vocsep

Fosfrancesco's group workspace

Group: Voice Separation - split

test_f1

1-20

of 70

Tags

Notes

Author

melkisedeath

State

Finished

Start time

January 5th, 2024 6:13:23 PM

Runtime

25m 31s

Tracked hours

25m 25s

Run path

vocsep/MusGConv/52p75beh

Linux-4.15.0-213-generic-x86_64-with-glibc2.27

Python version

3.9.18

Git repository

git clone git@github.com:manoskary/MusGConv.git

Git state

git checkout -b "MusGConv-2x64-woLN-lr=0.003-wd=0.005-dr=0.5-rl=auto-jk=False" 24696d3a9746ad8ccd25692e58824162fc9b5aa1

Command

/share/hel/home/manos/codes/MusGConv/experiments/voice_separation.py --gpus 1 --use_wandb --wandb_entity vocsep --model MusGConv --use_reledge --aggregation mul --seed 30

System Hardware

CPU count	20
Logical CPU count	40
GPU count	4
GPU type	GeForce GTX 1080 Ti

W&B CLI Version

0.15.11

Group

Voice Separation - split

Job Type

MusGConv-woPE-wEF-woSEF-woEE-noCat

Config parameters are your model's inputs. Learn more

▶
Config parameters:{} 35 keys
- activation:
  "relu"
- aggregation:
  "mul"
- batch_size:
  28
- block:
  "MusGConv"
- collection:
  "split"
- dropout:
  0.5
- force_reload:
  false
- gpus:
  "1"
- homogeneous:
  false
- in_feats:
  23
- jk:
  false
- linear_assignment:
  false
- load_from_checkpoint:
  false
- lr:
  0.003
- model:
  "MusGConv"
- module:
  "MetricalLinkPredictionModel"
- n_epochs:
  100
- n_hidden:
  64
- n_layers:
  2
- num_workers:
  20
- pitch_embedding:
  null
- pot_edges_max_dist:
  2
- reg_loss_type:
  "la"
- reg_loss_weight:
  "auto"
- return_edge_emb:
  false
- seed:
  30
- stack_convs:
  false
- tags:
  ""
- use_jk:
  false
- use_metrical:
  false
- use_reledge:
  true
- use_signed_features:
  false
- use_wandb:
  true
- wandb_entity:
  "vocsep"
- weight_decay:
  0.005

Summary metrics are your model's outputs. Learn more

▶
Summary metrics:{} 23 keys
- epoch:
  100
- lr-AdamW:
  0.000000003
- regloss_weight:
  0.009900002740323544
- regloss_weighted:
  0.32524657249450684
- test_fscore:
  0.8015381693840027
- test_fscore_la:
  0.8830262422561646
- test_la_score:
  1.4962475299835205
- test_precision:
  0.740257203578949
- test_precision_la:
  0.935525119304657
- test_recall:
  0.8825506567955017
- test_recall_la:
  0.8381925225257874
- train_acc:
  0.9192688465118408
- train_joinloss:
  0.7125110030174255
- train_loss_epoch:
  0.3872644007205963
- train_loss_step:
  0.38086438179016113
- train_regloss_epoch:
  32.85319519042969
- train_regloss_step:
  14.869146347045898
- trainer/global_step:
  2,700
- val_fscore:
  0.7492945790290833
- val_la_score:
  2.1626503467559814
- val_loss:
  0.31564760208129883
- val_precision:
  0.689441978931427
- val_recall:
  0.9087728261947632

This run consumed these artifacts as inputs. Learn more

job

job-git_github.com_manoskary_MusGConv.git_experiments_voice_separation.py:v14

This run produced these artifacts as outputs. Total: 2. Learn more

model

model-52p75beh:v0

wandb-history

run-52p75beh-history:v0