dillon_stuff/TalkEdit
1
0
Fork 0
Code Issues 1 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

fix: Remove torch from requirements.txt and use flexible versions

Browse Source 
- Remove torch/torchaudio/torchvision from requirements.txt (installed separately in Docker)
- Use >= instead of == for most packages to avoid version conflicts
- Install numpy before other requirements
- Add setuptools and wheel to pip upgrade step
This commit is contained in:
Your Name
2026-01-05 11:25:12 -05:00
parent 78e9df31e6
commit 4dd3c7600e
2 changed files with 23 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
Dockerfile.gpu
View File

@ -15,16 +15,17 @@ RUN apt-get update && apt-get install -y \
# Copy requirements first for better Docker layer caching # Copy requirements first for better Docker layer caching
COPY requirements.txt . COPY requirements.txt .
# Upgrade pip first to avoid installation issues # Upgrade pip and install build tools
RUN pip install --upgrade pip RUN pip install --upgrade pip setuptools wheel
# Install CUDA-optimized PyTorch FIRST (before other requirements) # Install CUDA-optimized PyTorch FIRST (before other requirements)
# Using latest versions from cu118 index for SpeechBrain 1.0 / pyannote diarization compatibility # Using latest versions from cu118 index for SpeechBrain 1.0 / pyannote diarization compatibility
# Not pinning versions - pip will resolve latest compatible versions from the index RUN pip install --no-cache-dir torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
RUN pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
# Install numpy first as many packages depend on it
RUN pip install --no-cache-dir "numpy>=1.24.0"
# Install remaining dependencies from requirements.txt # Install remaining dependencies from requirements.txt
# Skip torch/torchvision/torchaudio since they're already installed
RUN pip install --no-cache-dir -r requirements.txt RUN pip install --no-cache-dir -r requirements.txt
# Copy application code # Copy application code

36
requirements.txt
View File

@ -1,32 +1,30 @@
# OBS Recording Transcriber Dependencies # OBS Recording Transcriber Dependencies
# Core dependencies with pinned compatible versions # Core dependencies
streamlit==1.26.0 streamlit>=1.26.0
moviepy==1.0.3 moviepy>=1.0.3
openai-whisper==20231117 openai-whisper>=20231117
requests>=2.28.0 requests>=2.28.0
humanize>=4.6.0 humanize>=4.6.0
# PyTorch ecosystem - updated for SpeechBrain 1.0 compatibility # PyTorch ecosystem - DO NOT include here for Docker builds
# These are installed separately with CUDA support in Dockerfile.gpu
# For local installs: pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
# torchaudio >= 2.1.0 is REQUIRED for diarization to work properly # torchaudio >= 2.1.0 is REQUIRED for diarization to work properly
# NOTE: For Docker GPU builds, these are installed separately from cu118 index
torch>=2.1.0
torchaudio>=2.1.0
torchvision>=0.16.0
# Transformers ecosystem - compatible versions # Transformers ecosystem
transformers==4.35.0 transformers>=4.35.0
tokenizers>=0.14.0 tokenizers>=0.14.0
# ML dependencies with compatible versions # ML dependencies - use flexible versions for compatibility
numpy==1.24.3 numpy>=1.24.0
scipy==1.10.1 scipy>=1.10.0
scikit-learn==1.3.0 scikit-learn>=1.3.0
# Audio processing and ML models # Audio processing and ML models
# Pin speechbrain for stability with pyannote.audio # speechbrain 1.0+ required for pyannote compatibility
speechbrain==1.0.0 speechbrain>=1.0.0
pyannote.audio==3.1.1 pyannote.audio>=3.1.1
pytorch-lightning==2.1.0 pytorch-lightning>=2.0.0
# Other dependencies # Other dependencies
iso639>=0.1.4 iso639>=0.1.4