Switch to Gradio SDK with DiariZen git install in requirements.txt

- Changed from Docker SDK to Gradio SDK to avoid build complexity
- Added @spaces.GPU decorator for GPU acceleration
- Install DiariZen directly from git in requirements.txt
- Removed Dockerfile (not needed with Gradio SDK)
- Added spaces package for GPU decorator support

Dockerfile

@@ -1,55 +0,0 @@
-# Use NVIDIA CUDA base image (comes with CUDA pre-installed)
-FROM nvidia/cuda:12.1.0-cudnn8-runtime-ubuntu22.04
-
-# Install Python 3.10 and system dependencies
-RUN apt-get update && apt-get install -y \
-    python3.10 \
-    python3-pip \
-    git \
-    ffmpeg \
-    && rm -rf /var/lib/apt/lists/*
-
-# Set Python 3.10 as default
-RUN update-alternatives --install /usr/bin/python python /usr/bin/python3.10 1
-RUN update-alternatives --install /usr/bin/pip pip /usr/bin/pip3 1
-
-# Set working directory
-WORKDIR /app
-
-# Install PyTorch (will use pre-installed CUDA from base image)
-RUN pip install --no-cache-dir torch==2.1.1 torchvision==0.16.1 torchaudio==2.1.1 --index-url https://download.pytorch.org/whl/cu121
-
-# Clone DiariZen repository
-RUN git clone https://github.com/BUTSpeechFIT/DiariZen.git /tmp/diarizen
-
-# Install DiariZen requirements
-WORKDIR /tmp/diarizen
-RUN pip install --no-cache-dir -r requirements.txt
-
-# Install DiariZen package
-RUN pip install --no-cache-dir -e .
-
-# Initialize and update git submodules
-RUN git submodule init && git submodule update
-
-# Install bundled pyannote-audio
-WORKDIR /tmp/diarizen/pyannote-audio
-RUN pip install --no-cache-dir -e .
-
-# Install Gradio (spaces package not needed for Docker SDK)
-WORKDIR /app
-RUN pip install --no-cache-dir gradio==4.27.0 huggingface_hub<0.20.0
-
-# Copy application files
-COPY app.py /app/
-COPY README.md /app/
-
-# Expose port 7860
-EXPOSE 7860
-
-# Set up Gradio environment variables
-ENV GRADIO_SERVER_NAME="0.0.0.0"
-ENV GRADIO_SERVER_PORT=7860
-
-# Run the application
-CMD ["python", "app.py"]

README.md

@@ -3,12 +3,12 @@ title: DiariZen Speaker Diarization
 emoji: 🎙️
 colorFrom: blue
 colorTo: purple
-sdk: docker
-app_port: 7860
+sdk: gradio
+sdk_version: 4.27.0
+app_file: app.py
 suggested_hardware: t4-small
 pinned: false
 license: mit
-short_description: High-performance speaker diarization using DiariZen
 ---
 
 # 🎙️ DiariZen Speaker Diarization

app.py

@@ -1,4 +1,5 @@
 import gradio as gr
+import spaces
 import os
 import torch
 import tempfile
@@ -65,8 +66,9 @@ def save_rttm(annotations, audio_filename):
 
     return str(rttm_path)
 
+@spaces.GPU(duration=120)
 def diarize_audio(audio_file, model_choice):
-    """Main diarization function - GPU available via Docker container"""
+    """Main diarization function with GPU acceleration"""
     if not DIARIZEN_AVAILABLE:
         return "❌ Error: DiariZen not installed. Please install from https://github.com/BUTSpeechFIT/DiariZen", None
 

requirements.txt

@@ -1,2 +1,7 @@
-# This Space uses a Dockerfile for installation
-# See Dockerfile for DiariZen installation steps
+torch==2.1.1
+torchvision==0.16.1
+torchaudio==2.1.1
+gradio==4.27.0
+huggingface_hub<0.20.0
+spaces==0.30.2
+git+https://github.com/BUTSpeechFIT/DiariZen.git