Spaces:
Runtime error
Runtime error
Commit ·
7329fdd
1
Parent(s): 727bef6
asfasd
Browse files- Dockerfile +14 -19
Dockerfile
CHANGED
|
@@ -8,27 +8,24 @@ ENV HOME="/home/user" \
|
|
| 8 |
DOCLING_ARTIFACTS_PATH="$HOME/.cache/docling/models" \
|
| 9 |
OMP_NUM_THREADS=2
|
| 10 |
|
| 11 |
-
# 2) Create
|
| 12 |
RUN useradd -m -u 1000 user
|
| 13 |
|
| 14 |
-
# 3)
|
| 15 |
RUN apt-get update && apt-get install -y \
|
| 16 |
build-essential git libpoppler-cpp-dev poppler-utils \
|
| 17 |
libmagic-dev python3-dev \
|
| 18 |
&& rm -rf /var/lib/apt/lists/*
|
| 19 |
|
| 20 |
-
# 4)
|
| 21 |
WORKDIR $HOME/app
|
| 22 |
-
COPY --chown=user:
|
| 23 |
-
|
| 24 |
-
# 5) Install Python libs as root (to leverage caching)
|
| 25 |
RUN pip install --no-cache-dir \
|
| 26 |
torch torchvision --extra-index-url https://download.pytorch.org/whl/cpu \
|
| 27 |
-
huggingface_hub \
|
| 28 |
-
|
| 29 |
-
&& pip install --no-cache-dir -r requirements.txt
|
| 30 |
|
| 31 |
-
#
|
| 32 |
RUN mkdir -p \
|
| 33 |
$HOME/.cache/docling/models \
|
| 34 |
$HF_HUB_CACHE \
|
|
@@ -38,24 +35,22 @@ RUN mkdir -p \
|
|
| 38 |
$HOME/.cache/huggingface \
|
| 39 |
$HOME/app/model_cache
|
| 40 |
|
| 41 |
-
#
|
| 42 |
USER user
|
| 43 |
RUN pip install --no-cache-dir docling \
|
| 44 |
&& docling-tools models download
|
| 45 |
|
| 46 |
-
#
|
| 47 |
RUN python - <<EOF
|
| 48 |
from huggingface_hub import snapshot_download
|
|
|
|
| 49 |
snapshot_download(
|
| 50 |
repo_id="numind/NuExtract-1.5-tiny",
|
| 51 |
local_dir="$HOME/app/model_cache",
|
| 52 |
-
cache_dir="
|
| 53 |
)
|
| 54 |
EOF
|
| 55 |
|
| 56 |
-
#
|
| 57 |
-
COPY --chown=user: . .
|
| 58 |
-
|
| 59 |
-
# 10) Expose and run
|
| 60 |
-
ENV PORT=7860
|
| 61 |
-
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]
|
|
|
|
| 8 |
DOCLING_ARTIFACTS_PATH="$HOME/.cache/docling/models" \
|
| 9 |
OMP_NUM_THREADS=2
|
| 10 |
|
| 11 |
+
# 2) Create non-root user
|
| 12 |
RUN useradd -m -u 1000 user
|
| 13 |
|
| 14 |
+
# 3) System deps
|
| 15 |
RUN apt-get update && apt-get install -y \
|
| 16 |
build-essential git libpoppler-cpp-dev poppler-utils \
|
| 17 |
libmagic-dev python3-dev \
|
| 18 |
&& rm -rf /var/lib/apt/lists/*
|
| 19 |
|
| 20 |
+
# 4) Copy & install Python deps
|
| 21 |
WORKDIR $HOME/app
|
| 22 |
+
COPY --chown=user:user requirements.txt ./
|
|
|
|
|
|
|
| 23 |
RUN pip install --no-cache-dir \
|
| 24 |
torch torchvision --extra-index-url https://download.pytorch.org/whl/cpu \
|
| 25 |
+
huggingface_hub accelerate \
|
| 26 |
+
&& pip install --no-cache-dir -r requirements.txt
|
|
|
|
| 27 |
|
| 28 |
+
# 5) Prepare caches and ownership
|
| 29 |
RUN mkdir -p \
|
| 30 |
$HOME/.cache/docling/models \
|
| 31 |
$HF_HUB_CACHE \
|
|
|
|
| 35 |
$HOME/.cache/huggingface \
|
| 36 |
$HOME/app/model_cache
|
| 37 |
|
| 38 |
+
# 6) Switch to non-root and prefetch Docling models
|
| 39 |
USER user
|
| 40 |
RUN pip install --no-cache-dir docling \
|
| 41 |
&& docling-tools models download
|
| 42 |
|
| 43 |
+
# 7) Prefetch the NuExtract model
|
| 44 |
RUN python - <<EOF
|
| 45 |
from huggingface_hub import snapshot_download
|
| 46 |
+
import os
|
| 47 |
snapshot_download(
|
| 48 |
repo_id="numind/NuExtract-1.5-tiny",
|
| 49 |
local_dir="$HOME/app/model_cache",
|
| 50 |
+
cache_dir=os.getenv("HF_HUB_CACHE")
|
| 51 |
)
|
| 52 |
EOF
|
| 53 |
|
| 54 |
+
# 8) Copy application code and start
|
| 55 |
+
COPY --chown=user:user . .
|
| 56 |
+
CMD ["uvicorn","app:app","--host","0.0.0.0","--port","7860"]
|
|
|
|
|
|
|
|
|