Upload 11 files
Browse files- .gitattributes +2 -0
- results/moe_hash_with_lb/best_model.pt +3 -0
- results/moe_hash_with_lb/training_metrics.json +98 -0
- results/moe_token_choice/best_model.pt +3 -0
- results/moe_token_choice/expert_usage_history.json +427 -0
- results/moe_token_choice/moe_token_choice_test_predictions.json +3 -0
- results/moe_token_choice/training_curves.png +3 -0
- results/moe_token_choice_with_lb/best_model.pt +3 -0
- results/moe_token_choice_with_lb/training_metrics.json +98 -0
- results/moe_token_choice_without_lb/best_model.pt +3 -0
- results/moe_token_choice_without_lb/training_metrics.json +98 -0
- results/standard_mha_moe/best_model.pt +3 -0
.gitattributes
CHANGED
|
@@ -65,3 +65,5 @@ results/lb_comparison/expert_balance_comparison.png filter=lfs diff=lfs merge=lf
|
|
| 65 |
results/lb_comparison/loss_comparison.png filter=lfs diff=lfs merge=lfs -text
|
| 66 |
results/moe_hash/moe_hash_test_predictions.json filter=lfs diff=lfs merge=lfs -text
|
| 67 |
results/moe_hash/training_curves.png filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
| 65 |
results/lb_comparison/loss_comparison.png filter=lfs diff=lfs merge=lfs -text
|
| 66 |
results/moe_hash/moe_hash_test_predictions.json filter=lfs diff=lfs merge=lfs -text
|
| 67 |
results/moe_hash/training_curves.png filter=lfs diff=lfs merge=lfs -text
|
| 68 |
+
results/moe_token_choice/moe_token_choice_test_predictions.json filter=lfs diff=lfs merge=lfs -text
|
| 69 |
+
results/moe_token_choice/training_curves.png filter=lfs diff=lfs merge=lfs -text
|
results/moe_hash_with_lb/best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ec6145602df946505b720c0f174564accf1b6eabeee36fb85c86008af9ed553
|
| 3 |
+
size 2159273731
|
results/moe_hash_with_lb/training_metrics.json
ADDED
|
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"train_losses": [
|
| 3 |
+
7.284618572235107,
|
| 4 |
+
6.295677210235596,
|
| 5 |
+
5.8652772613525395,
|
| 6 |
+
5.56426231765747,
|
| 7 |
+
5.3282255226135256
|
| 8 |
+
],
|
| 9 |
+
"train_nll_losses": [
|
| 10 |
+
7.188395218658448,
|
| 11 |
+
6.206792694091797,
|
| 12 |
+
5.777505061340332,
|
| 13 |
+
5.47729527053833,
|
| 14 |
+
5.241658354949951
|
| 15 |
+
],
|
| 16 |
+
"train_lb_losses": [
|
| 17 |
+
0.0962233625292778,
|
| 18 |
+
0.08888451331853867,
|
| 19 |
+
0.08777220013141632,
|
| 20 |
+
0.08696705160140991,
|
| 21 |
+
0.08656716544628143
|
| 22 |
+
],
|
| 23 |
+
"val_losses": [
|
| 24 |
+
6.504456277877566,
|
| 25 |
+
5.991641816638765,
|
| 26 |
+
5.692757720039005,
|
| 27 |
+
5.5011179106576105,
|
| 28 |
+
5.334456035069057
|
| 29 |
+
],
|
| 30 |
+
"use_load_balancer": true,
|
| 31 |
+
"expert_balance_history": [
|
| 32 |
+
{
|
| 33 |
+
"epoch": 1,
|
| 34 |
+
"cv_scores": {
|
| 35 |
+
"encoder_layer_0": 0.9899597063375694,
|
| 36 |
+
"encoder_layer_1": 0.7990518141604487,
|
| 37 |
+
"encoder_layer_2": 0.5562532583831334,
|
| 38 |
+
"encoder_layer_3": 0.9468495465629408,
|
| 39 |
+
"decoder_layer_0": 0.21338860568053053,
|
| 40 |
+
"decoder_layer_1": 0.434910226517034,
|
| 41 |
+
"decoder_layer_2": 0.9061288024067989,
|
| 42 |
+
"decoder_layer_3": 0.8562114158633968
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 2,
|
| 47 |
+
"cv_scores": {
|
| 48 |
+
"encoder_layer_0": 0.9942137591217549,
|
| 49 |
+
"encoder_layer_1": 0.9034253039673751,
|
| 50 |
+
"encoder_layer_2": 0.7761734095185961,
|
| 51 |
+
"encoder_layer_3": 0.9762803535163799,
|
| 52 |
+
"decoder_layer_0": 0.1967641643874591,
|
| 53 |
+
"decoder_layer_1": 0.24658043997511517,
|
| 54 |
+
"decoder_layer_2": 0.9482812902409864,
|
| 55 |
+
"decoder_layer_3": 0.9350409567436199
|
| 56 |
+
}
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 3,
|
| 60 |
+
"cv_scores": {
|
| 61 |
+
"encoder_layer_0": 0.9979227118321601,
|
| 62 |
+
"encoder_layer_1": 0.9387816740785491,
|
| 63 |
+
"encoder_layer_2": 0.853512993646018,
|
| 64 |
+
"encoder_layer_3": 0.9866695375658572,
|
| 65 |
+
"decoder_layer_0": 0.20424998009552645,
|
| 66 |
+
"decoder_layer_1": 0.23413879522609943,
|
| 67 |
+
"decoder_layer_2": 0.9657628202085438,
|
| 68 |
+
"decoder_layer_3": 0.9512014746936421
|
| 69 |
+
}
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 4,
|
| 73 |
+
"cv_scores": {
|
| 74 |
+
"encoder_layer_0": 1.0002050910590432,
|
| 75 |
+
"encoder_layer_1": 0.9561595695534928,
|
| 76 |
+
"encoder_layer_2": 0.8924241572167828,
|
| 77 |
+
"encoder_layer_3": 0.9920346981354939,
|
| 78 |
+
"decoder_layer_0": 0.21548889996954693,
|
| 79 |
+
"decoder_layer_1": 0.23986793581997143,
|
| 80 |
+
"decoder_layer_2": 0.9749462380467006,
|
| 81 |
+
"decoder_layer_3": 0.9621878323536873
|
| 82 |
+
}
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 5,
|
| 86 |
+
"cv_scores": {
|
| 87 |
+
"encoder_layer_0": 1.0016175709974362,
|
| 88 |
+
"encoder_layer_1": 0.9665781082053273,
|
| 89 |
+
"encoder_layer_2": 0.9152427659781934,
|
| 90 |
+
"encoder_layer_3": 0.9951101417418388,
|
| 91 |
+
"decoder_layer_0": 0.2274634215213949,
|
| 92 |
+
"decoder_layer_1": 0.24629436643568103,
|
| 93 |
+
"decoder_layer_2": 0.9802201580055626,
|
| 94 |
+
"decoder_layer_3": 0.9680230320659131
|
| 95 |
+
}
|
| 96 |
+
}
|
| 97 |
+
]
|
| 98 |
+
}
|
results/moe_token_choice/best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a497228c41fbb0013e1d9f5a1514316e8c7f171f2b28625340468200505c4f06
|
| 3 |
+
size 2159613443
|
results/moe_token_choice/expert_usage_history.json
ADDED
|
@@ -0,0 +1,427 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"epoch": 1,
|
| 4 |
+
"usage": {
|
| 5 |
+
"encoder_layer_0": {
|
| 6 |
+
"0": 140993937,
|
| 7 |
+
"1": 135712261,
|
| 8 |
+
"2": 139518919,
|
| 9 |
+
"3": 139655991,
|
| 10 |
+
"4": 136781290,
|
| 11 |
+
"5": 137254831,
|
| 12 |
+
"6": 135476784,
|
| 13 |
+
"7": 137302659
|
| 14 |
+
},
|
| 15 |
+
"encoder_layer_1": {
|
| 16 |
+
"0": 153395951,
|
| 17 |
+
"1": 155997088,
|
| 18 |
+
"2": 135728108,
|
| 19 |
+
"3": 115885780,
|
| 20 |
+
"4": 147534633,
|
| 21 |
+
"5": 113059645,
|
| 22 |
+
"6": 139938622,
|
| 23 |
+
"7": 141156845
|
| 24 |
+
},
|
| 25 |
+
"encoder_layer_2": {
|
| 26 |
+
"0": 135360959,
|
| 27 |
+
"1": 143571480,
|
| 28 |
+
"2": 130426240,
|
| 29 |
+
"3": 138121467,
|
| 30 |
+
"4": 143331476,
|
| 31 |
+
"5": 143239623,
|
| 32 |
+
"6": 143669414,
|
| 33 |
+
"7": 124976013
|
| 34 |
+
},
|
| 35 |
+
"encoder_layer_3": {
|
| 36 |
+
"0": 138108424,
|
| 37 |
+
"1": 143624638,
|
| 38 |
+
"2": 136522770,
|
| 39 |
+
"3": 143495589,
|
| 40 |
+
"4": 136073789,
|
| 41 |
+
"5": 135460995,
|
| 42 |
+
"6": 133884692,
|
| 43 |
+
"7": 135525775
|
| 44 |
+
},
|
| 45 |
+
"decoder_layer_0": {
|
| 46 |
+
"0": 11885357,
|
| 47 |
+
"1": 11801320,
|
| 48 |
+
"2": 13060220,
|
| 49 |
+
"3": 12342825,
|
| 50 |
+
"4": 12361295,
|
| 51 |
+
"5": 13069915,
|
| 52 |
+
"6": 11787056,
|
| 53 |
+
"7": 13554748
|
| 54 |
+
},
|
| 55 |
+
"decoder_layer_1": {
|
| 56 |
+
"0": 11862357,
|
| 57 |
+
"1": 12259029,
|
| 58 |
+
"2": 11620687,
|
| 59 |
+
"3": 14837367,
|
| 60 |
+
"4": 11779914,
|
| 61 |
+
"5": 13099878,
|
| 62 |
+
"6": 11683397,
|
| 63 |
+
"7": 12720107
|
| 64 |
+
},
|
| 65 |
+
"decoder_layer_2": {
|
| 66 |
+
"0": 12201210,
|
| 67 |
+
"1": 12112379,
|
| 68 |
+
"2": 14274767,
|
| 69 |
+
"3": 12285646,
|
| 70 |
+
"4": 12123051,
|
| 71 |
+
"5": 12171337,
|
| 72 |
+
"6": 12135313,
|
| 73 |
+
"7": 12559033
|
| 74 |
+
},
|
| 75 |
+
"decoder_layer_3": {
|
| 76 |
+
"0": 12632673,
|
| 77 |
+
"1": 15462911,
|
| 78 |
+
"2": 11907600,
|
| 79 |
+
"3": 12328118,
|
| 80 |
+
"4": 11866897,
|
| 81 |
+
"5": 11858883,
|
| 82 |
+
"6": 11895272,
|
| 83 |
+
"7": 11910382
|
| 84 |
+
}
|
| 85 |
+
}
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 2,
|
| 89 |
+
"usage": {
|
| 90 |
+
"encoder_layer_0": {
|
| 91 |
+
"0": 140993937,
|
| 92 |
+
"1": 135712261,
|
| 93 |
+
"2": 139518919,
|
| 94 |
+
"3": 139655991,
|
| 95 |
+
"4": 136781290,
|
| 96 |
+
"5": 137254831,
|
| 97 |
+
"6": 135476784,
|
| 98 |
+
"7": 137302659
|
| 99 |
+
},
|
| 100 |
+
"encoder_layer_1": {
|
| 101 |
+
"0": 153395951,
|
| 102 |
+
"1": 155997088,
|
| 103 |
+
"2": 135728108,
|
| 104 |
+
"3": 115885780,
|
| 105 |
+
"4": 147534633,
|
| 106 |
+
"5": 113059645,
|
| 107 |
+
"6": 139938622,
|
| 108 |
+
"7": 141156845
|
| 109 |
+
},
|
| 110 |
+
"encoder_layer_2": {
|
| 111 |
+
"0": 135360959,
|
| 112 |
+
"1": 143571480,
|
| 113 |
+
"2": 130426240,
|
| 114 |
+
"3": 138121467,
|
| 115 |
+
"4": 143331476,
|
| 116 |
+
"5": 143239623,
|
| 117 |
+
"6": 143669414,
|
| 118 |
+
"7": 124976013
|
| 119 |
+
},
|
| 120 |
+
"encoder_layer_3": {
|
| 121 |
+
"0": 138108424,
|
| 122 |
+
"1": 143624638,
|
| 123 |
+
"2": 136522770,
|
| 124 |
+
"3": 143495589,
|
| 125 |
+
"4": 136073789,
|
| 126 |
+
"5": 135460995,
|
| 127 |
+
"6": 133884692,
|
| 128 |
+
"7": 135525775
|
| 129 |
+
},
|
| 130 |
+
"decoder_layer_0": {
|
| 131 |
+
"0": 11885357,
|
| 132 |
+
"1": 11801320,
|
| 133 |
+
"2": 13060220,
|
| 134 |
+
"3": 12342825,
|
| 135 |
+
"4": 12361295,
|
| 136 |
+
"5": 13069915,
|
| 137 |
+
"6": 11787056,
|
| 138 |
+
"7": 13554748
|
| 139 |
+
},
|
| 140 |
+
"decoder_layer_1": {
|
| 141 |
+
"0": 11862357,
|
| 142 |
+
"1": 12259029,
|
| 143 |
+
"2": 11620687,
|
| 144 |
+
"3": 14837367,
|
| 145 |
+
"4": 11779914,
|
| 146 |
+
"5": 13099878,
|
| 147 |
+
"6": 11683397,
|
| 148 |
+
"7": 12720107
|
| 149 |
+
},
|
| 150 |
+
"decoder_layer_2": {
|
| 151 |
+
"0": 12201210,
|
| 152 |
+
"1": 12112379,
|
| 153 |
+
"2": 14274767,
|
| 154 |
+
"3": 12285646,
|
| 155 |
+
"4": 12123051,
|
| 156 |
+
"5": 12171337,
|
| 157 |
+
"6": 12135313,
|
| 158 |
+
"7": 12559033
|
| 159 |
+
},
|
| 160 |
+
"decoder_layer_3": {
|
| 161 |
+
"0": 12632673,
|
| 162 |
+
"1": 15462911,
|
| 163 |
+
"2": 11907600,
|
| 164 |
+
"3": 12328118,
|
| 165 |
+
"4": 11866897,
|
| 166 |
+
"5": 11858883,
|
| 167 |
+
"6": 11895272,
|
| 168 |
+
"7": 11910382
|
| 169 |
+
}
|
| 170 |
+
}
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 3,
|
| 174 |
+
"usage": {
|
| 175 |
+
"encoder_layer_0": {
|
| 176 |
+
"0": 140993937,
|
| 177 |
+
"1": 135712261,
|
| 178 |
+
"2": 139518919,
|
| 179 |
+
"3": 139655991,
|
| 180 |
+
"4": 136781290,
|
| 181 |
+
"5": 137254831,
|
| 182 |
+
"6": 135476784,
|
| 183 |
+
"7": 137302659
|
| 184 |
+
},
|
| 185 |
+
"encoder_layer_1": {
|
| 186 |
+
"0": 153395951,
|
| 187 |
+
"1": 155997088,
|
| 188 |
+
"2": 135728108,
|
| 189 |
+
"3": 115885780,
|
| 190 |
+
"4": 147534633,
|
| 191 |
+
"5": 113059645,
|
| 192 |
+
"6": 139938622,
|
| 193 |
+
"7": 141156845
|
| 194 |
+
},
|
| 195 |
+
"encoder_layer_2": {
|
| 196 |
+
"0": 135360959,
|
| 197 |
+
"1": 143571480,
|
| 198 |
+
"2": 130426240,
|
| 199 |
+
"3": 138121467,
|
| 200 |
+
"4": 143331476,
|
| 201 |
+
"5": 143239623,
|
| 202 |
+
"6": 143669414,
|
| 203 |
+
"7": 124976013
|
| 204 |
+
},
|
| 205 |
+
"encoder_layer_3": {
|
| 206 |
+
"0": 138108424,
|
| 207 |
+
"1": 143624638,
|
| 208 |
+
"2": 136522770,
|
| 209 |
+
"3": 143495589,
|
| 210 |
+
"4": 136073789,
|
| 211 |
+
"5": 135460995,
|
| 212 |
+
"6": 133884692,
|
| 213 |
+
"7": 135525775
|
| 214 |
+
},
|
| 215 |
+
"decoder_layer_0": {
|
| 216 |
+
"0": 11885357,
|
| 217 |
+
"1": 11801320,
|
| 218 |
+
"2": 13060220,
|
| 219 |
+
"3": 12342825,
|
| 220 |
+
"4": 12361295,
|
| 221 |
+
"5": 13069915,
|
| 222 |
+
"6": 11787056,
|
| 223 |
+
"7": 13554748
|
| 224 |
+
},
|
| 225 |
+
"decoder_layer_1": {
|
| 226 |
+
"0": 11862357,
|
| 227 |
+
"1": 12259029,
|
| 228 |
+
"2": 11620687,
|
| 229 |
+
"3": 14837367,
|
| 230 |
+
"4": 11779914,
|
| 231 |
+
"5": 13099878,
|
| 232 |
+
"6": 11683397,
|
| 233 |
+
"7": 12720107
|
| 234 |
+
},
|
| 235 |
+
"decoder_layer_2": {
|
| 236 |
+
"0": 12201210,
|
| 237 |
+
"1": 12112379,
|
| 238 |
+
"2": 14274767,
|
| 239 |
+
"3": 12285646,
|
| 240 |
+
"4": 12123051,
|
| 241 |
+
"5": 12171337,
|
| 242 |
+
"6": 12135313,
|
| 243 |
+
"7": 12559033
|
| 244 |
+
},
|
| 245 |
+
"decoder_layer_3": {
|
| 246 |
+
"0": 12632673,
|
| 247 |
+
"1": 15462911,
|
| 248 |
+
"2": 11907600,
|
| 249 |
+
"3": 12328118,
|
| 250 |
+
"4": 11866897,
|
| 251 |
+
"5": 11858883,
|
| 252 |
+
"6": 11895272,
|
| 253 |
+
"7": 11910382
|
| 254 |
+
}
|
| 255 |
+
}
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 4,
|
| 259 |
+
"usage": {
|
| 260 |
+
"encoder_layer_0": {
|
| 261 |
+
"0": 140993937,
|
| 262 |
+
"1": 135712261,
|
| 263 |
+
"2": 139518919,
|
| 264 |
+
"3": 139655991,
|
| 265 |
+
"4": 136781290,
|
| 266 |
+
"5": 137254831,
|
| 267 |
+
"6": 135476784,
|
| 268 |
+
"7": 137302659
|
| 269 |
+
},
|
| 270 |
+
"encoder_layer_1": {
|
| 271 |
+
"0": 153395951,
|
| 272 |
+
"1": 155997088,
|
| 273 |
+
"2": 135728108,
|
| 274 |
+
"3": 115885780,
|
| 275 |
+
"4": 147534633,
|
| 276 |
+
"5": 113059645,
|
| 277 |
+
"6": 139938622,
|
| 278 |
+
"7": 141156845
|
| 279 |
+
},
|
| 280 |
+
"encoder_layer_2": {
|
| 281 |
+
"0": 135360959,
|
| 282 |
+
"1": 143571480,
|
| 283 |
+
"2": 130426240,
|
| 284 |
+
"3": 138121467,
|
| 285 |
+
"4": 143331476,
|
| 286 |
+
"5": 143239623,
|
| 287 |
+
"6": 143669414,
|
| 288 |
+
"7": 124976013
|
| 289 |
+
},
|
| 290 |
+
"encoder_layer_3": {
|
| 291 |
+
"0": 138108424,
|
| 292 |
+
"1": 143624638,
|
| 293 |
+
"2": 136522770,
|
| 294 |
+
"3": 143495589,
|
| 295 |
+
"4": 136073789,
|
| 296 |
+
"5": 135460995,
|
| 297 |
+
"6": 133884692,
|
| 298 |
+
"7": 135525775
|
| 299 |
+
},
|
| 300 |
+
"decoder_layer_0": {
|
| 301 |
+
"0": 11885357,
|
| 302 |
+
"1": 11801320,
|
| 303 |
+
"2": 13060220,
|
| 304 |
+
"3": 12342825,
|
| 305 |
+
"4": 12361295,
|
| 306 |
+
"5": 13069915,
|
| 307 |
+
"6": 11787056,
|
| 308 |
+
"7": 13554748
|
| 309 |
+
},
|
| 310 |
+
"decoder_layer_1": {
|
| 311 |
+
"0": 11862357,
|
| 312 |
+
"1": 12259029,
|
| 313 |
+
"2": 11620687,
|
| 314 |
+
"3": 14837367,
|
| 315 |
+
"4": 11779914,
|
| 316 |
+
"5": 13099878,
|
| 317 |
+
"6": 11683397,
|
| 318 |
+
"7": 12720107
|
| 319 |
+
},
|
| 320 |
+
"decoder_layer_2": {
|
| 321 |
+
"0": 12201210,
|
| 322 |
+
"1": 12112379,
|
| 323 |
+
"2": 14274767,
|
| 324 |
+
"3": 12285646,
|
| 325 |
+
"4": 12123051,
|
| 326 |
+
"5": 12171337,
|
| 327 |
+
"6": 12135313,
|
| 328 |
+
"7": 12559033
|
| 329 |
+
},
|
| 330 |
+
"decoder_layer_3": {
|
| 331 |
+
"0": 12632673,
|
| 332 |
+
"1": 15462911,
|
| 333 |
+
"2": 11907600,
|
| 334 |
+
"3": 12328118,
|
| 335 |
+
"4": 11866897,
|
| 336 |
+
"5": 11858883,
|
| 337 |
+
"6": 11895272,
|
| 338 |
+
"7": 11910382
|
| 339 |
+
}
|
| 340 |
+
}
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 5,
|
| 344 |
+
"usage": {
|
| 345 |
+
"encoder_layer_0": {
|
| 346 |
+
"0": 140993937,
|
| 347 |
+
"1": 135712261,
|
| 348 |
+
"2": 139518919,
|
| 349 |
+
"3": 139655991,
|
| 350 |
+
"4": 136781290,
|
| 351 |
+
"5": 137254831,
|
| 352 |
+
"6": 135476784,
|
| 353 |
+
"7": 137302659
|
| 354 |
+
},
|
| 355 |
+
"encoder_layer_1": {
|
| 356 |
+
"0": 153395951,
|
| 357 |
+
"1": 155997088,
|
| 358 |
+
"2": 135728108,
|
| 359 |
+
"3": 115885780,
|
| 360 |
+
"4": 147534633,
|
| 361 |
+
"5": 113059645,
|
| 362 |
+
"6": 139938622,
|
| 363 |
+
"7": 141156845
|
| 364 |
+
},
|
| 365 |
+
"encoder_layer_2": {
|
| 366 |
+
"0": 135360959,
|
| 367 |
+
"1": 143571480,
|
| 368 |
+
"2": 130426240,
|
| 369 |
+
"3": 138121467,
|
| 370 |
+
"4": 143331476,
|
| 371 |
+
"5": 143239623,
|
| 372 |
+
"6": 143669414,
|
| 373 |
+
"7": 124976013
|
| 374 |
+
},
|
| 375 |
+
"encoder_layer_3": {
|
| 376 |
+
"0": 138108424,
|
| 377 |
+
"1": 143624638,
|
| 378 |
+
"2": 136522770,
|
| 379 |
+
"3": 143495589,
|
| 380 |
+
"4": 136073789,
|
| 381 |
+
"5": 135460995,
|
| 382 |
+
"6": 133884692,
|
| 383 |
+
"7": 135525775
|
| 384 |
+
},
|
| 385 |
+
"decoder_layer_0": {
|
| 386 |
+
"0": 11885357,
|
| 387 |
+
"1": 11801320,
|
| 388 |
+
"2": 13060220,
|
| 389 |
+
"3": 12342825,
|
| 390 |
+
"4": 12361295,
|
| 391 |
+
"5": 13069915,
|
| 392 |
+
"6": 11787056,
|
| 393 |
+
"7": 13554748
|
| 394 |
+
},
|
| 395 |
+
"decoder_layer_1": {
|
| 396 |
+
"0": 11862357,
|
| 397 |
+
"1": 12259029,
|
| 398 |
+
"2": 11620687,
|
| 399 |
+
"3": 14837367,
|
| 400 |
+
"4": 11779914,
|
| 401 |
+
"5": 13099878,
|
| 402 |
+
"6": 11683397,
|
| 403 |
+
"7": 12720107
|
| 404 |
+
},
|
| 405 |
+
"decoder_layer_2": {
|
| 406 |
+
"0": 12201210,
|
| 407 |
+
"1": 12112379,
|
| 408 |
+
"2": 14274767,
|
| 409 |
+
"3": 12285646,
|
| 410 |
+
"4": 12123051,
|
| 411 |
+
"5": 12171337,
|
| 412 |
+
"6": 12135313,
|
| 413 |
+
"7": 12559033
|
| 414 |
+
},
|
| 415 |
+
"decoder_layer_3": {
|
| 416 |
+
"0": 12632673,
|
| 417 |
+
"1": 15462911,
|
| 418 |
+
"2": 11907600,
|
| 419 |
+
"3": 12328118,
|
| 420 |
+
"4": 11866897,
|
| 421 |
+
"5": 11858883,
|
| 422 |
+
"6": 11895272,
|
| 423 |
+
"7": 11910382
|
| 424 |
+
}
|
| 425 |
+
}
|
| 426 |
+
}
|
| 427 |
+
]
|
results/moe_token_choice/moe_token_choice_test_predictions.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aebe85737a13dc11530e579868d68fa7fd58f6f8d6cebea9fde02914242db6f4
|
| 3 |
+
size 27675266
|
results/moe_token_choice/training_curves.png
ADDED
|
Git LFS Details
|
results/moe_token_choice_with_lb/best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fafe0d552784dc65ed6caa6870f8689c28b465d177ef3d7daf596c1fed171cd
|
| 3 |
+
size 2159617603
|
results/moe_token_choice_with_lb/training_metrics.json
ADDED
|
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"train_losses": [
|
| 3 |
+
7.0997042610168455,
|
| 4 |
+
6.098364649200439,
|
| 5 |
+
5.656473810577393,
|
| 6 |
+
5.355114995574951,
|
| 7 |
+
5.108737034606934
|
| 8 |
+
],
|
| 9 |
+
"train_nll_losses": [
|
| 10 |
+
7.0122965110778805,
|
| 11 |
+
6.014897843170166,
|
| 12 |
+
5.573608850860595,
|
| 13 |
+
5.272693781280518,
|
| 14 |
+
5.026728285980225
|
| 15 |
+
],
|
| 16 |
+
"train_lb_losses": [
|
| 17 |
+
0.08740775455236435,
|
| 18 |
+
0.08346681421995163,
|
| 19 |
+
0.08286495605707168,
|
| 20 |
+
0.08242120879888534,
|
| 21 |
+
0.08200874214172363
|
| 22 |
+
],
|
| 23 |
+
"val_losses": [
|
| 24 |
+
6.270021446167477,
|
| 25 |
+
5.788443149082244,
|
| 26 |
+
5.507700261615572,
|
| 27 |
+
5.303510824839274,
|
| 28 |
+
5.156257523430718
|
| 29 |
+
],
|
| 30 |
+
"use_load_balancer": true,
|
| 31 |
+
"expert_balance_history": [
|
| 32 |
+
{
|
| 33 |
+
"epoch": 1,
|
| 34 |
+
"cv_scores": {
|
| 35 |
+
"encoder_layer_0": 0.02239848952710913,
|
| 36 |
+
"encoder_layer_1": 0.011894839454450919,
|
| 37 |
+
"encoder_layer_2": 0.014883410628884161,
|
| 38 |
+
"encoder_layer_3": 0.03680761393831419,
|
| 39 |
+
"decoder_layer_0": 0.07264102945333789,
|
| 40 |
+
"decoder_layer_1": 0.09091776210725255,
|
| 41 |
+
"decoder_layer_2": 0.07472949755270192,
|
| 42 |
+
"decoder_layer_3": 0.10471215665870788
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 2,
|
| 47 |
+
"cv_scores": {
|
| 48 |
+
"encoder_layer_0": 0.02293213506202102,
|
| 49 |
+
"encoder_layer_1": 0.01635183137600546,
|
| 50 |
+
"encoder_layer_2": 0.014198928648696423,
|
| 51 |
+
"encoder_layer_3": 0.024659876381341785,
|
| 52 |
+
"decoder_layer_0": 0.05170881803637949,
|
| 53 |
+
"decoder_layer_1": 0.058714955010434354,
|
| 54 |
+
"decoder_layer_2": 0.08058705108512597,
|
| 55 |
+
"decoder_layer_3": 0.051674624864711804
|
| 56 |
+
}
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 3,
|
| 60 |
+
"cv_scores": {
|
| 61 |
+
"encoder_layer_0": 0.01790022341165699,
|
| 62 |
+
"encoder_layer_1": 0.03365750521155693,
|
| 63 |
+
"encoder_layer_2": 0.020147216041616556,
|
| 64 |
+
"encoder_layer_3": 0.028766130854617978,
|
| 65 |
+
"decoder_layer_0": 0.06359621508791356,
|
| 66 |
+
"decoder_layer_1": 0.0664411660102753,
|
| 67 |
+
"decoder_layer_2": 0.08864758361507494,
|
| 68 |
+
"decoder_layer_3": 0.030620321433448113
|
| 69 |
+
}
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 4,
|
| 73 |
+
"cv_scores": {
|
| 74 |
+
"encoder_layer_0": 0.016398893255619877,
|
| 75 |
+
"encoder_layer_1": 0.03261250392944865,
|
| 76 |
+
"encoder_layer_2": 0.02526192742709379,
|
| 77 |
+
"encoder_layer_3": 0.03652289327142253,
|
| 78 |
+
"decoder_layer_0": 0.0825089900217896,
|
| 79 |
+
"decoder_layer_1": 0.07279900493183922,
|
| 80 |
+
"decoder_layer_2": 0.09745846649526438,
|
| 81 |
+
"decoder_layer_3": 0.0278281985663954
|
| 82 |
+
}
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 5,
|
| 86 |
+
"cv_scores": {
|
| 87 |
+
"encoder_layer_0": 0.015321057262938304,
|
| 88 |
+
"encoder_layer_1": 0.027677770571459974,
|
| 89 |
+
"encoder_layer_2": 0.036855431757315545,
|
| 90 |
+
"encoder_layer_3": 0.03604612322392209,
|
| 91 |
+
"decoder_layer_0": 0.09712146666634115,
|
| 92 |
+
"decoder_layer_1": 0.07458012114520648,
|
| 93 |
+
"decoder_layer_2": 0.10375502366450497,
|
| 94 |
+
"decoder_layer_3": 0.027890078551332173
|
| 95 |
+
}
|
| 96 |
+
}
|
| 97 |
+
]
|
| 98 |
+
}
|
results/moe_token_choice_without_lb/best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cdb1c7a7e405d1836d99ee7a323c6a17a1d44dd810e89ce3b31aeddaed1dbdb6
|
| 3 |
+
size 2159617603
|
results/moe_token_choice_without_lb/training_metrics.json
ADDED
|
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"train_losses": [
|
| 3 |
+
6.9957088943481445,
|
| 4 |
+
5.946418069458008,
|
| 5 |
+
5.539425620269776,
|
| 6 |
+
5.257929954528809,
|
| 7 |
+
5.025598110198975
|
| 8 |
+
],
|
| 9 |
+
"train_nll_losses": [
|
| 10 |
+
6.9957088943481445,
|
| 11 |
+
5.946418069458008,
|
| 12 |
+
5.539425620269776,
|
| 13 |
+
5.257929954528809,
|
| 14 |
+
5.025598110198975
|
| 15 |
+
],
|
| 16 |
+
"train_lb_losses": [
|
| 17 |
+
0.0,
|
| 18 |
+
0.0,
|
| 19 |
+
0.0,
|
| 20 |
+
0.0,
|
| 21 |
+
0.0
|
| 22 |
+
],
|
| 23 |
+
"val_losses": [
|
| 24 |
+
6.208024940793476,
|
| 25 |
+
5.756550130389986,
|
| 26 |
+
5.503833710201203,
|
| 27 |
+
5.309833534180172,
|
| 28 |
+
5.157609190259661
|
| 29 |
+
],
|
| 30 |
+
"use_load_balancer": false,
|
| 31 |
+
"expert_balance_history": [
|
| 32 |
+
{
|
| 33 |
+
"epoch": 1,
|
| 34 |
+
"cv_scores": {
|
| 35 |
+
"encoder_layer_0": 0.47773956186164035,
|
| 36 |
+
"encoder_layer_1": 1.2283993740562353,
|
| 37 |
+
"encoder_layer_2": 1.2066402900273532,
|
| 38 |
+
"encoder_layer_3": 1.1506108166570812,
|
| 39 |
+
"decoder_layer_0": 1.1311583795369742,
|
| 40 |
+
"decoder_layer_1": 1.031120870253164,
|
| 41 |
+
"decoder_layer_2": 1.2149890152604113,
|
| 42 |
+
"decoder_layer_3": 1.359569701461952
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 2,
|
| 47 |
+
"cv_scores": {
|
| 48 |
+
"encoder_layer_0": 0.5317931290413349,
|
| 49 |
+
"encoder_layer_1": 1.1592380871171228,
|
| 50 |
+
"encoder_layer_2": 1.048192812774701,
|
| 51 |
+
"encoder_layer_3": 1.0848774885298542,
|
| 52 |
+
"decoder_layer_0": 1.2000397098222888,
|
| 53 |
+
"decoder_layer_1": 0.973016538594419,
|
| 54 |
+
"decoder_layer_2": 1.2892343882605066,
|
| 55 |
+
"decoder_layer_3": 1.272053691280787
|
| 56 |
+
}
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 3,
|
| 60 |
+
"cv_scores": {
|
| 61 |
+
"encoder_layer_0": 0.5360830744020194,
|
| 62 |
+
"encoder_layer_1": 1.1426186109630772,
|
| 63 |
+
"encoder_layer_2": 0.9655670214289462,
|
| 64 |
+
"encoder_layer_3": 1.0901768668859588,
|
| 65 |
+
"decoder_layer_0": 1.2455368246203251,
|
| 66 |
+
"decoder_layer_1": 0.9588685599967219,
|
| 67 |
+
"decoder_layer_2": 1.3105986019784088,
|
| 68 |
+
"decoder_layer_3": 1.2299858663796666
|
| 69 |
+
}
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 4,
|
| 73 |
+
"cv_scores": {
|
| 74 |
+
"encoder_layer_0": 0.5186909345945561,
|
| 75 |
+
"encoder_layer_1": 1.1361607950311845,
|
| 76 |
+
"encoder_layer_2": 0.9141624769231046,
|
| 77 |
+
"encoder_layer_3": 1.1254102549849538,
|
| 78 |
+
"decoder_layer_0": 1.2753951147459726,
|
| 79 |
+
"decoder_layer_1": 0.9487367217372488,
|
| 80 |
+
"decoder_layer_2": 1.3281710137430134,
|
| 81 |
+
"decoder_layer_3": 1.205282174955813
|
| 82 |
+
}
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 5,
|
| 86 |
+
"cv_scores": {
|
| 87 |
+
"encoder_layer_0": 0.5177979354789254,
|
| 88 |
+
"encoder_layer_1": 1.1347360622293354,
|
| 89 |
+
"encoder_layer_2": 0.8721613047751007,
|
| 90 |
+
"encoder_layer_3": 1.1427533114030508,
|
| 91 |
+
"decoder_layer_0": 1.2888345008752309,
|
| 92 |
+
"decoder_layer_1": 0.9444531131949591,
|
| 93 |
+
"decoder_layer_2": 1.3459587847130505,
|
| 94 |
+
"decoder_layer_3": 1.18326321074551
|
| 95 |
+
}
|
| 96 |
+
}
|
| 97 |
+
]
|
| 98 |
+
}
|
results/standard_mha_moe/best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fdbdbfa8c37f4df106382200b2501689f11a215806c88cfdd7eb1b6b254e1a34
|
| 3 |
+
size 720563261
|