Update README.md

README.md

@@ -168,4 +168,4 @@ The app looks like below in the emulator. It displays the highest confident resu
 [Mobile_App.mp4](https://user-images.githubusercontent.com/89344987/210223484-44402679-ed00-46f1-bd63-ace5d8d21cf4.mp4)
 
 ### Week 13 (26<sup>th</sup> December to 2<sup>nd</sup> December)
-Now that the mobile app is finalized I tried to convert the vgg_16 model I trained earlier to a `.tflite` model. This caused some unexpected issues. When I run the code snippet which converts the model to a `.tflite` model the colab runs out of RAM. I was using the free version of colab which gives 12GB memory. I looked for solutions in online forums and tried limiting the memory growth, converting using a saved model, converting after freeing up the memory by deleting some data and reducing the batch size as mentioned in the given forums. [solution 01](https://github.com/tensorflow/models/issues/1817), [solution 02](https://github.com/tensorflow/tensorflow/issues/40760). But none of them seem to work. Also I was not able to find a project done by converting a vgg_16 model to a `.tflite` model. And it was mentioned in the [official documentation of TensorFlow lite](https://www.tensorflow.org/lite/guide/ops_compatibility) that certain types of models cannot be converted to `.tflite` format. So I decided that this model may not be usable in my project. Then I tried to train and convert the [InceptionV3](https://keras.io/api/applications/inceptionv3/) model which also couldn't be converted. Then I ran across some online tutorials to convert a [MobileNetV3](https://paperswithcode.com/method/mobilenetv3) model and use in edge devices. So I started using that model and was able to successfully convert it to a `.tflite` and implement it inside my mobile app. The accuracy value is still pretty low, so the rest of the will be spent to fine tune the accuracy.   
+Now that the mobile app is finalized I tried to convert the vgg_16 model I trained earlier to a `.tflite` model. This caused some unexpected issues. When I run the code snippet which converts the model to a `.tflite` model the colab runs out of RAM. I was using the free version of colab which gives 12GB memory. I looked for solutions in online forums and tried limiting the memory growth, converting using a saved model, converting after freeing up the memory by deleting some data and reducing the batch size as mentioned in the given forums. [solution 01](https://github.com/tensorflow/models/issues/1817), [solution 02](https://github.com/tensorflow/tensorflow/issues/40760). But none of them seem to work. Also I was not able to find a project done by converting a vgg_16 model to a `.tflite` model. And it was mentioned in the [official documentation of TensorFlow lite](https://www.tensorflow.org/lite/guide/ops_compatibility) that certain types of models cannot be converted to `.tflite` format. So I decided that this model may not be usable in my project. Then I tried to train and convert the [InceptionV3](https://keras.io/api/applications/inceptionv3/) model which also couldn't be converted. Then I ran across some online tutorials to convert a [MobileNetV3](https://paperswithcode.com/method/mobilenetv3) model and use in edge devices. So I started using that model and was able to successfully convert it to a `.tflite` and implement it inside my mobile app. The accuracy value is still pretty low, so the rest of the weeks will be spent to fine tune the accuracy.