-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathexample_runs_standard_memory.sh
43 lines (34 loc) · 2.81 KB
/
example_runs_standard_memory.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
#!/bin/bash
source ~/anaconda3/etc/profile.d/conda.sh
conda activate research
# MNIST Standard Memory Seed 0
python launch.py --experiment_name "MNIST_MEMO1_SEED0" --memo_per_class_context "50" --context_layers 0 1 2 3 4 \
--context_learner "LogisticRegression(random_state=0, max_iter=50, C=0.4)" --dataset "MNIST" --number_of_tasks "5" \
--model "CNN_MNIST" --seed "0" --learning_rate "0.01" \
--batch_size "32" --weight_decay "0.0" --phase_epochs "5" --activation_perc "95.0" --max_phases "5"
# FashionMNIST Standard Memory Seed 0
python launch.py --experiment_name "FashionMNIST_MEMO1_SEED0" --memo_per_class_context "50" --context_layers 0 1 2 3 4 \
--context_learner "LogisticRegression(random_state=0, max_iter=50, C=0.4)" --dataset "FMNIST" --number_of_tasks "5" \
--model "CNN_MNIST" --seed "0" --learning_rate "0.005" \
--batch_size "32" --weight_decay "0.0" --phase_epochs "5" --activation_perc "95.0" --max_phases "5"
# EMNIST Standard Memory Seed 0
python launch.py --experiment_name "EMNIST_MEMO1_SEED0" --memo_per_class_context "50" --context_layers 0 1 2 3 4 \
--context_learner "LogisticRegression(random_state=0, max_iter=20, C=0.15)" --dataset "EMNIST" --number_of_tasks "13" \
--model "CNN_MNIST" --seed "0" --learning_rate "0.005" --batch_size "32" --weight_decay "0.0" --phase_epochs "5" \
--activation_perc "95.0" --max_phases "5"
# CIFAR10 Standard Memory Seed 0
python launch.py --experiment_name "CIFAR10_MEMO1_SEED0" --memo_per_class_context "150" \
--context_layers 0 1 2 3 4 5 6 7 8 9 10 11 --context_learner "LogisticRegression(random_state=0, max_iter=20, C=0.01)" \
--dataset "CIFAR10" --number_of_tasks "5" --model "VGG11_SLIM" --seed "0" --learning_rate "0.01" --batch_size "32" \
--weight_decay "0.0001" --phase_epochs "10" --activation_perc "95.0" --max_phases "5"
# CIFAR100 Standard Memory Seed 0
python launch.py --experiment_name "CIFAR100_MEMO1_SEED0" --memo_per_class_context "50" \
--context_layers 0 1 2 3 4 5 6 7 8 9 10 11 --context_learner "LogisticRegression(random_state=0, max_iter=20, C=0.005)" \
--dataset "CIFAR100" --number_of_tasks "10" --model "VGG11_SLIM" --seed "0" --learning_rate "0.005" --batch_size "32" \
--weight_decay "0.001" --phase_epochs "15" --activation_perc "95.0" --max_phases "5"
# TinyImageNet Standard Memory Seed 0
python launch.py --experiment_name "TinyImagenet_MEMO1_SEED0" --memo_per_class_context "25" \
--context_layers 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 \
--context_learner "LogisticRegression(random_state=0, max_iter=25, C=0.001)" --dataset "TinyImagenet" \
--number_of_tasks "5" --model "ResNet18" --seed "0" --learning_rate "0.01" --batch_size "64" --weight_decay "0.0" \
--phase_epochs "20" --activation_perc "97.5" --max_phases "5"