File size: 2,279 Bytes
c700bd5 755842e ab4421a c700bd5 6458940 c700bd5 6458940 c700bd5 ad94d02 7b879b9 ad94d02 c700bd5 7b879b9 c8f8ff0 b87e018 35b038c c700bd5 5d1b13c b87e018 0543208 b87e018 c6d4d5b 8fc2ed8 0274aed 75ad871 5d1b13c 35b038c 5d1b13c 35b038c af95e31 c700bd5 4448c19 ec3b313 c9fb95e b972116 8fc2ed8 4448c19 8fc2ed8 edbb35f 488a20c 84bffde 56c49c7 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 |
# PyTorch + CUDA 12.1 + cuDNN 8 (passt zur L4)
FROM pytorch/pytorch:2.2.2-cuda12.1-cudnn8-runtime
# -- System-Pakete minimal, aber was wir wirklich brauchen --
RUN apt-get update && \
apt-get install -y --no-install-recommends \
git-lfs build-essential portaudio19-dev ffmpeg && \
rm -rf /var/lib/apt/lists/*
# Non-root-User (Spaces-Empfehlung)
RUN useradd -m -u 1000 user
USER user
WORKDIR /app
ENV PATH="/home/user/.local/bin:$PATH"
# Orpheus-/SNAC-Code + Server
COPY --chown=user . /app
ENV HF_HOME=/app/.cache
ENV VLLM_USE_LM_FORMAT_ENFORCER=0
# GPU-freundliches Torch-Upgrade (falls gewünscht)
#RUN pip install --no-cache-dir \
# torch==2.3.1+cu121 torchaudio==2.3.1+cu121 \
# --index-url https://download.pytorch.org/whl/cu121
RUN pip install --no-cache-dir "transformers==4.40.2" "lm-format-enforcer==0.9.8"
RUN pip install --no-cache-dir vllm>=0.9.0
# Python-Abhängigkeiten
COPY --chown=user requirements.txt .
RUN pip install --upgrade pip && \
pip install --no-cache-dir -r requirements.txt
RUN pip install --no-cache-dir "realtimetts[system]>=0.5.5"
RUN pip install --no-cache-dir flashinfer-cu121-preview || echo "FlashInfer not available – continuing without."
# nur *diese* Engine-Pflicht nachliefern, aber ohne Resolver:
# RUN pip install --no-cache-dir pyttsx3==2.90 --no-deps
# optional, um Warn-Spam zu reduzieren
# RUN pip install --no-cache-dir azure-cognitiveservices-speech==1.33.0 --no-deps \
# tqdm==4.66.1 --no-deps
EXPOSE 7860
# ───── Environment ───────────────────────────────────────
ENV ORPHEUS_MODEL=SebastianBodza/Kartoffel_Orpheus-3B_german_synthetic-v0.1
ENV MODEL_ID="SebastianBodza/Kartoffel_Orpheus-3B_german_synthetic-v0.1"
ENV ORPHEUS_API_URL=http://127.0.0.1:1234
# ───── Entrypoint ────────────────────────────────────────
CMD bash -c "\
python -m vllm.entrypoints.openai.api_server \
--model ${MODEL_ID} \
--port 1234 \
--dtype bfloat16 \
--gpu-memory-utilization 0.85 \
--max-model-len 8192 & \
uvicorn app:app --host 0.0.0.0 --port 7860"
|