compile time tests script added

compile_expts.sh

@@ -0,0 +1,20 @@
+# TinyLlama
+touch TinyLlama-1.1B-Chat-v1.0.txt
+for i in {1..10}; do
+    echo "TinyLlama $i"
+    python compile_test.py ../genAI/models/TinyLlama-1.1B-Chat-v1.0
+done
+
+# Llama-2-7b
+touch Llama-2-7b-hf.txt
+for i in {1..10}; do
+    echo "Llama-2-7b $i"
+    python compile_test.py ../genAI/models/Llama-2-7b-hf
+done
+
+# Llama-3-8B
+touch Meta-Llama-3-8B-Instruct.txt
+for i in {1..10}; do
+    echo "Llama-3-8B $i"
+    python compile_test.py ../genAI/models/Meta-Llama-3-8B-Instruct
+done

compile_test.py

@@ -0,0 +1,34 @@
+import numpy as np
+import openvino as ov
+import openvino.properties.hint as hints
+import openvino.properties as props
+import time
+import sys
+import os
+
+from transformers import AutoTokenizer
+from tqdm import tqdm
+
+
+def main(model_dir):
+    core = ov.Core()
+    core.set_property("CPU", {hints.execution_mode: hints.ExecutionMode.PERFORMANCE})
+
+    config = {
+        props.inference_num_threads: 64,
+        hints.inference_precision: "f16", 
+    }
+
+    start = time.time()
+    model = core.compile_model(model_dir + "/openvino_model.xml", "CPU", config)
+    end = time.time()
+
+    model_name = os.path.basename(model_dir)
+    with open(f"{model_name}.txt", "a") as f:
+        f.writelines(f"{end - start},")
+
+    print(f"{model_name}: {end - start} sec")
+
+
+if __name__ == "__main__":
+    main(sys.argv[1])