[Refactor] Output Processor Phase 2: separate multimodal output channel (#1601) (#2744)

Signed-off-by: Megha Agarwal <agarwalmegha1308@gmail.com>
Signed-off-by: meghaagr13 <divyanshsinghvi@gmail.com>
Signed-off-by: Taichang Zhou <tzhouam@connect.ust.hk>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: meghaagr13 <divyanshsinghvi@gmail.com>
Co-authored-by: Boao Shi <aoibosh@connect.hku.hk>
Co-authored-by: Taichang Zhou <tzhouam@connect.ust.hk>
Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
authored and GitHub committed 2026-06-12 15:19:14 +08:00
1 parent 1be1a71156
commit 8f869c8ea5
49 files changed
+985 -327

No files matched your search

@@ -1,4 +1,4 @@
from collections.abc import Iterable
from collections.abc import Iterable, Mapping
from dataclasses import replace
from functools import cached_property
from pathlib import Path
@@ -220,7 +220,7 @@ class VoxtralTTSForConditionalGeneration(
def tts_postprocess(self, hidden_states: torch.Tensor, multimodal_outputs: object, **info_dict: object | None):
update_dict = {}
if isinstance(multimodal_outputs, dict) and "audio" in multimodal_outputs:
if isinstance(multimodal_outputs, Mapping) and "audio" in multimodal_outputs:
assert self.post_process_idx < len(multimodal_outputs["audio"]), (
f"Expect {self.post_process_idx=} < {len(multimodal_outputs['audio'])=}"
)