Some experiments cannot be reproduced

[invisprints]

I followed the readme for data preparation, but some experimental results in some papers could not be reproduced

MICIL in scenario E3:

python main_MICIL.py --name_experiment MICIL --mode MICIL_train --scenario sceE3 --method MICIL --wn
python main_MICIL.py --name_experiment MICIL --mode MICIL_test --scenario sceE3

And get following results:

----- Loading data (exp = 0) -----
100%|██████████████████| 58/58 [00:00<00:00, 98.15it/s]
----- Testing (exp = 0) -----
ACC  = 0.8793
BACC = 0.8939
F1S  = 0.8793
[[26  7  0  0  0 0]
[ 0 25  0  0  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]]
--- VALIDATION IN ALL EXPERIENCES ---
ACC  = 0.8793
BACC = 0.8939
F1S  = 0.8793
[[26  7  0  0  0  0]
[ 0 25  0  0  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]]
----- Loading data (exp = 1) -----
100%|███████████████| 41/41 [00:00<00:00, 46.64it/s]
----- Testing (exp = 1) -----
ACC  = 0.1724
BACC = 0.1758
F1S  = 0.2982
[[ 5  0 11 17  0  0]
[ 0  5 10 10  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]
[ 0  0  0  0  0  0]]
ACC  = 0.878
BACC = 0.881
F1S  = 0.887
[[ 0  0  0  0  0  0]
 [ 0  0  0  0  0  0]
 [ 0  0 20  0  0  0]
 [ 0  1  4 16  0  0]
 [ 0  0  0  0  0  0]
 [ 0  0  0  0  0  0]]
 
--- VALIDATION IN ALL EXPERIENCES ---
ACC  = 0.4646
BACC = 0.5284
F1S  = 0.4253
[[ 5  0 11 17  0  0]
 [ 0  5 10 10  0  0]
 [ 0  0 20  0  0  0]
 [ 0  1  4 16  0  0]
 [ 0  0  0  0  0  0]
 [ 0  0  0  0  0  0]]
----- Loading data (exp = 2) -----
100%|████████████████████| 20/20 [00:00<00:00, 53.68it/s]
----- Testing (exp = 2) -----
ACC  = 0.0172
BACC = 0.02
F1S  = 0.0385
[[ 0  0  0  0  8 25]
 [ 0  1  0  0  3 21]
 [ 0  0  0  0  0  0]
 [ 0  0  0  0  0  0]
 [ 0  0  0  0  0  0]
 [ 0  0  0  0  0  0]]
ACC  = 0.0
BACC = 0.0
F1S  = 0.0
[[ 0  0  0  0  0  0]
 [ 0  0  0  0  0  0]
 [ 0  0  0  0  5 15]
 [ 0  1  0  0  4 16]
 [ 0  0  0  0  0  0]
 [ 0  0  0  0  0  0]]
ACC  = 0.85
BACC = 0.8535
F1S  = 0.8496
[[0 0 0 0 0 0]
 [0 0 0 0 0 0]
 [0 0 0 0 0 0]
 [0 0 0 0 0 0]
 [0 0 0 0 9 2]
 [0 0 0 0 1 8]]
--- VALIDATION IN ALL EXPERIENCES ---
ACC  = 0.1513
BACC = 0.2912
F1S  = 0.1133
[[ 0  0  0  0  8 25]
 [ 0  1  0  0  3 21]
 [ 0  0  0  0  5 15]
 [ 0  1  0  0  4 16]
 [ 0  0  0  0  9  2]
 [ 0  0  0  0  1  8]]

In the exp=2, the average acc = 0.1513, much lower than the results in the paper. Anything wrong in my setting?