Replace progressbar with tqdm (#3494)

Signed-off-by: Kyunggeun Lee <quic_kyunggeu@quicinc.com>
quic · Nov 16, 2024 · cdfd5eb · cdfd5eb
1 parent a56920a
commit cdfd5eb
Show file tree

Hide file tree

Showing 18 changed files with 56 additions and 89 deletions.
diff --git a/Examples/onnx/utils/image_net_evaluator.py b/Examples/onnx/utils/image_net_evaluator.py
@@ -40,7 +40,7 @@
 """
 import logging
 
-import progressbar
+from tqdm import tqdm
 import torch
 import onnxruntime as ort
 
@@ -92,25 +92,19 @@ def evaluate(self, sess: ort.InferenceSession, iterations: int = None) -> float:
         logger.info("Evaluating nn.Module for %d iterations with batch_size %d",
                     iterations, self._val_data_loader.batch_size)
 
-        batch_cntr = 1
-        with progressbar.ProgressBar(max_value=iterations) as progress_bar:
-            for input_data, target_data in self._val_data_loader:
+        for i, (input_data, target_data) in tqdm(enumerate(self._val_data_loader), total=10):
+            if i == 10:
+                break
 
-                inputs_batch = input_data.numpy()
+            inputs_batch = input_data.numpy()
 
-                predicted_batch = sess.run(None, {input_name : inputs_batch})[0]
+            predicted_batch = sess.run(None, {input_name : inputs_batch})[0]
 
-                batch_avg_top_1_5 = accuracy(output=torch.from_numpy(predicted_batch), target=target_data,
-                                             topk=(1, 5))
+            batch_avg_top_1_5 = accuracy(output=torch.from_numpy(predicted_batch), target=target_data,
+                                         topk=(1, 5))
 
-                acc_top1 += batch_avg_top_1_5[0].item()
-                acc_top5 += batch_avg_top_1_5[1].item()
-
-                progress_bar.update(batch_cntr)
-
-                batch_cntr += 1
-                if batch_cntr > iterations:
-                    break
+            acc_top1 += batch_avg_top_1_5[0].item()
+            acc_top5 += batch_avg_top_1_5[1].item()
 
         acc_top1 /= iterations
         acc_top5 /= iterations

diff --git a/Examples/torch/quantization/qat_range_learning_ddp_eval.py b/Examples/torch/quantization/qat_range_learning_ddp_eval.py
@@ -52,7 +52,7 @@
 import socket
 import argparse
 
-import progressbar
+from tqdm import tqdm
 import torch
 import torch.distributed as dist
 import torch.multiprocessing as mp
@@ -129,15 +129,13 @@ def evaluate_quant(model, _):
         model.to(device)
         model.eval()
 
-        with progressbar.ProgressBar(max_value=100) as progress_bar:
-            with torch.no_grad():
-                for i, (images, _) in enumerate(val_loader):
-                    images = images.to(device)
-                    # compute output
-                    _ = model(images)
-                    progress_bar.update(i)
-                    if i == 100:
-                        break
+        with torch.no_grad():
+            for i, (images, _) in tqdm(enumerate(val_loader), total=100):
+                images = images.to(device)
+                # compute output
+                _ = model(images)
+                if i == 100:
+                    break
 
     return evaluate_quant
 

diff --git a/Examples/torch/utils/image_net_evaluator.py b/Examples/torch/utils/image_net_evaluator.py
@@ -40,7 +40,7 @@
 """
 import logging
 
-import progressbar
+from tqdm import tqdm
 import torch
 from torch import nn
 
@@ -104,27 +104,21 @@ def evaluate(self, model: nn.Module, iterations: int = None, use_cuda: bool = Fa
         model = model.to(device)
         model = model.eval()
 
-        batch_cntr = 1
-        with progressbar.ProgressBar(max_value=iterations) as progress_bar:
-            with torch.no_grad():
-                for input_data, target_data in self._val_data_loader:
+        with torch.no_grad():
+            for i, (input_data, target_data) in tqdm(enumerate(self._val_data_loader), total=iterations):
+                if i == iterations:
+                    break
 
-                    inputs_batch = input_data.to(device)
-                    target_batch = target_data.to(device)
+                inputs_batch = input_data.to(device)
+                target_batch = target_data.to(device)
 
-                    predicted_batch = model(inputs_batch)
+                predicted_batch = model(inputs_batch)
 
-                    batch_avg_top_1_5 = accuracy(output=predicted_batch, target=target_batch,
-                                                 topk=(1, 5))
+                batch_avg_top_1_5 = accuracy(output=predicted_batch, target=target_batch,
+                                             topk=(1, 5))
 
-                    acc_top1 += batch_avg_top_1_5[0].item()
-                    acc_top5 += batch_avg_top_1_5[1].item()
-
-                    progress_bar.update(batch_cntr)
-
-                    batch_cntr += 1
-                    if batch_cntr > iterations:
-                        break
+                acc_top1 += batch_avg_top_1_5[0].item()
+                acc_top5 += batch_avg_top_1_5[1].item()
 
         acc_top1 /= iterations
         acc_top5 /= iterations

diff --git a/Examples/torch/utils/image_net_trainer.py b/Examples/torch/utils/image_net_trainer.py
@@ -40,7 +40,7 @@
 """
 import logging
 
-import progressbar
+from tqdm import tqdm
 import torch
 from torch import nn, optim
 
@@ -103,36 +103,31 @@ def _train_loop(self, model: nn.Module, criterion: torch.nn.modules.loss, optimi
         model.train()
 
         avg_loss = 0.0
-        curr_iter = 1
-
-        with progressbar.ProgressBar(max_value=max_iterations) as progress_bar:
-            for images, target in self._train_loader:
-                images = images.to(device)
-                target = target.to(device)
-
-                # compute model output
-                output = model(images)
-                loss = criterion(output, target)
-                avg_loss += loss.item()
-
-                # compute gradient and do SGD step
-                optimizer.zero_grad()
-                loss.backward()
-                optimizer.step()
-
-                progress_bar.update(curr_iter)
-
-                if curr_iter % debug_steps == 0:
-                    eval_accuracy = self._evaluator.evaluate(model, use_cuda=use_cuda)
-                    logger.info('Epoch #%d/%d: iteration #%d/%d: Global Avg Loss=%f, Eval Accuracy=%f',
-                                current_epoch, max_epochs, curr_iter, max_iterations,
-                                avg_loss / curr_iter, eval_accuracy)
-                    # switch to training mode after evaluation
-                    model.train()
-
-                curr_iter += 1
-                if curr_iter > max_iterations:
-                    break
+
+        for i, (images, target) in tqdm(enumerate(self._train_loader), total=max_iterations):
+            if i == max_iterations:
+                break
+
+            images = images.to(device)
+            target = target.to(device)
+
+            # compute model output
+            output = model(images)
+            loss = criterion(output, target)
+            avg_loss += loss.item()
+
+            # compute gradient and do SGD step
+            optimizer.zero_grad()
+            loss.backward()
+            optimizer.step()
+
+            if (i+1) % debug_steps == 0:
+                eval_accuracy = self._evaluator.evaluate(model, use_cuda=use_cuda)
+                logger.info('Epoch #%d/%d: iteration #%d/%d: Global Avg Loss=%f, Eval Accuracy=%f',
+                            current_epoch, max_epochs, curr_iter, max_iterations,
+                            avg_loss / curr_iter, eval_accuracy)
+                # switch to training mode after evaluation
+                model.train()
 
         eval_accuracy = self._evaluator.evaluate(model, use_cuda=use_cuda)
         print("eval : ", eval_accuracy)

diff --git a/Jenkins/Dockerfile.onnx-cpu b/Jenkins/Dockerfile.onnx-cpu
@@ -170,7 +170,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.20.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/Dockerfile.onnx-gpu b/Jenkins/Dockerfile.onnx-gpu
@@ -181,7 +181,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.20.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/Dockerfile.tf-cpu b/Jenkins/Dockerfile.tf-cpu
@@ -153,7 +153,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.19.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/Dockerfile.tf-gpu b/Jenkins/Dockerfile.tf-gpu
@@ -164,7 +164,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.19.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/Dockerfile.tf-torch-cpu b/Jenkins/Dockerfile.tf-torch-cpu
@@ -158,7 +158,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         # protobuf==3.20.1 \
         psutil \
         ptflops \

diff --git a/Jenkins/Dockerfile.torch-cpu b/Jenkins/Dockerfile.torch-cpu
@@ -169,7 +169,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.20.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/Dockerfile.torch-cpu-pt113 b/Jenkins/Dockerfile.torch-cpu-pt113
@@ -169,7 +169,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.20.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/Dockerfile.torch-gpu b/Jenkins/Dockerfile.torch-gpu
@@ -181,7 +181,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.20.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/Dockerfile.torch-gpu-pt113 b/Jenkins/Dockerfile.torch-gpu-pt113
@@ -181,7 +181,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.20.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/fast-release/Dockerfile.torch-gpu-pt21-py310 b/Jenkins/fast-release/Dockerfile.torch-gpu-pt21-py310
@@ -178,7 +178,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.20.2 \
         psutil \
         ptflops \

diff --git a/Jenkins/fast-release/Dockerfile.torch-gpu-pt21-py38 b/Jenkins/fast-release/Dockerfile.torch-gpu-pt21-py38
@@ -174,7 +174,6 @@ RUN python3 -m pip --no-cache-dir install \
         piccolo-theme \
         Pillow==9.3.0 \
         pluggy==0.12.0 \
-        progressbar2 \
         protobuf==3.20.2 \
         psutil \
         ptflops \

diff --git a/packaging/dependencies/fast-release/torch-gpu-pt21-py310/reqs_pip_torch_gpu.txt b/packaging/dependencies/fast-release/torch-gpu-pt21-py310/reqs_pip_torch_gpu.txt
@@ -13,7 +13,6 @@ onnx==1.14.1
 onnxsim
 osqp
 pandas==1.5.3
-progressbar2
 protobuf==3.20.2
 psutil
 pybind11

diff --git a/packaging/dependencies/fast-release/torch-gpu-pt21-py38/reqs_pip_torch_gpu.txt b/packaging/dependencies/fast-release/torch-gpu-pt21-py38/reqs_pip_torch_gpu.txt
@@ -13,7 +13,6 @@ onnx==1.14.1
 onnxsim
 osqp
 pandas==1.5.3
-progressbar2
 protobuf==3.20.2
 psutil
 pybind11

diff --git a/packaging/dependencies/reqs_pip_common.txt b/packaging/dependencies/reqs_pip_common.txt
@@ -6,7 +6,6 @@ hvplot==0.9.2
 Jinja2==3.0.3
 jsonschema
 osqp
-progressbar2
 pybind11
 PyYAML
 scikit-learn==1.1.3