Upload run_server.sh with huggingface_hub

run_server.sh

@@ -0,0 +1,82 @@
+#!/bin/bash
+
+# Detoxify-Small - GGUF Model Server Launcher
+# Generated on 2025-09-17 20:07:11
+
+# Configuration
+MODEL_FILE="model.gguf"
+MODEL_NAME="Detoxify-Small"
+HOST="127.0.0.1"
+PORT="8000"
+GPU_LAYERS="0"
+
+# Colors for output
+RED=''
+GREEN=''
+YELLOW=''
+BLUE=''
+NC='' # No Color
+
+echo -e "========================================"
+echo -e "  Detoxify-Small Server"
+echo -e "========================================"
+echo
+
+# Check if model file exists
+if [ ! -f "$MODEL_FILE" ]; then
+    echo -e "Error: Model file '$MODEL_FILE' not found!"
+    echo -e "Make sure you're running this script from the model directory."
+    exit 1
+fi
+
+echo -e "✓ Model file found: $MODEL_FILE"
+echo -e "✓ Model: $MODEL_NAME"
+echo -e "✓ Server: http://$HOST:$PORT"
+echo
+
+# Check if port is available
+if lsof -Pi :$PORT -sTCP:LISTEN -t >/dev/null ; then
+    echo -e "Warning: Port $PORT is already in use."
+    echo -e "Trying alternative port 8001..."
+    PORT="8001"
+fi
+
+# Detect system capabilities
+if [[ "$OSTYPE" == "darwin"* ]]; then
+    echo -e "✓ macOS detected - Metal GPU acceleration available"
+    GPU_LAYERS="35"  # Reasonable default for M1/M2/M3
+elif command -v nvidia-smi &> /dev/null; then
+    echo -e "✓ NVIDIA GPU detected - CUDA acceleration available"
+    GPU_LAYERS="35"  # Reasonable default for CUDA
+else
+    echo -e "No GPU acceleration detected - using CPU only"
+    GPU_LAYERS="0"
+fi
+
+echo -e "✓ GPU layers: $GPU_LAYERS"
+echo
+
+# Check for llama-server
+if ! command -v llama-server &> /dev/null; then
+    echo -e "Error: llama-server not found in PATH!"
+    echo -e "Please install llama.cpp and add it to your PATH:"
+    echo -e "  https://github.com/ggerganov/llama.cpp"
+    exit 1
+fi
+
+echo -e "✓ llama-server found"
+echo
+
+echo -e "Starting server..."
+echo -e "Press Ctrl+C to stop"
+echo
+
+# Start the server
+exec llama-server \
+    -m "$MODEL_FILE" \
+    --host "$HOST" \
+    --port "$PORT" \
+    --n-gpu-layers "$GPU_LAYERS" \
+    --ctx-size 4096 \
+    --chat-template "" \
+    --log-disable