Changes fix model transfer bugs

onnx2tflite.py

@@ -1,31 +1,13 @@
-import onnx
+from ultralytics import YOLO
-import tensorflow as tf
-from onnx_tf.backend import prepare
 
+# Load the YOLO11 model
+model = YOLO("runs/detect/train/weights/best.pt")
 
-def onnx_to_tflite(onnx_model_path, tflite_model_path):
+# Export the model to TFLite format
-    # 加载 ONNX 模型
+model.export(format="tflite")  # creates 'yolo11n_float32.tflite'
-    onnx_model = onnx.load(onnx_model_path)
 
-    # 使用 onnx-tf 将 ONNX 模型转换为 TensorFlow 模型
+# Load the exported TFLite model
-    tf_rep = prepare(onnx_model)
+tflite_model = YOLO("runs/detect/train/weights/best.tflite")
 
-    # 保存 TensorFlow 模型
+# Run inference
-    saved_model_dir = 'runs/detect/train/weights/best.pt'
+results = tflite_model("datasets/valid/images/2007_000042_jpg.rf.dadefef10ec4b9422a1689f7ce40588e.jpg")
-    tf_rep.export_graph(saved_model_dir)
-
-    # 加载 TensorFlow 模型并转换为 TFLite 模型
-    converter = tf.lite.TFLiteConverter.from_saved_model(saved_model_dir)
-    tflite_model = converter.convert()
-
-    # 保存 TFLite 模型
-    with open(tflite_model_path, 'wb') as f:
-        f.write(tflite_model)
-    print(f"模型已成功转换并保存为 {tflite_model_path}")
-
-
-# 使用示例
-onnx_model_path = 'model.onnx'  # 替换为你的 ONNX 模型文件路径
-tflite_model_path = 'model.tflite'  # 替换为你希望保存的 TFLite 模型路径
-
-onnx_to_tflite(onnx_model_path, tflite_model_path)

test_tf.py

@@ -0,0 +1,3 @@
+import tensorflow as tf
+print("TF Version:", tf.__version__)
+print("Is GPU available:", tf.test.is_gpu_available())