Skip to content

GitLab

"docs/source/philosophy.mdx" did not exist on "b0c7d2ec58a8080780fb21327cc7cd7e1a177780"
Unverified Commit 2295d783 authored by Sylvain Gugger's avatar Sylvain Gugger Committed by GitHub
Browse files

Copy tokenizer files in each of their repo (#10624) 

* Move tokenizer files in each repo

* Fix mBART50 tests

* Fix mBART tests

* Fix Marian tests

* Update templates
parent d26b37e7
Show whitespace changes
Inline Side-by-side
Showing with 137 additions and 108 deletions
src/transformers/models/m2m_100/tokenization_m2m_100.py
View file @ 2295d783
...@@ -37,17 +37,21 @@ VOCAB_FILES_NAMES = { ...@@ -37,17 +37,21 @@ VOCAB_FILES_NAMES = {
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"facebook/m2m100_418M": "https://huggingface.co/facebook/m2m100_418M/resolve/main/vocab.json", "facebook/m2m100_418M": "https://huggingface.co/facebook/m2m100_418M/resolve/main/vocab.json",
"facebook/m2m100_1.2B": "https://huggingface.co/facebook/m2m100_1.2B/resolve/main/vocab.json",
}, },
"spm_file": { "spm_file": {
"facebook/m2m100_418M": "https://huggingface.co/facebook/m2m100_418M/resolve/main/sentencepiece.bpe.model", "facebook/m2m100_418M": "https://huggingface.co/facebook/m2m100_418M/resolve/main/sentencepiece.bpe.model",
"facebook/m2m100_1.2B": "https://huggingface.co/facebook/m2m100_1.2B/resolve/main/sentencepiece.bpe.model",
}, },
"tokenizer_config_file": { "tokenizer_config_file": {
"facebook/m2m100_418M": "https://huggingface.co/facebook/m2m100_418M/resolve/main/tokenizer_config.json", "facebook/m2m100_418M": "https://huggingface.co/facebook/m2m100_418M/resolve/main/tokenizer_config.json",
"facebook/m2m100_1.2B": "https://huggingface.co/facebook/m2m100_1.2B/resolve/main/tokenizer_config.json",
}, },
} }
ALL_M2M100_MODELS = ["facebook/m2m100_418M", "facebook/m2m100_1.2B"] PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
SPM_URL = "https://huggingface.co/facebook/m2m100_418M/resolve/main/sentence.bpe.model" "facebook/m2m100_418M": 1024,
}
# fmt: off # fmt: off
FAIRSEQ_LANGUAGE_CODES = ["af", "am", "ar", "ast", "az", "ba", "be", "bg", "bn", "br", "bs", "ca", "ceb", "cs", "cy", "da", "de", "el", "en", "es", "et", "fa", "ff", "fi", "fr", "fy", "ga", "gd", "gl", "gu", "ha", "he", "hi", "hr", "ht", "hu", "hy", "id", "ig", "ilo", "is", "it", "ja", "jv", "ka", "kk", "km", "kn", "ko", "lb", "lg", "ln", "lo", "lt", "lv", "mg", "mk", "ml", "mn", "mr", "ms", "my", "ne", "nl", "no", "ns", "oc", "or", "pa", "pl", "ps", "pt", "ro", "ru", "sd", "si", "sk", "sl", "so", "sq", "sr", "ss", "su", "sv", "sw", "ta", "th", "tl", "tn", "tr", "uk", "ur", "uz", "vi", "wo", "xh", "yi", "yo", "zh", "zu"] FAIRSEQ_LANGUAGE_CODES = ["af", "am", "ar", "ast", "az", "ba", "be", "bg", "bn", "br", "bs", "ca", "ceb", "cs", "cy", "da", "de", "el", "en", "es", "et", "fa", "ff", "fi", "fr", "fy", "ga", "gd", "gl", "gu", "ha", "he", "hi", "hr", "ht", "hu", "hy", "id", "ig", "ilo", "is", "it", "ja", "jv", "ka", "kk", "km", "kn", "ko", "lb", "lg", "ln", "lo", "lt", "lv", "mg", "mk", "ml", "mn", "mr", "ms", "my", "ne", "nl", "no", "ns", "oc", "or", "pa", "pl", "ps", "pt", "ro", "ru", "sd", "si", "sk", "sl", "so", "sq", "sr", "ss", "su", "sv", "sw", "ta", "th", "tl", "tn", "tr", "uk", "ur", "uz", "vi", "wo", "xh", "yi", "yo", "zh", "zu"]
...@@ -96,7 +100,7 @@ class M2M100Tokenizer(PreTrainedTokenizer): ...@@ -96,7 +100,7 @@ class M2M100Tokenizer(PreTrainedTokenizer):
""" """
vocab_files_names = VOCAB_FILES_NAMES vocab_files_names = VOCAB_FILES_NAMES
max_model_input_sizes = {m: 1024 for m in ALL_M2M100_MODELS} max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
model_input_names = ["input_ids", "attention_mask"] model_input_names = ["input_ids", "attention_mask"]
......
src/transformers/models/marian/tokenization_marian.py
View file @ 2295d783
...@@ -25,7 +25,7 @@ import sentencepiece ...@@ -25,7 +25,7 @@ import sentencepiece
from ...tokenization_utils import PreTrainedTokenizer from ...tokenization_utils import PreTrainedTokenizer
vocab_files_names = { VOCAB_FILES_NAMES = {
"source_spm": "source.spm", "source_spm": "source.spm",
"target_spm": "target.spm", "target_spm": "target.spm",
"vocab": "vocab.json", "vocab": "vocab.json",
...@@ -33,11 +33,17 @@ vocab_files_names = { ...@@ -33,11 +33,17 @@ vocab_files_names = {
} }
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"source_spm": {"Helsinki-NLP/opus-mt-en-de": "https://cdn.huggingface.co/Helsinki-NLP/opus-mt-en-de/source.spm"}, "source_spm": {
"target_spm": {"Helsinki-NLP/opus-mt-en-de": "https://cdn.huggingface.co/Helsinki-NLP/opus-mt-en-de/target.spm"}, "Helsinki-NLP/opus-mt-en-de": "https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/source.spm"
"vocab": {"Helsinki-NLP/opus-mt-en-de": "https://cdn.huggingface.co/Helsinki-NLP/opus-mt-en-de/vocab.json"}, },
"target_spm": {
"Helsinki-NLP/opus-mt-en-de": "https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/target.spm"
},
"vocab": {
"Helsinki-NLP/opus-mt-en-de": "https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/vocab.json"
},
"tokenizer_config_file": { "tokenizer_config_file": {
"Helsinki-NLP/opus-mt-en-de": "https://cdn.huggingface.co/Helsinki-NLP/opus-mt-en-de/tokenizer_config.json" "Helsinki-NLP/opus-mt-en-de": "https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/tokenizer_config.json"
}, },
} }
...@@ -91,7 +97,7 @@ class MarianTokenizer(PreTrainedTokenizer): ...@@ -91,7 +97,7 @@ class MarianTokenizer(PreTrainedTokenizer):
>>> outputs = model(**inputs) should work >>> outputs = model(**inputs) should work
""" """
vocab_files_names = vocab_files_names vocab_files_names = VOCAB_FILES_NAMES
pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
pretrained_init_configuration = PRETRAINED_INIT_CONFIGURATION pretrained_init_configuration = PRETRAINED_INIT_CONFIGURATION
max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
......
src/transformers/models/mbart/tokenization_mbart.py
View file @ 2295d783
...@@ -23,8 +23,20 @@ from ..xlm_roberta.tokenization_xlm_roberta import XLMRobertaTokenizer ...@@ -23,8 +23,20 @@ from ..xlm_roberta.tokenization_xlm_roberta import XLMRobertaTokenizer
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
_all_mbart_models = ["facebook/mbart-large-en-ro", "facebook/mbart-large-cc25"]
SPM_URL = "https://huggingface.co/facebook/mbart-large-en-ro/resolve/main/sentence.bpe.model" VOCAB_FILES_NAMES = {"vocab_file": "sentencepiece.bpe.model"}
PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": {
"facebook/mbart-large-en-ro": "https://huggingface.co/facebook/mbart-large-en-ro/resolve/main/sentencepiece.bpe.model",
"facebook/mbart-large-cc25": "https://huggingface.co/facebook/mbart-large-cc25/resolve/main/sentencepiece.bpe.model",
}
}
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"facebook/mbart-large-en-ro": 1024,
"facebook/mbart-large-cc25": 1024,
}
FAIRSEQ_LANGUAGE_CODES = [ FAIRSEQ_LANGUAGE_CODES = [
"ar_AR", "ar_AR",
...@@ -78,9 +90,9 @@ class MBartTokenizer(XLMRobertaTokenizer): ...@@ -78,9 +90,9 @@ class MBartTokenizer(XLMRobertaTokenizer):
>>> inputs["labels"] = labels["input_ids"] >>> inputs["labels"] = labels["input_ids"]
""" """
vocab_files_names = {"vocab_file": "sentencepiece.bpe.model"} vocab_files_names = VOCAB_FILES_NAMES
max_model_input_sizes = {m: 1024 for m in _all_mbart_models} max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
pretrained_vocab_files_map = {"vocab_file": {m: SPM_URL for m in _all_mbart_models}} pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
prefix_tokens: List[int] = [] prefix_tokens: List[int] = []
suffix_tokens: List[int] = [] suffix_tokens: List[int] = []
......
src/transformers/models/mbart/tokenization_mbart50.py
View file @ 2295d783
...@@ -30,8 +30,15 @@ SPIECE_UNDERLINE = "▁" ...@@ -30,8 +30,15 @@ SPIECE_UNDERLINE = "▁"
VOCAB_FILES_NAMES = {"vocab_file": "sentencepiece.bpe.model"} VOCAB_FILES_NAMES = {"vocab_file": "sentencepiece.bpe.model"}
_all_mbart50_models = ["facebook/mbart-large-50-one-to-many-mmt"] PRETRAINED_VOCAB_FILES_MAP = {
SPM_URL = "https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/sentencepiece.bpe.model" "vocab_file": {
"facebook/mbart-large-50-one-to-many-mmt": "https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/sentencepiece.bpe.model",
}
}
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"facebook/mbart-large-50-one-to-many-mmt": 1024,
}
# fmt: off # fmt: off
FAIRSEQ_LANGUAGE_CODES = ["ar_AR", "cs_CZ", "de_DE", "en_XX", "es_XX", "et_EE", "fi_FI", "fr_XX", "gu_IN", "hi_IN", "it_IT", "ja_XX", "kk_KZ", "ko_KR", "lt_LT", "lv_LV", "my_MM", "ne_NP", "nl_XX", "ro_RO", "ru_RU", "si_LK", "tr_TR", "vi_VN", "zh_CN", "af_ZA", "az_AZ", "bn_IN", "fa_IR", "he_IL", "hr_HR", "id_ID", "ka_GE", "km_KH", "mk_MK", "ml_IN", "mn_MN", "mr_IN", "pl_PL", "ps_AF", "pt_XX", "sv_SE", "sw_KE", "ta_IN", "te_IN", "th_TH", "tl_XX", "uk_UA", "ur_PK", "xh_ZA", "gl_ES", "sl_SI"] FAIRSEQ_LANGUAGE_CODES = ["ar_AR", "cs_CZ", "de_DE", "en_XX", "es_XX", "et_EE", "fi_FI", "fr_XX", "gu_IN", "hi_IN", "it_IT", "ja_XX", "kk_KZ", "ko_KR", "lt_LT", "lv_LV", "my_MM", "ne_NP", "nl_XX", "ro_RO", "ru_RU", "si_LK", "tr_TR", "vi_VN", "zh_CN", "af_ZA", "az_AZ", "bn_IN", "fa_IR", "he_IL", "hr_HR", "id_ID", "ka_GE", "km_KH", "mk_MK", "ml_IN", "mn_MN", "mr_IN", "pl_PL", "ps_AF", "pt_XX", "sv_SE", "sw_KE", "ta_IN", "te_IN", "th_TH", "tl_XX", "uk_UA", "ur_PK", "xh_ZA", "gl_ES", "sl_SI"]
...@@ -83,8 +90,8 @@ class MBart50Tokenizer(PreTrainedTokenizer): ...@@ -83,8 +90,8 @@ class MBart50Tokenizer(PreTrainedTokenizer):
""" """
vocab_files_names = VOCAB_FILES_NAMES vocab_files_names = VOCAB_FILES_NAMES
max_model_input_sizes = {m: 1024 for m in _all_mbart50_models} max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
pretrained_vocab_files_map = {"vocab_file": {m: SPM_URL for m in _all_mbart50_models}} pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
model_input_names = ["input_ids", "attention_mask"] model_input_names = ["input_ids", "attention_mask"]
prefix_tokens: List[int] = [] prefix_tokens: List[int] = []
......
src/transformers/models/mbart/tokenization_mbart50_fast.py
View file @ 2295d783
...@@ -36,9 +36,18 @@ logger = logging.get_logger(__name__) ...@@ -36,9 +36,18 @@ logger = logging.get_logger(__name__)
VOCAB_FILES_NAMES = {"vocab_file": "sentencepiece.bpe.model", "tokenizer_file": "tokenizer.json"} VOCAB_FILES_NAMES = {"vocab_file": "sentencepiece.bpe.model", "tokenizer_file": "tokenizer.json"}
_all_mbart50_models = ["facebook/mbart-large-50-one-to-many-mmt"] PRETRAINED_VOCAB_FILES_MAP = {
SPM_URL = "https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/sentencepiece.bpe.model" "vocab_file": {
tokenizer_URL = "https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/tokenizer.json" "facebook/mbart-large-50-one-to-many-mmt": "https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/sentencepiece.bpe.model",
},
"tokenizer_file": {
"facebook/mbart-large-50-one-to-many-mmt": "https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/tokenizer.json",
},
}
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"facebook/mbart-large-50-one-to-many-mmt": 1024,
}
# fmt: off # fmt: off
FAIRSEQ_LANGUAGE_CODES = ["ar_AR", "cs_CZ", "de_DE", "en_XX", "es_XX", "et_EE", "fi_FI", "fr_XX", "gu_IN", "hi_IN", "it_IT", "ja_XX", "kk_KZ", "ko_KR", "lt_LT", "lv_LV", "my_MM", "ne_NP", "nl_XX", "ro_RO", "ru_RU", "si_LK", "tr_TR", "vi_VN", "zh_CN", "af_ZA", "az_AZ", "bn_IN", "fa_IR", "he_IL", "hr_HR", "id_ID", "ka_GE", "km_KH", "mk_MK", "ml_IN", "mn_MN", "mr_IN", "pl_PL", "ps_AF", "pt_XX", "sv_SE", "sw_KE", "ta_IN", "te_IN", "th_TH", "tl_XX", "uk_UA", "ur_PK", "xh_ZA", "gl_ES", "sl_SI"] FAIRSEQ_LANGUAGE_CODES = ["ar_AR", "cs_CZ", "de_DE", "en_XX", "es_XX", "et_EE", "fi_FI", "fr_XX", "gu_IN", "hi_IN", "it_IT", "ja_XX", "kk_KZ", "ko_KR", "lt_LT", "lv_LV", "my_MM", "ne_NP", "nl_XX", "ro_RO", "ru_RU", "si_LK", "tr_TR", "vi_VN", "zh_CN", "af_ZA", "az_AZ", "bn_IN", "fa_IR", "he_IL", "hr_HR", "id_ID", "ka_GE", "km_KH", "mk_MK", "ml_IN", "mn_MN", "mr_IN", "pl_PL", "ps_AF", "pt_XX", "sv_SE", "sw_KE", "ta_IN", "te_IN", "th_TH", "tl_XX", "uk_UA", "ur_PK", "xh_ZA", "gl_ES", "sl_SI"]
...@@ -91,8 +100,8 @@ class MBart50TokenizerFast(PreTrainedTokenizerFast): ...@@ -91,8 +100,8 @@ class MBart50TokenizerFast(PreTrainedTokenizerFast):
""" """
vocab_files_names = VOCAB_FILES_NAMES vocab_files_names = VOCAB_FILES_NAMES
max_model_input_sizes = {m: 1024 for m in _all_mbart50_models} max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
pretrained_vocab_files_map = {"vocab_file": {m: SPM_URL for m in _all_mbart50_models}} pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
model_input_names = ["input_ids", "attention_mask"] model_input_names = ["input_ids", "attention_mask"]
slow_tokenizer_class = MBart50Tokenizer slow_tokenizer_class = MBart50Tokenizer
......
src/transformers/models/mbart/tokenization_mbart_fast.py
View file @ 2295d783
...@@ -32,9 +32,24 @@ else: ...@@ -32,9 +32,24 @@ else:
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
_all_mbart_models = ["facebook/mbart-large-en-ro", "facebook/mbart-large-cc25"]
SPM_URL = "https://huggingface.co/facebook/mbart-large-en-ro/resolve/main/sentence.bpe.model" VOCAB_FILES_NAMES = {"vocab_file": "sentencepiece.bpe.model", "tokenizer_file": "tokenizer.json"}
tokenizer_URL = "https://huggingface.co/facebook/mbart-large-en-ro/resolve/main/tokenizer.json"
PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": {
"facebook/mbart-large-en-ro": "https://huggingface.co/facebook/mbart-large-en-ro/resolve/main/sentencepiece.bpe.model",
"facebook/mbart-large-cc25": "https://huggingface.co/facebook/mbart-large-cc25/resolve/main/sentencepiece.bpe.model",
},
"tokenizer_file": {
"facebook/mbart-large-en-ro": "https://huggingface.co/facebook/mbart-large-en-ro/resolve/main/tokenizer.json",
"facebook/mbart-large-cc25": "https://huggingface.co/facebook/mbart-large-cc25/resolve/main/tokenizer.json",
},
}
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"facebook/mbart-large-en-ro": 1024,
"facebook/mbart-large-cc25": 1024,
}
FAIRSEQ_LANGUAGE_CODES = [ FAIRSEQ_LANGUAGE_CODES = [
"ar_AR", "ar_AR",
...@@ -89,9 +104,9 @@ class MBartTokenizerFast(XLMRobertaTokenizerFast): ...@@ -89,9 +104,9 @@ class MBartTokenizerFast(XLMRobertaTokenizerFast):
>>> inputs["labels"] = labels["input_ids"] >>> inputs["labels"] = labels["input_ids"]
""" """
vocab_files_names = {"vocab_file": "sentencepiece.bpe.model"} vocab_files_names = VOCAB_FILES_NAMES
max_model_input_sizes = {m: 1024 for m in _all_mbart_models} max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
pretrained_vocab_files_map = {"vocab_file": {m: SPM_URL for m in _all_mbart_models}} pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
slow_tokenizer_class = MBartTokenizer slow_tokenizer_class = MBartTokenizer
prefix_tokens: List[int] = [] prefix_tokens: List[int] = []
......
src/transformers/models/reformer/tokenization_reformer.py
View file @ 2295d783
...@@ -27,28 +27,17 @@ from ...utils import logging ...@@ -27,28 +27,17 @@ from ...utils import logging
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
SPIECE_UNDERLINE = "▁"
SPIECE_UNDERLINE = "▁"
####################################################
# Mapping from the keyword arguments names of Tokenizer `__init__`
# to file names for serializing Tokenizer instances
####################################################
VOCAB_FILES_NAMES = {"vocab_file": "spiece.model"} VOCAB_FILES_NAMES = {"vocab_file": "spiece.model"}
####################################################
# Mapping from the keyword arguments names of Tokenizer `__init__`
# to pretrained vocabulary URL for all the model ids.
####################################################
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"google/reformer-crime-and-punishment": "https://huggingface.co/google/reformer-crime-and-punishment/resolve/main/spiece.model" "google/reformer-crime-and-punishment": "https://huggingface.co/google/reformer-crime-and-punishment/resolve/main/spiece.model"
} }
} }
####################################################
# Mapping from model ids to max length of inputs
####################################################
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = { PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"google/reformer-crime-and-punishment": 524288, "google/reformer-crime-and-punishment": 524288,
} }
......
src/transformers/models/reformer/tokenization_reformer_fast.py
View file @ 2295d783
...@@ -32,19 +32,11 @@ else: ...@@ -32,19 +32,11 @@ else:
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
SPIECE_UNDERLINE = "▁"
SPIECE_UNDERLINE = "▁"
####################################################
# Mapping from the keyword arguments names of Tokenizer `__init__`
# to file names for serializing Tokenizer instances
####################################################
VOCAB_FILES_NAMES = {"vocab_file": "spiece.model", "tokenizer_file": "tokenizer.json"} VOCAB_FILES_NAMES = {"vocab_file": "spiece.model", "tokenizer_file": "tokenizer.json"}
####################################################
# Mapping from the keyword arguments names of Tokenizer `__init__`
# to pretrained vocabulary URL for all the model ids.
####################################################
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"google/reformer-crime-and-punishment": "https://huggingface.co/google/reformer-crime-and-punishment/resolve/main/spiece.model" "google/reformer-crime-and-punishment": "https://huggingface.co/google/reformer-crime-and-punishment/resolve/main/spiece.model"
...@@ -54,9 +46,6 @@ PRETRAINED_VOCAB_FILES_MAP = { ...@@ -54,9 +46,6 @@ PRETRAINED_VOCAB_FILES_MAP = {
}, },
} }
####################################################
# Mapping from model ids to max length of inputs
####################################################
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = { PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"google/reformer-crime-and-punishment": 524288, "google/reformer-crime-and-punishment": 524288,
} }
......
src/transformers/models/retribert/tokenization_retribert.py
View file @ 2295d783
...@@ -24,7 +24,7 @@ VOCAB_FILES_NAMES = {"vocab_file": "vocab.txt"} ...@@ -24,7 +24,7 @@ VOCAB_FILES_NAMES = {"vocab_file": "vocab.txt"}
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"yjernite/retribert-base-uncased": "https://huggingface.co/bert-base-uncased/resolve/main/vocab.txt", "yjernite/retribert-base-uncased": "https://huggingface.co/yjernite/retribert-base-uncased/resolve/main/vocab.txt",
} }
} }
......
src/transformers/models/retribert/tokenization_retribert_fast.py
View file @ 2295d783
...@@ -25,10 +25,10 @@ VOCAB_FILES_NAMES = {"vocab_file": "vocab.txt", "tokenizer_file": "tokenizer.jso ...@@ -25,10 +25,10 @@ VOCAB_FILES_NAMES = {"vocab_file": "vocab.txt", "tokenizer_file": "tokenizer.jso
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"yjernite/retribert-base-uncased": "https://huggingface.co/bert-base-uncased/resolve/main/vocab.txt", "yjernite/retribert-base-uncased": "https://huggingface.co/yjernite/retribert-base-uncased/resolve/main/vocab.txt",
}, },
"tokenizer_file": { "tokenizer_file": {
"yjernite/retribert-base-uncased": "https://huggingface.co/bert-base-uncased/resolve/main/tokenizer.json", "yjernite/retribert-base-uncased": "https://huggingface.co/yjernite/retribert-base-uncased/resolve/main/tokenizer.json",
}, },
} }
......
src/transformers/models/roberta/tokenization_roberta.py
View file @ 2295d783
...@@ -34,16 +34,16 @@ PRETRAINED_VOCAB_FILES_MAP = { ...@@ -34,16 +34,16 @@ PRETRAINED_VOCAB_FILES_MAP = {
"roberta-large": "https://huggingface.co/roberta-large/resolve/main/vocab.json", "roberta-large": "https://huggingface.co/roberta-large/resolve/main/vocab.json",
"roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/vocab.json", "roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/vocab.json",
"distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/vocab.json", "distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/vocab.json",
"roberta-base-openai-detector": "https://huggingface.co/roberta-base/resolve/main/vocab.json", "roberta-base-openai-detector": "https://huggingface.co/roberta-base-openai-detector/resolve/main/vocab.json",
"roberta-large-openai-detector": "https://huggingface.co/roberta-large/resolve/main/vocab.json", "roberta-large-openai-detector": "https://huggingface.co/roberta-large-openai-detector/resolve/main/vocab.json",
}, },
"merges_file": { "merges_file": {
"roberta-base": "https://huggingface.co/roberta-base/resolve/main/merges.txt", "roberta-base": "https://huggingface.co/roberta-base/resolve/main/merges.txt",
"roberta-large": "https://huggingface.co/roberta-large/resolve/main/merges.txt", "roberta-large": "https://huggingface.co/roberta-large/resolve/main/merges.txt",
"roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/merges.txt", "roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/merges.txt",
"distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/merges.txt", "distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/merges.txt",
"roberta-base-openai-detector": "https://huggingface.co/roberta-base/resolve/main/merges.txt", "roberta-base-openai-detector": "https://huggingface.co/roberta-base-openai-detector/resolve/main/merges.txt",
"roberta-large-openai-detector": "https://huggingface.co/roberta-large/resolve/main/merges.txt", "roberta-large-openai-detector": "https://huggingface.co/roberta-large-openai-detector/resolve/main/merges.txt",
}, },
} }
......
src/transformers/models/roberta/tokenization_roberta_fast.py
View file @ 2295d783
...@@ -32,24 +32,24 @@ PRETRAINED_VOCAB_FILES_MAP = { ...@@ -32,24 +32,24 @@ PRETRAINED_VOCAB_FILES_MAP = {
"roberta-large": "https://huggingface.co/roberta-large/resolve/main/vocab.json", "roberta-large": "https://huggingface.co/roberta-large/resolve/main/vocab.json",
"roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/vocab.json", "roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/vocab.json",
"distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/vocab.json", "distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/vocab.json",
"roberta-base-openai-detector": "https://huggingface.co/roberta-base/resolve/main/vocab.json", "roberta-base-openai-detector": "https://huggingface.co/roberta-base-openai-detector/resolve/main/vocab.json",
"roberta-large-openai-detector": "https://huggingface.co/roberta-large/resolve/main/vocab.json", "roberta-large-openai-detector": "https://huggingface.co/roberta-large-openai-detector/resolve/main/vocab.json",
}, },
"merges_file": { "merges_file": {
"roberta-base": "https://huggingface.co/roberta-base/resolve/main/merges.txt", "roberta-base": "https://huggingface.co/roberta-base/resolve/main/merges.txt",
"roberta-large": "https://huggingface.co/roberta-large/resolve/main/merges.txt", "roberta-large": "https://huggingface.co/roberta-large/resolve/main/merges.txt",
"roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/merges.txt", "roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/merges.txt",
"distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/merges.txt", "distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/merges.txt",
"roberta-base-openai-detector": "https://huggingface.co/roberta-base/resolve/main/merges.txt", "roberta-base-openai-detector": "https://huggingface.co/roberta-base-openai-detector/resolve/main/merges.txt",
"roberta-large-openai-detector": "https://huggingface.co/roberta-large/resolve/main/merges.txt", "roberta-large-openai-detector": "https://huggingface.co/roberta-large-openai-detector/resolve/main/merges.txt",
}, },
"tokenizer_file": { "tokenizer_file": {
"roberta-base": "https://huggingface.co/roberta-base/resolve/main/tokenizer.json", "roberta-base": "https://huggingface.co/roberta-base/resolve/main/tokenizer.json",
"roberta-large": "https://huggingface.co/roberta-large/resolve/main/tokenizer.json", "roberta-large": "https://huggingface.co/roberta-large/resolve/main/tokenizer.json",
"roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/tokenizer.json", "roberta-large-mnli": "https://huggingface.co/roberta-large-mnli/resolve/main/tokenizer.json",
"distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/tokenizer.json", "distilroberta-base": "https://huggingface.co/distilroberta-base/resolve/main/tokenizer.json",
"roberta-base-openai-detector": "https://huggingface.co/roberta-base/resolve/main/tokenizer.json", "roberta-base-openai-detector": "https://huggingface.co/roberta-base-openai-detector/resolve/main/tokenizer.json",
"roberta-large-openai-detector": "https://huggingface.co/roberta-large/resolve/main/tokenizer.json", "roberta-large-openai-detector": "https://huggingface.co/roberta-large-openai-detector/resolve/main/tokenizer.json",
}, },
} }
......
src/transformers/models/t5/tokenization_t5.py
View file @ 2295d783
...@@ -29,16 +29,8 @@ from ...utils import logging ...@@ -29,16 +29,8 @@ from ...utils import logging
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
####################################################
# Mapping from the keyword arguments names of Tokenizer `__init__`
# to file names for serializing Tokenizer instances
####################################################
VOCAB_FILES_NAMES = {"vocab_file": "spiece.model"} VOCAB_FILES_NAMES = {"vocab_file": "spiece.model"}
####################################################
# Mapping from the keyword arguments names of Tokenizer `__init__`
# to pretrained vocabulary URL for all the model ids.
####################################################
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"t5-small": "https://huggingface.co/t5-small/resolve/main/spiece.model", "t5-small": "https://huggingface.co/t5-small/resolve/main/spiece.model",
...@@ -49,9 +41,6 @@ PRETRAINED_VOCAB_FILES_MAP = { ...@@ -49,9 +41,6 @@ PRETRAINED_VOCAB_FILES_MAP = {
} }
} }
####################################################
# Mapping from model ids to max length of inputs
####################################################
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = { PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"t5-small": 512, "t5-small": 512,
"t5-base": 512, "t5-base": 512,
......
src/transformers/models/t5/tokenization_t5_fast.py
View file @ 2295d783
...@@ -32,16 +32,8 @@ else: ...@@ -32,16 +32,8 @@ else:
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
####################################################
# Mapping from the keyword arguments names of Tokenizer `__init__`
# to file names for serializing Tokenizer instances
####################################################
VOCAB_FILES_NAMES = {"vocab_file": "spiece.model", "tokenizer_file": "tokenizer.json"} VOCAB_FILES_NAMES = {"vocab_file": "spiece.model", "tokenizer_file": "tokenizer.json"}
####################################################
# Mapping from the keyword arguments names of Tokenizer `__init__`
# to pretrained vocabulary URL for all the model ids.
####################################################
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"t5-small": "https://huggingface.co/t5-small/resolve/main/spiece.model", "t5-small": "https://huggingface.co/t5-small/resolve/main/spiece.model",
...@@ -59,9 +51,6 @@ PRETRAINED_VOCAB_FILES_MAP = { ...@@ -59,9 +51,6 @@ PRETRAINED_VOCAB_FILES_MAP = {
}, },
} }
####################################################
# Mapping from model ids to max length of inputs
####################################################
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = { PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"t5-small": 512, "t5-small": 512,
"t5-base": 512, "t5-base": 512,
......
src/transformers/models/wav2vec2/tokenization_wav2vec2.py
View file @ 2295d783
...@@ -37,6 +37,17 @@ VOCAB_FILES_NAMES = { ...@@ -37,6 +37,17 @@ VOCAB_FILES_NAMES = {
"tokenizer_config_file": "tokenizer_config.json", "tokenizer_config_file": "tokenizer_config.json",
} }
PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": {
"facebook/wav2vec2-base-960h": "https://huggingface.co/facebook/wav2vec2-base-960h/resolve/main/vocab.json",
},
"tokenizer_config_file": {
"facebook/wav2vec2-base-960h": "https://huggingface.co/facebook/wav2vec2-base-960h/resolve/main/tokenizer_config.json",
},
}
# Wav2Vec2 has no max input length
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {"facebook/wav2vec2-base-960h": sys.maxsize}
WAV2VEC2_KWARGS_DOCSTRING = r""" WAV2VEC2_KWARGS_DOCSTRING = r"""
padding (:obj:`bool`, :obj:`str` or :class:`~transformers.file_utils.PaddingStrategy`, `optional`, defaults to :obj:`False`): padding (:obj:`bool`, :obj:`str` or :class:`~transformers.file_utils.PaddingStrategy`, `optional`, defaults to :obj:`False`):
...@@ -98,16 +109,8 @@ class Wav2Vec2CTCTokenizer(PreTrainedTokenizer): ...@@ -98,16 +109,8 @@ class Wav2Vec2CTCTokenizer(PreTrainedTokenizer):
""" """
vocab_files_names = VOCAB_FILES_NAMES vocab_files_names = VOCAB_FILES_NAMES
pretrained_vocab_files_map = { pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
"vocab_file": { max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
"facebook/wav2vec2-base-960h": "https://huggingface.co/facebook/wav2vec2-base-960h/resolve/main/vocab.json"
},
"tokenizer_config_file": {
"facebook/wav2vec2-base-960h": "https://huggingface.co/facebook/wav2vec2-base-960h/resolve/main/tokenizer.json",
},
}
# Wav2Vec2 has no max input length
max_model_input_sizes = {"facebook/wav2vec2-base-960h": sys.maxsize}
model_input_names = ["input_ids", "attention_mask"] model_input_names = ["input_ids", "attention_mask"]
def __init__( def __init__(
......
templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/tokenization_fast_{{cookiecutter.lowercase_modelname}}.py
View file @ 2295d783
...@@ -65,6 +65,8 @@ from .tokenization_{{cookiecutter.lowercase_modelname}} import {{cookiecutter.ca ...@@ -65,6 +65,8 @@ from .tokenization_{{cookiecutter.lowercase_modelname}} import {{cookiecutter.ca
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
VOCAB_FILES_NAMES = {"vocab_file": "vocab.json", "merges_file": "merges.txt", "tokenizer_file": "tokenizer.json"}
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"{{cookiecutter.checkpoint_identifier}}": "https://huggingface.co/{{cookiecutter.checkpoint_identifier}}/resolve/main/vocab.json", "{{cookiecutter.checkpoint_identifier}}": "https://huggingface.co/{{cookiecutter.checkpoint_identifier}}/resolve/main/vocab.json",
...@@ -93,6 +95,7 @@ class {{cookiecutter.camelcase_modelname}}TokenizerFast(BartTokenizerFast): ...@@ -93,6 +95,7 @@ class {{cookiecutter.camelcase_modelname}}TokenizerFast(BartTokenizerFast):
parameters. parameters.
""" """
vocab_files_names = VOCAB_FILES_NAMES
pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
slow_tokenizer_class = {{cookiecutter.camelcase_modelname}}Tokenizer slow_tokenizer_class = {{cookiecutter.camelcase_modelname}}Tokenizer
...@@ -109,9 +112,16 @@ from .tokenization_{{cookiecutter.lowercase_modelname}} import {{cookiecutter.ca ...@@ -109,9 +112,16 @@ from .tokenization_{{cookiecutter.lowercase_modelname}} import {{cookiecutter.ca
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
VOCAB_FILES_NAMES = {} VOCAB_FILES_NAMES = {"vocab_file": "vocab.txt", "tokenizer_file": "tokenizer.json"}
PRETRAINED_VOCAB_FILES_MAP = {} PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": {
"{{cookiecutter.checkpoint_identifier}}": "https://huggingface.co/{{cookiecutter.checkpoint_identifier}}/resolve/main/vocab.txt",
},
"tokenizer_file": {
"{{cookiecutter.checkpoint_identifier}}": "https://huggingface.co/{{cookiecutter.checkpoint_identifier}}/resolve/main/tokenizer.json",
},
}
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = { PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"{{cookiecutter.checkpoint_identifier}}": 1024, "{{cookiecutter.checkpoint_identifier}}": 1024,
......
templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/tokenization_{{cookiecutter.lowercase_modelname}}.py
View file @ 2295d783
...@@ -62,6 +62,8 @@ from ..bart.tokenization_bart import BartTokenizer ...@@ -62,6 +62,8 @@ from ..bart.tokenization_bart import BartTokenizer
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
VOCAB_FILES_NAMES = {"vocab_file": "vocab.json", "merges_file": "merges.txt", "tokenizer_file": "tokenizer.json"}
PRETRAINED_VOCAB_FILES_MAP = { PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": { "vocab_file": {
"{{cookiecutter.checkpoint_identifier}}": "https://huggingface.co/{{cookiecutter.checkpoint_identifier}}/resolve/main/vocab.json", "{{cookiecutter.checkpoint_identifier}}": "https://huggingface.co/{{cookiecutter.checkpoint_identifier}}/resolve/main/vocab.json",
...@@ -90,6 +92,7 @@ class {{cookiecutter.camelcase_modelname}}Tokenizer(BartTokenizer): ...@@ -90,6 +92,7 @@ class {{cookiecutter.camelcase_modelname}}Tokenizer(BartTokenizer):
parameters. parameters.
""" """
vocab_files_names = VOCAB_FILES_NAMES
pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
...@@ -105,9 +108,13 @@ from ...utils import logging ...@@ -105,9 +108,13 @@ from ...utils import logging
logger = logging.get_logger(__name__) logger = logging.get_logger(__name__)
VOCAB_FILES_NAMES = {} VOCAB_FILES_NAMES = {"vocab_file": "vocab.txt"}
PRETRAINED_VOCAB_FILES_MAP = {} PRETRAINED_VOCAB_FILES_MAP = {
"vocab_file": {
"{{cookiecutter.checkpoint_identifier}}": "https://huggingface.co/{{cookiecutter.checkpoint_identifier}}/resolve/main/vocab.txt",
},
}
PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = { PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
"{{cookiecutter.checkpoint_identifier}}": 1024, "{{cookiecutter.checkpoint_identifier}}": 1024,
......
tests/test_tokenization_marian.py
View file @ 2295d783
...@@ -26,7 +26,7 @@ from transformers.testing_utils import require_sentencepiece ...@@ -26,7 +26,7 @@ from transformers.testing_utils import require_sentencepiece
if is_sentencepiece_available(): if is_sentencepiece_available():
from transformers.models.marian.tokenization_marian import save_json, vocab_files_names from transformers.models.marian.tokenization_marian import VOCAB_FILES_NAMES, save_json
from .test_tokenization_common import TokenizerTesterMixin from .test_tokenization_common import TokenizerTesterMixin
...@@ -50,11 +50,11 @@ class MarianTokenizationTest(TokenizerTesterMixin, unittest.TestCase): ...@@ -50,11 +50,11 @@ class MarianTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
vocab = ["</s>", "<unk>", "▁This", "▁is", "▁a", "▁t", "est", "\u0120", "<pad>"] vocab = ["</s>", "<unk>", "▁This", "▁is", "▁a", "▁t", "est", "\u0120", "<pad>"]
vocab_tokens = dict(zip(vocab, range(len(vocab)))) vocab_tokens = dict(zip(vocab, range(len(vocab))))
save_dir = Path(self.tmpdirname) save_dir = Path(self.tmpdirname)
save_json(vocab_tokens, save_dir / vocab_files_names["vocab"]) save_json(vocab_tokens, save_dir / VOCAB_FILES_NAMES["vocab"])
save_json(mock_tokenizer_config, save_dir / vocab_files_names["tokenizer_config_file"]) save_json(mock_tokenizer_config, save_dir / VOCAB_FILES_NAMES["tokenizer_config_file"])
if not (save_dir / vocab_files_names["source_spm"]).exists(): if not (save_dir / VOCAB_FILES_NAMES["source_spm"]).exists():
copyfile(SAMPLE_SP, save_dir / vocab_files_names["source_spm"]) copyfile(SAMPLE_SP, save_dir / VOCAB_FILES_NAMES["source_spm"])
copyfile(SAMPLE_SP, save_dir / vocab_files_names["target_spm"]) copyfile(SAMPLE_SP, save_dir / VOCAB_FILES_NAMES["target_spm"])
tokenizer = MarianTokenizer.from_pretrained(self.tmpdirname) tokenizer = MarianTokenizer.from_pretrained(self.tmpdirname)
tokenizer.save_pretrained(self.tmpdirname) tokenizer.save_pretrained(self.tmpdirname)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment