Update README.md

README.md

@@ -1,13 +1,118 @@
----
-title: YoloV3 PASCAL VOC
-emoji: 🏃
-colorFrom: indigo
-colorTo: pink
-sdk: gradio
-sdk_version: 3.39.0
-app_file: app.py
-pinned: false
-license: mit
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# [GithubREPO](https://github.com/deepanshudashora/ERAV1/tree/master/session13)
+
+# Training Procedure
+
+1. The model is trained on Tesla T4 (15GB GPU memory)
+2. The training is completed in two phases
+3. The first phase contains 20 epochs and second phase contains another 20 epochs
+4. In the first training we see loss dropping correctly but in the second training it drops less
+5. We run our two training loops separately and do not run any kind of validation on them, except for validation loss
+6. Later we evaluate the model and get the numbers
+7. The lightning generally saves the model as .ckpt format, so we convert it to torch format by saving state dict as .pt format
+8. For doing this we use these two lines of code
+
+```
+  best_model = torch.load(weights_path)
+  torch.save(best_model['state_dict'], f'best_model.pth')
+  litemodel = YOLOv3(num_classes=num_classes)
+  litemodel.load_state_dict(torch.load("best_model.pth",map_location='cpu'))
+  device = "cpu"
+  torch.save(litemodel.state_dict(), PATH)
+```
+   
+
+8. The model starts overfitting on the dataset after 30 epochs
+9. Future Improvements
+     1. Train the model in 1 shot instead of two different phases
+     2. Keep a better batch size (Basically earn more money and buy a good GPU)
+     3. Data transformation also plays a vital role here
+     4. OneCycle LR range needs to be appropriately modified for a better LR
+
+# Data Transformation
+
+Along with the transforms mentioned in the [config file](https://github.com/deepanshudashora/ERAV1/blob/master/session13/lightning_version/config.py), we also apply **mosaic transform** on 75% images 
+
+[Reference](https://www.kaggle.com/code/nvnnghia/awesome-augmentation/notebook)
+
+# Accuracy Report
+
+```
+Class accuracy is: 82.999725%
+No obj accuracy is: 96.828300%
+Obj accuracy is: 76.898473%
+
+MAP: 0.29939851760864258
+
+```
+
+# [Training Logs](https://github.com/deepanshudashora/ERAV1/blob/master/session13/lightning_version/merged_logs.csv)
+
+#### For faster execution we run the validation step after 20 epochs for the first 20 epochs of training and after that after every 5 epochs till 40 epochs
+
+```
+      Unnamed: 0   lr-Adam    step  train_loss  epoch  val_loss
+8150        8150       NaN  164299    4.186745   39.0       NaN
+8151        8151  0.000132  164349         NaN    NaN       NaN
+8152        8152       NaN  164349    2.936086   39.0       NaN
+8153        8153  0.000132  164399         NaN    NaN       NaN
+8154        8154       NaN  164399    4.777130   39.0       NaN
+8155        8155  0.000132  164449         NaN    NaN       NaN
+8156        8156       NaN  164449    3.139145   39.0       NaN
+8157        8157  0.000132  164499         NaN    NaN       NaN
+8158        8158       NaN  164499    4.596097   39.0       NaN
+8159        8159  0.000132  164549         NaN    NaN       NaN
+8160        8160       NaN  164549    5.587294   39.0       NaN
+8161        8161  0.000132  164599         NaN    NaN       NaN
+8162        8162       NaN  164599    4.592830   39.0       NaN
+8163        8163  0.000132  164649         NaN    NaN       NaN
+8164        8164       NaN  164649    3.914468   39.0       NaN
+8165        8165  0.000132  164699         NaN    NaN       NaN
+8166        8166       NaN  164699    3.180615   39.0       NaN
+8167        8167  0.000132  164749         NaN    NaN       NaN
+8168        8168       NaN  164749    5.772174   39.0       NaN
+8169        8169  0.000132  164799         NaN    NaN       NaN
+8170        8170       NaN  164799    2.894014   39.0       NaN
+8171        8171  0.000132  164849         NaN    NaN       NaN
+8172        8172       NaN  164849    4.473828   39.0       NaN
+8173        8173  0.000132  164899         NaN    NaN       NaN
+8174        8174       NaN  164899    6.397766   39.0       NaN
+8175        8175  0.000132  164949         NaN    NaN       NaN
+8176        8176       NaN  164949    3.789242   39.0       NaN
+8177        8177  0.000132  164999         NaN    NaN       NaN
+8178        8178       NaN  164999    5.182691   39.0       NaN
+8179        8179  0.000132  165049         NaN    NaN       NaN
+8180        8180       NaN  165049    4.845749   39.0       NaN
+8181        8181  0.000132  165099         NaN    NaN       NaN
+8182        8182       NaN  165099    3.672542   39.0       NaN
+8183        8183  0.000132  165149         NaN    NaN       NaN
+8184        8184       NaN  165149    4.230726   39.0       NaN
+8185        8185  0.000132  165199         NaN    NaN       NaN
+8186        8186       NaN  165199    4.625024   39.0       NaN
+8187        8187  0.000132  165249         NaN    NaN       NaN
+8188        8188       NaN  165249    6.549682   39.0       NaN
+8189        8189  0.000132  165299         NaN    NaN       NaN
+8190        8190       NaN  165299    5.040627   39.0       NaN
+8191        8191  0.000132  165349         NaN    NaN       NaN
+8192        8192       NaN  165349    5.857126   39.0       NaN
+8193        8193  0.000132  165399         NaN    NaN       NaN
+8194        8194       NaN  165399    3.081895   39.0       NaN
+8195        8195  0.000132  165449         NaN    NaN       NaN
+8196        8196       NaN  165449    3.945353   39.0       NaN
+8197        8197  0.000132  165499         NaN    NaN       NaN
+8198        8198       NaN  165499    3.513420   39.0       NaN
+8199        8199       NaN  165519         NaN   39.0  6.084875
+
+
+```
+
+# Results
+
+## For epochs 0 to 19
+![accuracy_per_class.png](images/train_logs_1.png.png)
+
+## From 19 to 20
+![accuracy_per_class.png](images/train_logs_2.png)
+
+## Full training logs for loss
+
+![accuracy_per_class.png](images/full_training.png)