include requirement txt

=2.0

@@ -0,0 +1,191 @@
+Requirement already satisfied: gradio in /usr/local/lib/python3.10/site-packages (5.41.0)
+Collecting transformers
+  Downloading transformers-4.55.0-py3-none-any.whl.metadata (39 kB)
+Collecting torch
+  Downloading torch-2.8.0-cp310-cp310-manylinux_2_28_x86_64.whl.metadata (30 kB)
+Collecting faster-whisper
+  Downloading faster_whisper-1.2.0-py3-none-any.whl.metadata (16 kB)
+Requirement already satisfied: huggingface-hub in /usr/local/lib/python3.10/site-packages (0.34.3)
+Requirement already satisfied: numpy in /usr/local/lib/python3.10/site-packages (2.2.6)
+Collecting scipy
+  Downloading scipy-1.15.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (61 kB)
+Collecting soundfile
+  Downloading soundfile-0.13.1-py2.py3-none-manylinux_2_28_x86_64.whl.metadata (16 kB)
+Collecting ffmpeg-python
+  Downloading ffmpeg_python-0.2.0-py3-none-any.whl.metadata (1.7 kB)
+Requirement already satisfied: aiofiles<25.0,>=22.0 in /usr/local/lib/python3.10/site-packages (from gradio) (24.1.0)
+Requirement already satisfied: anyio<5.0,>=3.0 in /usr/local/lib/python3.10/site-packages (from gradio) (4.10.0)
+Requirement already satisfied: brotli>=1.1.0 in /usr/local/lib/python3.10/site-packages (from gradio) (1.1.0)
+Requirement already satisfied: fastapi<1.0,>=0.115.2 in /usr/local/lib/python3.10/site-packages (from gradio) (0.116.1)
+Requirement already satisfied: ffmpy in /usr/local/lib/python3.10/site-packages (from gradio) (0.6.1)
+Requirement already satisfied: gradio-client==1.11.0 in /usr/local/lib/python3.10/site-packages (from gradio) (1.11.0)
+Requirement already satisfied: groovy~=0.1 in /usr/local/lib/python3.10/site-packages (from gradio) (0.1.2)
+Requirement already satisfied: httpx<1.0,>=0.24.1 in /usr/local/lib/python3.10/site-packages (from gradio) (0.28.1)
+Requirement already satisfied: jinja2<4.0 in /usr/local/lib/python3.10/site-packages (from gradio) (3.1.6)
+Requirement already satisfied: markupsafe<4.0,>=2.0 in /usr/local/lib/python3.10/site-packages (from gradio) (3.0.2)
+Requirement already satisfied: orjson~=3.0 in /usr/local/lib/python3.10/site-packages (from gradio) (3.11.1)
+Requirement already satisfied: packaging in /usr/local/lib/python3.10/site-packages (from gradio) (25.0)
+Requirement already satisfied: pandas<3.0,>=1.0 in /usr/local/lib/python3.10/site-packages (from gradio) (2.3.1)
+Requirement already satisfied: pillow<12.0,>=8.0 in /usr/local/lib/python3.10/site-packages (from gradio) (11.3.0)
+Requirement already satisfied: pydantic<2.12,>=2.0 in /usr/local/lib/python3.10/site-packages (from gradio) (2.11.7)
+Requirement already satisfied: pydub in /usr/local/lib/python3.10/site-packages (from gradio) (0.25.1)
+Requirement already satisfied: python-multipart>=0.0.18 in /usr/local/lib/python3.10/site-packages (from gradio) (0.0.20)
+Requirement already satisfied: pyyaml<7.0,>=5.0 in /usr/local/lib/python3.10/site-packages (from gradio) (6.0.2)
+Requirement already satisfied: ruff>=0.9.3 in /usr/local/lib/python3.10/site-packages (from gradio) (0.12.7)
+Requirement already satisfied: safehttpx<0.2.0,>=0.1.6 in /usr/local/lib/python3.10/site-packages (from gradio) (0.1.6)
+Requirement already satisfied: semantic-version~=2.0 in /usr/local/lib/python3.10/site-packages (from gradio) (2.10.0)
+Requirement already satisfied: starlette<1.0,>=0.40.0 in /usr/local/lib/python3.10/site-packages (from gradio) (0.47.2)
+Requirement already satisfied: tomlkit<0.14.0,>=0.12.0 in /usr/local/lib/python3.10/site-packages (from gradio) (0.13.3)
+Requirement already satisfied: typer<1.0,>=0.12 in /usr/local/lib/python3.10/site-packages (from gradio) (0.16.0)
+Requirement already satisfied: typing-extensions~=4.0 in /usr/local/lib/python3.10/site-packages (from gradio) (4.14.1)
+Requirement already satisfied: uvicorn>=0.14.0 in /usr/local/lib/python3.10/site-packages (from gradio) (0.35.0)
+Requirement already satisfied: fsspec in /usr/local/lib/python3.10/site-packages (from gradio-client==1.11.0->gradio) (2025.3.0)
+Requirement already satisfied: websockets<16.0,>=10.0 in /usr/local/lib/python3.10/site-packages (from gradio-client==1.11.0->gradio) (15.0.1)
+Requirement already satisfied: filelock in /usr/local/lib/python3.10/site-packages (from huggingface-hub) (3.18.0)
+Requirement already satisfied: requests in /usr/local/lib/python3.10/site-packages (from huggingface-hub) (2.32.4)
+Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.10/site-packages (from huggingface-hub) (4.67.1)
+Requirement already satisfied: hf-xet<2.0.0,>=1.1.3 in /usr/local/lib/python3.10/site-packages (from huggingface-hub) (1.1.7)
+Requirement already satisfied: exceptiongroup>=1.0.2 in /usr/local/lib/python3.10/site-packages (from anyio<5.0,>=3.0->gradio) (1.3.0)
+Requirement already satisfied: idna>=2.8 in /usr/local/lib/python3.10/site-packages (from anyio<5.0,>=3.0->gradio) (3.10)
+Requirement already satisfied: sniffio>=1.1 in /usr/local/lib/python3.10/site-packages (from anyio<5.0,>=3.0->gradio) (1.3.1)
+Requirement already satisfied: certifi in /usr/local/lib/python3.10/site-packages (from httpx<1.0,>=0.24.1->gradio) (2025.8.3)
+Requirement already satisfied: httpcore==1.* in /usr/local/lib/python3.10/site-packages (from httpx<1.0,>=0.24.1->gradio) (1.0.9)
+Requirement already satisfied: h11>=0.16 in /usr/local/lib/python3.10/site-packages (from httpcore==1.*->httpx<1.0,>=0.24.1->gradio) (0.16.0)
+Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.10/site-packages (from pandas<3.0,>=1.0->gradio) (2.9.0.post0)
+Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/site-packages (from pandas<3.0,>=1.0->gradio) (2025.2)
+Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.10/site-packages (from pandas<3.0,>=1.0->gradio) (2025.2)
+Requirement already satisfied: annotated-types>=0.6.0 in /usr/local/lib/python3.10/site-packages (from pydantic<2.12,>=2.0->gradio) (0.7.0)
+Requirement already satisfied: pydantic-core==2.33.2 in /usr/local/lib/python3.10/site-packages (from pydantic<2.12,>=2.0->gradio) (2.33.2)
+Requirement already satisfied: typing-inspection>=0.4.0 in /usr/local/lib/python3.10/site-packages (from pydantic<2.12,>=2.0->gradio) (0.4.1)
+Requirement already satisfied: click>=8.0.0 in /usr/local/lib/python3.10/site-packages (from typer<1.0,>=0.12->gradio) (8.0.4)
+Requirement already satisfied: shellingham>=1.3.0 in /usr/local/lib/python3.10/site-packages (from typer<1.0,>=0.12->gradio) (1.5.4)
+Requirement already satisfied: rich>=10.11.0 in /usr/local/lib/python3.10/site-packages (from typer<1.0,>=0.12->gradio) (14.1.0)
+Collecting regex!=2019.12.17 (from transformers)
+  Downloading regex-2025.7.34-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl.metadata (40 kB)
+Collecting tokenizers<0.22,>=0.21 (from transformers)
+  Downloading tokenizers-0.21.4-cp39-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (6.7 kB)
+Collecting safetensors>=0.4.3 (from transformers)
+  Downloading safetensors-0.6.1-cp38-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (4.1 kB)
+Collecting sympy>=1.13.3 (from torch)
+  Downloading sympy-1.14.0-py3-none-any.whl.metadata (12 kB)
+Collecting networkx (from torch)
+  Downloading networkx-3.4.2-py3-none-any.whl.metadata (6.3 kB)
+Collecting nvidia-cuda-nvrtc-cu12==12.8.93 (from torch)
+  Downloading nvidia_cuda_nvrtc_cu12-12.8.93-py3-none-manylinux2010_x86_64.manylinux_2_12_x86_64.whl.metadata (1.7 kB)
+Collecting nvidia-cuda-runtime-cu12==12.8.90 (from torch)
+  Downloading nvidia_cuda_runtime_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.7 kB)
+Collecting nvidia-cuda-cupti-cu12==12.8.90 (from torch)
+  Downloading nvidia_cuda_cupti_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.7 kB)
+Collecting nvidia-cudnn-cu12==9.10.2.21 (from torch)
+  Downloading nvidia_cudnn_cu12-9.10.2.21-py3-none-manylinux_2_27_x86_64.whl.metadata (1.8 kB)
+Collecting nvidia-cublas-cu12==12.8.4.1 (from torch)
+  Downloading nvidia_cublas_cu12-12.8.4.1-py3-none-manylinux_2_27_x86_64.whl.metadata (1.7 kB)
+Collecting nvidia-cufft-cu12==11.3.3.83 (from torch)
+  Downloading nvidia_cufft_cu12-11.3.3.83-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.7 kB)
+Collecting nvidia-curand-cu12==10.3.9.90 (from torch)
+  Downloading nvidia_curand_cu12-10.3.9.90-py3-none-manylinux_2_27_x86_64.whl.metadata (1.7 kB)
+Collecting nvidia-cusolver-cu12==11.7.3.90 (from torch)
+  Downloading nvidia_cusolver_cu12-11.7.3.90-py3-none-manylinux_2_27_x86_64.whl.metadata (1.8 kB)
+Collecting nvidia-cusparse-cu12==12.5.8.93 (from torch)
+  Downloading nvidia_cusparse_cu12-12.5.8.93-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.8 kB)
+Collecting nvidia-cusparselt-cu12==0.7.1 (from torch)
+  Downloading nvidia_cusparselt_cu12-0.7.1-py3-none-manylinux2014_x86_64.whl.metadata (7.0 kB)
+Collecting nvidia-nccl-cu12==2.27.3 (from torch)
+  Downloading nvidia_nccl_cu12-2.27.3-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (2.0 kB)
+Collecting nvidia-nvtx-cu12==12.8.90 (from torch)
+  Downloading nvidia_nvtx_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.8 kB)
+Collecting nvidia-nvjitlink-cu12==12.8.93 (from torch)
+  Downloading nvidia_nvjitlink_cu12-12.8.93-py3-none-manylinux2010_x86_64.manylinux_2_12_x86_64.whl.metadata (1.7 kB)
+Collecting nvidia-cufile-cu12==1.13.1.3 (from torch)
+  Downloading nvidia_cufile_cu12-1.13.1.3-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (1.7 kB)
+Collecting triton==3.4.0 (from torch)
+  Downloading triton-3.4.0-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (1.7 kB)
+Requirement already satisfied: setuptools>=40.8.0 in /usr/local/lib/python3.10/site-packages (from triton==3.4.0->torch) (65.5.1)
+Collecting ctranslate2<5,>=4.0 (from faster-whisper)
+  Downloading ctranslate2-4.6.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (10 kB)
+Collecting onnxruntime<2,>=1.14 (from faster-whisper)
+  Downloading onnxruntime-1.22.1-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (4.6 kB)
+Collecting av>=11 (from faster-whisper)
+  Downloading av-15.0.0-cp310-cp310-manylinux_2_28_x86_64.whl.metadata (4.6 kB)
+Collecting coloredlogs (from onnxruntime<2,>=1.14->faster-whisper)
+  Downloading coloredlogs-15.0.1-py2.py3-none-any.whl.metadata (12 kB)
+Collecting flatbuffers (from onnxruntime<2,>=1.14->faster-whisper)
+  Downloading flatbuffers-25.2.10-py2.py3-none-any.whl.metadata (875 bytes)
+Requirement already satisfied: protobuf in /usr/local/lib/python3.10/site-packages (from onnxruntime<2,>=1.14->faster-whisper) (3.20.3)
+Requirement already satisfied: cffi>=1.0 in /usr/local/lib/python3.10/site-packages (from soundfile) (1.17.1)
+Collecting future (from ffmpeg-python)
+  Downloading future-1.0.0-py3-none-any.whl.metadata (4.0 kB)
+Requirement already satisfied: pycparser in /usr/local/lib/python3.10/site-packages (from cffi>=1.0->soundfile) (2.22)
+Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/site-packages (from python-dateutil>=2.8.2->pandas<3.0,>=1.0->gradio) (1.17.0)
+Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.10/site-packages (from rich>=10.11.0->typer<1.0,>=0.12->gradio) (3.0.0)
+Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.10/site-packages (from rich>=10.11.0->typer<1.0,>=0.12->gradio) (2.19.2)
+Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.10/site-packages (from markdown-it-py>=2.2.0->rich>=10.11.0->typer<1.0,>=0.12->gradio) (0.1.2)
+Collecting mpmath<1.4,>=1.1.0 (from sympy>=1.13.3->torch)
+  Downloading mpmath-1.3.0-py3-none-any.whl.metadata (8.6 kB)
+Collecting humanfriendly>=9.1 (from coloredlogs->onnxruntime<2,>=1.14->faster-whisper)
+  Downloading humanfriendly-10.0-py2.py3-none-any.whl.metadata (9.2 kB)
+Requirement already satisfied: charset_normalizer<4,>=2 in /usr/local/lib/python3.10/site-packages (from requests->huggingface-hub) (3.4.2)
+Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/site-packages (from requests->huggingface-hub) (2.5.0)
+Downloading transformers-4.55.0-py3-none-any.whl (11.3 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.3/11.3 MB 157.2 MB/s  0:00:00
+Downloading tokenizers-0.21.4-cp39-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (3.1 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 3.1/3.1 MB 168.3 MB/s  0:00:00
+Downloading torch-2.8.0-cp310-cp310-manylinux_2_28_x86_64.whl (888.0 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 888.0/888.0 MB 105.1 MB/s  0:00:06
+Downloading nvidia_cublas_cu12-12.8.4.1-py3-none-manylinux_2_27_x86_64.whl (594.3 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 594.3/594.3 MB 258.5 MB/s  0:00:01
+Downloading nvidia_cuda_cupti_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (10.2 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 10.2/10.2 MB 226.6 MB/s  0:00:00
+Downloading nvidia_cuda_nvrtc_cu12-12.8.93-py3-none-manylinux2010_x86_64.manylinux_2_12_x86_64.whl (88.0 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 88.0/88.0 MB 378.9 MB/s  0:00:00
+Downloading nvidia_cuda_runtime_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (954 kB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 954.8/954.8 kB 69.2 MB/s  0:00:00
+Downloading nvidia_cudnn_cu12-9.10.2.21-py3-none-manylinux_2_27_x86_64.whl (706.8 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 706.8/706.8 MB 286.2 MB/s  0:00:01
+Downloading nvidia_cufft_cu12-11.3.3.83-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (193.1 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 193.1/193.1 MB 163.9 MB/s  0:00:01
+Downloading nvidia_cufile_cu12-1.13.1.3-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (1.2 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.2/1.2 MB 89.5 MB/s  0:00:00
+Downloading nvidia_curand_cu12-10.3.9.90-py3-none-manylinux_2_27_x86_64.whl (63.6 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 63.6/63.6 MB 303.5 MB/s  0:00:00
+Downloading nvidia_cusolver_cu12-11.7.3.90-py3-none-manylinux_2_27_x86_64.whl (267.5 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 267.5/267.5 MB 265.9 MB/s  0:00:00
+Downloading nvidia_cusparse_cu12-12.5.8.93-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (288.2 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 288.2/288.2 MB 383.7 MB/s  0:00:00
+Downloading nvidia_cusparselt_cu12-0.7.1-py3-none-manylinux2014_x86_64.whl (287.2 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 287.2/287.2 MB 389.0 MB/s  0:00:00
+Downloading nvidia_nccl_cu12-2.27.3-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (322.4 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 322.4/322.4 MB 298.2 MB/s  0:00:01
+Downloading nvidia_nvjitlink_cu12-12.8.93-py3-none-manylinux2010_x86_64.manylinux_2_12_x86_64.whl (39.3 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 39.3/39.3 MB 378.8 MB/s  0:00:00
+Downloading nvidia_nvtx_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (89 kB)
+Downloading triton-3.4.0-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (155.4 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 155.4/155.4 MB 163.6 MB/s  0:00:00
+Downloading faster_whisper-1.2.0-py3-none-any.whl (1.1 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.1/1.1 MB 76.0 MB/s  0:00:00
+Downloading ctranslate2-4.6.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (38.4 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 38.4/38.4 MB 161.9 MB/s  0:00:00
+Downloading onnxruntime-1.22.1-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (16.5 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 16.5/16.5 MB 163.1 MB/s  0:00:00
+Downloading scipy-1.15.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (37.7 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 37.7/37.7 MB 395.4 MB/s  0:00:00
+Downloading soundfile-0.13.1-py2.py3-none-manylinux_2_28_x86_64.whl (1.3 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.3/1.3 MB 33.6 MB/s  0:00:00
+Downloading ffmpeg_python-0.2.0-py3-none-any.whl (25 kB)
+Downloading av-15.0.0-cp310-cp310-manylinux_2_28_x86_64.whl (39.2 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 39.2/39.2 MB 163.0 MB/s  0:00:00
+Downloading regex-2025.7.34-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (789 kB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 789.8/789.8 kB 50.5 MB/s  0:00:00
+Downloading safetensors-0.6.1-cp38-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (485 kB)
+Downloading sympy-1.14.0-py3-none-any.whl (6.3 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 6.3/6.3 MB 209.0 MB/s  0:00:00
+Downloading mpmath-1.3.0-py3-none-any.whl (536 kB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 536.2/536.2 kB 18.0 MB/s  0:00:00
+Downloading coloredlogs-15.0.1-py2.py3-none-any.whl (46 kB)
+Downloading humanfriendly-10.0-py2.py3-none-any.whl (86 kB)
+Downloading flatbuffers-25.2.10-py2.py3-none-any.whl (30 kB)
+Downloading future-1.0.0-py3-none-any.whl (491 kB)
+Downloading networkx-3.4.2-py3-none-any.whl (1.7 MB)
+   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.7/1.7 MB 129.1 MB/s  0:00:00
+Installing collected packages: nvidia-cusparselt-cu12, mpmath, flatbuffers, triton, sympy, scipy, safetensors, regex, nvidia-nvtx-cu12, nvidia-nvjitlink-cu12, nvidia-nccl-cu12, nvidia-curand-cu12, nvidia-cufile-cu12, nvidia-cuda-runtime-cu12, nvidia-cuda-nvrtc-cu12, nvidia-cuda-cupti-cu12, nvidia-cublas-cu12, networkx, humanfriendly, future, ctranslate2, av, soundfile, nvidia-cusparse-cu12, nvidia-cufft-cu12, nvidia-cudnn-cu12, ffmpeg-python, coloredlogs, tokenizers, onnxruntime, nvidia-cusolver-cu12, transformers, torch, faster-whisper
+

app.py

@@ -1,59 +1,239 @@
+# app.py
+# Minimal Gradio voice-assistant demo for Hugging Face Spaces.
+# - STT: faster-whisper (WhisperModel)
+# - Sentiment: transformers pipeline
+# - LLM: local HF model (small) or HF Inference API (recommended for huge models)
+# - TTS: bosonai/higgs-audio-v2 (when available)
+#
+# NOTES:
+# - Replace model names with versions you prefer.
+# - If you want streaming / low-latency WebRTC, see the "Advanced: WebRTC / streaming" comments below.
+
+import os
+import tempfile
+import uuid
+from pathlib import Path
+from faster_whisper import WhisperModel
+from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
+import torch
 import gradio as gr
-from transformers import pipeline
-import numpy as np
-
-transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-base.en")
-
-def transcribe(stream, new_chunk):
-    sr, y = new_chunk
-    
-    # Convert to mono if stereo
-    if y.ndim > 1:
-        y = y.mean(axis=1)
-        
-    y = y.astype(np.float32)
-    y /= np.max(np.abs(y))
-
-    if stream is not None:
-        stream = np.concatenate([stream, y])
+from huggingface_hub import InferenceApi, hf_hub_download
+
+# --------------------
+# CONFIG / ENV VARS
+# --------------------
+HUGGINGFACEHUB_API_TOKEN = os.environ.get("HUGGINGFACEHUB_API_TOKEN")  # optional for hosted inference
+USE_HOSTED_LLM = bool(os.environ.get("USE_HOSTED_LLM", ""))  # set to "1" to use hosted HF inference instead of a local model
+LLM_MODEL_NAME = os.environ.get("LLM_MODEL_NAME", "gpt2")  # default small model for Spaces demo
+TTS_MODEL = os.environ.get("TTS_MODEL", "bosonai/higgs-audio-v2-generation-3B-base")  # example; may need HF token
+DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
+
+# --------------------
+# Initialize models
+# --------------------
+
+# Faster Whisper STT
+# Choose model size: "small", "medium", "large-v2", etc. Smaller -> lower latency.
+whisper_model_size = "small"
+whisper = WhisperModel(whisper_model_size, device=DEVICE, compute_type="int8_float16")  # compute_type optional based on your hardware
+
+# Sentiment / conversation cue analysis
+sentiment = pipeline("sentiment-analysis", device=0 if DEVICE == "cuda" else -1)
+
+# LLM (local fallback)
+local_llm = None
+local_tokenizer = None
+if not USE_HOSTED_LLM:
+    try:
+        # Keep this small on Spaces by default. Swap to a bigger model if you have the hardware.
+        local_llm_name = LLM_MODEL_NAME  # e.g., "gpt2", "EleutherAI/gpt-neo-125M", or an HF chat model
+        local_tokenizer = AutoTokenizer.from_pretrained(local_llm_name)
+        local_llm = AutoModelForCausalLM.from_pretrained(local_llm_name, device_map="auto" if DEVICE=="cuda" else None)
+    except Exception as e:
+        print("Failed to load local LLM:", e)
+        local_llm = None
+
+# Hosted inference (optional)
+hf_inference = None
+if USE_HOSTED_LLM and HUGGINGFACEHUB_API_TOKEN:
+    hf_inference = InferenceApi(repo_id=os.environ.get("HF_INFERENCE_MODEL", "openai/gpt-oss-7b"), token=HUGGINGFACEHUB_API_TOKEN)
+
+
+# TTS: This is an illustration. For actual use, replace with a tested TTS pipeline suited to the model.
+def tts_generate(text: str, out_path: str) -> str:
+    """
+    Generate TTS audio for `text` and save to out_path.
+    For Higgs Audio v2 (BosonAI) you likely need a specific inference client or HF pipeline that supports that repo.
+    This function is a stub illustrating where TTS would be called.
+    """
+    # Option A: If the TTS repo provides a pipeline via transformers (pseudo):
+    # from transformers import pipeline
+    # tts_pipe = pipeline("text-to-speech", model=TTS_MODEL, use_auth_token=HUGGINGFACEHUB_API_TOKEN)
+    # audio_array, sr = tts_pipe(text)
+    # save audio to out_path here...
+    #
+    # Option B: If using hosted HF inference:
+    if HUGGINGFACEHUB_API_TOKEN and os.environ.get("USE_HOSTED_TTS"):
+        # Using HF Inference API (simplified)
+        from huggingface_hub import InferenceApi
+        inf = InferenceApi(repo_id=TTS_MODEL, token=HUGGINGFACEHUB_API_TOKEN)
+        res = inf(text, {"wait_for_model": True})
+        # hf returns audio in base64 or a URL depending on model; handle accordingly (pseudo)
+        # Save binary to out_path...
+    # Fallback: return an empty audio or a short beep (silence)
+    # For demo, create a tiny silent WAV so Gradio can play something:
+    import wave, struct
+    with wave.open(out_path, "wb") as wf:
+        n_channels = 1
+        sampwidth = 2
+        framerate = 16000
+        n_frames = framerate // 10  # 0.1s silence
+        wf.setnchannels(n_channels)
+        wf.setsampwidth(sampwidth)
+        wf.setframerate(framerate)
+        wf.writeframes(b"\x00\x00" * n_frames)
+    return out_path
+
+# --------------------
+# Core pipeline functions
+# --------------------
+
+def transcribe_audio(audio_file: str):
+    """
+    audio_file: path to uploaded audio (wav/m4a/ogg)
+    returns: transcription string and whisper segments (list)
+    """
+    segments = []
+    text_acc = []
+    # faster-whisper: transcribe file
+    # model.transcribe returns generator of segments when streaming; we call full transcription here
+    try:
+        result = whisper.transcribe(audio_file, beam_size=5, word_timestamps=False)
+        # result is a list of segments or object depending on version. We'll handle simple case:
+        # If result has 'segments' attribute (older/variant) handle it:
+        if hasattr(result, "segments"):
+            for seg in result.segments:
+                segments.append({"start": seg.start, "end": seg.end, "text": seg.text})
+                text_acc.append(seg.text)
+        else:
+            # result may be tuple/text depending on faster-whisper version
+            if isinstance(result, dict) and "text" in result:
+                text_acc.append(result["text"])
+            elif isinstance(result, str):
+                text_acc.append(result)
+    except Exception as e:
+        print("STT error:", e)
+        return "",""
+
+    transcription = " ".join(text_acc).strip()
+    return transcription, segments
+
+def analyze_sentiment(text: str):
+    if not text:
+        return {"label":"NEUTRAL", "score": 0.0}
+    try:
+        out = sentiment(text[:1000])  # limit to first 1000 chars
+        return out[0] if isinstance(out, list) and out else out
+    except Exception as e:
+        print("Sentiment error:", e)
+        return {"label":"UNKNOWN", "score": 0.0}
+
+def llm_respond(prompt: str, conversation_history=None, max_length=256):
+    """
+    Returns text response for prompt. Uses hosted HF inference if USE_HOSTED_LLM else local model.
+    conversation_history: list of (user, assistant) tuples (not deeply used in demo)
+    """
+    if USE_HOSTED_LLM and hf_inference:
+        # Use HF Inference API (simplified)
+        payload = {"inputs": prompt}
+        try:
+            result = hf_inference(prompt)
+            # result could be a dict with 'generated_text' or a string
+            if isinstance(result, dict) and "generated_text" in result:
+                return result["generated_text"]
+            elif isinstance(result, str):
+                return result
+            else:
+                return str(result)
+        except Exception as e:
+            print("Hosted LLM error:", e)
+            return "Sorry — I couldn't reach the hosted model."
+    elif local_llm and local_tokenizer:
+        # Local generation (small models suitable for Spaces)
+        inputs = local_tokenizer(prompt, return_tensors="pt").to(DEVICE)
+        outputs = local_llm.generate(**inputs, max_new_tokens=max_length)
+        return local_tokenizer.decode(outputs[0], skip_special_tokens=True)
+    else:
+        return "No LLM available. Set USE_HOSTED_LLM=1 and provide a HUGGINGFACEHUB_API_TOKEN or load a local model."
+
+# --------------------
+# Gradio UI callbacks
+# --------------------
+
+def process_audio_and_respond(audio):
+    """
+    audio: a tuple or path from gradio Audio component
+    Returns: transcription, sentiment label, assistant text, tts_audio_path
+    """
+    if audio is None:
+        return "","", "No audio received.", None
+    # gradio audio file can be a dict or tuple; handle both
+    if isinstance(audio, tuple) or isinstance(audio, list):
+        # (sample_rate, numpy_array) OR (path, sample_rate)
+        # gradio sends a path when type="filepath"
+        audio_path = audio[0] if isinstance(audio[0], str) else None
+    elif isinstance(audio, str) and Path(audio).exists():
+        audio_path = audio
     else:
-        stream = y
-    return stream
-
-demo = gr.Interface(
-    transcribe,
-    ["state", gr.Audio(sources=["microphone"], streaming=True)],
-    ["state", "text"],
-    live=True,
-)
-
-demo.launch()
-
-# transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-base.en")
-
-# def transcribe(stream, new_chunk):
-#     sr, y = new_chunk
-    
-#     # Convert to mono if stereo
-#     if y.ndim > 1:
-#         y = y.mean(axis=1)
-        
-#     y = y.astype(np.float32)
-#     y /= np.max(np.abs(y))
-
-#     if stream is not None:
-#         stream = np.concatenate([stream, y])
-#     else:
-#         stream = y
-    
-#     # Return the stream as state and a string representation of the array for display
-#     return stream, str(stream)
-
-# demo = gr.Interface(
-#     transcribe,
-#     ["state", gr.Audio(sources=["microphone"], streaming=True)],
-#     ["state", "text"],
-#     live=True,
-# )
-
-# demo.launch()
+        # try to handle (np_array, sr)
+        import soundfile as sf
+        tmp = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)
+        sf.write(tmp.name, audio[0], audio[1])
+        audio_path = tmp.name
+
+    transcription, segments = transcribe_audio(audio_path)
+    sentiment_res = analyze_sentiment(transcription)
+    sentiment_label = sentiment_res.get("label", "UNKNOWN")
+    # Build a prompt for LLM
+    prompt = f"User: {transcription}\n\nAssistant:"
+    assistant_text = llm_respond(prompt)
+
+    # Synthesize TTS (non-blocking improvement: spawn background job)
+    out_audio_path = f"/tmp/assistant_tts_{uuid.uuid4().hex}.wav"
+    tts_generate(assistant_text, out_audio_path)
+
+    return transcription, sentiment_label, assistant_text, out_audio_path
+
+# --------------------
+# GRADIO APP
+# --------------------
+with gr.Blocks(title="Voice Assistant (Faster Whisper + LLM + Higgs Audio)") as demo:
+    gr.Markdown("# Voice Assistant demo")
+    with gr.Row():
+        with gr.Column(scale=2):
+            audio_in = gr.Audio(source="microphone", type="filepath", label="Speak (record)", show_label=True)
+            btn = gr.Button("Send")
+            status = gr.Label("Status: idle")
+        with gr.Column(scale=3):
+            transcript_out = gr.Textbox(label="Transcription", interactive=False)
+            sentiment_out = gr.Textbox(label="Sentiment", interactive=False)
+            assistant_out = gr.Textbox(label="Assistant response", interactive=False)
+            audio_out = gr.Audio(label="Assistant voice", interactive=False)
+
+    def on_click(audio):
+        status.value = "Processing..."
+        transcription, sentiment_label, assistant_text, tts_path = process_audio_and_respond(audio)
+        status.value = "Done"
+        return gr.update(value=transcription), gr.update(value=sentiment_label), gr.update(value=assistant_text), gr.update(value=tts_path)
+
+    btn.click(on_click, inputs=[audio_in], outputs=[transcript_out, sentiment_out, assistant_out, audio_out])
+
+    gr.Markdown(
+        "### Notes\n"
+        "- This demo uses Faster Whisper for STT, a transformers sentiment classifier, and a small LLM by default.\n"
+        "- Set `USE_HOSTED_LLM=1` & provide `HUGGINGFACEHUB_API_TOKEN` to use a hosted LLM (e.g. gpt-oss variants) via the HF Inference API.\n"
+        "- To enable richer TTS (Higgs Audio v2), set `USE_HOSTED_TTS=1` and make sure your HF token has access to the model if it's gated.\n"
+    )
+
+if __name__ == '__main__':
+    demo.launch(server_name="0.0.0.0", server_port=int(os.environ.get("PORT", 7860)))

requirements.txt

@@ -0,0 +1,11 @@
+gradio>=3.30
+transformers>=4.30
+torch>=2.0
+faster-whisper
+huggingface-hub
+numpy
+scipy
+soundfile
+ffmpeg-python
+# optional for more advanced WebRTC:
+# fastrtc