Upload runpod_setup.sh with huggingface_hub

runpod_setup.sh

@@ -1,27 +1,26 @@
 #!/bin/bash
-# RunPod vLLM Setup Script for Ministral 14B
+# RunPod SGLang Setup Script for Ministral 14B
 # Usage: curl -sSL https://huggingface.co/datasets/RoleModel/scripts/raw/main/runpod_setup.sh | bash
 
 set -e
 
-echo "=== Stopping any existing vLLM processes ==="
-pkill -f "vllm.entrypoints" 2>/dev/null || true
+echo "=== Stopping any existing processes ==="
+pkill -f "sglang" 2>/dev/null || true
+pkill -f "vllm" 2>/dev/null || true
 sleep 2
 
-echo "=== Installing dependencies ==="
-pip install -q "numpy<2.3" --quiet
-pip install -q vllm "transformers>=5.0.0rc0" "mistral-common>=1.8.6" --quiet
+echo "=== Installing SGLang ==="
+pip install -q "sglang[all]" "transformers>=5.0.0rc0" "mistral-common>=1.8.6" --quiet
 
-echo "=== Starting vLLM server ==="
+echo "=== Starting SGLang server ==="
 echo "Model: RoleModel/ministral-14b-merged-official"
 echo "Port: 8000"
 echo ""
 
-python -m vllm.entrypoints.openai.api_server \
+python -m sglang.launch_server \
   --model RoleModel/ministral-14b-merged-official \
-  --tokenizer mistralai/Ministral-3-14B-Base-2512 \
   --dtype bfloat16 \
-  --max-model-len 8192 \
+  --context-length 8192 \
   --port 8000 \
   --trust-remote-code \
   --served-model-name ministral-14b