Skip to main content
torch.js has not been released yet.
torch.js
Learn
Playground
Contact
Search...
⌘ K
Login
Back to Examples
Loading...
Back to Examples
mnist
Train a neural network to recognize handwritten digits
Start
Save Model
Load Model
MLP
109.39K params
Idle
Epoch 0/3 • Batch 0
-
loss
-
Training Progress
-
-
-
-
-
Step
Loss
-
Hyperparameters
Learning Rate
0.100
Batch Size
64
Epochs
3
Optimizer
SGD
Adam
28×28
Predicted
—
Confidence
-
Probabilities
0
1
2
3
4
5
6
7
8
9
Predicted:
0
Confidence:
0.0%
Logs
No output yet...
Analysis
Confusion Matrix
Accuracy:
-
0
1
2
3
4
5
6
7
8
9
0
-
-
-
-
-
-
-
-
-
-
1
-
-
-
-
-
-
-
-
-
-
2
-
-
-
-
-
-
-
-
-
-
3
-
-
-
-
-
-
-
-
-
-
4
-
-
-
-
-
-
-
-
-
-
5
-
-
-
-
-
-
-
-
-
-
6
-
-
-
-
-
-
-
-
-
-
7
-
-
-
-
-
-
-
-
-
-
8
-
-
-
-
-
-
-
-
-
-
9
-
-
-
-
-
-
-
-
-
-
Actual ↓ / Predicted →
Total: -
FC1 Weight Distribution
-
-
-
-
-
-
-
μ:
-
σ:
-
n:
-
Input
784
Linear
128
100.5K
ReLU
Linear
64
8.3K
ReLU
Linear
10
650
Output
10
input
output
conv
linear
activation
norm
pool
dropout