DistributedDataParallel in PyTorch - Model Metrics & Evaluation

Metrics & Evaluation - DistributedDataParallel

Which metric matters for DistributedDataParallel and WHY

When using DistributedDataParallel (DDP), the main goal is to train a model faster by splitting data across multiple devices. The key metrics to watch are training loss and validation accuracy. These show if the model learns well and generalizes correctly. Also, speedup and scalability metrics matter to check if adding more devices really helps training time without hurting model quality.

Confusion matrix or equivalent visualization

DistributedDataParallel itself does not change the confusion matrix of the model predictions. However, here is an example confusion matrix from a classification model trained with DDP:

      | Predicted Positive | Predicted Negative |
      |--------------------|--------------------|
      | True Positive (TP) = 90  | False Negative (FN) = 10 |
      | False Positive (FP) = 5  | True Negative (TN) = 95  |
    

Totals: TP + FP + TN + FN = 90 + 5 + 95 + 10 = 200 samples.

Precision = 90 / (90 + 5) = 0.947

Recall = 90 / (90 + 10) = 0.9

F1 Score = 2 * (0.947 * 0.9) / (0.947 + 0.9) ≈ 0.923

Precision vs Recall tradeoff with concrete examples

In DDP training, the model's precision and recall depend on the data and model, not DDP itself. But DDP helps train faster and on more data, which can improve both.

Example 1: Spam filter - High precision is important to avoid marking good emails as spam. DDP can help train a better model faster.

Example 2: Medical diagnosis - High recall is critical to catch all disease cases. DDP allows training on large datasets to improve recall.

DDP helps by enabling faster training on bigger data, which can improve the balance between precision and recall.

What "good" vs "bad" metric values look like for DistributedDataParallel use

Good metrics after training with DDP:

• Training loss steadily decreases and matches single-device training loss.
• Validation accuracy is similar or better than single-device training.
• Speedup close to the number of devices used (e.g., 4 GPUs -> ~4x faster).
• Consistent precision and recall values without degradation.

Bad metrics:

• Training loss does not decrease or is unstable compared to single-device.
• Validation accuracy drops significantly, indicating poor model quality.
• Speedup is very low, showing overhead or bottlenecks.
• Precision or recall drops, possibly due to synchronization issues.

Common pitfalls in metrics when using DistributedDataParallel

• Data leakage: If data is not properly split across devices, the model may see the same samples multiple times, inflating accuracy.
• Overfitting: Faster training can cause overfitting if early stopping or validation checks are ignored.
• Synchronization issues: Incorrect gradient averaging can cause training instability and poor metrics.
• Accuracy paradox: High accuracy may hide poor recall or precision on important classes.
• Unequal batch sizes: If batches differ across devices, metrics may be skewed.

Self-check question

Your model trained with DistributedDataParallel has 98% accuracy but only 12% recall on the fraud class. Is it good for production? Why or why not?

Answer: No, it is not good. Although accuracy is high, the very low recall means the model misses most fraud cases. For fraud detection, recall is critical because missing fraud is costly. The model needs improvement to catch more fraud cases before production use.

Key Result

DistributedDataParallel aims for similar model quality metrics as single-device training while achieving near-linear speedup.

Practice

1. What is the main purpose of using DistributedDataParallel in PyTorch?

easy

A. To automatically tune hyperparameters during training

B. To reduce the size of the model for deployment

C. To convert a model to run only on CPU

D. To speed up training by using multiple GPUs or machines simultaneously

Solution

1. Step 1: Understand the role of DistributedDataParallel

   DistributedDataParallel (DDP) is designed to run training across multiple GPUs or machines to speed up the process.

2. Step 2: Compare options with DDP's purpose

   Options B, C, and D do not relate to parallel training or speed-up. Only To speed up training by using multiple GPUs or machines simultaneously correctly describes DDP's main purpose.

3. Final Answer:

   To speed up training by using multiple GPUs or machines simultaneously -> Option D

4. Quick Check:

   DDP speeds training = A [OK]

Hint: DDP means training faster with many GPUs [OK]

Common Mistakes:

• Thinking DDP reduces model size
• Confusing DDP with CPU-only training
• Assuming DDP tunes hyperparameters automatically

2. Which of the following is the correct way to wrap a PyTorch model with DistributedDataParallel assuming model is your neural network and device is the GPU device?

easy

A. model = torch.nn.DataParallel(model, device_ids=[device])

B. model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[device])

C. model = torch.nn.DistributedDataParallel(model, device=device)

D. model = torch.nn.parallel.DistributedDataParallel(model, devices=device)

Solution

1. Step 1: Recall the correct syntax for DistributedDataParallel

   The correct class is torch.nn.parallel.DistributedDataParallel and it takes device_ids as a list of devices.

2. Step 2: Check each option for syntax correctness

   model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[device]) uses the correct class and parameter device_ids=[device]. model = torch.nn.DataParallel(model, device_ids=[device]) uses DataParallel, which is different. Options C and D use incorrect parameter names or class names.

3. Final Answer:

   model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[device]) -> Option B

4. Quick Check:

   Correct class and device_ids list = A [OK]

Hint: Use torch.nn.parallel.DistributedDataParallel with device_ids list [OK]

Common Mistakes:

• Using DataParallel instead of DistributedDataParallel
• Passing device instead of device_ids list
• Wrong class or parameter names

3. Consider the following code snippet for a distributed training setup:

import torch
import torch.distributed as dist
from torch.nn.parallel import DistributedDataParallel as DDP

def setup(rank, world_size):
    dist.init_process_group('gloo', rank=rank, world_size=world_size)

model = torch.nn.Linear(10, 1).to(rank)
model = DDP(model, device_ids=[rank])

print(model.module.weight.device)

What will be the output device of model.module.weight.device if rank=1 and GPUs are available?

medium

A. cuda:1

B. cpu

C. cuda:0

D. cuda:rank

Solution

1. Step 1: Understand device assignment in the code

   The model is moved to rank device with .to(rank), so for rank=1, it is on cuda:1.

2. Step 2: Check DistributedDataParallel wrapping

   DDP wraps the model and uses device_ids=[rank], so the model stays on cuda:1.

3. Final Answer:

   cuda:1 -> Option A

4. Quick Check:

   Model device matches rank GPU = B [OK]

Hint: Model device matches rank GPU in DDP setup [OK]

Common Mistakes:

• Assuming model stays on CPU
• Confusing device index with rank variable
• Thinking device is always cuda:0

4. You run a distributed training script using DistributedDataParallel but get the error: RuntimeError: Expected to have finished reduction in the prior iteration before starting a new one. What is the most likely cause?

medium

A. You forgot to call optimizer.step() after loss.backward()

B. You did not wrap the model with DistributedDataParallel

C. You used the wrong backend in init_process_group

D. You did not call dist.destroy_process_group() at the end

Solution

1. Step 1: Understand the error about unfinished reduction

   This error usually means gradients were not synchronized properly before the next iteration.

2. Step 2: Identify missing optimizer step

   Not calling optimizer.step() after loss.backward() can cause gradients to not be updated and reduction to remain unfinished.

3. Final Answer:

   You forgot to call optimizer.step() after loss.backward() -> Option A

4. Quick Check:

   Missing optimizer.step() causes reduction error = C [OK]

Hint: Always call optimizer.step() after backward() in DDP [OK]

Common Mistakes:

• Skipping optimizer.step() call
• Ignoring proper model wrapping
• Confusing backend errors with sync errors

5. You want to train a model on 4 GPUs using DistributedDataParallel across 2 machines, each with 2 GPUs. Which of the following is the correct approach to set up the training?

hard

A. Launch 4 processes on one machine only, ignoring the second machine

B. Launch 2 processes total, each process uses 2 GPUs, wrap model with DataParallel on each process

C. Launch 4 processes total, each process uses one GPU, initialize process group with correct rank and world_size, wrap model with DDP on each process

D. Launch 1 process per machine, use DDP with device_ids=[0,1] on each machine

Solution

1. Step 1: Understand multi-machine multi-GPU DDP setup

   Each GPU should have one process. For 4 GPUs across 2 machines, launch 4 processes total.

2. Step 2: Confirm process group and model wrapping

   Initialize the process group with correct rank and world_size=4. Wrap the model with DDP on each process using the assigned GPU.

3. Final Answer:

   Launch 4 processes total, each process uses one GPU, initialize process group with correct rank and world_size, wrap model with DDP on each process -> Option C

4. Quick Check:

   One process per GPU with correct init = D [OK]

Hint: One process per GPU, init process group correctly [OK]

Common Mistakes:

• Using fewer processes than GPUs
• Using DataParallel instead of DDP for multi-machine
• Ignoring second machine in setup