Redesign Docker build system to single-stage architecture for reliable model loading

Replaced complex two-stage build (builder → runtime) with single-stage
Dockerfile that builds and runs Ollama in one image. This fixes model
loading issues caused by missing CUDA libraries and LD_LIBRARY_PATH
mismatches in the previous multi-stage design.

Changes:
- Add docker/Dockerfile: Single-stage build with GCC 10, CMake 4, Go 1.25.3, CUDA 11.4
- Clone source from https://github.com/dogkeeper886/ollama37
- Compile Ollama with "CUDA 11" preset for Tesla K80 (compute capability 3.7)
- Keep complete CUDA toolkit and all libraries in final image (~20GB)
- Update docker-compose.yml: Simplified config, use ollama37:latest image
- Update docker/README.md: New build instructions and architecture docs

Trade-off: Larger image size (~20GB vs ~3GB) for guaranteed compatibility
and reliable GPU backend operation. All libraries remain accessible with
correct paths, ensuring models load properly on Tesla K80.

Tested: Successfully runs gemma3:1b on Tesla K80

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

docker/docker-compose.yml

@@ -2,10 +2,9 @@ version: "3.8"
 
 services:
   ollama:
-    image: ollama37-runtime:latest
-    container_name: ollama37-runtime
+    image: ollama37:latest
+    container_name: ollama37
     runtime: nvidia
-    user: "${UID:-1000}:${GID:-1000}"
     deploy:
       resources:
         reservations:
@@ -16,9 +15,8 @@ services:
     ports:
       - "11434:11434"
     volumes:
-      - ${HOME}/.ollama:${HOME}/.ollama
+      - ollama-data:/root/.ollama
     environment:
-      - HOME=${HOME}
       - OLLAMA_HOST=0.0.0.0:11434
       - NVIDIA_VISIBLE_DEVICES=all
       - NVIDIA_DRIVER_CAPABILITIES=compute,utility
@@ -29,6 +27,7 @@ services:
       timeout: 10s
       retries: 3
       start_period: 5s
+
 volumes:
   ollama-data:
     name: ollama-data