSzp18UeS = 0.1
n4ljua2gi1Pr.h3BUtwwQ_ZW5 = oIL3U1EOcJgs.AttentionType.LOCAL_2D
n4ljua2gi1Pr.bOgwkN3Z_Ukr = (ehT0Px3KOsy9(chr(322 - 274) + '\x6f' + chr(0b110010) + '\060', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b110101 + 0o72) + chr(0b110010) + chr(0b101111 + 0o1), 8))
n4ljua2gi1Pr.BpQobI3VuWq4 = (ehT0Px3KOsy9(chr(0b1001 + 0o47) + '\157' + '\061' + chr(961 - 913), 8), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110001) + '\x30', 8))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer2d_base
|
def img2img_transformer2d_base():
"""Base params for img2img 2d attention."""
hparams = image_transformer2d_base()
# learning related flags
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
# This version seems to benefit from a higher learning rate.
hparams.learning_rate = 0.2
hparams.layer_prepostprocess_dropout = 0.1
hparams.learning_rate_warmup_steps = 12000
hparams.filter_size = 2048
hparams.num_encoder_layers = 4
hparams.num_decoder_layers = 8
hparams.bottom["inputs"] = modalities.image_channel_embeddings_bottom
hparams.dec_attention_type = cia.AttentionType.LOCAL_2D
hparams.block_raster_scan = True
return hparams
|
python
|
def img2img_transformer2d_base():
"""Base params for img2img 2d attention."""
hparams = image_transformer2d_base()
# learning related flags
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
# This version seems to benefit from a higher learning rate.
hparams.learning_rate = 0.2
hparams.layer_prepostprocess_dropout = 0.1
hparams.learning_rate_warmup_steps = 12000
hparams.filter_size = 2048
hparams.num_encoder_layers = 4
hparams.num_decoder_layers = 8
hparams.bottom["inputs"] = modalities.image_channel_embeddings_bottom
hparams.dec_attention_type = cia.AttentionType.LOCAL_2D
hparams.block_raster_scan = True
return hparams
|
[
"def",
"img2img_transformer2d_base",
"(",
")",
":",
"hparams",
"=",
"image_transformer2d_base",
"(",
")",
"# learning related flags",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"n\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"da\"",
"# This version seems to benefit from a higher learning rate.",
"hparams",
".",
"learning_rate",
"=",
"0.2",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.1",
"hparams",
".",
"learning_rate_warmup_steps",
"=",
"12000",
"hparams",
".",
"filter_size",
"=",
"2048",
"hparams",
".",
"num_encoder_layers",
"=",
"4",
"hparams",
".",
"num_decoder_layers",
"=",
"8",
"hparams",
".",
"bottom",
"[",
"\"inputs\"",
"]",
"=",
"modalities",
".",
"image_channel_embeddings_bottom",
"hparams",
".",
"dec_attention_type",
"=",
"cia",
".",
"AttentionType",
".",
"LOCAL_2D",
"hparams",
".",
"block_raster_scan",
"=",
"True",
"return",
"hparams"
] |
Base params for img2img 2d attention.
|
[
"Base",
"params",
"for",
"img2img",
"2d",
"attention",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L585-L601
|
train
|
Base params for img2img 2d attention.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + '\x37' + '\060', 27875 - 27867), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b1001 + 0o55) + '\x36', 40972 - 40964), ehT0Px3KOsy9(chr(0b1 + 0o57) + '\x6f' + '\062' + '\x35' + chr(0b1000 + 0o52), 2042 - 2034), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + chr(0b10 + 0o57) + chr(51), 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(51) + chr(0b110101) + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b111001 + 0o66) + chr(1147 - 1098) + chr(2128 - 2078) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\x32' + chr(357 - 308), 0o10), ehT0Px3KOsy9(chr(1899 - 1851) + '\x6f' + chr(50) + chr(0b110100) + chr(1049 - 994), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011010 + 0o25) + chr(0b110100 + 0o3) + chr(0b110011), 20018 - 20010), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(0b1101100 + 0o3) + chr(51) + chr(0b110000) + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(1627 - 1579) + chr(0b1101111) + '\063' + '\064' + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b11001 + 0o31) + chr(51) + chr(577 - 524), 44007 - 43999), ehT0Px3KOsy9(chr(1035 - 987) + chr(0b1101110 + 0o1) + chr(0b110011) + '\x30' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(50) + chr(48) + chr(0b110010 + 0o3), 12446 - 12438), ehT0Px3KOsy9(chr(2031 - 1983) + chr(9639 - 9528) + chr(0b11101 + 0o24) + '\063' + chr(0b11011 + 0o30), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(1928 - 1877) + '\x30' + chr(295 - 247), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(1361 - 1306), 0b1000), ehT0Px3KOsy9(chr(2082 - 2034) + chr(10866 - 10755) + chr(0b110011) + chr(2756 - 2702) + '\063', 48281 - 48273), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b1001 + 0o52) + chr(53) + chr(0b1010 + 0o46), 0b1000), ehT0Px3KOsy9(chr(716 - 668) + '\157' + chr(0b110010) + chr(53) + chr(0b101011 + 0o13), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x33' + chr(0b110111) + '\x34', 13165 - 13157), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + '\060' + '\065', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\067' + chr(0b110000), 0b1000), ehT0Px3KOsy9(chr(0b101100 + 0o4) + '\x6f' + chr(399 - 349) + '\x30' + '\062', 0o10), ehT0Px3KOsy9(chr(2230 - 2182) + '\x6f' + chr(2218 - 2167) + chr(0b110100) + '\x35', 8), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + chr(0b100000 + 0o27) + chr(0b11001 + 0o34), 7464 - 7456), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\061', 10312 - 10304), ehT0Px3KOsy9(chr(263 - 215) + '\x6f' + chr(2162 - 2111) + chr(0b110111) + '\x31', 41905 - 41897), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(8568 - 8457) + '\061' + chr(153 - 105) + chr(2095 - 2040), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(3565 - 3454) + chr(0b11001 + 0o32) + '\x34' + chr(2894 - 2840), ord("\x08")), ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(0b1001111 + 0o40) + chr(217 - 168) + chr(0b110 + 0o56) + chr(728 - 673), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(2156 - 2107) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(89 - 41) + '\x6f' + chr(50) + chr(1350 - 1302) + '\061', 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(50) + chr(226 - 174) + '\062', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b11010 + 0o30) + chr(2322 - 2272) + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\157' + chr(1697 - 1646) + '\061' + chr(357 - 307), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b10110 + 0o131) + chr(0b100011 + 0o16) + '\062' + '\065', 42447 - 42439), ehT0Px3KOsy9(chr(0b100011 + 0o15) + '\157' + chr(0b110001) + chr(1959 - 1909) + chr(1766 - 1717), 0b1000), ehT0Px3KOsy9(chr(173 - 125) + chr(0b101110 + 0o101) + chr(1487 - 1435) + '\061', 33842 - 33834), ehT0Px3KOsy9(chr(1282 - 1234) + chr(111) + chr(0b110001) + chr(49), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b1101111) + '\065' + '\060', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xd6'), '\144' + chr(0b11 + 0o142) + chr(0b1100011) + chr(0b1101100 + 0o3) + chr(9936 - 9836) + chr(0b111101 + 0o50))(chr(0b1110101) + chr(0b1110100) + chr(102) + chr(45) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def AnTsaAZ4phGl():
n4ljua2gi1Pr = Z9uDCoX3BqFw()
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'\x96'), '\144' + chr(101) + '\143' + chr(111) + chr(0b1000001 + 0o43) + chr(4520 - 4419))(chr(1857 - 1740) + chr(8588 - 8472) + '\x66' + chr(0b101 + 0o50) + chr(0b11001 + 0o37))
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'\x9cW'), chr(2471 - 2371) + '\145' + chr(0b1100011) + '\157' + chr(0b1011000 + 0o14) + '\x65')(chr(117) + chr(0b1110100) + chr(8912 - 8810) + chr(45) + '\070')
n4ljua2gi1Pr.QGSIpd_yUNzU = 0.2
n4ljua2gi1Pr.RW_xSzp18UeS = 0.1
n4ljua2gi1Pr.fHyhoyGmdvM9 = ehT0Px3KOsy9(chr(0b101 + 0o53) + chr(0b1101100 + 0o3) + chr(0b1100 + 0o46) + '\x37' + chr(0b110011) + chr(52) + chr(0b11100 + 0o24), 0o10)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9('\x30' + chr(0b11 + 0o154) + chr(0b110100) + chr(48) + '\x30' + chr(0b110000), 10093 - 10085)
n4ljua2gi1Pr.RS6YkARoTleN = ehT0Px3KOsy9(chr(0b0 + 0o60) + '\157' + chr(52), 0b1000)
n4ljua2gi1Pr.pRi6YFAYEnH4 = ehT0Px3KOsy9(chr(0b110000) + chr(3357 - 3246) + '\061' + '\x30', ord("\x08"))
n4ljua2gi1Pr.kXxsZxlIQUSQ[xafqLlk3kkUe(SXOLrMavuUCe(b'\x91X\xda\xfe\x9a '), chr(7475 - 7375) + chr(0b1100101) + '\143' + '\157' + chr(0b1100100) + '\x65')(chr(0b1110101) + chr(116) + chr(5933 - 5831) + '\x2d' + chr(56))] = PuPeNl0CuqOQ.image_channel_embeddings_bottom
n4ljua2gi1Pr.h3BUtwwQ_ZW5 = oIL3U1EOcJgs.AttentionType.LOCAL_2D
n4ljua2gi1Pr.wG71c38XLLpE = ehT0Px3KOsy9('\060' + chr(0b1001010 + 0o45) + chr(0b10100 + 0o35), 8)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer2d_q3
|
def img2img_transformer2d_q3():
"""Current best hparams for local 2d."""
hparams = img2img_transformer2d_q1()
hparams.batch_size = 2
hparams.query_shape = (8, 16)
hparams.memory_flange = (8, 32)
return hparams
|
python
|
def img2img_transformer2d_q3():
"""Current best hparams for local 2d."""
hparams = img2img_transformer2d_q1()
hparams.batch_size = 2
hparams.query_shape = (8, 16)
hparams.memory_flange = (8, 32)
return hparams
|
[
"def",
"img2img_transformer2d_q3",
"(",
")",
":",
"hparams",
"=",
"img2img_transformer2d_q1",
"(",
")",
"hparams",
".",
"batch_size",
"=",
"2",
"hparams",
".",
"query_shape",
"=",
"(",
"8",
",",
"16",
")",
"hparams",
".",
"memory_flange",
"=",
"(",
"8",
",",
"32",
")",
"return",
"hparams"
] |
Current best hparams for local 2d.
|
[
"Current",
"best",
"hparams",
"for",
"local",
"2d",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L627-L633
|
train
|
Current best hparams for local 2d.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(837 - 789) + chr(111) + '\x31' + chr(48) + chr(714 - 665), 0o10), ehT0Px3KOsy9(chr(2214 - 2166) + chr(0b1010110 + 0o31) + chr(0b110001) + chr(0b11001 + 0o34) + '\063', 0b1000), ehT0Px3KOsy9(chr(0b1110 + 0o42) + '\157' + chr(55) + chr(53), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(1401 - 1348) + chr(625 - 577), 0b1000), ehT0Px3KOsy9(chr(48) + chr(10349 - 10238) + '\x31' + chr(0b11010 + 0o27) + chr(0b101010 + 0o10), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(510 - 399) + chr(0b100011 + 0o20) + '\x32' + chr(1579 - 1528), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(1357 - 1308) + chr(0b110001) + chr(801 - 751), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110001) + chr(925 - 873) + chr(0b110110), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(606 - 556) + chr(1491 - 1443) + chr(48), 353 - 345), ehT0Px3KOsy9('\060' + chr(111) + '\x32' + chr(0b101001 + 0o7) + '\x32', 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + chr(55) + chr(52), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b11111 + 0o22) + chr(49) + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x31' + chr(50) + chr(55), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(50) + chr(0b101111 + 0o2) + chr(0b10010 + 0o42), 27002 - 26994), ehT0Px3KOsy9('\060' + chr(111) + '\061' + chr(51) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(0b100110 + 0o12) + '\x6f' + chr(0b110001) + chr(0b110011) + chr(0b110110), 0o10), ehT0Px3KOsy9('\060' + chr(710 - 599) + chr(0b110010) + '\x33' + chr(89 - 40), 60043 - 60035), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b100011 + 0o20) + chr(0b100101 + 0o15) + chr(0b100000 + 0o21), 20282 - 20274), ehT0Px3KOsy9(chr(989 - 941) + chr(0b1100110 + 0o11) + '\062' + chr(51) + chr(0b100 + 0o54), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110001) + chr(1164 - 1114) + '\061', 0b1000), ehT0Px3KOsy9('\060' + chr(0b101001 + 0o106) + chr(50) + '\x36' + '\x32', 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110010) + chr(0b110010 + 0o3) + chr(0b110111), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + chr(53) + chr(52), 0o10), ehT0Px3KOsy9(chr(0b1110 + 0o42) + chr(1621 - 1510) + chr(50) + '\060' + '\067', 35970 - 35962), ehT0Px3KOsy9('\x30' + '\157' + chr(2037 - 1987) + chr(0b110000) + '\067', 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b110111 + 0o70) + chr(0b101000 + 0o11) + '\065' + chr(0b110010), 49571 - 49563), ehT0Px3KOsy9('\x30' + chr(0b111000 + 0o67) + chr(51), 35826 - 35818), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\x6f' + chr(0b110010) + chr(53) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\061' + chr(0b110001) + chr(2029 - 1977), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(4702 - 4591) + chr(0b110010) + '\x31' + chr(2790 - 2735), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x32' + chr(50) + chr(599 - 548), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b110 + 0o151) + '\x31' + '\x36' + '\060', 0o10), ehT0Px3KOsy9(chr(1891 - 1843) + chr(0b1101111) + '\x35' + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(757 - 709) + chr(396 - 285) + chr(49) + '\x36' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b11001 + 0o27) + '\x6f' + chr(0b110010) + chr(110 - 59), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(5560 - 5449) + chr(0b110001) + chr(0b110010) + '\x35', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(6629 - 6518) + chr(0b10011 + 0o36) + chr(1283 - 1233) + '\062', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + chr(0b110010) + chr(53), 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(49) + chr(48) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(1105 - 1057) + chr(111) + chr(480 - 430) + '\x35' + '\x35', 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(0b10110 + 0o131) + chr(0b101 + 0o60) + chr(0b110000), 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'o'), '\144' + chr(0b1011110 + 0o7) + '\143' + chr(0b111110 + 0o61) + chr(100) + chr(0b1100101))('\x75' + chr(0b1110100) + '\146' + chr(0b101101) + chr(0b1100 + 0o54)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def iWh0bxpkdGT2():
n4ljua2gi1Pr = fpHHqmj3_uzE()
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9('\060' + '\157' + '\x32', ord("\x08"))
n4ljua2gi1Pr.bOgwkN3Z_Ukr = (ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(49) + chr(48), 14272 - 14264), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + '\x30', 62866 - 62858))
n4ljua2gi1Pr.BpQobI3VuWq4 = (ehT0Px3KOsy9(chr(48) + chr(0b111110 + 0o61) + chr(0b11101 + 0o24) + '\060', 8), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110100) + '\x30', 6622 - 6614))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer_base
|
def img2img_transformer_base():
"""Base params for local1d attention."""
hparams = image_transformer2d_base()
# learning related flags
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
# This version seems to benefit from a higher learning rate.
hparams.learning_rate = 0.2
hparams.layer_prepostprocess_dropout = 0.1
hparams.learning_rate_warmup_steps = 12000
hparams.filter_size = 2048
hparams.num_encoder_layers = 4
hparams.num_decoder_layers = 8
hparams.block_length = 256
hparams.block_width = 256
hparams.dec_attention_type = cia.AttentionType.LOCAL_1D
hparams.block_raster_scan = False
return hparams
|
python
|
def img2img_transformer_base():
"""Base params for local1d attention."""
hparams = image_transformer2d_base()
# learning related flags
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
# This version seems to benefit from a higher learning rate.
hparams.learning_rate = 0.2
hparams.layer_prepostprocess_dropout = 0.1
hparams.learning_rate_warmup_steps = 12000
hparams.filter_size = 2048
hparams.num_encoder_layers = 4
hparams.num_decoder_layers = 8
hparams.block_length = 256
hparams.block_width = 256
hparams.dec_attention_type = cia.AttentionType.LOCAL_1D
hparams.block_raster_scan = False
return hparams
|
[
"def",
"img2img_transformer_base",
"(",
")",
":",
"hparams",
"=",
"image_transformer2d_base",
"(",
")",
"# learning related flags",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"n\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"da\"",
"# This version seems to benefit from a higher learning rate.",
"hparams",
".",
"learning_rate",
"=",
"0.2",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.1",
"hparams",
".",
"learning_rate_warmup_steps",
"=",
"12000",
"hparams",
".",
"filter_size",
"=",
"2048",
"hparams",
".",
"num_encoder_layers",
"=",
"4",
"hparams",
".",
"num_decoder_layers",
"=",
"8",
"hparams",
".",
"block_length",
"=",
"256",
"hparams",
".",
"block_width",
"=",
"256",
"hparams",
".",
"dec_attention_type",
"=",
"cia",
".",
"AttentionType",
".",
"LOCAL_1D",
"hparams",
".",
"block_raster_scan",
"=",
"False",
"return",
"hparams"
] |
Base params for local1d attention.
|
[
"Base",
"params",
"for",
"local1d",
"attention",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L637-L654
|
train
|
Base params for local1d attention.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110001) + chr(0b110000) + chr(0b111 + 0o51), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + '\066' + chr(0b11001 + 0o31), 46298 - 46290), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b101111 + 0o3) + chr(0b110011) + chr(0b11000 + 0o32), ord("\x08")), ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\x6f' + chr(105 - 54) + chr(0b101 + 0o54) + chr(0b101011 + 0o14), ord("\x08")), ehT0Px3KOsy9(chr(1970 - 1922) + chr(10467 - 10356) + chr(866 - 817) + chr(1237 - 1186) + '\x36', 9220 - 9212), ehT0Px3KOsy9(chr(0b10000 + 0o40) + '\157' + chr(0b110001) + chr(2172 - 2124) + '\x31', 28497 - 28489), ehT0Px3KOsy9('\060' + '\157' + chr(0b100010 + 0o20) + chr(0b110010 + 0o0) + chr(272 - 217), 14419 - 14411), ehT0Px3KOsy9(chr(319 - 271) + '\x6f' + '\061' + chr(0b11001 + 0o32) + chr(968 - 918), 61307 - 61299), ehT0Px3KOsy9(chr(956 - 908) + chr(111) + chr(49) + '\060' + chr(53), 36673 - 36665), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b10011 + 0o37) + chr(54) + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(51) + '\063' + '\x31', 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(51) + '\x32' + '\067', 0o10), ehT0Px3KOsy9('\x30' + chr(4103 - 3992) + '\061' + chr(0b100111 + 0o20), 0o10), ehT0Px3KOsy9(chr(0b11100 + 0o24) + '\x6f' + '\063' + chr(0b110000 + 0o7) + chr(0b10111 + 0o34), 57957 - 57949), ehT0Px3KOsy9(chr(48) + chr(1622 - 1511) + '\x33' + '\064' + '\x31', 0o10), ehT0Px3KOsy9(chr(255 - 207) + '\157' + chr(0b110101) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(54) + chr(2191 - 2138), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\061' + '\065' + '\x30', 0o10), ehT0Px3KOsy9(chr(0b1110 + 0o42) + '\x6f' + '\061' + chr(0b110111) + chr(0b110100), 20929 - 20921), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110010) + chr(0b110001 + 0o4) + chr(55), 13045 - 13037), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(49) + chr(112 - 63) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x34' + '\x30', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1010110 + 0o31) + chr(50) + chr(0b1111 + 0o41) + '\060', 24958 - 24950), ehT0Px3KOsy9('\x30' + '\157' + chr(1214 - 1165) + '\063', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011) + '\x37' + chr(0b110101 + 0o1), 55823 - 55815), ehT0Px3KOsy9(chr(0b110000) + chr(0b101101 + 0o102) + '\061' + '\067' + chr(0b110011), 0b1000), ehT0Px3KOsy9('\060' + chr(4194 - 4083) + chr(2237 - 2186) + chr(0b100011 + 0o23) + chr(0b110100), 5220 - 5212), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(51) + chr(1953 - 1900) + chr(55), 32437 - 32429), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\157' + chr(0b10010 + 0o37) + chr(2610 - 2556) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(2174 - 2126) + chr(0b111000 + 0o67) + chr(1198 - 1149) + '\066' + chr(363 - 308), 15105 - 15097), ehT0Px3KOsy9(chr(404 - 356) + '\157' + chr(0b1011 + 0o47) + chr(0b110011) + '\x31', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + '\x36' + chr(55), 8), ehT0Px3KOsy9(chr(48) + chr(12080 - 11969) + '\x31' + '\x35' + chr(0b110000), 8), ehT0Px3KOsy9('\060' + '\157' + '\061' + chr(51) + '\x32', 8), ehT0Px3KOsy9('\x30' + chr(0b1001001 + 0o46) + chr(0b110110) + chr(0b11100 + 0o31), 8), ehT0Px3KOsy9(chr(48) + '\157' + '\x32' + chr(0b11011 + 0o31) + chr(0b110001), 55792 - 55784), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(111) + chr(0b1110 + 0o45) + chr(53) + chr(2327 - 2278), 0o10), ehT0Px3KOsy9(chr(96 - 48) + chr(0b1101111) + chr(1756 - 1706) + chr(0b0 + 0o61), 34032 - 34024), ehT0Px3KOsy9('\060' + '\x6f' + '\062' + chr(1753 - 1704) + chr(0b11 + 0o64), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(2228 - 2179) + chr(1085 - 1035), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + '\157' + '\065' + '\x30', ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'n'), chr(0b1100100) + '\145' + chr(0b1100011) + '\157' + '\x64' + chr(0b1100101))(chr(0b1011110 + 0o27) + chr(0b1101110 + 0o6) + chr(10024 - 9922) + chr(0b11111 + 0o16) + chr(0b10000 + 0o50)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def iylO1AJJesvo():
n4ljua2gi1Pr = Z9uDCoX3BqFw()
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'.'), chr(0b1100100) + chr(0b1100101) + chr(99) + chr(0b1000110 + 0o51) + chr(8091 - 7991) + chr(101))('\x75' + '\x74' + '\146' + chr(278 - 233) + '\070')
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'$\xd3'), '\144' + chr(101) + chr(0b1100011) + chr(3484 - 3373) + chr(5306 - 5206) + '\145')(chr(0b1011001 + 0o34) + '\x74' + '\x66' + chr(45) + '\x38')
n4ljua2gi1Pr.QGSIpd_yUNzU = 0.2
n4ljua2gi1Pr.RW_xSzp18UeS = 0.1
n4ljua2gi1Pr.fHyhoyGmdvM9 = ehT0Px3KOsy9('\060' + chr(4969 - 4858) + chr(236 - 186) + chr(1708 - 1653) + '\x33' + '\x34' + chr(861 - 813), 53269 - 53261)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9(chr(0b110000) + chr(10703 - 10592) + chr(1648 - 1596) + chr(0b11001 + 0o27) + chr(48) + '\060', 0o10)
n4ljua2gi1Pr.RS6YkARoTleN = ehT0Px3KOsy9('\x30' + chr(111) + chr(1650 - 1598), ord("\x08"))
n4ljua2gi1Pr.pRi6YFAYEnH4 = ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110001) + '\060', 55266 - 55258)
n4ljua2gi1Pr.MMwtQ0bPonxt = ehT0Px3KOsy9('\x30' + '\x6f' + chr(1034 - 982) + '\060' + chr(0b11 + 0o55), 4097 - 4089)
n4ljua2gi1Pr.H_cF2TKAb4ed = ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(52) + chr(48) + chr(0b1 + 0o57), 8)
n4ljua2gi1Pr.h3BUtwwQ_ZW5 = oIL3U1EOcJgs.AttentionType.LOCAL_1D
n4ljua2gi1Pr.wG71c38XLLpE = ehT0Px3KOsy9('\x30' + '\157' + chr(48), ord("\x08"))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer_b3
|
def img2img_transformer_b3():
"""Current best hparams for local 1d."""
hparams = img2img_transformer_base()
hparams.batch_size = 2
hparams.layer_preprocess_sequence = "none"
hparams.layer_postprocess_sequence = "dan"
hparams.block_length = 128
hparams.sampling_temp = 0.9
return hparams
|
python
|
def img2img_transformer_b3():
"""Current best hparams for local 1d."""
hparams = img2img_transformer_base()
hparams.batch_size = 2
hparams.layer_preprocess_sequence = "none"
hparams.layer_postprocess_sequence = "dan"
hparams.block_length = 128
hparams.sampling_temp = 0.9
return hparams
|
[
"def",
"img2img_transformer_b3",
"(",
")",
":",
"hparams",
"=",
"img2img_transformer_base",
"(",
")",
"hparams",
".",
"batch_size",
"=",
"2",
"hparams",
".",
"layer_preprocess_sequence",
"=",
"\"none\"",
"hparams",
".",
"layer_postprocess_sequence",
"=",
"\"dan\"",
"hparams",
".",
"block_length",
"=",
"128",
"hparams",
".",
"sampling_temp",
"=",
"0.9",
"return",
"hparams"
] |
Current best hparams for local 1d.
|
[
"Current",
"best",
"hparams",
"for",
"local",
"1d",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L678-L686
|
train
|
Current best hparams for local 1d.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + '\157' + '\x35', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b111 + 0o53) + chr(917 - 862) + chr(2222 - 2171), 64096 - 64088), ehT0Px3KOsy9(chr(48) + chr(111) + '\x35' + '\x35', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(50) + chr(1762 - 1709) + chr(1723 - 1673), 60434 - 60426), ehT0Px3KOsy9(chr(0b110000) + chr(0b101101 + 0o102) + chr(49) + '\067' + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b111010 + 0o65) + '\061' + chr(2153 - 2099) + '\061', 0o10), ehT0Px3KOsy9('\060' + chr(1845 - 1734) + chr(49) + '\x33' + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b110 + 0o151) + chr(1087 - 1037) + chr(51) + chr(50), 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110011) + chr(51) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(0b1101111) + chr(0b1000 + 0o50), ord("\x08")), ehT0Px3KOsy9('\060' + chr(135 - 24) + chr(49) + '\062' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(0b101111 + 0o3) + chr(0b100001 + 0o25), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b11 + 0o60) + '\064' + '\x35', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + chr(50) + '\x37', 0o10), ehT0Px3KOsy9(chr(0b10010 + 0o36) + '\157' + chr(51) + chr(1475 - 1421), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011011 + 0o24) + chr(1904 - 1853) + chr(50) + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(111) + chr(1496 - 1447) + '\062' + '\063', 0b1000), ehT0Px3KOsy9(chr(314 - 266) + chr(0b1101111) + chr(323 - 274) + '\060', 55323 - 55315), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b1101111) + '\x32' + chr(2421 - 2366) + chr(0b100101 + 0o14), 46532 - 46524), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\061' + '\060' + chr(48), 33075 - 33067), ehT0Px3KOsy9(chr(0b11011 + 0o25) + chr(0b1101111) + '\x33' + chr(51) + chr(0b110100), 28795 - 28787), ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(0b1000001 + 0o56) + chr(54) + '\x31', 0b1000), ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(111) + chr(51) + chr(52) + chr(1361 - 1307), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(49) + '\x30' + chr(897 - 846), 10266 - 10258), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + '\x33' + chr(49), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x32' + chr(0b110110) + chr(0b10010 + 0o36), 0b1000), ehT0Px3KOsy9(chr(0b100000 + 0o20) + chr(0b1101111) + chr(51) + chr(0b110110) + chr(0b110100), 3791 - 3783), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(0b110000 + 0o3) + chr(0b10001 + 0o42), 0b1000), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(0b1101111) + chr(0b101110 + 0o4) + '\x31' + chr(859 - 805), 12307 - 12299), ehT0Px3KOsy9(chr(1579 - 1531) + chr(111) + '\x33' + chr(0b110101) + chr(0b110101 + 0o0), 0o10), ehT0Px3KOsy9(chr(1553 - 1505) + chr(111) + '\x33' + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(2094 - 2046) + '\x6f' + '\x32' + '\064' + '\064', 7187 - 7179), ehT0Px3KOsy9(chr(0b110000) + chr(0b1010000 + 0o37) + '\x32' + '\064' + chr(0b1111 + 0o44), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101001 + 0o6) + '\x31' + chr(2170 - 2120) + chr(2250 - 2202), 0b1000), ehT0Px3KOsy9(chr(1316 - 1268) + '\x6f' + chr(1685 - 1636) + chr(373 - 324) + chr(749 - 695), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011011 + 0o24) + '\x31' + chr(0b110100) + chr(0b11000 + 0o34), 47893 - 47885), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\063' + '\064' + chr(307 - 257), ord("\x08")), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(111) + chr(0b111 + 0o54) + '\x36' + chr(48), 44931 - 44923), ehT0Px3KOsy9(chr(1885 - 1837) + chr(0b111100 + 0o63) + chr(0b11101 + 0o26) + chr(48) + chr(0b10101 + 0o36), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(51) + chr(1420 - 1367) + '\067', 15266 - 15258)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(2187 - 2139) + chr(4503 - 4392) + '\065' + chr(1955 - 1907), 26924 - 26916)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'+'), '\144' + '\x65' + chr(9132 - 9033) + chr(0b1101111) + chr(100) + chr(9121 - 9020))(chr(0b1010100 + 0o41) + '\x74' + '\x66' + chr(0b1000 + 0o45) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def otoZWObNM1MI():
n4ljua2gi1Pr = iylO1AJJesvo()
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(1235 - 1187) + chr(0b101 + 0o152) + chr(50), 16099 - 16091)
n4ljua2gi1Pr.WjY1aZ7lwLOu = xafqLlk3kkUe(SXOLrMavuUCe(b'k\xd4C\xcf'), chr(0b1000010 + 0o42) + chr(101) + chr(99) + chr(111) + chr(0b1001100 + 0o30) + chr(101))('\x75' + chr(0b11000 + 0o134) + '\x66' + chr(0b101101) + chr(0b111000))
n4ljua2gi1Pr.s6T_PoakASTI = xafqLlk3kkUe(SXOLrMavuUCe(b'a\xdaC'), chr(6095 - 5995) + chr(0b1100101) + '\x63' + '\157' + chr(100) + '\145')(chr(0b1110101) + chr(3794 - 3678) + '\x66' + '\055' + chr(652 - 596))
n4ljua2gi1Pr.MMwtQ0bPonxt = ehT0Px3KOsy9(chr(0b110000) + '\157' + '\062' + chr(1067 - 1019) + chr(0b110000), 14393 - 14385)
n4ljua2gi1Pr.Ep30xVZP6Jij = 0.9
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer_dilated
|
def img2img_transformer_dilated():
"""Try dilated."""
hparams = img2img_transformer_base()
hparams.add_hparam("num_memory_blocks", 1)
hparams.num_heads = 8
hparams.attention_key_channels = hparams.attention_value_channels = 0
hparams.hidden_size = 512
hparams.filter_size = 2048
hparams.num_decoder_layers = 8
hparams.sampling_method = "random"
hparams.gap_sizes = [0, 16, 64, 0, 16, 64, 128, 0]
hparams.dec_attention_type = cia.AttentionType.DILATED
hparams.img_len = 64
hparams.block_length = 128
hparams.block_width = 128
return hparams
|
python
|
def img2img_transformer_dilated():
"""Try dilated."""
hparams = img2img_transformer_base()
hparams.add_hparam("num_memory_blocks", 1)
hparams.num_heads = 8
hparams.attention_key_channels = hparams.attention_value_channels = 0
hparams.hidden_size = 512
hparams.filter_size = 2048
hparams.num_decoder_layers = 8
hparams.sampling_method = "random"
hparams.gap_sizes = [0, 16, 64, 0, 16, 64, 128, 0]
hparams.dec_attention_type = cia.AttentionType.DILATED
hparams.img_len = 64
hparams.block_length = 128
hparams.block_width = 128
return hparams
|
[
"def",
"img2img_transformer_dilated",
"(",
")",
":",
"hparams",
"=",
"img2img_transformer_base",
"(",
")",
"hparams",
".",
"add_hparam",
"(",
"\"num_memory_blocks\"",
",",
"1",
")",
"hparams",
".",
"num_heads",
"=",
"8",
"hparams",
".",
"attention_key_channels",
"=",
"hparams",
".",
"attention_value_channels",
"=",
"0",
"hparams",
".",
"hidden_size",
"=",
"512",
"hparams",
".",
"filter_size",
"=",
"2048",
"hparams",
".",
"num_decoder_layers",
"=",
"8",
"hparams",
".",
"sampling_method",
"=",
"\"random\"",
"hparams",
".",
"gap_sizes",
"=",
"[",
"0",
",",
"16",
",",
"64",
",",
"0",
",",
"16",
",",
"64",
",",
"128",
",",
"0",
"]",
"hparams",
".",
"dec_attention_type",
"=",
"cia",
".",
"AttentionType",
".",
"DILATED",
"hparams",
".",
"img_len",
"=",
"64",
"hparams",
".",
"block_length",
"=",
"128",
"hparams",
".",
"block_width",
"=",
"128",
"return",
"hparams"
] |
Try dilated.
|
[
"Try",
"dilated",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L760-L775
|
train
|
Try dilated.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(1303 - 1192) + '\061' + chr(48), ord("\x08")), ehT0Px3KOsy9('\060' + chr(674 - 563) + chr(0b110001) + chr(370 - 321) + '\061', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b101111 + 0o6) + '\064', 0b1000), ehT0Px3KOsy9(chr(48) + chr(8210 - 8099) + '\x32' + '\x30' + chr(0b100 + 0o57), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + '\x37' + chr(50), 0o10), ehT0Px3KOsy9('\060' + chr(2389 - 2278) + chr(0b100100 + 0o16) + chr(0b101011 + 0o14) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(0b1101111) + chr(0b110001) + '\065' + '\x34', 48206 - 48198), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(225 - 173) + chr(0b110001), 29997 - 29989), ehT0Px3KOsy9('\060' + '\157' + chr(53) + chr(0b1111 + 0o44), 41924 - 41916), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + chr(1720 - 1609) + chr(0b1011 + 0o52) + '\065', ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110011) + chr(0b110100), 31355 - 31347), ehT0Px3KOsy9(chr(531 - 483) + chr(0b1000000 + 0o57) + '\x33' + chr(0b110111) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + '\062' + chr(0b110110) + chr(0b110100), 29169 - 29161), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110001) + '\067' + chr(55), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + '\x34' + '\066', 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b100001 + 0o22) + '\060' + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1101111) + chr(49) + chr(0b101010 + 0o13) + chr(567 - 513), 0o10), ehT0Px3KOsy9(chr(0b1111 + 0o41) + '\157' + '\061' + '\x35' + chr(2773 - 2718), 0b1000), ehT0Px3KOsy9(chr(0b100101 + 0o13) + '\157' + chr(0b110010) + chr(446 - 398) + chr(52), 53303 - 53295), ehT0Px3KOsy9(chr(48) + chr(674 - 563) + chr(50) + '\065' + chr(53), 2503 - 2495), ehT0Px3KOsy9(chr(1954 - 1906) + chr(9176 - 9065) + chr(499 - 448) + chr(2067 - 2017), 1304 - 1296), ehT0Px3KOsy9(chr(48) + chr(111) + '\x34' + chr(0b10111 + 0o31), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062' + chr(2170 - 2117) + '\060', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b0 + 0o61) + '\x33' + '\x35', 40092 - 40084), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33' + chr(52) + chr(0b11 + 0o55), 0o10), ehT0Px3KOsy9('\060' + chr(716 - 605) + chr(0b10001 + 0o40) + chr(2313 - 2261) + '\x32', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b100100 + 0o113) + chr(50) + '\062' + chr(0b100101 + 0o14), ord("\x08")), ehT0Px3KOsy9(chr(404 - 356) + chr(0b1100110 + 0o11) + chr(0b110010) + chr(2134 - 2079) + chr(0b100111 + 0o16), 10644 - 10636), ehT0Px3KOsy9(chr(48) + chr(0b1000100 + 0o53) + chr(51) + chr(0b1101 + 0o46), 0o10), ehT0Px3KOsy9(chr(48) + chr(11310 - 11199) + chr(0b110010) + chr(0b10111 + 0o40) + chr(0b100011 + 0o20), 23849 - 23841), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110001) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(114 - 66) + chr(0b1101111) + chr(0b110001) + chr(0b101001 + 0o11) + '\x37', 0o10), ehT0Px3KOsy9('\060' + chr(927 - 816) + '\x34' + '\x33', 57168 - 57160), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110011) + chr(51) + '\x31', 10238 - 10230), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x33' + '\x30' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(221 - 173) + chr(0b1101111) + '\x36' + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(9160 - 9049) + '\061' + chr(48) + chr(688 - 639), 0b1000), ehT0Px3KOsy9('\060' + chr(9953 - 9842) + '\063' + chr(50) + chr(0b10011 + 0o40), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110 + 0o55) + '\x33' + chr(0b110001), 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(185 - 74) + chr(0b110101) + chr(0b10010 + 0o36), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'&'), chr(100) + '\x65' + chr(0b1100011) + '\157' + chr(2157 - 2057) + '\145')('\x75' + chr(9435 - 9319) + '\146' + chr(45) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def tF_SduQfCPX_():
n4ljua2gi1Pr = iylO1AJJesvo()
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'i-H\xdc\x92\xca\r\x93J`'), chr(0b1001110 + 0o26) + chr(0b1100101) + chr(2987 - 2888) + chr(5201 - 5090) + chr(0b1100010 + 0o2) + chr(0b110000 + 0o65))(chr(117) + chr(116) + chr(0b10101 + 0o121) + '\x2d' + chr(201 - 145)))(xafqLlk3kkUe(SXOLrMavuUCe(b'f<A\xdc\x97\xdf\x01\x8eYtr~\x91\xdb\x1c\r\xb2'), chr(0b1100100) + chr(0b1000110 + 0o37) + chr(99) + chr(0b1101111) + chr(100) + chr(0b11100 + 0o111))(chr(117) + chr(0b1110100) + '\146' + chr(0b100101 + 0o10) + chr(0b100001 + 0o27)), ehT0Px3KOsy9('\x30' + chr(0b10111 + 0o130) + chr(1951 - 1902), ord("\x08")))
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(111) + chr(49) + chr(48), 8)
n4ljua2gi1Pr.Hj_JCZasfmqG = n4ljua2gi1Pr.PZHUuenu09ti = ehT0Px3KOsy9('\x30' + chr(111) + chr(641 - 593), 0b1000)
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9(chr(1813 - 1765) + chr(12246 - 12135) + chr(0b11100 + 0o25) + chr(48) + chr(1230 - 1182) + chr(554 - 506), 0b1000)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9(chr(0b10110 + 0o32) + '\157' + chr(1740 - 1688) + chr(882 - 834) + chr(0b110000) + chr(0b110000), 0o10)
n4ljua2gi1Pr.pRi6YFAYEnH4 = ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(49) + chr(48), 8)
n4ljua2gi1Pr.Ud1InQ7hapop = xafqLlk3kkUe(SXOLrMavuUCe(b'z(B\xe7\x95\xd7'), '\x64' + chr(0b1100101) + chr(0b1100011) + '\x6f' + chr(0b1100100) + '\x65')('\x75' + chr(3007 - 2891) + '\146' + '\055' + '\070')
n4ljua2gi1Pr.KoBu_7YSNAF7 = [ehT0Px3KOsy9(chr(48) + chr(2092 - 1981) + chr(0b110000), 8), ehT0Px3KOsy9(chr(867 - 819) + chr(0b11100 + 0o123) + chr(0b1111 + 0o43) + chr(48), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110001) + '\x30' + '\x30', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\060', 8), ehT0Px3KOsy9('\x30' + chr(10304 - 10193) + chr(0b1111 + 0o43) + chr(790 - 742), 8), ehT0Px3KOsy9('\x30' + chr(0b1001010 + 0o45) + chr(1116 - 1067) + chr(48) + chr(48), 8), ehT0Px3KOsy9('\x30' + chr(9654 - 9543) + chr(400 - 350) + chr(48) + chr(48), 28909 - 28901), ehT0Px3KOsy9(chr(950 - 902) + chr(4350 - 4239) + chr(0b110000), 8)]
n4ljua2gi1Pr.h3BUtwwQ_ZW5 = oIL3U1EOcJgs.AttentionType.DILATED
n4ljua2gi1Pr.laxD7jy5y7k1 = ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110001) + chr(0b100001 + 0o17) + '\060', 8)
n4ljua2gi1Pr.MMwtQ0bPonxt = ehT0Px3KOsy9('\x30' + chr(5931 - 5820) + chr(1762 - 1712) + chr(48) + chr(1725 - 1677), 8)
n4ljua2gi1Pr.H_cF2TKAb4ed = ehT0Px3KOsy9(chr(48) + '\157' + chr(1811 - 1761) + chr(48) + '\x30', 8)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer_base_tpu
|
def img2img_transformer_base_tpu():
"""Hparams for training img2img_transformer on tpu."""
hparams = img2img_transformer_base()
update_hparams_for_tpu(hparams)
hparams.batch_size = 2
hparams.num_heads = 4 # heads are expensive on tpu
hparams.num_decoder_layers = 8
hparams.num_encoder_layers = 4
hparams.shared_embedding_and_softmax_weights = False
return hparams
|
python
|
def img2img_transformer_base_tpu():
"""Hparams for training img2img_transformer on tpu."""
hparams = img2img_transformer_base()
update_hparams_for_tpu(hparams)
hparams.batch_size = 2
hparams.num_heads = 4 # heads are expensive on tpu
hparams.num_decoder_layers = 8
hparams.num_encoder_layers = 4
hparams.shared_embedding_and_softmax_weights = False
return hparams
|
[
"def",
"img2img_transformer_base_tpu",
"(",
")",
":",
"hparams",
"=",
"img2img_transformer_base",
"(",
")",
"update_hparams_for_tpu",
"(",
"hparams",
")",
"hparams",
".",
"batch_size",
"=",
"2",
"hparams",
".",
"num_heads",
"=",
"4",
"# heads are expensive on tpu",
"hparams",
".",
"num_decoder_layers",
"=",
"8",
"hparams",
".",
"num_encoder_layers",
"=",
"4",
"hparams",
".",
"shared_embedding_and_softmax_weights",
"=",
"False",
"return",
"hparams"
] |
Hparams for training img2img_transformer on tpu.
|
[
"Hparams",
"for",
"training",
"img2img_transformer",
"on",
"tpu",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L794-L803
|
train
|
Hparams for training img2img_transformer on tpu.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(1532 - 1484) + chr(111) + chr(487 - 437) + '\x31' + chr(0b11011 + 0o31), 54256 - 54248), ehT0Px3KOsy9(chr(48) + chr(0b11 + 0o154) + chr(382 - 329) + chr(51), 25614 - 25606), ehT0Px3KOsy9('\x30' + chr(5450 - 5339) + '\x33' + '\x33' + chr(52), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b11111 + 0o26) + '\x30', 26177 - 26169), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\063' + '\064' + chr(2643 - 2591), 10702 - 10694), ehT0Px3KOsy9(chr(2278 - 2230) + '\157' + '\x31' + '\064', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + '\064', 8), ehT0Px3KOsy9(chr(1807 - 1759) + chr(0b1101111) + '\x32' + chr(0b110101) + chr(0b110000), 59956 - 59948), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b101100 + 0o6) + chr(507 - 457) + chr(54), 24856 - 24848), ehT0Px3KOsy9(chr(48) + chr(0b1100 + 0o143) + chr(1210 - 1161) + chr(48) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b101011 + 0o104) + chr(0b101111 + 0o2) + chr(49), 0o10), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(0b1101111) + '\062' + chr(0b110100) + chr(55), 0b1000), ehT0Px3KOsy9(chr(708 - 660) + chr(111) + '\060', ord("\x08")), ehT0Px3KOsy9('\060' + chr(5350 - 5239) + chr(0b101001 + 0o11) + '\063', 669 - 661), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110010) + chr(53) + chr(0b1000 + 0o51), 28618 - 28610), ehT0Px3KOsy9(chr(48) + chr(0b111 + 0o150) + chr(906 - 856) + chr(50) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(51) + '\x33' + '\x35', 0b1000), ehT0Px3KOsy9(chr(1503 - 1455) + '\157' + chr(49) + chr(0b110110) + chr(0b11010 + 0o32), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(8071 - 7960) + chr(0b101000 + 0o12) + chr(2536 - 2484) + chr(0b10110 + 0o32), 0b1000), ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(5264 - 5153) + chr(0b110010) + chr(54) + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011 + 0o144) + chr(0b110011) + chr(0b101110 + 0o2) + '\x34', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b111100 + 0o63) + chr(1735 - 1686) + chr(0b110011) + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b100100 + 0o113) + chr(0b110001 + 0o1) + '\064' + chr(2315 - 2265), 37435 - 37427), ehT0Px3KOsy9(chr(0b101011 + 0o5) + chr(0b110101 + 0o72) + chr(0b100111 + 0o13) + chr(267 - 219), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(1691 - 1580) + chr(0b110000 + 0o1) + '\x37' + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(427 - 379) + chr(111) + chr(2595 - 2544) + chr(0b1100 + 0o47) + chr(54), 0o10), ehT0Px3KOsy9('\x30' + chr(11156 - 11045) + chr(0b110011) + chr(2070 - 2021) + chr(0b101000 + 0o11), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b10100 + 0o133) + chr(1885 - 1836) + '\067' + chr(0b110000), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x32' + chr(798 - 745) + '\x30', 8), ehT0Px3KOsy9('\x30' + '\157' + chr(0b1110 + 0o44) + chr(55) + chr(157 - 106), 0b1000), ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(0b1100010 + 0o15) + chr(1630 - 1581) + '\067' + chr(2670 - 2616), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(7061 - 6950) + chr(0b110011 + 0o0) + '\065' + '\x34', 17143 - 17135), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(0b1101111) + '\x31' + '\x37' + chr(978 - 923), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(49), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + '\065', 61483 - 61475), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\061' + chr(0b110101) + chr(1969 - 1914), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + '\061' + chr(1532 - 1480), 55286 - 55278), ehT0Px3KOsy9(chr(48) + chr(111) + '\062' + chr(0b101111 + 0o7) + '\x33', 0b1000), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(0b1101111) + '\x33' + chr(0b110010), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + chr(49) + chr(1464 - 1415), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b101011 + 0o12) + chr(0b110000), 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xa2'), chr(100) + '\x65' + '\x63' + chr(0b1101111) + chr(0b1100100) + chr(0b1111 + 0o126))(chr(0b1110101) + chr(11837 - 11721) + chr(0b1000010 + 0o44) + chr(0b101101) + chr(0b10110 + 0o42)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def pdvTXKdij0VA():
n4ljua2gi1Pr = iylO1AJJesvo()
gWr33mh0VbqT(n4ljua2gi1Pr)
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(0b110000) + chr(0b111010 + 0o65) + '\062', ord("\x08"))
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(111) + '\x34', 0o10)
n4ljua2gi1Pr.pRi6YFAYEnH4 = ehT0Px3KOsy9(chr(0b110 + 0o52) + chr(111) + chr(399 - 350) + '\x30', 0b1000)
n4ljua2gi1Pr.RS6YkARoTleN = ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110100), 8)
n4ljua2gi1Pr.qVamxim0L2I1 = ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(10903 - 10792) + chr(48), 8)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer2d_n31
|
def img2img_transformer2d_n31():
"""Set of hyperparameters."""
hparams = img2img_transformer2d_base()
hparams.batch_size = 1
hparams.num_encoder_layers = 6
hparams.num_decoder_layers = 12
hparams.num_heads = 8
hparams.query_shape = (16, 32)
hparams.memory_flange = (16, 32)
return hparams
|
python
|
def img2img_transformer2d_n31():
"""Set of hyperparameters."""
hparams = img2img_transformer2d_base()
hparams.batch_size = 1
hparams.num_encoder_layers = 6
hparams.num_decoder_layers = 12
hparams.num_heads = 8
hparams.query_shape = (16, 32)
hparams.memory_flange = (16, 32)
return hparams
|
[
"def",
"img2img_transformer2d_n31",
"(",
")",
":",
"hparams",
"=",
"img2img_transformer2d_base",
"(",
")",
"hparams",
".",
"batch_size",
"=",
"1",
"hparams",
".",
"num_encoder_layers",
"=",
"6",
"hparams",
".",
"num_decoder_layers",
"=",
"12",
"hparams",
".",
"num_heads",
"=",
"8",
"hparams",
".",
"query_shape",
"=",
"(",
"16",
",",
"32",
")",
"hparams",
".",
"memory_flange",
"=",
"(",
"16",
",",
"32",
")",
"return",
"hparams"
] |
Set of hyperparameters.
|
[
"Set",
"of",
"hyperparameters",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L829-L838
|
train
|
Hparams for img2img_transformer2d_n31.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + chr(2690 - 2579) + chr(49) + '\x35' + chr(49), 0o10), ehT0Px3KOsy9('\060' + '\157' + '\x33' + chr(48) + '\x34', 46504 - 46496), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x33' + chr(52) + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b10101 + 0o36) + chr(1169 - 1120) + chr(55), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + chr(0b110000 + 0o3) + chr(0b110110), 0o10), ehT0Px3KOsy9('\x30' + chr(0b100110 + 0o111) + '\061' + chr(52) + chr(2099 - 2050), 0o10), ehT0Px3KOsy9('\060' + chr(0b1001 + 0o146) + chr(0b110 + 0o53) + '\x36' + '\065', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\062' + '\x34' + chr(0b101010 + 0o10), 62506 - 62498), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(111) + chr(0b10000 + 0o41) + '\x35' + '\x36', 50155 - 50147), ehT0Px3KOsy9(chr(51 - 3) + '\157' + chr(0b0 + 0o61) + '\061' + chr(0b101101 + 0o12), ord("\x08")), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(0b1101111) + chr(0b110010) + chr(1996 - 1943) + chr(55), 0b1000), ehT0Px3KOsy9(chr(1680 - 1632) + chr(0b100110 + 0o111) + chr(2443 - 2392) + chr(0b110001) + chr(51), 53864 - 53856), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110001) + chr(936 - 883) + chr(0b100001 + 0o25), 8), ehT0Px3KOsy9(chr(400 - 352) + chr(0b1010011 + 0o34) + chr(0b110010) + '\x37' + chr(55), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b100110 + 0o21) + chr(1995 - 1943), 11006 - 10998), ehT0Px3KOsy9(chr(48) + chr(6839 - 6728) + chr(49) + '\x36' + chr(0b10010 + 0o43), 8), ehT0Px3KOsy9(chr(48) + '\157' + chr(1085 - 1035) + chr(490 - 441) + chr(0b11100 + 0o33), 0o10), ehT0Px3KOsy9(chr(48) + chr(847 - 736) + '\x32' + chr(0b100100 + 0o20) + '\064', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1183 - 1132) + chr(847 - 795) + chr(0b1100 + 0o45), 0o10), ehT0Px3KOsy9(chr(1215 - 1167) + chr(0b1101111) + '\x32' + chr(526 - 471), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x32' + chr(0b10011 + 0o42) + chr(49), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(1918 - 1868) + chr(0b110111) + chr(0b1010 + 0o50), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b11010 + 0o125) + '\063' + '\065' + chr(0b110000), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + chr(0b11111 + 0o23) + '\065', ord("\x08")), ehT0Px3KOsy9(chr(2030 - 1982) + chr(0b1101111) + '\063' + chr(52), 9001 - 8993), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b100111 + 0o15) + chr(0b11001 + 0o33), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1100110 + 0o11) + chr(894 - 844) + '\x31' + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(0b10111 + 0o31) + '\x6f' + chr(0b110000 + 0o2) + chr(0b10100 + 0o42) + '\x32', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b11101 + 0o27), 0o10), ehT0Px3KOsy9('\060' + chr(5336 - 5225) + chr(0b110010) + chr(55) + chr(51), 0o10), ehT0Px3KOsy9('\x30' + chr(8444 - 8333) + chr(1669 - 1618) + chr(0b110100), 8), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(4509 - 4398) + '\x33' + chr(0b100001 + 0o23), 8), ehT0Px3KOsy9('\060' + '\157' + chr(0b100 + 0o55) + chr(48) + chr(0b101101 + 0o7), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(2447 - 2397) + chr(55) + chr(0b110111), 8), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(0b1101111) + chr(0b110011), 53607 - 53599), ehT0Px3KOsy9('\060' + chr(111) + '\x32' + chr(165 - 113) + '\x33', 1635 - 1627), ehT0Px3KOsy9(chr(0b11000 + 0o30) + chr(0b1001001 + 0o46) + chr(0b110011) + chr(48), 58037 - 58029), ehT0Px3KOsy9('\060' + '\x6f' + '\x32' + chr(0b101010 + 0o11) + '\064', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\061' + chr(816 - 763) + chr(53), 0b1000), ehT0Px3KOsy9('\060' + chr(111) + '\063' + chr(2254 - 2206), 8)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x35' + chr(0b110000), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'G'), chr(100) + '\x65' + chr(4294 - 4195) + chr(0b1101001 + 0o6) + '\x64' + chr(0b1100101))(chr(0b1010111 + 0o36) + chr(0b11001 + 0o133) + chr(0b100110 + 0o100) + '\055' + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def alUyWBM1bEdL():
n4ljua2gi1Pr = AnTsaAZ4phGl()
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(102 - 54) + chr(7893 - 7782) + chr(1581 - 1532), 0o10)
n4ljua2gi1Pr.RS6YkARoTleN = ehT0Px3KOsy9('\x30' + chr(0b1100010 + 0o15) + '\066', 37997 - 37989)
n4ljua2gi1Pr.pRi6YFAYEnH4 = ehT0Px3KOsy9(chr(798 - 750) + chr(111) + chr(49) + '\064', 0b1000)
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9('\060' + chr(111) + chr(0b10100 + 0o35) + chr(0b110000), 0o10)
n4ljua2gi1Pr.bOgwkN3Z_Ukr = (ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(111) + chr(0b110010) + chr(0b100100 + 0o14), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1100010 + 0o15) + chr(0b10000 + 0o44) + chr(2155 - 2107), ord("\x08")))
n4ljua2gi1Pr.BpQobI3VuWq4 = (ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(722 - 611) + chr(0b110010) + chr(0b101010 + 0o6), 8), ehT0Px3KOsy9('\x30' + '\157' + chr(1817 - 1765) + chr(0b110000), 8))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer2d_n24
|
def img2img_transformer2d_n24():
"""Set of hyperparameters."""
hparams = img2img_transformer2d_base()
hparams.batch_size = 1
hparams.hidden_size = 1024
hparams.filter_size = 2048
hparams.layer_prepostprocess_dropout = 0.2
hparams.num_decoder_layers = 8
hparams.query_shape = (8, 16)
hparams.memory_flange = (8, 32)
return hparams
|
python
|
def img2img_transformer2d_n24():
"""Set of hyperparameters."""
hparams = img2img_transformer2d_base()
hparams.batch_size = 1
hparams.hidden_size = 1024
hparams.filter_size = 2048
hparams.layer_prepostprocess_dropout = 0.2
hparams.num_decoder_layers = 8
hparams.query_shape = (8, 16)
hparams.memory_flange = (8, 32)
return hparams
|
[
"def",
"img2img_transformer2d_n24",
"(",
")",
":",
"hparams",
"=",
"img2img_transformer2d_base",
"(",
")",
"hparams",
".",
"batch_size",
"=",
"1",
"hparams",
".",
"hidden_size",
"=",
"1024",
"hparams",
".",
"filter_size",
"=",
"2048",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.2",
"hparams",
".",
"num_decoder_layers",
"=",
"8",
"hparams",
".",
"query_shape",
"=",
"(",
"8",
",",
"16",
")",
"hparams",
".",
"memory_flange",
"=",
"(",
"8",
",",
"32",
")",
"return",
"hparams"
] |
Set of hyperparameters.
|
[
"Set",
"of",
"hyperparameters",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L842-L852
|
train
|
Hparams for img2img_transformer2d_n24.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b1011101 + 0o22) + chr(0b110001) + chr(0b110000) + chr(55), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + '\x31' + chr(55) + chr(245 - 190), 0o10), ehT0Px3KOsy9('\060' + '\157' + '\x31' + chr(0b110111) + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(815 - 767) + '\157' + chr(51) + chr(0b110000) + chr(1527 - 1477), 0o10), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(111) + chr(0b11010 + 0o30) + '\x30' + '\x32', 22421 - 22413), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010) + '\064' + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x33' + chr(0b110101) + '\066', 0b1000), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(111) + chr(49) + chr(54) + '\065', 19165 - 19157), ehT0Px3KOsy9(chr(0b110000) + chr(10952 - 10841) + '\x31' + chr(48) + chr(2191 - 2138), 36602 - 36594), ehT0Px3KOsy9(chr(432 - 384) + chr(8492 - 8381) + chr(0b101111 + 0o7) + chr(0b101000 + 0o15), ord("\x08")), ehT0Px3KOsy9(chr(1694 - 1646) + chr(0b1011101 + 0o22) + chr(0b110010) + chr(318 - 269) + '\x35', 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b11010 + 0o30) + '\x37' + chr(448 - 393), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(50) + chr(55) + chr(0b100100 + 0o21), 0b1000), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(8353 - 8242) + chr(0b110001) + '\x32' + chr(0b10010 + 0o40), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b1100 + 0o45) + chr(0b1011 + 0o47) + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51) + chr(362 - 310) + '\x33', 26663 - 26655), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(0b1101111) + '\063' + '\067' + chr(48), 0b1000), ehT0Px3KOsy9(chr(995 - 947) + chr(111) + '\x33' + '\062' + chr(0b110011), 13807 - 13799), ehT0Px3KOsy9(chr(0b101011 + 0o5) + '\x6f' + chr(0b1111 + 0o44) + '\x35' + chr(0b110000), 10806 - 10798), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\061' + chr(0b110111) + '\060', 0b1000), ehT0Px3KOsy9(chr(820 - 772) + chr(111) + chr(0b1110 + 0o44) + '\062', 32345 - 32337), ehT0Px3KOsy9('\060' + '\157' + '\x32' + chr(2126 - 2074) + chr(0b110000), 38694 - 38686), ehT0Px3KOsy9(chr(0b11010 + 0o26) + '\157' + chr(0b100100 + 0o17) + chr(0b110100) + chr(1613 - 1560), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(49) + chr(54) + chr(1040 - 985), 0o10), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(111) + chr(0b110011) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b111 + 0o150) + chr(0b100001 + 0o21) + chr(691 - 636) + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(5824 - 5713) + chr(0b110001) + '\x35' + chr(0b110011), 5341 - 5333), ehT0Px3KOsy9(chr(0b0 + 0o60) + chr(0b110011 + 0o74) + chr(51) + '\x35' + chr(0b10010 + 0o43), 0b1000), ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(0b1101111) + '\063' + chr(0b10110 + 0o33) + '\065', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b111110 + 0o61) + chr(1927 - 1876) + chr(0b11111 + 0o25) + '\065', 8), ehT0Px3KOsy9(chr(48) + chr(0b1000000 + 0o57) + chr(0b110011) + chr(52) + chr(1062 - 1013), 0b1000), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(0b1101100 + 0o3) + chr(50) + chr(0b110111 + 0o0) + '\x33', 0o10), ehT0Px3KOsy9('\x30' + chr(111) + '\061' + '\067' + '\067', 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(300 - 249) + '\x31' + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b1011 + 0o45) + '\157' + chr(0b110001) + '\067' + chr(0b110111), 8), ehT0Px3KOsy9(chr(2114 - 2066) + chr(0b1101111) + chr(51) + chr(2178 - 2130) + chr(0b110 + 0o60), 0o10), ehT0Px3KOsy9('\060' + chr(0b1001000 + 0o47) + '\x37' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\x36' + '\062', 46642 - 46634), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110110) + chr(0b11010 + 0o30), 8), ehT0Px3KOsy9(chr(678 - 630) + chr(3099 - 2988) + '\x31' + '\067' + chr(0b10010 + 0o36), 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(1354 - 1306) + chr(0b1101111) + '\x35' + chr(0b110000), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'i'), chr(100) + '\145' + chr(0b1011011 + 0o10) + '\157' + '\144' + '\145')(chr(4699 - 4582) + chr(116) + '\146' + chr(0b101101) + chr(942 - 886)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def jpwVQ_kC7eQO():
n4ljua2gi1Pr = AnTsaAZ4phGl()
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(48) + '\157' + chr(49), ord("\x08"))
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9(chr(957 - 909) + chr(0b1101111) + chr(796 - 746) + chr(0b110000) + '\x30' + chr(48), 0o10)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9(chr(374 - 326) + chr(0b110011 + 0o74) + chr(0b110100) + chr(2257 - 2209) + chr(0b1010 + 0o46) + chr(48), ord("\x08"))
n4ljua2gi1Pr.RW_xSzp18UeS = 0.2
n4ljua2gi1Pr.pRi6YFAYEnH4 = ehT0Px3KOsy9('\060' + '\x6f' + chr(49) + chr(0b110000), 0b1000)
n4ljua2gi1Pr.bOgwkN3Z_Ukr = (ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x31' + chr(199 - 151), 8), ehT0Px3KOsy9(chr(48) + chr(0b1011001 + 0o26) + '\062' + '\x30', 0o10))
n4ljua2gi1Pr.BpQobI3VuWq4 = (ehT0Px3KOsy9('\060' + '\157' + '\061' + chr(2081 - 2033), 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x34' + '\x30', ord("\x08")))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer2d_tiny
|
def img2img_transformer2d_tiny():
"""Tiny params."""
hparams = img2img_transformer2d_base()
hparams.num_decoder_layers = 2
hparams.hidden_size = 128
hparams.batch_size = 4
hparams.max_length = 128
hparams.attention_key_channels = hparams.attention_value_channels = 0
hparams.filter_size = 128
hparams.num_heads = 4
hparams.pos = "timing"
hparams.img_len = 32
return hparams
|
python
|
def img2img_transformer2d_tiny():
"""Tiny params."""
hparams = img2img_transformer2d_base()
hparams.num_decoder_layers = 2
hparams.hidden_size = 128
hparams.batch_size = 4
hparams.max_length = 128
hparams.attention_key_channels = hparams.attention_value_channels = 0
hparams.filter_size = 128
hparams.num_heads = 4
hparams.pos = "timing"
hparams.img_len = 32
return hparams
|
[
"def",
"img2img_transformer2d_tiny",
"(",
")",
":",
"hparams",
"=",
"img2img_transformer2d_base",
"(",
")",
"hparams",
".",
"num_decoder_layers",
"=",
"2",
"hparams",
".",
"hidden_size",
"=",
"128",
"hparams",
".",
"batch_size",
"=",
"4",
"hparams",
".",
"max_length",
"=",
"128",
"hparams",
".",
"attention_key_channels",
"=",
"hparams",
".",
"attention_value_channels",
"=",
"0",
"hparams",
".",
"filter_size",
"=",
"128",
"hparams",
".",
"num_heads",
"=",
"4",
"hparams",
".",
"pos",
"=",
"\"timing\"",
"hparams",
".",
"img_len",
"=",
"32",
"return",
"hparams"
] |
Tiny params.
|
[
"Tiny",
"params",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L880-L892
|
train
|
Tiny params.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(0b110 + 0o151) + chr(0b101001 + 0o10) + chr(54) + '\x35', 65467 - 65459), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b111 + 0o150) + '\063' + chr(0b110010) + chr(336 - 286), 58166 - 58158), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010) + chr(54) + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51) + '\066' + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(636 - 588) + chr(0b1101111) + '\x33', 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\066' + chr(0b100101 + 0o15), 26527 - 26519), ehT0Px3KOsy9('\060' + chr(111) + chr(0b1001 + 0o52) + chr(0b110010) + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110010) + chr(49) + '\060', 0o10), ehT0Px3KOsy9('\060' + chr(0b1001110 + 0o41) + chr(0b110010) + chr(0b101111 + 0o10) + '\x31', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101110 + 0o1) + '\x33' + '\061' + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(0b101111 + 0o1) + '\157' + chr(50) + '\x36', 0o10), ehT0Px3KOsy9(chr(973 - 925) + chr(6815 - 6704) + '\x33' + chr(0b110001) + chr(0b110000), 27775 - 27767), ehT0Px3KOsy9(chr(0b110000) + chr(0b1000101 + 0o52) + chr(1074 - 1025) + chr(0b110011) + chr(801 - 746), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b1 + 0o61) + chr(51) + chr(1086 - 1033), 0o10), ehT0Px3KOsy9(chr(476 - 428) + chr(111) + chr(761 - 711) + chr(48) + chr(0b110011), 0o10), ehT0Px3KOsy9('\060' + chr(7954 - 7843) + chr(51) + chr(0b10110 + 0o41), 0o10), ehT0Px3KOsy9(chr(0b10010 + 0o36) + '\x6f' + chr(857 - 806) + chr(50) + chr(49), 6387 - 6379), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\061' + chr(964 - 915) + chr(196 - 143), 0b1000), ehT0Px3KOsy9(chr(0b10010 + 0o36) + chr(0b1001 + 0o146) + chr(0b101101 + 0o5) + chr(52) + '\x33', 35447 - 35439), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110101 + 0o2) + '\x30', 39042 - 39034), ehT0Px3KOsy9('\060' + chr(0b1010101 + 0o32) + chr(0b111 + 0o52), ord("\x08")), ehT0Px3KOsy9(chr(1190 - 1142) + '\x6f' + chr(50) + '\067' + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b10110 + 0o33) + '\x37' + '\x31', 0o10), ehT0Px3KOsy9(chr(0b110 + 0o52) + '\157' + chr(2355 - 2302) + '\x31', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\061' + chr(50), 0o10), ehT0Px3KOsy9('\x30' + chr(11541 - 11430) + chr(0b110001) + chr(0b101110 + 0o4) + chr(0b1 + 0o57), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(1186 - 1137) + chr(51) + chr(697 - 644), 0b1000), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\x6f' + chr(0b100100 + 0o15) + '\067' + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(52) + '\x30', 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(686 - 637) + chr(2320 - 2270) + chr(1115 - 1067), 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + chr(0b110100) + chr(0b10111 + 0o35), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(6754 - 6643) + chr(0b110110) + chr(0b100110 + 0o13), ord("\x08")), ehT0Px3KOsy9(chr(913 - 865) + chr(6886 - 6775) + chr(1353 - 1298) + chr(2189 - 2137), ord("\x08")), ehT0Px3KOsy9(chr(2145 - 2097) + chr(1608 - 1497) + chr(0b110001) + chr(54) + '\x33', 26920 - 26912), ehT0Px3KOsy9('\060' + chr(6818 - 6707) + chr(1708 - 1657) + chr(0b10001 + 0o46) + chr(0b10110 + 0o40), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1010111 + 0o30) + chr(1163 - 1112) + chr(0b110010) + chr(48), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001) + chr(0b100 + 0o55) + chr(0b110101), 8), ehT0Px3KOsy9('\x30' + chr(3619 - 3508) + chr(0b110011 + 0o0) + '\064' + '\064', ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + '\066' + chr(160 - 111), 8), ehT0Px3KOsy9('\060' + chr(2617 - 2506) + chr(0b10011 + 0o37) + chr(55), 14944 - 14936)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + '\157' + chr(53) + chr(0b110000), 9009 - 9001)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'['), chr(100) + '\145' + '\x63' + chr(111) + chr(100) + chr(101))(chr(564 - 447) + chr(0b111100 + 0o70) + chr(102) + '\x2d' + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def nrb2SArWrlxz():
n4ljua2gi1Pr = AnTsaAZ4phGl()
n4ljua2gi1Pr.pRi6YFAYEnH4 = ehT0Px3KOsy9(chr(48) + '\x6f' + chr(50), 0o10)
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9('\060' + '\x6f' + '\062' + chr(0b110000) + chr(0b110000), ord("\x08"))
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(48) + '\157' + chr(52), ord("\x08"))
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x32' + chr(1482 - 1434) + chr(0b100010 + 0o16), 8)
n4ljua2gi1Pr.Hj_JCZasfmqG = n4ljua2gi1Pr.PZHUuenu09ti = ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b10111 + 0o31), ord("\x08"))
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x32' + chr(0b110000) + chr(0b110000), 8)
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9('\x30' + '\x6f' + chr(2239 - 2187), 8)
n4ljua2gi1Pr.NXd0aqYJd4lK = xafqLlk3kkUe(SXOLrMavuUCe(b'\x01\xbb\xdc\x9emU'), chr(0b1100100) + chr(0b110100 + 0o61) + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + chr(0b1100101))('\165' + '\x74' + chr(102) + '\x2d' + '\070')
n4ljua2gi1Pr.laxD7jy5y7k1 = ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110100) + '\x30', ord("\x08"))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/image_transformer_2d.py
|
img2img_transformer_tiny
|
def img2img_transformer_tiny():
"""Tiny params."""
hparams = img2img_transformer2d_base()
hparams.num_hidden_layers = 2
hparams.hidden_size = 128
hparams.batch_size = 4
hparams.max_length = 128
hparams.attention_key_channels = hparams.attention_value_channels = 0
hparams.filter_size = 128
hparams.num_heads = 1
hparams.pos = "timing"
return hparams
|
python
|
def img2img_transformer_tiny():
"""Tiny params."""
hparams = img2img_transformer2d_base()
hparams.num_hidden_layers = 2
hparams.hidden_size = 128
hparams.batch_size = 4
hparams.max_length = 128
hparams.attention_key_channels = hparams.attention_value_channels = 0
hparams.filter_size = 128
hparams.num_heads = 1
hparams.pos = "timing"
return hparams
|
[
"def",
"img2img_transformer_tiny",
"(",
")",
":",
"hparams",
"=",
"img2img_transformer2d_base",
"(",
")",
"hparams",
".",
"num_hidden_layers",
"=",
"2",
"hparams",
".",
"hidden_size",
"=",
"128",
"hparams",
".",
"batch_size",
"=",
"4",
"hparams",
".",
"max_length",
"=",
"128",
"hparams",
".",
"attention_key_channels",
"=",
"hparams",
".",
"attention_value_channels",
"=",
"0",
"hparams",
".",
"filter_size",
"=",
"128",
"hparams",
".",
"num_heads",
"=",
"1",
"hparams",
".",
"pos",
"=",
"\"timing\"",
"return",
"hparams"
] |
Tiny params.
|
[
"Tiny",
"params",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/image_transformer_2d.py#L896-L907
|
train
|
Tiny params.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b10010 + 0o36) + chr(10292 - 10181) + chr(0b101110 + 0o4) + '\064' + chr(199 - 146), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1000010 + 0o55) + chr(51) + chr(51) + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(373 - 322) + chr(51), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + '\x31' + chr(783 - 732) + '\062', 33771 - 33763), ehT0Px3KOsy9(chr(48) + chr(2238 - 2127) + '\x33' + '\x35' + '\067', 53190 - 53182), ehT0Px3KOsy9(chr(1241 - 1193) + chr(8759 - 8648) + chr(0b1 + 0o60) + chr(1716 - 1664) + chr(0b10011 + 0o40), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(5035 - 4924) + '\066' + '\066', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1001110 + 0o41) + '\061' + chr(925 - 870) + '\060', 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + '\x35' + chr(0b10001 + 0o45), 46197 - 46189), ehT0Px3KOsy9('\060' + chr(5857 - 5746) + '\061' + chr(0b110101) + '\x35', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1010101 + 0o32) + chr(590 - 538) + chr(2256 - 2206), 6296 - 6288), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(0b1101111) + chr(0b110001) + chr(50) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + chr(0b11111 + 0o30) + chr(54), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(50), 20293 - 20285), ehT0Px3KOsy9(chr(1290 - 1242) + chr(9782 - 9671) + chr(50) + chr(0b100011 + 0o22) + chr(2510 - 2456), ord("\x08")), ehT0Px3KOsy9(chr(0b11111 + 0o21) + chr(0b1101111) + '\x33' + chr(331 - 281) + chr(0b11001 + 0o35), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110011) + chr(0b101111 + 0o5) + '\064', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(6157 - 6046) + '\061' + chr(48) + chr(1095 - 1040), 7251 - 7243), ehT0Px3KOsy9(chr(1919 - 1871) + '\157' + chr(0b110000 + 0o1) + chr(49) + '\x30', 19993 - 19985), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(0b110110) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(731 - 683) + '\157' + chr(0b11100 + 0o32) + chr(55), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b11110 + 0o23) + chr(0b110000) + chr(102 - 52), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b0 + 0o67), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(4544 - 4433) + '\063' + chr(48), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(4890 - 4779) + chr(50) + chr(0b110101) + '\066', 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b101101 + 0o102) + chr(0b101000 + 0o11) + chr(0b10001 + 0o45), 39605 - 39597), ehT0Px3KOsy9('\060' + chr(3206 - 3095) + '\063' + chr(1256 - 1204) + chr(983 - 931), 8), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b0 + 0o157) + chr(947 - 898) + chr(0b110000) + chr(0b110111), 8), ehT0Px3KOsy9('\060' + chr(111) + chr(1507 - 1458) + '\063' + '\067', 28647 - 28639), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011000 + 0o27) + chr(0b110010) + chr(0b110010 + 0o2) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(0b11111 + 0o21) + chr(0b1000010 + 0o55) + chr(0b110010) + '\066' + '\060', 48430 - 48422), ehT0Px3KOsy9('\060' + chr(0b1101010 + 0o5) + chr(0b11 + 0o60) + chr(51) + '\062', 8), ehT0Px3KOsy9(chr(1667 - 1619) + chr(0b1101111) + chr(51) + '\061' + chr(1255 - 1206), 18629 - 18621), ehT0Px3KOsy9(chr(297 - 249) + chr(9006 - 8895) + chr(0b110011 + 0o0) + '\060' + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(111) + chr(0b100100 + 0o15) + chr(48) + chr(1345 - 1297), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b101001 + 0o106) + '\062' + '\x32' + '\x37', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b11111 + 0o23) + chr(50) + '\065', 63539 - 63531), ehT0Px3KOsy9(chr(2000 - 1952) + chr(0b1101111) + chr(50) + chr(426 - 375) + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(640 - 591) + chr(0b110 + 0o61) + '\061', 15184 - 15176), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33' + '\x35' + '\x32', ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + '\x6f' + '\065' + chr(0b110000), 62758 - 62750)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x8b'), '\x64' + chr(0b110001 + 0o64) + chr(0b1100011) + chr(0b1101111) + '\x64' + chr(0b1100101))(chr(0b1110101) + chr(116) + '\x66' + chr(246 - 201) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def PsEigfaDe8X2():
n4ljua2gi1Pr = AnTsaAZ4phGl()
n4ljua2gi1Pr.jZh5_pLUoOoZ = ehT0Px3KOsy9('\060' + chr(1047 - 936) + chr(0b110010), 8)
n4ljua2gi1Pr.qzoyXN3kdhDL = ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + '\060' + chr(0b10111 + 0o31), 18488 - 18480)
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110100), ord("\x08"))
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(50) + chr(408 - 360) + chr(48), 8)
n4ljua2gi1Pr.Hj_JCZasfmqG = n4ljua2gi1Pr.PZHUuenu09ti = ehT0Px3KOsy9(chr(0b110000) + chr(6978 - 6867) + chr(0b110000), 0b1000)
n4ljua2gi1Pr.deybX8NJ0oEI = ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110010) + '\x30' + chr(0b11101 + 0o23), 8)
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(598 - 550) + '\x6f' + chr(573 - 524), 0o10)
n4ljua2gi1Pr.NXd0aqYJd4lK = xafqLlk3kkUe(SXOLrMavuUCe(b'\xd1\xb363\x81\x04'), chr(0b1100100) + chr(0b1100101) + '\143' + chr(0b10011 + 0o134) + '\144' + '\x65')('\165' + chr(0b1110100) + chr(9965 - 9863) + '\x2d' + chr(0b111000))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/trax/models/transformer.py
|
ResidualFeedForward
|
def ResidualFeedForward(feature_depth,
feedforward_depth,
dropout,
mode):
"""Residual feed-forward layer with normalization at start."""
return layers.Residual(
layers.LayerNorm(),
layers.Dense(feedforward_depth),
layers.Relu(),
layers.Dropout(rate=dropout, mode=mode),
layers.Dense(feature_depth),
layers.Dropout(rate=dropout, mode=mode)
)
|
python
|
def ResidualFeedForward(feature_depth,
feedforward_depth,
dropout,
mode):
"""Residual feed-forward layer with normalization at start."""
return layers.Residual(
layers.LayerNorm(),
layers.Dense(feedforward_depth),
layers.Relu(),
layers.Dropout(rate=dropout, mode=mode),
layers.Dense(feature_depth),
layers.Dropout(rate=dropout, mode=mode)
)
|
[
"def",
"ResidualFeedForward",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"dropout",
",",
"mode",
")",
":",
"return",
"layers",
".",
"Residual",
"(",
"layers",
".",
"LayerNorm",
"(",
")",
",",
"layers",
".",
"Dense",
"(",
"feedforward_depth",
")",
",",
"layers",
".",
"Relu",
"(",
")",
",",
"layers",
".",
"Dropout",
"(",
"rate",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
",",
"layers",
".",
"Dense",
"(",
"feature_depth",
")",
",",
"layers",
".",
"Dropout",
"(",
"rate",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
")"
] |
Residual feed-forward layer with normalization at start.
|
[
"Residual",
"feed",
"-",
"forward",
"layer",
"with",
"normalization",
"at",
"start",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/trax/models/transformer.py#L24-L36
|
train
|
Residual feed - forward layer with normalization at start.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b100101 + 0o13) + '\x6f' + '\062' + chr(3016 - 2961) + chr(2399 - 2349), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(49) + chr(0b110100) + '\063', 0o10), ehT0Px3KOsy9(chr(48) + chr(9602 - 9491) + chr(49) + chr(54) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(1194 - 1145) + chr(52) + '\067', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1000000 + 0o57) + chr(0b110010) + '\060' + '\x34', 0o10), ehT0Px3KOsy9(chr(1104 - 1056) + chr(0b1101111) + '\061' + chr(913 - 863) + chr(0b10 + 0o57), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1100001 + 0o16) + chr(0b1101 + 0o47) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(665 - 617) + chr(0b1101111) + chr(0b101001 + 0o12) + chr(1783 - 1734) + chr(0b11000 + 0o35), 22179 - 22171), ehT0Px3KOsy9(chr(934 - 886) + chr(111) + '\x32' + chr(0b110101) + '\067', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(7221 - 7110) + chr(0b1100 + 0o47) + '\061' + chr(428 - 373), 5218 - 5210), ehT0Px3KOsy9(chr(0b11010 + 0o26) + chr(11678 - 11567) + '\x32' + chr(0b101110 + 0o4) + '\x30', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(5280 - 5169) + chr(0b110011) + '\x37' + chr(53), ord("\x08")), ehT0Px3KOsy9('\060' + chr(9606 - 9495) + '\x31' + chr(0b110100) + '\060', 0b1000), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(9177 - 9066) + chr(49) + chr(0b10001 + 0o40) + chr(2066 - 2014), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10001 + 0o40) + chr(0b100110 + 0o14), ord("\x08")), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(2417 - 2306) + '\061' + chr(0b100011 + 0o24) + chr(1701 - 1649), 58005 - 57997), ehT0Px3KOsy9(chr(0b1110 + 0o42) + chr(0b1101111) + chr(0b110011) + chr(569 - 518) + chr(2588 - 2535), 0b1000), ehT0Px3KOsy9(chr(0b100010 + 0o16) + '\157' + chr(0b110101) + chr(0b110010), 33281 - 33273), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b100000 + 0o23) + '\x36' + chr(0b1 + 0o61), 40001 - 39993), ehT0Px3KOsy9(chr(0b110000) + chr(0b11101 + 0o122) + chr(0b1001 + 0o51), 0b1000), ehT0Px3KOsy9('\x30' + chr(11625 - 11514) + chr(50) + chr(0b110101) + chr(217 - 166), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(51) + '\x35' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1001010 + 0o45) + '\x35' + chr(55), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110010) + '\064' + '\x33', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1000011 + 0o54) + '\x33' + chr(48) + '\067', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(10187 - 10076) + chr(0b110011) + chr(2152 - 2098), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x36' + '\066', 0o10), ehT0Px3KOsy9('\060' + '\x6f' + '\x31' + chr(0b110111) + chr(49), 54758 - 54750), ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(0b100001 + 0o116) + chr(1792 - 1742) + chr(49) + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b100 + 0o54) + '\157' + '\063' + chr(738 - 684) + chr(197 - 149), 14846 - 14838), ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(0b100000 + 0o117) + chr(1706 - 1655) + '\067' + chr(789 - 735), ord("\x08")), ehT0Px3KOsy9(chr(0b1111 + 0o41) + '\x6f' + chr(0b10011 + 0o37) + chr(55) + chr(0b110010), 8), ehT0Px3KOsy9(chr(0b100100 + 0o14) + '\157' + '\061' + chr(661 - 608) + chr(0b11 + 0o57), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(2248 - 2198) + chr(0b11111 + 0o23) + '\x37', 0b1000), ehT0Px3KOsy9(chr(0b101101 + 0o3) + '\x6f' + '\x31' + chr(739 - 691) + chr(0b110 + 0o56), 0o10), ehT0Px3KOsy9(chr(2118 - 2070) + chr(0b1101111) + '\062' + chr(0b11001 + 0o30) + '\062', 8), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(0b1101111) + chr(1386 - 1335) + chr(50) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(1968 - 1920) + '\x6f' + chr(0b110001) + '\x34', 55580 - 55572), ehT0Px3KOsy9(chr(297 - 249) + '\157' + chr(0b110010) + chr(0b110110) + chr(358 - 305), 0o10), ehT0Px3KOsy9('\060' + chr(111) + '\062' + chr(54) + chr(0b110001), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b1010 + 0o46) + chr(111) + chr(53) + chr(0b110000), 16956 - 16948)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x94'), chr(0b1010100 + 0o20) + chr(0b1100101) + chr(99) + chr(0b1010011 + 0o34) + chr(0b1100010 + 0o2) + '\x65')('\165' + chr(0b1110100) + '\146' + chr(280 - 235) + '\x38') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def Z01zxSkrfh7f(E1c_5v_Zd9l8, xXh07emJaiUD, ag0mwEgWzjYv, holLFgwB7vsP):
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe8\x06\x02\x0e\xe0$D\xa1'), '\x64' + chr(5716 - 5615) + chr(6793 - 6694) + '\157' + chr(0b1100100) + chr(0b1010011 + 0o22))('\x75' + chr(0b1010111 + 0o35) + chr(102) + chr(0b0 + 0o55) + '\070'))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xf6\x02\x08\x02\xf6\x1fJ\xbf\xc2'), '\144' + chr(8336 - 8235) + '\x63' + '\157' + chr(3730 - 3630) + chr(0b1100101))(chr(117) + chr(116) + chr(0b10100 + 0o122) + chr(0b101101) + chr(3050 - 2994)))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe\x06\x1f\x14\xe1'), chr(9193 - 9093) + '\x65' + chr(5253 - 5154) + '\x6f' + '\x64' + '\x65')(chr(117) + chr(0b1110100) + '\x66' + chr(0b101101) + '\x38'))(xXh07emJaiUD), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xe8\x06\x1d\x12'), '\x64' + chr(0b1100101) + '\143' + chr(4150 - 4039) + chr(0b1000011 + 0o41) + chr(0b1100101))(chr(0b1001011 + 0o52) + chr(116) + chr(0b1100110) + chr(45) + chr(2737 - 2681)))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe\x11\x1e\x17\xeb$Q'), chr(0b1000110 + 0o36) + chr(8957 - 8856) + chr(6978 - 6879) + '\x6f' + chr(7416 - 7316) + chr(7951 - 7850))(chr(0b1110101) + chr(0b1110000 + 0o4) + '\x66' + chr(0b1100 + 0o41) + chr(0b110 + 0o62)))(rate=ag0mwEgWzjYv, mode=holLFgwB7vsP), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe\x06\x1f\x14\xe1'), chr(100) + chr(101) + '\x63' + '\157' + chr(0b1100100) + chr(0b1100101))('\165' + chr(12730 - 12614) + chr(0b1100110) + chr(789 - 744) + '\x38'))(E1c_5v_Zd9l8), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xfe\x11\x1e\x17\xeb$Q'), '\x64' + '\x65' + '\143' + chr(0b1101111) + '\144' + chr(0b1100101))('\x75' + '\164' + '\x66' + chr(45) + chr(56)))(rate=ag0mwEgWzjYv, mode=holLFgwB7vsP))
|
tensorflow/tensor2tensor
|
tensor2tensor/trax/models/transformer.py
|
EncoderLayer
|
def EncoderLayer(feature_depth,
feedforward_depth,
num_heads,
dropout,
mode):
"""Transformer encoder layer.
The input to the encoder is a pair (embedded source, mask) where
the mask is created from the original source to prevent attending
to the padding part of the input.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
mode: str: 'train' or 'eval'
Returns:
the layer, returning a pair (actiavtions, mask).
"""
# The encoder block expects (activation, mask) as input and returns
# the new activations only, we add the mask back to output next.
encoder_block = layers.Serial(
layers.Residual( # Attention block here.
layers.Parallel(layers.LayerNorm(), layers.Identity()),
layers.MultiHeadedAttention(feature_depth, num_heads=num_heads,
dropout=dropout, mode=mode),
layers.Dropout(rate=dropout, mode=mode),
shortcut=layers.FirstBranch()
),
ResidualFeedForward(feature_depth, feedforward_depth, dropout, mode=mode)
)
# Now we add the mask back.
return layers.Serial(
layers.Reorder(output=((0, 1), 1)), # (x, mask) --> ((x, mask), mask)
layers.Parallel(encoder_block, layers.Identity())
)
|
python
|
def EncoderLayer(feature_depth,
feedforward_depth,
num_heads,
dropout,
mode):
"""Transformer encoder layer.
The input to the encoder is a pair (embedded source, mask) where
the mask is created from the original source to prevent attending
to the padding part of the input.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
mode: str: 'train' or 'eval'
Returns:
the layer, returning a pair (actiavtions, mask).
"""
# The encoder block expects (activation, mask) as input and returns
# the new activations only, we add the mask back to output next.
encoder_block = layers.Serial(
layers.Residual( # Attention block here.
layers.Parallel(layers.LayerNorm(), layers.Identity()),
layers.MultiHeadedAttention(feature_depth, num_heads=num_heads,
dropout=dropout, mode=mode),
layers.Dropout(rate=dropout, mode=mode),
shortcut=layers.FirstBranch()
),
ResidualFeedForward(feature_depth, feedforward_depth, dropout, mode=mode)
)
# Now we add the mask back.
return layers.Serial(
layers.Reorder(output=((0, 1), 1)), # (x, mask) --> ((x, mask), mask)
layers.Parallel(encoder_block, layers.Identity())
)
|
[
"def",
"EncoderLayer",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"num_heads",
",",
"dropout",
",",
"mode",
")",
":",
"# The encoder block expects (activation, mask) as input and returns",
"# the new activations only, we add the mask back to output next.",
"encoder_block",
"=",
"layers",
".",
"Serial",
"(",
"layers",
".",
"Residual",
"(",
"# Attention block here.",
"layers",
".",
"Parallel",
"(",
"layers",
".",
"LayerNorm",
"(",
")",
",",
"layers",
".",
"Identity",
"(",
")",
")",
",",
"layers",
".",
"MultiHeadedAttention",
"(",
"feature_depth",
",",
"num_heads",
"=",
"num_heads",
",",
"dropout",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
",",
"layers",
".",
"Dropout",
"(",
"rate",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
",",
"shortcut",
"=",
"layers",
".",
"FirstBranch",
"(",
")",
")",
",",
"ResidualFeedForward",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
"# Now we add the mask back.",
"return",
"layers",
".",
"Serial",
"(",
"layers",
".",
"Reorder",
"(",
"output",
"=",
"(",
"(",
"0",
",",
"1",
")",
",",
"1",
")",
")",
",",
"# (x, mask) --> ((x, mask), mask)",
"layers",
".",
"Parallel",
"(",
"encoder_block",
",",
"layers",
".",
"Identity",
"(",
")",
")",
")"
] |
Transformer encoder layer.
The input to the encoder is a pair (embedded source, mask) where
the mask is created from the original source to prevent attending
to the padding part of the input.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
mode: str: 'train' or 'eval'
Returns:
the layer, returning a pair (actiavtions, mask).
|
[
"Transformer",
"encoder",
"layer",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/trax/models/transformer.py#L39-L76
|
train
|
Transformer encoder layer.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(3997 - 3886) + chr(0b100010 + 0o17) + '\067' + '\x32', 49521 - 49513), ehT0Px3KOsy9('\060' + chr(9375 - 9264) + '\x33' + chr(2332 - 2277) + chr(285 - 235), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + chr(2303 - 2255) + chr(1827 - 1776), 53351 - 53343), ehT0Px3KOsy9(chr(0b101111 + 0o1) + '\x6f' + chr(0b110001) + '\x37' + chr(54), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x31' + chr(1541 - 1489) + chr(0b101111 + 0o5), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b1011 + 0o46) + chr(0b110001) + chr(55), 24017 - 24009), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(50) + chr(0b100100 + 0o22) + chr(0b101110 + 0o6), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(50) + chr(1031 - 978), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\061' + chr(0b10001 + 0o37) + '\060', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b1011 + 0o50) + chr(699 - 645) + '\x31', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(49) + chr(48) + chr(51), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(5358 - 5247) + chr(0b10100 + 0o35) + '\x36' + '\x36', 0o10), ehT0Px3KOsy9(chr(720 - 672) + chr(0b1101111) + chr(0b10000 + 0o43) + '\065' + '\x31', 58196 - 58188), ehT0Px3KOsy9('\060' + chr(7875 - 7764) + chr(50) + chr(691 - 643), 60131 - 60123), ehT0Px3KOsy9(chr(316 - 268) + '\157' + chr(0b110010) + '\063' + '\x37', 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x32' + chr(0b101101 + 0o6) + chr(49), 0b1000), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(0b1100100 + 0o13) + chr(0b110010) + '\x35' + '\066', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + '\062' + chr(53), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110001) + '\x30', 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(2061 - 2010) + chr(0b110000 + 0o6) + '\x37', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101110 + 0o1) + '\x33' + '\x33' + chr(0b101101 + 0o11), 64302 - 64294), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(206 - 156) + chr(52) + '\063', 0o10), ehT0Px3KOsy9(chr(0b1100 + 0o44) + chr(1604 - 1493) + '\061' + chr(0b110101) + chr(0b1001 + 0o51), ord("\x08")), ehT0Px3KOsy9(chr(1568 - 1520) + '\157' + chr(726 - 675) + chr(0b110011), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x35' + chr(0b110101 + 0o0), 0b1000), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(111) + chr(0b100001 + 0o22) + chr(1482 - 1431) + chr(0b110010 + 0o2), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + chr(53) + '\061', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1001111 + 0o40) + '\061', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(49) + chr(0b110010) + chr(51), 11760 - 11752), ehT0Px3KOsy9(chr(48) + chr(111) + '\062' + chr(48) + chr(2048 - 1998), ord("\x08")), ehT0Px3KOsy9(chr(603 - 555) + chr(0b1101000 + 0o7) + chr(49) + '\063' + chr(0b10011 + 0o40), 32960 - 32952), ehT0Px3KOsy9(chr(1585 - 1537) + '\157' + '\x36' + chr(1054 - 1004), 0b1000), ehT0Px3KOsy9(chr(2290 - 2242) + chr(7957 - 7846) + '\x32' + '\x35' + chr(0b1011 + 0o47), 0b1000), ehT0Px3KOsy9(chr(459 - 411) + chr(111) + '\x31' + '\067', 0b1000), ehT0Px3KOsy9('\060' + chr(0b1100101 + 0o12) + '\063' + chr(0b1011 + 0o47) + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(411 - 360) + chr(2092 - 2043) + chr(381 - 328), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(49) + chr(2474 - 2419) + chr(0b101 + 0o62), 14839 - 14831), ehT0Px3KOsy9('\060' + '\157' + '\x32' + '\061' + '\x37', 0o10), ehT0Px3KOsy9(chr(1088 - 1040) + chr(0b110111 + 0o70) + '\x31' + '\x36', 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33' + chr(0b110110) + chr(51), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(53) + '\x30', 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'T'), chr(100) + chr(0b1100101 + 0o0) + chr(0b1100011) + '\157' + chr(0b1011010 + 0o12) + chr(0b11001 + 0o114))('\x75' + chr(12597 - 12481) + chr(0b1011011 + 0o13) + chr(0b101101) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def JLKzX_kJQyck(E1c_5v_Zd9l8, xXh07emJaiUD, vRVqPOZ1hUG7, ag0mwEgWzjYv, holLFgwB7vsP):
h5UHUNC6umBU = sGi5Aql23May.Serial(sGi5Aql23May.Residual(sGi5Aql23May.Parallel(sGi5Aql23May.LayerNorm(), sGi5Aql23May.Identity()), sGi5Aql23May.MultiHeadedAttention(E1c_5v_Zd9l8, num_heads=vRVqPOZ1hUG7, dropout=ag0mwEgWzjYv, mode=holLFgwB7vsP), sGi5Aql23May.Dropout(rate=ag0mwEgWzjYv, mode=holLFgwB7vsP), shortcut=sGi5Aql23May.FirstBranch()), Z01zxSkrfh7f(E1c_5v_Zd9l8, xXh07emJaiUD, ag0mwEgWzjYv, mode=holLFgwB7vsP))
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b')\x9f\xb1\x17\x05#'), chr(0b1001 + 0o133) + '\x65' + chr(0b1100011) + '\x6f' + chr(0b1000000 + 0o44) + chr(0b1001010 + 0o33))('\165' + chr(0b1110100) + chr(0b1000101 + 0o41) + '\055' + chr(0b111000)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'(\x9f\xac\x0c\x00*\xc6'), '\x64' + chr(0b1100101) + chr(99) + chr(5648 - 5537) + chr(2817 - 2717) + '\x65')(chr(117) + chr(116) + '\x66' + chr(0b101101) + chr(0b100111 + 0o21)))(output=((ehT0Px3KOsy9(chr(0b110000) + chr(7254 - 7143) + '\060', 21082 - 21074), ehT0Px3KOsy9('\x30' + chr(0b1000001 + 0o56) + chr(1424 - 1375), 8)), ehT0Px3KOsy9('\060' + chr(111) + chr(2120 - 2071), 8))), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'*\x9b\xb1\x1f\x08#\xd1y'), chr(4707 - 4607) + chr(0b10101 + 0o120) + chr(3948 - 3849) + '\x6f' + chr(0b1100100) + chr(0b1100101))(chr(0b1110101) + chr(0b1110100) + chr(2583 - 2481) + chr(0b101101) + chr(0b111000)))(h5UHUNC6umBU, xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'3\x9e\xa6\x10\x10&\xc0l'), '\x64' + chr(0b1110 + 0o127) + chr(0b1000 + 0o133) + chr(0b1011100 + 0o23) + chr(100) + '\145')(chr(0b10100 + 0o141) + '\164' + chr(102) + '\055' + chr(0b111000)))()))
|
tensorflow/tensor2tensor
|
tensor2tensor/trax/models/transformer.py
|
TransformerEncoder
|
def TransformerEncoder(vocab_size,
num_classes=10,
feature_depth=512,
feedforward_depth=2048,
num_layers=6,
num_heads=8,
dropout=0.1,
max_len=2048,
mode='train'):
"""Transformer encoder.
Args:
vocab_size: int: vocab size
num_classes: how many classes on output
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the Transformer encoder layer.
"""
input_embedding = layers.Serial(
layers.Embedding(feature_depth, vocab_size),
layers.Dropout(rate=dropout, mode=mode),
layers.PositionalEncoding(max_len=max_len)
)
return layers.Serial(
layers.Branch(), # Branch input to create embedding and mask.
layers.Parallel(input_embedding, layers.PaddingMask()),
layers.Serial(*[EncoderLayer(feature_depth, feedforward_depth, num_heads,
dropout, mode)
for _ in range(num_layers)]),
layers.FirstBranch(), # Drop the mask.
layers.LayerNorm(),
layers.Mean(axis=1), # Average on length.
layers.Dense(num_classes),
layers.LogSoftmax()
)
|
python
|
def TransformerEncoder(vocab_size,
num_classes=10,
feature_depth=512,
feedforward_depth=2048,
num_layers=6,
num_heads=8,
dropout=0.1,
max_len=2048,
mode='train'):
"""Transformer encoder.
Args:
vocab_size: int: vocab size
num_classes: how many classes on output
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the Transformer encoder layer.
"""
input_embedding = layers.Serial(
layers.Embedding(feature_depth, vocab_size),
layers.Dropout(rate=dropout, mode=mode),
layers.PositionalEncoding(max_len=max_len)
)
return layers.Serial(
layers.Branch(), # Branch input to create embedding and mask.
layers.Parallel(input_embedding, layers.PaddingMask()),
layers.Serial(*[EncoderLayer(feature_depth, feedforward_depth, num_heads,
dropout, mode)
for _ in range(num_layers)]),
layers.FirstBranch(), # Drop the mask.
layers.LayerNorm(),
layers.Mean(axis=1), # Average on length.
layers.Dense(num_classes),
layers.LogSoftmax()
)
|
[
"def",
"TransformerEncoder",
"(",
"vocab_size",
",",
"num_classes",
"=",
"10",
",",
"feature_depth",
"=",
"512",
",",
"feedforward_depth",
"=",
"2048",
",",
"num_layers",
"=",
"6",
",",
"num_heads",
"=",
"8",
",",
"dropout",
"=",
"0.1",
",",
"max_len",
"=",
"2048",
",",
"mode",
"=",
"'train'",
")",
":",
"input_embedding",
"=",
"layers",
".",
"Serial",
"(",
"layers",
".",
"Embedding",
"(",
"feature_depth",
",",
"vocab_size",
")",
",",
"layers",
".",
"Dropout",
"(",
"rate",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
",",
"layers",
".",
"PositionalEncoding",
"(",
"max_len",
"=",
"max_len",
")",
")",
"return",
"layers",
".",
"Serial",
"(",
"layers",
".",
"Branch",
"(",
")",
",",
"# Branch input to create embedding and mask.",
"layers",
".",
"Parallel",
"(",
"input_embedding",
",",
"layers",
".",
"PaddingMask",
"(",
")",
")",
",",
"layers",
".",
"Serial",
"(",
"*",
"[",
"EncoderLayer",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"num_heads",
",",
"dropout",
",",
"mode",
")",
"for",
"_",
"in",
"range",
"(",
"num_layers",
")",
"]",
")",
",",
"layers",
".",
"FirstBranch",
"(",
")",
",",
"# Drop the mask.",
"layers",
".",
"LayerNorm",
"(",
")",
",",
"layers",
".",
"Mean",
"(",
"axis",
"=",
"1",
")",
",",
"# Average on length.",
"layers",
".",
"Dense",
"(",
"num_classes",
")",
",",
"layers",
".",
"LogSoftmax",
"(",
")",
")"
] |
Transformer encoder.
Args:
vocab_size: int: vocab size
num_classes: how many classes on output
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the Transformer encoder layer.
|
[
"Transformer",
"encoder",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/trax/models/transformer.py#L79-L120
|
train
|
Transformer encoder.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x33' + chr(0b110110) + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110101) + chr(48), 0o10), ehT0Px3KOsy9('\x30' + '\157' + '\x31' + chr(0b110000 + 0o7) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(0b10010 + 0o36) + chr(4746 - 4635) + '\063' + chr(0b11011 + 0o34), ord("\x08")), ehT0Px3KOsy9(chr(0b10100 + 0o34) + chr(111) + chr(49) + chr(0b110001), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(0b100011 + 0o23) + '\060', 0b1000), ehT0Px3KOsy9('\x30' + '\157' + chr(759 - 708) + chr(1920 - 1870) + chr(49), 0o10), ehT0Px3KOsy9(chr(2002 - 1954) + chr(0b110000 + 0o77) + chr(739 - 690) + chr(0b110001) + chr(50), 0b1000), ehT0Px3KOsy9(chr(1049 - 1001) + chr(0b1101111) + chr(317 - 267) + chr(53) + chr(0b100100 + 0o21), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(50) + chr(1190 - 1138) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(0b1110 + 0o42) + chr(6235 - 6124) + '\062' + chr(51) + chr(0b10111 + 0o34), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\061' + '\066' + chr(694 - 643), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(50) + '\x30' + chr(1859 - 1811), 55542 - 55534), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110011) + chr(54) + chr(0b110010), 16970 - 16962), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(111) + chr(1246 - 1195) + chr(0b110000) + chr(84 - 30), 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + '\063' + chr(2450 - 2399) + '\062', 59973 - 59965), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100010 + 0o15) + chr(1899 - 1849) + chr(48) + chr(0b10000 + 0o42), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b10011 + 0o40) + chr(0b110001 + 0o4) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(1093 - 982) + '\x33' + chr(0b1000 + 0o50) + chr(0b110000 + 0o7), ord("\x08")), ehT0Px3KOsy9(chr(1980 - 1932) + chr(4560 - 4449) + '\x31' + chr(0b0 + 0o61) + chr(48), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\063' + '\060' + chr(1262 - 1212), 60461 - 60453), ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(9509 - 9398) + chr(0b110111) + chr(0b101011 + 0o12), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\x33' + '\x32' + chr(0b110101), 26063 - 26055), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100011 + 0o14) + '\x31' + '\x32' + '\066', 12014 - 12006), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(0b1101111) + chr(0b110011) + '\064' + chr(53), 0o10), ehT0Px3KOsy9(chr(1935 - 1887) + '\x6f' + chr(1366 - 1316) + '\x34' + chr(1125 - 1072), 0b1000), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(111) + chr(1915 - 1864) + chr(933 - 885) + chr(0b11110 + 0o22), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1100100 + 0o13) + chr(54) + '\065', 0o10), ehT0Px3KOsy9(chr(0b10010 + 0o36) + '\x6f' + chr(0b1100 + 0o46) + '\064' + chr(0b101010 + 0o12), 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b110011) + chr(50), 41047 - 41039), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x33' + chr(53) + chr(0b11011 + 0o34), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + '\x35' + chr(1042 - 991), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + chr(747 - 696) + chr(0b110110) + chr(0b110001), 58261 - 58253), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\062' + '\061' + '\x30', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x35' + chr(2050 - 1999), ord("\x08")), ehT0Px3KOsy9(chr(69 - 21) + '\x6f' + chr(1791 - 1742) + chr(51) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(471 - 423) + chr(0b1101111) + chr(1515 - 1463) + '\060', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(6050 - 5939) + '\x33' + chr(54) + chr(2258 - 2205), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + '\062' + '\x34' + chr(0b110101), 8), ehT0Px3KOsy9(chr(0b11111 + 0o21) + '\x6f' + chr(51) + '\062' + '\060', 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(0b11011 + 0o124) + chr(0b110101) + chr(48), 8)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xdd'), '\144' + '\145' + chr(1084 - 985) + '\157' + '\144' + chr(0b11011 + 0o112))('\x75' + chr(0b1100 + 0o150) + chr(102) + chr(510 - 465) + chr(0b100001 + 0o27)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def zz1tCO9Y0Jv5(CeyMIoSyrpkQ, i6loyAgxUM2t=ehT0Px3KOsy9(chr(48) + chr(111) + '\x31' + '\x32', 57385 - 57377), E1c_5v_Zd9l8=ehT0Px3KOsy9('\060' + '\x6f' + chr(49) + chr(48) + chr(1267 - 1219) + '\x30', 4320 - 4312), xXh07emJaiUD=ehT0Px3KOsy9(chr(918 - 870) + '\157' + chr(0b110100) + chr(48) + chr(0b110000) + chr(0b110000), 52024 - 52016), uftkTXJyNORO=ehT0Px3KOsy9(chr(1327 - 1279) + '\x6f' + chr(0b110101 + 0o1), 0b1000), vRVqPOZ1hUG7=ehT0Px3KOsy9(chr(0b100100 + 0o14) + '\157' + chr(0b110001) + '\060', 23036 - 23028), ag0mwEgWzjYv=0.1, qbKO12mgagKE=ehT0Px3KOsy9(chr(539 - 491) + chr(0b110100 + 0o73) + '\x34' + chr(1023 - 975) + '\060' + chr(0b110000), 8), holLFgwB7vsP=xafqLlk3kkUe(SXOLrMavuUCe(b'\x87Z\x1bvg'), chr(100) + chr(101) + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + '\x65')(chr(117) + chr(0b1101111 + 0o5) + chr(0b1100110) + chr(1940 - 1895) + '\070')):
dE6NM1fGXzm0 = sGi5Aql23May.Serial(sGi5Aql23May.Embedding(E1c_5v_Zd9l8, CeyMIoSyrpkQ), sGi5Aql23May.Dropout(rate=ag0mwEgWzjYv, mode=holLFgwB7vsP), sGi5Aql23May.PositionalEncoding(max_len=qbKO12mgagKE))
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xa0M\x08vh\\'), '\x64' + chr(6179 - 6078) + chr(0b1000111 + 0o34) + '\x6f' + chr(2441 - 2341) + chr(0b100100 + 0o101))(chr(0b1110100 + 0o1) + chr(4950 - 4834) + chr(0b1100110) + chr(381 - 336) + '\x38'))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb1Z\x1bqjX'), chr(0b10011 + 0o121) + chr(0b1100101) + chr(0b11000 + 0o113) + '\157' + '\144' + '\x65')(chr(117) + chr(0b1110100) + chr(0b101001 + 0o75) + chr(400 - 355) + chr(56)))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xa3I\x08~e\\\xb4&'), chr(100) + '\145' + chr(0b1100011) + chr(111) + '\x64' + chr(0b111 + 0o136))(chr(13098 - 12981) + chr(0b1110100) + chr(0b1100110) + '\x2d' + chr(147 - 91)))(dE6NM1fGXzm0, xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xa3I\x1e{`^\xb6\x07\xcc\xd7\xf0'), chr(100) + chr(0b1100101) + chr(0b1100011) + '\x6f' + chr(0b1011100 + 0o10) + chr(0b1100101))('\x75' + chr(10933 - 10817) + '\146' + chr(0b101101) + '\x38'))()), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xa0M\x08vh\\'), chr(0b111 + 0o135) + chr(0b1100101) + chr(1153 - 1054) + chr(0b1001110 + 0o41) + chr(100) + chr(101))('\x75' + chr(12606 - 12490) + chr(0b1011101 + 0o11) + chr(0b101100 + 0o1) + '\070'))(*[JLKzX_kJQyck(E1c_5v_Zd9l8, xXh07emJaiUD, vRVqPOZ1hUG7, ag0mwEgWzjYv, holLFgwB7vsP) for VNGQdHSFPrso in vQr8gNKaIaWE(uftkTXJyNORO)]), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb5A\x08l}r\xa3+\xc3\xc7\xf3'), chr(100) + chr(0b1100101) + '\143' + '\157' + '\x64' + '\145')(chr(0b1110101) + chr(116) + chr(1076 - 974) + chr(45) + chr(2272 - 2216)))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xbfI\x03z{~\xbe8\xc0'), chr(100) + chr(0b1010111 + 0o16) + chr(0b1100011) + chr(111) + chr(0b1100100) + '\145')('\165' + '\164' + '\x66' + chr(653 - 608) + '\x38'))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xbeM\x1bq'), '\x64' + chr(101) + '\143' + chr(1076 - 965) + chr(100) + chr(0b1100011 + 0o2))('\165' + '\x74' + chr(0b1001011 + 0o33) + chr(0b101101) + '\070'))(axis=ehT0Px3KOsy9('\060' + chr(111) + chr(49), 43770 - 43762)), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7M\x14ll'), '\x64' + chr(0b1001100 + 0o31) + '\x63' + '\x6f' + chr(0b1100100) + chr(0b1100 + 0o131))(chr(6264 - 6147) + chr(116) + chr(0b110 + 0o140) + chr(0b101101) + '\x38'))(i6loyAgxUM2t), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b"\xbfG\x1dLfV\xa5'\xcc\xdc"), '\144' + '\x65' + chr(5929 - 5830) + chr(0b1101111) + chr(100) + '\x65')('\165' + chr(0b1110 + 0o146) + chr(0b1000001 + 0o45) + chr(0b101101) + '\070'))())
|
tensorflow/tensor2tensor
|
tensor2tensor/trax/models/transformer.py
|
DecoderLayer
|
def DecoderLayer(feature_depth,
feedforward_depth,
num_heads,
dropout,
mode):
"""Transformer decoder layer.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
mode: str: 'train' or 'eval'
Returns:
the layer.
"""
return layers.Serial(
layers.Residual( # Self-attention block.
layers.LayerNorm(),
layers.Branch(),
layers.Parallel(layers.Identity(), # activation for (q, k, v)
layers.CausalMask(axis=-2)), # attention mask
layers.MultiHeadedAttention(feature_depth, num_heads=num_heads,
dropout=dropout, mode=mode),
layers.Dropout(rate=dropout, mode=mode)
),
ResidualFeedForward(feature_depth, feedforward_depth, dropout, mode=mode)
)
|
python
|
def DecoderLayer(feature_depth,
feedforward_depth,
num_heads,
dropout,
mode):
"""Transformer decoder layer.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
mode: str: 'train' or 'eval'
Returns:
the layer.
"""
return layers.Serial(
layers.Residual( # Self-attention block.
layers.LayerNorm(),
layers.Branch(),
layers.Parallel(layers.Identity(), # activation for (q, k, v)
layers.CausalMask(axis=-2)), # attention mask
layers.MultiHeadedAttention(feature_depth, num_heads=num_heads,
dropout=dropout, mode=mode),
layers.Dropout(rate=dropout, mode=mode)
),
ResidualFeedForward(feature_depth, feedforward_depth, dropout, mode=mode)
)
|
[
"def",
"DecoderLayer",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"num_heads",
",",
"dropout",
",",
"mode",
")",
":",
"return",
"layers",
".",
"Serial",
"(",
"layers",
".",
"Residual",
"(",
"# Self-attention block.",
"layers",
".",
"LayerNorm",
"(",
")",
",",
"layers",
".",
"Branch",
"(",
")",
",",
"layers",
".",
"Parallel",
"(",
"layers",
".",
"Identity",
"(",
")",
",",
"# activation for (q, k, v)",
"layers",
".",
"CausalMask",
"(",
"axis",
"=",
"-",
"2",
")",
")",
",",
"# attention mask",
"layers",
".",
"MultiHeadedAttention",
"(",
"feature_depth",
",",
"num_heads",
"=",
"num_heads",
",",
"dropout",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
",",
"layers",
".",
"Dropout",
"(",
"rate",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
",",
"ResidualFeedForward",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"dropout",
",",
"mode",
"=",
"mode",
")",
")"
] |
Transformer decoder layer.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
mode: str: 'train' or 'eval'
Returns:
the layer.
|
[
"Transformer",
"decoder",
"layer",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/trax/models/transformer.py#L123-L151
|
train
|
Transformer decoder layer.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + '\157' + '\063' + chr(49) + chr(53), 64043 - 64035), ehT0Px3KOsy9(chr(48) + chr(111) + chr(64 - 15) + chr(0b110010) + chr(0b1 + 0o64), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\062' + chr(2185 - 2136) + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(48) + chr(4541 - 4430) + chr(52) + chr(860 - 810), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\063' + '\061' + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(51) + chr(0b10000 + 0o42) + chr(49), 38956 - 38948), ehT0Px3KOsy9(chr(0b1011 + 0o45) + chr(10124 - 10013) + chr(0b110010) + '\x37', 0b1000), ehT0Px3KOsy9(chr(1794 - 1746) + chr(0b1101111) + chr(0b101010 + 0o7) + chr(53) + chr(0b10001 + 0o37), 0o10), ehT0Px3KOsy9(chr(48) + chr(8968 - 8857) + chr(0b110001) + chr(0b110111), 15608 - 15600), ehT0Px3KOsy9(chr(1970 - 1922) + chr(0b1101111) + chr(1194 - 1144) + chr(0b100101 + 0o17) + chr(0b100101 + 0o22), 0b1000), ehT0Px3KOsy9(chr(1213 - 1165) + '\x6f' + '\066' + '\x30', 0b1000), ehT0Px3KOsy9(chr(48) + chr(3445 - 3334) + chr(1668 - 1619) + chr(954 - 903) + chr(0b10001 + 0o42), ord("\x08")), ehT0Px3KOsy9(chr(349 - 301) + '\157' + chr(2185 - 2135) + '\066' + chr(2720 - 2665), 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(50) + '\061', 38255 - 38247), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(10225 - 10114) + '\063' + chr(2237 - 2189) + '\x32', 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + '\061' + chr(1019 - 967) + chr(48), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(50) + chr(0b10011 + 0o44) + '\067', 0o10), ehT0Px3KOsy9(chr(1788 - 1740) + '\157' + '\x37' + chr(0b11101 + 0o26), 0o10), ehT0Px3KOsy9(chr(0b111 + 0o51) + '\157' + chr(0b11000 + 0o34) + chr(0b110010), 8), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110001) + chr(53) + chr(1249 - 1195), 64468 - 64460), ehT0Px3KOsy9(chr(1779 - 1731) + '\157' + chr(0b1 + 0o63) + chr(285 - 235), 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110 + 0o54) + '\x33' + '\x37', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b10101 + 0o36) + chr(0b110001) + '\065', 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\x31' + chr(55) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110001) + '\x34' + chr(0b110010 + 0o4), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + '\062' + '\062' + chr(0b11101 + 0o27), 0o10), ehT0Px3KOsy9('\060' + chr(2468 - 2357) + chr(0b110011) + chr(0b110000) + chr(1645 - 1595), 8), ehT0Px3KOsy9(chr(2195 - 2147) + chr(0b1101111) + '\x31' + '\065' + chr(50), 0b1000), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(0b1101111) + '\x33' + '\x35' + '\064', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(51) + chr(2140 - 2091) + '\061', 8), ehT0Px3KOsy9('\x30' + '\x6f' + '\x33' + '\x30' + chr(54), 0o10), ehT0Px3KOsy9(chr(993 - 945) + chr(111) + '\x32' + chr(0b110100) + chr(0b10101 + 0o33), 13753 - 13745), ehT0Px3KOsy9(chr(278 - 230) + '\157' + '\061' + chr(0b110 + 0o60) + chr(1252 - 1204), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(49) + '\x33' + chr(988 - 939), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\065', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x32' + chr(0b101000 + 0o17) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b101001 + 0o7) + '\x6f' + chr(0b100010 + 0o20) + chr(1928 - 1878) + chr(1903 - 1853), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(1771 - 1721) + chr(0b101111 + 0o10) + '\065', 8), ehT0Px3KOsy9(chr(0b1000 + 0o50) + '\157' + chr(0b110110) + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110110) + chr(51), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110101) + chr(952 - 904), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xe3'), '\x64' + '\x65' + chr(0b1010100 + 0o17) + chr(0b111011 + 0o64) + '\x64' + chr(0b11 + 0o142))('\x75' + '\164' + chr(0b1100110) + chr(45) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def vK9Z7_1dGHpp(E1c_5v_Zd9l8, xXh07emJaiUD, vRVqPOZ1hUG7, ag0mwEgWzjYv, holLFgwB7vsP):
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x9e0\xfdK\xa8\xbd'), '\144' + chr(0b1011011 + 0o12) + chr(9170 - 9071) + '\x6f' + '\x64' + chr(0b1011111 + 0o6))(chr(0b10111 + 0o136) + chr(0b11000 + 0o134) + '\x66' + chr(0b11100 + 0o21) + chr(1356 - 1300)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x9f0\xfcK\xad\xa4v\x04'), chr(0b11111 + 0o105) + '\x65' + '\x63' + chr(4569 - 4458) + chr(100) + '\x65')(chr(0b111000 + 0o75) + '\164' + chr(0b1100110) + chr(0b101101) + chr(56)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x814\xf6G\xbb\x9fx\x1ab'), chr(7104 - 7004) + chr(0b111100 + 0o51) + '\143' + chr(111) + chr(3105 - 3005) + chr(0b10011 + 0o122))(chr(0b1100101 + 0o20) + chr(0b1110100) + chr(1224 - 1122) + '\x2d' + chr(1686 - 1630)))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b"\x8f'\xeeL\xaa\xb9"), chr(8968 - 8868) + chr(101) + chr(0b1100011) + chr(111) + chr(100) + chr(0b1010101 + 0o20))(chr(0b1110101) + '\x74' + chr(0b1100110) + chr(45) + chr(0b110 + 0o62)))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x9d4\xfdC\xa5\xbdr\x04'), chr(565 - 465) + chr(6788 - 6687) + chr(8352 - 8253) + chr(0b1001000 + 0o47) + chr(100) + '\145')('\x75' + '\164' + chr(0b111011 + 0o53) + chr(0b101101) + chr(0b111000)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x841\xeaL\xbd\xb8c\x11'), chr(100) + '\145' + chr(0b1100011) + chr(4853 - 4742) + chr(0b1100100) + chr(101))('\165' + chr(116) + chr(0b10 + 0o144) + '\055' + '\070'))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x8e4\xfaQ\xa8\xbdZ\t|\xcc'), chr(0b1100100) + chr(0b1001111 + 0o26) + chr(0b101101 + 0o66) + '\157' + '\x64' + chr(0b1100101))(chr(1397 - 1280) + chr(0b101101 + 0o107) + chr(7532 - 7430) + chr(0b10110 + 0o27) + '\x38'))(axis=-ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b1011 + 0o47), 48622 - 48614))), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x80 \xe3V\xa0\x99r\tk\xc2T\x1a\xc3e\xa7\x0e\xcbRM\x00'), '\144' + chr(8320 - 8219) + '\143' + chr(0b1101111) + chr(3770 - 3670) + chr(0b1100101))(chr(117) + chr(116) + chr(420 - 318) + chr(0b101101) + chr(0b111000)))(E1c_5v_Zd9l8, num_heads=vRVqPOZ1hUG7, dropout=ag0mwEgWzjYv, mode=holLFgwB7vsP), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b"\x89'\xe0R\xa6\xa4c"), '\144' + chr(0b1100101) + '\x63' + chr(8759 - 8648) + chr(100) + chr(101))(chr(0b1110101) + chr(116) + '\x66' + chr(1016 - 971) + chr(0b111000)))(rate=ag0mwEgWzjYv, mode=holLFgwB7vsP)), Z01zxSkrfh7f(E1c_5v_Zd9l8, xXh07emJaiUD, ag0mwEgWzjYv, mode=holLFgwB7vsP))
|
tensorflow/tensor2tensor
|
tensor2tensor/trax/models/transformer.py
|
TransformerLM
|
def TransformerLM(vocab_size,
feature_depth=512,
feedforward_depth=2048,
num_layers=6,
num_heads=8,
dropout=0.1,
max_len=2048,
mode='train'):
"""Transformer language model (only uses the decoder part of Transformer).
Args:
vocab_size: int: vocab size
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the layer.
"""
return layers.Serial(
layers.ShiftRight(),
layers.Embedding(feature_depth, vocab_size),
layers.Dropout(rate=dropout, mode=mode),
layers.PositionalEncoding(max_len=max_len),
layers.Serial(*[DecoderLayer(feature_depth, feedforward_depth, num_heads,
dropout, mode)
for _ in range(num_layers)]),
layers.LayerNorm(),
layers.Dense(vocab_size),
layers.LogSoftmax()
)
|
python
|
def TransformerLM(vocab_size,
feature_depth=512,
feedforward_depth=2048,
num_layers=6,
num_heads=8,
dropout=0.1,
max_len=2048,
mode='train'):
"""Transformer language model (only uses the decoder part of Transformer).
Args:
vocab_size: int: vocab size
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the layer.
"""
return layers.Serial(
layers.ShiftRight(),
layers.Embedding(feature_depth, vocab_size),
layers.Dropout(rate=dropout, mode=mode),
layers.PositionalEncoding(max_len=max_len),
layers.Serial(*[DecoderLayer(feature_depth, feedforward_depth, num_heads,
dropout, mode)
for _ in range(num_layers)]),
layers.LayerNorm(),
layers.Dense(vocab_size),
layers.LogSoftmax()
)
|
[
"def",
"TransformerLM",
"(",
"vocab_size",
",",
"feature_depth",
"=",
"512",
",",
"feedforward_depth",
"=",
"2048",
",",
"num_layers",
"=",
"6",
",",
"num_heads",
"=",
"8",
",",
"dropout",
"=",
"0.1",
",",
"max_len",
"=",
"2048",
",",
"mode",
"=",
"'train'",
")",
":",
"return",
"layers",
".",
"Serial",
"(",
"layers",
".",
"ShiftRight",
"(",
")",
",",
"layers",
".",
"Embedding",
"(",
"feature_depth",
",",
"vocab_size",
")",
",",
"layers",
".",
"Dropout",
"(",
"rate",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
",",
"layers",
".",
"PositionalEncoding",
"(",
"max_len",
"=",
"max_len",
")",
",",
"layers",
".",
"Serial",
"(",
"*",
"[",
"DecoderLayer",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"num_heads",
",",
"dropout",
",",
"mode",
")",
"for",
"_",
"in",
"range",
"(",
"num_layers",
")",
"]",
")",
",",
"layers",
".",
"LayerNorm",
"(",
")",
",",
"layers",
".",
"Dense",
"(",
"vocab_size",
")",
",",
"layers",
".",
"LogSoftmax",
"(",
")",
")"
] |
Transformer language model (only uses the decoder part of Transformer).
Args:
vocab_size: int: vocab size
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the layer.
|
[
"Transformer",
"language",
"model",
"(",
"only",
"uses",
"the",
"decoder",
"part",
"of",
"Transformer",
")",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/trax/models/transformer.py#L154-L188
|
train
|
Transformer language model.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + '\x6f' + '\065' + '\063', 0b1000), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(10121 - 10010) + chr(0b100110 + 0o13), ord("\x08")), ehT0Px3KOsy9('\060' + chr(8802 - 8691) + chr(0b10101 + 0o35) + chr(48) + '\063', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b11 + 0o154) + '\063' + '\061' + chr(0b10110 + 0o35), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\062' + '\066', 16678 - 16670), ehT0Px3KOsy9(chr(686 - 638) + chr(0b1101111) + '\062' + chr(50) + '\061', ord("\x08")), ehT0Px3KOsy9(chr(0b101000 + 0o10) + chr(0b1101111) + chr(0b110111) + chr(0b110111), 0o10), ehT0Px3KOsy9(chr(1789 - 1741) + chr(4933 - 4822) + '\063' + '\064' + chr(0b110 + 0o60), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b110011 + 0o74) + chr(0b110001) + chr(0b110100) + chr(0b110011 + 0o4), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + chr(59 - 10) + chr(0b110010), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(3213 - 3102) + '\061' + '\064' + chr(133 - 81), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1990 - 1936) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\063' + '\063' + chr(53), 20793 - 20785), ehT0Px3KOsy9(chr(48) + chr(111) + chr(133 - 82) + chr(50) + chr(2546 - 2493), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(3810 - 3699) + chr(0b110011) + chr(48) + chr(49), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110101) + '\x31', 32687 - 32679), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\064' + chr(0b11111 + 0o21), ord("\x08")), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1101111) + chr(51) + chr(0b1101 + 0o52), 17785 - 17777), ehT0Px3KOsy9(chr(282 - 234) + '\157' + chr(0b101111 + 0o3) + chr(48) + '\x33', 8), ehT0Px3KOsy9(chr(48) + chr(681 - 570) + '\062' + chr(0b110110 + 0o0), 8), ehT0Px3KOsy9(chr(0b110000) + chr(1696 - 1585) + '\x37' + '\063', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(50) + chr(54) + chr(0b110100), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(1383 - 1334) + chr(975 - 920) + '\x31', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(1005 - 894) + '\x33' + chr(51) + chr(49), 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(50) + '\067' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(0b1010 + 0o46) + chr(111) + '\062' + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(7936 - 7825) + chr(0b100011 + 0o17) + chr(53) + chr(52), 12755 - 12747), ehT0Px3KOsy9(chr(0b110000) + chr(4681 - 4570) + '\x31' + '\x37', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(50) + '\063' + '\067', 49218 - 49210), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b101110 + 0o3) + '\x32' + chr(0b10000 + 0o47), 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + '\x35' + chr(0b110010), 0b1000), ehT0Px3KOsy9('\060' + chr(0b110001 + 0o76) + chr(0b100011 + 0o20) + chr(0b110111) + '\x32', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(51) + '\x32' + '\x31', 37612 - 37604), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + '\x33' + '\066', 0b1000), ehT0Px3KOsy9(chr(960 - 912) + chr(11347 - 11236) + '\x36' + chr(0b110100), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110010) + chr(0b101010 + 0o11) + '\060', 0o10), ehT0Px3KOsy9('\060' + '\157' + '\x31' + '\063' + chr(0b10010 + 0o43), 0o10), ehT0Px3KOsy9(chr(0b100111 + 0o11) + '\x6f' + '\063' + chr(0b11010 + 0o26) + chr(0b101110 + 0o2), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(51) + '\067' + chr(0b10000 + 0o47), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\064' + chr(0b110101), 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9('\060' + chr(0b100001 + 0o116) + chr(0b110101) + chr(2106 - 2058), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x05'), chr(100) + chr(0b1100101) + '\143' + chr(111) + chr(100) + chr(0b1101 + 0o130))('\x75' + chr(116) + '\x66' + chr(0b11010 + 0o23) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def pBixmOEKogTN(CeyMIoSyrpkQ, E1c_5v_Zd9l8=ehT0Px3KOsy9(chr(0b10010 + 0o36) + chr(0b1101111) + chr(1490 - 1441) + '\060' + chr(0b100 + 0o54) + chr(0b11111 + 0o21), 0b1000), xXh07emJaiUD=ehT0Px3KOsy9(chr(0b10 + 0o56) + chr(111) + '\x34' + chr(0b0 + 0o60) + chr(48) + chr(0b11 + 0o55), 0o10), uftkTXJyNORO=ehT0Px3KOsy9('\060' + chr(0b1101111) + '\066', 0o10), vRVqPOZ1hUG7=ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(49) + chr(48), 0o10), ag0mwEgWzjYv=0.1, qbKO12mgagKE=ehT0Px3KOsy9(chr(1121 - 1073) + chr(0b11110 + 0o121) + '\064' + '\x30' + chr(530 - 482) + '\060', 8), holLFgwB7vsP=xafqLlk3kkUe(SXOLrMavuUCe(b'_s\xe2\xa2x'), chr(0b1100100) + chr(0b1100101) + chr(99) + chr(0b1101111) + chr(8355 - 8255) + chr(0b1100101))(chr(10323 - 10206) + chr(116) + chr(0b1010000 + 0o26) + '\055' + '\070')):
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'xd\xf1\xa2w\xfd'), chr(0b1 + 0o143) + '\x65' + chr(99) + chr(5199 - 5088) + chr(0b1010001 + 0o23) + chr(101))('\165' + chr(0b100100 + 0o120) + chr(102) + chr(0b101001 + 0o4) + chr(0b111000)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'xi\xea\xadb\xc3V\x81\x0f\xbe'), chr(100) + '\145' + '\143' + chr(10948 - 10837) + chr(0b1100100) + chr(101))(chr(117) + chr(0b1110100) + '\146' + '\055' + '\070'))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'nl\xe1\xaer\xf5V\x88\x00'), chr(0b1100100) + chr(0b1100101) + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + chr(0b1101 + 0o130))(chr(0b1010110 + 0o37) + chr(4429 - 4313) + chr(0b100110 + 0o100) + '\x2d' + chr(3127 - 3071)))(E1c_5v_Zd9l8, CeyMIoSyrpkQ), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'os\xec\xbby\xe4K'), chr(0b1000101 + 0o37) + chr(2490 - 2389) + chr(3501 - 3402) + '\x6f' + '\144' + '\x65')(chr(0b1110101) + '\x74' + chr(0b11000 + 0o116) + chr(45) + chr(938 - 882)))(rate=ag0mwEgWzjYv, mode=holLFgwB7vsP), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'{n\xf0\xa2b\xf8P\x88\x06\xa6![\xbe\xba\xa5@Nx'), chr(100) + chr(1220 - 1119) + chr(99) + chr(0b1101010 + 0o5) + '\x64' + '\x65')(chr(0b10001 + 0o144) + chr(11870 - 11754) + '\146' + chr(0b101101) + chr(56)))(max_len=qbKO12mgagKE), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'xd\xf1\xa2w\xfd'), '\144' + chr(3630 - 3529) + '\x63' + chr(5843 - 5732) + chr(2462 - 2362) + chr(0b111100 + 0o51))(chr(0b1000111 + 0o56) + chr(0b110 + 0o156) + chr(0b101100 + 0o72) + '\055' + '\x38'))(*[vK9Z7_1dGHpp(E1c_5v_Zd9l8, xXh07emJaiUD, vRVqPOZ1hUG7, ag0mwEgWzjYv, holLFgwB7vsP) for VNGQdHSFPrso in vQr8gNKaIaWE(uftkTXJyNORO)]), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'g`\xfa\xaed\xdfP\x94\n'), chr(100) + chr(0b1100101) + chr(3047 - 2948) + '\x6f' + chr(0b1100100) + chr(0b1100101))('\x75' + chr(116) + chr(0b1011010 + 0o14) + '\x2d' + '\x38'))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'od\xed\xb8s'), chr(0b1100100) + chr(101) + chr(0b1011111 + 0o4) + chr(7240 - 7129) + chr(0b110010 + 0o62) + chr(0b1001110 + 0o27))(chr(0b1110101) + '\164' + chr(102) + chr(458 - 413) + chr(0b110 + 0o62)))(CeyMIoSyrpkQ), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'gn\xe4\x98y\xf7K\x8b\x06\xb2'), chr(0b1001101 + 0o27) + chr(0b1100101) + chr(0b1100011) + '\157' + '\144' + '\145')('\x75' + chr(1679 - 1563) + '\x66' + chr(0b101101) + chr(0b10010 + 0o46)))())
|
tensorflow/tensor2tensor
|
tensor2tensor/trax/models/transformer.py
|
ChunkedDecoderLayer
|
def ChunkedDecoderLayer(feature_depth,
feedforward_depth,
num_heads,
dropout,
chunk_selector,
mode):
"""Transformer decoder layer operating on chunks.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
chunk_selector: a function from chunk number to list of chunks to attend.
mode: str: 'train' or 'eval'
Returns:
the layer.
"""
return layers.Serial(
layers.Residual( # Self-attention block.
layers.Map(layers.LayerNorm()),
layers.ChunkedCausalMultiHeadedAttention(
feature_depth, num_heads=num_heads, dropout=dropout,
chunk_selector=chunk_selector, mode=mode),
layers.Map(layers.Dropout(rate=dropout, mode=mode)),
),
layers.Map(ResidualFeedForward(
feature_depth, feedforward_depth, dropout, mode=mode))
)
|
python
|
def ChunkedDecoderLayer(feature_depth,
feedforward_depth,
num_heads,
dropout,
chunk_selector,
mode):
"""Transformer decoder layer operating on chunks.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
chunk_selector: a function from chunk number to list of chunks to attend.
mode: str: 'train' or 'eval'
Returns:
the layer.
"""
return layers.Serial(
layers.Residual( # Self-attention block.
layers.Map(layers.LayerNorm()),
layers.ChunkedCausalMultiHeadedAttention(
feature_depth, num_heads=num_heads, dropout=dropout,
chunk_selector=chunk_selector, mode=mode),
layers.Map(layers.Dropout(rate=dropout, mode=mode)),
),
layers.Map(ResidualFeedForward(
feature_depth, feedforward_depth, dropout, mode=mode))
)
|
[
"def",
"ChunkedDecoderLayer",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"num_heads",
",",
"dropout",
",",
"chunk_selector",
",",
"mode",
")",
":",
"return",
"layers",
".",
"Serial",
"(",
"layers",
".",
"Residual",
"(",
"# Self-attention block.",
"layers",
".",
"Map",
"(",
"layers",
".",
"LayerNorm",
"(",
")",
")",
",",
"layers",
".",
"ChunkedCausalMultiHeadedAttention",
"(",
"feature_depth",
",",
"num_heads",
"=",
"num_heads",
",",
"dropout",
"=",
"dropout",
",",
"chunk_selector",
"=",
"chunk_selector",
",",
"mode",
"=",
"mode",
")",
",",
"layers",
".",
"Map",
"(",
"layers",
".",
"Dropout",
"(",
"rate",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
",",
")",
",",
"layers",
".",
"Map",
"(",
"ResidualFeedForward",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
")"
] |
Transformer decoder layer operating on chunks.
Args:
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
chunk_selector: a function from chunk number to list of chunks to attend.
mode: str: 'train' or 'eval'
Returns:
the layer.
|
[
"Transformer",
"decoder",
"layer",
"operating",
"on",
"chunks",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/trax/models/transformer.py#L191-L220
|
train
|
Transformer decoder layer operating on chunks.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(2042 - 1994) + chr(0b1101111) + '\x31' + chr(0b11001 + 0o30) + chr(1658 - 1610), 0o10), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(1397 - 1349) + chr(111) + chr(1368 - 1318) + '\x33', 0b1000), ehT0Px3KOsy9(chr(48) + chr(1772 - 1661) + chr(0b110011) + chr(0b1 + 0o64) + chr(1351 - 1299), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + '\x32' + chr(49), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(2526 - 2475) + chr(0b110000) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(49) + chr(0b10011 + 0o37) + chr(55), 51204 - 51196), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\062' + chr(0b101111 + 0o10), 0b1000), ehT0Px3KOsy9(chr(80 - 32) + '\x6f' + chr(49) + chr(0b110000) + chr(0b110100), 22935 - 22927), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b1110 + 0o44) + '\061' + chr(0b110110), 0b1000), ehT0Px3KOsy9('\x30' + '\157' + '\x32' + chr(0b110011) + chr(0b101111 + 0o1), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(50) + '\061' + chr(0b110100), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + '\063' + '\x32' + chr(0b11001 + 0o31), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(753 - 698) + chr(621 - 567), 0b1000), ehT0Px3KOsy9(chr(0b11011 + 0o25) + '\x6f' + chr(1626 - 1576) + chr(1998 - 1947) + chr(0b110111), 8356 - 8348), ehT0Px3KOsy9('\060' + chr(111) + chr(0b10010 + 0o41) + chr(0b11000 + 0o31) + '\067', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1010101 + 0o32) + '\063' + chr(346 - 298), ord("\x08")), ehT0Px3KOsy9(chr(251 - 203) + '\157' + chr(50) + chr(50) + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(589 - 541) + chr(0b1101111) + chr(0b111 + 0o54) + chr(2111 - 2058) + chr(63 - 14), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110011 + 0o0) + chr(0b110000 + 0o5) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1010111 + 0o30) + chr(0b110011) + chr(2102 - 2054) + '\065', 0o10), ehT0Px3KOsy9(chr(1867 - 1819) + chr(1238 - 1127) + chr(0b1110 + 0o43) + chr(50) + chr(0b101000 + 0o13), ord("\x08")), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(111) + '\x32' + chr(0b110001) + chr(834 - 783), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(2352 - 2241) + chr(0b110001) + chr(0b110110) + chr(54), 0b1000), ehT0Px3KOsy9(chr(1200 - 1152) + chr(0b1101111) + chr(0b110111) + '\x32', 0o10), ehT0Px3KOsy9(chr(511 - 463) + chr(0b1101111) + '\x33' + chr(51) + chr(1165 - 1111), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110000 + 0o2) + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(0b1001 + 0o47) + chr(0b1101111) + '\x31' + '\061' + chr(0b110011), 49197 - 49189), ehT0Px3KOsy9(chr(871 - 823) + chr(0b111100 + 0o63) + chr(0b110001) + chr(0b100 + 0o56) + '\066', ord("\x08")), ehT0Px3KOsy9(chr(1625 - 1577) + chr(9108 - 8997) + chr(50) + chr(0b11111 + 0o26) + chr(52), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b1100 + 0o47) + '\x32', 0o10), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(3846 - 3735) + chr(1963 - 1914) + chr(0b110110) + chr(0b110000), 16110 - 16102), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(50) + chr(0b110001 + 0o5) + chr(1449 - 1399), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(51) + '\x32' + chr(49), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b101011 + 0o11) + '\061', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x33' + chr(51) + chr(0b100001 + 0o23), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(11398 - 11287) + '\062' + chr(0b10101 + 0o42), 8), ehT0Px3KOsy9(chr(654 - 606) + chr(111) + chr(0b10110 + 0o34) + chr(0b110011) + '\x33', 0b1000), ehT0Px3KOsy9(chr(0b10010 + 0o36) + '\x6f' + chr(0b110001) + '\x34', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110001) + chr(0b110001) + chr(2957 - 2902), 52455 - 52447)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b0 + 0o60) + '\157' + '\065' + '\060', ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'f'), chr(7349 - 7249) + chr(9083 - 8982) + chr(0b111110 + 0o45) + chr(752 - 641) + '\144' + '\145')(chr(0b1101000 + 0o15) + '\x74' + chr(0b1000 + 0o136) + chr(2016 - 1971) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def mdXp8pM6AUVj(E1c_5v_Zd9l8, xXh07emJaiUD, vRVqPOZ1hUG7, ag0mwEgWzjYv, X91zauzfCPzF, holLFgwB7vsP):
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x1bca\x85p\xaf'), chr(0b1100100) + '\x65' + '\143' + '\x6f' + '\x64' + '\x65')(chr(0b1110101) + '\164' + chr(102) + chr(45) + chr(0b110011 + 0o5)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x1ac`\x85u\xb66{'), '\x64' + chr(4982 - 4881) + '\143' + '\x6f' + chr(0b100000 + 0o104) + chr(0b1100101))(chr(0b1001100 + 0o51) + '\x74' + chr(0b1100110) + chr(45) + '\070'))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x05gc'), chr(0b1100001 + 0o3) + chr(0b1100101) + chr(3569 - 3470) + '\x6f' + chr(100) + '\x65')(chr(3648 - 3531) + '\164' + chr(0b11110 + 0o110) + chr(0b1101 + 0o40) + chr(56)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x04gj\x89c\x8d8e)'), '\x64' + chr(3746 - 3645) + chr(0b1100011) + '\157' + '\x64' + '\145')(chr(117) + chr(0b1110100) + '\x66' + chr(0b11110 + 0o17) + '\070'))()), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x0bnf\x82z\xa63T%\xfb\xeb\xed\xbes\xea\xa3l\xfd\xa1\x8a\xa47\xee\x12{\xaaf.8\xd8s\x1f\xdc'), chr(0b1001001 + 0o33) + chr(0b110010 + 0o63) + chr(0b11 + 0o140) + chr(0b1001000 + 0o47) + '\144' + '\x65')(chr(0b1110101) + chr(700 - 584) + '\146' + chr(45) + '\070'))(E1c_5v_Zd9l8, num_heads=vRVqPOZ1hUG7, dropout=ag0mwEgWzjYv, chunk_selector=X91zauzfCPzF, mode=holLFgwB7vsP), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x05gc'), chr(0b1000 + 0o134) + chr(7935 - 7834) + chr(4169 - 4070) + '\x6f' + chr(0b1100100) + chr(101))('\165' + chr(385 - 269) + chr(102) + '\x2d' + '\x38'))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x0ct|\x9c~\xb6#'), chr(8011 - 7911) + chr(9518 - 9417) + chr(99) + chr(111) + chr(0b110000 + 0o64) + chr(0b110011 + 0o62))(chr(0b1011111 + 0o26) + chr(116) + '\x66' + '\055' + chr(802 - 746)))(rate=ag0mwEgWzjYv, mode=holLFgwB7vsP))), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x05gc'), chr(0b1001 + 0o133) + chr(0b1011 + 0o132) + '\x63' + '\x6f' + chr(9906 - 9806) + chr(0b1001111 + 0o26))(chr(117) + chr(116) + chr(0b1100011 + 0o3) + chr(1136 - 1091) + '\070'))(Z01zxSkrfh7f(E1c_5v_Zd9l8, xXh07emJaiUD, ag0mwEgWzjYv, mode=holLFgwB7vsP)))
|
tensorflow/tensor2tensor
|
tensor2tensor/trax/models/transformer.py
|
ChunkedTransformerLM
|
def ChunkedTransformerLM(vocab_size,
feature_depth=512,
feedforward_depth=2048,
num_layers=6,
num_heads=8,
dropout=0.1,
chunk_selector=None,
max_len=2048,
mode='train'):
"""Transformer language model operating on chunks.
The input to this model is a sequence presented as a list or tuple of chunks:
(chunk1, chunk2, chunks3, ..., chunkN).
Each chunk should have the same shape (batch, chunk-length) and together they
represent a long sequence that's a concatenation chunk1,chunk2,...,chunkN.
Chunked Transformer emulates the operation of a Transformer on this long
sequence except for the chunked attention layer, which may attend to only
a subset of the chunks to reduce memory use.
Args:
vocab_size: int: vocab size
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
chunk_selector: a function from chunk number to list of chunks to attend
(if None, attends to the previous chunks which is equivalent to setting
chunk_selector(x) = [] if x < 1 else [x-1] (TransformerXL); we attend
to the current chunk with a causal mask too, selected chunks unmasked).
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the layer.
"""
stack = [ChunkedDecoderLayer(feature_depth, feedforward_depth, num_heads,
dropout, chunk_selector, mode)
for _ in range(num_layers)]
# Below each Map(L) applies the layer L to each chunk independently.
return layers.Serial(
layers.ShiftRight(),
layers.Map(layers.Embedding(feature_depth, vocab_size)),
layers.Map(layers.Dropout(rate=dropout, mode=mode)),
layers.PositionalEncoding(max_len=max_len),
layers.Serial(*stack),
layers.Map(layers.LayerNorm()),
layers.Map(layers.Dense(vocab_size)),
layers.Map(layers.LogSoftmax()),
)
|
python
|
def ChunkedTransformerLM(vocab_size,
feature_depth=512,
feedforward_depth=2048,
num_layers=6,
num_heads=8,
dropout=0.1,
chunk_selector=None,
max_len=2048,
mode='train'):
"""Transformer language model operating on chunks.
The input to this model is a sequence presented as a list or tuple of chunks:
(chunk1, chunk2, chunks3, ..., chunkN).
Each chunk should have the same shape (batch, chunk-length) and together they
represent a long sequence that's a concatenation chunk1,chunk2,...,chunkN.
Chunked Transformer emulates the operation of a Transformer on this long
sequence except for the chunked attention layer, which may attend to only
a subset of the chunks to reduce memory use.
Args:
vocab_size: int: vocab size
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
chunk_selector: a function from chunk number to list of chunks to attend
(if None, attends to the previous chunks which is equivalent to setting
chunk_selector(x) = [] if x < 1 else [x-1] (TransformerXL); we attend
to the current chunk with a causal mask too, selected chunks unmasked).
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the layer.
"""
stack = [ChunkedDecoderLayer(feature_depth, feedforward_depth, num_heads,
dropout, chunk_selector, mode)
for _ in range(num_layers)]
# Below each Map(L) applies the layer L to each chunk independently.
return layers.Serial(
layers.ShiftRight(),
layers.Map(layers.Embedding(feature_depth, vocab_size)),
layers.Map(layers.Dropout(rate=dropout, mode=mode)),
layers.PositionalEncoding(max_len=max_len),
layers.Serial(*stack),
layers.Map(layers.LayerNorm()),
layers.Map(layers.Dense(vocab_size)),
layers.Map(layers.LogSoftmax()),
)
|
[
"def",
"ChunkedTransformerLM",
"(",
"vocab_size",
",",
"feature_depth",
"=",
"512",
",",
"feedforward_depth",
"=",
"2048",
",",
"num_layers",
"=",
"6",
",",
"num_heads",
"=",
"8",
",",
"dropout",
"=",
"0.1",
",",
"chunk_selector",
"=",
"None",
",",
"max_len",
"=",
"2048",
",",
"mode",
"=",
"'train'",
")",
":",
"stack",
"=",
"[",
"ChunkedDecoderLayer",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"num_heads",
",",
"dropout",
",",
"chunk_selector",
",",
"mode",
")",
"for",
"_",
"in",
"range",
"(",
"num_layers",
")",
"]",
"# Below each Map(L) applies the layer L to each chunk independently.",
"return",
"layers",
".",
"Serial",
"(",
"layers",
".",
"ShiftRight",
"(",
")",
",",
"layers",
".",
"Map",
"(",
"layers",
".",
"Embedding",
"(",
"feature_depth",
",",
"vocab_size",
")",
")",
",",
"layers",
".",
"Map",
"(",
"layers",
".",
"Dropout",
"(",
"rate",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
",",
"layers",
".",
"PositionalEncoding",
"(",
"max_len",
"=",
"max_len",
")",
",",
"layers",
".",
"Serial",
"(",
"*",
"stack",
")",
",",
"layers",
".",
"Map",
"(",
"layers",
".",
"LayerNorm",
"(",
")",
")",
",",
"layers",
".",
"Map",
"(",
"layers",
".",
"Dense",
"(",
"vocab_size",
")",
")",
",",
"layers",
".",
"Map",
"(",
"layers",
".",
"LogSoftmax",
"(",
")",
")",
",",
")"
] |
Transformer language model operating on chunks.
The input to this model is a sequence presented as a list or tuple of chunks:
(chunk1, chunk2, chunks3, ..., chunkN).
Each chunk should have the same shape (batch, chunk-length) and together they
represent a long sequence that's a concatenation chunk1,chunk2,...,chunkN.
Chunked Transformer emulates the operation of a Transformer on this long
sequence except for the chunked attention layer, which may attend to only
a subset of the chunks to reduce memory use.
Args:
vocab_size: int: vocab size
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_layers: int: number of encoder/decoder layers
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
chunk_selector: a function from chunk number to list of chunks to attend
(if None, attends to the previous chunks which is equivalent to setting
chunk_selector(x) = [] if x < 1 else [x-1] (TransformerXL); we attend
to the current chunk with a causal mask too, selected chunks unmasked).
max_len: int: maximum symbol length for positional encoding
mode: str: 'train' or 'eval'
Returns:
the layer.
|
[
"Transformer",
"language",
"model",
"operating",
"on",
"chunks",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/trax/models/transformer.py#L223-L273
|
train
|
Transformer language model operating on chunks.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b100111 + 0o11) + '\x6f' + chr(0b10111 + 0o35), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(11514 - 11403) + chr(2695 - 2641) + chr(1497 - 1446), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1100010 + 0o15) + '\x35' + chr(2189 - 2140), ord("\x08")), ehT0Px3KOsy9(chr(1072 - 1024) + '\157' + '\x30', 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b101 + 0o54) + '\060' + chr(48), 913 - 905), ehT0Px3KOsy9(chr(853 - 805) + chr(111) + '\x33' + '\x31' + chr(0b110000), ord("\x08")), ehT0Px3KOsy9('\060' + chr(4422 - 4311) + '\x31' + chr(410 - 360) + chr(0b110 + 0o61), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(4820 - 4709) + '\061' + chr(0b110001) + '\060', 0o10), ehT0Px3KOsy9(chr(48) + chr(12165 - 12054) + '\061' + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(402 - 354) + chr(0b10010 + 0o135) + '\061' + chr(529 - 475) + chr(48), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(814 - 764) + chr(2502 - 2449) + chr(1674 - 1625), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110111) + chr(1794 - 1740), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x32' + chr(50) + '\x34', ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110001) + chr(1749 - 1695) + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(818 - 769) + chr(0b0 + 0o64) + chr(1240 - 1190), 31060 - 31052), ehT0Px3KOsy9(chr(535 - 487) + chr(0b11011 + 0o124) + chr(0b110010) + '\x32' + '\063', 25410 - 25402), ehT0Px3KOsy9(chr(0b1101 + 0o43) + '\157' + chr(831 - 782) + chr(0b100110 + 0o15) + chr(168 - 113), 22514 - 22506), ehT0Px3KOsy9('\060' + chr(5081 - 4970) + chr(0b110001) + '\065' + '\x31', 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b101101 + 0o5) + chr(1793 - 1738) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(0b11111 + 0o21) + chr(5746 - 5635) + chr(0b110011) + chr(1436 - 1385) + '\x31', ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\062' + chr(167 - 119) + chr(0b110100), 0b1000), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(0b1101111) + chr(50) + '\063' + chr(1971 - 1919), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1111 + 0o140) + '\064' + chr(0b110001), ord("\x08")), ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(0b1101111) + chr(0b1011 + 0o50) + '\067' + '\x32', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110001) + chr(0b10 + 0o61) + '\060', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(51) + '\063' + chr(1891 - 1842), 8), ehT0Px3KOsy9(chr(1591 - 1543) + '\x6f' + chr(0b100000 + 0o22) + chr(0b1110 + 0o47) + '\063', 52543 - 52535), ehT0Px3KOsy9(chr(1331 - 1283) + '\x6f' + chr(0b101 + 0o54) + '\062', 65496 - 65488), ehT0Px3KOsy9(chr(48) + chr(111) + chr(299 - 249) + chr(744 - 691) + '\x33', 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(312 - 262), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(5114 - 5003) + chr(0b110010 + 0o1) + '\x34' + chr(1127 - 1079), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x32' + '\067' + chr(633 - 579), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(2028 - 1978) + chr(48) + chr(1924 - 1875), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b10 + 0o61) + chr(49) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(0b11100 + 0o24) + '\157' + chr(988 - 934) + '\x35', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110110) + chr(0b110011), 8), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b101011 + 0o6) + chr(0b110111) + '\060', 0b1000), ehT0Px3KOsy9(chr(717 - 669) + chr(111) + chr(0b10010 + 0o41) + chr(0b100001 + 0o25) + chr(0b110101), 0o10), ehT0Px3KOsy9('\060' + chr(7316 - 7205) + chr(2501 - 2450) + chr(0b101 + 0o57) + chr(0b101110 + 0o10), 0o10), ehT0Px3KOsy9('\060' + '\157' + '\062' + chr(1745 - 1695) + chr(0b110010), 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b100010 + 0o23) + chr(2278 - 2230), 25047 - 25039)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'*'), '\144' + chr(3603 - 3502) + chr(0b111111 + 0o44) + chr(0b1101111) + '\x64' + chr(6274 - 6173))(chr(0b1110101) + '\164' + chr(2545 - 2443) + chr(45) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def _BnL7C_0SsLs(CeyMIoSyrpkQ, E1c_5v_Zd9l8=ehT0Px3KOsy9(chr(0b110000) + '\157' + '\061' + chr(0b10111 + 0o31) + chr(48) + chr(0b110000), ord("\x08")), xXh07emJaiUD=ehT0Px3KOsy9('\x30' + chr(0b1101010 + 0o5) + chr(1666 - 1614) + chr(48) + chr(1467 - 1419) + '\060', 0b1000), uftkTXJyNORO=ehT0Px3KOsy9(chr(0b100110 + 0o12) + chr(0b10110 + 0o131) + chr(54), 58623 - 58615), vRVqPOZ1hUG7=ehT0Px3KOsy9(chr(0b110000) + chr(6488 - 6377) + chr(2268 - 2219) + '\060', 0o10), ag0mwEgWzjYv=0.1, X91zauzfCPzF=None, qbKO12mgagKE=ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x34' + '\x30' + chr(48) + chr(0b100100 + 0o14), 8), holLFgwB7vsP=xafqLlk3kkUe(SXOLrMavuUCe(b'pAHi.'), chr(0b1100100) + '\145' + '\143' + chr(0b1110 + 0o141) + chr(100) + '\x65')(chr(0b1110101) + chr(0b1110100) + chr(0b100001 + 0o105) + chr(1540 - 1495) + chr(0b101111 + 0o11))):
rFoCQMjVYqWa = [mdXp8pM6AUVj(E1c_5v_Zd9l8, xXh07emJaiUD, vRVqPOZ1hUG7, ag0mwEgWzjYv, X91zauzfCPzF, holLFgwB7vsP) for VNGQdHSFPrso in vQr8gNKaIaWE(uftkTXJyNORO)]
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'WV[i!\xa4'), chr(0b1100100) + '\145' + chr(99) + chr(0b1000000 + 0o57) + chr(9260 - 9160) + chr(0b101111 + 0o66))('\x75' + chr(116) + chr(7506 - 7404) + chr(0b101101) + chr(2660 - 2604)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'W[@f4\x9a>/e\x04'), chr(0b110000 + 0o64) + chr(2112 - 2011) + chr(0b1000000 + 0o43) + '\x6f' + '\144' + chr(0b1100101))(chr(0b1101101 + 0o10) + chr(0b1110100) + chr(0b1100110) + chr(45) + chr(1429 - 1373)))(), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'IRY'), chr(100) + '\x65' + '\143' + chr(11207 - 11096) + '\144' + '\x65')(chr(0b1100 + 0o151) + '\164' + chr(0b1100110) + '\x2d' + chr(0b111000)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'A^Ke$\xac>&j'), '\144' + chr(5979 - 5878) + '\x63' + '\157' + chr(100) + chr(101))(chr(10849 - 10732) + chr(0b110100 + 0o100) + chr(102) + '\055' + chr(56)))(E1c_5v_Zd9l8, CeyMIoSyrpkQ)), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'IRY'), chr(0b101000 + 0o74) + chr(0b1001101 + 0o30) + chr(0b1100011) + chr(9526 - 9415) + '\x64' + chr(5614 - 5513))(chr(0b1011101 + 0o30) + chr(0b1110100) + '\146' + chr(0b101101) + chr(99 - 43)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'@AFp/\xbd#'), chr(0b101010 + 0o72) + chr(0b1100101) + chr(4700 - 4601) + chr(0b1101111) + chr(0b1000111 + 0o35) + '\x65')('\165' + '\164' + chr(0b1100110) + '\055' + '\x38'))(rate=ag0mwEgWzjYv, mode=holLFgwB7vsP)), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'T\\Zi4\xa18&l\x1c\xecP\xf7\x1d\x06\xfa1\x0e'), chr(2899 - 2799) + chr(0b1100101) + '\x63' + chr(319 - 208) + chr(6040 - 5940) + '\x65')(chr(0b11100 + 0o131) + chr(0b1110100) + chr(102) + chr(1991 - 1946) + '\x38'))(max_len=qbKO12mgagKE), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'WV[i!\xa4'), chr(100) + chr(0b1100101) + '\143' + '\157' + chr(0b1100100) + chr(101))(chr(1548 - 1431) + '\x74' + chr(0b1100110) + '\x2d' + chr(0b111000)))(*rFoCQMjVYqWa), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'IRY'), chr(0b1100100) + '\145' + chr(99) + chr(3395 - 3284) + chr(0b1100100) + '\x65')('\165' + chr(116) + '\x66' + chr(0b101101) + chr(0b111000)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'HRPe2\x868:`'), chr(0b1100100) + chr(0b1100101) + chr(0b10010 + 0o121) + chr(7752 - 7641) + '\144' + '\x65')('\165' + '\164' + '\x66' + '\055' + chr(56)))()), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'IRY'), chr(0b111010 + 0o52) + chr(101) + chr(0b1100011) + chr(0b1110 + 0o141) + chr(5540 - 5440) + chr(101))(chr(117) + chr(0b1000111 + 0o55) + chr(0b1010110 + 0o20) + chr(0b11000 + 0o25) + chr(0b1101 + 0o53)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'@VGs%'), chr(0b111111 + 0o45) + '\x65' + '\143' + chr(111) + chr(7212 - 7112) + '\145')(chr(0b1110101) + chr(0b1101110 + 0o6) + chr(4359 - 4257) + '\055' + '\070'))(CeyMIoSyrpkQ)), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'IRY'), chr(0b1100100) + chr(9908 - 9807) + chr(99) + chr(0b1101111) + '\144' + '\x65')('\165' + '\164' + chr(0b1000 + 0o136) + '\055' + chr(1204 - 1148)))(xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'H\\NS/\xae#%l\x08'), chr(0b1011001 + 0o13) + chr(0b1001 + 0o134) + chr(0b1100011) + chr(0b1101111) + '\x64' + chr(2091 - 1990))(chr(0b1 + 0o164) + chr(0b1011011 + 0o31) + chr(0b1011110 + 0o10) + chr(0b100 + 0o51) + chr(0b111000)))()))
|
tensorflow/tensor2tensor
|
tensor2tensor/trax/models/transformer.py
|
Transformer
|
def Transformer(source_vocab_size,
target_vocab_size,
mode='train',
num_layers=6,
feature_depth=512,
feedforward_depth=2048,
num_heads=8,
dropout=0.1,
shared_embedding=True,
max_len=200,
return_evals=False):
"""Transformer model.
Args:
source_vocab_size: int: source vocab size
target_vocab_size: int: target vocab size
mode: str: 'train' or 'eval'
num_layers: int: number of encoder/decoder layers
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
shared_embedding: bool: specify whether source/target embeddings are tied.
max_len: int: maximum symbol length for positional encoding
return_evals: bool: whether to generate decode-time evaluation functions
Returns:
A namedtuple containing model 'init' and 'apply' functions for training and
the 'evals' functions that itself returns a namedtuple containing evaluation
functions for the trained encoder, decoder, and generator substax.
"""
# Input embedding and positional encoding
inject_position = layers.Serial(
layers.Dropout(dropout, mode=mode),
layers.PositionalEncoding(feature_depth, max_len=max_len)
)
if shared_embedding:
assert source_vocab_size == target_vocab_size
# Weight-shared Embedding
embedding = layers.Share(layers.Embedding(feature_depth, source_vocab_size))
source_embedding_layer = layers.Serial(embedding, inject_position)
target_embedding_layer = source_embedding_layer
else:
source_embedding = layers.Embedding(feature_depth, source_vocab_size)
target_embedding = layers.Embedding(feature_depth, target_vocab_size)
source_embedding_layer = layers.Serial(source_embedding, inject_position)
target_embedding_layer = layers.Serial(target_embedding, inject_position)
# Multi-headed Attention and Feed-forward layers
multi_attention = layers.MultiHeadedAttention(
feature_depth, num_heads=num_heads, dropout=dropout, mode=mode)
# Encoder
@layers.Lambda
def Encoder(source, source_mask):
"""Transformer encoder stack.
Args:
source: layer variable: raw source sequences
source_mask: layer variable: self-attention mask
Returns:
Layer variable that outputs encoded source.
"""
encoder_layer = layers.Serial(
# input attends to self
layers.Residual(layers.LayerNorm(),
layers.Branch(size=4),
layers.Parallel(layers.Identity(), # query
layers.Identity(), # key
layers.Identity(), # value
source_mask), # attention mask
multi_attention,
layers.Dropout(dropout, mode=mode)),
# feed-forward
ResidualFeedForward(
feature_depth, feedforward_depth, dropout, mode=mode),
)
return layers.Serial(
source,
source_embedding_layer,
layers.repeat(encoder_layer, num_layers),
layers.LayerNorm(),
)
# Decoder
@layers.Lambda
def Decoder(memory, target, target_mask, memory_mask):
"""Transformer decoder stack.
Args:
memory: layer variable: encoded source sequences
target: layer variable: raw target sequences
target_mask: layer variable: self-attention mask
memory_mask: layer variable: memory attention mask
Returns:
Layer variable that outputs encoded source.
"""
decoder_layer = layers.Serial(
# target attends to self
layers.Residual(layers.LayerNorm(),
layers.Branch(size=4),
layers.Parallel(layers.Identity(), # query
layers.Identity(), # key
layers.Identity(), # value
target_mask), # attention mask
multi_attention,
layers.Dropout(dropout, mode=mode)),
# target attends to encoded source
layers.Residual(layers.LayerNorm(),
layers.Branch(size=4),
layers.Parallel(layers.Identity(), # query
memory, # key
memory, # value
memory_mask), # attention mask
multi_attention,
layers.Dropout(dropout, mode=mode)),
# feed-forward
ResidualFeedForward(
feature_depth, feedforward_depth, dropout, mode=mode)
)
return layers.Serial(
target,
target_embedding_layer,
layers.repeat(decoder_layer, num_layers),
layers.LayerNorm(),
)
# The Transformer
@layers.Lambda
def transformer(source, target, source_mask, target_mask, memory_mask): # pylint: disable=invalid-name
encoded_source = Encoder(source, source_mask)
return Decoder(encoded_source, target, target_mask, memory_mask)
# Finally, bind the generator transform to use later for inference.
@layers.Lambda
def Generator(encoded_target):
return layers.Serial(
encoded_target,
layers.Dense(target_vocab_size),
layers.LogSoftmax
)
# Model-Building and Evaluation Functions
# Get entire model's the layer pair
top_init, top_apply = Generator(transformer)
# By default act as a normal constructor and emit an (init, apply) pair.
if not return_evals:
return (top_init, top_apply)
else:
raise ValueError('inference in this model is still a work in progress')
|
python
|
def Transformer(source_vocab_size,
target_vocab_size,
mode='train',
num_layers=6,
feature_depth=512,
feedforward_depth=2048,
num_heads=8,
dropout=0.1,
shared_embedding=True,
max_len=200,
return_evals=False):
"""Transformer model.
Args:
source_vocab_size: int: source vocab size
target_vocab_size: int: target vocab size
mode: str: 'train' or 'eval'
num_layers: int: number of encoder/decoder layers
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
shared_embedding: bool: specify whether source/target embeddings are tied.
max_len: int: maximum symbol length for positional encoding
return_evals: bool: whether to generate decode-time evaluation functions
Returns:
A namedtuple containing model 'init' and 'apply' functions for training and
the 'evals' functions that itself returns a namedtuple containing evaluation
functions for the trained encoder, decoder, and generator substax.
"""
# Input embedding and positional encoding
inject_position = layers.Serial(
layers.Dropout(dropout, mode=mode),
layers.PositionalEncoding(feature_depth, max_len=max_len)
)
if shared_embedding:
assert source_vocab_size == target_vocab_size
# Weight-shared Embedding
embedding = layers.Share(layers.Embedding(feature_depth, source_vocab_size))
source_embedding_layer = layers.Serial(embedding, inject_position)
target_embedding_layer = source_embedding_layer
else:
source_embedding = layers.Embedding(feature_depth, source_vocab_size)
target_embedding = layers.Embedding(feature_depth, target_vocab_size)
source_embedding_layer = layers.Serial(source_embedding, inject_position)
target_embedding_layer = layers.Serial(target_embedding, inject_position)
# Multi-headed Attention and Feed-forward layers
multi_attention = layers.MultiHeadedAttention(
feature_depth, num_heads=num_heads, dropout=dropout, mode=mode)
# Encoder
@layers.Lambda
def Encoder(source, source_mask):
"""Transformer encoder stack.
Args:
source: layer variable: raw source sequences
source_mask: layer variable: self-attention mask
Returns:
Layer variable that outputs encoded source.
"""
encoder_layer = layers.Serial(
# input attends to self
layers.Residual(layers.LayerNorm(),
layers.Branch(size=4),
layers.Parallel(layers.Identity(), # query
layers.Identity(), # key
layers.Identity(), # value
source_mask), # attention mask
multi_attention,
layers.Dropout(dropout, mode=mode)),
# feed-forward
ResidualFeedForward(
feature_depth, feedforward_depth, dropout, mode=mode),
)
return layers.Serial(
source,
source_embedding_layer,
layers.repeat(encoder_layer, num_layers),
layers.LayerNorm(),
)
# Decoder
@layers.Lambda
def Decoder(memory, target, target_mask, memory_mask):
"""Transformer decoder stack.
Args:
memory: layer variable: encoded source sequences
target: layer variable: raw target sequences
target_mask: layer variable: self-attention mask
memory_mask: layer variable: memory attention mask
Returns:
Layer variable that outputs encoded source.
"""
decoder_layer = layers.Serial(
# target attends to self
layers.Residual(layers.LayerNorm(),
layers.Branch(size=4),
layers.Parallel(layers.Identity(), # query
layers.Identity(), # key
layers.Identity(), # value
target_mask), # attention mask
multi_attention,
layers.Dropout(dropout, mode=mode)),
# target attends to encoded source
layers.Residual(layers.LayerNorm(),
layers.Branch(size=4),
layers.Parallel(layers.Identity(), # query
memory, # key
memory, # value
memory_mask), # attention mask
multi_attention,
layers.Dropout(dropout, mode=mode)),
# feed-forward
ResidualFeedForward(
feature_depth, feedforward_depth, dropout, mode=mode)
)
return layers.Serial(
target,
target_embedding_layer,
layers.repeat(decoder_layer, num_layers),
layers.LayerNorm(),
)
# The Transformer
@layers.Lambda
def transformer(source, target, source_mask, target_mask, memory_mask): # pylint: disable=invalid-name
encoded_source = Encoder(source, source_mask)
return Decoder(encoded_source, target, target_mask, memory_mask)
# Finally, bind the generator transform to use later for inference.
@layers.Lambda
def Generator(encoded_target):
return layers.Serial(
encoded_target,
layers.Dense(target_vocab_size),
layers.LogSoftmax
)
# Model-Building and Evaluation Functions
# Get entire model's the layer pair
top_init, top_apply = Generator(transformer)
# By default act as a normal constructor and emit an (init, apply) pair.
if not return_evals:
return (top_init, top_apply)
else:
raise ValueError('inference in this model is still a work in progress')
|
[
"def",
"Transformer",
"(",
"source_vocab_size",
",",
"target_vocab_size",
",",
"mode",
"=",
"'train'",
",",
"num_layers",
"=",
"6",
",",
"feature_depth",
"=",
"512",
",",
"feedforward_depth",
"=",
"2048",
",",
"num_heads",
"=",
"8",
",",
"dropout",
"=",
"0.1",
",",
"shared_embedding",
"=",
"True",
",",
"max_len",
"=",
"200",
",",
"return_evals",
"=",
"False",
")",
":",
"# Input embedding and positional encoding",
"inject_position",
"=",
"layers",
".",
"Serial",
"(",
"layers",
".",
"Dropout",
"(",
"dropout",
",",
"mode",
"=",
"mode",
")",
",",
"layers",
".",
"PositionalEncoding",
"(",
"feature_depth",
",",
"max_len",
"=",
"max_len",
")",
")",
"if",
"shared_embedding",
":",
"assert",
"source_vocab_size",
"==",
"target_vocab_size",
"# Weight-shared Embedding",
"embedding",
"=",
"layers",
".",
"Share",
"(",
"layers",
".",
"Embedding",
"(",
"feature_depth",
",",
"source_vocab_size",
")",
")",
"source_embedding_layer",
"=",
"layers",
".",
"Serial",
"(",
"embedding",
",",
"inject_position",
")",
"target_embedding_layer",
"=",
"source_embedding_layer",
"else",
":",
"source_embedding",
"=",
"layers",
".",
"Embedding",
"(",
"feature_depth",
",",
"source_vocab_size",
")",
"target_embedding",
"=",
"layers",
".",
"Embedding",
"(",
"feature_depth",
",",
"target_vocab_size",
")",
"source_embedding_layer",
"=",
"layers",
".",
"Serial",
"(",
"source_embedding",
",",
"inject_position",
")",
"target_embedding_layer",
"=",
"layers",
".",
"Serial",
"(",
"target_embedding",
",",
"inject_position",
")",
"# Multi-headed Attention and Feed-forward layers",
"multi_attention",
"=",
"layers",
".",
"MultiHeadedAttention",
"(",
"feature_depth",
",",
"num_heads",
"=",
"num_heads",
",",
"dropout",
"=",
"dropout",
",",
"mode",
"=",
"mode",
")",
"# Encoder",
"@",
"layers",
".",
"Lambda",
"def",
"Encoder",
"(",
"source",
",",
"source_mask",
")",
":",
"\"\"\"Transformer encoder stack.\n\n Args:\n source: layer variable: raw source sequences\n source_mask: layer variable: self-attention mask\n\n Returns:\n Layer variable that outputs encoded source.\n \"\"\"",
"encoder_layer",
"=",
"layers",
".",
"Serial",
"(",
"# input attends to self",
"layers",
".",
"Residual",
"(",
"layers",
".",
"LayerNorm",
"(",
")",
",",
"layers",
".",
"Branch",
"(",
"size",
"=",
"4",
")",
",",
"layers",
".",
"Parallel",
"(",
"layers",
".",
"Identity",
"(",
")",
",",
"# query",
"layers",
".",
"Identity",
"(",
")",
",",
"# key",
"layers",
".",
"Identity",
"(",
")",
",",
"# value",
"source_mask",
")",
",",
"# attention mask",
"multi_attention",
",",
"layers",
".",
"Dropout",
"(",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
",",
"# feed-forward",
"ResidualFeedForward",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"dropout",
",",
"mode",
"=",
"mode",
")",
",",
")",
"return",
"layers",
".",
"Serial",
"(",
"source",
",",
"source_embedding_layer",
",",
"layers",
".",
"repeat",
"(",
"encoder_layer",
",",
"num_layers",
")",
",",
"layers",
".",
"LayerNorm",
"(",
")",
",",
")",
"# Decoder",
"@",
"layers",
".",
"Lambda",
"def",
"Decoder",
"(",
"memory",
",",
"target",
",",
"target_mask",
",",
"memory_mask",
")",
":",
"\"\"\"Transformer decoder stack.\n\n Args:\n memory: layer variable: encoded source sequences\n target: layer variable: raw target sequences\n target_mask: layer variable: self-attention mask\n memory_mask: layer variable: memory attention mask\n\n Returns:\n Layer variable that outputs encoded source.\n \"\"\"",
"decoder_layer",
"=",
"layers",
".",
"Serial",
"(",
"# target attends to self",
"layers",
".",
"Residual",
"(",
"layers",
".",
"LayerNorm",
"(",
")",
",",
"layers",
".",
"Branch",
"(",
"size",
"=",
"4",
")",
",",
"layers",
".",
"Parallel",
"(",
"layers",
".",
"Identity",
"(",
")",
",",
"# query",
"layers",
".",
"Identity",
"(",
")",
",",
"# key",
"layers",
".",
"Identity",
"(",
")",
",",
"# value",
"target_mask",
")",
",",
"# attention mask",
"multi_attention",
",",
"layers",
".",
"Dropout",
"(",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
",",
"# target attends to encoded source",
"layers",
".",
"Residual",
"(",
"layers",
".",
"LayerNorm",
"(",
")",
",",
"layers",
".",
"Branch",
"(",
"size",
"=",
"4",
")",
",",
"layers",
".",
"Parallel",
"(",
"layers",
".",
"Identity",
"(",
")",
",",
"# query",
"memory",
",",
"# key",
"memory",
",",
"# value",
"memory_mask",
")",
",",
"# attention mask",
"multi_attention",
",",
"layers",
".",
"Dropout",
"(",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
",",
"# feed-forward",
"ResidualFeedForward",
"(",
"feature_depth",
",",
"feedforward_depth",
",",
"dropout",
",",
"mode",
"=",
"mode",
")",
")",
"return",
"layers",
".",
"Serial",
"(",
"target",
",",
"target_embedding_layer",
",",
"layers",
".",
"repeat",
"(",
"decoder_layer",
",",
"num_layers",
")",
",",
"layers",
".",
"LayerNorm",
"(",
")",
",",
")",
"# The Transformer",
"@",
"layers",
".",
"Lambda",
"def",
"transformer",
"(",
"source",
",",
"target",
",",
"source_mask",
",",
"target_mask",
",",
"memory_mask",
")",
":",
"# pylint: disable=invalid-name",
"encoded_source",
"=",
"Encoder",
"(",
"source",
",",
"source_mask",
")",
"return",
"Decoder",
"(",
"encoded_source",
",",
"target",
",",
"target_mask",
",",
"memory_mask",
")",
"# Finally, bind the generator transform to use later for inference.",
"@",
"layers",
".",
"Lambda",
"def",
"Generator",
"(",
"encoded_target",
")",
":",
"return",
"layers",
".",
"Serial",
"(",
"encoded_target",
",",
"layers",
".",
"Dense",
"(",
"target_vocab_size",
")",
",",
"layers",
".",
"LogSoftmax",
")",
"# Model-Building and Evaluation Functions",
"# Get entire model's the layer pair",
"top_init",
",",
"top_apply",
"=",
"Generator",
"(",
"transformer",
")",
"# By default act as a normal constructor and emit an (init, apply) pair.",
"if",
"not",
"return_evals",
":",
"return",
"(",
"top_init",
",",
"top_apply",
")",
"else",
":",
"raise",
"ValueError",
"(",
"'inference in this model is still a work in progress'",
")"
] |
Transformer model.
Args:
source_vocab_size: int: source vocab size
target_vocab_size: int: target vocab size
mode: str: 'train' or 'eval'
num_layers: int: number of encoder/decoder layers
feature_depth: int: depth of embedding
feedforward_depth: int: depth of feed-forward layer
num_heads: int: number of attention heads
dropout: float: dropout rate (how much to drop out)
shared_embedding: bool: specify whether source/target embeddings are tied.
max_len: int: maximum symbol length for positional encoding
return_evals: bool: whether to generate decode-time evaluation functions
Returns:
A namedtuple containing model 'init' and 'apply' functions for training and
the 'evals' functions that itself returns a namedtuple containing evaluation
functions for the trained encoder, decoder, and generator substax.
|
[
"Transformer",
"model",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/trax/models/transformer.py#L279-L431
|
train
|
Transformer model.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b100000 + 0o23) + '\x37' + '\x33', ord("\x08")), ehT0Px3KOsy9('\060' + chr(10777 - 10666) + chr(49) + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + '\065' + chr(55), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1001111 + 0o40) + chr(951 - 900) + chr(0b110101) + chr(48), 9180 - 9172), ehT0Px3KOsy9(chr(833 - 785) + '\157' + chr(0b110011) + chr(953 - 904) + chr(0b110110), 0o10), ehT0Px3KOsy9(chr(1589 - 1541) + chr(1503 - 1392) + chr(1617 - 1566) + chr(52) + chr(1517 - 1468), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\x32' + chr(0b100 + 0o63) + '\062', 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + '\x30' + '\x34', 0o10), ehT0Px3KOsy9(chr(0b11001 + 0o27) + '\x6f' + '\x33' + chr(0b11111 + 0o25) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110100), 50832 - 50824), ehT0Px3KOsy9(chr(48) + chr(2318 - 2207) + '\x32' + '\064' + chr(0b110001 + 0o5), 45098 - 45090), ehT0Px3KOsy9('\x30' + chr(111) + '\063' + chr(0b110101) + chr(0b10011 + 0o37), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(777 - 666) + chr(0b110010) + '\x30' + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\157' + chr(2135 - 2086) + chr(0b110110) + chr(1255 - 1207), 4801 - 4793), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10011 + 0o37) + chr(0b110100) + '\067', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(0b10101 + 0o33), 40543 - 40535), ehT0Px3KOsy9(chr(48) + chr(111) + chr(49) + chr(0b110011) + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(10310 - 10199) + chr(0b110100) + '\064', 0b1000), ehT0Px3KOsy9('\060' + '\157' + chr(0b110100 + 0o3) + chr(0b110111), ord("\x08")), ehT0Px3KOsy9(chr(0b1 + 0o57) + chr(6405 - 6294) + '\064', 8), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110011) + chr(48) + '\x33', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + '\x31' + chr(53), 0o10), ehT0Px3KOsy9('\060' + chr(148 - 37) + chr(527 - 478) + chr(52) + chr(804 - 755), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110001) + chr(50) + '\x31', 58017 - 58009), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(2027 - 1977) + chr(0b110110) + chr(0b101001 + 0o15), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(50) + chr(0b11101 + 0o31) + chr(0b101001 + 0o7), 0o10), ehT0Px3KOsy9(chr(0b10000 + 0o40) + '\157' + chr(0b110010) + chr(0b110101) + '\061', 0o10), ehT0Px3KOsy9('\x30' + '\x6f' + '\x34' + chr(487 - 433), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110010) + chr(2508 - 2456) + chr(48), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(6483 - 6372) + '\063' + chr(0b110010) + '\x32', 5301 - 5293), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\x33' + chr(0b110011) + chr(0b110100), 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b101001 + 0o12) + chr(0b110101) + chr(50), 8), ehT0Px3KOsy9(chr(834 - 786) + '\x6f' + chr(0b10101 + 0o36) + '\065', 36626 - 36618), ehT0Px3KOsy9(chr(48) + '\x6f' + '\x32' + '\065' + '\066', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b100001 + 0o116) + chr(51) + '\x37' + chr(2095 - 2044), 8), ehT0Px3KOsy9('\x30' + '\157' + chr(1240 - 1189) + chr(194 - 146) + chr(0b101100 + 0o12), 48990 - 48982), ehT0Px3KOsy9('\060' + '\157' + '\064' + '\064', 8), ehT0Px3KOsy9(chr(484 - 436) + chr(111) + '\065' + '\x37', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(51) + chr(0b110100) + chr(0b1010 + 0o50), 0o10), ehT0Px3KOsy9('\060' + '\157' + '\060', 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\065' + chr(2045 - 1997), 0b1000)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xd5'), '\x64' + chr(0b1100001 + 0o4) + chr(99) + chr(9275 - 9164) + '\x64' + chr(5958 - 5857))(chr(0b11101 + 0o130) + chr(4147 - 4031) + '\146' + '\055' + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def BBoBUISS8Jxr(XUqE0GHx9ndw, HHpQzG13Xmp7, holLFgwB7vsP=xafqLlk3kkUe(SXOLrMavuUCe(b'\x8f|\x8e\x81\xa0'), chr(1630 - 1530) + chr(0b1100101) + '\143' + chr(187 - 76) + '\144' + chr(101))(chr(0b1110101) + chr(116) + chr(705 - 603) + '\x2d' + chr(0b111000)), uftkTXJyNORO=ehT0Px3KOsy9(chr(0b1010 + 0o46) + chr(111) + '\066', 12145 - 12137), E1c_5v_Zd9l8=ehT0Px3KOsy9('\060' + chr(111) + '\x31' + chr(0b110000) + chr(0b110000) + '\060', ord("\x08")), xXh07emJaiUD=ehT0Px3KOsy9(chr(2188 - 2140) + chr(111) + chr(748 - 696) + '\x30' + '\x30' + chr(0b110000), 0o10), vRVqPOZ1hUG7=ehT0Px3KOsy9(chr(1055 - 1007) + '\157' + chr(0b110001) + '\060', ord("\x08")), ag0mwEgWzjYv=0.1, f7bdxoAgzo_R=ehT0Px3KOsy9(chr(0b11 + 0o55) + '\x6f' + '\061', 54398 - 54390), qbKO12mgagKE=ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(51) + chr(1893 - 1844) + chr(0b110000), 7941 - 7933), HStumbLH46tl=ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110000), 8)):
PhU_C1cgYIh2 = sGi5Aql23May.Serial(sGi5Aql23May.Dropout(ag0mwEgWzjYv, mode=holLFgwB7vsP), sGi5Aql23May.PositionalEncoding(E1c_5v_Zd9l8, max_len=qbKO12mgagKE))
if f7bdxoAgzo_R:
assert XUqE0GHx9ndw == HHpQzG13Xmp7
lwIir85sFEJp = sGi5Aql23May.Share(sGi5Aql23May.Embedding(E1c_5v_Zd9l8, XUqE0GHx9ndw))
wqEu0qNdxfOe = sGi5Aql23May.Serial(lwIir85sFEJp, PhU_C1cgYIh2)
J_ePOe8eJp9N = wqEu0qNdxfOe
else:
SdSEWTUNHFRT = sGi5Aql23May.Embedding(E1c_5v_Zd9l8, XUqE0GHx9ndw)
hiRoIvkoXk4B = sGi5Aql23May.Embedding(E1c_5v_Zd9l8, HHpQzG13Xmp7)
wqEu0qNdxfOe = sGi5Aql23May.Serial(SdSEWTUNHFRT, PhU_C1cgYIh2)
J_ePOe8eJp9N = sGi5Aql23May.Serial(hiRoIvkoXk4B, PhU_C1cgYIh2)
pTYjgQGkeuZ5 = sGi5Aql23May.MultiHeadedAttention(E1c_5v_Zd9l8, num_heads=vRVqPOZ1hUG7, dropout=ag0mwEgWzjYv, mode=holLFgwB7vsP)
@xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7o\x82\x8a\xaa\x80'), chr(0b1100100) + chr(101) + chr(99) + chr(7157 - 7046) + chr(0b1100100) + chr(101))(chr(117) + chr(116) + chr(0b111100 + 0o52) + '\x2d' + '\070'))
def NFMOEd9Vt1Jc(Qas9W3D0Xbzi, dR8yeI_DNUv3):
lM0wAWVIOgKw = sGi5Aql23May.Serial(sGi5Aql23May.Residual(sGi5Aql23May.LayerNorm(), sGi5Aql23May.Branch(size=ehT0Px3KOsy9(chr(48) + chr(0b1100010 + 0o15) + '\x34', 8)), sGi5Aql23May.Parallel(sGi5Aql23May.Identity(), sGi5Aql23May.Identity(), sGi5Aql23May.Identity(), dR8yeI_DNUv3), pTYjgQGkeuZ5, sGi5Aql23May.Dropout(ag0mwEgWzjYv, mode=holLFgwB7vsP)), Z01zxSkrfh7f(E1c_5v_Zd9l8, xXh07emJaiUD, ag0mwEgWzjYv, mode=holLFgwB7vsP))
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xa8k\x9d\x81\xaf\x8d'), chr(0b1000 + 0o134) + chr(5302 - 5201) + chr(0b11011 + 0o110) + '\157' + chr(0b110100 + 0o60) + chr(2562 - 2461))('\x75' + chr(116) + '\146' + chr(45) + chr(2483 - 2427)))(Qas9W3D0Xbzi, wqEu0qNdxfOe, xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89k\x9f\x8d\xaf\x95'), chr(100) + chr(0b1100101) + '\143' + '\157' + chr(100) + '\145')('\165' + chr(116) + chr(0b1100110) + chr(0b100011 + 0o12) + '\x38'))(lM0wAWVIOgKw, uftkTXJyNORO), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7o\x96\x8d\xbc\xaf\xd56\x8b'), '\144' + chr(0b1100101) + chr(99) + chr(111) + chr(100) + '\x65')('\x75' + '\x74' + chr(102) + '\055' + chr(0b101100 + 0o14)))())
@xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7o\x82\x8a\xaa\x80'), '\144' + chr(1416 - 1315) + chr(0b11010 + 0o111) + '\x6f' + chr(0b100 + 0o140) + '\x65')(chr(0b1110101) + '\x74' + chr(102) + chr(1371 - 1326) + chr(1945 - 1889)))
def PX4n8nzWQqG9(KcR7WgfLppqF, GR1581dR5rDS, ibztgOsJi_kJ, BtD8adbU5dQx):
GxnokQLrinKi = sGi5Aql23May.Serial(sGi5Aql23May.Residual(sGi5Aql23May.LayerNorm(), sGi5Aql23May.Branch(size=ehT0Px3KOsy9(chr(0b110000) + '\157' + '\064', 8)), sGi5Aql23May.Parallel(sGi5Aql23May.Identity(), sGi5Aql23May.Identity(), sGi5Aql23May.Identity(), ibztgOsJi_kJ), pTYjgQGkeuZ5, sGi5Aql23May.Dropout(ag0mwEgWzjYv, mode=holLFgwB7vsP)), sGi5Aql23May.Residual(sGi5Aql23May.LayerNorm(), sGi5Aql23May.Branch(size=ehT0Px3KOsy9('\060' + chr(0b110001 + 0o76) + chr(52), 8)), sGi5Aql23May.Parallel(sGi5Aql23May.Identity(), KcR7WgfLppqF, KcR7WgfLppqF, BtD8adbU5dQx), pTYjgQGkeuZ5, sGi5Aql23May.Dropout(ag0mwEgWzjYv, mode=holLFgwB7vsP)), Z01zxSkrfh7f(E1c_5v_Zd9l8, xXh07emJaiUD, ag0mwEgWzjYv, mode=holLFgwB7vsP))
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xa8k\x9d\x81\xaf\x8d'), '\144' + chr(0b11000 + 0o115) + '\x63' + chr(0b1101111) + '\144' + '\145')(chr(117) + chr(116) + chr(0b1000001 + 0o45) + '\x2d' + chr(0b11010 + 0o36)))(GR1581dR5rDS, J_ePOe8eJp9N, xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89k\x9f\x8d\xaf\x95'), '\144' + '\x65' + chr(0b0 + 0o143) + chr(111) + chr(0b111001 + 0o53) + chr(101))(chr(0b11010 + 0o133) + chr(877 - 761) + chr(102) + chr(45) + chr(56)))(GxnokQLrinKi, uftkTXJyNORO), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7o\x96\x8d\xbc\xaf\xd56\x8b'), '\144' + '\x65' + chr(0b1100011) + chr(111) + chr(100) + '\145')('\x75' + chr(0b1110100) + '\x66' + chr(45) + '\x38'))())
@xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7o\x82\x8a\xaa\x80'), chr(0b1100100) + '\145' + '\x63' + chr(9940 - 9829) + chr(100) + chr(9021 - 8920))('\x75' + chr(116) + '\146' + '\055' + chr(105 - 49)))
def Nk9m9eKr4iuF(Qas9W3D0Xbzi, GR1581dR5rDS, dR8yeI_DNUv3, ibztgOsJi_kJ, BtD8adbU5dQx):
bBsYDXWid6t7 = NFMOEd9Vt1Jc(Qas9W3D0Xbzi, dR8yeI_DNUv3)
return PX4n8nzWQqG9(bBsYDXWid6t7, GR1581dR5rDS, ibztgOsJi_kJ, BtD8adbU5dQx)
@xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7o\x82\x8a\xaa\x80'), chr(0b111111 + 0o45) + '\145' + chr(0b1100011) + chr(0b1101111) + '\x64' + '\x65')(chr(117) + chr(116) + chr(0b101101 + 0o71) + '\055' + chr(0b111000)))
def f1mM0j6jV_vs(U1521jDd3v5d):
return xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xa8k\x9d\x81\xaf\x8d'), '\x64' + chr(101) + chr(8128 - 8029) + chr(0b100101 + 0o112) + chr(7759 - 7659) + chr(0b10000 + 0o125))(chr(117) + '\x74' + chr(0b110000 + 0o66) + chr(45) + chr(0b101011 + 0o15)))(U1521jDd3v5d, xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xbfk\x81\x9b\xab'), chr(100) + '\145' + chr(7338 - 7239) + chr(111) + chr(100) + chr(9387 - 9286))(chr(117) + '\x74' + chr(102) + chr(0b10110 + 0o27) + '\070'))(HHpQzG13Xmp7), xafqLlk3kkUe(sGi5Aql23May, xafqLlk3kkUe(SXOLrMavuUCe(b'\xb7a\x88\xbb\xa1\x87\xce)\x87|'), chr(0b1100100) + chr(0b1100101) + '\143' + chr(3844 - 3733) + chr(0b1100100) + chr(101))(chr(0b100 + 0o161) + chr(116) + chr(102) + '\055' + '\070')))
(lZeqOzdNvpfO, PcgYsWhp9Z37) = f1mM0j6jV_vs(Nk9m9eKr4iuF)
if not HStumbLH46tl:
return (lZeqOzdNvpfO, PcgYsWhp9Z37)
else:
raise q1QCh3W88sgk(xafqLlk3kkUe(SXOLrMavuUCe(b"\x92`\x89\x8d\xbc\x84\xd4'\x83$\xcf\x84\xa6\x04\xcfi(\x04Rk\xa7h\rq\xdf\xc3\x89U\xd4\xbb\xb0\x86=\xcf\xdb\xb1K]\x89 \x92`\xcf\x98\xbc\x8e\xdd6\x83w\xd5"), chr(2529 - 2429) + chr(101) + chr(0b1100011) + chr(0b1101111) + chr(100) + '\x65')(chr(0b1110101) + chr(0b1011000 + 0o34) + '\146' + chr(419 - 374) + '\x38'))
|
tensorflow/tensor2tensor
|
tensor2tensor/models/mtf_transformer.py
|
mtf_transformer_base
|
def mtf_transformer_base():
"""Set of hyperparameters."""
hparams = common_hparams.basic_params1()
hparams.no_data_parallelism = True
hparams.use_fixed_batch_size = True
hparams.add_hparam("mtf_mode", True)
hparams.batch_size = 64
hparams.max_length = 256
hparams.add_hparam("d_model", 512)
hparams.add_hparam("d_kv", 128)
hparams.add_hparam("local_attention_window_size", 128)
hparams.label_smoothing = 0.1
# 8-way model-parallelism
hparams.add_hparam("mesh_shape", "model:8")
hparams.add_hparam("layout", "batch:batch;vocab:model;d_ff:model;heads:model")
hparams.add_hparam("num_heads", 8)
hparams.add_hparam("d_ff", 2048)
hparams.add_hparam("encoder_replicate_factor", 1)
hparams.add_hparam("decoder_replicate_factor", 1)
hparams.add_hparam("encoder_layers", ["att", "drd"] * 6)
hparams.add_hparam("decoder_layers", ["att", "enc_att", "drd"] * 6)
hparams.add_hparam("attention_dropout", 0.1)
hparams.add_hparam("relu_dropout", 0.1)
hparams.layer_prepostprocess_dropout = 0.1
# Describes what model architecture:
# "encdec": encoder + autoregressive decoder
# "decoder": single-stack autoregressive sequence model.
# "encoder": single-stack non-autoregressive model
# with equal-length inputs and outputs.
hparams.add_hparam("transformer_type", "encdec")
# What does the decoder do:
# "autoregressive": Decoder left to right
# "denoising": Fills in masked-out values simultaneously
hparams.add_hparam("decoder_type", "autoregressive")
# Parameters describing the noising algorithm for denoising decoders
hparams.add_hparam("noising_spec_train", {"type": "mask", "prob": 0.15})
hparams.add_hparam("noising_spec_eval", {"type": "mask", "prob": 0.15})
# during training, we use the eval noiser with this probability
hparams.add_hparam("noising_use_eval_during_train", 0.1)
# round up vocab sizes to be a multiple of this value
hparams.vocab_divisor = 128
# options are dense_relu_dense, moe, hmoe
hparams.add_hparam("feedforward_layer", "drd")
# If True, then reuse targets_embedding_var * rsqrt(d_model) as softmax_var
# If hparams.transformer_type == "encoder", then there is no targets embedding
# so we reuse the inputs embedding instead.
hparams.shared_embedding_and_softmax_weights = True
# Reuse targets_embedding_var as inputs_embedding_var
# relevant only if hparams.transformer_type == "encdec"
hparams.shared_embedding = True
hparams.optimizer = "Adafactor"
hparams.learning_rate_schedule = "linear_warmup*rsqrt_decay*linear_decay"
hparams.learning_rate_warmup_steps = 10000
hparams.add_hparam("master_dtype", "bfloat16")
hparams.add_hparam("slice_dtype", "float32")
hparams.activation_dtype = "bfloat16"
# These parameters make Transformer model compatible with MtfTransformer
# Do not override these, as mtf_transformer does not support other options.
hparams.clip_grad_norm = 0. # i.e. no gradient clipping
hparams.bottom = {
"inputs": modalities.identity_bottom,
"targets": modalities.identity_bottom,
}
hparams.top = {
"targets": modalities.identity_top,
}
# Parameters for computing the maximum decode length in beam search.
# Maximum decode length is:
# min(max_length,
# decode_length_multiplier * input_length + decode_length_constant)
hparams.add_hparam("decode_length_multiplier", 1.5)
hparams.add_hparam("decode_length_constant", 10.0)
# If nonzero, we split the batch across two tensor-dimensions named
# "outer_batch" and "inner_batch", allowing for splitting across two mesh
# dimensions. This is necessary for hierarchical mixture of experts.
# The two tensor dimensions have sizes hparams.outer_batch_size and
# hparams.batch_size // hparams.outer_batch_size.
hparams.add_hparam("outer_batch_size", 0)
# TODO(noam): file a bug
hparams.add_hparam("reshape_logits_hack", False)
hparams.add_hparam("compression_factor", 4)
return hparams
|
python
|
def mtf_transformer_base():
"""Set of hyperparameters."""
hparams = common_hparams.basic_params1()
hparams.no_data_parallelism = True
hparams.use_fixed_batch_size = True
hparams.add_hparam("mtf_mode", True)
hparams.batch_size = 64
hparams.max_length = 256
hparams.add_hparam("d_model", 512)
hparams.add_hparam("d_kv", 128)
hparams.add_hparam("local_attention_window_size", 128)
hparams.label_smoothing = 0.1
# 8-way model-parallelism
hparams.add_hparam("mesh_shape", "model:8")
hparams.add_hparam("layout", "batch:batch;vocab:model;d_ff:model;heads:model")
hparams.add_hparam("num_heads", 8)
hparams.add_hparam("d_ff", 2048)
hparams.add_hparam("encoder_replicate_factor", 1)
hparams.add_hparam("decoder_replicate_factor", 1)
hparams.add_hparam("encoder_layers", ["att", "drd"] * 6)
hparams.add_hparam("decoder_layers", ["att", "enc_att", "drd"] * 6)
hparams.add_hparam("attention_dropout", 0.1)
hparams.add_hparam("relu_dropout", 0.1)
hparams.layer_prepostprocess_dropout = 0.1
# Describes what model architecture:
# "encdec": encoder + autoregressive decoder
# "decoder": single-stack autoregressive sequence model.
# "encoder": single-stack non-autoregressive model
# with equal-length inputs and outputs.
hparams.add_hparam("transformer_type", "encdec")
# What does the decoder do:
# "autoregressive": Decoder left to right
# "denoising": Fills in masked-out values simultaneously
hparams.add_hparam("decoder_type", "autoregressive")
# Parameters describing the noising algorithm for denoising decoders
hparams.add_hparam("noising_spec_train", {"type": "mask", "prob": 0.15})
hparams.add_hparam("noising_spec_eval", {"type": "mask", "prob": 0.15})
# during training, we use the eval noiser with this probability
hparams.add_hparam("noising_use_eval_during_train", 0.1)
# round up vocab sizes to be a multiple of this value
hparams.vocab_divisor = 128
# options are dense_relu_dense, moe, hmoe
hparams.add_hparam("feedforward_layer", "drd")
# If True, then reuse targets_embedding_var * rsqrt(d_model) as softmax_var
# If hparams.transformer_type == "encoder", then there is no targets embedding
# so we reuse the inputs embedding instead.
hparams.shared_embedding_and_softmax_weights = True
# Reuse targets_embedding_var as inputs_embedding_var
# relevant only if hparams.transformer_type == "encdec"
hparams.shared_embedding = True
hparams.optimizer = "Adafactor"
hparams.learning_rate_schedule = "linear_warmup*rsqrt_decay*linear_decay"
hparams.learning_rate_warmup_steps = 10000
hparams.add_hparam("master_dtype", "bfloat16")
hparams.add_hparam("slice_dtype", "float32")
hparams.activation_dtype = "bfloat16"
# These parameters make Transformer model compatible with MtfTransformer
# Do not override these, as mtf_transformer does not support other options.
hparams.clip_grad_norm = 0. # i.e. no gradient clipping
hparams.bottom = {
"inputs": modalities.identity_bottom,
"targets": modalities.identity_bottom,
}
hparams.top = {
"targets": modalities.identity_top,
}
# Parameters for computing the maximum decode length in beam search.
# Maximum decode length is:
# min(max_length,
# decode_length_multiplier * input_length + decode_length_constant)
hparams.add_hparam("decode_length_multiplier", 1.5)
hparams.add_hparam("decode_length_constant", 10.0)
# If nonzero, we split the batch across two tensor-dimensions named
# "outer_batch" and "inner_batch", allowing for splitting across two mesh
# dimensions. This is necessary for hierarchical mixture of experts.
# The two tensor dimensions have sizes hparams.outer_batch_size and
# hparams.batch_size // hparams.outer_batch_size.
hparams.add_hparam("outer_batch_size", 0)
# TODO(noam): file a bug
hparams.add_hparam("reshape_logits_hack", False)
hparams.add_hparam("compression_factor", 4)
return hparams
|
[
"def",
"mtf_transformer_base",
"(",
")",
":",
"hparams",
"=",
"common_hparams",
".",
"basic_params1",
"(",
")",
"hparams",
".",
"no_data_parallelism",
"=",
"True",
"hparams",
".",
"use_fixed_batch_size",
"=",
"True",
"hparams",
".",
"add_hparam",
"(",
"\"mtf_mode\"",
",",
"True",
")",
"hparams",
".",
"batch_size",
"=",
"64",
"hparams",
".",
"max_length",
"=",
"256",
"hparams",
".",
"add_hparam",
"(",
"\"d_model\"",
",",
"512",
")",
"hparams",
".",
"add_hparam",
"(",
"\"d_kv\"",
",",
"128",
")",
"hparams",
".",
"add_hparam",
"(",
"\"local_attention_window_size\"",
",",
"128",
")",
"hparams",
".",
"label_smoothing",
"=",
"0.1",
"# 8-way model-parallelism",
"hparams",
".",
"add_hparam",
"(",
"\"mesh_shape\"",
",",
"\"model:8\"",
")",
"hparams",
".",
"add_hparam",
"(",
"\"layout\"",
",",
"\"batch:batch;vocab:model;d_ff:model;heads:model\"",
")",
"hparams",
".",
"add_hparam",
"(",
"\"num_heads\"",
",",
"8",
")",
"hparams",
".",
"add_hparam",
"(",
"\"d_ff\"",
",",
"2048",
")",
"hparams",
".",
"add_hparam",
"(",
"\"encoder_replicate_factor\"",
",",
"1",
")",
"hparams",
".",
"add_hparam",
"(",
"\"decoder_replicate_factor\"",
",",
"1",
")",
"hparams",
".",
"add_hparam",
"(",
"\"encoder_layers\"",
",",
"[",
"\"att\"",
",",
"\"drd\"",
"]",
"*",
"6",
")",
"hparams",
".",
"add_hparam",
"(",
"\"decoder_layers\"",
",",
"[",
"\"att\"",
",",
"\"enc_att\"",
",",
"\"drd\"",
"]",
"*",
"6",
")",
"hparams",
".",
"add_hparam",
"(",
"\"attention_dropout\"",
",",
"0.1",
")",
"hparams",
".",
"add_hparam",
"(",
"\"relu_dropout\"",
",",
"0.1",
")",
"hparams",
".",
"layer_prepostprocess_dropout",
"=",
"0.1",
"# Describes what model architecture:",
"# \"encdec\": encoder + autoregressive decoder",
"# \"decoder\": single-stack autoregressive sequence model.",
"# \"encoder\": single-stack non-autoregressive model",
"# with equal-length inputs and outputs.",
"hparams",
".",
"add_hparam",
"(",
"\"transformer_type\"",
",",
"\"encdec\"",
")",
"# What does the decoder do:",
"# \"autoregressive\": Decoder left to right",
"# \"denoising\": Fills in masked-out values simultaneously",
"hparams",
".",
"add_hparam",
"(",
"\"decoder_type\"",
",",
"\"autoregressive\"",
")",
"# Parameters describing the noising algorithm for denoising decoders",
"hparams",
".",
"add_hparam",
"(",
"\"noising_spec_train\"",
",",
"{",
"\"type\"",
":",
"\"mask\"",
",",
"\"prob\"",
":",
"0.15",
"}",
")",
"hparams",
".",
"add_hparam",
"(",
"\"noising_spec_eval\"",
",",
"{",
"\"type\"",
":",
"\"mask\"",
",",
"\"prob\"",
":",
"0.15",
"}",
")",
"# during training, we use the eval noiser with this probability",
"hparams",
".",
"add_hparam",
"(",
"\"noising_use_eval_during_train\"",
",",
"0.1",
")",
"# round up vocab sizes to be a multiple of this value",
"hparams",
".",
"vocab_divisor",
"=",
"128",
"# options are dense_relu_dense, moe, hmoe",
"hparams",
".",
"add_hparam",
"(",
"\"feedforward_layer\"",
",",
"\"drd\"",
")",
"# If True, then reuse targets_embedding_var * rsqrt(d_model) as softmax_var",
"# If hparams.transformer_type == \"encoder\", then there is no targets embedding",
"# so we reuse the inputs embedding instead.",
"hparams",
".",
"shared_embedding_and_softmax_weights",
"=",
"True",
"# Reuse targets_embedding_var as inputs_embedding_var",
"# relevant only if hparams.transformer_type == \"encdec\"",
"hparams",
".",
"shared_embedding",
"=",
"True",
"hparams",
".",
"optimizer",
"=",
"\"Adafactor\"",
"hparams",
".",
"learning_rate_schedule",
"=",
"\"linear_warmup*rsqrt_decay*linear_decay\"",
"hparams",
".",
"learning_rate_warmup_steps",
"=",
"10000",
"hparams",
".",
"add_hparam",
"(",
"\"master_dtype\"",
",",
"\"bfloat16\"",
")",
"hparams",
".",
"add_hparam",
"(",
"\"slice_dtype\"",
",",
"\"float32\"",
")",
"hparams",
".",
"activation_dtype",
"=",
"\"bfloat16\"",
"# These parameters make Transformer model compatible with MtfTransformer",
"# Do not override these, as mtf_transformer does not support other options.",
"hparams",
".",
"clip_grad_norm",
"=",
"0.",
"# i.e. no gradient clipping",
"hparams",
".",
"bottom",
"=",
"{",
"\"inputs\"",
":",
"modalities",
".",
"identity_bottom",
",",
"\"targets\"",
":",
"modalities",
".",
"identity_bottom",
",",
"}",
"hparams",
".",
"top",
"=",
"{",
"\"targets\"",
":",
"modalities",
".",
"identity_top",
",",
"}",
"# Parameters for computing the maximum decode length in beam search.",
"# Maximum decode length is:",
"# min(max_length,",
"# decode_length_multiplier * input_length + decode_length_constant)",
"hparams",
".",
"add_hparam",
"(",
"\"decode_length_multiplier\"",
",",
"1.5",
")",
"hparams",
".",
"add_hparam",
"(",
"\"decode_length_constant\"",
",",
"10.0",
")",
"# If nonzero, we split the batch across two tensor-dimensions named",
"# \"outer_batch\" and \"inner_batch\", allowing for splitting across two mesh",
"# dimensions. This is necessary for hierarchical mixture of experts.",
"# The two tensor dimensions have sizes hparams.outer_batch_size and",
"# hparams.batch_size // hparams.outer_batch_size.",
"hparams",
".",
"add_hparam",
"(",
"\"outer_batch_size\"",
",",
"0",
")",
"# TODO(noam): file a bug",
"hparams",
".",
"add_hparam",
"(",
"\"reshape_logits_hack\"",
",",
"False",
")",
"hparams",
".",
"add_hparam",
"(",
"\"compression_factor\"",
",",
"4",
")",
"return",
"hparams"
] |
Set of hyperparameters.
|
[
"Set",
"of",
"hyperparameters",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/mtf_transformer.py#L791-L883
|
train
|
Set of hyperparameters for Mt - Transformer model.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b1101111) + chr(0b11000 + 0o33) + chr(2442 - 2389) + '\x30', 12051 - 12043), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(988 - 937) + chr(1571 - 1522) + chr(0b110000), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101 + 0o142) + chr(0b110010) + chr(0b1001 + 0o53) + '\x36', 47690 - 47682), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\062' + chr(0b11010 + 0o34), 22290 - 22282), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(5443 - 5332) + chr(0b110001) + chr(0b110000) + chr(2829 - 2775), 0b1000), ehT0Px3KOsy9(chr(0b11001 + 0o27) + '\x6f' + chr(49) + chr(51) + chr(52), 22455 - 22447), ehT0Px3KOsy9(chr(1734 - 1686) + chr(5022 - 4911) + '\x32' + chr(1685 - 1630) + chr(278 - 228), 14409 - 14401), ehT0Px3KOsy9(chr(0b11010 + 0o26) + chr(4620 - 4509) + '\x31' + '\x30' + chr(0b100000 + 0o20), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110010) + chr(1594 - 1541) + '\063', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(1634 - 1523) + chr(426 - 376) + '\064' + chr(55), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110010) + '\062' + chr(0b110010), 57555 - 57547), ehT0Px3KOsy9(chr(402 - 354) + chr(0b110101 + 0o72) + chr(0b110000 + 0o3) + '\x30' + '\x32', 46246 - 46238), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b1 + 0o60) + chr(50), 0o10), ehT0Px3KOsy9(chr(0b100110 + 0o12) + '\157' + '\x33' + chr(639 - 588) + chr(55), 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + chr(681 - 631) + '\x32', 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010) + chr(0b110 + 0o57) + '\066', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101011 + 0o4) + chr(1889 - 1840) + chr(0b110000) + chr(0b11000 + 0o30), 8), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b101111 + 0o3) + chr(0b110000) + chr(55), 23889 - 23881), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110001) + chr(55) + '\064', 20922 - 20914), ehT0Px3KOsy9('\x30' + chr(3821 - 3710) + chr(0b110001) + chr(0b110101) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b110010 + 0o75) + '\067' + chr(51), 1897 - 1889), ehT0Px3KOsy9('\060' + chr(0b100011 + 0o114) + '\061' + chr(49) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b1110 + 0o45) + '\x32' + chr(0b10000 + 0o40), 0b1000), ehT0Px3KOsy9(chr(0b11010 + 0o26) + chr(1328 - 1217) + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(50) + '\x31' + chr(50), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(49) + chr(48) + chr(174 - 119), ord("\x08")), ehT0Px3KOsy9(chr(0b11100 + 0o24) + '\157' + '\063' + chr(0b100 + 0o55) + chr(55), 53815 - 53807), ehT0Px3KOsy9(chr(0b111 + 0o51) + chr(11494 - 11383) + chr(0b100111 + 0o12) + chr(54) + '\064', 16747 - 16739), ehT0Px3KOsy9('\060' + chr(0b100100 + 0o113) + '\x32' + '\064', 0o10), ehT0Px3KOsy9(chr(48) + chr(888 - 777) + '\x32' + chr(0b10010 + 0o36) + '\x36', 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + '\x33' + chr(50), 0o10), ehT0Px3KOsy9(chr(712 - 664) + '\x6f' + '\061' + chr(0b110001) + chr(0b1111 + 0o42), 41254 - 41246), ehT0Px3KOsy9(chr(0b110000) + chr(0b1000001 + 0o56) + chr(50) + '\x32' + chr(51), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\063' + chr(177 - 126), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + chr(0b110111) + chr(49), 0b1000), ehT0Px3KOsy9(chr(1422 - 1374) + '\x6f' + chr(53) + chr(0b100 + 0o62), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(55), 9922 - 9914), ehT0Px3KOsy9(chr(0b110000) + chr(0b101001 + 0o106) + chr(0b100000 + 0o21) + chr(552 - 502) + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(318 - 269) + chr(0b110010 + 0o5) + '\x36', 0b1000), ehT0Px3KOsy9(chr(0b11110 + 0o22) + '\157' + chr(1364 - 1313) + chr(87 - 34) + chr(0b101101 + 0o3), 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(53) + chr(0b110000), 51974 - 51966)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xc6'), chr(0b11100 + 0o110) + chr(0b1010011 + 0o22) + '\143' + chr(0b1000 + 0o147) + chr(0b1100100) + chr(0b1010001 + 0o24))(chr(0b1011001 + 0o34) + chr(0b1110100) + '\x66' + chr(0b101101) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def Hc6ONEMgq9eN():
n4ljua2gi1Pr = vLnG3ZpOXWXZ.basic_params1()
n4ljua2gi1Pr.ahN6YYm9NJTr = ehT0Px3KOsy9(chr(2246 - 2198) + chr(0b110011 + 0o74) + chr(0b110001), 0o10)
n4ljua2gi1Pr.V9YwhDsFOlGK = ehT0Px3KOsy9('\x30' + '\x6f' + chr(849 - 800), 8)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\144' + chr(0b10011 + 0o122) + chr(0b1100011) + chr(0b1011011 + 0o24) + chr(100) + '\x65')(chr(10244 - 10127) + chr(0b1110100) + chr(0b1100110) + chr(1177 - 1132) + chr(0b101010 + 0o16)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x85\xbc\xc0I+3\xde%'), chr(0b1100100) + '\x65' + chr(0b1100011) + chr(9612 - 9501) + chr(0b111010 + 0o52) + chr(9537 - 9436))('\165' + '\x74' + '\146' + chr(0b11011 + 0o22) + '\070'), ehT0Px3KOsy9(chr(48) + chr(111) + '\061', 8))
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(49) + chr(0b101010 + 0o6) + chr(48), 8)
n4ljua2gi1Pr._o7pVXAdOCRy = ehT0Px3KOsy9(chr(786 - 738) + chr(1284 - 1173) + chr(0b110100) + chr(0b110000) + '\x30', ord("\x08"))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(8240 - 8140) + chr(3770 - 3669) + '\143' + chr(111) + chr(0b1011101 + 0o7) + chr(0b111111 + 0o46))(chr(117) + chr(116) + chr(102) + '\x2d' + chr(0b100000 + 0o30)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\x97\xcby"9\xd6'), chr(0b1000011 + 0o41) + '\x65' + chr(3593 - 3494) + '\157' + chr(3907 - 3807) + chr(4405 - 4304))(chr(11976 - 11859) + chr(116) + '\146' + '\x2d' + '\070'), ehT0Px3KOsy9(chr(0b10011 + 0o35) + chr(8612 - 8501) + chr(872 - 823) + '\x30' + '\060' + chr(48), 0o10))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b1100100) + '\x65' + chr(0b1011100 + 0o7) + chr(0b1101011 + 0o4) + '\144' + '\x65')(chr(13371 - 13254) + chr(11216 - 11100) + '\146' + '\x2d' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\x97\xcd`'), chr(100) + chr(4773 - 4672) + '\x63' + chr(3874 - 3763) + '\x64' + chr(0b1100101))(chr(12134 - 12017) + chr(0b1110100) + chr(0b11 + 0o143) + chr(0b101101) + '\070'), ehT0Px3KOsy9('\060' + chr(0b11001 + 0o126) + '\x32' + chr(0b110000) + chr(48), 61190 - 61182))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\x64' + '\x65' + chr(99) + chr(0b1101111) + chr(0b1100100) + '\x65')(chr(117) + '\x74' + '\x66' + chr(0b100011 + 0o12) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x84\xa7\xc5w*\x03\xdb4\xdf\xc2\xfc\xb6c\xb0\xfc\xf17\xee\x12\tT9\x8fw\xe3=\xaa'), chr(0b1100100) + chr(101) + chr(99) + '\157' + chr(6518 - 6418) + chr(0b1100101))(chr(13615 - 13498) + '\x74' + '\146' + chr(436 - 391) + '\x38'), ehT0Px3KOsy9(chr(0b100100 + 0o14) + '\157' + chr(50) + '\x30' + '\060', 8))
n4ljua2gi1Pr.FSjUgdaczzRk = 0.1
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\144' + '\145' + chr(1483 - 1384) + chr(0b10000 + 0o137) + '\x64' + chr(101))(chr(117) + '\x74' + chr(0b111110 + 0o50) + chr(0b1001 + 0o44) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x85\xad\xd5~\x19/\xd2!\xdb\xc2'), '\144' + chr(0b1101 + 0o130) + chr(3036 - 2937) + '\x6f' + '\144' + chr(0b11000 + 0o115))(chr(9244 - 9127) + chr(0b110111 + 0o75) + chr(102) + chr(0b101101) + '\070'), xafqLlk3kkUe(SXOLrMavuUCe(b'\x85\xa7\xc2s*f\x82'), chr(1769 - 1669) + chr(101) + chr(99) + chr(6307 - 6196) + chr(0b101011 + 0o71) + chr(101))(chr(0b1110101) + '\164' + '\x66' + '\055' + chr(0b111000)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\144' + chr(101) + chr(0b1100011) + '\x6f' + chr(0b1100100) + '\145')(chr(0b1110101) + '\x74' + '\x66' + chr(1081 - 1036) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x84\xa9\xdfy3('), chr(100) + chr(0b1100101) + chr(0b1100011) + '\157' + chr(100) + '\145')(chr(117) + chr(116) + chr(0b1100110) + '\x2d' + chr(0b111000)), xafqLlk3kkUe(SXOLrMavuUCe(b'\x8a\xa9\xd2u.f\xd8!\xdf\xc4\xfa\xf9|\xb0\xf1\xcf"\xbd\x11\x02_+\xbc?\xee\x18\xa9\x12.\xeb\xf5-\xf6w\x82Fb0\x1a\x9b\xd2\xa5\xc9r#0'), '\144' + chr(0b1100101) + chr(99) + chr(0b1101111) + '\144' + chr(0b10111 + 0o116))(chr(877 - 760) + chr(0b1110100) + '\x66' + chr(45) + chr(2131 - 2075)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b1001111 + 0o25) + '\x65' + chr(9384 - 9285) + chr(0b101000 + 0o107) + chr(100) + chr(0b101000 + 0o75))(chr(0b1110101) + chr(0b1111 + 0o145) + chr(0b1001010 + 0o34) + '\055' + chr(2893 - 2837)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x86\xbd\xcbI.9\xdb$\xd8'), chr(100) + chr(101) + chr(99) + chr(0b1101111) + chr(100) + chr(101))('\x75' + chr(0b1110100) + '\x66' + chr(0b1100 + 0o41) + chr(0b0 + 0o70)), ehT0Px3KOsy9(chr(915 - 867) + chr(6515 - 6404) + chr(0b110000 + 0o1) + chr(48), 0b1000))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\144' + chr(0b11 + 0o142) + chr(6642 - 6543) + chr(0b1010 + 0o145) + chr(6680 - 6580) + chr(2440 - 2339))(chr(0b10000 + 0o145) + chr(116) + chr(102) + chr(0b101101) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\x97\xc0p'), chr(2647 - 2547) + '\145' + chr(99) + chr(0b1101101 + 0o2) + chr(0b11010 + 0o112) + chr(101))(chr(117) + chr(0b110111 + 0o75) + chr(0b1011000 + 0o16) + chr(0b101101) + chr(56)), ehT0Px3KOsy9(chr(85 - 37) + '\x6f' + chr(0b110100) + '\060' + chr(48) + chr(48), 0o10))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(5194 - 5094) + '\x65' + chr(0b1011001 + 0o12) + '\157' + chr(0b1001001 + 0o33) + chr(101))(chr(12325 - 12208) + chr(0b1001000 + 0o54) + '\x66' + chr(45) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8d\xa6\xc5y"9\xc8\x1f\xd9\xc2\xe2\xaec\xbc\xf3\xda%\xd8\x1a\x0cX:\xbfv'), chr(100) + chr(0b11010 + 0o113) + chr(0b1100011) + '\157' + chr(0b10000 + 0o124) + chr(101))(chr(10271 - 10154) + chr(340 - 224) + chr(3841 - 3739) + chr(45) + '\070'), ehT0Px3KOsy9(chr(0b1101 + 0o43) + '\157' + chr(0b110001), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b1010010 + 0o22) + chr(101) + '\x63' + '\x6f' + chr(0b1100100) + chr(101))(chr(0b1101001 + 0o14) + chr(116) + '\x66' + chr(0b101000 + 0o5) + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\xad\xc5y"9\xc8\x1f\xd9\xc2\xe2\xaec\xbc\xf3\xda%\xd8\x1a\x0cX:\xbfv'), chr(100) + '\145' + chr(0b11000 + 0o113) + chr(3873 - 3762) + '\144' + '\145')(chr(0b101000 + 0o115) + chr(13361 - 13245) + '\146' + '\x2d' + '\070'), ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b110111 + 0o70) + '\061', 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b1001100 + 0o30) + chr(0b1011111 + 0o6) + chr(8012 - 7913) + chr(111) + chr(1552 - 1452) + '\145')(chr(0b1110101) + chr(8546 - 8430) + chr(0b10100 + 0o122) + chr(45) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8d\xa6\xc5y"9\xc8\x1f\xc7\xc6\xeb\xa7x\xac'), chr(100) + '\x65' + chr(1222 - 1123) + chr(0b1101111) + chr(4277 - 4177) + '\x65')(chr(12918 - 12801) + '\x74' + chr(0b1100110) + chr(0b100010 + 0o13) + chr(56)), [xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xbc\xd2'), chr(0b10011 + 0o121) + '\145' + '\x63' + '\157' + chr(0b1100100) + chr(1705 - 1604))(chr(1844 - 1727) + '\x74' + '\x66' + chr(576 - 531) + chr(0b111000)), xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\xba\xc2'), chr(0b101101 + 0o67) + '\145' + chr(0b11 + 0o140) + '\157' + chr(6145 - 6045) + '\145')(chr(13299 - 13182) + chr(0b1110100) + chr(10082 - 9980) + '\x2d' + chr(56))] * ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(2125 - 2071), 0o10))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b10011 + 0o121) + chr(101) + '\143' + chr(0b10111 + 0o130) + chr(2815 - 2715) + chr(101))(chr(0b1110101) + chr(6497 - 6381) + '\146' + chr(1198 - 1153) + chr(1928 - 1872)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\xad\xc5y"9\xc8\x1f\xc7\xc6\xeb\xa7x\xac'), chr(100) + chr(0b1100101) + chr(0b1100011) + chr(10764 - 10653) + '\x64' + '\x65')(chr(0b1110101) + chr(0b101 + 0o157) + chr(9842 - 9740) + chr(45) + chr(178 - 122)), [xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xbc\xd2'), '\144' + '\145' + chr(2742 - 2643) + chr(0b1101111) + chr(100) + '\x65')(chr(0b10001 + 0o144) + '\164' + chr(102) + chr(0b101101) + chr(56)), xafqLlk3kkUe(SXOLrMavuUCe(b"\x8d\xa6\xc5I'(\xce"), '\x64' + '\x65' + chr(99) + chr(0b1101111) + chr(7964 - 7864) + chr(0b1001111 + 0o26))(chr(0b1110101) + chr(116) + chr(0b1100110) + chr(0b101101) + '\x38'), xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\xba\xc2'), chr(0b101011 + 0o71) + chr(101) + chr(0b1000111 + 0o34) + chr(0b1101111) + '\144' + chr(101))(chr(0b1110101) + '\x74' + chr(0b11011 + 0o113) + '\x2d' + chr(0b111000))] * ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b100110 + 0o111) + chr(0b110110), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b10100 + 0o120) + chr(0b111 + 0o136) + chr(0b1100011) + '\x6f' + '\x64' + chr(0b1010001 + 0o24))('\165' + chr(0b1110100) + chr(0b1100110) + '\055' + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xbc\xd2s((\xd3/\xc5\xf8\xf6\xb0e\xaf\xfd\xdb4'), '\144' + chr(0b101 + 0o140) + chr(0b1100011) + chr(0b1101111) + chr(3245 - 3145) + chr(0b1100101))(chr(0b1110101) + chr(0b1110100) + chr(9827 - 9725) + chr(45) + chr(56)), 0.1)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(100) + chr(0b111011 + 0o52) + chr(9150 - 9051) + chr(10171 - 10060) + chr(0b1100100) + '\x65')('\x75' + '\164' + chr(3567 - 3465) + chr(0b100110 + 0o7) + chr(0b101100 + 0o14)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x9a\xad\xcac\x198\xc8/\xdb\xc8\xe7\xb6'), '\x64' + chr(0b11101 + 0o110) + chr(0b1100011) + chr(9430 - 9319) + chr(1746 - 1646) + chr(101))(chr(0b1110101) + '\164' + '\146' + chr(45) + chr(0b1000 + 0o60)), 0.1)
n4ljua2gi1Pr.RW_xSzp18UeS = 0.1
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\x64' + '\x65' + '\x63' + chr(0b1101111) + chr(100) + chr(101))(chr(11464 - 11347) + chr(0b1101111 + 0o5) + '\x66' + chr(0b10100 + 0o31) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x9c\xba\xc7x5:\xd52\xc6\xc2\xe0\x9d~\xa6\xe2\xcb'), chr(0b1100100) + chr(101) + chr(99) + '\x6f' + chr(100) + '\x65')(chr(11207 - 11090) + chr(0b1110100) + chr(0b11001 + 0o115) + chr(0b10 + 0o53) + chr(56)), xafqLlk3kkUe(SXOLrMavuUCe(b'\x8d\xa6\xc5r#?'), chr(0b1100100) + chr(8775 - 8674) + chr(2908 - 2809) + '\157' + '\x64' + chr(101))('\165' + '\x74' + chr(0b1100110) + '\x2d' + chr(84 - 28)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b1100100) + chr(101) + chr(9157 - 9058) + chr(4911 - 4800) + chr(9606 - 9506) + chr(5019 - 4918))(chr(3787 - 3670) + chr(0b10001 + 0o143) + '\146' + chr(0b100 + 0o51) + chr(855 - 799)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\xad\xc5y"9\xc8\x1f\xdf\xde\xe2\xa7'), chr(0b1100100) + '\145' + chr(0b110110 + 0o55) + chr(111) + chr(100) + chr(0b100011 + 0o102))('\165' + chr(1849 - 1733) + chr(0b1100110) + chr(0b101011 + 0o2) + '\x38'), xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xbd\xd2y49\xdd2\xce\xd4\xe1\xab|\xba'), chr(7776 - 7676) + '\145' + chr(0b1100011) + chr(0b1011001 + 0o26) + chr(100) + chr(101))(chr(1517 - 1400) + '\164' + chr(1799 - 1697) + chr(45) + chr(1145 - 1089)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b10 + 0o142) + chr(2963 - 2862) + chr(0b1110 + 0o125) + chr(3125 - 3014) + chr(100) + chr(101))(chr(117) + chr(116) + '\x66' + chr(783 - 738) + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x86\xa7\xcfe/2\xdd\x1f\xd8\xd7\xf7\xa1U\xab\xe0\xcf)\xe9'), '\x64' + '\x65' + '\x63' + '\x6f' + chr(0b1100100) + chr(101))('\x75' + chr(0b1 + 0o163) + '\146' + chr(1739 - 1694) + '\x38'), {xafqLlk3kkUe(SXOLrMavuUCe(b'\x9c\xb1\xd6s'), chr(0b1100100) + chr(101) + chr(0b1100011) + chr(0b101011 + 0o104) + '\144' + chr(7727 - 7626))('\165' + '\x74' + chr(102) + chr(45) + chr(0b111000)): xafqLlk3kkUe(SXOLrMavuUCe(b'\x85\xa9\xd5}'), chr(0b1100100) + '\x65' + '\143' + chr(0b1101111) + chr(9505 - 9405) + chr(0b101010 + 0o73))('\165' + '\x74' + chr(0b1100110) + chr(45) + '\x38'), xafqLlk3kkUe(SXOLrMavuUCe(b'\x98\xba\xc9t'), chr(0b11100 + 0o110) + chr(0b1100101) + '\x63' + '\x6f' + '\144' + chr(8822 - 8721))(chr(0b1110101) + chr(0b1110100) + chr(9306 - 9204) + chr(0b101101) + chr(2838 - 2782)): 0.15})
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b0 + 0o144) + chr(7265 - 7164) + chr(0b1100011) + '\x6f' + '\144' + chr(0b11110 + 0o107))(chr(117) + '\164' + chr(0b1100110) + chr(45) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x86\xa7\xcfe/2\xdd\x1f\xd8\xd7\xf7\xa1U\xba\xe4\xcf,'), chr(0b1100100) + chr(0b1100101) + chr(99) + '\157' + chr(3875 - 3775) + chr(0b1100101))('\x75' + '\164' + chr(0b1001010 + 0o34) + chr(0b100110 + 0o7) + chr(56)), {xafqLlk3kkUe(SXOLrMavuUCe(b'\x9c\xb1\xd6s'), '\144' + '\145' + chr(3025 - 2926) + '\x6f' + chr(0b1100100) + chr(1002 - 901))(chr(0b1000000 + 0o65) + chr(0b1110100) + chr(0b1100110) + chr(45) + chr(0b111000)): xafqLlk3kkUe(SXOLrMavuUCe(b'\x85\xa9\xd5}'), chr(100) + chr(101) + chr(0b1100011) + chr(0b1010100 + 0o33) + chr(9725 - 9625) + '\145')(chr(0b101111 + 0o106) + '\164' + '\x66' + '\x2d' + chr(56)), xafqLlk3kkUe(SXOLrMavuUCe(b'\x98\xba\xc9t'), chr(0b1100100) + '\x65' + chr(0b1100011) + chr(0b1101111) + chr(0b1100100) + chr(4265 - 4164))('\x75' + '\164' + '\146' + '\x2d' + chr(0b101111 + 0o11)): 0.15})
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\x64' + chr(0b1100101) + '\143' + chr(0b1101111) + chr(390 - 290) + '\x65')(chr(5908 - 5791) + chr(116) + chr(6438 - 6336) + '\055' + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x86\xa7\xcfe/2\xdd\x1f\xde\xd4\xf7\x9do\xa9\xf3\xc2\x1f\xe3\t\x1fR \xb7[\xfe5\xae\x1dz'), chr(0b1011010 + 0o12) + chr(0b1001001 + 0o34) + chr(0b11100 + 0o107) + chr(10794 - 10683) + chr(0b1100100) + chr(101))(chr(0b1110011 + 0o2) + chr(116) + chr(8653 - 8551) + chr(45) + chr(0b1111 + 0o51)), 0.1)
n4ljua2gi1Pr.uDahGBSW4HJn = ehT0Px3KOsy9('\x30' + '\157' + '\x32' + '\x30' + '\x30', 8)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\x64' + chr(101) + chr(7181 - 7082) + '\157' + chr(0b111111 + 0o45) + chr(101))('\165' + chr(2518 - 2402) + chr(102) + chr(0b101101) + chr(0b1110 + 0o52)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8e\xad\xc3r 3\xc87\xca\xd5\xf6\x9df\xbe\xeb\xcb2'), chr(100) + '\x65' + '\x63' + chr(0b1101111) + chr(0b100110 + 0o76) + chr(0b1101 + 0o130))('\x75' + chr(0b1010110 + 0o36) + '\x66' + chr(195 - 150) + chr(404 - 348)), xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\xba\xc2'), chr(100) + chr(0b1100101) + chr(0b1100011) + chr(10724 - 10613) + chr(100) + chr(0b1100101))(chr(5814 - 5697) + chr(0b11 + 0o161) + chr(0b11101 + 0o111) + chr(0b100111 + 0o6) + '\070'))
n4ljua2gi1Pr.qVamxim0L2I1 = ehT0Px3KOsy9('\x30' + chr(111) + chr(139 - 90), 8)
n4ljua2gi1Pr.f7bdxoAgzo_R = ehT0Px3KOsy9(chr(0b101101 + 0o3) + chr(10399 - 10288) + chr(49), 8)
n4ljua2gi1Pr.XdKNcYRObPK3 = xafqLlk3kkUe(SXOLrMavuUCe(b"\xa9\xac\xc7p'?\xce/\xd9"), '\144' + '\145' + chr(99) + chr(0b1101111) + chr(5017 - 4917) + chr(0b1100101))(chr(4656 - 4539) + '\164' + chr(0b1100110) + chr(0b101101) + chr(0b111000))
n4ljua2gi1Pr.Lz_s7neUzM5V = xafqLlk3kkUe(SXOLrMavuUCe(b"\x84\xa1\xc8s'.\xe57\xca\xd5\xff\xb7z\xf5\xe0\xdd1\xf5\x082_+\xb3e\xf3m\xa3\x1dz\xe3\xfb;\xcc\x7f\xdcMf("), chr(100) + '\x65' + chr(99) + chr(4789 - 4678) + chr(9566 - 9466) + '\x65')(chr(0b1110101) + chr(3458 - 3342) + '\x66' + chr(0b101100 + 0o1) + chr(0b110001 + 0o7))
n4ljua2gi1Pr.fHyhoyGmdvM9 = ehT0Px3KOsy9(chr(684 - 636) + chr(0b1011001 + 0o26) + chr(50) + chr(0b1 + 0o62) + chr(0b110100) + chr(0b110010) + chr(48), ord("\x08"))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b1000110 + 0o36) + '\x65' + chr(99) + chr(2008 - 1897) + chr(4116 - 4016) + chr(101))(chr(0b1110101) + '\164' + chr(6433 - 6331) + chr(45) + chr(969 - 913)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x85\xa9\xd5b#.\xe5$\xdf\xde\xe2\xa7'), chr(100) + chr(7215 - 7114) + chr(0b1010101 + 0o16) + chr(111) + '\x64' + chr(101))(chr(0b11010 + 0o133) + chr(116) + chr(0b1111 + 0o127) + chr(509 - 464) + '\070'), xafqLlk3kkUe(SXOLrMavuUCe(b"\x8a\xae\xcay'(\x8bv"), chr(100) + chr(0b1100101) + chr(0b1011 + 0o130) + chr(0b10 + 0o155) + chr(0b10100 + 0o120) + chr(3302 - 3201))(chr(5162 - 5045) + chr(0b1110100) + '\146' + chr(1742 - 1697) + chr(56)))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b100 + 0o140) + '\145' + '\x63' + chr(0b1101111) + chr(0b1010010 + 0o22) + chr(0b10 + 0o143))(chr(117) + chr(3923 - 3807) + chr(0b100101 + 0o101) + '\x2d' + '\070'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x9b\xa4\xcfu#\x03\xde4\xd2\xd7\xf7'), chr(6142 - 6042) + '\145' + chr(0b1100011) + '\157' + '\144' + chr(2261 - 2160))(chr(0b1110101) + chr(8275 - 8159) + chr(6503 - 6401) + '\055' + chr(0b111000)), xafqLlk3kkUe(SXOLrMavuUCe(b'\x8e\xa4\xc9w2o\x88'), '\144' + '\x65' + chr(0b1100011) + '\157' + chr(2505 - 2405) + '\145')('\x75' + chr(116) + '\146' + '\055' + chr(2476 - 2420)))
n4ljua2gi1Pr.n6ZCgJ7AKd3U = xafqLlk3kkUe(SXOLrMavuUCe(b"\x8a\xae\xcay'(\x8bv"), '\144' + chr(101) + chr(99) + chr(5473 - 5362) + chr(0b1100100) + '\x65')(chr(117) + '\x74' + chr(102) + chr(45) + chr(56))
n4ljua2gi1Pr.SdNSZNVkVjLh = 0.0
n4ljua2gi1Pr.kXxsZxlIQUSQ = {xafqLlk3kkUe(SXOLrMavuUCe(b'\x81\xa6\xd6c2/'), chr(100) + chr(0b1100101) + '\143' + chr(111) + chr(8960 - 8860) + chr(101))('\x75' + chr(0b1110100) + '\146' + chr(45) + '\070'): PuPeNl0CuqOQ.identity_bottom, xafqLlk3kkUe(SXOLrMavuUCe(b'\x9c\xa9\xd4q#(\xc9'), chr(2740 - 2640) + chr(4789 - 4688) + chr(0b1100011) + chr(6084 - 5973) + chr(100) + '\x65')('\165' + '\x74' + chr(8077 - 7975) + '\x2d' + chr(2778 - 2722)): PuPeNl0CuqOQ.identity_bottom}
n4ljua2gi1Pr.qxrVBjeryNEZ = {xafqLlk3kkUe(SXOLrMavuUCe(b'\x9c\xa9\xd4q#(\xc9'), '\144' + chr(0b1000000 + 0o45) + '\143' + '\x6f' + '\x64' + '\x65')(chr(0b1000 + 0o155) + chr(4059 - 3943) + chr(0b110001 + 0o65) + chr(45) + '\x38'): PuPeNl0CuqOQ.identity_top}
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), '\x64' + chr(0b1100101) + '\x63' + '\157' + chr(9748 - 9648) + '\x65')(chr(117) + '\x74' + chr(0b1100110) + chr(0b101101) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\xad\xc5y"9\xe5,\xce\xc9\xf5\xb6b\x80\xff\xdb,\xf3\x15\x1dW\'\xb5v'), chr(100) + chr(101) + '\143' + chr(0b1101111) + chr(0b1100100) + chr(0b11101 + 0o110))('\x75' + chr(5497 - 5381) + '\x66' + chr(263 - 218) + chr(2080 - 2024)), 1.5)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(7386 - 7286) + chr(101) + chr(99) + chr(111) + chr(3087 - 2987) + chr(4326 - 4225))('\x75' + chr(10674 - 10558) + chr(6246 - 6144) + chr(0b10110 + 0o27) + '\x38'))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8c\xad\xc5y"9\xe5,\xce\xc9\xf5\xb6b\x80\xf1\xc1.\xf4\x08\x0cU:'), '\x64' + '\145' + chr(0b1100011) + chr(0b11101 + 0o122) + chr(0b1100100) + chr(9615 - 9514))('\165' + chr(0b110111 + 0o75) + chr(0b1100110) + '\x2d' + '\x38'), 10.0)
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b1100100) + chr(101) + chr(0b1011011 + 0o10) + chr(0b1101111) + chr(0b1100100) + '\x65')(chr(117) + '\x74' + '\x66' + chr(661 - 616) + chr(3136 - 3080)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x87\xbd\xd2s4\x03\xd8!\xdf\xc4\xfa\x9dy\xb6\xe8\xcb'), '\x64' + '\145' + chr(5165 - 5066) + chr(111) + chr(0b1010010 + 0o22) + chr(9526 - 9425))('\165' + chr(0b1110100) + chr(7449 - 7347) + chr(0b1110 + 0o37) + '\x38'), ehT0Px3KOsy9('\060' + chr(2560 - 2449) + '\060', 0b1000))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(0b1000011 + 0o41) + chr(0b1100101) + chr(9405 - 9306) + chr(0b1101111) + chr(0b10100 + 0o120) + chr(0b1010011 + 0o22))('\x75' + chr(0b1110100) + '\x66' + '\x2d' + chr(56)))(xafqLlk3kkUe(SXOLrMavuUCe(b"\x9a\xad\xd5~',\xdf\x1f\xc7\xc8\xf5\xab~\xac\xcd\xc6!\xe4\x17"), '\144' + '\x65' + chr(0b11 + 0o140) + '\157' + '\x64' + chr(101))('\165' + chr(0b1110100) + chr(0b1100110) + chr(45) + chr(125 - 69)), ehT0Px3KOsy9(chr(2197 - 2149) + '\157' + chr(0b11001 + 0o27), 8))
xafqLlk3kkUe(n4ljua2gi1Pr, xafqLlk3kkUe(SXOLrMavuUCe(b'\x89\xac\xc2I.,\xdb2\xca\xca'), chr(3827 - 3727) + chr(0b110010 + 0o63) + chr(5320 - 5221) + chr(1073 - 962) + '\144' + chr(0b1100101))(chr(117) + chr(0b1110100) + chr(9802 - 9700) + chr(0b110 + 0o47) + chr(0b111000)))(xafqLlk3kkUe(SXOLrMavuUCe(b'\x8b\xa7\xcbf49\xc93\xc2\xc8\xfc\x9dl\xbe\xf1\xda/\xf5'), chr(5156 - 5056) + '\x65' + chr(0b1100011) + chr(111) + chr(873 - 773) + '\x65')('\x75' + chr(0b1011000 + 0o34) + chr(0b1010111 + 0o17) + '\x2d' + chr(2370 - 2314)), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x34', 8))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/mtf_transformer.py
|
mtf_transformer_tiny
|
def mtf_transformer_tiny():
"""Catch bugs locally..."""
hparams = mtf_transformer_base()
hparams.d_model = 128
hparams.d_ff = 512
hparams.batch_size = 8
hparams.encoder_layers = ["att", "drd"] * 2
hparams.decoder_layers = ["att", "enc_att", "drd"] * 2
hparams.num_heads = 8
# data parallelism and model-parallelism
hparams.mesh_shape = "batch:2;model:4"
hparams.activation_dtype = "float32"
return hparams
|
python
|
def mtf_transformer_tiny():
"""Catch bugs locally..."""
hparams = mtf_transformer_base()
hparams.d_model = 128
hparams.d_ff = 512
hparams.batch_size = 8
hparams.encoder_layers = ["att", "drd"] * 2
hparams.decoder_layers = ["att", "enc_att", "drd"] * 2
hparams.num_heads = 8
# data parallelism and model-parallelism
hparams.mesh_shape = "batch:2;model:4"
hparams.activation_dtype = "float32"
return hparams
|
[
"def",
"mtf_transformer_tiny",
"(",
")",
":",
"hparams",
"=",
"mtf_transformer_base",
"(",
")",
"hparams",
".",
"d_model",
"=",
"128",
"hparams",
".",
"d_ff",
"=",
"512",
"hparams",
".",
"batch_size",
"=",
"8",
"hparams",
".",
"encoder_layers",
"=",
"[",
"\"att\"",
",",
"\"drd\"",
"]",
"*",
"2",
"hparams",
".",
"decoder_layers",
"=",
"[",
"\"att\"",
",",
"\"enc_att\"",
",",
"\"drd\"",
"]",
"*",
"2",
"hparams",
".",
"num_heads",
"=",
"8",
"# data parallelism and model-parallelism",
"hparams",
".",
"mesh_shape",
"=",
"\"batch:2;model:4\"",
"hparams",
".",
"activation_dtype",
"=",
"\"float32\"",
"return",
"hparams"
] |
Catch bugs locally...
|
[
"Catch",
"bugs",
"locally",
"..."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/mtf_transformer.py#L897-L909
|
train
|
Catch bugs locally...
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(111) + chr(0b110010) + chr(0b110011) + '\062', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b110111) + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110111) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1316 - 1266) + '\x36' + chr(51), 26966 - 26958), ehT0Px3KOsy9(chr(1130 - 1082) + chr(0b1101111) + '\063' + chr(51) + chr(48), 0o10), ehT0Px3KOsy9(chr(1871 - 1823) + '\x6f' + chr(50) + chr(0b1010 + 0o46) + chr(0b110100), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1010010 + 0o35) + chr(0b110010) + '\x32' + '\x33', 0o10), ehT0Px3KOsy9('\060' + chr(0b101101 + 0o102) + '\x32' + '\x30' + '\060', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b110000 + 0o77) + chr(49) + '\x30' + chr(54), 0o10), ehT0Px3KOsy9('\060' + chr(0b1001111 + 0o40) + chr(0b110001) + '\062' + chr(2184 - 2132), ord("\x08")), ehT0Px3KOsy9(chr(0b101100 + 0o4) + chr(111) + '\061' + chr(825 - 773) + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b110001) + chr(2358 - 2308) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100011 + 0o14) + chr(243 - 193) + chr(0b11010 + 0o33) + '\062', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(51) + chr(1149 - 1099) + chr(52), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(1393 - 1343) + chr(539 - 484) + chr(50), ord("\x08")), ehT0Px3KOsy9(chr(117 - 69) + '\x6f' + chr(0b101110 + 0o5) + chr(0b110101) + chr(2452 - 2398), 0o10), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\157' + '\x31' + '\x37' + chr(1084 - 1032), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33' + '\x33' + chr(0b110100), 42657 - 42649), ehT0Px3KOsy9(chr(2148 - 2100) + chr(7867 - 7756) + chr(0b101011 + 0o6) + '\x33' + chr(0b100000 + 0o21), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(65 - 16) + chr(0b110100) + chr(2183 - 2131), 46937 - 46929), ehT0Px3KOsy9(chr(0b10101 + 0o33) + chr(0b110011 + 0o74) + chr(0b101100 + 0o6) + chr(0b110001) + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(0b101001 + 0o7) + chr(111) + chr(344 - 291) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b1000 + 0o52) + chr(0b110010) + '\066', 0b1000), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(2956 - 2845) + chr(0b11000 + 0o32) + chr(0b110100) + chr(0b110000), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\063' + chr(0b110001) + chr(2329 - 2276), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110010) + '\061' + '\064', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b10001 + 0o40) + chr(0b110011) + chr(50), 58002 - 57994), ehT0Px3KOsy9(chr(1005 - 957) + chr(111) + chr(2218 - 2167) + chr(674 - 619) + chr(52), 0o10), ehT0Px3KOsy9('\x30' + '\157' + chr(1187 - 1138) + chr(0b100101 + 0o14) + chr(0b100001 + 0o24), 0b1000), ehT0Px3KOsy9('\060' + chr(10730 - 10619) + '\x33' + chr(0b110010) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x32' + chr(0b0 + 0o61) + '\064', 8), ehT0Px3KOsy9(chr(787 - 739) + '\157' + chr(763 - 714) + chr(51) + chr(115 - 61), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + '\x33', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b101010 + 0o7) + '\x37' + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(0b10001 + 0o37) + chr(3216 - 3105) + chr(50) + chr(50) + '\062', ord("\x08")), ehT0Px3KOsy9(chr(0b10110 + 0o32) + '\x6f' + chr(0b110001) + chr(0b110001), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(7495 - 7384) + chr(0b110011) + chr(53) + chr(0b100010 + 0o22), 26607 - 26599), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(2389 - 2340) + chr(1206 - 1155) + '\x32', 8), ehT0Px3KOsy9('\060' + chr(0b1100111 + 0o10) + chr(51) + chr(0b101 + 0o62) + '\x32', 32843 - 32835), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b11101 + 0o24) + chr(48) + '\067', 0o10)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(1303 - 1192) + chr(0b110101) + chr(0b110000), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xb4'), '\x64' + '\x65' + chr(0b1100011) + chr(0b1011010 + 0o25) + '\x64' + chr(0b1100011 + 0o2))(chr(10562 - 10445) + chr(116) + chr(102) + chr(834 - 789) + chr(2583 - 2527)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def yRga0c19np5L():
n4ljua2gi1Pr = Hc6ONEMgq9eN()
n4ljua2gi1Pr.dHIk6a7HYqLO = ehT0Px3KOsy9(chr(0b110000) + chr(9495 - 9384) + chr(0b110010) + '\060' + chr(48), 8)
n4ljua2gi1Pr.EpyOHjLLhjxL = ehT0Px3KOsy9(chr(0b1111 + 0o41) + chr(111) + chr(49) + '\060' + '\060' + '\060', 57271 - 57263)
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(1551 - 1503) + chr(111) + '\061' + chr(315 - 267), 29206 - 29198)
n4ljua2gi1Pr.pbeC7au6N1jQ = [xafqLlk3kkUe(SXOLrMavuUCe(b'\xfbJO'), '\x64' + chr(0b1100101) + chr(99) + chr(0b1011111 + 0o20) + '\144' + '\x65')(chr(117) + '\164' + chr(0b1100110) + chr(45) + chr(2649 - 2593)), xafqLlk3kkUe(SXOLrMavuUCe(b'\xfeL_'), chr(0b1001011 + 0o31) + chr(0b1100101) + '\143' + '\157' + chr(1049 - 949) + '\145')('\165' + chr(0b1110100) + chr(0b1000 + 0o136) + chr(0b101101) + chr(0b10110 + 0o42))] * ehT0Px3KOsy9(chr(1530 - 1482) + chr(0b1101111) + chr(50), 0b1000)
n4ljua2gi1Pr.DuMwu3fbieF4 = [xafqLlk3kkUe(SXOLrMavuUCe(b'\xfbJO'), '\x64' + '\145' + '\x63' + '\x6f' + chr(8495 - 8395) + chr(101))(chr(4909 - 4792) + chr(116) + '\x66' + chr(1395 - 1350) + chr(0b111000)), xafqLlk3kkUe(SXOLrMavuUCe(b'\xffPX\xec\xb9\xf0\xe7'), '\x64' + chr(3671 - 3570) + '\x63' + '\x6f' + chr(0b1100100) + chr(0b1100101))('\x75' + chr(9883 - 9767) + chr(102) + chr(0b10100 + 0o31) + '\x38'), xafqLlk3kkUe(SXOLrMavuUCe(b'\xfeL_'), '\x64' + '\145' + chr(99) + chr(111) + '\144' + chr(0b1100101))('\x75' + '\x74' + chr(9940 - 9838) + chr(0b101101) + '\x38')] * ehT0Px3KOsy9(chr(1576 - 1528) + chr(0b1101111) + chr(0b110010), 8)
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(chr(1958 - 1910) + chr(11866 - 11755) + chr(49) + chr(1836 - 1788), 8)
n4ljua2gi1Pr.GnGMnRt7o0q6 = xafqLlk3kkUe(SXOLrMavuUCe(b'\xf8_O\xd0\xb0\xbe\xa1\xbb+;\x047\xc6\xee\x8e'), chr(0b1100100) + chr(3000 - 2899) + chr(0b11110 + 0o105) + chr(0b11100 + 0o123) + '\x64' + chr(4708 - 4607))(chr(0b1110101) + chr(0b1110100) + chr(0b100001 + 0o105) + chr(45) + chr(56))
n4ljua2gi1Pr.n6ZCgJ7AKd3U = xafqLlk3kkUe(SXOLrMavuUCe(b'\xfcRT\xd2\xac\xb7\xa1'), chr(100) + chr(0b1000011 + 0o42) + chr(0b1100011) + chr(111) + chr(0b10011 + 0o121) + '\x65')(chr(117) + '\164' + chr(102) + '\055' + '\070')
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/mtf_transformer.py
|
mtf_transformer_paper_lm
|
def mtf_transformer_paper_lm(size):
"""Config for language-model experiments.
Train these on languagemodel_lm1b32k_packed for 136000 steps (10 epochs)
The size parameter is an integer that controls the number of heads and the
size of the size of the feedforward hidden layers. Increasing size by 1
doubles each of these.
Results:
size params/10^9 log-ppl(per-token)
-1 0.14 3.209
0 0.22 3.119
1 0.37 3.037
2 0.67 2.969
3 1.28 2.912
4 2.48 2.874
5 4.90 2.871
(to get word-level log-ppl, multiply by 1.1078)
Args:
size: an integer
Returns:
a hparams object
"""
n = 2 ** size
hparams = mtf_transformer_base_lm()
hparams.batch_size = 256
hparams.d_model = 1024
hparams.d_ff = int(8192 * n)
hparams.d_kv = 256
hparams.num_heads = int(8 * n)
hparams.shared_embedding_and_softmax_weights = False
# one epoch for languagemodel_lm1b32k_packed = 13600 steps
hparams.learning_rate_decay_steps = 13600
return hparams
|
python
|
def mtf_transformer_paper_lm(size):
"""Config for language-model experiments.
Train these on languagemodel_lm1b32k_packed for 136000 steps (10 epochs)
The size parameter is an integer that controls the number of heads and the
size of the size of the feedforward hidden layers. Increasing size by 1
doubles each of these.
Results:
size params/10^9 log-ppl(per-token)
-1 0.14 3.209
0 0.22 3.119
1 0.37 3.037
2 0.67 2.969
3 1.28 2.912
4 2.48 2.874
5 4.90 2.871
(to get word-level log-ppl, multiply by 1.1078)
Args:
size: an integer
Returns:
a hparams object
"""
n = 2 ** size
hparams = mtf_transformer_base_lm()
hparams.batch_size = 256
hparams.d_model = 1024
hparams.d_ff = int(8192 * n)
hparams.d_kv = 256
hparams.num_heads = int(8 * n)
hparams.shared_embedding_and_softmax_weights = False
# one epoch for languagemodel_lm1b32k_packed = 13600 steps
hparams.learning_rate_decay_steps = 13600
return hparams
|
[
"def",
"mtf_transformer_paper_lm",
"(",
"size",
")",
":",
"n",
"=",
"2",
"**",
"size",
"hparams",
"=",
"mtf_transformer_base_lm",
"(",
")",
"hparams",
".",
"batch_size",
"=",
"256",
"hparams",
".",
"d_model",
"=",
"1024",
"hparams",
".",
"d_ff",
"=",
"int",
"(",
"8192",
"*",
"n",
")",
"hparams",
".",
"d_kv",
"=",
"256",
"hparams",
".",
"num_heads",
"=",
"int",
"(",
"8",
"*",
"n",
")",
"hparams",
".",
"shared_embedding_and_softmax_weights",
"=",
"False",
"# one epoch for languagemodel_lm1b32k_packed = 13600 steps",
"hparams",
".",
"learning_rate_decay_steps",
"=",
"13600",
"return",
"hparams"
] |
Config for language-model experiments.
Train these on languagemodel_lm1b32k_packed for 136000 steps (10 epochs)
The size parameter is an integer that controls the number of heads and the
size of the size of the feedforward hidden layers. Increasing size by 1
doubles each of these.
Results:
size params/10^9 log-ppl(per-token)
-1 0.14 3.209
0 0.22 3.119
1 0.37 3.037
2 0.67 2.969
3 1.28 2.912
4 2.48 2.874
5 4.90 2.871
(to get word-level log-ppl, multiply by 1.1078)
Args:
size: an integer
Returns:
a hparams object
|
[
"Config",
"for",
"language",
"-",
"model",
"experiments",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/mtf_transformer.py#L953-L989
|
train
|
Config for language - model experiments.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\062' + chr(1277 - 1224) + '\x32', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b100010 + 0o20) + '\x37' + '\x30', ord("\x08")), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\157' + chr(1914 - 1863) + chr(0b110000) + '\063', 0o10), ehT0Px3KOsy9('\060' + chr(0b1101111) + '\x37' + chr(0b1011 + 0o51), 0b1000), ehT0Px3KOsy9(chr(48) + chr(12167 - 12056) + '\x34' + chr(616 - 568), 55095 - 55087), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010) + chr(274 - 223) + '\064', ord("\x08")), ehT0Px3KOsy9(chr(1417 - 1369) + chr(0b1101111) + '\061' + chr(0b101110 + 0o3) + '\x31', ord("\x08")), ehT0Px3KOsy9(chr(0b11001 + 0o27) + chr(111) + '\x36' + '\x37', 26849 - 26841), ehT0Px3KOsy9(chr(0b110000) + chr(0b1100111 + 0o10) + chr(0b110001) + chr(0b110011) + '\x37', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(2831 - 2777) + chr(51), 0b1000), ehT0Px3KOsy9(chr(437 - 389) + '\x6f' + chr(454 - 404) + chr(0b10110 + 0o36) + chr(1144 - 1090), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\062' + chr(0b110011) + chr(0b11010 + 0o27), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + chr(0b110011) + '\x36' + '\x35', 0b1000), ehT0Px3KOsy9(chr(48) + chr(3655 - 3544) + '\062' + '\x36' + '\063', 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b10110 + 0o34) + chr(0b101000 + 0o16) + chr(843 - 790), 0b1000), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110010) + chr(0b110100) + '\062', 49427 - 49419), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(0b11011 + 0o124) + '\x33' + chr(0b11100 + 0o30) + '\060', 0o10), ehT0Px3KOsy9(chr(0b101110 + 0o2) + '\x6f' + chr(489 - 439) + '\062' + chr(1455 - 1401), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b10010 + 0o37) + chr(1526 - 1473) + chr(0b10000 + 0o44), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\063' + '\x37' + '\x36', 0o10), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010) + '\x33' + chr(48), 41316 - 41308), ehT0Px3KOsy9(chr(2133 - 2085) + chr(111) + '\061' + chr(0b101100 + 0o6) + chr(2156 - 2108), 6313 - 6305), ehT0Px3KOsy9('\x30' + '\157' + chr(0b11110 + 0o25) + chr(51) + chr(0b1111 + 0o42), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(0b10011 + 0o40) + chr(2500 - 2447) + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(0b11110 + 0o22) + chr(0b1101100 + 0o3) + '\x32' + '\x33' + '\066', 0b1000), ehT0Px3KOsy9('\060' + chr(111) + '\x33' + chr(0b110111) + '\066', 8), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x33' + chr(0b110100) + chr(55), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010) + '\063' + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(2123 - 2073) + chr(55) + chr(0b110111), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1000100 + 0o53) + '\x32' + '\x30' + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(0b11111 + 0o21) + '\x6f' + '\061' + chr(1469 - 1419), 0o10), ehT0Px3KOsy9(chr(0b100010 + 0o16) + '\x6f' + chr(0b100111 + 0o12) + chr(0b110001) + chr(0b100011 + 0o22), 42598 - 42590), ehT0Px3KOsy9(chr(0b11111 + 0o21) + chr(2111 - 2000) + chr(0b101010 + 0o11) + chr(0b110000) + chr(52), 0o10), ehT0Px3KOsy9(chr(0b1001 + 0o47) + chr(0b1001111 + 0o40) + chr(0b110010) + chr(0b10011 + 0o43) + chr(0b110101), 8), ehT0Px3KOsy9(chr(48) + chr(8696 - 8585) + chr(49) + chr(0b110011) + chr(0b110101), 57770 - 57762), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(1248 - 1199) + chr(0b110010) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b1000 + 0o50) + '\x6f' + chr(0b110001) + chr(0b110011) + '\x36', 0o10), ehT0Px3KOsy9(chr(0b10110 + 0o32) + '\x6f' + '\x32' + chr(0b1110 + 0o42), ord("\x08")), ehT0Px3KOsy9('\060' + '\157' + '\x32' + '\x36' + '\x35', 8), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b110010) + chr(53) + '\062', 8)][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x35' + chr(407 - 359), 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x84'), chr(6125 - 6025) + '\145' + chr(99) + chr(0b1101111) + '\x64' + chr(6911 - 6810))('\165' + '\x74' + chr(6034 - 5932) + chr(45) + '\070') + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def dM5o03B9SMXY(NLcc3BCJnQka):
m1NkCryOw9Bx = ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010), ord("\x08")) ** NLcc3BCJnQka
n4ljua2gi1Pr = PQlKMm71kedX()
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(48) + '\157' + chr(0b11011 + 0o31) + '\x30' + chr(1690 - 1642), 0b1000)
n4ljua2gi1Pr.dHIk6a7HYqLO = ehT0Px3KOsy9(chr(366 - 318) + '\x6f' + chr(2351 - 2301) + '\060' + chr(0b110000) + '\x30', 0b1000)
n4ljua2gi1Pr.EpyOHjLLhjxL = ehT0Px3KOsy9(ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\062' + chr(0b110000) + chr(48) + chr(453 - 405) + chr(587 - 539), 0b1000) * m1NkCryOw9Bx)
n4ljua2gi1Pr.cboJxztkwgnV = ehT0Px3KOsy9(chr(0b111 + 0o51) + chr(0b1101111) + chr(2240 - 2188) + chr(2250 - 2202) + '\x30', 8)
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(ehT0Px3KOsy9(chr(0b100101 + 0o13) + chr(5008 - 4897) + '\061' + '\060', 26521 - 26513) * m1NkCryOw9Bx)
n4ljua2gi1Pr.qVamxim0L2I1 = ehT0Px3KOsy9(chr(2185 - 2137) + chr(9750 - 9639) + chr(0b110000), 0b1000)
n4ljua2gi1Pr.YBAB1XyoxOc5 = ehT0Px3KOsy9(chr(0b110000 + 0o0) + chr(0b1101001 + 0o6) + '\x33' + chr(0b110010) + chr(1439 - 1387) + chr(52) + chr(0b11111 + 0o21), 0b1000)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/mtf_transformer.py
|
mtf_transformer_paper_tr
|
def mtf_transformer_paper_tr(size):
"""Config for translation experiments.
Train these on translate_enfr_wmt32k_packed for 154000 steps (3 epochs)
The size parameter is an integer that controls the number of heads and the
size of the size of the feedforward hidden layers. Increasing size by 1
doubles each of these.
Args:
size: an integer
Returns:
a hparams object
"""
n = 2 ** size
hparams = mtf_transformer_base()
hparams.label_smoothing = 0.1
hparams.batch_size = 128
hparams.d_model = 1024
hparams.d_ff = int(4096 * n)
hparams.num_heads = int(8 * n)
hparams.shared_embedding_and_softmax_weights = False
# one epoch for translate_enfr_wmt32k_packed = 51400 steps
hparams.learning_rate_decay_steps = 51400
return hparams
|
python
|
def mtf_transformer_paper_tr(size):
"""Config for translation experiments.
Train these on translate_enfr_wmt32k_packed for 154000 steps (3 epochs)
The size parameter is an integer that controls the number of heads and the
size of the size of the feedforward hidden layers. Increasing size by 1
doubles each of these.
Args:
size: an integer
Returns:
a hparams object
"""
n = 2 ** size
hparams = mtf_transformer_base()
hparams.label_smoothing = 0.1
hparams.batch_size = 128
hparams.d_model = 1024
hparams.d_ff = int(4096 * n)
hparams.num_heads = int(8 * n)
hparams.shared_embedding_and_softmax_weights = False
# one epoch for translate_enfr_wmt32k_packed = 51400 steps
hparams.learning_rate_decay_steps = 51400
return hparams
|
[
"def",
"mtf_transformer_paper_tr",
"(",
"size",
")",
":",
"n",
"=",
"2",
"**",
"size",
"hparams",
"=",
"mtf_transformer_base",
"(",
")",
"hparams",
".",
"label_smoothing",
"=",
"0.1",
"hparams",
".",
"batch_size",
"=",
"128",
"hparams",
".",
"d_model",
"=",
"1024",
"hparams",
".",
"d_ff",
"=",
"int",
"(",
"4096",
"*",
"n",
")",
"hparams",
".",
"num_heads",
"=",
"int",
"(",
"8",
"*",
"n",
")",
"hparams",
".",
"shared_embedding_and_softmax_weights",
"=",
"False",
"# one epoch for translate_enfr_wmt32k_packed = 51400 steps",
"hparams",
".",
"learning_rate_decay_steps",
"=",
"51400",
"return",
"hparams"
] |
Config for translation experiments.
Train these on translate_enfr_wmt32k_packed for 154000 steps (3 epochs)
The size parameter is an integer that controls the number of heads and the
size of the size of the feedforward hidden layers. Increasing size by 1
doubles each of these.
Args:
size: an integer
Returns:
a hparams object
|
[
"Config",
"for",
"translation",
"experiments",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/mtf_transformer.py#L1041-L1065
|
train
|
Config for translation experiments.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9('\060' + chr(111) + chr(0b1 + 0o60) + chr(0b101110 + 0o6) + chr(0b101110 + 0o2), 0o10), ehT0Px3KOsy9('\060' + chr(0b101011 + 0o104) + chr(50) + chr(0b110111) + chr(2718 - 2665), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(1886 - 1775) + '\061' + chr(0b1101 + 0o46) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(8844 - 8733) + chr(1011 - 962) + '\x31' + chr(52), 0b1000), ehT0Px3KOsy9(chr(732 - 684) + '\157' + chr(54) + '\065', 0b1000), ehT0Px3KOsy9(chr(0b1010 + 0o46) + '\x6f' + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011101 + 0o22) + chr(0b10101 + 0o36) + chr(0b110000) + chr(2074 - 2019), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\066' + '\x36', 0b1000), ehT0Px3KOsy9('\x30' + '\x6f' + chr(51) + chr(0b1011 + 0o54) + chr(0b100100 + 0o17), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x33' + '\060' + '\063', ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + '\062' + '\x35' + '\062', ord("\x08")), ehT0Px3KOsy9(chr(1796 - 1748) + '\157' + '\067' + '\064', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + '\062' + '\060' + chr(995 - 946), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b101010 + 0o7) + chr(1323 - 1275) + chr(0b1011 + 0o53), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + chr(2330 - 2280), 22106 - 22098), ehT0Px3KOsy9(chr(1358 - 1310) + chr(0b1101111 + 0o0) + chr(0b11111 + 0o24) + chr(0b110001) + '\x32', 0o10), ehT0Px3KOsy9(chr(1104 - 1056) + '\157' + chr(1186 - 1137) + '\062' + chr(0b110000), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b110000 + 0o77) + '\061' + '\062' + chr(549 - 495), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110001) + chr(52) + chr(829 - 777), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(5771 - 5660) + '\x32' + '\x33' + chr(0b110010), 10178 - 10170), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + '\060' + chr(0b111 + 0o51), 0b1000), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(48), 44678 - 44670), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\063' + chr(53), 9441 - 9433), ehT0Px3KOsy9(chr(0b1000 + 0o50) + chr(0b1000101 + 0o52) + chr(49) + chr(51) + chr(48), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + chr(51) + chr(0b110101) + chr(52), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\062' + chr(50) + chr(0b110001), 0o10), ehT0Px3KOsy9('\060' + chr(0b11000 + 0o127) + '\x32' + '\x32' + '\x30', 27949 - 27941), ehT0Px3KOsy9('\x30' + chr(0b11011 + 0o124) + chr(0b110001) + chr(48) + chr(49), 11024 - 11016), ehT0Px3KOsy9(chr(48) + chr(0b1100111 + 0o10) + '\062' + '\x31' + chr(0b110001 + 0o6), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(49) + chr(54) + chr(0b110001), 20106 - 20098), ehT0Px3KOsy9('\x30' + '\x6f' + chr(49) + chr(2079 - 2028) + chr(48), 8), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b11110 + 0o23) + '\x32' + chr(849 - 799), 38482 - 38474), ehT0Px3KOsy9(chr(0b11 + 0o55) + '\157' + chr(1486 - 1437) + '\x30', 4173 - 4165), ehT0Px3KOsy9('\060' + chr(11354 - 11243) + chr(270 - 220) + '\x34', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(111) + chr(50) + '\x32' + '\063', 0b1000), ehT0Px3KOsy9(chr(0b10000 + 0o40) + '\x6f' + '\061' + '\061' + chr(0b110101), 0o10), ehT0Px3KOsy9(chr(1377 - 1329) + '\x6f' + chr(49) + '\062' + '\x36', 8), ehT0Px3KOsy9('\060' + chr(111) + chr(51) + '\x31' + chr(51), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(51) + chr(0b110000), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\x32' + chr(0b110101) + '\x32', 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b110100 + 0o73) + '\065' + chr(0b110000), 61667 - 61659)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'N'), '\x64' + chr(4278 - 4177) + chr(99) + chr(111) + chr(100) + chr(0b100111 + 0o76))(chr(0b1110101) + chr(5654 - 5538) + '\x66' + chr(1161 - 1116) + chr(0b111000)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def sU8frzVFhDEf(NLcc3BCJnQka):
m1NkCryOw9Bx = ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110010), 8) ** NLcc3BCJnQka
n4ljua2gi1Pr = Hc6ONEMgq9eN()
n4ljua2gi1Pr.FSjUgdaczzRk = 0.1
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b100 + 0o56) + chr(48) + chr(0b0 + 0o60), ord("\x08"))
n4ljua2gi1Pr.dHIk6a7HYqLO = ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b1100 + 0o46) + '\060' + chr(0b110000) + '\x30', 0b1000)
n4ljua2gi1Pr.EpyOHjLLhjxL = ehT0Px3KOsy9(ehT0Px3KOsy9('\060' + '\x6f' + '\x31' + '\x30' + chr(0b1000 + 0o50) + '\x30' + '\x30', 0o10) * m1NkCryOw9Bx)
n4ljua2gi1Pr.vRVqPOZ1hUG7 = ehT0Px3KOsy9(ehT0Px3KOsy9(chr(2196 - 2148) + chr(0b110111 + 0o70) + '\061' + chr(1993 - 1945), 8) * m1NkCryOw9Bx)
n4ljua2gi1Pr.qVamxim0L2I1 = ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110000), 8)
n4ljua2gi1Pr.YBAB1XyoxOc5 = ehT0Px3KOsy9(chr(48) + '\157' + '\x31' + chr(2360 - 2308) + '\x34' + '\063' + '\061' + chr(0b100 + 0o54), 0b1000)
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/models/mtf_transformer.py
|
mtf_transformer_lm_baseline
|
def mtf_transformer_lm_baseline():
"""Small language model to run on 1 TPU.
Run this on 2x2 on languagemodel_lm1b32k_packed for 272000 steps (10 epochs)
Results:
params/10^9 log-ppl(per-token)
0.14 3.202
Returns:
a hparams
"""
hparams = mtf_transformer_paper_lm(-1)
hparams.batch_size = 128
hparams.learning_rate_decay_steps = 27200 # one epoch on lm1b
hparams.mesh_shape = "batch:8"
return hparams
|
python
|
def mtf_transformer_lm_baseline():
"""Small language model to run on 1 TPU.
Run this on 2x2 on languagemodel_lm1b32k_packed for 272000 steps (10 epochs)
Results:
params/10^9 log-ppl(per-token)
0.14 3.202
Returns:
a hparams
"""
hparams = mtf_transformer_paper_lm(-1)
hparams.batch_size = 128
hparams.learning_rate_decay_steps = 27200 # one epoch on lm1b
hparams.mesh_shape = "batch:8"
return hparams
|
[
"def",
"mtf_transformer_lm_baseline",
"(",
")",
":",
"hparams",
"=",
"mtf_transformer_paper_lm",
"(",
"-",
"1",
")",
"hparams",
".",
"batch_size",
"=",
"128",
"hparams",
".",
"learning_rate_decay_steps",
"=",
"27200",
"# one epoch on lm1b",
"hparams",
".",
"mesh_shape",
"=",
"\"batch:8\"",
"return",
"hparams"
] |
Small language model to run on 1 TPU.
Run this on 2x2 on languagemodel_lm1b32k_packed for 272000 steps (10 epochs)
Results:
params/10^9 log-ppl(per-token)
0.14 3.202
Returns:
a hparams
|
[
"Small",
"language",
"model",
"to",
"run",
"on",
"1",
"TPU",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/models/mtf_transformer.py#L1171-L1186
|
train
|
Small language model to run on 1 TPU.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1118 - 1067) + chr(54) + '\x32', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b110100 + 0o73) + '\x33' + '\x37' + '\064', 15953 - 15945), ehT0Px3KOsy9('\x30' + chr(0b100 + 0o153) + '\x31' + chr(1972 - 1917) + '\067', 9304 - 9296), ehT0Px3KOsy9('\060' + '\157' + chr(2184 - 2133) + '\064', ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(0b10011 + 0o37) + chr(50) + '\x30', 62443 - 62435), ehT0Px3KOsy9('\x30' + chr(8138 - 8027) + '\x33' + '\x31', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x31' + chr(1313 - 1258) + chr(0b100101 + 0o13), 0o10), ehT0Px3KOsy9(chr(0b100011 + 0o15) + chr(0b1101111) + '\x33' + '\066' + chr(557 - 508), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(54) + chr(1530 - 1480), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(49) + chr(0b110001) + '\x33', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b111111 + 0o60) + '\061' + chr(0b110101) + chr(49), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1000101 + 0o52) + chr(0b1000 + 0o57) + '\x35', 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(55), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1919 - 1869) + '\x30' + '\063', 62672 - 62664), ehT0Px3KOsy9(chr(48) + chr(111) + '\x32' + '\x31' + chr(55), ord("\x08")), ehT0Px3KOsy9(chr(48) + '\157' + chr(1197 - 1148) + chr(321 - 266) + '\x35', 11966 - 11958), ehT0Px3KOsy9(chr(0b11100 + 0o24) + chr(12314 - 12203) + '\x32' + chr(0b11000 + 0o32), 0b1000), ehT0Px3KOsy9(chr(827 - 779) + '\x6f' + '\063' + chr(0b10 + 0o56), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b111111 + 0o60) + '\x32' + '\x36' + chr(0b110110), 35153 - 35145), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\x33' + '\061' + chr(2214 - 2164), 0o10), ehT0Px3KOsy9(chr(753 - 705) + chr(111) + chr(1182 - 1131) + chr(0b110001) + chr(0b110110), 0o10), ehT0Px3KOsy9('\x30' + chr(0b1100101 + 0o12) + '\x32' + chr(0b11001 + 0o36) + chr(2352 - 2303), 0b1000), ehT0Px3KOsy9(chr(0b100000 + 0o20) + '\x6f' + '\062' + chr(51) + chr(54), 10604 - 10596), ehT0Px3KOsy9('\060' + '\x6f' + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(455 - 407) + '\157' + '\061' + chr(50) + '\x37', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(7025 - 6914) + '\061' + chr(0b100000 + 0o20) + chr(1391 - 1337), 29424 - 29416), ehT0Px3KOsy9('\x30' + chr(111) + chr(201 - 147) + '\x33', 0b1000), ehT0Px3KOsy9(chr(0b111 + 0o51) + '\157' + '\x31' + chr(0b10110 + 0o41) + '\x32', 16135 - 16127), ehT0Px3KOsy9(chr(2180 - 2132) + '\157' + chr(0b110010) + chr(0b11010 + 0o31) + '\063', ord("\x08")), ehT0Px3KOsy9(chr(1231 - 1183) + chr(0b1101111) + chr(49) + '\x35' + chr(0b110110), 0b1000), ehT0Px3KOsy9(chr(1887 - 1839) + chr(0b1101111) + '\063' + chr(0b1100 + 0o53) + '\x34', 8), ehT0Px3KOsy9(chr(228 - 180) + chr(111) + chr(0b101 + 0o54) + chr(0b10011 + 0o37) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(0b101101 + 0o102) + chr(49) + chr(52) + chr(0b110011), ord("\x08")), ehT0Px3KOsy9('\x30' + '\157' + chr(389 - 338) + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + chr(111) + chr(1934 - 1883) + chr(1960 - 1912) + chr(49), 0b1000), ehT0Px3KOsy9(chr(1921 - 1873) + chr(4995 - 4884) + chr(0b110111) + chr(51), 0o10), ehT0Px3KOsy9(chr(0b11 + 0o55) + '\157' + chr(0b11111 + 0o23) + chr(0b1100 + 0o45) + chr(0b110100), ord("\x08")), ehT0Px3KOsy9(chr(1215 - 1167) + chr(0b10000 + 0o137) + chr(0b110011) + chr(0b101001 + 0o15) + '\065', 0o10), ehT0Px3KOsy9(chr(0b111 + 0o51) + chr(0b1101111) + '\062' + '\061' + chr(48), 0b1000), ehT0Px3KOsy9(chr(0b0 + 0o60) + '\157' + '\x33' + '\066', 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\065' + chr(0b100110 + 0o12), 29087 - 29079)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xdc'), chr(0b1100100) + chr(101) + chr(2480 - 2381) + chr(111) + '\144' + chr(0b0 + 0o145))(chr(0b1110101) + '\164' + chr(0b1001110 + 0o30) + chr(45) + chr(0b1101 + 0o53)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def d_V5t5yywFGf():
n4ljua2gi1Pr = dM5o03B9SMXY(-ehT0Px3KOsy9(chr(48) + chr(111) + '\x31', 0b1000))
n4ljua2gi1Pr.ix9dZyeAmUxY = ehT0Px3KOsy9('\060' + chr(111) + chr(50) + chr(48) + '\060', 29352 - 29344)
n4ljua2gi1Pr.YBAB1XyoxOc5 = ehT0Px3KOsy9('\060' + chr(7213 - 7102) + '\066' + chr(0b110101) + chr(618 - 569) + chr(0b1000 + 0o50) + chr(48), ord("\x08"))
n4ljua2gi1Pr.GnGMnRt7o0q6 = xafqLlk3kkUe(SXOLrMavuUCe(b'\x90\x9d\x0b\x7f\xf8#@'), '\x64' + chr(0b1100101) + '\143' + chr(0b1101111) + chr(3721 - 3621) + '\x65')(chr(117) + '\164' + '\x66' + '\055' + chr(986 - 930))
return n4ljua2gi1Pr
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/message_passing_attention.py
|
multihead_graph_attention
|
def multihead_graph_attention(query_antecedent,
memory_antecedent,
bias,
total_key_depth,
total_value_depth,
output_depth,
num_heads,
dropout_rate,
image_shapes=None,
attention_type="edge_vector",
name="multihead_graph_attention",
save_weights_to=None,
make_image_summary=True,
dropout_broadcast_dims=None,
adjacency_matrix=None,
num_edge_types=5,
vars_3d=False,
**kwargs):
"""Multihead scaled-dot-product attention with input/output transformations.
Args:
query_antecedent: a Tensor with shape [batch, length_q, channels]
memory_antecedent: a Tensor with shape [batch, length_m, channels] or None
bias: bias Tensor (see attention_bias())
total_key_depth: an integer
total_value_depth: an integer
output_depth: an integer
num_heads: an integer dividing total_key_depth and total_value_depth
dropout_rate: a floating point number
image_shapes: optional tuple of integer scalars.
see comments for attention_image_summary()
attention_type: a string, either "dot_product", "dot_product_relative",
"local_mask_right", "local_unmasked", "masked_dilated_1d",
"unmasked_dilated_1d", graph, or any attention function
with the signature (query, key, value, **kwargs)
name: an optional string.
save_weights_to: an optional dictionary to capture attention weights
for vizualization; the weights tensor will be appended there under
a string key created from the variable scope (including name).
make_image_summary: Whether to make an attention image summary.
dropout_broadcast_dims: an optional list of integers less than 4
specifying in which dimensions to broadcast the dropout decisions.
saves memory.
adjacency_matrix: an optional tensor of shape [batch, len_q, len_q]
containing edge vectors for attention
num_edge_types: number of edge types, an int
vars_3d: use 3-dimensional variables for input/output transformations
**kwargs (dict): Parameters for the attention function
Returns:
The result of the attention transformation. The output shape is
[batch_size, length_q, output_depth]
Raises:
ValueError: if the key depth or value depth are not divisible by the
number of attention heads.
"""
if total_key_depth % num_heads != 0:
raise ValueError("Key depth (%d) must be divisible by the number of "
"attention heads (%d)." % (total_key_depth, num_heads))
if total_value_depth % num_heads != 0:
raise ValueError("Value depth (%d) must be divisible by the number of "
"attention heads (%d)." % (total_value_depth, num_heads))
vars_3d_num_heads = num_heads if vars_3d else None
with tf.variable_scope(
name,
default_name="multihead_attention",
values=[query_antecedent, memory_antecedent]):
q, k, v = common_attention.compute_qkv(
query_antecedent,
memory_antecedent,
total_key_depth,
total_value_depth,
vars_3d_num_heads=vars_3d_num_heads)
q = common_attention.split_heads(q, num_heads)
k = common_attention.split_heads(k, num_heads)
v = common_attention.split_heads(v, num_heads)
key_depth_per_head = total_key_depth // num_heads
if not vars_3d:
q *= key_depth_per_head**-0.5
additional_returned_value = None
if callable(attention_type): # Generic way to extend multihead_attention
x = attention_type(q, k, v, **kwargs)
if isinstance(x, tuple):
x, additional_returned_value = x # Unpack
elif attention_type == "edge_vector":
x = graph_attention(
q,
k,
v,
bias,
dropout_rate,
image_shapes,
save_weights_to=save_weights_to,
make_image_summary=make_image_summary,
dropout_broadcast_dims=dropout_broadcast_dims,
adjacency_matrix=adjacency_matrix,
num_edge_types=num_edge_types)
x = common_attention.combine_heads(x)
# Set last dim specifically.
x.set_shape(x.shape.as_list()[:-1] + [total_value_depth])
if vars_3d:
o_var = tf.get_variable(
"o", [num_heads, total_value_depth // num_heads, output_depth])
o_var = tf.reshape(o_var, [total_value_depth, output_depth])
x = tf.tensordot(x, o_var, axes=1)
else:
x = common_layers.dense(
x, output_depth, use_bias=False, name="output_transform")
if additional_returned_value is not None:
return x, additional_returned_value
return x
|
python
|
def multihead_graph_attention(query_antecedent,
memory_antecedent,
bias,
total_key_depth,
total_value_depth,
output_depth,
num_heads,
dropout_rate,
image_shapes=None,
attention_type="edge_vector",
name="multihead_graph_attention",
save_weights_to=None,
make_image_summary=True,
dropout_broadcast_dims=None,
adjacency_matrix=None,
num_edge_types=5,
vars_3d=False,
**kwargs):
"""Multihead scaled-dot-product attention with input/output transformations.
Args:
query_antecedent: a Tensor with shape [batch, length_q, channels]
memory_antecedent: a Tensor with shape [batch, length_m, channels] or None
bias: bias Tensor (see attention_bias())
total_key_depth: an integer
total_value_depth: an integer
output_depth: an integer
num_heads: an integer dividing total_key_depth and total_value_depth
dropout_rate: a floating point number
image_shapes: optional tuple of integer scalars.
see comments for attention_image_summary()
attention_type: a string, either "dot_product", "dot_product_relative",
"local_mask_right", "local_unmasked", "masked_dilated_1d",
"unmasked_dilated_1d", graph, or any attention function
with the signature (query, key, value, **kwargs)
name: an optional string.
save_weights_to: an optional dictionary to capture attention weights
for vizualization; the weights tensor will be appended there under
a string key created from the variable scope (including name).
make_image_summary: Whether to make an attention image summary.
dropout_broadcast_dims: an optional list of integers less than 4
specifying in which dimensions to broadcast the dropout decisions.
saves memory.
adjacency_matrix: an optional tensor of shape [batch, len_q, len_q]
containing edge vectors for attention
num_edge_types: number of edge types, an int
vars_3d: use 3-dimensional variables for input/output transformations
**kwargs (dict): Parameters for the attention function
Returns:
The result of the attention transformation. The output shape is
[batch_size, length_q, output_depth]
Raises:
ValueError: if the key depth or value depth are not divisible by the
number of attention heads.
"""
if total_key_depth % num_heads != 0:
raise ValueError("Key depth (%d) must be divisible by the number of "
"attention heads (%d)." % (total_key_depth, num_heads))
if total_value_depth % num_heads != 0:
raise ValueError("Value depth (%d) must be divisible by the number of "
"attention heads (%d)." % (total_value_depth, num_heads))
vars_3d_num_heads = num_heads if vars_3d else None
with tf.variable_scope(
name,
default_name="multihead_attention",
values=[query_antecedent, memory_antecedent]):
q, k, v = common_attention.compute_qkv(
query_antecedent,
memory_antecedent,
total_key_depth,
total_value_depth,
vars_3d_num_heads=vars_3d_num_heads)
q = common_attention.split_heads(q, num_heads)
k = common_attention.split_heads(k, num_heads)
v = common_attention.split_heads(v, num_heads)
key_depth_per_head = total_key_depth // num_heads
if not vars_3d:
q *= key_depth_per_head**-0.5
additional_returned_value = None
if callable(attention_type): # Generic way to extend multihead_attention
x = attention_type(q, k, v, **kwargs)
if isinstance(x, tuple):
x, additional_returned_value = x # Unpack
elif attention_type == "edge_vector":
x = graph_attention(
q,
k,
v,
bias,
dropout_rate,
image_shapes,
save_weights_to=save_weights_to,
make_image_summary=make_image_summary,
dropout_broadcast_dims=dropout_broadcast_dims,
adjacency_matrix=adjacency_matrix,
num_edge_types=num_edge_types)
x = common_attention.combine_heads(x)
# Set last dim specifically.
x.set_shape(x.shape.as_list()[:-1] + [total_value_depth])
if vars_3d:
o_var = tf.get_variable(
"o", [num_heads, total_value_depth // num_heads, output_depth])
o_var = tf.reshape(o_var, [total_value_depth, output_depth])
x = tf.tensordot(x, o_var, axes=1)
else:
x = common_layers.dense(
x, output_depth, use_bias=False, name="output_transform")
if additional_returned_value is not None:
return x, additional_returned_value
return x
|
[
"def",
"multihead_graph_attention",
"(",
"query_antecedent",
",",
"memory_antecedent",
",",
"bias",
",",
"total_key_depth",
",",
"total_value_depth",
",",
"output_depth",
",",
"num_heads",
",",
"dropout_rate",
",",
"image_shapes",
"=",
"None",
",",
"attention_type",
"=",
"\"edge_vector\"",
",",
"name",
"=",
"\"multihead_graph_attention\"",
",",
"save_weights_to",
"=",
"None",
",",
"make_image_summary",
"=",
"True",
",",
"dropout_broadcast_dims",
"=",
"None",
",",
"adjacency_matrix",
"=",
"None",
",",
"num_edge_types",
"=",
"5",
",",
"vars_3d",
"=",
"False",
",",
"*",
"*",
"kwargs",
")",
":",
"if",
"total_key_depth",
"%",
"num_heads",
"!=",
"0",
":",
"raise",
"ValueError",
"(",
"\"Key depth (%d) must be divisible by the number of \"",
"\"attention heads (%d).\"",
"%",
"(",
"total_key_depth",
",",
"num_heads",
")",
")",
"if",
"total_value_depth",
"%",
"num_heads",
"!=",
"0",
":",
"raise",
"ValueError",
"(",
"\"Value depth (%d) must be divisible by the number of \"",
"\"attention heads (%d).\"",
"%",
"(",
"total_value_depth",
",",
"num_heads",
")",
")",
"vars_3d_num_heads",
"=",
"num_heads",
"if",
"vars_3d",
"else",
"None",
"with",
"tf",
".",
"variable_scope",
"(",
"name",
",",
"default_name",
"=",
"\"multihead_attention\"",
",",
"values",
"=",
"[",
"query_antecedent",
",",
"memory_antecedent",
"]",
")",
":",
"q",
",",
"k",
",",
"v",
"=",
"common_attention",
".",
"compute_qkv",
"(",
"query_antecedent",
",",
"memory_antecedent",
",",
"total_key_depth",
",",
"total_value_depth",
",",
"vars_3d_num_heads",
"=",
"vars_3d_num_heads",
")",
"q",
"=",
"common_attention",
".",
"split_heads",
"(",
"q",
",",
"num_heads",
")",
"k",
"=",
"common_attention",
".",
"split_heads",
"(",
"k",
",",
"num_heads",
")",
"v",
"=",
"common_attention",
".",
"split_heads",
"(",
"v",
",",
"num_heads",
")",
"key_depth_per_head",
"=",
"total_key_depth",
"//",
"num_heads",
"if",
"not",
"vars_3d",
":",
"q",
"*=",
"key_depth_per_head",
"**",
"-",
"0.5",
"additional_returned_value",
"=",
"None",
"if",
"callable",
"(",
"attention_type",
")",
":",
"# Generic way to extend multihead_attention",
"x",
"=",
"attention_type",
"(",
"q",
",",
"k",
",",
"v",
",",
"*",
"*",
"kwargs",
")",
"if",
"isinstance",
"(",
"x",
",",
"tuple",
")",
":",
"x",
",",
"additional_returned_value",
"=",
"x",
"# Unpack",
"elif",
"attention_type",
"==",
"\"edge_vector\"",
":",
"x",
"=",
"graph_attention",
"(",
"q",
",",
"k",
",",
"v",
",",
"bias",
",",
"dropout_rate",
",",
"image_shapes",
",",
"save_weights_to",
"=",
"save_weights_to",
",",
"make_image_summary",
"=",
"make_image_summary",
",",
"dropout_broadcast_dims",
"=",
"dropout_broadcast_dims",
",",
"adjacency_matrix",
"=",
"adjacency_matrix",
",",
"num_edge_types",
"=",
"num_edge_types",
")",
"x",
"=",
"common_attention",
".",
"combine_heads",
"(",
"x",
")",
"# Set last dim specifically.",
"x",
".",
"set_shape",
"(",
"x",
".",
"shape",
".",
"as_list",
"(",
")",
"[",
":",
"-",
"1",
"]",
"+",
"[",
"total_value_depth",
"]",
")",
"if",
"vars_3d",
":",
"o_var",
"=",
"tf",
".",
"get_variable",
"(",
"\"o\"",
",",
"[",
"num_heads",
",",
"total_value_depth",
"//",
"num_heads",
",",
"output_depth",
"]",
")",
"o_var",
"=",
"tf",
".",
"reshape",
"(",
"o_var",
",",
"[",
"total_value_depth",
",",
"output_depth",
"]",
")",
"x",
"=",
"tf",
".",
"tensordot",
"(",
"x",
",",
"o_var",
",",
"axes",
"=",
"1",
")",
"else",
":",
"x",
"=",
"common_layers",
".",
"dense",
"(",
"x",
",",
"output_depth",
",",
"use_bias",
"=",
"False",
",",
"name",
"=",
"\"output_transform\"",
")",
"if",
"additional_returned_value",
"is",
"not",
"None",
":",
"return",
"x",
",",
"additional_returned_value",
"return",
"x"
] |
Multihead scaled-dot-product attention with input/output transformations.
Args:
query_antecedent: a Tensor with shape [batch, length_q, channels]
memory_antecedent: a Tensor with shape [batch, length_m, channels] or None
bias: bias Tensor (see attention_bias())
total_key_depth: an integer
total_value_depth: an integer
output_depth: an integer
num_heads: an integer dividing total_key_depth and total_value_depth
dropout_rate: a floating point number
image_shapes: optional tuple of integer scalars.
see comments for attention_image_summary()
attention_type: a string, either "dot_product", "dot_product_relative",
"local_mask_right", "local_unmasked", "masked_dilated_1d",
"unmasked_dilated_1d", graph, or any attention function
with the signature (query, key, value, **kwargs)
name: an optional string.
save_weights_to: an optional dictionary to capture attention weights
for vizualization; the weights tensor will be appended there under
a string key created from the variable scope (including name).
make_image_summary: Whether to make an attention image summary.
dropout_broadcast_dims: an optional list of integers less than 4
specifying in which dimensions to broadcast the dropout decisions.
saves memory.
adjacency_matrix: an optional tensor of shape [batch, len_q, len_q]
containing edge vectors for attention
num_edge_types: number of edge types, an int
vars_3d: use 3-dimensional variables for input/output transformations
**kwargs (dict): Parameters for the attention function
Returns:
The result of the attention transformation. The output shape is
[batch_size, length_q, output_depth]
Raises:
ValueError: if the key depth or value depth are not divisible by the
number of attention heads.
|
[
"Multihead",
"scaled",
"-",
"dot",
"-",
"product",
"attention",
"with",
"input",
"/",
"output",
"transformations",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/message_passing_attention.py#L28-L146
|
train
|
Multihead scaled - dot - product attention with input and output transformations.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x31' + chr(1923 - 1874) + '\067', 52138 - 52130), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010) + chr(2908 - 2854) + chr(53), ord("\x08")), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(0b100110 + 0o111) + chr(0b101011 + 0o7) + chr(0b100111 + 0o14), ord("\x08")), ehT0Px3KOsy9(chr(1320 - 1272) + '\x6f' + '\x31' + chr(0b110000 + 0o5) + chr(54), 32736 - 32728), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110011) + '\x34' + '\064', ord("\x08")), ehT0Px3KOsy9('\060' + chr(813 - 702) + chr(52), 37092 - 37084), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110100) + chr(0b110000 + 0o0), 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b10100 + 0o37) + '\x33', 57994 - 57986), ehT0Px3KOsy9(chr(48) + chr(111) + chr(0b110010) + chr(293 - 244) + chr(530 - 481), 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1011110 + 0o21) + '\x32' + '\x30' + chr(0b101000 + 0o12), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(11065 - 10954) + chr(51) + chr(847 - 794) + chr(53), 0b1000), ehT0Px3KOsy9(chr(0b100 + 0o54) + chr(2098 - 1987) + chr(154 - 103) + chr(52) + chr(0b110101 + 0o0), 0o10), ehT0Px3KOsy9(chr(0b1 + 0o57) + '\x6f' + chr(134 - 84), 4650 - 4642), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + chr(0b101000 + 0o10) + chr(0b100101 + 0o13), 0o10), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b110001) + chr(1754 - 1703), ord("\x08")), ehT0Px3KOsy9('\060' + '\x6f' + chr(0b11111 + 0o26) + chr(0b110001), ord("\x08")), ehT0Px3KOsy9('\x30' + chr(11930 - 11819) + '\x32' + '\062' + chr(856 - 807), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(180 - 126) + chr(51), 0o10), ehT0Px3KOsy9(chr(1477 - 1429) + '\x6f' + '\x37', 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(1377 - 1324), 32217 - 32209), ehT0Px3KOsy9('\060' + chr(111) + '\x31' + chr(0b110110) + chr(0b101000 + 0o12), 0b1000), ehT0Px3KOsy9(chr(0b11110 + 0o22) + chr(111) + chr(0b110010 + 0o0) + '\x33' + chr(55), 8576 - 8568), ehT0Px3KOsy9(chr(330 - 282) + '\x6f' + '\061' + '\066' + '\065', 51210 - 51202), ehT0Px3KOsy9('\060' + chr(10731 - 10620) + chr(398 - 348) + chr(0b1010 + 0o51) + chr(763 - 714), 47357 - 47349), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(1749 - 1638) + '\063' + chr(0b101111 + 0o3) + '\x31', 0o10), ehT0Px3KOsy9(chr(0b1100 + 0o44) + '\157' + chr(51) + chr(54), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1000111 + 0o50) + '\065', 8), ehT0Px3KOsy9(chr(48) + '\157' + '\x37' + '\065', 46984 - 46976), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b1110 + 0o45) + chr(2075 - 2025) + chr(0b11111 + 0o25), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b100110 + 0o13) + chr(0b11010 + 0o34) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\063' + chr(0b110111) + chr(1802 - 1749), 0b1000), ehT0Px3KOsy9(chr(48) + chr(7176 - 7065) + chr(0b10010 + 0o43) + '\x37', 0b1000), ehT0Px3KOsy9(chr(1904 - 1856) + chr(111) + chr(0b110001) + chr(0b101000 + 0o16) + '\x33', 0o10), ehT0Px3KOsy9(chr(790 - 742) + '\x6f' + chr(50) + chr(2614 - 2561) + chr(53), 0b1000), ehT0Px3KOsy9(chr(578 - 530) + chr(111) + chr(51) + chr(0b101010 + 0o10) + chr(2455 - 2403), 8), ehT0Px3KOsy9(chr(0b10010 + 0o36) + '\x6f' + '\062' + '\061' + chr(0b110000), 7875 - 7867), ehT0Px3KOsy9('\060' + chr(111) + '\066' + '\x33', 8), ehT0Px3KOsy9('\060' + '\157' + chr(0b110010) + '\x30' + chr(88 - 38), 8), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110011) + '\x36' + '\061', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x32' + '\063' + '\061', 8)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(48) + chr(0b1000000 + 0o57) + chr(0b100010 + 0o23) + chr(0b110000), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'a'), chr(1069 - 969) + '\145' + chr(99) + chr(1074 - 963) + chr(0b1100011 + 0o1) + '\145')(chr(0b1110101) + '\164' + '\x66' + '\055' + chr(0b110000 + 0o10)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def nIbPm3qvdKMe(ENas6b3HzFya, LWkuqV72y7LV, IKTrMTySqz10, _jxqy0P17UFy, F9lUBuHPQMmX, EkLrr8g1UZri, vRVqPOZ1hUG7, iI9Z069HML_u, IHMu1EGwZgDx=None, lZ1GB4L2oMeG=xafqLlk3kkUe(SXOLrMavuUCe(b'*\xd1t\x0b\xbbrEx\xfd\xed\x9f'), '\x64' + chr(3473 - 3372) + chr(99) + '\x6f' + chr(9265 - 9165) + chr(101))(chr(8563 - 8446) + chr(116) + '\146' + chr(45) + '\x38'), AIvJRzLdDfgF=xafqLlk3kkUe(SXOLrMavuUCe(b'"\xc0\x7f\x1a\x8dlEz\xed\xdd\x8a\x97c0cv\xf0Gs`\x1c\xeb\x1c\xf6\xbf'), '\144' + chr(101) + chr(0b1100011) + '\x6f' + chr(0b111 + 0o135) + chr(0b1100101))(chr(117) + '\x74' + chr(102) + chr(869 - 824) + '\x38'), zWaF_2VBEDjk=None, NC2xHNLwzxcH=ehT0Px3KOsy9('\x30' + chr(0b1011 + 0o144) + '\x31', 0o10), Tovc3lDEHg6s=None, l1TxoI57nKGl=None, Rzg3d1CR9gzK=ehT0Px3KOsy9(chr(48) + chr(0b1101111) + '\065', 8), ATCcziaCSlIY=ehT0Px3KOsy9(chr(399 - 351) + chr(111) + chr(0b10001 + 0o37), 10160 - 10152), **M8EIoTs2GJXE):
if _jxqy0P17UFy % vRVqPOZ1hUG7 != ehT0Px3KOsy9(chr(0b110000) + chr(999 - 888) + chr(1250 - 1202), 8):
raise q1QCh3W88sgk(xafqLlk3kkUe(SXOLrMavuUCe(b'\x04\xd0jN\x80aPo\xe1\xa2\xc5\xc0fi+D\xe4@s%\x10\xfaU\xfd\xb8hlN\xbd\x12\x91JS\xcf\xad\xa8G\xea\x94\xb9!\xc0~\x0c\x81v\x00t\xef\xa2\x8c\x91v%e]\xf8\\i%\x1a\xfa\x14\xfd\xa2>-\x18\xb0Y\xd3'), '\x64' + '\x65' + '\143' + '\157' + chr(100) + chr(101))(chr(117) + '\x74' + chr(0b1000111 + 0o37) + chr(0b1010 + 0o43) + chr(0b100010 + 0o26)) % (_jxqy0P17UFy, vRVqPOZ1hUG7))
if F9lUBuHPQMmX % vRVqPOZ1hUG7 != ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(1120 - 1072), 8):
raise q1QCh3W88sgk(xafqLlk3kkUe(SXOLrMavuUCe(b'\x19\xd4\x7f\x1b\x81$D~\xf9\xf6\x85\xc5*eo\x00\xb1^rv\x06\xbf\x17\xfc\xf1zlK\xbd\x03\x94M\x1f\xc8\xf4\xeaJ\xa2\x85\xf1*\x95}\x1b\x89fEi\xa9\xed\x8b\xc5c4\x7fL\xffGnj\x1c\xbf\x1d\xfc\xb0zv\x1d\xfcU\x99\x06]'), '\x64' + chr(101) + '\x63' + '\x6f' + chr(0b1100100) + chr(101))(chr(0b1110101) + chr(0b111111 + 0o65) + chr(0b1100110) + '\x2d' + chr(0b111000)) % (F9lUBuHPQMmX, vRVqPOZ1hUG7))
UlWL1V3BA5Ze = vRVqPOZ1hUG7 if ATCcziaCSlIY else None
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'9\xd4a\x07\x85fL~\xd6\xf1\x8e\x8ar%'), chr(1004 - 904) + '\145' + '\143' + '\157' + chr(0b1100100) + '\x65')(chr(117) + chr(0b1000101 + 0o57) + chr(102) + chr(0b100001 + 0o14) + '\x38'))(AIvJRzLdDfgF, default_name=xafqLlk3kkUe(SXOLrMavuUCe(b'"\xc0\x7f\x1a\x8dlEz\xed\xdd\x8c\x91v%e]\xf8\\i'), '\x64' + '\x65' + '\143' + '\157' + chr(100) + '\x65')('\x75' + '\x74' + chr(0b1100110) + chr(1129 - 1084) + chr(2739 - 2683)), values=[ENas6b3HzFya, LWkuqV72y7LV]):
(WtwjCI_b3w8O, OolUPRJhRaJd, cMbll0QYhULo) = WOnrfm4dlYcf.compute_qkv(ENas6b3HzFya, LWkuqV72y7LV, _jxqy0P17UFy, F9lUBuHPQMmX, vars_3d_num_heads=UlWL1V3BA5Ze)
WtwjCI_b3w8O = WOnrfm4dlYcf.split_heads(WtwjCI_b3w8O, vRVqPOZ1hUG7)
OolUPRJhRaJd = WOnrfm4dlYcf.split_heads(OolUPRJhRaJd, vRVqPOZ1hUG7)
cMbll0QYhULo = WOnrfm4dlYcf.split_heads(cMbll0QYhULo, vRVqPOZ1hUG7)
SwYdOFe3xV4H = _jxqy0P17UFy // vRVqPOZ1hUG7
if not ATCcziaCSlIY:
WtwjCI_b3w8O *= SwYdOFe3xV4H ** (-0.5)
b02C5juXis9p = None
if tzcpInYwBvYW(lZ1GB4L2oMeG):
OeWW0F1dBPRQ = lZ1GB4L2oMeG(WtwjCI_b3w8O, OolUPRJhRaJd, cMbll0QYhULo, **M8EIoTs2GJXE)
if PlSM16l2KDPD(OeWW0F1dBPRQ, KNyTy8rYcwji):
(OeWW0F1dBPRQ, b02C5juXis9p) = OeWW0F1dBPRQ
elif lZ1GB4L2oMeG == xafqLlk3kkUe(SXOLrMavuUCe(b'*\xd1t\x0b\xbbrEx\xfd\xed\x9f'), chr(9949 - 9849) + '\x65' + chr(99) + '\157' + chr(4953 - 4853) + '\145')(chr(753 - 636) + chr(11740 - 11624) + chr(0b110011 + 0o63) + chr(313 - 268) + chr(1073 - 1017)):
OeWW0F1dBPRQ = AeSoc87kLTbY(WtwjCI_b3w8O, OolUPRJhRaJd, cMbll0QYhULo, IKTrMTySqz10, iI9Z069HML_u, IHMu1EGwZgDx, save_weights_to=zWaF_2VBEDjk, make_image_summary=NC2xHNLwzxcH, dropout_broadcast_dims=Tovc3lDEHg6s, adjacency_matrix=l1TxoI57nKGl, num_edge_types=Rzg3d1CR9gzK)
OeWW0F1dBPRQ = WOnrfm4dlYcf.combine_heads(OeWW0F1dBPRQ)
xafqLlk3kkUe(OeWW0F1dBPRQ, xafqLlk3kkUe(SXOLrMavuUCe(b'<\xd0g1\x97lAk\xec'), chr(0b1010 + 0o132) + '\x65' + chr(0b111101 + 0o46) + '\157' + chr(0b111110 + 0o46) + '\x65')(chr(0b1110101) + '\x74' + '\x66' + chr(0b100101 + 0o10) + '\070'))(xafqLlk3kkUe(OeWW0F1dBPRQ.shape, xafqLlk3kkUe(SXOLrMavuUCe(b'.\xc6L\x02\x8dwT'), '\144' + chr(0b10 + 0o143) + '\143' + '\157' + chr(0b1100100) + chr(0b1001011 + 0o32))('\x75' + chr(2651 - 2535) + '\x66' + '\055' + '\070'))()[:-ehT0Px3KOsy9('\060' + chr(11325 - 11214) + '\061', 8)] + [F9lUBuHPQMmX])
if ATCcziaCSlIY:
JUtal7GNKsm2 = IDJ2eXGCBCDu.get_variable(xafqLlk3kkUe(SXOLrMavuUCe(b' '), chr(8557 - 8457) + chr(8068 - 7967) + chr(0b101100 + 0o67) + chr(111) + chr(0b1100100) + chr(101))(chr(0b101101 + 0o110) + chr(116) + chr(102) + chr(45) + '\070'), [vRVqPOZ1hUG7, F9lUBuHPQMmX // vRVqPOZ1hUG7, EkLrr8g1UZri])
JUtal7GNKsm2 = IDJ2eXGCBCDu.reshape(JUtal7GNKsm2, [F9lUBuHPQMmX, EkLrr8g1UZri])
OeWW0F1dBPRQ = IDJ2eXGCBCDu.tensordot(OeWW0F1dBPRQ, JUtal7GNKsm2, axes=ehT0Px3KOsy9('\x30' + chr(0b1100 + 0o143) + chr(703 - 654), 8))
else:
OeWW0F1dBPRQ = jSKPaHwSAfVv.dense(OeWW0F1dBPRQ, EkLrr8g1UZri, use_bias=ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110000), 8), name=xafqLlk3kkUe(SXOLrMavuUCe(b' \xc0g\x1e\x91p\x7fo\xfb\xe3\x83\x96d/yD'), '\144' + chr(0b1100101) + chr(0b1100011) + chr(11055 - 10944) + chr(100) + '\x65')(chr(0b1110101) + '\x74' + chr(2007 - 1905) + '\055' + chr(0b111000)))
if b02C5juXis9p is not None:
return (OeWW0F1dBPRQ, b02C5juXis9p)
return OeWW0F1dBPRQ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/message_passing_attention.py
|
graph_attention
|
def graph_attention(q,
k,
v,
bias,
dropout_rate=0.0,
image_shapes=None,
name=None,
make_image_summary=True,
save_weights_to=None,
dropout_broadcast_dims=None,
adjacency_matrix=None,
num_edge_types=5):
"""graph attention.
Args:
q: a Tensor with shape [batch, heads, length_q, depth_k]
k: a Tensor with shape [batch, heads, length_kv, depth_k]
v: a Tensor with shape [batch, heads, length_kv, depth_v]
bias: bias Tensor (see attention_bias())
dropout_rate: a floating point number
image_shapes: optional tuple of integer scalars.
see comments for attention_image_summary()
name: an optional string
make_image_summary: True if you want an image summary.
save_weights_to: an optional dictionary to capture attention weights
for vizualization; the weights tensor will be appended there under
a string key created from the variable scope (including name).
dropout_broadcast_dims: an optional list of integers less than 4
specifying in which dimensions to broadcast the dropout decisions.
saves memory.
adjacency_matrix: optional matrix of [batch, length, length] ids indicating
edge type
num_edge_types: an int indicating number of edge types
Returns:
A Tensor of shape [batch, length, depth(q)]
"""
with tf.variable_scope(
name, default_name="dot_product_attention", values=[q, k, v]) as scope:
# [batch, num_heads, query_length, memory_length]
logits = tf.matmul(q, k, transpose_b=True)
if adjacency_matrix is not None:
key_head_depth = common_layers.shape_list(q)[-1]
adjacency_vectors = make_edge_vectors(
adjacency_matrix,
num_edge_types,
key_head_depth,
name=name)
# transposing q to be [batch, length_q, heads, depth_k]
# to allow for matmul with [batch, length_q, length_q, depth_k]
q_t = tf.transpose(q, [0, 2, 1, 3])
adj_logits = tf.matmul(q_t, adjacency_vectors, transpose_b=True)
logits += tf.transpose(adj_logits, [0, 2, 1, 3])
# [batch, depth, num_nodes, num_nodes]
if bias is not None:
logits += bias
weights = tf.nn.softmax(logits, name="attention_weights")
if save_weights_to is not None:
save_weights_to[scope.name] = weights
# dropping out the attention links for each of the heads
weights = common_layers.dropout_with_broadcast_dims(
weights, 1.0 - dropout_rate, broadcast_dims=dropout_broadcast_dims)
if common_layers.should_generate_summaries() and make_image_summary:
common_attention.attention_image_summary(weights, image_shapes)
return tf.matmul(weights, v)
|
python
|
def graph_attention(q,
k,
v,
bias,
dropout_rate=0.0,
image_shapes=None,
name=None,
make_image_summary=True,
save_weights_to=None,
dropout_broadcast_dims=None,
adjacency_matrix=None,
num_edge_types=5):
"""graph attention.
Args:
q: a Tensor with shape [batch, heads, length_q, depth_k]
k: a Tensor with shape [batch, heads, length_kv, depth_k]
v: a Tensor with shape [batch, heads, length_kv, depth_v]
bias: bias Tensor (see attention_bias())
dropout_rate: a floating point number
image_shapes: optional tuple of integer scalars.
see comments for attention_image_summary()
name: an optional string
make_image_summary: True if you want an image summary.
save_weights_to: an optional dictionary to capture attention weights
for vizualization; the weights tensor will be appended there under
a string key created from the variable scope (including name).
dropout_broadcast_dims: an optional list of integers less than 4
specifying in which dimensions to broadcast the dropout decisions.
saves memory.
adjacency_matrix: optional matrix of [batch, length, length] ids indicating
edge type
num_edge_types: an int indicating number of edge types
Returns:
A Tensor of shape [batch, length, depth(q)]
"""
with tf.variable_scope(
name, default_name="dot_product_attention", values=[q, k, v]) as scope:
# [batch, num_heads, query_length, memory_length]
logits = tf.matmul(q, k, transpose_b=True)
if adjacency_matrix is not None:
key_head_depth = common_layers.shape_list(q)[-1]
adjacency_vectors = make_edge_vectors(
adjacency_matrix,
num_edge_types,
key_head_depth,
name=name)
# transposing q to be [batch, length_q, heads, depth_k]
# to allow for matmul with [batch, length_q, length_q, depth_k]
q_t = tf.transpose(q, [0, 2, 1, 3])
adj_logits = tf.matmul(q_t, adjacency_vectors, transpose_b=True)
logits += tf.transpose(adj_logits, [0, 2, 1, 3])
# [batch, depth, num_nodes, num_nodes]
if bias is not None:
logits += bias
weights = tf.nn.softmax(logits, name="attention_weights")
if save_weights_to is not None:
save_weights_to[scope.name] = weights
# dropping out the attention links for each of the heads
weights = common_layers.dropout_with_broadcast_dims(
weights, 1.0 - dropout_rate, broadcast_dims=dropout_broadcast_dims)
if common_layers.should_generate_summaries() and make_image_summary:
common_attention.attention_image_summary(weights, image_shapes)
return tf.matmul(weights, v)
|
[
"def",
"graph_attention",
"(",
"q",
",",
"k",
",",
"v",
",",
"bias",
",",
"dropout_rate",
"=",
"0.0",
",",
"image_shapes",
"=",
"None",
",",
"name",
"=",
"None",
",",
"make_image_summary",
"=",
"True",
",",
"save_weights_to",
"=",
"None",
",",
"dropout_broadcast_dims",
"=",
"None",
",",
"adjacency_matrix",
"=",
"None",
",",
"num_edge_types",
"=",
"5",
")",
":",
"with",
"tf",
".",
"variable_scope",
"(",
"name",
",",
"default_name",
"=",
"\"dot_product_attention\"",
",",
"values",
"=",
"[",
"q",
",",
"k",
",",
"v",
"]",
")",
"as",
"scope",
":",
"# [batch, num_heads, query_length, memory_length]",
"logits",
"=",
"tf",
".",
"matmul",
"(",
"q",
",",
"k",
",",
"transpose_b",
"=",
"True",
")",
"if",
"adjacency_matrix",
"is",
"not",
"None",
":",
"key_head_depth",
"=",
"common_layers",
".",
"shape_list",
"(",
"q",
")",
"[",
"-",
"1",
"]",
"adjacency_vectors",
"=",
"make_edge_vectors",
"(",
"adjacency_matrix",
",",
"num_edge_types",
",",
"key_head_depth",
",",
"name",
"=",
"name",
")",
"# transposing q to be [batch, length_q, heads, depth_k]",
"# to allow for matmul with [batch, length_q, length_q, depth_k]",
"q_t",
"=",
"tf",
".",
"transpose",
"(",
"q",
",",
"[",
"0",
",",
"2",
",",
"1",
",",
"3",
"]",
")",
"adj_logits",
"=",
"tf",
".",
"matmul",
"(",
"q_t",
",",
"adjacency_vectors",
",",
"transpose_b",
"=",
"True",
")",
"logits",
"+=",
"tf",
".",
"transpose",
"(",
"adj_logits",
",",
"[",
"0",
",",
"2",
",",
"1",
",",
"3",
"]",
")",
"# [batch, depth, num_nodes, num_nodes]",
"if",
"bias",
"is",
"not",
"None",
":",
"logits",
"+=",
"bias",
"weights",
"=",
"tf",
".",
"nn",
".",
"softmax",
"(",
"logits",
",",
"name",
"=",
"\"attention_weights\"",
")",
"if",
"save_weights_to",
"is",
"not",
"None",
":",
"save_weights_to",
"[",
"scope",
".",
"name",
"]",
"=",
"weights",
"# dropping out the attention links for each of the heads",
"weights",
"=",
"common_layers",
".",
"dropout_with_broadcast_dims",
"(",
"weights",
",",
"1.0",
"-",
"dropout_rate",
",",
"broadcast_dims",
"=",
"dropout_broadcast_dims",
")",
"if",
"common_layers",
".",
"should_generate_summaries",
"(",
")",
"and",
"make_image_summary",
":",
"common_attention",
".",
"attention_image_summary",
"(",
"weights",
",",
"image_shapes",
")",
"return",
"tf",
".",
"matmul",
"(",
"weights",
",",
"v",
")"
] |
graph attention.
Args:
q: a Tensor with shape [batch, heads, length_q, depth_k]
k: a Tensor with shape [batch, heads, length_kv, depth_k]
v: a Tensor with shape [batch, heads, length_kv, depth_v]
bias: bias Tensor (see attention_bias())
dropout_rate: a floating point number
image_shapes: optional tuple of integer scalars.
see comments for attention_image_summary()
name: an optional string
make_image_summary: True if you want an image summary.
save_weights_to: an optional dictionary to capture attention weights
for vizualization; the weights tensor will be appended there under
a string key created from the variable scope (including name).
dropout_broadcast_dims: an optional list of integers less than 4
specifying in which dimensions to broadcast the dropout decisions.
saves memory.
adjacency_matrix: optional matrix of [batch, length, length] ids indicating
edge type
num_edge_types: an int indicating number of edge types
Returns:
A Tensor of shape [batch, length, depth(q)]
|
[
"graph",
"attention",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/message_passing_attention.py#L185-L248
|
train
|
graph attention.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\065' + chr(53), 18614 - 18606), ehT0Px3KOsy9(chr(781 - 733) + chr(0b110000 + 0o77) + chr(0b0 + 0o62) + '\061' + '\061', 0o10), ehT0Px3KOsy9(chr(1526 - 1478) + chr(0b1100 + 0o143) + chr(0b10110 + 0o34) + chr(0b110011) + '\x32', 30875 - 30867), ehT0Px3KOsy9(chr(371 - 323) + '\157' + chr(2378 - 2327) + '\x36', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b10001 + 0o41) + chr(0b110111) + chr(48), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\x32' + chr(0b100000 + 0o20) + '\061', 65050 - 65042), ehT0Px3KOsy9(chr(0b110000) + chr(494 - 383) + chr(0b10 + 0o61) + chr(50), 0b1000), ehT0Px3KOsy9(chr(2205 - 2157) + '\157' + chr(0b101101 + 0o5) + chr(0b110100) + '\063', 0o10), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x37' + chr(49), 40184 - 40176), ehT0Px3KOsy9(chr(0b110000) + chr(0b1011 + 0o144) + chr(0b11 + 0o60) + chr(50) + '\x33', 0b1000), ehT0Px3KOsy9('\060' + chr(111) + chr(53) + chr(0b110011 + 0o1), ord("\x08")), ehT0Px3KOsy9(chr(1851 - 1803) + chr(111) + chr(901 - 852) + '\066' + '\x33', 0b1000), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + chr(2648 - 2594) + chr(50), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + '\x32' + chr(52) + chr(0b110011), 8), ehT0Px3KOsy9(chr(0b0 + 0o60) + chr(6016 - 5905) + chr(0b100110 + 0o15) + chr(0b110100) + '\x31', 11520 - 11512), ehT0Px3KOsy9(chr(0b101111 + 0o1) + chr(111) + '\x32' + '\066' + '\066', 0b1000), ehT0Px3KOsy9(chr(0b100111 + 0o11) + chr(0b1000010 + 0o55) + chr(2051 - 2000) + chr(0b110000 + 0o4) + chr(0b110001), 8), ehT0Px3KOsy9(chr(0b10100 + 0o34) + '\x6f' + chr(51) + chr(0b1001 + 0o55) + '\067', 13082 - 13074), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110001) + chr(2378 - 2323) + chr(0b1000 + 0o57), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(11511 - 11400) + chr(0b110001) + chr(0b110010) + chr(49), 22970 - 22962), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + '\062' + '\x30' + chr(50), 0b1000), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b1001 + 0o52) + chr(50) + chr(52), 56419 - 56411), ehT0Px3KOsy9(chr(1623 - 1575) + chr(0b1001010 + 0o45) + chr(2013 - 1964) + chr(0b110111) + chr(1247 - 1193), 54700 - 54692), ehT0Px3KOsy9(chr(48) + chr(0b1010010 + 0o35) + chr(0b101111 + 0o4) + chr(51), 35485 - 35477), ehT0Px3KOsy9(chr(0b110000) + chr(10729 - 10618) + chr(0b110101) + chr(0b110111), 0b1000), ehT0Px3KOsy9(chr(890 - 842) + '\157' + '\x32' + '\x32' + '\x35', 0b1000), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(1923 - 1812) + chr(267 - 216) + chr(55) + '\061', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(1575 - 1525) + chr(0b1 + 0o63), 0b1000), ehT0Px3KOsy9(chr(48) + chr(777 - 666) + '\x32' + chr(0b110110) + chr(0b110110), 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + '\x35' + chr(348 - 293), 39269 - 39261), ehT0Px3KOsy9(chr(0b101110 + 0o2) + chr(9172 - 9061) + chr(0b110011) + chr(2281 - 2233) + chr(54), 50891 - 50883), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110010) + chr(0b101 + 0o54) + '\067', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110010) + '\066' + '\062', 0b1000), ehT0Px3KOsy9(chr(615 - 567) + chr(333 - 222) + '\x31' + '\066' + chr(55), 61682 - 61674), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110011) + '\x37' + chr(0b100101 + 0o20), 0o10), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(50) + chr(48) + chr(54), 8584 - 8576), ehT0Px3KOsy9(chr(0b110000) + chr(3903 - 3792) + '\x32' + chr(0b101001 + 0o13) + chr(0b110001), 0b1000), ehT0Px3KOsy9('\060' + chr(0b0 + 0o157) + chr(0b110100) + '\061', 50966 - 50958), ehT0Px3KOsy9('\x30' + chr(0b101000 + 0o107) + '\x31' + chr(1041 - 986) + chr(2349 - 2299), 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(0b110001) + '\x31' + chr(0b101001 + 0o10), ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9('\x30' + chr(0b1010000 + 0o37) + chr(53) + chr(1682 - 1634), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\x03'), chr(100) + chr(101) + '\143' + chr(0b1101111) + chr(9011 - 8911) + chr(0b1100101))('\165' + '\164' + '\146' + chr(1330 - 1285) + chr(56)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def AeSoc87kLTbY(WtwjCI_b3w8O, OolUPRJhRaJd, cMbll0QYhULo, IKTrMTySqz10, iI9Z069HML_u=0.0, IHMu1EGwZgDx=None, AIvJRzLdDfgF=None, NC2xHNLwzxcH=ehT0Px3KOsy9(chr(48) + chr(111) + '\x31', ord("\x08")), zWaF_2VBEDjk=None, Tovc3lDEHg6s=None, l1TxoI57nKGl=None, Rzg3d1CR9gzK=ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(53), 64731 - 64723)):
with xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'[\xe8\xe8w\xd9\xe3v\xc6f\xa0O\x1c\x82\xc6'), chr(0b111110 + 0o46) + chr(0b111101 + 0o50) + chr(3197 - 3098) + chr(111) + chr(0b1000000 + 0o44) + chr(0b1100101))('\165' + chr(3688 - 3572) + chr(102) + '\055' + chr(0b111000)))(AIvJRzLdDfgF, default_name=xafqLlk3kkUe(SXOLrMavuUCe(b'I\xe6\xeeA\xc8\xf3u\xc7L\xb0X,\x93\xd7\x95\xd3\x8f\x83\x16>\xec'), '\144' + chr(101) + '\x63' + '\x6f' + chr(0b1100001 + 0o3) + '\x65')('\x75' + chr(116) + chr(102) + chr(702 - 657) + '\070'), values=[WtwjCI_b3w8O, OolUPRJhRaJd, cMbll0QYhULo]) as CJBHNoj4zKoT:
wF9nmvjsKjYM = IDJ2eXGCBCDu.matmul(WtwjCI_b3w8O, OolUPRJhRaJd, transpose_b=ehT0Px3KOsy9(chr(0b110000) + chr(8549 - 8438) + chr(0b110001), 8))
if l1TxoI57nKGl is not None:
zja63PgJFPEd = jSKPaHwSAfVv.shape_list(WtwjCI_b3w8O)[-ehT0Px3KOsy9(chr(1014 - 966) + '\157' + chr(0b110001), 8)]
saoJkVOtukhv = HNKgJFbx17r0(l1TxoI57nKGl, Rzg3d1CR9gzK, zja63PgJFPEd, name=AIvJRzLdDfgF)
baebVmZ0s0YE = IDJ2eXGCBCDu.transpose(WtwjCI_b3w8O, [ehT0Px3KOsy9('\x30' + chr(111) + '\060', 1499 - 1491), ehT0Px3KOsy9(chr(48) + chr(111) + '\062', 46982 - 46974), ehT0Px3KOsy9(chr(542 - 494) + chr(0b1101111) + chr(0b110001), 8), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b110011), 9585 - 9577)])
paxuAdGOJMbn = IDJ2eXGCBCDu.matmul(baebVmZ0s0YE, saoJkVOtukhv, transpose_b=ehT0Px3KOsy9(chr(0b110000) + chr(8847 - 8736) + '\061', 8))
wF9nmvjsKjYM += IDJ2eXGCBCDu.transpose(paxuAdGOJMbn, [ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(286 - 238), 8), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(0b1101111) + chr(0b110000 + 0o2), 8), ehT0Px3KOsy9(chr(1444 - 1396) + chr(2293 - 2182) + chr(49), 8), ehT0Px3KOsy9(chr(1348 - 1300) + '\x6f' + chr(0b10101 + 0o36), 8)])
if IKTrMTySqz10 is not None:
wF9nmvjsKjYM += IKTrMTySqz10
ZurHTci57aXw = IDJ2eXGCBCDu.nn.softmax(wF9nmvjsKjYM, name=xafqLlk3kkUe(SXOLrMavuUCe(b'L\xfd\xee{\xd6\xf5s\xccW\x8c[\x16\x9b\xc4\x89\xc2\x92'), '\144' + '\145' + '\x63' + '\157' + chr(3545 - 3445) + chr(0b111010 + 0o53))('\165' + chr(3323 - 3207) + '\x66' + chr(45) + chr(0b100111 + 0o21)))
if zWaF_2VBEDjk is not None:
zWaF_2VBEDjk[CJBHNoj4zKoT.AIvJRzLdDfgF] = ZurHTci57aXw
ZurHTci57aXw = jSKPaHwSAfVv.dropout_with_broadcast_dims(ZurHTci57aXw, 1.0 - iI9Z069HML_u, broadcast_dims=Tovc3lDEHg6s)
if xafqLlk3kkUe(jSKPaHwSAfVv, xafqLlk3kkUe(SXOLrMavuUCe(b'^\xe1\xf5k\xd4\xe5E\xc4\\\xbdI\x01\x93\xd7\x84\xe9\x92\x82\x12<\xe3\xa6\x17~\\'), chr(0b111 + 0o135) + chr(0b101110 + 0o67) + chr(5139 - 5040) + '\x6f' + '\x64' + chr(0b1100101))(chr(117) + '\164' + '\146' + '\055' + chr(969 - 913)))() and NC2xHNLwzxcH:
xafqLlk3kkUe(WOnrfm4dlYcf, xafqLlk3kkUe(SXOLrMavuUCe(b'L\xfd\xee{\xd6\xf5s\xccW\x8cE\x1e\x93\xc4\x84\xe9\x92\x82\x12<\xe3\xa6\x07'), chr(6426 - 6326) + chr(101) + chr(99) + chr(0b1001011 + 0o44) + '\144' + chr(0b1100101))(chr(6619 - 6502) + '\164' + chr(0b1100110) + '\055' + chr(0b111000)))(ZurHTci57aXw, IHMu1EGwZgDx)
return xafqLlk3kkUe(IDJ2eXGCBCDu, xafqLlk3kkUe(SXOLrMavuUCe(b'@\xe8\xees\xcd\xed'), '\144' + chr(0b110101 + 0o60) + chr(0b1100011) + chr(0b110000 + 0o77) + chr(0b10010 + 0o122) + chr(809 - 708))(chr(0b11101 + 0o130) + chr(0b1101111 + 0o5) + chr(102) + chr(0b1111 + 0o36) + chr(0b111000)))(ZurHTci57aXw, cMbll0QYhULo)
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/message_passing_attention.py
|
_compute_edge_transforms
|
def _compute_edge_transforms(node_states,
depth,
num_transforms,
name="transform"):
"""Helper function that computes transformation for keys and values.
Let B be the number of batches.
Let N be the number of nodes in the graph.
Let D be the size of the node hidden states.
Let K be the size of the attention keys/queries (total_key_depth).
Let V be the size of the attention values (total_value_depth).
Let T be the total number of transforms (num_transforms).
Computes the transforms for keys or values for attention.
* For each node N_j and edge type t, a key K_jt of size K is computed. When an
edge of type t goes from node N_j to any other node, K_jt is the key that is
in the attention process.
* For each node N_j and edge type t, a value V_jt of size V is computed. When
an edge of type t goes from node N_j to node N_i, Attention(Q_i, K_jt)
produces a weight w_ijt. The message sent along this edge is w_ijt * V_jt.
Args:
node_states: A tensor of shape [B, L, D]
depth: An integer (K or V)
num_transforms: An integer (T),
name: A name for the function
Returns:
x: A The attention keys or values for each node and edge type
(shape [B, N*T, K or V])
"""
node_shapes = common_layers.shape_list(node_states)
x = common_layers.dense(
node_states,
depth * num_transforms,
use_bias=False,
name=name)
batch = node_shapes[0] # B.
length = node_shapes[1] # N.
# Making the fourth dimension explicit by separating the vectors of size
# K*T (in k) and V*T (in v) into two-dimensional matrices with shape [K, T]
# (in k) and [V, T] in v.
#
x = tf.reshape(x, [batch, length, num_transforms, depth])
# Flatten out the fourth dimension.
x = tf.reshape(x, [batch, length * num_transforms, depth])
return x
|
python
|
def _compute_edge_transforms(node_states,
depth,
num_transforms,
name="transform"):
"""Helper function that computes transformation for keys and values.
Let B be the number of batches.
Let N be the number of nodes in the graph.
Let D be the size of the node hidden states.
Let K be the size of the attention keys/queries (total_key_depth).
Let V be the size of the attention values (total_value_depth).
Let T be the total number of transforms (num_transforms).
Computes the transforms for keys or values for attention.
* For each node N_j and edge type t, a key K_jt of size K is computed. When an
edge of type t goes from node N_j to any other node, K_jt is the key that is
in the attention process.
* For each node N_j and edge type t, a value V_jt of size V is computed. When
an edge of type t goes from node N_j to node N_i, Attention(Q_i, K_jt)
produces a weight w_ijt. The message sent along this edge is w_ijt * V_jt.
Args:
node_states: A tensor of shape [B, L, D]
depth: An integer (K or V)
num_transforms: An integer (T),
name: A name for the function
Returns:
x: A The attention keys or values for each node and edge type
(shape [B, N*T, K or V])
"""
node_shapes = common_layers.shape_list(node_states)
x = common_layers.dense(
node_states,
depth * num_transforms,
use_bias=False,
name=name)
batch = node_shapes[0] # B.
length = node_shapes[1] # N.
# Making the fourth dimension explicit by separating the vectors of size
# K*T (in k) and V*T (in v) into two-dimensional matrices with shape [K, T]
# (in k) and [V, T] in v.
#
x = tf.reshape(x, [batch, length, num_transforms, depth])
# Flatten out the fourth dimension.
x = tf.reshape(x, [batch, length * num_transforms, depth])
return x
|
[
"def",
"_compute_edge_transforms",
"(",
"node_states",
",",
"depth",
",",
"num_transforms",
",",
"name",
"=",
"\"transform\"",
")",
":",
"node_shapes",
"=",
"common_layers",
".",
"shape_list",
"(",
"node_states",
")",
"x",
"=",
"common_layers",
".",
"dense",
"(",
"node_states",
",",
"depth",
"*",
"num_transforms",
",",
"use_bias",
"=",
"False",
",",
"name",
"=",
"name",
")",
"batch",
"=",
"node_shapes",
"[",
"0",
"]",
"# B.",
"length",
"=",
"node_shapes",
"[",
"1",
"]",
"# N.",
"# Making the fourth dimension explicit by separating the vectors of size",
"# K*T (in k) and V*T (in v) into two-dimensional matrices with shape [K, T]",
"# (in k) and [V, T] in v.",
"#",
"x",
"=",
"tf",
".",
"reshape",
"(",
"x",
",",
"[",
"batch",
",",
"length",
",",
"num_transforms",
",",
"depth",
"]",
")",
"# Flatten out the fourth dimension.",
"x",
"=",
"tf",
".",
"reshape",
"(",
"x",
",",
"[",
"batch",
",",
"length",
"*",
"num_transforms",
",",
"depth",
"]",
")",
"return",
"x"
] |
Helper function that computes transformation for keys and values.
Let B be the number of batches.
Let N be the number of nodes in the graph.
Let D be the size of the node hidden states.
Let K be the size of the attention keys/queries (total_key_depth).
Let V be the size of the attention values (total_value_depth).
Let T be the total number of transforms (num_transforms).
Computes the transforms for keys or values for attention.
* For each node N_j and edge type t, a key K_jt of size K is computed. When an
edge of type t goes from node N_j to any other node, K_jt is the key that is
in the attention process.
* For each node N_j and edge type t, a value V_jt of size V is computed. When
an edge of type t goes from node N_j to node N_i, Attention(Q_i, K_jt)
produces a weight w_ijt. The message sent along this edge is w_ijt * V_jt.
Args:
node_states: A tensor of shape [B, L, D]
depth: An integer (K or V)
num_transforms: An integer (T),
name: A name for the function
Returns:
x: A The attention keys or values for each node and edge type
(shape [B, N*T, K or V])
|
[
"Helper",
"function",
"that",
"computes",
"transformation",
"for",
"keys",
"and",
"values",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/message_passing_attention.py#L251-L301
|
train
|
Helper function that computes the transforms for keys and values.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(51) + '\x31' + chr(1555 - 1507), 0b1000), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110010 + 0o5) + '\066', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(0b110001) + chr(2733 - 2678) + chr(830 - 782), 0o10), ehT0Px3KOsy9('\060' + chr(1790 - 1679) + '\x32' + chr(0b1111 + 0o44) + chr(0b110110), ord("\x08")), ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(0b1101111) + chr(1083 - 1033) + chr(49) + chr(0b110011), 0b1000), ehT0Px3KOsy9(chr(0b110 + 0o52) + '\157' + chr(486 - 435) + '\063' + chr(0b110010 + 0o0), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(0b110010) + chr(0b101100 + 0o13) + '\066', 0b1000), ehT0Px3KOsy9('\060' + chr(5523 - 5412) + chr(50) + chr(49) + chr(0b110101), 47359 - 47351), ehT0Px3KOsy9(chr(282 - 234) + chr(0b1101111) + chr(2448 - 2397) + chr(0b110100) + chr(48), 0b1000), ehT0Px3KOsy9(chr(0b11110 + 0o22) + chr(0b1101111) + chr(0b110001) + chr(0b10111 + 0o37) + '\x35', 52274 - 52266), ehT0Px3KOsy9(chr(48) + chr(0b1101000 + 0o7) + chr(0b1010 + 0o55) + '\062', 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(0b110011) + chr(0b110111) + '\x32', 0b1000), ehT0Px3KOsy9(chr(1667 - 1619) + '\x6f' + '\062' + chr(53) + '\x34', 18923 - 18915), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\x31' + chr(0b110101) + '\x36', 0o10), ehT0Px3KOsy9(chr(0b11101 + 0o23) + chr(0b1101100 + 0o3) + chr(50) + chr(0b110101), ord("\x08")), ehT0Px3KOsy9(chr(0b100000 + 0o20) + chr(0b1101111) + '\061' + chr(0b110000) + chr(471 - 420), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + chr(498 - 448) + '\063' + chr(0b110110), 8), ehT0Px3KOsy9(chr(48) + chr(0b111110 + 0o61) + chr(0b10111 + 0o34) + '\061' + chr(2243 - 2189), 24758 - 24750), ehT0Px3KOsy9('\060' + chr(111) + chr(0b1111 + 0o42) + '\061' + chr(439 - 384), 0b1000), ehT0Px3KOsy9('\060' + chr(0b1000000 + 0o57) + '\063' + chr(0b100100 + 0o15) + chr(0b10101 + 0o40), 0o10), ehT0Px3KOsy9(chr(48) + '\x6f' + chr(2589 - 2538) + '\067' + '\060', 3431 - 3423), ehT0Px3KOsy9('\060' + '\157' + chr(0b11110 + 0o24) + chr(0b1010 + 0o52) + chr(582 - 529), 0b1000), ehT0Px3KOsy9(chr(0b11 + 0o55) + '\x6f' + chr(50) + chr(0b110001) + chr(48), 39367 - 39359), ehT0Px3KOsy9('\x30' + chr(0b1100100 + 0o13) + chr(50) + chr(52) + chr(50), 18437 - 18429), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(2396 - 2343) + chr(0b111 + 0o56), ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + '\061' + '\x35' + '\x37', 23889 - 23881), ehT0Px3KOsy9(chr(1289 - 1241) + chr(1646 - 1535) + chr(0b101010 + 0o11) + chr(560 - 506) + '\x34', 63862 - 63854), ehT0Px3KOsy9('\060' + chr(6661 - 6550) + chr(0b110010) + '\x30' + '\x30', 0b1000), ehT0Px3KOsy9(chr(1991 - 1943) + '\157' + chr(51) + chr(0b100110 + 0o14) + chr(0b110001), 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(0b10110 + 0o131) + chr(0b110010) + '\x35' + chr(49), 0b1000), ehT0Px3KOsy9(chr(1445 - 1397) + '\157' + '\x36' + '\065', ord("\x08")), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\066' + '\067', 0b1000), ehT0Px3KOsy9(chr(0b110000) + chr(119 - 8) + '\x32' + '\064' + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(1138 - 1090) + '\157' + chr(0b101010 + 0o11) + chr(0b1011 + 0o46) + '\065', 8), ehT0Px3KOsy9(chr(0b110000) + '\157' + '\x32' + chr(874 - 822) + '\064', 8), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\061' + chr(477 - 424) + chr(0b11111 + 0o26), 0b1000), ehT0Px3KOsy9(chr(873 - 825) + '\x6f' + chr(0b110001) + '\064' + '\067', 14027 - 14019), ehT0Px3KOsy9(chr(527 - 479) + chr(111) + '\065' + chr(0b100110 + 0o20), 0b1000), ehT0Px3KOsy9(chr(1605 - 1557) + chr(0b1101111) + '\062' + chr(1929 - 1876) + chr(52), 8), ehT0Px3KOsy9(chr(0b10111 + 0o31) + chr(111) + '\061' + chr(0b110011) + '\066', 0b1000)][WVxHKyX45z_L % ehT0Px3KOsy9(chr(0b100101 + 0o13) + '\x6f' + chr(0b1001 + 0o54) + '\x30', 0o10)] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xe6'), chr(100) + chr(1265 - 1164) + chr(99) + '\157' + '\x64' + chr(0b100011 + 0o102))(chr(0b1100000 + 0o25) + chr(0b1100101 + 0o17) + chr(102) + chr(0b100111 + 0o6) + chr(2884 - 2828)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def bs2jGu0mCIzC(PrV0svX_Vz1w, UEys4_lSwsID, bDLImSNrRa8X, AIvJRzLdDfgF=xafqLlk3kkUe(SXOLrMavuUCe(b'\xbcL\x19\xf0\xf8\xbc\xd1\xff\x8d'), chr(100) + '\x65' + chr(0b111 + 0o134) + '\x6f' + '\x64' + chr(0b11001 + 0o114))('\x75' + chr(0b1101000 + 0o14) + '\146' + '\055' + chr(56))):
xrYmzdfeL9DB = jSKPaHwSAfVv.shape_list(PrV0svX_Vz1w)
OeWW0F1dBPRQ = jSKPaHwSAfVv.dense(PrV0svX_Vz1w, UEys4_lSwsID * bDLImSNrRa8X, use_bias=ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(48), 0b1000), name=AIvJRzLdDfgF)
dNwAahu8tvoY = xrYmzdfeL9DB[ehT0Px3KOsy9(chr(0b110000) + chr(7640 - 7529) + chr(1468 - 1420), 8)]
CHAOgk5VCHH_ = xrYmzdfeL9DB[ehT0Px3KOsy9('\060' + chr(0b1101111) + chr(0b110001), 0o10)]
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reshape(OeWW0F1dBPRQ, [dNwAahu8tvoY, CHAOgk5VCHH_, bDLImSNrRa8X, UEys4_lSwsID])
OeWW0F1dBPRQ = IDJ2eXGCBCDu.reshape(OeWW0F1dBPRQ, [dNwAahu8tvoY, CHAOgk5VCHH_ * bDLImSNrRa8X, UEys4_lSwsID])
return OeWW0F1dBPRQ
|
tensorflow/tensor2tensor
|
tensor2tensor/layers/message_passing_attention.py
|
compute_mpnn_qkv
|
def compute_mpnn_qkv(node_states,
total_key_depth,
total_value_depth,
num_transforms):
"""Computes query, key and value for edge matrices.
Let B be the number of batches.
Let N be the number of nodes in the graph.
Let D be the size of the node hidden states.
Let K be the size of the attention keys/queries (total_key_depth).
Let V be the size of the attention values (total_value_depth).
Let T be the total number of transforms (num_transforms).
Computes the queries, keys, and values for attention.
* For each node N_i in the graph, a query Q_i of size K is computed. This
query is used to determine the relative weights to give to each of the
node's incoming edges.
* For each node N_j and edge type t, a key K_jt of size K is computed. When an
edge of type t goes from node N_j to any other node, K_jt is the key that is
in the attention process.
* For each node N_j and edge type t, a value V_jt of size V is computed. When
an edge of type t goes from node N_j to node N_i, Attention(Q_i, K_jt)
produces a weight w_ijt. The message sent along this edge is w_ijt * V_jt.
Args:
node_states: A Tensor with shape [B, N, D].
total_key_depth: an integer (K).
total_value_depth: an integer (V).
num_transforms: a integer specifying number of transforms (T). This is
typically the number of edge types.
Returns:
q: The attention queries for each destination node (shape [B, N, K]).
k: The attention keys for each node and edge type (shape [B, N*T, K]).
v: The attention values for each node and edge type (shape [B, N*T, V]).
"""
# node_states is initially a tensor with shape [B, N, D]. The call to dense
# creates a D x K kernel that serves as a fully-connected layer.
#
# For each possible batch b and node n in the first two dimensions of
# node_states, the corresponding size-D vector (the third dimension of
# node_states) is the hidden state for node n in batch b. Each of these size-D
# vectors is multiplied by the kernel to produce an attention query of size K.
# The result is a tensor of size [B, N, K] containing the attention queries
# for each node in each batch.
q = common_layers.dense(
node_states, total_key_depth, use_bias=False, name="q_mpnn")
# Creates the attention keys in a manner similar to the process of creating
# the attention queries. One key is created for each type of outgoing edge the
# corresponding node might have, meaning k will have shape [B, N, K*T].
k = _compute_edge_transforms(node_states,
total_key_depth,
num_transforms,
name="k_mpnn")
v = _compute_edge_transforms(node_states,
total_value_depth,
num_transforms,
name="v_mpnn")
return q, k, v
|
python
|
def compute_mpnn_qkv(node_states,
total_key_depth,
total_value_depth,
num_transforms):
"""Computes query, key and value for edge matrices.
Let B be the number of batches.
Let N be the number of nodes in the graph.
Let D be the size of the node hidden states.
Let K be the size of the attention keys/queries (total_key_depth).
Let V be the size of the attention values (total_value_depth).
Let T be the total number of transforms (num_transforms).
Computes the queries, keys, and values for attention.
* For each node N_i in the graph, a query Q_i of size K is computed. This
query is used to determine the relative weights to give to each of the
node's incoming edges.
* For each node N_j and edge type t, a key K_jt of size K is computed. When an
edge of type t goes from node N_j to any other node, K_jt is the key that is
in the attention process.
* For each node N_j and edge type t, a value V_jt of size V is computed. When
an edge of type t goes from node N_j to node N_i, Attention(Q_i, K_jt)
produces a weight w_ijt. The message sent along this edge is w_ijt * V_jt.
Args:
node_states: A Tensor with shape [B, N, D].
total_key_depth: an integer (K).
total_value_depth: an integer (V).
num_transforms: a integer specifying number of transforms (T). This is
typically the number of edge types.
Returns:
q: The attention queries for each destination node (shape [B, N, K]).
k: The attention keys for each node and edge type (shape [B, N*T, K]).
v: The attention values for each node and edge type (shape [B, N*T, V]).
"""
# node_states is initially a tensor with shape [B, N, D]. The call to dense
# creates a D x K kernel that serves as a fully-connected layer.
#
# For each possible batch b and node n in the first two dimensions of
# node_states, the corresponding size-D vector (the third dimension of
# node_states) is the hidden state for node n in batch b. Each of these size-D
# vectors is multiplied by the kernel to produce an attention query of size K.
# The result is a tensor of size [B, N, K] containing the attention queries
# for each node in each batch.
q = common_layers.dense(
node_states, total_key_depth, use_bias=False, name="q_mpnn")
# Creates the attention keys in a manner similar to the process of creating
# the attention queries. One key is created for each type of outgoing edge the
# corresponding node might have, meaning k will have shape [B, N, K*T].
k = _compute_edge_transforms(node_states,
total_key_depth,
num_transforms,
name="k_mpnn")
v = _compute_edge_transforms(node_states,
total_value_depth,
num_transforms,
name="v_mpnn")
return q, k, v
|
[
"def",
"compute_mpnn_qkv",
"(",
"node_states",
",",
"total_key_depth",
",",
"total_value_depth",
",",
"num_transforms",
")",
":",
"# node_states is initially a tensor with shape [B, N, D]. The call to dense",
"# creates a D x K kernel that serves as a fully-connected layer.",
"#",
"# For each possible batch b and node n in the first two dimensions of",
"# node_states, the corresponding size-D vector (the third dimension of",
"# node_states) is the hidden state for node n in batch b. Each of these size-D",
"# vectors is multiplied by the kernel to produce an attention query of size K.",
"# The result is a tensor of size [B, N, K] containing the attention queries",
"# for each node in each batch.",
"q",
"=",
"common_layers",
".",
"dense",
"(",
"node_states",
",",
"total_key_depth",
",",
"use_bias",
"=",
"False",
",",
"name",
"=",
"\"q_mpnn\"",
")",
"# Creates the attention keys in a manner similar to the process of creating",
"# the attention queries. One key is created for each type of outgoing edge the",
"# corresponding node might have, meaning k will have shape [B, N, K*T].",
"k",
"=",
"_compute_edge_transforms",
"(",
"node_states",
",",
"total_key_depth",
",",
"num_transforms",
",",
"name",
"=",
"\"k_mpnn\"",
")",
"v",
"=",
"_compute_edge_transforms",
"(",
"node_states",
",",
"total_value_depth",
",",
"num_transforms",
",",
"name",
"=",
"\"v_mpnn\"",
")",
"return",
"q",
",",
"k",
",",
"v"
] |
Computes query, key and value for edge matrices.
Let B be the number of batches.
Let N be the number of nodes in the graph.
Let D be the size of the node hidden states.
Let K be the size of the attention keys/queries (total_key_depth).
Let V be the size of the attention values (total_value_depth).
Let T be the total number of transforms (num_transforms).
Computes the queries, keys, and values for attention.
* For each node N_i in the graph, a query Q_i of size K is computed. This
query is used to determine the relative weights to give to each of the
node's incoming edges.
* For each node N_j and edge type t, a key K_jt of size K is computed. When an
edge of type t goes from node N_j to any other node, K_jt is the key that is
in the attention process.
* For each node N_j and edge type t, a value V_jt of size V is computed. When
an edge of type t goes from node N_j to node N_i, Attention(Q_i, K_jt)
produces a weight w_ijt. The message sent along this edge is w_ijt * V_jt.
Args:
node_states: A Tensor with shape [B, N, D].
total_key_depth: an integer (K).
total_value_depth: an integer (V).
num_transforms: a integer specifying number of transforms (T). This is
typically the number of edge types.
Returns:
q: The attention queries for each destination node (shape [B, N, K]).
k: The attention keys for each node and edge type (shape [B, N*T, K]).
v: The attention values for each node and edge type (shape [B, N*T, V]).
|
[
"Computes",
"query",
"key",
"and",
"value",
"for",
"edge",
"matrices",
"."
] |
272500b6efe353aeb638d2745ed56e519462ca31
|
https://github.com/tensorflow/tensor2tensor/blob/272500b6efe353aeb638d2745ed56e519462ca31/tensor2tensor/layers/message_passing_attention.py#L304-L364
|
train
|
This function computes the query key and value for edge matrices.
|
Pu7Z6IJCgH3a,vcEHXBQXuDuh,sHOWSIAKtU58,ZVWAAMjVVHHl,qRin5pdYOdbB,IySsVMyKT3tF,FwEHNICjJCy0,yISIa0MMKKfB,GAtvbI59wr0o,OmNM6rT0Sgul,gu1MSKhYvigU,S2TTo9DhhiSh,aaLV7ZjAfkcR,ker4pIJmdvxf,WaQEaQCVMQ03,xV97BFGi0hY9,YnM1HtHE4j7G,X5FyJb4ToTo6,jLmadlzMdunT,GGFwFLsDF9Fv,prtR0Uw1GMh5,oNamnshN4dFG,QZzQeAYvsoum,VHAt7CcYKC2T,cKsTbNGLtp_O,sR2sPcm7Zrfn,yROw0HWBk0Qc,j9rjMYnN2BMp,hIlP7994qj8O,_fsda0v2_OKU,o0CgT5HPthxA,DXjfarvgFnbl,RQ6CSRrFArYB,RouZF7bjEXAv,jIl9qoALCRyb,bdLuls3EQFSd,FXUco0R3m83n,V5s4UV3vwoyK,Q6d3QdTENfxw,sbc9gub6LIFp,QWgp4ELTmqy4,_zJ24Vce7wp0,KlPSljPzIJ_u,N5Ee6d9YGQ_x,yDcnbVVBZ5VZ,OTstrxJfIC1n,GXwwnDRMCHJX,a9IKoVgO_m3w,GNd6AVvhYicE,ixtrydDuthdu,n0ZkatoveZpF,eh4BeXwijHpf,ZMHESMWYyt8h,hr2QaoivbFQ2,Iiw8L0MH5qfg,koCeDPYTrOFe,qqrhSmCSbbqk,pz9FlfzsWoy1,BXIwDASQ0Qkq,NL8dtWOpbcjF,_bikzMuRfbJG,sznFqDbNBHlx,ZsDPvpP4xdo3,cW7yQuyEnJ6E,KOHQGQ8qLDWm,NE1Yam2HHroQ,ygAzbDzrvRMh,SBRjvOU1ufVC,hOkXjmluKZfJ,q1QCh3W88sgk,TLbJ60djyws0,rIcPej9ZqMqV,WTxpD_zsEOh2,LgE_IO_tHXvM,Kk1hd194VKEC,OZYzwAeSQh7N,jFWsnpHpAUWz,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp,Lt3jp3Wjtj_1,OgxWTx4GSNFx,Dl48nj1rbi23,gUjKZptQBOom,UVSi4XW7eBIM,TtvdWC885wQi,hyjPAJYKYCCT,WbBjf8Y7v9VN,LXFmLC1F9ebP,QC9iu2kLpS8s,QOfmzcVJsrp8,tzcpInYwBvYW,iDQ_gSK8V7h0,Rurm1zTRfSmY,reqGiMiVQ77y,bsS9P6_LpdIe,sbGAZlkZOtyh,Cf_Qef15s3_F,eX02hlZjMfR0,wLqBDw8l0eIm,g1Uy6IV0tyJQ,f9CsFWzvg0Vq,YlkZvXL8qwsX,MCqssyYhLtLC,bpgWCAbiJWkL,CMUdZtaORwo4,hi1V0ySZcNds,kkSX4ccExqw4,V4roHaS3Ppej,o8rvoPw8ep3k,xafqLlk3kkUe,h0qciNl3EEEj,lot1PSoAwYhj,xfhwxiBOH72k,HcyiPkCViZiX,fOIXYo9a1WNS,z8EhBlYI2Bx4,Y3jVKaC8LEDU,ehT0Px3KOsy9,PlSM16l2KDPD,J6u1YyThfhgG,ZdP978XkGspL,c2A0yzQpDQB3,I7ZO3Ma9cXBb,YyaZ4tpXu4lf,eHmS9durw_Vs,abA97kOQKaLo,tsdjvlgh9gDP,VTYZGD68sBIs,Dx22bkKPdt5d,nSwwHEeM4cxI,sR_24x3xd4bh,xmV2riMOClNT,_fwkIVCGgtAN,Jp8aZ6mjyZZT,eO8Xfv8UVFey,zLUzGokYBM2Z,FL7SmUoxlR9h,k6bl9sLammpH,vQr8gNKaIaWE,S6hV9M2g7fO0,RFiwrCZH9Ie6,jB_HdqgHmVpI,MVEN8G6CxlvR,t0rOMsrOC7R_,W3g84rNiEdDQ,vUlqIvNSaRMa,gDnh40_OUDCn,M8_cKLkHVB2V,xkxBmo49x2An,KNx0Ujaz9UM0,KNyTy8rYcwji,wmQmyeWBmUpv,p1G5VS3dE_Ss,pZ0NK2y6HRbn,HByLaO1XdVEe,pgRJLRS7Iy8j,OZYzwAeSQh7N,tmzuw0hjv33u,RwRZiUMA3VWp,Gbej4oZqKLA6,TqkAMbUz4aLg,rw68imZ2Ikxp=ArithmeticError,AssertionError,AttributeError,BaseException,BlockingIOError,BrokenPipeError,BufferError,BytesWarning,ChildProcessError,ConnectionAbortedError,ConnectionError,ConnectionRefusedError,ConnectionResetError,DeprecationWarning,EOFError,Ellipsis,EncodingWarning,EnvironmentError,Exception,False,FileExistsError,FileNotFoundError,FloatingPointError,FutureWarning,GeneratorExit,IOError,ImportError,ImportWarning,IndentationError,IndexError,InterruptedError,IsADirectoryError,KeyError,KeyboardInterrupt,LookupError,MemoryError,ModuleNotFoundError,NameError,None,NotADirectoryError,NotImplemented,NotImplementedError,OSError,OverflowError,PendingDeprecationWarning,PermissionError,ProcessLookupError,RecursionError,ReferenceError,ResourceWarning,RuntimeError,RuntimeWarning,StopAsyncIteration,StopIteration,SyntaxError,SyntaxWarning,SystemError,SystemExit,TabError,TimeoutError,True,TypeError,UnboundLocalError,UnicodeDecodeError,UnicodeEncodeError,UnicodeError,UnicodeTranslateError,UnicodeWarning,UserWarning,ValueError,Warning,WindowsError,ZeroDivisionError,__build_class__,__debug__,__doc__,__import__,__loader__,__name__,__package__,__spec__,abs,aiter,all,anext,any,ascii,bin,bool,breakpoint,bytearray,bytes,callable,chr,classmethod,compile,complex,copyright,credits,delattr,dict,dir,divmod,enumerate,eval,exec,exit,filter,float,format,frozenset,getattr,globals,hasattr,hash,help,hex,id,input,int,isinstance,issubclass,iter,len,license,list,locals,map,max,memoryview,min,next,object,oct,open,ord,pow,print,property,quit,range,repr,reversed,round,set,setattr,slice,sorted,staticmethod,str,sum,super,tuple,type,vars,zip,__builtins__,__cached__,__doc__,__file__,__loader__,__name__,__package__,__spec__
SXOLrMavuUCe = lambda XbwU38w7NW8n: QOfmzcVJsrp8([OeWW0F1dBPRQ ^ [ehT0Px3KOsy9(chr(0b110000) + chr(0b101001 + 0o106) + chr(0b11110 + 0o25) + '\062' + chr(54), 0b1000), ehT0Px3KOsy9(chr(1076 - 1028) + chr(0b1101111) + chr(1243 - 1193) + '\061' + chr(52), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(531 - 480) + '\061' + chr(0b110010), ord("\x08")), ehT0Px3KOsy9(chr(875 - 827) + chr(0b1001110 + 0o41) + '\063' + chr(1935 - 1886) + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b100100 + 0o14) + chr(111) + chr(0b1101 + 0o44) + '\x30' + chr(0b110010), 0b1000), ehT0Px3KOsy9(chr(48) + chr(6264 - 6153) + chr(0b110011) + chr(50) + chr(0b1110 + 0o42), 61032 - 61024), ehT0Px3KOsy9(chr(0b11 + 0o55) + chr(9726 - 9615) + chr(0b0 + 0o61) + chr(0b10000 + 0o46) + chr(52), 0o10), ehT0Px3KOsy9(chr(0b1111 + 0o41) + '\157' + chr(50) + '\060' + '\x31', 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\062' + chr(50), 0b1000), ehT0Px3KOsy9(chr(48) + '\157' + '\061' + '\x31' + chr(55), 0o10), ehT0Px3KOsy9(chr(2167 - 2119) + chr(111) + chr(0b10010 + 0o40) + chr(0b110001) + chr(0b10100 + 0o41), ord("\x08")), ehT0Px3KOsy9('\060' + chr(111) + chr(50) + chr(2675 - 2621) + chr(53), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(0b1101111) + '\x32' + chr(48) + chr(549 - 501), 55717 - 55709), ehT0Px3KOsy9('\x30' + '\157' + chr(0b110010) + '\063' + chr(52), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(111) + '\063' + chr(53) + chr(1670 - 1620), 65533 - 65525), ehT0Px3KOsy9('\060' + chr(0b10001 + 0o136) + '\x36' + '\063', 51792 - 51784), ehT0Px3KOsy9(chr(0b110000) + '\x6f' + chr(165 - 114) + chr(0b11111 + 0o26) + '\x34', 52944 - 52936), ehT0Px3KOsy9(chr(0b1101 + 0o43) + chr(2518 - 2407) + chr(0b101111 + 0o2) + '\063' + chr(55), 57112 - 57104), ehT0Px3KOsy9(chr(1132 - 1084) + '\157' + '\x31' + chr(0b10101 + 0o33), 0b1000), ehT0Px3KOsy9(chr(232 - 184) + chr(3987 - 3876) + '\x33' + chr(52) + '\x31', 25171 - 25163), ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(111) + '\x31' + '\064' + '\x34', ord("\x08")), ehT0Px3KOsy9(chr(1706 - 1658) + chr(7165 - 7054) + chr(49) + chr(55) + '\066', 0o10), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(0b10100 + 0o36) + '\x33' + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(0b10000 + 0o40) + chr(111) + '\x32' + '\x31' + chr(0b110100), 8), ehT0Px3KOsy9(chr(0b100000 + 0o20) + chr(0b1101111) + chr(0b110001) + '\061' + '\x36', 0o10), ehT0Px3KOsy9('\x30' + chr(111) + chr(0b110011) + chr(808 - 757) + chr(0b10000 + 0o47), 0o10), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(1881 - 1831) + chr(1835 - 1781) + chr(0b1 + 0o60), 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1101111) + '\067', 0b1000), ehT0Px3KOsy9('\x30' + chr(0b1001100 + 0o43) + '\063' + '\061' + chr(2045 - 1997), ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(2006 - 1956) + '\067' + '\x36', ord("\x08")), ehT0Px3KOsy9(chr(48) + chr(0b1101111) + chr(55) + chr(1088 - 1034), 0b1000), ehT0Px3KOsy9(chr(0b101010 + 0o6) + '\x6f' + '\063' + chr(0b100 + 0o61) + '\x33', ord("\x08")), ehT0Px3KOsy9(chr(0b110000) + chr(111) + chr(0b101 + 0o55) + chr(0b1001 + 0o50) + chr(0b1110 + 0o47), 8), ehT0Px3KOsy9('\060' + chr(0b1100110 + 0o11) + '\x31' + '\x33' + chr(0b110101), 0b1000), ehT0Px3KOsy9(chr(0b111 + 0o51) + '\157' + chr(51) + chr(584 - 533) + chr(52), 0o10), ehT0Px3KOsy9(chr(985 - 937) + chr(0b1101111) + chr(1656 - 1607) + chr(1133 - 1084) + chr(0b110010), 0o10), ehT0Px3KOsy9(chr(0b100010 + 0o16) + chr(0b1101000 + 0o7) + chr(0b110001) + '\x34' + '\063', 0o10), ehT0Px3KOsy9('\060' + chr(111) + chr(49) + chr(0b1011 + 0o46) + chr(0b110100), 0o10), ehT0Px3KOsy9(chr(48) + '\157' + chr(0b1101 + 0o44), ord("\x08")), ehT0Px3KOsy9('\x30' + '\x6f' + chr(49) + chr(0b101100 + 0o11) + '\x36', ord("\x08"))][WVxHKyX45z_L % ehT0Px3KOsy9(chr(417 - 369) + chr(1445 - 1334) + '\065' + chr(48), ord("\x08"))] for (WVxHKyX45z_L, OeWW0F1dBPRQ) in YlkZvXL8qwsX(XbwU38w7NW8n)])
def NPPHb59961Bv(RqocVGOryNPv, _CF03Rifpmdh):
try:
return jFWsnpHpAUWz(RqocVGOryNPv + xafqLlk3kkUe(SXOLrMavuUCe(b'\xf8'), '\144' + chr(0b1001011 + 0o32) + '\143' + chr(7267 - 7156) + chr(5006 - 4906) + '\145')(chr(10248 - 10131) + chr(0b1110100) + chr(4689 - 4587) + chr(45) + chr(0b101111 + 0o11)) + _CF03Rifpmdh)
except yROw0HWBk0Qc:
return jFWsnpHpAUWz(RqocVGOryNPv)
def xCbuoIQ9Z7cv(PrV0svX_Vz1w, _jxqy0P17UFy, F9lUBuHPQMmX, bDLImSNrRa8X):
WtwjCI_b3w8O = jSKPaHwSAfVv.dense(PrV0svX_Vz1w, _jxqy0P17UFy, use_bias=ehT0Px3KOsy9(chr(0b101010 + 0o6) + chr(4103 - 3992) + chr(0b110000), ord("\x08")), name=xafqLlk3kkUe(SXOLrMavuUCe(b'\xa7\xd3\xa7\xbd,\xbe'), chr(100) + chr(4536 - 4435) + chr(5810 - 5711) + chr(0b1001100 + 0o43) + '\x64' + chr(101))(chr(117) + chr(116) + '\x66' + chr(0b101101) + '\x38'))
OolUPRJhRaJd = bs2jGu0mCIzC(PrV0svX_Vz1w, _jxqy0P17UFy, bDLImSNrRa8X, name=xafqLlk3kkUe(SXOLrMavuUCe(b'\xbd\xd3\xa7\xbd,\xbe'), '\x64' + '\x65' + chr(0b11011 + 0o110) + '\157' + chr(0b1100010 + 0o2) + chr(0b111100 + 0o51))('\x75' + '\x74' + '\x66' + chr(0b0 + 0o55) + '\x38'))
cMbll0QYhULo = bs2jGu0mCIzC(PrV0svX_Vz1w, F9lUBuHPQMmX, bDLImSNrRa8X, name=xafqLlk3kkUe(SXOLrMavuUCe(b'\xa0\xd3\xa7\xbd,\xbe'), '\144' + chr(0b100100 + 0o101) + '\143' + chr(9226 - 9115) + '\x64' + '\x65')('\x75' + chr(116) + '\x66' + chr(262 - 217) + '\070'))
return (WtwjCI_b3w8O, OolUPRJhRaJd, cMbll0QYhULo)
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.