Skip to content

[ vllm x v5] nit #44971

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 6 commits into from
Mar 24, 2026
Merged

[ vllm x v5] nit #44971

Show file tree
Hide file tree
Changes from 5 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
afb355c
remove class
ArthurZucker Mar 24, 2026
f7b99ea
Merge branch 'main' into small-nit
ArthurZucker Mar 24, 2026
7d90d1e
remove more
ArthurZucker Mar 24, 2026
d84813e
Merge branch 'small-nit' of github.com:huggingface/transformers into …
ArthurZucker Mar 24, 2026
dbc13eb
fix repo
ArthurZucker Mar 24, 2026
ece5f2a
up up
ArthurZucker Mar 24, 2026
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
4 changes: 0 additions & 4 deletions src/transformers/configuration_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,6 @@
from .generation.configuration_utils import GenerationConfig
from .modeling_gguf_pytorch_utils import load_gguf_checkpoint
from .modeling_rope_utils import RotaryEmbeddingConfigMixin
from .tokenization_utils_base import PreTrainedTokenizerBase
from .utils import (
CONFIG_NAME,
PushToHubMixin,
Expand Down Expand Up @@ -234,9 +233,6 @@ class PreTrainedConfig(PushToHubMixin, RotaryEmbeddingConfigMixin):
label2id: dict[str, int] | dict[str, str] | None = None
problem_type: Literal["regression", "single_label_classification", "multi_label_classification"] | None = None

Comment on lines 231 to 235

Copilot AI Mar 24, 2026

Copy link

Choose a reason for hiding this comment

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

Removing tokenizer_class from PreTrainedConfig will break existing tests/utilities that treat it as a common config kwarg/field (e.g. tests/utils/test_configuration_utils.py::test_config_common_kwargs_is_complete expects tokenizer_class to be present in PreTrainedConfig().__dict__). Please update the corresponding test expectations (and any shared config_common_kwargs/common-config logic) to reflect the new base config surface, or keep a backward-compatible tokenizer_class field if it’s still considered part of the common config contract.

Copilot uses AI. Check for mistakes.
# Tokenizer kwargs
tokenizer_class: str | PreTrainedTokenizerBase | None = None

def __post_init__(self, **kwargs):
# BC for the `torch_dtype` argument instead of the simpler `dtype`
# Do not warn, as it would otherwise always be triggered since most configs on the hub have `torch_dtype`
Expand Down
3 changes: 0 additions & 3 deletions src/transformers/models/mt5/configuration_mt5.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,8 +29,6 @@ class MT5Config(PreTrainedConfig):
The maximum distance of the longer sequences for the bucket separation.
feed_forward_proj (`str`, *optional*, defaults to `"gated-gelu"`):
Type of feed forward layer to be used. Should be one of `"relu"` or `"gated-gelu"`.
tokenizer_class (`str`, *optional*, defaults to `"T5Tokenizer"`):
The tokenizer's class name.
"""

model_type = "mt5"
Expand All @@ -57,7 +55,6 @@ class MT5Config(PreTrainedConfig):
feed_forward_proj: str = "gated-gelu"
is_encoder_decoder: bool = True
use_cache: bool = True
tokenizer_class: str = "T5Tokenizer"
tie_word_embeddings: bool = True
bos_token_id: int | None = None
pad_token_id: int | None = 0
Expand Down
3 changes: 0 additions & 3 deletions src/transformers/models/umt5/configuration_umt5.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,8 +29,6 @@ class UMT5Config(PreTrainedConfig):
The maximum distance of the longer sequences for the bucket separation.
feed_forward_proj (`str`, *optional*, defaults to `"gated-gelu"`):
Type of feed forward layer to be used. Should be one of `"relu"` or `"gated-gelu"`.
tokenizer_class (`str`, *optional*, defaults to `"T5Tokenizer"`):
The tokenizer's class name
"""

model_type = "umt5"
Expand All @@ -57,7 +55,6 @@ class UMT5Config(PreTrainedConfig):
feed_forward_proj: str = "gated-gelu"
is_encoder_decoder: bool = True
use_cache: bool = True
tokenizer_class: str = "T5Tokenizer"
pad_token_id: int | None = 0
eos_token_id: int | list[int] | None = 1
decoder_start_token_id: int | None = 0
Expand Down
Loading