adapt voxtral (#31095)

Signed-off-by: Patrick von Platen <patrick.v.platen@gmail.com>

tests/models/multimodal/generation/test_voxtral.py

@@ -111,4 +111,5 @@ async def test_online_serving(client, audio_assets: AudioTestAssets):
 
     assert len(chat_completion.choices) == 1
     choice = chat_completion.choices[0]
+    assert choice.message.content == "In the first audio clip, you hear a brief"
     assert choice.finish_reason == "length"

tests/models/registry.py

@@ -860,6 +860,11 @@ _MULTIMODAL_EXAMPLE_MODELS = {
         # disable this temporarily until we support HF format
         is_available_online=False,
     ),
+    "VoxtralStreamingGeneration": _HfExamplesInfo(
+        "<place-holder>",
+        # disable this temporarily until we support HF format
+        is_available_online=False,
+    ),
     # [Encoder-decoder]
     "WhisperForConditionalGeneration": _HfExamplesInfo(
         "openai/whisper-large-v3-turbo",