MLOps Assignment - Observations & Artifact Management

This report contains observations from hyperparameter tuning, model performance analysis, and the importance of artifact management for reproducibility and version control.

Anish Iyer (G24AIT135)

Created on February 14|Last edited on April 4

Comment

﻿
﻿
val_loss v. learning_rate
val_loss v. learning_rate
0.020.040.060.080.10learning_rate0.511.52val_loss
val_accuracy v. batch_size
val_accuracy v. batch_size
354045505560batch_size0.20.40.60.8val_accuracy
val_accuracy v. learning_rate
val_accuracy v. learning_rate
0.020.040.060.080.10learning_rate0.20.40.60.8val_accuracy
Parameter importance with respect toepoch
1-5
 of 5
Config parameter
Importance
Correlation
Runtime
Runtime
hidden_layer_neurons
hidden_layer_neurons
learning_rate
learning_rate
batch_size
batch_size
epochs
epochs
Loading...
batch_size
batch_size
generous-disco-67gentle-sweep-2azure-sweep-1serene-bush-63final-trained-modelfallen-sweep-4silvery-sweep-3earthy-sweep-2fresh-sweep-1usual-sweep-30204060
epoch
epoch
01234567Step12345
final-trained-model
fallen-sweep-4
silvery-sweep-3
earthy-sweep-2
fresh-sweep-1
fine-sweep-3
cherubic-chocolate-37
smitten-date-32
Run set47
﻿
﻿

Add a comment