feat: CUDA 12.6

Dockerfile

@@ -1,4 +1,4 @@
-FROM python:3.12.3-slim
+FROM python:3.12.7-slim
 WORKDIR /app
 COPY . /app
 RUN pip install --upgrade pip

Dockerfile.cuda

@@ -1,8 +1,9 @@
-FROM nvidia/cuda:12.5.0-runtime-ubuntu22.04
+FROM nvidia/cuda:12.6.2-runtime-ubuntu24.04
 WORKDIR /app
 COPY . /app
 RUN apt-get update && \
-    apt-get install -y python3-pip && \
-    pip install --upgrade pip
-RUN pip install --no-cache-dir -r requirements.txt --extra-index-url https://download.pytorch.org/whl/cu121
-CMD ["fastapi", "run", "main.py", "--proxy-headers", "--host", "0.0.0.0", "--port", "8000"]
+    apt-get install -y python3-pip python3-venv && \
+    python3 -m venv /app/venv && \
+    /app/venv/bin/pip install --upgrade pip && \
+    /app/venv/bin/pip install --no-cache-dir -r requirements.txt --extra-index-url https://download.pytorch.org/whl/cu121
+CMD ["/app/venv/bin/python", "-m", "fastapi", "run", "main.py", "--proxy-headers", "--host", "0.0.0.0", "--port", "8000"]

docker-compose.dev.cuda.yml

@@ -0,0 +1,24 @@
+services:
+  cuda:
+    build:
+      context: .
+      dockerfile: ./Dockerfile.cuda
+    ports:
+      - "8000:8000"
+    volumes:
+      - ./models:/root/.cache/huggingface/hub:rw
+    environment:
+      - DEFAULT_MODEL_NAME
+      - BATCH_SIZE
+      - ACCESS_TOKEN
+      - DEFAULT_SCORE
+      - USE_API_KEYS
+      - API_KEYS
+    restart: unless-stopped
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: all
+              capabilities: [ gpu ]

docker-compose.dev.yml

@@ -0,0 +1,17 @@
+services:
+  cpu:
+    build:
+      context: .
+      dockerfile: ./Dockerfile
+    ports:
+      - "8000:8000"
+    volumes:
+      - ./models:/root/.cache/huggingface/hub:rw
+    environment:
+      - DEFAULT_MODEL_NAME
+      - BATCH_SIZE
+      - ACCESS_TOKEN
+      - DEFAULT_SCORE
+      - USE_API_KEYS
+      - API_KEYS
+    restart: unless-stopped

docker-compose.yml

@@ -1,10 +1,10 @@
 services:
-  image_video_classification:
+  cpu:
     image: ghcr.io/doppeltilde/image_video_classification:latest
     ports:
       - "8000:8000"
     volumes:
-      - models:/root/.cache/huggingface/hub:rw
+      - ./models:/root/.cache/huggingface/hub:rw
     environment:
       - DEFAULT_MODEL_NAME
       - BATCH_SIZE
@@ -14,12 +14,12 @@ services:
       - API_KEYS
     restart: unless-stopped
 
-  image_video_classification_cuda:
+  cuda:
     image: ghcr.io/doppeltilde/image_video_classification:latest-cuda
     ports:
       - "8000:8000"
     volumes:
-      - models:/root/.cache/huggingface/hub:rw
+      - ./models:/root/.cache/huggingface/hub:rw
     environment:
       - DEFAULT_MODEL_NAME
       - BATCH_SIZE
@@ -35,6 +35,3 @@ services:
             - driver: nvidia
               count: all
               capabilities: [ gpu ]
-
-volumes:
-  models:

requirements.txt

@@ -1,8 +1,8 @@
-fastapi[standard]==0.112.0
-Pillow==10.4.0
-transformers==4.44.0
-torch==2.4.0
+fastapi[standard]==0.115.3
+Pillow==11.0.0
+transformers==4.46.0
+torch==2.5.0
 python-dotenv==1.0.1
 filetype==1.2.0
 opencv-python-headless==4.10.0.84
-python-multipart==0.0.9
+python-multipart==0.0.12