Cherry-pick upstream bugfixes PRs #4718, #4720, #4723, #4740, #4741, #4743, #4749

docs/serving/audio_generate_api.md

@@ -66,7 +66,7 @@ Content-Type: application/json
 |-----------|------|---------|-------------|
 | `input` | string | **required** | Text prompt describing the audio to generate |
 | `model` | string | server's model | Model to use (optional, should match server if specified) |
-| `response_format` | string | "wav" | Audio format: wav, mp3, flac, pcm, aac, opus |
+| `response_format` | string | "wav" | Audio format: wav, mp3, flac, pcm, opus |
 | `speed` | float | 1.0 | Playback speed (0.25 - 4.0) |
 
 #### Diffusion Parameters
@@ -90,7 +90,6 @@ Returns binary audio data with the appropriate `Content-Type` header:
 | `mp3` | `audio/mpeg` |
 | `flac` | `audio/flac` |
 | `pcm` | `audio/pcm` |
-| `aac` | `audio/aac` |
 | `opus` | `audio/opus` |
 
 ## Examples
@@ -285,7 +284,7 @@ Pydantic validation failure (e.g. invalid `response_format`, `speed` out of rang
     "detail": [
         {
             "type": "literal_error",
-            "msg": "Input should be 'wav', 'pcm', 'flac', 'mp3', 'aac' or 'opus'",
+            "msg": "Input should be 'wav', 'pcm', 'flac', 'mp3' or 'opus'",
             ...
         }
     ]

docs/serving/speech_api.md

@@ -102,7 +102,7 @@ Content-Type: application/json
 | `input` | string | **required** | The text to synthesize into speech |
 | `model` | string | server's model | Model to use (optional, should match server if specified) |
 | `voice` | string | "vivian" | Speaker name (e.g., vivian, ryan, aiden) |
-| `response_format` | string | "wav" | Audio format: wav, mp3, flac, pcm, aac, opus |
+| `response_format` | string | "wav" | Audio format: wav, mp3, flac, pcm, opus |
 | `speed` | float | 1.0 | Playback speed (0.25-4.0) |
 
 #### vLLM-Omni Extension Parameters

docs/user_guide/examples/online_serving/text_to_audio.md

@@ -139,7 +139,7 @@ Content-Type: application/json
 |-----------|------|---------|-------------|
 | `input` | string | **required** | Text prompt describing the audio to generate |
 | `model` | string | server's model | Model to use (optional, should match server if specified) |
-| `response_format` | string | "wav" | Audio format: wav, mp3, flac, pcm, aac, opus |
+| `response_format` | string | "wav" | Audio format: wav, mp3, flac, pcm, opus |
 | `speed` | float | 1.0 | Playback speed (0.25 - 4.0) |
 | `audio_length` | float | null | Audio duration in seconds (max ~47s for `stable-audio-open-1.0`) |
 | `audio_start` | float | 0.0 | Audio start time in seconds |

tests/dfx/reliability/invalid_param_test/test_invalid_audio_diffusion.py

@@ -11,9 +11,6 @@
 
 pytestmark = [pytest.mark.slow, pytest.mark.diffusion]
 
-# https://github.com/vllm-project/vllm-omni/issues/3649 — invalid inputs still HTTP 200 / timeout / missing validation.
-_SKIP_ISSUE_3649 = pytest.mark.skip(reason="https://github.com/vllm-project/vllm-omni/issues/3649")
-
 _PARAMS = [
     pytest.param(
         OmniServerParams(
@@ -34,10 +31,8 @@
 @pytest.mark.parametrize(
     "overrides, err_message",
     [
-        pytest.param({"input": ""}, ("cannot be empty", "input"), id="input_empty", marks=_SKIP_ISSUE_3649),
-        pytest.param(
-            {"input": "   "}, ("cannot be empty", "input"), id="input_whitespace_only", marks=_SKIP_ISSUE_3649
-        ),
+        pytest.param({"input": ""}, ("cannot be empty", "input"), id="input_empty"),
+        pytest.param({"input": "   "}, ("cannot be empty", "input"), id="input_whitespace_only"),
         pytest.param({"input": 123}, ("input", "a valid string"), id="input_wrong_type"),
         pytest.param(
             {"response_format": "mpeg"},
@@ -52,10 +47,8 @@
             ("sse", "stream_format", "value_error", "not a supported stream_format"),
             id="stream_format_sse_blocked",
         ),
-        pytest.param(
-            {"audio_length": -1.0}, ("audio_end_in_s", "higher than", "audio_start_in_s"), id="audio_length_negative"
-        ),
-        pytest.param({"audio_length": 0}, "audio_length", id="audio_length_zero", marks=_SKIP_ISSUE_3649),
+        pytest.param({"audio_length": -1.0}, ("audio_length", "greater_than", "0"), id="audio_length_negative"),
+        pytest.param({"audio_length": 0}, ("audio_length", "greater_than", "0"), id="audio_length_zero"),
         # ``OmniOpenAIServingAudioGenerate`` only forwards ``audio_start`` when ``audio_length`` is set.
         pytest.param(
             {"audio_start": -0.5, "audio_length": 5.0},
@@ -65,22 +58,27 @@
         pytest.param(
             {"negative_prompt": ["noise"]}, ("negative_prompt", "a valid string"), id="negative_prompt_wrong_type"
         ),
-        pytest.param({"guidance_scale": -1.0}, "guidance_scale", id="guidance_scale_negative", marks=_SKIP_ISSUE_3649),
-        pytest.param({"guidance_scale": 0}, "guidance_scale", id="guidance_scale_zero", marks=_SKIP_ISSUE_3649),
+        pytest.param({"guidance_scale": -1.0}, ("guidance_scale", "greater_than", "0"), id="guidance_scale_negative"),
+        pytest.param({"guidance_scale": 0}, ("guidance_scale", "greater_than", "0"), id="guidance_scale_zero"),
         pytest.param(
-            {"num_inference_steps": 0}, "num_inference_steps", id="num_inference_steps_zero", marks=_SKIP_ISSUE_3649
+            {"num_inference_steps": 0},
+            ("num_inference_steps", "greater_than_equal", "1"),
+            id="num_inference_steps_zero",
         ),
         pytest.param(
-            {"num_inference_steps": -1}, ("number of steps", "non-negative"), id="num_inference_steps_negative"
+            {"num_inference_steps": -1},
+            ("num_inference_steps", "greater_than_equal", "1"),
+            id="num_inference_steps_negative",
         ),
         pytest.param(
             {"num_inference_steps": 6000},
-            "num_inference_steps",
+            ("num_inference_steps", "less_than_equal", "1000"),
             id="num_inference_steps_above_max",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
-            {"guidance_scale": 1001.0}, "guidance_scale", id="guidance_scale_above_max", marks=_SKIP_ISSUE_3649
+            {"guidance_scale": 1001.0},
+            ("guidance_scale", "less_than_equal", "1000"),
+            id="guidance_scale_above_max",
         ),
         pytest.param(
             {"audio_length": 86401.0}, ("Requested audio length", "exceeds maximum"), id="audio_length_above_max"

tests/dfx/reliability/invalid_param_test/test_invalid_audio_speech.py

@@ -37,11 +37,9 @@
 # as "unset". Strip clone fields from the batch envelope before asserting CustomVoice preset errors.
 _SPEECH_BATCH_NO_CLONE_FIELDS = "__speech_batch_no_clone_fields__"
 
-_SKIP_ISSUE_3649 = pytest.mark.skip(reason="https://github.com/vllm-project/vllm-omni/issues/3649")
-
-
 pytestmark = [pytest.mark.slow, pytest.mark.tts]
 
+_SKIP_ISSUE_3649 = pytest.mark.skip(reason="https://github.com/vllm-project/vllm-omni/issues/3649")
 _L4_SPEECH_HW = hardware_marks(res={"cuda": "L4"})
 _SPEECH_SERVER_ARGS = ["--trust-remote-code", "--disable-log-stats"]
 
@@ -114,7 +112,7 @@ def test_speech_missing_required_fields(omni_server: OmniServer, openai_client:
     [
         pytest.param({"input": ""}, ("input", "empty"), id="input_empty"),
         pytest.param({"input": "   "}, ("input", "empty"), id="input_whitespace_only"),
-        pytest.param({"voice": ""}, "Invalid voice", id="voice_empty", marks=_SKIP_ISSUE_3649),
+        pytest.param({"voice": ""}, "Invalid voice", id="voice_empty"),
         pytest.param(
             {"instructions": 123}, ("instructions", "string_type", "valid string"), id="instructions_wrong_type"
         ),
@@ -273,16 +271,15 @@ def test_speech_batch_missing_items(omni_server: OmniServer, openai_client: Open
 @pytest.mark.parametrize(
     "loc, overrides, err_message",
     [
-        pytest.param("item", {"input": ""}, ("input", "empty"), id="item_input_empty", marks=_SKIP_ISSUE_3649),
+        pytest.param("item", {"input": ""}, ("input", "empty"), id="item_input_empty"),
         pytest.param(
             "item",
             {"input": "   "},
             ("input", "empty"),
             id="item_input_whitespace_only",
-            marks=_SKIP_ISSUE_3649,
         ),
-        pytest.param("batch", {"voice": ""}, "Invalid voice", id="batch_voice_empty", marks=_SKIP_ISSUE_3649),
-        pytest.param("item", {"voice": ""}, "Invalid voice", id="item_voice_empty", marks=_SKIP_ISSUE_3649),
+        pytest.param("batch", {"voice": ""}, "Invalid voice", id="batch_voice_empty"),
+        pytest.param("item", {"voice": ""}, "Invalid voice", id="item_voice_empty"),
         pytest.param(
             "batch",
             {
@@ -293,7 +290,6 @@ def test_speech_batch_missing_items(omni_server: OmniServer, openai_client: Open
             },
             "Invalid voice",
             id="batch_voice_unknown_preset",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "item",
@@ -304,7 +300,6 @@ def test_speech_batch_missing_items(omni_server: OmniServer, openai_client: Open
             },
             "Invalid voice",
             id="item_voice_unknown_preset",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "batch",
@@ -323,14 +318,12 @@ def test_speech_batch_missing_items(omni_server: OmniServer, openai_client: Open
             {"instructions": "x" * (_SPEECH_API_MAX_INSTRUCTIONS_CHARS + 1)},
             ("instructions", "too long"),
             id="batch_instructions_exceed_api_limit",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "item",
             {"instructions": "x" * (_SPEECH_API_MAX_INSTRUCTIONS_CHARS + 1)},
             ("instructions", "too long"),
             id="item_instructions_exceed_api_limit",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "batch",
@@ -363,42 +356,36 @@ def test_speech_batch_missing_items(omni_server: OmniServer, openai_client: Open
             {"language": ""},
             ("language", "invalid language", "chinese"),
             id="batch_language_empty",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "item",
             {"language": ""},
             ("language", "invalid language", "chinese"),
             id="item_language_empty",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "batch",
             {"ref_audio": "ftp://example.com/a.wav"},
             ("ref_audio", "URL"),
             id="batch_ref_audio_bad_scheme",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "item",
             {"ref_audio": "ftp://example.com/a.wav"},
             ("ref_audio", "URL"),
             id="item_ref_audio_bad_scheme",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "batch",
             {"ref_audio": "not_a_valid_uri"},
             ("ref_audio", "url"),
             id="batch_ref_audio_invalid_uri",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "item",
             {"ref_audio": "not_a_valid_uri"},
             ("ref_audio", "url"),
             id="item_ref_audio_invalid_uri",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "batch",
@@ -429,28 +416,24 @@ def test_speech_batch_missing_items(omni_server: OmniServer, openai_client: Open
             {"max_new_tokens": 0},
             ("max_new_tokens", "least 1"),
             id="batch_max_new_tokens_below_min",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "item",
             {"max_new_tokens": 0},
             ("max_new_tokens", "least 1"),
             id="item_max_new_tokens_below_min",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "batch",
             {"max_new_tokens": _SPEECH_API_MAX_NEW_TOKENS + 1},
             ("max_new_tokens", "exceed 4096"),
             id="batch_max_new_tokens_above_max",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "item",
             {"max_new_tokens": _SPEECH_API_MAX_NEW_TOKENS + 1},
             ("max_new_tokens", "exceed 4096"),
             id="item_max_new_tokens_above_max",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "batch",
@@ -474,7 +457,6 @@ def test_speech_batch_missing_items(omni_server: OmniServer, openai_client: Open
             },
             ("ref_text", "base task", "non-empty"),
             id="batch_ref_text_whitespace_only_base_clone",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "item",
@@ -485,7 +467,6 @@ def test_speech_batch_missing_items(omni_server: OmniServer, openai_client: Open
             },
             ("ref_text", "base task", "non-empty"),
             id="item_ref_text_whitespace_only_base_clone",
-            marks=_SKIP_ISSUE_3649,
         ),
     ],
 )
@@ -649,21 +630,18 @@ def _voices_upload_multipart_files(kind: str | None) -> dict[str, Any] | None:
             {"consent": "   ", "name": "v_ws_consent_{uuid}"},
             "consent",
             id="whitespace_consent",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "wav_ok",
             {"consent": "bad/consent", "name": "v_bad_cons_{uuid}"},
             "consent",
             id="consent_path_sep",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "wav_ok",
             {"consent": _VF_LONG_CONSENT, "name": "v_long_cons_{uuid}"},
             ("consent", "too long", "Failed to save"),
             id="consent_too_long",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "wav_ok",
@@ -683,7 +661,6 @@ def _voices_upload_multipart_files(kind: str | None) -> dict[str, Any] | None:
             {"consent": "consent_ok", "name": "v_long_ref_{uuid}", "ref_text": _VF_LONG_REF_TEXT},
             "ref_text",
             id="ref_text_too_long",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "wav_ok",
@@ -694,7 +671,6 @@ def _voices_upload_multipart_files(kind: str | None) -> dict[str, Any] | None:
             },
             "speaker_description",
             id="speaker_description_too_long",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "wav_pdf_type",

tests/dfx/reliability/invalid_param_test/test_invalid_image_generation.py

@@ -13,8 +13,6 @@
 
 pytestmark = [pytest.mark.slow, pytest.mark.diffusion]
 
-_SKIP_ISSUE_3649 = pytest.mark.skip(reason="https://github.com/vllm-project/vllm-omni/issues/3649")
-
 _PARAMS = [
     pytest.param(
         OmniServerParams(model="Qwen/Qwen-Image"),
@@ -45,8 +43,8 @@ def _minimal_images_gen_json(omni_server: OmniServer) -> dict[str, object]:
             ("model", "mismatch"),
             id="model_mismatch",
         ),
-        pytest.param({"prompt": ""}, ("prompt", "empty"), id="prompt_empty", marks=_SKIP_ISSUE_3649),
-        pytest.param({"prompt": "   "}, ("prompt", "empty"), id="prompt_whitespace_only", marks=_SKIP_ISSUE_3649),
+        pytest.param({"prompt": ""}, ("prompt", "empty"), id="prompt_empty"),
+        pytest.param({"prompt": "   "}, ("prompt", "empty"), id="prompt_whitespace_only"),
         pytest.param({"prompt": 123}, ("prompt", "string_type", "valid string"), id="prompt_wrong_type"),
         pytest.param({"n": 0}, ("n", "greater_than_equal", "1"), id="n_below_min"),
         pytest.param({"n": 11}, ("n", "less_than_equal", "10"), id="n_above_max"),
@@ -89,18 +87,16 @@ def _minimal_images_gen_json(omni_server: OmniServer) -> dict[str, object]:
         ),
         pytest.param({"layers": 2}, ("layers", "value_error", "3", "10"), id="layers_below_min"),
         pytest.param({"layers": 11}, ("layers", "value_error", "3", "10"), id="layers_above_max"),
-        pytest.param({"seed": -1}, ("seed", "greater_than_equal", "0"), id="seed_negative", marks=_SKIP_ISSUE_3649),
+        pytest.param({"seed": -1}, ("seed", "greater_than_equal", "0"), id="seed_negative"),
         pytest.param(
             {"seed": 2**32},
             ("seed", "less_than_equal", "4294967295"),
             id="seed_above_uint32",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             {"output_format": "gif"},
-            ("output_format", "value_error", "b64_json"),
+            ("output_format", "value_error", "png"),
             id="output_format_invalid",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             {"vae_use_slicing": "wrong_type"},

tests/dfx/reliability/invalid_param_test/test_invalid_omni_chat.py

@@ -15,9 +15,6 @@
 
 pytestmark = [pytest.mark.slow, pytest.mark.omni]
 
-_SKIP_ISSUE_3649 = pytest.mark.skip(reason="https://github.com/vllm-project/vllm-omni/issues/3649")
-
-
 def _minimal_chat_json(omni_server: OmniServer) -> dict[str, object]:
     """Minimal valid chat body; individual tests override one offending field."""
     return {
@@ -112,17 +109,16 @@ def _chat_completions_request_without_expectations(omni_server: OmniServer, case
         pytest.param(
             "modalities_list_bad",
             400,
-            ("modalities", "value_error", ""),
+            ("modalities", "list of strings"),
             id="modalities_list_bad_element",
-            marks=_SKIP_ISSUE_3649,
         ),
         pytest.param(
             "response_format_json_schema_incomplete",
             400,
             ("response_format", "value_error", "json_schema"),
             id="invalid_response_format_json_schema",
         ),
-        pytest.param("logprobs_wrong_type", 400, "logprobs", id="logprobs_wrong_type", marks=_SKIP_ISSUE_3649),
+        pytest.param("logprobs_wrong_type", 400, ("logprobs", "boolean"), id="logprobs_wrong_type"),
         pytest.param(
             "logprobs_top_without_enabled",
             400,