forked from yuexujiang/MUTarget
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathrun_hellbender.sh
87 lines (68 loc) · 3.88 KB
/
run_hellbender.sh
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
#!/bin/bash
#SBATCH --mem 50G
#SBATCH -n 1
#SBATCH --gres gpu:A100:1
#SBATCH --time 2-00:00:00 #Time for the job to run
#SBATCH --job-name mutarget
##SBATCH -p gpu
#SBATCH -p xudong-gpu
#SBATCH -A xudong-lab
module load miniconda3
module load miniconda3
# Activate the Conda environment
#source activate /home/wangdu/.conda/envs/pytorch1.13.0
source activate /home/wangdu/data/env/splm_gvp
export TORCH_HOME=/cluster/pixstor/xudong-lab/duolin/torch_cache/
export HF_HOME=/cluster/pixstor/xudong-lab/duolin/transformers_cache/
#python train.py --config_path ./configs/config_supcon_hardfalse.yaml --result_path ./results_supcon_hardFalse_pft2/b20_p2_n4
#1535976 after finished should change into b10_p2_n4!!!!
#no supcon loss
#python train.py --config_path ./configs/config_supcon_hardtrue.yaml --result_path ./results_supcon_hardTrue_pft2/b10_p2_n4
#1535980 no supcon loss
#before this are all used supcon data sampling and supcon datasample, anchor may from same class in one batch
#python train.py --config_path ./configs/config_supcon_hardtrue.yaml --result_path ./results_supcon_hardTrue_pft2/cleanloader/b8_p3_n9
#1556043
#python train.py --config_path ./configs/config_supcon_hardtrue_esm8.yaml --result_path ./results_supcon_hardTrue_pft2_esm8/cleanloader/b8_p4_n12
#1556054
#python train.py --config_path ./configs/config_supcon_hardtrue_esm8.yaml \
#--predict 1 \
#--resume_path /cluster/pixstor/xudong-lab/duolin/MUTargetCLEAN/results_supcon_hardTrue_pft2_esm8/cleanloader/b8_p4_n12/2024-03-27__00-32-00/checkpoints/best_model.pth \
#--result_path ./results_supcon_hardTrue_pft2_esm8/cleanloader/b8_p4_n12/predict
#python train.py --config_path ./configs/config_supcon_hardtrue.yaml \
#--predict 1 \
#--resume_path /cluster/pixstor/xudong-lab/duolin/MUTargetCLEAN/results_supcon_hardTrue_pft2/cleanloader/b8_p3_n9/2024-03-27__00-26-05/checkpoints/best_model.pth \
#--result_path ./results_supcon_hardTrue_pft2/cleanloader/b8_p3_n9/predict
#python train.py --config_path ./configs/config_supcon_hardtrue_esmfix.yaml \
#--result_path ./config_supcon_hardtrue_esmfix/cleanloader/
#1565611 lr small not work!
#1565887 lr bigger not work!
#using batchsample dataloader not clean but clean's pos and neg, try original codes.
#python train.py --config_path ./configs/config_supcon_onlysampling.yaml --result_path ./results_supcon_hardTrue_onlysampling/b10_p2_n4
#1565886 stopped by error
#1566832 modified train.py
#using batchsample dataloader not clean but clean's pos and neg, try original codes. and original lr schedular
#python train.py --config_path ./configs/config_supcon_onlysampling.yaml \
#--result_path ./results_supcon_hardTrue_onlysampling/b16_p2_n2_adam_epoch20
#epoch 40 1571588
#epoch 20 1584450
# run in digbio using pytorch1.13.0 but transformers==4.38.2
#python train.py --config_path ./configs/config_supcon_onlysampling.yaml \
#--predict 1 \
#--resume_path /cluster/pixstor/xudong-lab/duolin/MUTargetCLEAN/results_supcon_hardTrue_onlysampling/b10_p2_n4/2024-03-28__11-36-17/checkpoints//best_model.pth \
#--result_path ./results_supcon_hardTrue_onlysampling/b10_p2_n4/predict
#python train.py --config_path ./configs/config_supcon_onlysampling_hardfalse.yaml \
#--result_path ./results_supcon_hardFalse_onlysampling/b16_p2_n2
#1571554
#original lr hardFalse
#python train.py --config_path ./configs/config_supcon_onlysampling_hardfalse.yaml \
#--result_path ./results_supcon_hardFalse_onlysampling/b16_p2_n2_adam
#1571590
#test the codes!
#python train.py --config_path ./configs/config_nosupcon.yaml \
#--result_path ./config_nosupcon/b16_epoch20_adam
#1584449
#python train.py --config_path ./config_mutarget.yaml
python train.py --config_path ./configs/config_supcon_onlysampling.yaml \
--resume_path /cluster/pixstor/xudong-lab/duolin/MUTargetCLEAN/results_supcon_hardTrue_onlysampling/b10_p2_n4/2024-03-28__11-36-17/checkpoints//best_model.pth \
--result_path ./results_supcon_hardTrue_onlysampling/b10_p2_n4/run2
#1594176