[CLEANUP]

servers/cogvlm/Dockerfile

@@ -7,7 +7,6 @@ ENV ARTIFACTS_PATH=/app/artifacts
 ENV STORAGE_PATH=/app/storage
 ENV HF_HUB_ENABLE_HF_TRANSFER=True
 
-
 # Set the working directory to the root
 WORKDIR /
 
@@ -26,20 +25,14 @@ COPY requirements.txt .
 # Install Python dependencies from requirements.txt
 RUN python3.10 -m pip install -r requirements.txt
 
-# Adjust the working directory to where your application's code will reside
-WORKDIR /swarms-cloud/servers/cogvlm
-
 # Copy the application's entire directory structure into the container
 COPY . /swarms-cloud
 
-# Copy the entrypoint script into the container
-COPY scripts/entry_point.sh /entry_point.sh
-
-# Make the entrypoint script executable
-RUN chmod +x /entry_point.sh
+# Set the working directory to the cogvlm directory
+WORKDIR /swarms-cloud/servers/cogvlm
 
 # Expose the port the app runs on
 EXPOSE 8000
 
-# Use the entrypoint script to configure environment variables and start the application
-ENTRYPOINT ["/entry_point.sh"]
+# Run the cogvlm.py file as a Uvicorn server on port 8000
+CMD ["uvicorn", "cogvlm:app", "--host", "0.0.0.0", "--port", "8000"]

servers/vllm_llm/config.yaml

@@ -0,0 +1,24 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: ttv-deployment
+spec:
+  replicas: 3
+  selector:
+    matchLabels:
+      app: ttv
+  template:
+    metadata:
+      labels:
+        app: ttv
+    spec:
+      containers:
+      - name: ttv
+        image: ttv
+        resources:
+          limits:
+            nvidia.com/gpu: 1
+  autoscale:
+    minReplicas: 1
+    maxReplicas: 10
+    targetCPUUtilizationPercentage: 80

servers/vllm_llm/sky_serve.yaml

@@ -112,13 +112,7 @@ resources:
 # workdir: ~/swarms-cloud/servers/cogvlm
 
 setup: |
-  conda create -n vllm python=3.9 -y
-  conda activate vllm
-  pip install vllm
+  docker build -t ttv .
 
 run: |
-  conda activate vllm
-  python -m vllm.entrypoints.openai.api_server \
-    --tensor-parallel-size $SKYPILOT_NUM_GPUS_PER_NODE \
-    --host 0.0.0.0 --port 8080 \
-    --model mistralai/Mixtral-8x7B-Instruct-v0.1
+  docker run ttv