tzhouamandClaude Fable 5.1
49a839ecf9
* qwen3_tts_talker_only pipeline (single terminal LLM_AR stage, final_output_type "latent") + vllm_omni/deploy/qwen3_tts_talker_only.yaml; a terminal talker publishes its per-step codes.audio rows to the client (omni_client_multimodal_output_keys). * /v1/audio/speech/stream: session.config.output_mode="codec_tokens" streams codec.start (JSON) + binary frames (<IHB seq/n_frames/flags + int16 rows, vllm_omni/entrypoints/openai/codec_stream.py) + codec.done; new input.cancel message aborts an in-flight generation for both codec and PCM sessions (single reader task + session inbox replaces direct socket reads); optional session.config.seed; audio.done gains "cancelled" only when it happened. * serving_speech._generate_codec_chunks aligns the cumulative code tensor to output token ids (drops prefill placeholder/EOS rows) and emits deltas. * TTS adapters: supports_codec_stream / codec_stream_spec (Qwen3-TTS: 16 codebooks, 2048 ids, 12.5 Hz, 24 kHz, decoder qwen3_tts_tokenizer_v2_12hz). * examples/online_serving/text_to_speech/qwen3_tts/edge_decoder_client.py decodes locally with Qwen3TTSTokenizerV2Decoder (ramp schedule, 72-frame context, cancel/reset), codec_parity_check.py and benchmarks/tts/codec_stream_e2e.py compare against the two-stage server. * docs/serving/codec_token_streaming.md; CPU unit tests for the wire format, alignment, session config and handler (frames, unsupported adapter, cancel); L3 e2e tests/e2e/online_serving/test_qwen3_tts_codec_stream.py. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
86 lines
3.3 KiB
Python
86 lines
3.3 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
|
"""Qwen3-TTS pipeline: Talker (text → RVQ codec) → Code2Wav (codec → audio).
|
|
|
|
Chunked vs end-to-end mode is dispatched from ``deploy.async_chunk``.
|
|
"""
|
|
|
|
from vllm_omni.config.stage_config import (
|
|
PipelineConfig,
|
|
StageExecutionType,
|
|
StagePipelineConfig,
|
|
)
|
|
|
|
_PROC = "vllm_omni.model_executor.stage_input_processors.qwen3_tts"
|
|
|
|
QWEN3_TTS_PIPELINE = PipelineConfig(
|
|
model_type="qwen3_tts",
|
|
default_deploy_config_name="qwen3_tts.yaml",
|
|
# Pipeline-level default; the code2wav stage overrides per-stage below.
|
|
model_arch="Qwen3TTSTalkerForConditionalGeneration",
|
|
stages=(
|
|
StagePipelineConfig(
|
|
stage_id=0,
|
|
model_stage="qwen3_tts",
|
|
execution_type=StageExecutionType.LLM_AR,
|
|
input_sources=(),
|
|
owns_tokenizer=True,
|
|
engine_output_type="latent",
|
|
async_chunk_process_next_stage_input_func=(f"{_PROC}.talker2code2wav_async_chunk"),
|
|
custom_process_next_stage_input_func=f"{_PROC}.talker2code2wav_full_payload",
|
|
sampling_constraints={
|
|
"detokenize": False,
|
|
"stop_token_ids": [2150],
|
|
},
|
|
),
|
|
StagePipelineConfig(
|
|
stage_id=1,
|
|
model_stage="code2wav",
|
|
execution_type=StageExecutionType.LLM_GENERATION,
|
|
input_sources=(0,),
|
|
final_output=True,
|
|
final_output_type="audio",
|
|
engine_output_type="audio",
|
|
model_arch="Qwen3TTSCode2Wav",
|
|
# ``sync_process_input_func`` is the only input-proc override for
|
|
# this stage in sync (non-async-chunk) mode: a length-only
|
|
# ``_token_only`` placeholder. The bulk codec payload itself
|
|
# ships via the worker connector from stage 0's
|
|
# ``talker2code2wav_full_payload`` producer. Under async_chunk
|
|
# mode no pre-stage processing is needed -- chunks deliver
|
|
# directly to the consumer.
|
|
sync_process_input_func=f"{_PROC}.talker2code2wav_token_only",
|
|
sampling_constraints={"detokenize": True},
|
|
extras={"tts_args": {"max_instructions_length": 500}},
|
|
requires_full_payload_input=True,
|
|
),
|
|
),
|
|
)
|
|
|
|
# Talker-only pipeline: stage 0 is terminal and its codec frames (codes.audio)
|
|
# are published to the client for on-device decoding (codec-token streaming,
|
|
# see entrypoints/openai/codec_stream.py). No next-stage processors, so no
|
|
# omni connector is created; ``final_output_type="latent"`` keeps the AR
|
|
# output path (audio-specific terminal handling is not applied).
|
|
QWEN3_TTS_TALKER_ONLY_PIPELINE = PipelineConfig(
|
|
model_type="qwen3_tts_talker_only",
|
|
default_deploy_config_name="qwen3_tts_talker_only.yaml",
|
|
model_arch="Qwen3TTSTalkerForConditionalGeneration",
|
|
stages=(
|
|
StagePipelineConfig(
|
|
stage_id=0,
|
|
model_stage="qwen3_tts",
|
|
execution_type=StageExecutionType.LLM_AR,
|
|
input_sources=(),
|
|
owns_tokenizer=True,
|
|
final_output=True,
|
|
final_output_type="latent",
|
|
engine_output_type="latent",
|
|
sampling_constraints={
|
|
"detokenize": False,
|
|
"stop_token_ids": [2150],
|
|
},
|
|
),
|
|
),
|
|
)
|