ChannelWise | kaggle-seti-exp Overview – Weights & Biases

Skip to main content

Ayush-thakur's group workspace

Group: ChannelWise

7

1-1

of 1

Tags

Notes

Train EfficientNetB0 by stacking all the spectrograms channel-wise

Tags

effnetb0

full-channel

Author

State

Finished

Start time

June 9th, 2021 12:29:32 AM

Runtime

5m 33s

Tracked hours

5m 32s

Run path

ayush-thakur/kaggle-seti-exp/28fwglnl

OS

Linux-4.19.0-16-cloud-amd64-x86_64-with-debian-10.9

Python version

3.7.10

Command

SETI/train-seti-experiments.ipynb

System Hardware

CPU count	8
GPU count	1
GPU type	Tesla V100-SXM2-16GB

W&B CLI Version

0.10.31

Group

Job Type

train

Config parameters are your model's inputs. Learn more

▶
Config parameters:{} 14 keys
- architecture:
  "CNN"
- batch_size:
  64
- epochs:
  100
- group:
  "ChannelWise"
- img_height:
  224
- img_width:
  224
- infra:
  "GCP"
- learning_rate:
  0.001
- model_name:
  "MobileNetV2"
- num_folds:
  5
- num_labels:
  1
- seed:
  42
- train_val_split:
  0.2
- wandb_kernel:
  true

Summary metrics are your model's outputs. Learn more

▶
Summary metrics:{} 9 keys
- auc:
  0.9953225255012512
- best_epoch:
  10
- best_val_loss:
  0.3281766474246979
- epoch:
  15
- ▶
  graph:{} 4 keys
  - _type:
    "graph-file"
  - path:
    "media/graph/graph_summary_d394fa6d.graph.json"
  - sha256:
    "d394fa6d4164c7b18668d43fa851a18983821abad6b4633cfb24f0aebdb4b9c7"
  - size:
    1,127
- loss:
  0.08309649676084518
- Val AUC-ROC:
  0.8464733362197876
- val_auc:
  0.8239621520042419
- val_loss:
  0.3843863904476166

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

run-28fwglnl-history:v0

Loading...