Add pydocstyle to the ruff rules (#52136)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Harry Mellor <hej.mellor@gmail.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Harry Mellor
2026-09-16 16:44:29 +00:00
committed by GitHub
co-authored by Claude Opus 5
parent 804e5377aa
commit 8c1557a79c
2002 changed files with 15024 additions and 14269 deletions
+3 -2
View File
@@ -6,14 +6,14 @@ from transformers import PretrainedConfig
class AXK1Config(PretrainedConfig):
r"""
This is the configuration class to store the configuration of a [`AXK1Model`].
r"""This is the configuration class to store the configuration of a [`AXK1Model`].
It is used to instantiate an A.X model according to the specified arguments,
defining the model architecture. Instantiating a configuration with the defaults
will yield a similar configuration to that of the A.X K1.
Configuration objects inherit from [`PretrainedConfig`] and can be used to control
the model outputs. Read the documentation from [`PretrainedConfig`] for more
information.
Args:
vocab_size (`int`, *optional*, defaults to 163840):
Vocabulary size of the A.X K1 model. Defines the number of different
@@ -112,6 +112,7 @@ class AXK1Config(PretrainedConfig):
layers during self-attention.
attention_dropout (`float`, *optional*, defaults to 0.0):
The dropout ratio for the attention probabilities.
"""
model_type = "axk1"
+1 -2
View File
@@ -1,7 +1,6 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""
Model configs may be defined in this directory for the following reasons:
"""Model configs may be defined in this directory for the following reasons:
- There is no configuration file defined by HF Hub or Transformers library.
- There is a need to override the existing config to support vLLM.
+1 -2
View File
@@ -1,7 +1,6 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""
ColPali configuration that extends PaliGemmaConfig with embedding projection
"""ColPali configuration that extends PaliGemmaConfig with embedding projection
fields. This allows ColPali models to be loaded without trust_remote_code
by mapping their custom model_type (colpali) to a standard config class
that vLLM understands.
+1 -2
View File
@@ -1,7 +1,6 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""
ColQwen3 configuration that extends Qwen3VLConfig with embedding projection
"""ColQwen3 configuration that extends Qwen3VLConfig with embedding projection
fields. This allows ColQwen3 models to be loaded without trust_remote_code
by mapping their custom model_type (colqwen3, ops_colqwen3, etc.) to a
standard config class that vLLM understands.
@@ -1,7 +1,8 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""Config definitions for ExtractHiddenStatesModel, to be used with
the extract_hidden_states spec decoding method."""
the extract_hidden_states spec decoding method.
"""
import os
+1 -1
View File
@@ -18,7 +18,7 @@
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Falcon configuration"""
"""Falcon configuration."""
from transformers.configuration_utils import PretrainedConfig
+2 -2
View File
@@ -6,8 +6,7 @@ from transformers.configuration_utils import PretrainedConfig
class HYV3Config(PretrainedConfig):
r"""
This is the configuration class to store the configuration of a [`HYV3Model`].
r"""This is the configuration class to store the configuration of a [`HYV3Model`].
It is used to instantiate a HYV3 model (HY V3 MoE language model) according to
the specified arguments.
@@ -93,6 +92,7 @@ class HYV3Config(PretrainedConfig):
>>> config = HYV3Config()
>>> model = HYV3Model(config)
```
"""
model_type = "hy_v3"
+1
View File
@@ -52,6 +52,7 @@ class HYV4Config(PreTrainedConfig):
below zero disable the clamp.
num_nextn_predict_layers: Number of MTP layers in the checkpoint.
mtp_loss_factor: Training-time MTP loss weight (unused at inference).
"""
model_type = "hy_v4"
@@ -21,8 +21,7 @@ from transformers.configuration_utils import PretrainedConfig
class HyperCLOVAXConfig(PretrainedConfig):
r"""
This is the configuration class to store the configuration of a
r"""This is the configuration class to store the configuration of a
[`HyperCLOVAXModel`]. It is used to instantiate a HyperCLOVAX model
according to the specified arguments, defining the model architecture.
Configuration objects inherit from [`PretrainedConfig`] and can be used
@@ -166,6 +165,7 @@ class HyperCLOVAXConfig(PretrainedConfig):
`get_rope`. When provided, takes precedence over `rope_theta`
and `rope_scaling`. If `None`, it is derived from `rope_theta`
and `rope_scaling` automatically.
"""
model_type = "hyperclovax"
+2 -2
View File
@@ -1,7 +1,6 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""
Kimi-K2.5 Model Configuration.
"""Kimi-K2.5 Model Configuration.
This configuration supports video-chunk as an internal modality type.
A video-chunk is the smallest independently processable unit of video.
@@ -73,6 +72,7 @@ class KimiK25Config(PretrainedConfig):
ignore_index: The ignore index for the loss function.
media_placeholder_token_id: The token ID for media placeholders.
pad_token_id: The token ID for padding.
"""
model_type = "kimi_k25"
+4 -3
View File
@@ -6,8 +6,7 @@ from transformers.configuration_utils import PretrainedConfig
class Lfm2MoeConfig(PretrainedConfig):
r"""
This is the configuration class to store the configuration of a [`Lfm2MoeModel`]. It is used to instantiate a LFM2 Moe
r"""This is the configuration class to store the configuration of a [`Lfm2MoeModel`]. It is used to instantiate a LFM2 Moe
model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
defaults will yield a similar configuration to that of the LFM2-8B-A1B model.
e.g. [LiquidAI/LFM2-8B-A1B](https://huggingface.co/LiquidAI/LFM2-8B-A1B)
@@ -85,7 +84,9 @@ class Lfm2MoeConfig(PretrainedConfig):
>>> # Accessing the model configuration
>>> configuration = model.config
```""" # noqa: E501
```
""" # noqa: E501
model_type = "lfm2_moe"
keys_to_ignore_at_inference = ["past_key_values"]
@@ -24,8 +24,7 @@ class MLPSpeculatorConfig(PretrainedConfig):
scale_input: bool = False,
**kwargs,
):
"""
Initialize an MLPSpeculatorConfig
"""Initialize an MLPSpeculatorConfig.
Args:
vocab_size: int
@@ -52,6 +51,7 @@ class MLPSpeculatorConfig(PretrainedConfig):
scale_input: bool
if True, will scale the initial hidden states from
the base model.
"""
if top_k_tokens_per_head is None:
top_k_tokens_per_head = [5, 4, 3]
@@ -148,5 +148,6 @@ class Moondream3Config(PretrainedConfig):
Args:
decoder: Ignored. Only used for encoder-decoder models.
"""
return self.text_config
+6 -7
View File
@@ -15,7 +15,7 @@
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Nemotron model configuration"""
"""Nemotron model configuration."""
from transformers import PretrainedConfig
from transformers.utils import logging
@@ -24,8 +24,7 @@ logger = logging.get_logger(__name__)
class NemotronConfig(PretrainedConfig):
r"""
This is the configuration class to store the configuration of a
r"""This is the configuration class to store the configuration of a
[`NemotronModel`]. It is used to instantiate a Nemotron model
according to the specified arguments, defining the model architecture.
Instantiating a configuration with the defaults will yield a similar
@@ -114,7 +113,9 @@ class NemotronConfig(PretrainedConfig):
>>> model = NemotronModel(configuration)
>>> # Accessing the model configuration
>>> configuration = model.config
```"""
```
"""
model_type = "nemotron"
keys_to_ignore_at_inference = ["past_key_values"]
@@ -193,9 +194,7 @@ class NemotronConfig(PretrainedConfig):
)
def _rope_parameters_validation(self):
"""
Validate the `rope_parameters` configuration.
"""
"""Validate the `rope_parameters` configuration."""
if self.rope_parameters is None:
return
@@ -15,7 +15,7 @@
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""NemotronH model configuration"""
"""NemotronH model configuration."""
import regex as re
from transformers.configuration_utils import PretrainedConfig
@@ -25,12 +25,12 @@ logger = logging.get_logger(__name__)
class NemotronHConfig(PretrainedConfig):
r"""
This is the configuration class to store the configuration of a
r"""This is the configuration class to store the configuration of a
[`NemotronHModel`]. It is used to instantiate a NemotronH model according
to the specified arguments, defining the model architecture. Instantiating
a configuration with the defaults will yield a similar configuration to
that of the NemotronH-v0.1 model.
Args:
vocab_size (`int`, *optional*, defaults to 131072):
Vocabulary size of the NemotronH model. Defines the number of
@@ -139,6 +139,7 @@ class NemotronHConfig(PretrainedConfig):
Size of chunks for Mamba processing.
rescale_prenorm_residual (`bool`, *optional*, defaults to `True`):
Whether to rescale the pre-normalization residual connections.
"""
model_type = "nemotron_h"
@@ -6,8 +6,7 @@ from transformers.configuration_utils import PretrainedConfig
class OlmoHybridConfig(PretrainedConfig):
r"""
Configuration class for [`OlmoHybridModel`]. It is used to
r"""Configuration class for [`OlmoHybridModel`]. It is used to
instantiate an OLMo Hybrid model according to the specified
arguments, defining the model architecture. Instantiating a
configuration with the defaults will yield a similar
@@ -19,7 +18,7 @@ class OlmoHybridConfig(PretrainedConfig):
can be used to control the model outputs. Read the
documentation from [`PreTrainedConfig`] for more information.
Args:
Args:
vocab_size (`int`, *optional*, defaults to 100352):
Vocabulary size of the OlmoHybrid model. Defines
the number of different tokens that can be
@@ -159,6 +158,7 @@ class OlmoHybridConfig(PretrainedConfig):
>>> model = OlmoHybridModel(configuration)
>>> configuration = model.config
```
"""
model_type = "olmo_hybrid"
+2
View File
@@ -14,6 +14,7 @@ class AIMv2Config(PretrainedConfig):
"""This is the configuration class to store the configuration of an [`AIMv2Model`].
Instantiating a configuration with the defaults will yield a similar configuration
to that of the [apple/aimv2-large-patch14-224](https://huggingface.co/apple/aimv2-large-patch14-224).
Args:
hidden_size: Dimension of the hidden representations.
intermediate_size: Dimension of the SwiGLU representations.
@@ -29,6 +30,7 @@ class AIMv2Config(PretrainedConfig):
qkv_bias: Whether to add a bias to the queries, keys and values.
use_bias: Whether to add a bias in the feed-forward and projection layers.
kwargs: Keyword arguments for the [`PretrainedConfig`].
"""
model_type: str = "aimv2"
+1 -1
View File
@@ -14,7 +14,7 @@
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Qwen3.5 model configuration"""
"""Qwen3.5 model configuration."""
from transformers.configuration_utils import PretrainedConfig
@@ -14,7 +14,7 @@
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Qwen3.5-MoE model configuration"""
"""Qwen3.5-MoE model configuration."""
from transformers.configuration_utils import PretrainedConfig
@@ -14,7 +14,7 @@
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Qwen3-Next model configuration"""
"""Qwen3-Next model configuration."""
from transformers.configuration_utils import PretrainedConfig
from transformers.utils import logging
@@ -23,8 +23,7 @@ logger = logging.get_logger(__name__)
class Qwen3NextConfig(PretrainedConfig):
r"""
This is the configuration class to store the configuration of a [`Qwen3NextModel`]. It is used to instantiate a
r"""This is the configuration class to store the configuration of a [`Qwen3NextModel`]. It is used to instantiate a
Qwen3-Next model according to the specified arguments, defining the model architecture.
Instantiating a configuration with the defaults will yield a similar configuration to that of
Qwen3-Next-80B-A3B-Instruct [Qwen/Qwen3-Next-80B-A3B-Instruct](https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Instruct).
@@ -157,6 +156,7 @@ class Qwen3NextConfig(PretrainedConfig):
>>> # Accessing the model configuration
>>> configuration = model.config
```
""" # noqa: E501
model_type = "qwen3_next"
+3 -3
View File
@@ -1,6 +1,6 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""Radio vision model configuration"""
"""Radio vision model configuration."""
from typing import Any
@@ -21,8 +21,7 @@ OPENAI_CLIP_STD = (0.26862954, 0.26130258, 0.27577711)
class RadioConfig(PretrainedConfig):
r"""
This is the configuration class to store the configuration of a Radio
r"""This is the configuration class to store the configuration of a Radio
vision model. It is used to instantiate a Radio model according to the
specified arguments, defining the model architecture.
@@ -55,6 +54,7 @@ class RadioConfig(PretrainedConfig):
dedicated video patch embedder (3*T*P*P -> hidden) separate from the
image embedder (3*P*P -> hidden). When False, a single embedder with
input size 3*T*P*P is used for both (images are duplicated T times).
"""
model_type = "radio"
@@ -14,8 +14,7 @@ def register_speculator(name):
@register_speculator("eagle3")
def update_eagle3(config_dict: dict, pre_trained_config: dict) -> None:
"""
Apply Eagle-3 specific configuration transformations to the `dict` used to
"""Apply Eagle-3 specific configuration transformations to the `dict` used to
construct the Transformers PreTrainedConfig.
Eagle-3 specific fields:
@@ -28,7 +27,6 @@ def update_eagle3(config_dict: dict, pre_trained_config: dict) -> None:
provide intermediate hidden states that help the drafter make better
predictions. This is the standard field used in Eagle3 checkpoints.
"""
pre_trained_config["draft_vocab_size"] = config_dict.get("draft_vocab_size")
if config_dict.get("target_hidden_size") is not None:
pre_trained_config["target_hidden_size"] = config_dict["target_hidden_size"]
@@ -54,8 +52,7 @@ def update_eagle3(config_dict: dict, pre_trained_config: dict) -> None:
@register_speculator("peagle")
def update_peagle(config_dict: dict, pre_trained_config: dict) -> None:
"""
Apply PEagle (Parallel Eagle) specific configuration transformations to
"""Apply PEagle (Parallel Eagle) specific configuration transformations to
the `dict` used to construct the Transformers PreTrainedConfig.
PEagle specific fields:
@@ -92,8 +89,7 @@ def update_peagle(config_dict: dict, pre_trained_config: dict) -> None:
@register_speculator("dflash")
def update_dflash(config_dict: dict, pre_trained_config: dict) -> None:
"""
Apply DFlash specific configuration transformations to the `dict` used to
"""Apply DFlash specific configuration transformations to the `dict` used to
construct the Transformers PreTrainedConfig.
DFlash specific fields:
@@ -132,8 +128,7 @@ def update_dflash(config_dict: dict, pre_trained_config: dict) -> None:
@register_speculator("dflash2")
def update_dflash2(config_dict: dict, pre_trained_config: dict) -> None:
"""
Apply DFlash2 specific configuration transformations to the `dict` used
"""Apply DFlash2 specific configuration transformations to the `dict` used
to construct the Transformers PreTrainedConfig.
DFlash2 extends DFlash with a grouped convolution layer and a
@@ -156,8 +151,7 @@ def update_dflash2(config_dict: dict, pre_trained_config: dict) -> None:
@register_speculator("dspark")
def update_dspark(config_dict: dict, pre_trained_config: dict) -> None:
"""
Apply DSpark specific configuration transformations to the `dict` used to
"""Apply DSpark specific configuration transformations to the `dict` used to
construct the Transformers PreTrainedConfig.
DSpark extends DFlash with a Markov logit-bias head, reusing the same
@@ -47,9 +47,7 @@ class SpeculatorsConfig(PretrainedConfig):
def extract_transformers_pre_trained_config(
cls, config_dict: dict[str, Any]
) -> dict[str, Any]:
"""
Extract standard Transformers PreTrainedConfig config from speculators config.
"""
"""Extract the Transformers `PreTrainedConfig` from a speculators config."""
speculators_model_type = config_dict.get("speculators_model_type")
if speculators_model_type not in SUPPORTED_SPECULATORS_TYPES:
raise ValueError(
@@ -99,8 +97,7 @@ class SpeculatorsConfig(PretrainedConfig):
def build_vllm_speculative_config(
cls, config_dict: dict[str, Any]
) -> dict[str, Any]:
"""
Build vLLM-compatible speculative configuration from speculators format.
"""Build vLLM-compatible speculative configuration from speculators format.
This method extracts and transforms speculative configuration from the
speculators format into the structure expected by vLLM.
@@ -110,6 +107,7 @@ class SpeculatorsConfig(PretrainedConfig):
Returns:
Dictionary with vLLM-compatible speculative configuration
"""
# Extract speculators configuration
spec_config = config_dict["speculators_config"]
+2 -2
View File
@@ -8,8 +8,7 @@ import transformers
class UltravoxConfig(transformers.PretrainedConfig):
r"""
This is the configuration class to store the configuration of a
r"""This is the configuration class to store the configuration of a
[`UltravoxForConditionalGeneration`]. It is used to instantiate an
Ultravox model according to the specified arguments, defining the model
architecture.
@@ -41,6 +40,7 @@ class UltravoxConfig(transformers.PretrainedConfig):
Whether to apply layer normalization at the middle of the
projector or at the end. Versions v0.4.1 and below
use `False`, but v0.5 and above use `True`.
"""
model_type = "ultravox"