Skip to content

Models with incorrect tokenizer_class in tokenization_config.json tha… #44179

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ArthurZucker merged 1 commit into from
Feb 23, 2026
+27 −0
Merged

Models with incorrect tokenizer_class in tokenization_config.json tha… #44179

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
27 changes: 27 additions & 0 deletions src/transformers/models/auto/tokenization_auto.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -185,6 +185,12 @@
("megatron-bert", "BertTokenizer" if is_tokenizers_available() else None),
("metaclip_2", "XLMRobertaTokenizer" if is_tokenizers_available() else None),
("mgp-str", "MgpstrTokenizer"),
(
"ministral",
"MistralCommonBackend"
if is_mistral_common_available()
else ("TokenizersBackend" if is_tokenizers_available() else None),
),
(
"ministral3",
"MistralCommonBackend"
Expand Down Expand Up @@ -331,6 +337,27 @@
]
)

# Models with incorrect tokenizer_class in their Hub tokenizer_config.json files.
# These models will be forced to use TokenizersBackend.
MODELS_WITH_INCORRECT_HUB_TOKENIZER_CLASS: set[str] = {
"arctic",
"deepseek_vl",
"deepseek_vl_hybrid",
"hyperclovax_vlm",
"janus",
"jamba",
"llava",
"llava_next",
Comment on lines +349 to +350
Copy link
Collaborator

@ArthurZucker ArthurZucker Feb 20, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

are these from the vllm ci as well?

"opencua",
"phi3",
"step3p5",
"vipllava",
}

for model_type in MODELS_WITH_INCORRECT_HUB_TOKENIZER_CLASS:
if model_type not in TOKENIZER_MAPPING_NAMES:
TOKENIZER_MAPPING_NAMES[model_type] = "TokenizersBackend" if is_tokenizers_available() else None

TOKENIZER_MAPPING = _LazyAutoMapping(CONFIG_MAPPING_NAMES, TOKENIZER_MAPPING_NAMES)

CONFIG_TO_TYPE = {v: k for k, v in CONFIG_MAPPING_NAMES.items()}
Expand Down