dal-nlp

Assenmacher-mat's group workspace

Group: bert_coreset_fnc1_15ep

1-5

of 5

Tags

Notes

Author

lura

State

Finished

Start time

February 17th, 2023 4:15:44 PM

Runtime

1h 55m 31s

Tracked hours

1h 55m 27s

Run path

dal-nlp/active-glae_old/fnkvejs4

Linux-4.15.0-197-generic-x86_64-with-glibc2.27

Python version

3.9.5

Command

/mnt/home/lrauch/projects/dal-toolbox/experiments/active_learning/al_txt.py model=bert dataset=fnc1 output_dir=/mnt/work/glae/glae-results/fnc1/bert/coreset/15_ep/N_INIT100__ACQ_SIZE100__N_ACQ15/seed4 random_seed=4 al_strategy=coreset al_cycle.n_init=100 al_cycle.acq_size=100 al_cycle.n_acq=15 wandb.group=bert_coreset_fnc1_15ep model.n_epochs=15

System Hardware

CPU count	128
Logical CPU count	256
GPU count	1
GPU type	NVIDIA A100-SXM4-40GB

W&B CLI Version

0.13.9

Group

bert_coreset_fnc1_15ep

Config parameters are your model's inputs. Learn more

▶
Config parameters:{} 13 keys
- ▶
  al_cycle:{} 4 keys
  - acq_size:
    100
  - cold_start:
    true
  - n_acq:
    15
  - n_init:
    100
- ▶
  al_strategy:{} 2 keys
  - name:
    "coreset"
  - subset_size:
    10,000
- check_freq_model:
  15
- ▶
  dataset:{} 5 keys
- dataset_path:
  "$HOME/.cache/huggingface/datasets"
- device:
  "cuda"
- ▶
  model:{} 5 keys
- output_dir:
  "/mnt/work/glae/glae-results/fnc1/bert/coreset/15_ep/N_INIT100__ACQ_SIZE100__N_ACQ15/seed4"
- print_freq_eval:
  25
- print_freq_train:
  25
- random_seed:
  4
- val_batch_size:
  128
- ▶
  wandb:{} 5 keys

Summary metrics are your model's outputs. Learn more

▶
Summary metrics:{} 7 keys
- final_auc_acc:
  87.574
- final_auc_acc_blc:
  0.623
- final_auc_f1:
  0.587
- test_batch_acc_balanced_epoch:
  0.7745775384675985
- test_batch_acc_epoch:
  93.53741494461556
- test_batch_f1_epoch:
  0.7467345463685934
- test_loss_epoch:
  0.35614422016910147

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

wandb-history

run-fnkvejs4-history:v0

Assenmacher-mat's group workspace

bert_coreset_fnc1_15ep

bert_coreset_fnc1#4