[Bugfix] (qwen3_tts): enable batched offline inference by fixing tensor slicing

Signed-off-by: Roman Koshkin <roman.koshkin@sbintuitions.co.jp>

Author: RomanKoshkin

.gitignore

@@ -245,3 +245,5 @@ tmp_test
 
 # output files
 *.wav
+examples/offline_inference/qwen3_tts/test.py
+examples/online_serving/qwen3_tts/Untitled.ipynb

READMEmy.md

@@ -0,0 +1,25 @@
+
+
+```bash
+cd /lustre/users/rkoshkin
+git clone https://github.com/vllm-project/vllm-omni.git
+cd vllm-omni
+uv venv --python 3.10 --seed
+source .venv/bin/activate
+cd ..
+git clone https://github.com/vllm-project/vllm.git
+cd vllm
+git checkout v0.16.0
+export VLLM_PRECOMPILED_WHEEL_LOCATION=https://github.com/vllm-project/vllm/releases/download/v0.16.0/vllm-0.16.0-cp38-abi3-manylinux_2_31_x86_64.whl
+uv pip install -e .
+cd ../vllm-omni
+uv pip install -e .
+```
+
+
+
+```bash
+# edit /lustre/users/rkoshkin/vllm-omni/vllm_omni/model_executor/stage_configs/qwen3_tts.yaml AS NECESSARY
+cd examples/online_serving/qwen3_tts
+./run_server.sh Base
+```