updated Readme and logs md

Readme.md

@@ -0,0 +1,48 @@
+# ResNet50 Model Implementation
+
+This implementation provides a customizable ResNet50 model for image classification tasks.
+
+## Model Architecture
+
+The model uses the ResNet50 architecture, which is a deep convolutional neural network with 50 layers. Key features include:
+
+- Based on the standard ResNet50 architecture
+- Customizable number of output classes
+- Modified final fully connected layer to match the desired number of classes
+- Initialized from scratch (no pre-training)
+
+## Functions
+
+### `get_model(num_classes)`
+Initializes a new ResNet50 model with a custom number of output classes.
+- **Input**: Number of classes (integer)
+- **Output**: Initialized ResNet50 model
+- **Note**: The model is initialized without pre-trained weights
+
+### `save_model(model, path)`
+Saves the model's state dictionary to a specified path.
+- **Input**: 
+  - model: Trained PyTorch model
+  - path: File path for saving the model
+
+### `load_model(num_classes, path)`
+Loads a previously saved model.
+- **Input**:
+  - num_classes: Number of output classes
+  - path: Path to the saved model file
+- **Output**: Loaded ResNet50 model
+
+## Usage Example
+
+```python
+model = get_model(num_classes=1000)
+save_model(model, 'model.pth')
+loaded_model = load_model(num_classes=1000, path='model.pth')
+```
+
+#Training Logs
+![Trained on AWS](https://github.com/srtangirala/resnet-training/blob/main/logs/images/aws.png)
+![Logs 1](https://github.com/srtangirala/resnet-training/blob/main/logs/images/logs-1.png)
+![Logs 2](https://github.com/srtangirala/resnet-training/blob/main/logs/images/logs-2.png)
+
+

logs/training_log_20250104_071323.md

@@ -1,56 +1,26 @@
 # Training Log
 
-| Epoch | Train Loss | Train Acc | Test Acc | Best Acc |
-|-------|------------|-----------|-----------|----------|
-|     1 | 2.101 | 21.95% | 29.35% | 0.00% |
-**New best accuracy achieved at epoch 1**
-
-|     2 | 1.875 | 30.13% | 34.70% | 29.35% |
-**New best accuracy achieved at epoch 2**
-
-|     3 | 1.778 | 34.90% | 34.60% | 34.70% |
-|     4 | 1.714 | 36.99% | 40.80% | 34.70% |
-**New best accuracy achieved at epoch 4**
-
-|     5 | 1.646 | 40.05% | 40.15% | 40.80% |
-|     6 | 1.587 | 42.39% | 43.00% | 40.80% |
-**New best accuracy achieved at epoch 6**
-
-|     7 | 1.513 | 45.59% | 47.30% | 43.00% |
-**New best accuracy achieved at epoch 7**
-
-|     8 | 1.436 | 48.44% | 51.85% | 47.30% |
-**New best accuracy achieved at epoch 8**
-
-|     9 | 1.347 | 51.07% | 53.50% | 51.85% |
-**New best accuracy achieved at epoch 9**
-
-|    10 | 1.294 | 53.89% | 55.95% | 53.50% |
-**New best accuracy achieved at epoch 10**
-
-|    11 | 1.220 | 56.07% | 55.75% | 55.95% |
-|    12 | 1.173 | 58.45% | 56.25% | 55.95% |
-**New best accuracy achieved at epoch 12**
-
-|    13 | 1.125 | 60.20% | 59.00% | 56.25% |
-**New best accuracy achieved at epoch 13**
-
-|    14 | 1.042 | 63.76% | 60.40% | 59.00% |
-**New best accuracy achieved at epoch 14**
-
-|    15 | 0.997 | 65.18% | 62.35% | 60.40% |
-**New best accuracy achieved at epoch 15**
-
-|    16 | 0.955 | 66.08% | 66.00% | 62.35% |
-**New best accuracy achieved at epoch 16**
-
-|    17 | 0.908 | 68.75% | 68.30% | 66.00% |
-**New best accuracy achieved at epoch 17**
-
-|    18 | 0.881 | 69.69% | 67.45% | 68.30% |
-|    19 | 0.844 | 70.84% | 73.30% | 68.30% |
-**New best accuracy achieved at epoch 19**
-
+| Epoch | Train Loss | Train Acc | Test Acc | Best Acc | Best Epoch |
+|-------|------------|-----------|-----------|----------|-------------|
+|     1 | 2.101 | 21.95% | 29.35% | 0.00% | 1 |
+|     2 | 1.875 | 30.13% | 34.70% | 29.35% | 2 |    
+|     3 | 1.778 | 34.90% | 34.60% | 34.70% | 3 |
+|     4 | 1.714 | 36.99% | 40.80% | 34.70% | 4 |
+|     5 | 1.646 | 40.05% | 40.15% | 40.80% | 5 |
+|     6 | 1.587 | 42.39% | 43.00% | 40.80% | 6 |
+|     7 | 1.513 | 45.59% | 47.30% | 43.00% | 7 |
+|     8 | 1.436 | 48.44% | 51.85% | 47.30% | 8 |
+|     9 | 1.347 | 51.07% | 53.50% | 51.85% | 9 |
+|    10 | 1.294 | 53.89% | 55.95% | 53.50% | 10 |
+|    11 | 1.220 | 56.07% | 55.75% | 55.95% | 11 |
+|    12 | 1.173 | 58.45% | 56.25% | 55.95% | 12 |
+|    13 | 1.125 | 60.20% | 59.00% | 56.25% | 13 |
+|    14 | 1.042 | 63.76% | 60.40% | 59.00% | 14 |
+|    15 | 0.997 | 65.18% | 62.35% | 60.40% | 15 |
+|    16 | 0.955 | 66.08% | 66.00% | 62.35% | 16 |
+|    17 | 0.908 | 68.75% | 68.30% | 66.00% | 17 |
+|    18 | 0.881 | 69.69% | 67.45% | 68.30% | 18 |
+|    19 | 0.844 | 70.84% | 73.30% | 68.30% | 19 |
 
 **Training stopped at epoch 19 after reaching target accuracy of 70%**