Datasets:
ainewtrend07
/
HyperLora

Dataset card Data Studio Files
xet
 Community
1
Model_name
stringclasses
16 values
Train_size
int64
50.8k
50.8k
Test_size
int64
12.7k
12.7k
arg
dict
lora
listlengths
1
9
Parameters
int64
110M
1.85B
Trainable_parameters
int64
9.27k
1.11B
r
int64
4
1.02k
Memory Allocation
stringlengths
5
7
Training Time
stringlengths
5
7
Performance
dict
facebook/bart-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
410,893,325
3,538,944
16
3924.14
1758.92
{ "accuracy": 0.7686531773632628, "f1_macro": 0.6826876092376734, "f1_weighted": 0.7440488208213536, "precision": 0.7771219124685611, "recall": 0.7016688662589734 }
google/rembert
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in" ]
804,343,322
228,407,949
512
8302.35
3755.15
{ "accuracy": 0.9070502687322163, "f1_macro": 0.903753915884788, "f1_weighted": 0.9072142536440623, "precision": 0.9054038543476592, "recall": 0.9023628313733296 }
FacebookAI/xlm-roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
563,325,978
3,422,221
16
4948.37
1990.08
{ "accuracy": 0.8811255137527664, "f1_macro": 0.8754385267226975, "f1_weighted": 0.8812890144378421, "precision": 0.8772433829458713, "recall": 0.8739726112198498 }
microsoft/deberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "in_proj", "pos_proj" ]
434,681,882
28,455,949
64
4032.45
2008.74
{ "accuracy": 0.8946411634524186, "f1_macro": 0.8886334259764349, "f1_weighted": 0.8948719249496236, "precision": 0.8887474171563118, "recall": 0.8887215649100207 }
FacebookAI/roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
394,184,730
38,811,661
256
3707.79
1601.62
{ "accuracy": 0.8920328801770471, "f1_macro": 0.886170143990757, "f1_weighted": 0.8922916127906928, "precision": 0.8862698244878935, "recall": 0.8862980811626541 }
facebook/bart-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
414,432,269
7,077,888
32
3979.72
1774.11
{ "accuracy": 0.8195542206765729, "f1_macro": 0.7891727330039158, "f1_weighted": 0.8146717438958958, "precision": 0.8268035907289701, "recall": 0.7814600362290562 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in", "ffn", "ffn_output", "pooler" ]
223,194,650
545,805
8
2295.33
3284.2
{ "accuracy": 0.8721150806196649, "f1_macro": 0.8625622384926562, "f1_weighted": 0.8724111635668815, "precision": 0.8649331079858922, "recall": 0.8608322995407953 }
FacebookAI/roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
431,933,466
76,560,397
512
4278.11
1782.09
{ "accuracy": 0.8944830856781536, "f1_macro": 0.8892429447437229, "f1_weighted": 0.8947368747528309, "precision": 0.8903031569499813, "recall": 0.888506894876618 }
FacebookAI/xlm-roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
565,685,274
5,781,517
32
4986.55
1994.05
{ "accuracy": 0.885630730319317, "f1_macro": 0.8799731038459591, "f1_weighted": 0.8857962947130573, "precision": 0.8815299608315463, "recall": 0.8787330687950922 }
microsoft/deberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "in_proj", "pos_proj" ]
463,124,506
56,898,573
128
4453.11
2149.45
{ "accuracy": 0.898593107809042, "f1_macro": 0.8934656945521743, "f1_weighted": 0.8988183229515778, "precision": 0.8938509342926158, "recall": 0.8932909876195411 }
facebook/bart-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
421,510,157
14,155,776
64
4092.16
1810.7
{ "accuracy": 0.8457951312045526, "f1_macro": 0.8273407352655966, "f1_weighted": 0.8443636531931306, "precision": 0.8406379183344572, "recall": 0.821503174684913 }
FacebookAI/xlm-roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
570,403,866
10,500,109
64
5042.73
2022.7
{ "accuracy": 0.8900569079987354, "f1_macro": 0.8845784610114992, "f1_weighted": 0.8902482199996962, "precision": 0.8857819603436438, "recall": 0.8836843549118483 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in", "ffn", "ffn_output", "pooler" ]
223,687,194
1,038,349
16
2298.76
3289.64
{ "accuracy": 0.8799399304457793, "f1_macro": 0.8727232599677692, "f1_weighted": 0.8802950315159862, "precision": 0.8742590790729176, "recall": 0.871615438463911 }
facebook/bart-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
435,665,933
28,311,552
128
4345.52
1886.56
{ "accuracy": 0.865396775213405, "f1_macro": 0.853898616921064, "f1_weighted": 0.8656075941731403, "precision": 0.8586767053713357, "recall": 0.8506964250902089 }
microsoft/deberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "in_proj", "pos_proj" ]
520,009,754
113,783,821
256
5287.04
2419.67
{ "accuracy": 0.9042839076825798, "f1_macro": 0.9001127737292516, "f1_weighted": 0.9044466485600137, "precision": 0.9011130035298595, "recall": 0.899312944790197 }
FacebookAI/xlm-roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
579,841,050
19,937,293
128
5227.61
2077.36
{ "accuracy": 0.8927442301612394, "f1_macro": 0.8881118485485311, "f1_weighted": 0.8930036566962143, "precision": 0.8891157359783974, "recall": 0.8874910420884368 }
facebook/bart-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
463,977,485
56,623,104
256
4590.42
2000.55
{ "accuracy": 0.8740910527979766, "f1_macro": 0.8641543455754344, "f1_weighted": 0.8744270309406653, "precision": 0.8671219020252846, "recall": 0.8621327576250177 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in", "ffn", "ffn_output", "pooler" ]
224,672,282
2,023,437
32
2305.6
3300.51
{ "accuracy": 0.8845241858994626, "f1_macro": 0.8782377066234525, "f1_weighted": 0.8848242879873667, "precision": 0.8796871831463196, "recall": 0.8771529823096995 }
microsoft/deberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "in_proj", "pos_proj" ]
633,780,250
227,554,317
512
6643.75
2962.36
{ "accuracy": 0.9059437243123617, "f1_macro": 0.901780645397186, "f1_weighted": 0.906162578965779, "precision": 0.9025106939554568, "recall": 0.9013069138412005 }
FacebookAI/xlm-roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
598,715,418
38,811,661
256
5342.99
2159.44
{ "accuracy": 0.8952734745494784, "f1_macro": 0.8906239745753451, "f1_weighted": 0.8955239582373958, "precision": 0.8918392821684102, "recall": 0.8897941213726537 }
facebook/bart-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
520,600,589
113,246,208
512
5351.74
2288.81
{ "accuracy": 0.8827853303825483, "f1_macro": 0.8758976206384835, "f1_weighted": 0.8831363942020529, "precision": 0.8767459425958344, "recall": 0.8755781920595305 }
FacebookAI/xlm-roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
636,464,154
76,560,397
512
5912.53
2346.36
{ "accuracy": 0.8951944356623459, "f1_macro": 0.891139520661247, "f1_weighted": 0.8954635995605096, "precision": 0.8920068253407829, "recall": 0.8905737461685924 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in", "ffn", "ffn_output", "pooler" ]
226,642,458
3,993,613
64
2319.3
3337.84
{ "accuracy": 0.8865001580777743, "f1_macro": 0.880670620925541, "f1_weighted": 0.8867478256411928, "precision": 0.8822391815354093, "recall": 0.8794013826768659 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in", "ffn", "ffn_output", "pooler" ]
230,582,810
7,933,965
128
2350.51
3408.55
{ "accuracy": 0.8899778691116029, "f1_macro": 0.8848448827524312, "f1_weighted": 0.8902174678074721, "precision": 0.8863394957145604, "recall": 0.88363733472323 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in", "ffn", "ffn_output", "pooler" ]
238,463,514
15,814,669
256
2415.6
3558.94
{ "accuracy": 0.8917957635156497, "f1_macro": 0.8872782814934173, "f1_weighted": 0.8920497025520682, "precision": 0.8886148427960074, "recall": 0.8862449585163625 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in", "ffn", "ffn_output", "pooler" ]
254,224,922
31,576,077
512
2546.18
3884.1
{ "accuracy": 0.8944040467910211, "f1_macro": 0.8899140715503804, "f1_weighted": 0.8946827645200242, "precision": 0.89115674006335, "recall": 0.8889881039788747 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
222,800,410
151,565
4
2029.66
3173.24
{ "accuracy": 0.8938507745810939, "f1_macro": 0.8896909649855667, "f1_weighted": 0.8941282671690112, "precision": 0.8908889293413487, "recall": 0.8887894611225947 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
222,898,714
249,869
8
2030.47
3172.55
{ "accuracy": 0.8941669301296238, "f1_macro": 0.8900942642970121, "f1_weighted": 0.8944363517219318, "precision": 0.8912690464855278, "recall": 0.8892007397859174 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "o_proj", "score", "up_proj" ]
597,523,456
1,733,632
4
3583.02
1205.81
{ "accuracy": 0.8495099588997786, "f1_macro": 0.8385755546299981, "f1_weighted": 0.8495874874778647, "precision": 0.8407753759458068, "recall": 0.8370620776967219 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "project_in", "project_out", "score" ]
332,401,664
1,198,592
4
3086.03
1390.6
{ "accuracy": 0.8519601644008853, "f1_macro": 0.8400036572109179, "f1_weighted": 0.8516378694260406, "precision": 0.8444827896517915, "recall": 0.8369213404978963 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "o_proj", "score", "up_proj" ]
599,243,776
3,453,952
8
3590.06
1210.68
{ "accuracy": 0.8663452418589946, "f1_macro": 0.858056087907067, "f1_weighted": 0.8664719543153707, "precision": 0.8611149474260551, "recall": 0.8557566195015335 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "project_in", "project_out", "score" ]
333,593,600
2,390,528
8
3096.52
1390.79
{ "accuracy": 0.8686373695858363, "f1_macro": 0.8605217107156121, "f1_weighted": 0.8686929583044234, "precision": 0.8630679751708743, "recall": 0.8584899433864929 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
223,095,322
446,477
16
2032.09
3174.3
{ "accuracy": 0.8937717356939614, "f1_macro": 0.8896708427798876, "f1_weighted": 0.8940213875261503, "precision": 0.8907364675520443, "recall": 0.888859879333826 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "o_proj", "score", "up_proj" ]
602,684,416
6,894,592
16
3638.32
1215.16
{ "accuracy": 0.8763831805248182, "f1_macro": 0.8695910232338727, "f1_weighted": 0.8765158045910105, "precision": 0.8719863576954388, "recall": 0.8678224582977548 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "project_in", "project_out", "score" ]
335,977,472
4,774,400
16
3121.49
1403.21
{ "accuracy": 0.8774897249446728, "f1_macro": 0.8705801386855827, "f1_weighted": 0.8775923924095402, "precision": 0.872747132809748, "recall": 0.8688451210009968 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "score" ]
125,701,632
452,352
4
1378.46
540.28
{ "accuracy": 0.8248498261144483, "f1_macro": 0.8076317807889607, "f1_weighted": 0.8238920177941255, "precision": 0.8155232634631306, "recall": 0.8029339499264776 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "o_proj", "score", "up_proj" ]
609,565,696
13,775,872
32
3734.77
1225.1
{ "accuracy": 0.8814416693012962, "f1_macro": 0.8752011202330032, "f1_weighted": 0.8814977171575402, "precision": 0.8775422088417311, "recall": 0.8733521359668068 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "score" ]
126,144,000
894,720
8
1374.35
527.61
{ "accuracy": 0.8480082200442618, "f1_macro": 0.8358974282612031, "f1_weighted": 0.8479896358328798, "precision": 0.840470532955232, "recall": 0.8326994722784696 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "project_in", "project_out", "score" ]
340,745,216
9,542,144
32
3193.19
1419.09
{ "accuracy": 0.8827062914954158, "f1_macro": 0.8768934512723614, "f1_weighted": 0.8828352969251948, "precision": 0.8786607488305744, "recall": 0.875449475408163 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "score" ]
127,028,736
1,779,456
16
1379.39
528.11
{ "accuracy": 0.8598640531141322, "f1_macro": 0.8504201573049784, "f1_weighted": 0.8600125629724613, "precision": 0.8532928865727251, "recall": 0.8482820092821041 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "score" ]
128,798,208
3,548,928
32
1405.3
534.23
{ "accuracy": 0.8691906417957636, "f1_macro": 0.8613274370403433, "f1_weighted": 0.869311286892757, "precision": 0.8631030330748762, "recall": 0.8599313850021404 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
223,488,538
839,693
32
2035.35
3176.8
{ "accuracy": 0.893692696806829, "f1_macro": 0.8894913672557245, "f1_weighted": 0.8939279846498257, "precision": 0.8904198404404744, "recall": 0.8887944729795864 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "o_proj", "score", "up_proj" ]
623,328,256
27,538,432
64
3928.78
1265.37
{ "accuracy": 0.8849193803351248, "f1_macro": 0.8793954905478482, "f1_weighted": 0.8849972025838169, "precision": 0.8812942310679491, "recall": 0.8779293386177012 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "score" ]
132,337,152
7,087,872
64
1469.39
547.29
{ "accuracy": 0.8754347138792286, "f1_macro": 0.8687603756871548, "f1_weighted": 0.8756111701057591, "precision": 0.8702787186967382, "recall": 0.8675559912638823 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "project_in", "project_out", "score" ]
350,280,704
19,077,632
64
3363.41
1460.02
{ "accuracy": 0.8881599747075561, "f1_macro": 0.8827488878990528, "f1_weighted": 0.8883235800097365, "precision": 0.8844188895216042, "recall": 0.881344735939886 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "score" ]
139,415,040
14,165,760
128
1475.69
574.65
{ "accuracy": 0.8808093582042364, "f1_macro": 0.8757351777767126, "f1_weighted": 0.8809718632830494, "precision": 0.8764827510513497, "recall": 0.8751809439754545 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "o_proj", "score", "up_proj" ]
650,853,376
55,063,552
128
3947.86
1323.75
{ "accuracy": 0.8887132469174834, "f1_macro": 0.882947403290091, "f1_weighted": 0.8888142104408974, "precision": 0.8840665457325508, "recall": 0.8821097577938244 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "score" ]
153,570,816
28,321,536
256
1656.05
630.84
{ "accuracy": 0.8843661081251976, "f1_macro": 0.8796034740786574, "f1_weighted": 0.8845291873945885, "precision": 0.879804498832286, "recall": 0.8795587848073895 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "project_in", "project_out", "score" ]
369,351,680
38,148,608
128
3452.37
1554.8
{ "accuracy": 0.889661713563073, "f1_macro": 0.884396470196353, "f1_weighted": 0.8898252253051071, "precision": 0.8858316003016513, "recall": 0.8832202253882099 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "score" ]
181,882,368
56,633,088
512
1966.66
746.49
{ "accuracy": 0.8881599747075561, "f1_macro": 0.8834452097809463, "f1_weighted": 0.8883324625177675, "precision": 0.883980720092082, "recall": 0.8830864093213949 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "o_proj", "score", "up_proj" ]
705,903,616
110,113,792
256
4706.61
1448.96
{ "accuracy": 0.8921119190641795, "f1_macro": 0.886943525562743, "f1_weighted": 0.8922698960012461, "precision": 0.8876251319231722, "recall": 0.8864827669316555 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
224,274,970
1,626,125
64
2041.86
3199.69
{ "accuracy": 0.8936136579196965, "f1_macro": 0.8895526505830134, "f1_weighted": 0.8938358820835779, "precision": 0.8905472499747631, "recall": 0.8887723878326236 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "project_in", "project_out", "score" ]
407,493,632
76,290,560
256
3952.12
1722.39
{ "accuracy": 0.8914796079671198, "f1_macro": 0.8864860765885818, "f1_weighted": 0.8916172746962457, "precision": 0.8875755728399223, "recall": 0.8855774754757261 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
125,480,448
231,168
4
1167.78
538.37
{ "accuracy": 0.8099114764464116, "f1_macro": 0.7926030797288386, "f1_weighted": 0.8088421694140228, "precision": 0.8025224335938597, "recall": 0.7872884396470996 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "o_proj", "score", "up_proj" ]
816,004,096
220,214,272
512
5846.99
1719.08
{ "accuracy": 0.8940878912424913, "f1_macro": 0.8890996213167982, "f1_weighted": 0.8942800605830346, "precision": 0.8896864247090098, "recall": 0.8887568816093899 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
125,701,632
452,352
8
1161.96
535.91
{ "accuracy": 0.8394720202339551, "f1_macro": 0.826812075612483, "f1_weighted": 0.8393632386550851, "precision": 0.8318628106440021, "recall": 0.82331669697969 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
126,144,000
894,720
16
1161.88
538.96
{ "accuracy": 0.85243439772368, "f1_macro": 0.8414666204710375, "f1_weighted": 0.8525160041286179, "precision": 0.8459804920353469, "recall": 0.8383450765753199 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "out_proj", "project_in", "project_out", "score" ]
483,777,536
152,574,464
512
5015.85
2045.48
{ "accuracy": 0.8950363578880809, "f1_macro": 0.8907249099691867, "f1_weighted": 0.8951831287169804, "precision": 0.8921445128882302, "recall": 0.8894959616314126 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
127,028,736
1,779,456
32
1187.1
539.7
{ "accuracy": 0.861681947518179, "f1_macro": 0.8523904511311535, "f1_weighted": 0.8617716493957568, "precision": 0.8561816760131653, "recall": 0.8496307119007345 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
225,847,834
3,198,989
128
2058.03
3235.83
{ "accuracy": 0.8936136579196965, "f1_macro": 0.889623441794932, "f1_weighted": 0.8938304921792675, "precision": 0.8906434669168403, "recall": 0.8888174749651816 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
128,798,208
3,548,928
64
1182.25
540.44
{ "accuracy": 0.8682421751501739, "f1_macro": 0.8600805458073466, "f1_weighted": 0.8684536360301094, "precision": 0.862802619210438, "recall": 0.8580443571669788 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
596,605,952
816,128
4
3587.4
1189.93
{ "accuracy": 0.8208188428706924, "f1_macro": 0.803803184777025, "f1_weighted": 0.8203842243011442, "precision": 0.8079681691965234, "recall": 0.8011812343313595 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
132,337,152
7,087,872
128
1219.23
544.92
{ "accuracy": 0.8740910527979766, "f1_macro": 0.8674076585919336, "f1_weighted": 0.874357452235864, "precision": 0.8692916124298063, "recall": 0.8660304771602927 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
139,415,040
14,165,760
256
1354.48
584.16
{ "accuracy": 0.8788333860259248, "f1_macro": 0.8732464662931819, "f1_weighted": 0.8790780327011927, "precision": 0.8740462746037198, "recall": 0.8727523994376197 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
153,570,816
28,321,536
512
1567.45
636.97
{ "accuracy": 0.8819949415112235, "f1_macro": 0.8767961948123558, "f1_weighted": 0.8821914765653006, "precision": 0.8778226120184408, "recall": 0.8760540127189675 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
597,408,768
1,618,944
8
3581.6
1197.55
{ "accuracy": 0.8503003477711034, "f1_macro": 0.8386897466533239, "f1_weighted": 0.8502870533049263, "precision": 0.841251221913875, "recall": 0.8366487479785492 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
599,014,400
3,224,576
16
3604.08
1199.89
{ "accuracy": 0.8627094530509011, "f1_macro": 0.8534614281798678, "f1_weighted": 0.8626890910596854, "precision": 0.8557828039968904, "recall": 0.8516010099180618 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
228,993,562
6,344,717
256
2085.53
3308.44
{ "accuracy": 0.8932184634840341, "f1_macro": 0.8890447211297917, "f1_weighted": 0.8933972874346584, "precision": 0.8902306849388082, "recall": 0.8880565057993695 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
602,225,664
6,435,840
32
3649.07
1203.88
{ "accuracy": 0.8706923806512804, "f1_macro": 0.8634239412029298, "f1_weighted": 0.8707698430841055, "precision": 0.8653924937639137, "recall": 0.8617939389226705 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
608,648,192
12,858,368
64
3739.18
1225.97
{ "accuracy": 0.8776478027189377, "f1_macro": 0.8708582194931862, "f1_weighted": 0.8777249091185308, "precision": 0.8731814987140946, "recall": 0.8689035853367599 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "key", "query", "value" ]
235,285,018
12,636,173
512
2136.19
3466.42
{ "accuracy": 0.893692696806829, "f1_macro": 0.8893260523230074, "f1_weighted": 0.8938325610555807, "precision": 0.8903718417564096, "recall": 0.8884453261973413 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
621,493,248
25,703,424
128
3939.04
1263.57
{ "accuracy": 0.8842080303509326, "f1_macro": 0.8781890456366821, "f1_weighted": 0.8843166512583319, "precision": 0.8801848180799959, "recall": 0.8765411101850726 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
647,183,360
51,393,536
256
4294.98
1314.67
{ "accuracy": 0.8883970913689535, "f1_macro": 0.8827218155926939, "f1_weighted": 0.8885255515340483, "precision": 0.8845067968995219, "recall": 0.8812372146737685 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "k_proj", "q_proj", "v_proj" ]
698,563,584
102,773,760
512
5071.22
1461.21
{ "accuracy": 0.8910844135314575, "f1_macro": 0.8856577081113475, "f1_weighted": 0.8912433184815193, "precision": 0.887238378087781, "recall": 0.8843873865959717 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "q_proj", "score", "v_proj" ]
125,922,816
673,536
4
1391.73
672.16
{ "accuracy": 0.8340973759089472, "f1_macro": 0.8195111310215177, "f1_weighted": 0.8336773801387918, "precision": 0.8248669361299604, "recall": 0.8160423355014034 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "q_proj", "score", "v_proj" ]
126,586,368
1,337,088
8
1391.12
664.79
{ "accuracy": 0.8570186531773633, "f1_macro": 0.8472144472803785, "f1_weighted": 0.8570195833928468, "precision": 0.8510754538262949, "recall": 0.8444453913553926 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "k_proj", "o_proj", "q_proj", "score", "up_proj", "v_proj" ]
598,326,272
2,536,448
4
3846.87
1414.81
{ "accuracy": 0.8550426809990516, "f1_macro": 0.8448131741359602, "f1_weighted": 0.8550439816954918, "precision": 0.847881108884049, "recall": 0.8426216972541946 }
microsoft/deberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "in_proj", "pos_proj", "pos_q_proj" ]
408,213,530
1,987,597
4
3615.83
1874.51
{ "accuracy": 0.8408947202023396, "f1_macro": 0.8195939408694818, "f1_weighted": 0.839153423365856, "precision": 0.8337740775349373, "recall": 0.814049656635298 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "project_in", "project_out", "q_proj", "score", "v_proj" ]
332,991,488
1,788,416
4
3104.24
1620.05
{ "accuracy": 0.8584413531457477, "f1_macro": 0.8481145846712672, "f1_weighted": 0.8584607781383096, "precision": 0.8507815774518181, "recall": 0.846161502238727 }
FacebookAI/roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "dense", "key", "out_proj", "query", "value" ]
358,205,466
2,832,397
4
3321.83
1617.65
{ "accuracy": 0.8765412582990831, "f1_macro": 0.8694366262701073, "f1_weighted": 0.8767632742969793, "precision": 0.8721242437505067, "recall": 0.8674943496723615 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "q_proj", "score", "v_proj" ]
127,913,472
2,664,192
16
1402.76
669.39
{ "accuracy": 0.8657129307619349, "f1_macro": 0.8579806861280955, "f1_weighted": 0.865831023499551, "precision": 0.8607545184375709, "recall": 0.855889511984076 }
google-t5/t5-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "dense", "k", "o", "out_proj", "q", "v", "wi", "wo" ]
743,068,737
4,337,716
4
2293.08
2656.68
{ "accuracy": 0.5262409105279797, "f1_macro": 0.43016474280409145, "f1_weighted": 0.4797732855072074, "precision": 0.4858722107603928, "recall": 0.4657348240972282 }
FacebookAI/xlm-roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "dense", "key", "out_proj", "query", "value" ]
562,736,154
2,832,397
4
4959.69
2193.45
{ "accuracy": 0.8746443250079039, "f1_macro": 0.8668971915516758, "f1_weighted": 0.8748557341267201, "precision": 0.869321478508235, "recall": 0.8650385070431643 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "q_proj", "score", "v_proj" ]
130,567,680
5,318,400
32
1456.57
676.94
{ "accuracy": 0.8739329750237117, "f1_macro": 0.8676273807919703, "f1_weighted": 0.874056711084974, "precision": 0.8698157332723009, "recall": 0.8658792727704271 }
albert/albert-xxlarge-v2
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "embedding_hidden_mapping_in", "ffn", "ffn_output", "key", "pooler", "query", "value" ]
223,046,682
397,837
4
2602.45
3246.11
{ "accuracy": 0.7702339551059121, "f1_macro": 0.7068562937124588, "f1_weighted": 0.7547953627231089, "precision": 0.7688068102381627, "recall": 0.7133370406264442 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "k_proj", "o_proj", "q_proj", "score", "up_proj", "v_proj" ]
600,849,408
5,059,584
8
3865.16
1412.02
{ "accuracy": 0.871798925071135, "f1_macro": 0.8638650825003127, "f1_weighted": 0.871920076925293, "precision": 0.8656606299049033, "recall": 0.8625151897950111 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "project_in", "project_out", "q_proj", "score", "v_proj" ]
334,773,248
3,570,176
8
3124.63
1614.67
{ "accuracy": 0.8721150806196649, "f1_macro": 0.8653212321486065, "f1_weighted": 0.8724240402917425, "precision": 0.8663330901029702, "recall": 0.8647637306409034 }
FacebookAI/roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "dense", "key", "out_proj", "query", "value" ]
359,974,938
4,601,869
8
3339.6
1622.92
{ "accuracy": 0.8855516914321846, "f1_macro": 0.8798053247021997, "f1_weighted": 0.8859125849695306, "precision": 0.8810081137909893, "recall": 0.8790351704529304 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "q_proj", "score", "v_proj" ]
135,876,096
10,626,816
64
1517.11
693.35
{ "accuracy": 0.8787543471387923, "f1_macro": 0.8728657588306669, "f1_weighted": 0.8789343596761923, "precision": 0.8739969364416118, "recall": 0.8720554387804194 }
microsoft/deberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "in_proj", "pos_proj", "pos_q_proj" ]
410,187,802
3,961,869
8
3634.97
1863.11
{ "accuracy": 0.872984508378122, "f1_macro": 0.8616691597569711, "f1_weighted": 0.8729036044932067, "precision": 0.8659197561697604, "recall": 0.858944919384401 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "q_proj", "score", "v_proj" ]
146,492,928
21,243,648
128
1565.05
722.75
{ "accuracy": 0.8838918748024028, "f1_macro": 0.8790234599280935, "f1_weighted": 0.884055055240144, "precision": 0.8798701835390108, "recall": 0.8784294749160899 }
FacebookAI/xlm-roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "dense", "key", "out_proj", "query", "value" ]
564,505,626
4,601,869
8
4979.78
2194.14
{ "accuracy": 0.8836547581410054, "f1_macro": 0.877149962434546, "f1_weighted": 0.8838700883763358, "precision": 0.8796309693409445, "recall": 0.8752070134820868 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "k_proj", "o_proj", "q_proj", "score", "up_proj", "v_proj" ]
605,895,680
10,105,856
16
3935.89
1417.21
{ "accuracy": 0.8787543471387923, "f1_macro": 0.8712871204100013, "f1_weighted": 0.8788483866217315, "precision": 0.8730807182058917, "recall": 0.8699772657021357 }
facebook/opt-350m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "project_in", "project_out", "q_proj", "score", "v_proj" ]
338,336,768
7,133,696
16
3175.46
1626.37
{ "accuracy": 0.8801770471071767, "f1_macro": 0.8740703684110246, "f1_weighted": 0.8804309344324178, "precision": 0.8746337653733508, "recall": 0.8737917682404827 }
google-t5/t5-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "dense", "k", "o", "out_proj", "q", "v", "wi", "wo" ]
747,406,453
8,675,432
8
2304.43
2691.91
{ "accuracy": 0.7349826114448309, "f1_macro": 0.6523196597934748, "f1_weighted": 0.7146716185446933, "precision": 0.7317995959886214, "recall": 0.6742453986509211 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "q_proj", "score", "v_proj" ]
167,726,592
42,477,312
256
1870.67
805.73
{ "accuracy": 0.8878438191590262, "f1_macro": 0.8832363607841981, "f1_weighted": 0.8880241382732339, "precision": 0.8838165911828101, "recall": 0.8828869540982952 }
FacebookAI/roberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "dense", "key", "out_proj", "query", "value" ]
363,513,882
8,140,813
16
3394.42
1635.31
{ "accuracy": 0.8893455580145432, "f1_macro": 0.88397372792545, "f1_weighted": 0.8896281135337569, "precision": 0.8849301816109053, "recall": 0.8833705253065253 }
microsoft/deberta-large
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "classifier", "dense", "in_proj", "pos_proj", "pos_q_proj" ]
414,136,346
7,910,413
16
3683.47
1880.62
{ "accuracy": 0.8842080303509326, "f1_macro": 0.8759611031149815, "f1_weighted": 0.8844413324469923, "precision": 0.8779137462714746, "recall": 0.8745282214468114 }
facebook/opt-125m
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "fc1", "fc2", "k_proj", "out_proj", "q_proj", "score", "v_proj" ]
210,193,920
84,944,640
512
2372.22
979.85
{ "accuracy": 0.8909263357571925, "f1_macro": 0.886209373607292, "f1_weighted": 0.8911053653716174, "precision": 0.8869214004263485, "recall": 0.8857133418480126 }
Qwen/Qwen3-Reranker-0.6B
50,775
12,652
{ "auto_find_batch_size": true, "gradient_accumulation_steps": 4, "learning_rate": 0.00005, "logging_steps": 1, "lr_scheduler_type": "linear", "num_train_epochs": 1, "optim": "adamw_8bit", "output_dir": "outputs", "report_to": "none", "save_strategy": "no", "save_total_limit": 0, "seed": 3407, ...
[ "down_proj", "gate_proj", "k_proj", "o_proj", "q_proj", "score", "up_proj", "v_proj" ]
615,988,224
20,198,400
32
4077.33
1435.45
{ "accuracy": 0.8852355358836548, "f1_macro": 0.8785310963213552, "f1_weighted": 0.8852918756115437, "precision": 0.8804424423088518, "recall": 0.8770859602217187 }