krenerd77

Krenerd77's group workspace

Group: CIFAR10-ResNet18-Asymmetric_noise

1-6

of 6

Tags

Notes

Author

krenerd77

State

Finished

Start time

July 20th, 2022 6:02:02 PM

Runtime

2h 46m 13s

Tracked hours

2h 46m 10s

Run path

krenerd77/CIFAR10-classification/1vy8sh3n

Linux-5.4.188+-x86_64-with-Ubuntu-18.04-bionic

Python version

3.7.13

Git repository

git clone https://github.com/krenerd/awesome-modular-pytorch-lightning

Git state

git checkout -b "CIFAR10-ResNet18-Asymmetric_noise-0.8-Jul20_18-01-54" 65ee2b53c5d193906f159e8a5c6bedfc13f3bdef

Command

train.py --name CIFAR10-ResNet18-Asymmetric_noise-0.8 --group CIFAR10-ResNet18-Asymmetric_noise --config configs/data/label-noise/8.yaml configs/vision/classification/resnet-cifar10.yaml configs/vision/models/resnet/resnet18-custom.yaml configs/data/cifar10-kuangliu.yaml configs/device/gpu.yaml configs/utils/wandb.yaml configs/utils/train.yaml

System Hardware

CPU count	2
GPU count	1
GPU type	Tesla T4

W&B CLI Version

0.12.21

Group

CIFAR10-ResNet18-Asymmetric_noise

Config parameters are your model's inputs. Learn more

▶
Config parameters:{} 79 keys
- ▶
  callbacks:[] 2 items
  - ▶
    0:{} 3 keys
    - ▶
      args:{} 4 keys
      - mode:
        "max"
      - monitor:
        "epoch_val/accuracy"
      - save_last:
        true
      - save_top_k:
        1
    - file:
      "lightning"
    - name:
      "ModelCheckpoint"
  - ▶
    1:{} 3 keys
    - ▶
      args:{} 1 key
      - logging_interval:
        "epoch"
    - file:
      "lightning"
    - name:
      "LearningRateMonitor"
- ▶
  const/label_map:[] 10 items
- ▶
  const/normalization_mean:[] 3 items
  - 0:
    0.4914
  - 1:
    0.4822
  - 2:
    0.4465
- ▶
  const/normalization_std:[] 3 items
  - 0:
    0.2023
  - 1:
    0.1994
  - 2:
    0.201
- const/num_classes:
  10
- const/task:
  "image classification"
- dataloader/base_dataloader/batch_size:
  64
- dataloader/base_dataloader/num_workers:
  8
- dataloader/base_dataloader/pin_memory:
  true
- dataloader/base_dataloader/prefetch_factor:
  2
- dataloader/base_dataloader/shuffle:
  true
- dataloader/trn/shuffle:
  true
- dataloader/val/drop_last:
  false
- dataloader/val/shuffle:
  false
- dataset/dataset_base_cfg/root:
  "./cifar10"
- dataset/dataset_subset_cfg/trn/download:
  true
- dataset/dataset_subset_cfg/trn/train:
  true
- dataset/dataset_subset_cfg/val/train:
  false
- dataset/initial_transform/name:
  "TupleToClassificationData"
- ▶
  dataset/mapping:[] 1 item
  - ▶
    0:[] 2 items
    - 0:
      "trn"
    - ▶
      1:[] 1 item
      - ▶
        0:{} 2 keys
        ▶
        args:{} 3 keys
        method:
        "asymmetric"
        num_classes:
        10
        transition_prob:
        0.8
        name:
        "InstanceIndependantNoise"
- dataset/MODE:
  "torchvision"
- dataset/NAME:
  "CIFAR10"
- DEBUG_MODE:
  false
- ▶
  debug/network_summary/input_shape:[] 3 items
  - 0:
    3
  - 1:
    224
  - 2:
    224
- debug/view_train_augmentation/plot_size:
  5
- ▶
  debug/view_train_augmentation/resize_to:[] 2 items
  - 0:
    256
  - 1:
    256
- debug/view_train_augmentation/save_to:
  "vis/training_samples.png"
- ▶
  debug/view_train_augmentation/subplot_dim:[] 2 items
  - 0:
    4
  - 1:
    4
- model/backbone/cfg/low_res:
  true
- model/backbone/ID:
  "ResNet18"
- model/backbone/out_features:
  512
- model/backbone/TYPE:
  "custom"
- model/modules/classifier/args/dropout:
  0.2
- model/modules/classifier/args/in_features:
  512
- model/modules/classifier/args/num_classes:
  10
- model/modules/classifier/args/reduction:
  "gap"
- model/modules/classifier/args/return_logits:
  true
- model/modules/classifier/input:
  "output"
- model/modules/classifier/name:
  "ClassificationHead"
- model/modules/loss_fn/name:
  "CrossEntropyLoss"
- name:
  "CIFAR10-ResNet18-Asymmetric_noise-0.8"
- seed:
  true
- trainer/accumulate_grad_batches:
  1
- trainer/check_val_every_n_epoch:
  5
- ▶
  trainer/gpus:[] 1 item
  - 0:
    0
- trainer/precision:
  32
- ▶
  transform:[] 2 items
  - ▶
    0:[] 2 items
    - 0:
      "trn"
    - ▶
      1:[] 2 items
      - ▶
        0:{} 2 keys
        ▶
        args:{} 2 keys
        ▶
        ARGS:{} 2 keys
        padding:
        4
        size:
        32
        NAME:
        "RandomCrop"
        name:
        "TorchTransforms"
      - ▶
        1:{} 2 keys
        ▶
        args:{} 1 key
        NAME:
        "RandomHorizontalFlip"
        name:
        "TorchTransforms"
  - ▶
    1:[] 2 items
    - 0:
      "trn,val"
    - ▶
      1:[] 2 items
      - ▶
        0:{} 2 keys
        args:{} 0 keys
        name:
        "ToTensor"
      - ▶
        1:{} 2 keys
        ▶
        args:{} 2 keys
        ▶
        mean:[] 3 items
        0:
        0.4914
        1:
        0.4822
        2:
        0.4465
        ▶
        std:[] 3 items
        0:
        0.2023
        1:
        0.1994
        2:
        0.201
        name:
        "Normalize"
- VERBOSE:
  "DEFAULT"
- wandb/group:
  "CIFAR10-ResNet18-Asymmetric_noise"
- wandb/project:
  "CIFAR10-classification"

Summary metrics are your model's outputs. Learn more

▶
Summary metrics:{} 14 keys
- epoch:
  200
- epoch_test/accuracy:
  0.1137000024318695
- ▶
  epoch_test/confusion_matrix:{} 7 keys
- epoch_test/loss:
  6.064535140991211
- epoch_trn/accuracy:
  0.9983000159263612
- ▶
  epoch_trn/confusion_matrix:{} 7 keys
- epoch_trn/loss:
  0.008925073780119419
- epoch_val/accuracy:
  0.1137000024318695
- ▶
  epoch_val/confusion_matrix:{} 7 keys
- epoch_val/loss:
  6.058297157287598
- lr-SGD:
  0.00000185051277509118
- step/main_loss:
  6.060582637786865
- step/train_loss:
  0.010406105779111384
- trainer/global_step:
  156,400

This run produced these artifacts as outputs. Total: 1. Learn more

wandb-history

run-1vy8sh3n-history:v0