gloamxun/modeling_zhaocui
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Train ATI-HO and finalize project outputs

Browse Source
This commit is contained in:
gloamxun
2026-09-26 16:05:44 +08:00
parent a86560da64
commit 9cdd604117
358 changed files with 10540 additions and 173 deletions
Download Patch File Download Diff File Expand all files Collapse all files
final/experiments/q3/ati_ho/models/A2/seed_2026/model_best.pt
BIN
View File
Binary file not shown.
final/experiments/q3/ati_ho/models/A2/seed_2026/training_history.csv
+7
View File
@@ -0,0 +1,7 @@
method,seed,epoch,train_loss,valid_selection_loss,valid_clean_loss,lambda_interaction,lambda_mask
A2,2026,1,0.9926471710205078,0.928315669953168,0.9239131356333639,0.001,0.0
A2,2026,2,0.820087315859618,0.8673638532777409,0.8601592170013176,0.001,0.0
A2,2026,3,0.7308831590193289,0.862087192279952,0.8561929890087673,0.001,0.0
A2,2026,4,0.6851603929643277,0.8666458973190287,0.8619590087251349,0.001,0.0
A2,2026,5,0.6444870498445299,0.88101957656525,0.8752215864894154,0.001,0.0
A2,2026,6,0.5834257365376861,0.9354289185542327,0.9353187752294017,0.001,0.0
1 method seed epoch train_loss valid_selection_loss valid_clean_loss lambda_interaction lambda_mask
2 A2 2026 1 0.9926471710205078 0.928315669953168 0.9239131356333639 0.001 0.0
3 A2 2026 2 0.820087315859618 0.8673638532777409 0.8601592170013176 0.001 0.0
4 A2 2026 3 0.7308831590193289 0.862087192279952 0.8561929890087673 0.001 0.0
5 A2 2026 4 0.6851603929643277 0.8666458973190287 0.8619590087251349 0.001 0.0
6 A2 2026 5 0.6444870498445299 0.88101957656525 0.8752215864894154 0.001 0.0
7 A2 2026 6 0.5834257365376861 0.9354289185542327 0.9353187752294017 0.001 0.0
final/experiments/q3/ati_ho/models/A2/seed_2026/training_manifest.json
+127
View File
@@ -0,0 +1,127 @@
{
"method": "A2",
"seed": 2026,
"best_epoch": 3,
"best_selection_loss": 0.862087192279952,
"elapsed_seconds": 9.472242687013932,
"batch_size": 64,
"epoch_limit": 12,
"patience": 3,
"optimizer": "AdamW",
"learning_rate": 0.0003,
"weight_decay": 0.001,
"gradient_clip_norm": 1.0,
"training_mask_rates": [
0.0,
0.1,
0.3,
0.5,
0.7
],
"training_mask_patterns": [
"single",
"sync",
"partial",
"async"
],
"training_mask_seed_base": 20261227,
"same_orders_and_masks_across_methods_for_same_seed": true,
"config": {
"name": "A2_anchored_pairwise",
"low_rank": true,
"cross_attention": true,
"anchored": true,
"lambda_interaction": 0.001,
"lambda_mask": 0.0,
"rank": 4,
"hidden": 64,
"gru_hidden_per_direction": 32,
"attention_heads": 4,
"attention_ffn": 128,
"eta_init": 0.1
},
"history": [
{
"method": "A2",
"seed": 2026,
"epoch": 1,
"train_loss": 0.9926471710205078,
"valid_selection_loss": 0.928315669953168,
"valid_clean_loss": 0.9239131356333639,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 2026,
"epoch": 2,
"train_loss": 0.820087315859618,
"valid_selection_loss": 0.8673638532777409,
"valid_clean_loss": 0.8601592170013176,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 2026,
"epoch": 3,
"train_loss": 0.7308831590193289,
"valid_selection_loss": 0.862087192279952,
"valid_clean_loss": 0.8561929890087673,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 2026,
"epoch": 4,
"train_loss": 0.6851603929643277,
"valid_selection_loss": 0.8666458973190287,
"valid_clean_loss": 0.8619590087251349,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 2026,
"epoch": 5,
"train_loss": 0.6444870498445299,
"valid_selection_loss": 0.88101957656525,
"valid_clean_loss": 0.8752215864894154,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 2026,
"epoch": 6,
"train_loss": 0.5834257365376861,
"valid_selection_loss": 0.9354289185542327,
"valid_clean_loss": 0.9353187752294017,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
}
],
"mask_counts": {
"0.1/async": 970,
"0.0/async": 995,
"0.5/sync": 1040,
"0.0/sync": 1012,
"0.5/single": 999,
"0.0/single": 1046,
"0.7/async": 1048,
"0.0/partial": 1063,
"0.7/sync": 1020,
"0.5/partial": 1022,
"0.3/partial": 1002,
"0.1/partial": 1022,
"0.3/single": 1004,
"0.3/async": 1026,
"0.5/async": 1028,
"0.1/single": 1048,
"0.7/single": 959,
"0.3/sync": 1030,
"0.1/sync": 1007,
"0.7/partial": 1029
}
}
final/experiments/q3/ati_ho/models/A2/seed_3407/model_best.pt
BIN
View File
Binary file not shown.
final/experiments/q3/ati_ho/models/A2/seed_3407/training_history.csv
+7
View File
@@ -0,0 +1,7 @@
method,seed,epoch,train_loss,valid_selection_loss,valid_clean_loss,lambda_interaction,lambda_mask
A2,3407,1,0.9647035289693762,0.9118595260840197,0.9059003676686969,0.001,0.0
A2,3407,2,0.8049548080673924,0.8821067369573719,0.8777574326965835,0.001,0.0
A2,3407,3,0.7312485785396011,0.880271397449158,0.8748716823347322,0.001,0.0
A2,3407,4,0.6844777796003554,0.88496050123985,0.881083797622513,0.001,0.0
A2,3407,5,0.6395798369690224,0.9139724308317835,0.9094404204861148,0.001,0.0
A2,3407,6,0.5926711849040456,0.9307978958873957,0.9263341426849365,0.001,0.0
1 method seed epoch train_loss valid_selection_loss valid_clean_loss lambda_interaction lambda_mask
2 A2 3407 1 0.9647035289693762 0.9118595260840197 0.9059003676686969 0.001 0.0
3 A2 3407 2 0.8049548080673924 0.8821067369573719 0.8777574326965835 0.001 0.0
4 A2 3407 3 0.7312485785396011 0.880271397449158 0.8748716823347322 0.001 0.0
5 A2 3407 4 0.6844777796003554 0.88496050123985 0.881083797622513 0.001 0.0
6 A2 3407 5 0.6395798369690224 0.9139724308317835 0.9094404204861148 0.001 0.0
7 A2 3407 6 0.5926711849040456 0.9307978958873957 0.9263341426849365 0.001 0.0
final/experiments/q3/ati_ho/models/A2/seed_3407/training_manifest.json
+127
View File
@@ -0,0 +1,127 @@
{
"method": "A2",
"seed": 3407,
"best_epoch": 3,
"best_selection_loss": 0.880271397449158,
"elapsed_seconds": 9.558659436006565,
"batch_size": 64,
"epoch_limit": 12,
"patience": 3,
"optimizer": "AdamW",
"learning_rate": 0.0003,
"weight_decay": 0.001,
"gradient_clip_norm": 1.0,
"training_mask_rates": [
0.0,
0.1,
0.3,
0.5,
0.7
],
"training_mask_patterns": [
"single",
"sync",
"partial",
"async"
],
"training_mask_seed_base": 20261227,
"same_orders_and_masks_across_methods_for_same_seed": true,
"config": {
"name": "A2_anchored_pairwise",
"low_rank": true,
"cross_attention": true,
"anchored": true,
"lambda_interaction": 0.001,
"lambda_mask": 0.0,
"rank": 4,
"hidden": 64,
"gru_hidden_per_direction": 32,
"attention_heads": 4,
"attention_ffn": 128,
"eta_init": 0.1
},
"history": [
{
"method": "A2",
"seed": 3407,
"epoch": 1,
"train_loss": 0.9647035289693762,
"valid_selection_loss": 0.9118595260840197,
"valid_clean_loss": 0.9059003676686969,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 3407,
"epoch": 2,
"train_loss": 0.8049548080673924,
"valid_selection_loss": 0.8821067369573719,
"valid_clean_loss": 0.8777574326965835,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 3407,
"epoch": 3,
"train_loss": 0.7312485785396011,
"valid_selection_loss": 0.880271397449158,
"valid_clean_loss": 0.8748716823347322,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 3407,
"epoch": 4,
"train_loss": 0.6844777796003554,
"valid_selection_loss": 0.88496050123985,
"valid_clean_loss": 0.881083797622513,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 3407,
"epoch": 5,
"train_loss": 0.6395798369690224,
"valid_selection_loss": 0.9139724308317835,
"valid_clean_loss": 0.9094404204861148,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 3407,
"epoch": 6,
"train_loss": 0.5926711849040456,
"valid_selection_loss": 0.9307978958873957,
"valid_clean_loss": 0.9263341426849365,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
}
],
"mask_counts": {
"0.7/async": 1003,
"0.0/async": 1014,
"0.5/sync": 1033,
"0.1/sync": 1024,
"0.5/partial": 996,
"0.3/partial": 1019,
"0.5/async": 1053,
"0.3/single": 1003,
"0.0/sync": 1042,
"0.7/partial": 1079,
"0.7/sync": 986,
"0.0/single": 1029,
"0.1/async": 959,
"0.3/sync": 1055,
"0.0/partial": 1016,
"0.1/single": 1011,
"0.1/partial": 988,
"0.5/single": 1076,
"0.7/single": 997,
"0.3/async": 987
}
}
final/experiments/q3/ati_ho/models/A2/seed_42/model_best.pt
BIN
View File
Binary file not shown.
final/experiments/q3/ati_ho/models/A2/seed_42/training_history.csv
+6
View File
@@ -0,0 +1,6 @@
method,seed,epoch,train_loss,valid_selection_loss,valid_clean_loss,lambda_interaction,lambda_mask
A2,42,1,0.9741679154060505,0.9116741605512388,0.9052737184933254,0.001,0.0
A2,42,2,0.7997706675970996,0.8634063961741689,0.8530163961452443,0.001,0.0
A2,42,3,0.7264242089456983,0.8717996972602802,0.8635974290606739,0.001,0.0
A2,42,4,0.6684221158976908,0.8713849653581996,0.8629672671412374,0.001,0.0
A2,42,5,0.6371411096166681,0.8887755983805918,0.8809175399633554,0.001,0.0
1 method seed epoch train_loss valid_selection_loss valid_clean_loss lambda_interaction lambda_mask
2 A2 42 1 0.9741679154060505 0.9116741605512388 0.9052737184933254 0.001 0.0
3 A2 42 2 0.7997706675970996 0.8634063961741689 0.8530163961452443 0.001 0.0
4 A2 42 3 0.7264242089456983 0.8717996972602802 0.8635974290606739 0.001 0.0
5 A2 42 4 0.6684221158976908 0.8713849653581996 0.8629672671412374 0.001 0.0
6 A2 42 5 0.6371411096166681 0.8887755983805918 0.8809175399633554 0.001 0.0
final/experiments/q3/ati_ho/models/A2/seed_42/training_manifest.json
+117
View File
@@ -0,0 +1,117 @@
{
"method": "A2",
"seed": 42,
"best_epoch": 2,
"best_selection_loss": 0.8634063961741689,
"elapsed_seconds": 8.171100930019747,
"batch_size": 64,
"epoch_limit": 12,
"patience": 3,
"optimizer": "AdamW",
"learning_rate": 0.0003,
"weight_decay": 0.001,
"gradient_clip_norm": 1.0,
"training_mask_rates": [
0.0,
0.1,
0.3,
0.5,
0.7
],
"training_mask_patterns": [
"single",
"sync",
"partial",
"async"
],
"training_mask_seed_base": 20261227,
"same_orders_and_masks_across_methods_for_same_seed": true,
"config": {
"name": "A2_anchored_pairwise",
"low_rank": true,
"cross_attention": true,
"anchored": true,
"lambda_interaction": 0.001,
"lambda_mask": 0.0,
"rank": 4,
"hidden": 64,
"gru_hidden_per_direction": 32,
"attention_heads": 4,
"attention_ffn": 128,
"eta_init": 0.1
},
"history": [
{
"method": "A2",
"seed": 42,
"epoch": 1,
"train_loss": 0.9741679154060505,
"valid_selection_loss": 0.9116741605512388,
"valid_clean_loss": 0.9052737184933254,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 42,
"epoch": 2,
"train_loss": 0.7997706675970996,
"valid_selection_loss": 0.8634063961741689,
"valid_clean_loss": 0.8530163961452443,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 42,
"epoch": 3,
"train_loss": 0.7264242089456983,
"valid_selection_loss": 0.8717996972602802,
"valid_clean_loss": 0.8635974290606739,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 42,
"epoch": 4,
"train_loss": 0.6684221158976908,
"valid_selection_loss": 0.8713849653581996,
"valid_clean_loss": 0.8629672671412374,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
},
{
"method": "A2",
"seed": 42,
"epoch": 5,
"train_loss": 0.6371411096166681,
"valid_selection_loss": 0.8887755983805918,
"valid_clean_loss": 0.8809175399633554,
"lambda_interaction": 0.001,
"lambda_mask": 0.0
}
],
"mask_counts": {
"0.0/single": 838,
"0.3/single": 841,
"0.5/single": 832,
"0.7/sync": 931,
"0.7/partial": 854,
"0.0/async": 818,
"0.3/async": 856,
"0.3/partial": 861,
"0.3/sync": 841,
"0.1/single": 857,
"0.5/async": 832,
"0.7/single": 872,
"0.5/sync": 848,
"0.5/partial": 854,
"0.0/partial": 870,
"0.0/sync": 827,
"0.7/async": 831,
"0.1/sync": 839,
"0.1/async": 847,
"0.1/partial": 826
}
}