Update entrypoint.sh

entrypoint.sh

@@ -17,7 +17,7 @@ done
 
 # Start text-generation-inference server in the background
 echo "Starting text-generation-inference server..."
-text-generation-launcher --model-id ${MODEL_NAME} --num-shard 1 --port 8080 --disable-custom-kernels &
+text-generation-launcher --model-id ${MODEL_NAME} --num-shard 1 --port 8080 --disable-custom-kernels --max-batch-prefill-tokens 4096 &
 
 # Wait for text-generation-inference to start
 until check_port 8080; do