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
+6 -5
View File
@@ -147,11 +147,10 @@ def _ensure_moss_audio_placeholder_tokens(tokenizer: object) -> None:
class MossAudioAudioInputs(TensorSchema):
"""
Dimensions:
- b: Batch size
- nmb: Number of mel bins
- t: Time frames
"""Dimensions:
- b: Batch size
- nmb: Number of mel bins
- t: Time frames
"""
audio_data: Annotated[torch.Tensor, TensorShape("b", "nmb", "t")]
@@ -1091,6 +1090,7 @@ class MossAudioProcessor:
-> audio_data has shape [1, mel_dim, max_time]
-> mel_dim is the number of mel filter-bank bins, 128 by default
-> audio_data_seqlens stores the unpadded mel length
"""
del kwargs
@@ -1723,6 +1723,7 @@ class MossAudioModel(nn.Module, SupportsMultiModal, SupportsPP, SupportsLoRA):
audio_data=[2, 128, 1200], audio_data_seqlens=[800, 1200]
-> returns (audio0_embeds, audio1_embeds), split by token length
-> DeepStack packs each item as [main, layer0, ...] on dim -1
"""
audio_data = audio_input["audio_data"]
audio_data_seqlens = audio_input["audio_data_seqlens"]