Hands on Machine Learning¶
Course: Spring 2025 CS391L at UT Austin
Instructor: Prof. Inderjit Dhillon
Presented by: Nilesh Gupta (TA)
Life of an ML Experiment¶
Machine Learning (ML) experiments are far more than just training a model—they encompass an entire lifecycle that includes problem formulation, data gathering, preprocessing, model building, training, evaluation, and continuous iteration. In this notebook, we’ll walk through a simplified version of this process using a binary image classification task.
Key Stages in the ML Experiment Lifecycle:
Defining Problem¶
Understand the problem you’re trying to solve. Define clear objectives and metrics that will be used to assess model performance.
Defining Solution¶
Concretizing the data representation, model, loss, and optimization choices. This stage involves mapping the problem to a concrete ML setup.
Implementing Solution¶
Setting up the ML environment: Install the necessary libraries and set up the environment. This typically involves installing libraries, loading data, and preparing the notebook for experimentation.
Data Preparation: Load and preprocess data. Data preprocessing is a crucial step that involves normalizing, augmenting, and splitting the data into training and validation sets.
Setting up Model and Optimization Process: Design and implement a neural network architecture. Choose the right loss function, optimizer.
Training: Set up the training loop that includes forward propagation, loss computation, backpropagation, and optimization. Effective training involves iterative improvements and fine-tuning of hyperparameters.
Evaluation & Metrics: Assess the model using a validation/test set. Evaluating performance with metrics such as loss and accuracy provides insights into where improvements are needed.
Experiment Tracking & Iteration: Document your findings, track experiments, and iterate on your design. Recording the results is crucial for reproducibility and helps guide subsequent refinements.
Defining Problem¶
Demo problem: classify images of dogs vs cats¶
Key definitions of a problem¶
What is my input ($x$)?¶
- An image of a dog or a cat
What is my output ($y$)?¶
- A label indicating whether the image is a cat or a dog
How do I make a collection of such inputs and outputs a.k.a create my dataset ($\mathcal{D}$)?¶
- We can gather images of cats and dogs from various sources and manually label them (someone already did this for us - CIFAR10 dataset)
How do I measure success?¶
- Accuracy: the percentage of correctly classified images
Defining Solution¶
How do I represent my input ($x$) in numbers?¶
- Represent an image as a 3D tensor (channels, height, width) where each value represents the pixel intensity
How do I represent my output ($y$) in numbers?¶
- Represent the output as a single number (0 for cat, 1 for dog)
How do I map my input to my output a.k.a what is my model ($\mathcal{M}$)?¶
- Simple fully connected neural network with one hidden layer
How do I measure success in a differentiable way a.k.a what is my loss function ($\mathcal{L}$)?¶
- Binary cross-entropy loss
NOTE: usually we quantify the opposite of success i.e. the cost of making a mistake
How do I train my model to minimize the loss a.k.a what is my optimizer ($\mathcal{O}$)?¶
- Use stochastic gradient descent (SGD) to minimize the loss
Setting up the Environment¶
Torch Imports¶
torchandtorchvisionare the primary libraries used for building and training neural networks in PyTorch.torch.nncontains the building blocks for creating neural network models.torch.nn.functionalprovides functions that are used as building blocks in neural networks.torch.optimprovides implementations of various optimization algorithms.torch.utils.datasimplifies the data loading process for training and evaluation.torchvision.datasetsandtorchvision.transformsprovide access to popular datasets and data transformations for computer vision tasks.
In [1]:
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
from torch.utils.data import DataLoader, Subset
import matplotlib.pyplot as plt
WandB¶
We use Weights & Biases (wandb) to log metrics and visualize results. Wandb is a tool that helps track experiments and collaborate with team members. It logs metrics, hyperparameters, and artifacts, and provides a dashboard to visualize and compare results.
API description:
wandb.init(): Initializes a new wandb run.wandb.log(): Logs metrics to the wandb dashboard.wandb.watch(): Logs gradients and parameters of the model to the wandb dashboard.wandb.finish(): Marks the end of the wandb run.
In [2]:
# Import wandb and initialize
import wandb
# Initialize wandb with a project name and configuration
wandb.init(
project="hands_on_ml_demo",
config={
"learning_rate": 0.01,
"epochs": 10,
"batch_size": 32,
"hidden_activation": "relu",
"hidden_size": 128,
"dataset": "CIFAR-10 (cats (0) vs dogs (1))",
"model": "SimpleBinaryNet"
}
)
config = wandb.config
wandb: Using wandb-core as the SDK backend. Please refer to https://wandb.me/wandb-core for more information. wandb: Currently logged in as: nilesh2797 to https://api.wandb.ai. Use `wandb login --relogin` to force relogin
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_145758-rv4p69xu View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
In [3]:
print(*[f"{k}: {v}" for k, v in config.items()], sep="\n")
learning_rate: 0.01 epochs: 10 batch_size: 32 hidden_activation: relu hidden_size: 128 dataset: CIFAR-10 (cats (0) vs dogs (1)) model: SimpleBinaryNet
Data preparation¶
In [4]:
# Setup Data Transformation
transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])
# Setup Target Transformation (0 for cats, 1 for dogs)
target_transform = lambda x: 0 if x == 3 else 1
In [5]:
# Load and Filter CIFAR-10 Data for Binary Classification (Ships vs Planes)
train_dataset_full = datasets.CIFAR10(root='./data', train=True, transform=transform, target_transform=target_transform, download=True)
test_dataset_full = datasets.CIFAR10(root='./data', train=False, transform=transform, target_transform=target_transform, download=True)
train_indices = [i for i, target in enumerate(train_dataset_full.targets) if target in [3, 5]]
test_indices = [i for i, target in enumerate(test_dataset_full.targets) if target in [3, 5]]
train_dataset = Subset(train_dataset_full, train_indices)
test_dataset = Subset(test_dataset_full, test_indices)
# Create Validation Dataset
train_size = int(0.8 * len(train_dataset))
val_size = len(train_dataset) - train_size
train_dataset, val_dataset = torch.utils.data.random_split(train_dataset, [train_size, val_size])
In [6]:
from torchvision.utils import make_grid
# Function to visualize sample images from the dataset
def visualize_samples(dataset, num_samples=16):
# Get a subset of the dataset
indices = torch.randperm(len(dataset))[:num_samples]
samples = [dataset[i][0] for i in indices]
# Create a grid of images
grid = make_grid(samples, nrow=4, normalize=True)
# Plot the grid
plt.figure(figsize=(8, 8))
plt.imshow(grid.permute(1, 2, 0))
plt.title('Sample Images from the Dataset')
plt.axis('off')
plt.show()
# Visualize sample images from the training dataset
visualize_samples(train_dataset)
Dataloader¶
DataLoader is used to load data in batches and provides various utilities for shuffling, batching, and loading data in parallel. Important parameters:
dataset: The dataset to load.batch_size: Number of samples in each batch.shuffle: If True, shuffles the data before creating batches.num_workers: Number of processes to use for data loading.
In [7]:
# Data Loaders
train_loader = DataLoader(train_dataset, batch_size=config.batch_size, shuffle=True)
val_loader = DataLoader(val_dataset, batch_size=config.batch_size, shuffle=False)
test_loader = DataLoader(test_dataset, batch_size=config.batch_size, shuffle=False)
Setting up Model, Loss, and Optimizer¶
In [8]:
# Define Simple Neural Network
class SimpleBinaryNet(nn.Module):
def __init__(self, input_size=3*32*32, hidden_size=128, hidden_activation='sigmoid'):
super(SimpleBinaryNet, self).__init__()
self.w1 = nn.Parameter(torch.randn(input_size, hidden_size) / input_size**0.5) # Shape: (3*32*32) x hidden_size
self.b1 = nn.Parameter(torch.zeros(hidden_size)) # Shape: hidden_size
self.w2 = nn.Parameter(torch.randn(hidden_size, 1) / hidden_size**0.5) # Shape: hidden_size
self.b2 = nn.Parameter(torch.zeros(1)) # Shape: 1
self.hidden_activation_fn = F.sigmoid if hidden_activation == 'sigmoid' else F.relu if hidden_activation == 'relu' else ValueError("Unknown activation function")
def forward(self, x):
'''Forward pass of the network
Args:
x: Input tensor of shape (batch_size, 3*32*32)
Returns:
logits: Raw logits of shape (batch_size, 1)
'''
x = x.view(x.size(0), -1) # Flatten the image, shape: (batch_size, 3*32*32)
z1 = torch.matmul(x, self.w1) + self.b1 # Shape: (batch_size, hidden_size)
a1 = self.hidden_activation_fn(z1) # Shape: (batch_size, hidden_size)
z2 = torch.matmul(a1, self.w2) + self.b2 # Shape: (batch_size, 1)
return z2
class NNLinearBinaryNet(nn.Module):
def __init__(self, input_size=3*32*32, hidden_size=128, hidden_activation='sigmoid'):
super(NNLinearBinaryNet, self).__init__()
self.fc1 = nn.Linear(input_size, hidden_size)
self.fc2 = nn.Linear(hidden_size, hidden_size // 2)
self.fc3 = nn.Linear(hidden_size // 2, 1)
self.hidden_activation_fn = F.sigmoid if hidden_activation == 'sigmoid' else F.relu if hidden_activation == 'relu' else ValueError("Unknown activation function")
def forward(self, x):
'''Forward pass of the network
Args:
x: Input tensor of shape (batch_size, 3*32*32)
Returns:
logits: Raw logits of shape (batch_size, 1)
'''
x = x.view(x.size(0), -1) # Flatten the image, shape: (batch_size, 3*32*32)
x = self.hidden_activation_fn(self.fc1(x)) # Shape: (batch_size, hidden_size)
x = self.hidden_activation_fn(self.fc2(x))
x = self.fc3(x)
return x
if config.model == "SimpleBinaryNet":
model = SimpleBinaryNet(input_size=3*32*32, hidden_size=config.hidden_size, hidden_activation=config.hidden_activation)
elif config.model == "NNLinearNet":
model = NNLinearBinaryNet(input_size=3*32*32, hidden_size=config.hidden_size, hidden_activation=config.hidden_activation)
else:
raise ValueError("Unknown model type")
# Use wandb.watch to log gradients and model parameters
wandb.watch(model, log="all", log_freq=100, log_graph=True)
wandb: logging graph, to disable use `wandb.watch(log_graph=False)`
In [9]:
# Define Loss Function and Optimizer
criterion = nn.BCEWithLogitsLoss()
optimizer = optim.SGD(model.parameters(), lr=config.learning_rate)
In [10]:
sample_batch = next(iter(train_loader))
In [11]:
from torchviz import make_dot
pred = model(sample_batch[0])
loss = criterion(pred, sample_batch[1].float().view(-1, 1))
make_dot(loss, params=dict(model.named_parameters()), show_attrs=False, show_saved=False)
Out[11]:
Setting up Training and Evaluation Loops¶
In [12]:
from tqdm import tqdm
# Define Training and Evaluation Functions
def train(model, device, train_loader, optimizer, criterion, epoch):
model.train()
running_loss = 0.0
running_num_batches = 0
pbar = tqdm(enumerate(train_loader), total=len(train_loader), desc=f'Epoch {epoch}', dynamic_ncols=True)
for batch_idx, (data, target) in pbar:
data, target = data.to(device), target.to(device).float().unsqueeze(1)
optimizer.zero_grad()
output = model(data)
loss = criterion(output, target)
loss.backward()
optimizer.step()
running_loss += loss.item()
running_num_batches += 1
# Log loss per batch
wandb.log({"train_loss": loss.item(), "epoch": epoch, "batch": batch_idx})
pbar.set_postfix({'mean loss': running_loss / running_num_batches})
avg_loss = running_loss / len(train_loader)
print(f'Average training loss for epoch {epoch}: {avg_loss:.6f}')
return avg_loss
In [13]:
def evaluate(model, device, data_loader, criterion, epoch, split="val"):
model.eval()
test_loss = 0
correct = 0
sample_images = [] # To collect sample images and predictions
sample_preds = []
sample_targets = []
with torch.no_grad():
for batch_idx, (data, target) in enumerate(data_loader):
data, target = data.to(device), target.to(device).float().unsqueeze(1)
output = model(data)
test_loss += criterion(output, target).item()
preds = (torch.sigmoid(output) >= 0.5).to(target.dtype) # threshold at 0.5
correct += preds.eq(target.byte()).sum().item()
# Save first batch sample predictions for visualization
if batch_idx == 0:
sample_images.append(data.cpu())
sample_preds.append(preds.cpu())
sample_targets.append(target.cpu())
test_loss /= len(data_loader)
accuracy = 100. * correct / len(data_loader.dataset)
print(f'\n{split.capitalize()} set: Average loss: {test_loss:.4f}, Accuracy: {accuracy:.2f}%\n')
# Log evaluation metrics to wandb
wandb.log({f"{split}_loss": test_loss, f"{split}_accuracy": accuracy, "epoch": epoch})
# Log sample predictions (showing first batch of test set)
if sample_images:
# Log a list of sample images with predictions and targets as captions
wandb.log({
f"{split}_sample_predictions": [wandb.Image(img, caption=f"Pred: {'Dog' if pred.item() else 'Cat'}, Gold: {'Dog' if target.item() else 'Cat'}")
for img, pred, target in zip(sample_images[0], sample_preds[0], sample_targets[0])]
})
return test_loss, accuracy
Training the Model¶
In [14]:
# Training and Evaluation Loop with wandb tracking
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)
epochs = config.epochs
train_losses = []
test_accuracies = []
for epoch in range(1, epochs + 1):
train_loss = train(model, device, train_loader, optimizer, criterion, epoch)
val_loss, val_acc = evaluate(model, device, val_loader, criterion, epoch)
train_loss, train_acc = evaluate(model, device, train_loader, criterion, epoch, split="train")
wandb.log({"train_loss": train_loss, "train_accuracy": train_acc, "val_loss": val_loss, "val_accuracy": val_acc, "epoch": epoch})
Epoch 1: 100%|██████████| 250/250 [00:01<00:00, 243.80it/s, mean loss=0.664]
Average training loss for epoch 1: 0.663668 Val set: Average loss: 0.6542, Accuracy: 60.50% Train set: Average loss: 0.6381, Accuracy: 63.21%
Epoch 2: 100%|██████████| 250/250 [00:00<00:00, 287.84it/s, mean loss=0.638]
Average training loss for epoch 2: 0.637992 Val set: Average loss: 0.6510, Accuracy: 61.95% Train set: Average loss: 0.6206, Accuracy: 65.42%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 296.58it/s, mean loss=0.624]
Average training loss for epoch 3: 0.624370 Val set: Average loss: 0.6471, Accuracy: 62.50% Train set: Average loss: 0.6063, Accuracy: 66.96%
Epoch 4: 100%|██████████| 250/250 [00:00<00:00, 285.91it/s, mean loss=0.612]
Average training loss for epoch 4: 0.612083 Val set: Average loss: 0.6469, Accuracy: 62.25% Train set: Average loss: 0.5976, Accuracy: 68.08%
Epoch 5: 100%|██████████| 250/250 [00:01<00:00, 230.08it/s, mean loss=0.603]
Average training loss for epoch 5: 0.603175 Val set: Average loss: 0.6406, Accuracy: 62.80% Train set: Average loss: 0.5835, Accuracy: 69.89%
Epoch 6: 100%|██████████| 250/250 [00:00<00:00, 367.70it/s, mean loss=0.593]
Average training loss for epoch 6: 0.592636 Val set: Average loss: 0.6471, Accuracy: 62.85% Train set: Average loss: 0.5771, Accuracy: 70.15%
Epoch 7: 100%|██████████| 250/250 [00:00<00:00, 348.59it/s, mean loss=0.582]
Average training loss for epoch 7: 0.582014 Val set: Average loss: 0.6437, Accuracy: 62.95% Train set: Average loss: 0.5698, Accuracy: 71.16%
Epoch 8: 100%|██████████| 250/250 [00:00<00:00, 368.10it/s, mean loss=0.573]
Average training loss for epoch 8: 0.573100 Val set: Average loss: 0.6394, Accuracy: 63.50% Train set: Average loss: 0.5544, Accuracy: 73.21%
Epoch 9: 100%|██████████| 250/250 [00:00<00:00, 369.93it/s, mean loss=0.562]
Average training loss for epoch 9: 0.562172 Val set: Average loss: 0.6506, Accuracy: 62.95% Train set: Average loss: 0.5490, Accuracy: 73.10%
Epoch 10: 100%|██████████| 250/250 [00:00<00:00, 344.07it/s, mean loss=0.552]
Average training loss for epoch 10: 0.552380 Val set: Average loss: 0.6467, Accuracy: 63.25% Train set: Average loss: 0.5327, Accuracy: 74.17%
Evaluating the Model¶
In [15]:
test_loss, test_acc = evaluate(model, device, test_loader, criterion, epoch, split="test")
wandb.finish()
Test set: Average loss: 0.6390, Accuracy: 64.10%
View run chocolate-dust-55 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/fmc6di72
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 672 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 672 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_133509-fmc6di72/logs Automatic Hyperparameter Sweep with wandb¶
Hyperparameter sweeps allow you to automatically explore different configurations to optimize your model's performance. In this section, we set up a sweep to tune key hyperparameters such as learning rate, batch size, and the size of the hidden layer. We use a Bayesian optimization approach to maximize the test accuracy.
The sweep will run multiple experiments, and wandb will aggregate the results so that you can easily compare and select the best configuration.
In [20]:
# Define the sweep configuration
sweep_configuration = {
'method': 'grid', # Bayesian optimization
'metric': {
'name': 'val_accuracy',
'goal': 'maximize'
},
'parameters': {
'learning_rate': {'values': [0.001, 0.0001, 0.01]},
'batch_size': {'values': [32, 64, 128]},
'hidden_size': {'values': [64, 128, 256]},
'epochs': {'value': 3} # keep epochs small for quick sweeps
}
}
# Initialize the sweep
sweep_id = wandb.sweep(sweep=sweep_configuration, project="hands_on_ml_demo")
print("Sweep ID:", sweep_id)
Create sweep with ID: 1sn1nvxk Sweep URL: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk Sweep ID: 1sn1nvxk
In [21]:
def sweep_train():
"""Training function for hyperparameter sweep runs."""
# Initialize wandb for this sweep run
wandb.init()
config = wandb.config
# Set up data loaders using the hyperparameter batch size
batch_size = config.batch_size
train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)
val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False)
# Initialize the model using the hyperparameter hidden_size
model = SimpleBinaryNet(hidden_size=config.hidden_size)
model.to(device)
# Define loss function and optimizer with hyperparameter learning_rate
criterion = nn.BCEWithLogitsLoss()
optimizer = optim.SGD(model.parameters(), lr=config.learning_rate)
# Run training for the specified number of epochs
for epoch in range(1, config.epochs + 1):
train_loss = train(model, device, train_loader, optimizer, criterion, epoch)
val_loss, val_accuracy = evaluate(model, device, val_loader, criterion, epoch)
# Log metrics for each epoch
wandb.log({
"epoch_train_loss": train_loss,
"epoch_val_loss": val_loss,
"val_accuracy": val_accuracy,
"epoch": epoch
})
# Finish the run
wandb.finish()
In [22]:
# Launch the sweep agent to run a specified number of runs
wandb.agent(sweep_id, function=sweep_train, count=None)
wandb: Agent Starting Run: 560ab4u8 with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134105-560ab4u8 Syncing run sweet-sweep-1 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:00<00:00, 302.01it/s, mean loss=0.707]
Average training loss for epoch 1: 0.707201 Val set: Average loss: 0.6967, Accuracy: 49.25%
Epoch 2: 100%|██████████| 250/250 [00:00<00:00, 324.50it/s, mean loss=0.69]
Average training loss for epoch 2: 0.690431 Val set: Average loss: 0.6871, Accuracy: 54.60%
Epoch 3: 100%|██████████| 250/250 [00:01<00:00, 235.25it/s, mean loss=0.684]
Average training loss for epoch 3: 0.684309 Val set: Average loss: 0.6818, Accuracy: 57.10%
View run sweet-sweep-1 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/560ab4u8
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134105-560ab4u8/logs wandb: Agent Starting Run: ae5tv3br with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134115-ae5tv3br Syncing run kind-sweep-2 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:00<00:00, 294.21it/s, mean loss=0.7]
Average training loss for epoch 1: 0.700002 Val set: Average loss: 0.7019, Accuracy: 49.30%
Epoch 2: 100%|██████████| 250/250 [00:00<00:00, 336.43it/s, mean loss=0.697]
Average training loss for epoch 2: 0.697356 Val set: Average loss: 0.6991, Accuracy: 49.60%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 312.14it/s, mean loss=0.695]
Average training loss for epoch 3: 0.695154 Val set: Average loss: 0.6967, Accuracy: 50.15%
View run kind-sweep-2 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/ae5tv3br
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134115-ae5tv3br/logs wandb: Agent Starting Run: l2c9max2 with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134125-l2c9max2 Syncing run fresh-sweep-3 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:00<00:00, 337.08it/s, mean loss=0.681]
Average training loss for epoch 1: 0.680718 Val set: Average loss: 0.6706, Accuracy: 59.40%
Epoch 2: 100%|██████████| 250/250 [00:00<00:00, 302.42it/s, mean loss=0.671]
Average training loss for epoch 2: 0.671256 Val set: Average loss: 0.6662, Accuracy: 59.50%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 356.63it/s, mean loss=0.667]
Average training loss for epoch 3: 0.667198 Val set: Average loss: 0.6635, Accuracy: 60.10%
View run fresh-sweep-3 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/l2c9max2
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134125-l2c9max2/logs wandb: Agent Starting Run: v502tr5o with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134136-v502tr5o Syncing run drawn-sweep-4 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:00<00:00, 295.53it/s, mean loss=0.696]
Average training loss for epoch 1: 0.695815 Val set: Average loss: 0.6864, Accuracy: 57.10%
Epoch 2: 100%|██████████| 250/250 [00:01<00:00, 229.86it/s, mean loss=0.686]
Average training loss for epoch 2: 0.685905 Val set: Average loss: 0.6821, Accuracy: 58.45%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 310.32it/s, mean loss=0.682]
Average training loss for epoch 3: 0.682092 Val set: Average loss: 0.6789, Accuracy: 59.35%
View run drawn-sweep-4 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/v502tr5o
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134136-v502tr5o/logs wandb: Agent Starting Run: 25l6y941 with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134146-25l6y941 Syncing run eternal-sweep-5 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:00<00:00, 308.08it/s, mean loss=0.694]
Average training loss for epoch 1: 0.694144 Val set: Average loss: 0.6940, Accuracy: 50.95%
Epoch 2: 100%|██████████| 250/250 [00:00<00:00, 280.28it/s, mean loss=0.693]
Average training loss for epoch 2: 0.692519 Val set: Average loss: 0.6929, Accuracy: 52.10%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 332.71it/s, mean loss=0.691]
Average training loss for epoch 3: 0.691221 Val set: Average loss: 0.6919, Accuracy: 51.95%
View run eternal-sweep-5 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/25l6y941
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134146-25l6y941/logs wandb: Agent Starting Run: dj06tolz with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134156-dj06tolz Syncing run valiant-sweep-6 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:01<00:00, 220.67it/s, mean loss=0.68]
Average training loss for epoch 1: 0.680385 Val set: Average loss: 0.6710, Accuracy: 59.30%
Epoch 2: 100%|██████████| 250/250 [00:00<00:00, 255.06it/s, mean loss=0.671]
Average training loss for epoch 2: 0.670618 Val set: Average loss: 0.6655, Accuracy: 59.75%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 282.68it/s, mean loss=0.667]
Average training loss for epoch 3: 0.666644 Val set: Average loss: 0.6645, Accuracy: 60.60%
View run valiant-sweep-6 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/dj06tolz
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134156-dj06tolz/logs wandb: Agent Starting Run: yhu2aq0f with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134207-yhu2aq0f Syncing run genial-sweep-7 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:01<00:00, 214.38it/s, mean loss=0.704]
Average training loss for epoch 1: 0.703914 Val set: Average loss: 0.6900, Accuracy: 52.00%
Epoch 2: 100%|██████████| 250/250 [00:00<00:00, 269.49it/s, mean loss=0.687]
Average training loss for epoch 2: 0.687089 Val set: Average loss: 0.6835, Accuracy: 55.95%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 310.40it/s, mean loss=0.682]
Average training loss for epoch 3: 0.682307 Val set: Average loss: 0.6793, Accuracy: 57.55%
View run genial-sweep-7 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/yhu2aq0f
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134207-yhu2aq0f/logs wandb: Agent Starting Run: 2g8ygd84 with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134217-2g8ygd84 Syncing run dry-sweep-8 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:00<00:00, 267.85it/s, mean loss=0.706]
Average training loss for epoch 1: 0.705731 Val set: Average loss: 0.7043, Accuracy: 49.20%
Epoch 2: 100%|██████████| 250/250 [00:00<00:00, 303.09it/s, mean loss=0.697]
Average training loss for epoch 2: 0.697079 Val set: Average loss: 0.6973, Accuracy: 49.10%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 299.53it/s, mean loss=0.693]
Average training loss for epoch 3: 0.692983 Val set: Average loss: 0.6936, Accuracy: 51.10%
View run dry-sweep-8 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/2g8ygd84
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134217-2g8ygd84/logs wandb: Agent Starting Run: noft849u with config: wandb: batch_size: 32 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134227-noft849u Syncing run denim-sweep-9 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 250/250 [00:00<00:00, 291.07it/s, mean loss=0.68]
Average training loss for epoch 1: 0.679796 Val set: Average loss: 0.6701, Accuracy: 59.10%
Epoch 2: 100%|██████████| 250/250 [00:01<00:00, 212.66it/s, mean loss=0.669]
Average training loss for epoch 2: 0.668984 Val set: Average loss: 0.6670, Accuracy: 59.15%
Epoch 3: 100%|██████████| 250/250 [00:00<00:00, 295.56it/s, mean loss=0.666]
Average training loss for epoch 3: 0.666017 Val set: Average loss: 0.6623, Accuracy: 60.65%
View run denim-sweep-9 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/noft849u
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 96 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134227-noft849u/logs wandb: Agent Starting Run: 8xilgzeo with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134238-8xilgzeo Syncing run different-sweep-10 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 148.54it/s, mean loss=0.755]
Average training loss for epoch 1: 0.754559 Val set: Average loss: 0.7158, Accuracy: 50.85%
Epoch 2: 100%|██████████| 125/125 [00:00<00:00, 180.99it/s, mean loss=0.71]
Average training loss for epoch 2: 0.709623 Val set: Average loss: 0.6937, Accuracy: 51.40%
Epoch 3: 100%|██████████| 125/125 [00:00<00:00, 159.29it/s, mean loss=0.694]
Average training loss for epoch 3: 0.693613 Val set: Average loss: 0.6858, Accuracy: 53.50%
View run different-sweep-10 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/8xilgzeo
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134238-8xilgzeo/logs wandb: Agent Starting Run: q9gjw0yy with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134248-q9gjw0yy Syncing run dry-sweep-11 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 211.51it/s, mean loss=0.751]
Average training loss for epoch 1: 0.751483 Val set: Average loss: 0.7382, Accuracy: 50.85%
Epoch 2: 100%|██████████| 125/125 [00:00<00:00, 226.99it/s, mean loss=0.745]
Average training loss for epoch 2: 0.744822 Val set: Average loss: 0.7323, Accuracy: 50.85%
Epoch 3: 100%|██████████| 125/125 [00:00<00:00, 212.04it/s, mean loss=0.739]
Average training loss for epoch 3: 0.738824 Val set: Average loss: 0.7271, Accuracy: 50.85%
View run dry-sweep-11 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/q9gjw0yy
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134248-q9gjw0yy/logs wandb: Agent Starting Run: 8qs4z78v with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134253-8qs4z78v Syncing run cosmic-sweep-12 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 164.58it/s, mean loss=0.69]
Average training loss for epoch 1: 0.689676 Val set: Average loss: 0.6750, Accuracy: 57.70%
Epoch 2: 100%|██████████| 125/125 [00:00<00:00, 204.79it/s, mean loss=0.674]
Average training loss for epoch 2: 0.673630 Val set: Average loss: 0.6710, Accuracy: 58.75%
Epoch 3: 100%|██████████| 125/125 [00:01<00:00, 120.30it/s, mean loss=0.67]
Average training loss for epoch 3: 0.670178 Val set: Average loss: 0.6666, Accuracy: 58.85%
View run cosmic-sweep-12 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/8qs4z78v
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134253-8qs4z78v/logs wandb: Agent Starting Run: e94baf7i with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134304-e94baf7i Syncing run likely-sweep-13 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 164.77it/s, mean loss=0.728]
Average training loss for epoch 1: 0.727522 Val set: Average loss: 0.7044, Accuracy: 49.55%
Epoch 2: 100%|██████████| 125/125 [00:00<00:00, 162.12it/s, mean loss=0.692]
Average training loss for epoch 2: 0.692493 Val set: Average loss: 0.6897, Accuracy: 52.10%
Epoch 3: 100%|██████████| 125/125 [00:00<00:00, 183.99it/s, mean loss=0.686]
Average training loss for epoch 3: 0.686100 Val set: Average loss: 0.6852, Accuracy: 54.80%
View run likely-sweep-13 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/e94baf7i
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134304-e94baf7i/logs wandb: Agent Starting Run: 1zhy8bue with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134314-1zhy8bue Syncing run rosy-sweep-14 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 170.18it/s, mean loss=0.766]
Average training loss for epoch 1: 0.766400 Val set: Average loss: 0.7680, Accuracy: 49.15%
Epoch 2: 100%|██████████| 125/125 [00:00<00:00, 170.85it/s, mean loss=0.752]
Average training loss for epoch 2: 0.751676 Val set: Average loss: 0.7535, Accuracy: 49.15%
Epoch 3: 100%|██████████| 125/125 [00:00<00:00, 213.51it/s, mean loss=0.74]
Average training loss for epoch 3: 0.739530 Val set: Average loss: 0.7415, Accuracy: 49.15%
View run rosy-sweep-14 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/1zhy8bue
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134314-1zhy8bue/logs wandb: Sweep Agent: Waiting for job. wandb: Job received. wandb: Agent Starting Run: vbbsyxnn with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134329-vbbsyxnn Syncing run stellar-sweep-15 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 159.82it/s, mean loss=0.687]
Average training loss for epoch 1: 0.687135 Val set: Average loss: 0.6741, Accuracy: 59.40%
Epoch 2: 100%|██████████| 125/125 [00:00<00:00, 200.16it/s, mean loss=0.673]
Average training loss for epoch 2: 0.673439 Val set: Average loss: 0.6690, Accuracy: 59.75%
Epoch 3: 100%|██████████| 125/125 [00:00<00:00, 187.83it/s, mean loss=0.67]
Average training loss for epoch 3: 0.669716 Val set: Average loss: 0.6666, Accuracy: 59.90%
View run stellar-sweep-15 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/vbbsyxnn
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134329-vbbsyxnn/logs wandb: Agent Starting Run: g527a0ji with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134335-g527a0ji Syncing run restful-sweep-16 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 177.77it/s, mean loss=0.708]
Average training loss for epoch 1: 0.708367 Val set: Average loss: 0.6948, Accuracy: 51.15%
Epoch 2: 100%|██████████| 125/125 [00:00<00:00, 161.74it/s, mean loss=0.693]
Average training loss for epoch 2: 0.692513 Val set: Average loss: 0.6902, Accuracy: 54.05%
Epoch 3: 100%|██████████| 125/125 [00:00<00:00, 162.21it/s, mean loss=0.689]
Average training loss for epoch 3: 0.688524 Val set: Average loss: 0.6867, Accuracy: 55.85%
View run restful-sweep-16 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/g527a0ji
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134335-g527a0ji/logs wandb: Sweep Agent: Waiting for job. wandb: Job received. wandb: Agent Starting Run: 2bre31fc with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134350-2bre31fc Syncing run jolly-sweep-17 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 170.30it/s, mean loss=0.696]
Average training loss for epoch 1: 0.696188 Val set: Average loss: 0.6986, Accuracy: 49.20%
Epoch 2: 100%|██████████| 125/125 [00:01<00:00, 99.87it/s, mean loss=0.695]
Average training loss for epoch 2: 0.695279 Val set: Average loss: 0.6975, Accuracy: 49.85%
Epoch 3: 100%|██████████| 125/125 [00:00<00:00, 165.05it/s, mean loss=0.695]
Average training loss for epoch 3: 0.694525 Val set: Average loss: 0.6966, Accuracy: 50.15%
View run jolly-sweep-17 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/2bre31fc
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134350-2bre31fc/logs wandb: Agent Starting Run: 43u8ob9f with config: wandb: batch_size: 64 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134401-43u8ob9f Syncing run ethereal-sweep-18 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 125/125 [00:00<00:00, 180.70it/s, mean loss=0.684]
Average training loss for epoch 1: 0.684060 Val set: Average loss: 0.6749, Accuracy: 59.05%
Epoch 2: 100%|██████████| 125/125 [00:00<00:00, 166.60it/s, mean loss=0.674]
Average training loss for epoch 2: 0.674402 Val set: Average loss: 0.6721, Accuracy: 58.95%
Epoch 3: 100%|██████████| 125/125 [00:00<00:00, 190.73it/s, mean loss=0.67]
Average training loss for epoch 3: 0.670307 Val set: Average loss: 0.6664, Accuracy: 59.95%
View run ethereal-sweep-18 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/43u8ob9f
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 192 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134401-43u8ob9f/logs wandb: Sweep Agent: Waiting for job. wandb: Job received. wandb: Agent Starting Run: 5b3dgkfh with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134416-5b3dgkfh Syncing run autumn-sweep-19 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 103.76it/s, mean loss=0.781] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.780697 Val set: Average loss: 0.7490, Accuracy: 50.85%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 128.28it/s, mean loss=0.74] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 2: 0.740295 Val set: Average loss: 0.7209, Accuracy: 50.85%
Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 92.86it/s, mean loss=0.717] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 3: 0.717317 Val set: Average loss: 0.7052, Accuracy: 50.95%
View run autumn-sweep-19 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/5b3dgkfh
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134416-5b3dgkfh/logs wandb: Agent Starting Run: rojkcdam with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134421-rojkcdam Syncing run clean-sweep-20 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 91.44it/s, mean loss=0.711] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.710936 Val set: Average loss: 0.7153, Accuracy: 49.25%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 118.12it/s, mean loss=0.709] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 2: 0.709387 Val set: Average loss: 0.7139, Accuracy: 49.30%
Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 127.82it/s, mean loss=0.709] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 3: 0.708667 Val set: Average loss: 0.7126, Accuracy: 49.35%
View run clean-sweep-20 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/rojkcdam
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134421-rojkcdam/logs wandb: Agent Starting Run: 7ojbpeoh with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 64 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134427-7ojbpeoh Syncing run smooth-sweep-21 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 91.10it/s, mean loss=0.687] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.686800 Val set: Average loss: 0.6829, Accuracy: 57.05%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 115.70it/s, mean loss=0.679]
Average training loss for epoch 2: 0.679163 Val set: Average loss: 0.6779, Accuracy: 58.45%
WARNING:root:Only 108 Image will be uploaded. Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 125.05it/s, mean loss=0.675] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 3: 0.674996 Val set: Average loss: 0.6745, Accuracy: 59.50%
View run smooth-sweep-21 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/7ojbpeoh
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134427-7ojbpeoh/logs wandb: Sweep Agent: Waiting for job. wandb: Job received. wandb: Agent Starting Run: 2witricv with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134442-2witricv Syncing run lively-sweep-22 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 64.98it/s, mean loss=0.692] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.692417 Val set: Average loss: 0.6930, Accuracy: 51.35%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 97.58it/s, mean loss=0.691] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 2: 0.690614 Val set: Average loss: 0.6912, Accuracy: 52.65%
Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 107.85it/s, mean loss=0.689]
Average training loss for epoch 3: 0.689081 Val set: Average loss: 0.6897, Accuracy: 52.85%
WARNING:root:Only 108 Image will be uploaded.
View run lively-sweep-22 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/2witricv
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134442-2witricv/logs wandb: Sweep Agent: Waiting for job. wandb: Job received. wandb: Agent Starting Run: 9gell266 with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134457-9gell266 Syncing run denim-sweep-23 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 115.67it/s, mean loss=0.739] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.739491 Val set: Average loss: 0.7273, Accuracy: 50.85%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 113.91it/s, mean loss=0.734]
Average training loss for epoch 2: 0.733902 Val set: Average loss: 0.7225, Accuracy: 50.85%
WARNING:root:Only 108 Image will be uploaded. Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 120.72it/s, mean loss=0.729] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 3: 0.729304 Val set: Average loss: 0.7181, Accuracy: 50.85%
View run denim-sweep-23 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/9gell266
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134457-9gell266/logs wandb: Agent Starting Run: 2tdo4gzd with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 128 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134503-2tdo4gzd Syncing run light-sweep-24 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 102.15it/s, mean loss=0.687] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.687019 Val set: Average loss: 0.6806, Accuracy: 58.15%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 111.31it/s, mean loss=0.678] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 2: 0.678201 Val set: Average loss: 0.6754, Accuracy: 59.35%
Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 111.20it/s, mean loss=0.674] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 3: 0.674119 Val set: Average loss: 0.6728, Accuracy: 59.35%
View run light-sweep-24 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/2tdo4gzd
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134503-2tdo4gzd/logs wandb: Sweep Agent: Waiting for job. wandb: Job received. wandb: Agent Starting Run: v1oz8xzc with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134518-v1oz8xzc Syncing run fresh-sweep-25 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 101.50it/s, mean loss=0.697] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.696679 Val set: Average loss: 0.6913, Accuracy: 52.70%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 98.77it/s, mean loss=0.69] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 2: 0.689895 Val set: Average loss: 0.6891, Accuracy: 55.15%
Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 96.86it/s, mean loss=0.688]
Average training loss for epoch 3: 0.687731 Val set: Average loss: 0.6874, Accuracy: 56.05%
WARNING:root:Only 108 Image will be uploaded.
View run fresh-sweep-25 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/v1oz8xzc
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134518-v1oz8xzc/logs wandb: Agent Starting Run: yfirxtau with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.0001
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134523-yfirxtau Syncing run bright-sweep-26 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 106.18it/s, mean loss=0.719] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.718968 Val set: Average loss: 0.7123, Accuracy: 50.85%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 95.19it/s, mean loss=0.714] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 2: 0.713748 Val set: Average loss: 0.7082, Accuracy: 50.90%
Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 112.09it/s, mean loss=0.709] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 3: 0.709374 Val set: Average loss: 0.7050, Accuracy: 50.95%
View run bright-sweep-26 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/yfirxtau
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134523-yfirxtau/logs wandb: Sweep Agent: Waiting for job. wandb: Job received. wandb: Agent Starting Run: q9fjdtk2 with config: wandb: batch_size: 128 wandb: epochs: 3 wandb: hidden_size: 256 wandb: learning_rate: 0.01
Tracking run with wandb version 0.19.7
Run data is saved locally in
/Users/nileshgupta/Downloads/work/Spring25/CS391L/HandsOnML/wandb/run-20250226_134538-q9fjdtk2 Syncing run scarlet-sweep-27 to Weights & Biases (docs)
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
Sweep page: https://wandb.ai/nilesh2797/hands_on_ml_demo/sweeps/1sn1nvxk
View project at https://wandb.ai/nilesh2797/hands_on_ml_demo
Epoch 1: 100%|██████████| 63/63 [00:00<00:00, 105.06it/s, mean loss=0.692] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 1: 0.692448 Val set: Average loss: 0.6810, Accuracy: 57.90%
Epoch 2: 100%|██████████| 63/63 [00:00<00:00, 112.83it/s, mean loss=0.679] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 2: 0.678632 Val set: Average loss: 0.6747, Accuracy: 59.35%
Epoch 3: 100%|██████████| 63/63 [00:00<00:00, 101.84it/s, mean loss=0.674] WARNING:root:Only 108 Image will be uploaded.
Average training loss for epoch 3: 0.674382 Val set: Average loss: 0.6720, Accuracy: 59.20%
View run scarlet-sweep-27 at: https://wandb.ai/nilesh2797/hands_on_ml_demo/runs/q9fjdtk2
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
View project at: https://wandb.ai/nilesh2797/hands_on_ml_demo
Synced 5 W&B file(s), 324 media file(s), 0 artifact file(s) and 0 other file(s)
Find logs at:
./wandb/run-20250226_134538-q9fjdtk2/logs wandb: Sweep Agent: Waiting for job. wandb: Sweep Agent: Exiting.
Debugging Your Neural Network¶
Training neural networks can be challenging. Here are a couple of strategies for debugging your model:
1. Monitor Gradient Norms¶
- Purpose: Ensure gradients are neither vanishing nor exploding.
- How: Compute the norm of gradients for each parameter after backpropagation and log them.
In [350]:
def get_gradient_norms(model):
norms = {}
for name, param in model.named_parameters():
if param.grad is not None:
norms[name] = param.grad.norm().item()
return norms
# Example: Log gradient norms during training
grad_norms = get_gradient_norms(model)
wandb.log({"gradient_norms": grad_norms, "epoch": epoch})
--------------------------------------------------------------------------- Error Traceback (most recent call last) Cell In[350], line 10 8 # Example: Log gradient norms during training 9 grad_norms = get_gradient_norms(model) ---> 10 wandb.log({"gradient_norms": grad_norms, "epoch": epoch}) File ~/miniconda3/lib/python3.11/site-packages/wandb/sdk/lib/preinit.py:36, in PreInitCallable.<locals>.preinit_wrapper(*args, **kwargs) 35 def preinit_wrapper(*args: Any, **kwargs: Any) -> Any: ---> 36 raise wandb.Error(f"You must call wandb.init() before {name}()") Error: You must call wandb.init() before wandb.log()
2. Inspect Activation Ranges¶
- Purpose: Verify that layer activations are within a sensible range (avoid saturation or dead neurons).
- How: Use forward hooks to log statistics (min, max, mean) of activations.
In [ ]:
def activation_hook(module, input, output):
wandb.log({
f"{module.__class__.__name__}_activation_min": output.min().item(),
f"{module.__class__.__name__}_activation_max": output.max().item(),
f"{module.__class__.__name__}_activation_mean": output.mean().item(),
})
# Register the hook for the first fully connected layer
model.w1.register_forward_hook(activation_hook)