Skip to content

Commit 18bb5bb

Browse files
authoredDec 2, 2024··
Update README.md
1 parent 9109367 commit 18bb5bb

File tree

1 file changed

+5
-5
lines changed

1 file changed

+5
-5
lines changed
 

‎README.md

+5-5
Original file line numberDiff line numberDiff line change
@@ -17,7 +17,7 @@ This project implements a Convolutional Neural Network (CNN) for MNIST digit cla
1717
- Parameter count verification (< 20,000 parameters)
1818
- Input shape validation (28x28)
1919
- Output shape validation (10 classes)
20-
- Accuracy threshold (> 95% on test set)
20+
- Accuracy threshold (> 99.3% on test set)
2121
- Automatic model versioning with timestamp and accuracy
2222
- CPU-only training support for both local and CI environments
2323

@@ -77,7 +77,7 @@ The pipeline is triggered on every push to the repository.
7777
## Model Training Details
7878
- Dataset: MNIST (60,000 training images, 10,000 test images)
7979
- Batch Size: 128
80-
- Optimizer: Adam (lr=0.01)
80+
- Optimizer: SGD (lr=0.01)
8181
- Loss Function: Negative Log Likelihood
8282
- Training: Single epoch with progress bar showing loss and accuracy
8383
- Model Saving: Automatic with timestamp and accuracy in filename
@@ -93,13 +93,13 @@ The pipeline is triggered on every push to the repository.
9393
## Testing
9494
The test suite (`test_model.py`) verifies:
9595
- Model architecture compatibility with 28x28 input images
96-
- Total parameter count (< 25,000)
97-
- Model accuracy on test set (> 95%)
96+
- Total parameter count (< 20,000)
97+
- Model accuracy on test set (> 99.3%)
9898
- Output shape (10 classes)
9999

100100
## Model Artifacts
101101
Trained models are saved in the `models/` directory with the name mnist_model_YYYYMMDD_HHMMSS_accXX.X.pth
102102
where:
103103
- YYYYMMDD: Date
104104
- HHMMSS: Time
105-
- XX.X: Achieved accuracy
105+
- XX.X: Achieved accuracy

0 commit comments

Comments
 (0)
Please sign in to comment.