re-trained model after eliminating annotation errors

README.md

@@ -1,4 +0,0 @@
----
-license: afl-3.0
-inference: false
----

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d5c5cbc732dc61f382394ca38be8c180004f09ebf568bf5f6c366f50c08d16a9
+oid sha256:9ec1ce3a63b75c6aa0715c8710c405d13c678f4147cedc324bcbac332a02aca8
 size 1008039837

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:72d93baa2b2bd83e9ad3d5a6dee38be0e6fdc3537acb4f533d9871c48a5cffca
+oid sha256:c4a34342441087075352a656fa4d85e95c5aa11fc4407c5367b8c0a76c1d5996
 size 504033325

rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4c81a7f6b3c22226527f74224efa5bfd6812e7643711f3363571e33b7048a5d4
-size 14567
+oid sha256:fa069ecc27b6d86cf02d066f167af631efe8c187ca951ac5ef9de95cfbf526f7
+size 14503

scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:501d80737344dcd2274196fc2df8440bccd24c52b8c2e1bc8bdc49a42cbf6651
+oid sha256:7dc4a8d7215a4f34ed15795f0b52176bdc71ccf4c978bd45cb3ccfcf2a049c02
 size 623

trainer_state.json

@@ -1,588 +1,272 @@
 {
-  "best_metric": 0.7370102490601179,
-  "best_model_checkpoint": "./CARES/checkpoints/roberta-stratified/run-4/checkpoint-5112",
-  "epoch": 36.255319148936174,
-  "global_step": 5112,
+  "best_metric": 0.8328207869559483,
+  "best_model_checkpoint": "./CARES/checkpoints/roberta-stratified/run-9/checkpoint-2272",
+  "epoch": 32.0,
+  "global_step": 2272,
   "is_hyper_param_search": true,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.01,
-      "eval_loss": 0.3023780882358551,
-      "eval_macro_f1": 0.04276761517615176,
-      "eval_macro_precision": 0.03250514933058703,
-      "eval_macro_recall": 0.0625,
-      "eval_micro_f1": 0.4057854560064283,
-      "eval_micro_precision": 0.5200823892893924,
-      "eval_micro_recall": 0.3326745718050066,
-      "eval_runtime": 2.6706,
-      "eval_samples_per_second": 363.595,
-      "eval_steps_per_second": 22.842,
+      "epoch": 2.0,
+      "eval_loss": 0.29745519161224365,
+      "eval_macro_f1": 0.0,
+      "eval_macro_precision": 0.0,
+      "eval_macro_recall": 0.0,
+      "eval_micro_f1": 0.0,
+      "eval_micro_precision": 0.0,
+      "eval_micro_recall": 0.0,
+      "eval_runtime": 2.6698,
+      "eval_samples_per_second": 361.828,
+      "eval_steps_per_second": 11.611,
       "step": 142
     },
     {
-      "epoch": 2.01,
-      "eval_loss": 0.2355797439813614,
-      "eval_macro_f1": 0.0478700416730977,
-      "eval_macro_precision": 0.10847007722007722,
-      "eval_macro_recall": 0.04781835896355593,
-      "eval_micro_f1": 0.3766552231486022,
-      "eval_micro_precision": 0.7370441458733206,
-      "eval_micro_recall": 0.25296442687747034,
-      "eval_runtime": 2.6743,
-      "eval_samples_per_second": 363.08,
-      "eval_steps_per_second": 22.809,
+      "epoch": 4.0,
+      "eval_loss": 0.2143602967262268,
+      "eval_macro_f1": 0.20760247776727883,
+      "eval_macro_precision": 0.2427021619684663,
+      "eval_macro_recall": 0.19452630677199997,
+      "eval_micro_f1": 0.5696619950535862,
+      "eval_micro_precision": 0.7601760176017601,
+      "eval_micro_recall": 0.45550428477257743,
+      "eval_runtime": 2.6814,
+      "eval_samples_per_second": 360.254,
+      "eval_steps_per_second": 11.561,
       "step": 284
     },
     {
-      "epoch": 3.02,
-      "eval_loss": 0.20839156210422516,
-      "eval_macro_f1": 0.1270678623977507,
-      "eval_macro_precision": 0.15146020789583886,
-      "eval_macro_recall": 0.11056835837347367,
-      "eval_micro_f1": 0.5080789946140036,
-      "eval_micro_precision": 0.7971830985915493,
-      "eval_micro_recall": 0.37285902503293805,
-      "eval_runtime": 2.6737,
-      "eval_samples_per_second": 363.172,
-      "eval_steps_per_second": 22.815,
+      "epoch": 6.0,
+      "eval_loss": 0.15995128452777863,
+      "eval_macro_f1": 0.32385167310781526,
+      "eval_macro_precision": 0.42763364413607774,
+      "eval_macro_recall": 0.2931228227099484,
+      "eval_micro_f1": 0.6900908014212397,
+      "eval_micro_precision": 0.860236220472441,
+      "eval_micro_recall": 0.5761371127224786,
+      "eval_runtime": 2.674,
+      "eval_samples_per_second": 361.255,
+      "eval_steps_per_second": 11.593,
       "step": 426
     },
     {
-      "epoch": 3.55,
-      "learning_rate": 4.760888771678367e-05,
-      "loss": 0.3067,
+      "epoch": 7.04,
+      "learning_rate": 4.40741152818786e-05,
+      "loss": 0.2773,
       "step": 500
     },
     {
-      "epoch": 4.03,
-      "eval_loss": 0.17458897829055786,
-      "eval_macro_f1": 0.2134043623340923,
-      "eval_macro_precision": 0.2555480902740095,
-      "eval_macro_recall": 0.1956241201896679,
-      "eval_micro_f1": 0.652869972666927,
-      "eval_micro_precision": 0.8015340364333653,
-      "eval_micro_recall": 0.5507246376811594,
-      "eval_runtime": 2.6714,
-      "eval_samples_per_second": 363.484,
-      "eval_steps_per_second": 22.835,
+      "epoch": 8.0,
+      "eval_loss": 0.13520723581314087,
+      "eval_macro_f1": 0.4425371105794407,
+      "eval_macro_precision": 0.49169850541474464,
+      "eval_macro_recall": 0.41330740987880255,
+      "eval_micro_f1": 0.7547309833024118,
+      "eval_micro_precision": 0.8633276740237691,
+      "eval_micro_recall": 0.6704021094264997,
+      "eval_runtime": 2.6984,
+      "eval_samples_per_second": 357.994,
+      "eval_steps_per_second": 11.488,
       "step": 568
     },
     {
-      "epoch": 5.04,
-      "eval_loss": 0.1594998985528946,
-      "eval_macro_f1": 0.322947699403185,
-      "eval_macro_precision": 0.33082406427783345,
-      "eval_macro_recall": 0.31592054350156284,
-      "eval_micro_f1": 0.7086383601756956,
-      "eval_micro_precision": 0.7973640856672158,
-      "eval_micro_recall": 0.6376811594202898,
-      "eval_runtime": 2.6774,
-      "eval_samples_per_second": 362.663,
-      "eval_steps_per_second": 22.783,
+      "epoch": 10.0,
+      "eval_loss": 0.10962910205125809,
+      "eval_macro_f1": 0.5655053778813368,
+      "eval_macro_precision": 0.6696587937430376,
+      "eval_macro_recall": 0.5338063461136978,
+      "eval_micro_f1": 0.8189806678383128,
+      "eval_micro_precision": 0.8772590361445783,
+      "eval_micro_recall": 0.7679630850362558,
+      "eval_runtime": 2.6744,
+      "eval_samples_per_second": 361.208,
+      "eval_steps_per_second": 11.592,
       "step": 710
     },
     {
-      "epoch": 6.04,
-      "eval_loss": 0.14328011870384216,
-      "eval_macro_f1": 0.35552541662372633,
-      "eval_macro_precision": 0.4154000509380286,
-      "eval_macro_recall": 0.3384352474665031,
-      "eval_micro_f1": 0.730332967435053,
-      "eval_micro_precision": 0.8213991769547325,
-      "eval_micro_recall": 0.6574440052700923,
-      "eval_runtime": 2.6759,
-      "eval_samples_per_second": 362.867,
-      "eval_steps_per_second": 22.796,
+      "epoch": 12.0,
+      "eval_loss": 0.0956677794456482,
+      "eval_macro_f1": 0.6602463081601572,
+      "eval_macro_precision": 0.799820247637967,
+      "eval_macro_recall": 0.5970945139878616,
+      "eval_micro_f1": 0.8437173686042465,
+      "eval_micro_precision": 0.8938053097345132,
+      "eval_micro_recall": 0.7989452867501649,
+      "eval_runtime": 2.6891,
+      "eval_samples_per_second": 359.223,
+      "eval_steps_per_second": 11.528,
       "step": 852
     },
     {
-      "epoch": 7.05,
-      "eval_loss": 0.1372288167476654,
-      "eval_macro_f1": 0.35769698646328396,
-      "eval_macro_precision": 0.45729587282684003,
-      "eval_macro_recall": 0.340956446807801,
-      "eval_micro_f1": 0.747014115092291,
-      "eval_micro_precision": 0.8289156626506025,
-      "eval_micro_recall": 0.6798418972332015,
-      "eval_runtime": 2.6665,
-      "eval_samples_per_second": 364.154,
-      "eval_steps_per_second": 22.877,
+      "epoch": 14.0,
+      "eval_loss": 0.09418635815382004,
+      "eval_macro_f1": 0.709171253471717,
+      "eval_macro_precision": 0.8508463365856882,
+      "eval_macro_recall": 0.6566230286073916,
+      "eval_micro_f1": 0.8551865799383773,
+      "eval_micro_precision": 0.8896011396011396,
+      "eval_micro_recall": 0.8233355306526038,
+      "eval_runtime": 2.6733,
+      "eval_samples_per_second": 361.353,
+      "eval_steps_per_second": 11.596,
       "step": 994
     },
     {
-      "epoch": 7.09,
-      "learning_rate": 4.585856096249015e-05,
-      "loss": 0.1366,
+      "epoch": 14.08,
+      "learning_rate": 4.073516715446356e-05,
+      "loss": 0.068,
       "step": 1000
     },
     {
-      "epoch": 8.06,
-      "eval_loss": 0.13587501645088196,
-      "eval_macro_f1": 0.4519999262177494,
-      "eval_macro_precision": 0.6383482484756009,
-      "eval_macro_recall": 0.4120606473790193,
-      "eval_micro_f1": 0.7476635514018692,
-      "eval_micro_precision": 0.8227848101265823,
-      "eval_micro_recall": 0.6851119894598156,
-      "eval_runtime": 2.6678,
-      "eval_samples_per_second": 363.976,
-      "eval_steps_per_second": 22.866,
+      "epoch": 16.0,
+      "eval_loss": 0.09525582939386368,
+      "eval_macro_f1": 0.776452344525862,
+      "eval_macro_precision": 0.8422867152488994,
+      "eval_macro_recall": 0.7394950574942071,
+      "eval_micro_f1": 0.8662207357859532,
+      "eval_micro_precision": 0.8791581805838425,
+      "eval_micro_recall": 0.8536585365853658,
+      "eval_runtime": 2.6899,
+      "eval_samples_per_second": 359.119,
+      "eval_steps_per_second": 11.525,
       "step": 1136
     },
     {
-      "epoch": 9.06,
-      "eval_loss": 0.1254434585571289,
-      "eval_macro_f1": 0.5752781473187552,
-      "eval_macro_precision": 0.6682540412600851,
-      "eval_macro_recall": 0.5248832453258128,
-      "eval_micro_f1": 0.7869884575026234,
-      "eval_micro_precision": 0.8389261744966443,
-      "eval_micro_recall": 0.741106719367589,
-      "eval_runtime": 2.6676,
-      "eval_samples_per_second": 364.002,
-      "eval_steps_per_second": 22.867,
+      "epoch": 18.0,
+      "eval_loss": 0.09120669960975647,
+      "eval_macro_f1": 0.7799395687308482,
+      "eval_macro_precision": 0.8259713713716451,
+      "eval_macro_recall": 0.7560566435810081,
+      "eval_micro_f1": 0.8660743665679499,
+      "eval_micro_precision": 0.8646517739816032,
+      "eval_micro_recall": 0.8675016479894528,
+      "eval_runtime": 2.6882,
+      "eval_samples_per_second": 359.349,
+      "eval_steps_per_second": 11.532,
       "step": 1278
     },
     {
-      "epoch": 10.07,
-      "eval_loss": 0.12716087698936462,
-      "eval_macro_f1": 0.5783160913404322,
-      "eval_macro_precision": 0.6669120855288475,
-      "eval_macro_recall": 0.5333612306839322,
-      "eval_micro_f1": 0.7954701441317777,
-      "eval_micro_precision": 0.8302292263610315,
-      "eval_micro_recall": 0.7635046113306982,
-      "eval_runtime": 2.6683,
-      "eval_samples_per_second": 363.896,
-      "eval_steps_per_second": 22.861,
+      "epoch": 20.0,
+      "eval_loss": 0.09322977066040039,
+      "eval_macro_f1": 0.7717526983031062,
+      "eval_macro_precision": 0.8213810289257493,
+      "eval_macro_recall": 0.747806253729357,
+      "eval_micro_f1": 0.8603205757278378,
+      "eval_micro_precision": 0.8538961038961039,
+      "eval_micro_recall": 0.8668424522083059,
+      "eval_runtime": 2.6911,
+      "eval_samples_per_second": 358.966,
+      "eval_steps_per_second": 11.52,
       "step": 1420
     },
     {
-      "epoch": 10.64,
-      "learning_rate": 4.410823420819664e-05,
-      "loss": 0.0661,
+      "epoch": 21.13,
+      "learning_rate": 3.739621902704851e-05,
+      "loss": 0.0222,
       "step": 1500
     },
     {
-      "epoch": 11.08,
-      "eval_loss": 0.12210354208946228,
-      "eval_macro_f1": 0.6129525814973475,
-      "eval_macro_precision": 0.694661979564102,
-      "eval_macro_recall": 0.5604968051311103,
-      "eval_micro_f1": 0.8090971743625087,
-      "eval_micro_precision": 0.8482658959537572,
-      "eval_micro_recall": 0.7733860342555995,
-      "eval_runtime": 2.6691,
-      "eval_samples_per_second": 363.79,
-      "eval_steps_per_second": 22.854,
+      "epoch": 22.0,
+      "eval_loss": 0.10442250967025757,
+      "eval_macro_f1": 0.765921285849591,
+      "eval_macro_precision": 0.7782427042161657,
+      "eval_macro_recall": 0.774471020810195,
+      "eval_micro_f1": 0.8516003879728419,
+      "eval_micro_precision": 0.8356598984771574,
+      "eval_micro_recall": 0.8681608437705999,
+      "eval_runtime": 2.6921,
+      "eval_samples_per_second": 358.824,
+      "eval_steps_per_second": 11.515,
       "step": 1562
     },
     {
-      "epoch": 12.09,
-      "eval_loss": 0.13834641873836517,
-      "eval_macro_f1": 0.59964421079272,
-      "eval_macro_precision": 0.6259545070217613,
-      "eval_macro_recall": 0.5849130461175929,
-      "eval_micro_f1": 0.7874705287975748,
-      "eval_micro_precision": 0.8056512749827704,
-      "eval_micro_recall": 0.7700922266139657,
-      "eval_runtime": 2.6674,
-      "eval_samples_per_second": 364.02,
-      "eval_steps_per_second": 22.868,
+      "epoch": 24.0,
+      "eval_loss": 0.09634574502706528,
+      "eval_macro_f1": 0.805432074935829,
+      "eval_macro_precision": 0.9021531097855335,
+      "eval_macro_recall": 0.7712696478949495,
+      "eval_micro_f1": 0.8704318936877077,
+      "eval_micro_precision": 0.8774279973208305,
+      "eval_micro_recall": 0.8635464733025708,
+      "eval_runtime": 2.6753,
+      "eval_samples_per_second": 361.083,
+      "eval_steps_per_second": 11.588,
       "step": 1704
     },
     {
-      "epoch": 13.09,
-      "eval_loss": 0.13302326202392578,
-      "eval_macro_f1": 0.6249414362192053,
-      "eval_macro_precision": 0.6603169616331872,
-      "eval_macro_recall": 0.6008160113233191,
-      "eval_micro_f1": 0.8105579685933846,
-      "eval_micro_precision": 0.8223728813559322,
-      "eval_micro_recall": 0.7990777338603425,
-      "eval_runtime": 2.6684,
-      "eval_samples_per_second": 363.887,
-      "eval_steps_per_second": 22.86,
+      "epoch": 26.0,
+      "eval_loss": 0.09631907194852829,
+      "eval_macro_f1": 0.8054621279402976,
+      "eval_macro_precision": 0.8819456215653025,
+      "eval_macro_recall": 0.782841919580731,
+      "eval_micro_f1": 0.8722700198544011,
+      "eval_micro_precision": 0.8757475083056478,
+      "eval_micro_recall": 0.8688200395517469,
+      "eval_runtime": 2.6758,
+      "eval_samples_per_second": 361.016,
+      "eval_steps_per_second": 11.585,
       "step": 1846
     },
     {
-      "epoch": 14.1,
-      "eval_loss": 0.13799144327640533,
-      "eval_macro_f1": 0.6256821918613437,
-      "eval_macro_precision": 0.6740205274811021,
-      "eval_macro_recall": 0.5938987995613589,
-      "eval_micro_f1": 0.8119891008174386,
-      "eval_micro_precision": 0.840620592383639,
-      "eval_micro_recall": 0.7852437417654808,
-      "eval_runtime": 2.6691,
-      "eval_samples_per_second": 363.79,
-      "eval_steps_per_second": 22.854,
+      "epoch": 28.0,
+      "eval_loss": 0.09827031195163727,
+      "eval_macro_f1": 0.824449032097833,
+      "eval_macro_precision": 0.8941626132550853,
+      "eval_macro_recall": 0.7913713438345684,
+      "eval_micro_f1": 0.8718459495351926,
+      "eval_micro_precision": 0.8782608695652174,
+      "eval_micro_recall": 0.8655240606460118,
+      "eval_runtime": 2.6758,
+      "eval_samples_per_second": 361.008,
+      "eval_steps_per_second": 11.585,
       "step": 1988
     },
     {
-      "epoch": 14.18,
-      "learning_rate": 4.235790745390312e-05,
-      "loss": 0.0324,
+      "epoch": 28.17,
+      "learning_rate": 3.4057270899633464e-05,
+      "loss": 0.011,
       "step": 2000
     },
     {
-      "epoch": 15.11,
-      "eval_loss": 0.13957080245018005,
-      "eval_macro_f1": 0.6541379860188454,
-      "eval_macro_precision": 0.7002594602789083,
-      "eval_macro_recall": 0.6359217043250158,
-      "eval_micro_f1": 0.803843605036448,
-      "eval_micro_precision": 0.8086666666666666,
-      "eval_micro_recall": 0.7990777338603425,
-      "eval_runtime": 2.6709,
-      "eval_samples_per_second": 363.55,
-      "eval_steps_per_second": 22.839,
+      "epoch": 30.0,
+      "eval_loss": 0.10059890896081924,
+      "eval_macro_f1": 0.8165877105799546,
+      "eval_macro_precision": 0.884478008536637,
+      "eval_macro_recall": 0.7957804401453603,
+      "eval_micro_f1": 0.8693088765149033,
+      "eval_micro_precision": 0.8639322916666666,
+      "eval_micro_recall": 0.8747528015820699,
+      "eval_runtime": 2.6759,
+      "eval_samples_per_second": 361.006,
+      "eval_steps_per_second": 11.585,
       "step": 2130
     },
     {
-      "epoch": 16.11,
-      "eval_loss": 0.13600043952465057,
-      "eval_macro_f1": 0.6528569810495737,
-      "eval_macro_precision": 0.7119806310239326,
-      "eval_macro_recall": 0.6199612374678921,
-      "eval_micro_f1": 0.8169491525423729,
-      "eval_micro_precision": 0.8414804469273743,
-      "eval_micro_recall": 0.7938076416337286,
-      "eval_runtime": 2.6686,
-      "eval_samples_per_second": 363.864,
-      "eval_steps_per_second": 22.859,
+      "epoch": 32.0,
+      "eval_loss": 0.10830199718475342,
+      "eval_macro_f1": 0.8328207869559483,
+      "eval_macro_precision": 0.8739781063350807,
+      "eval_macro_recall": 0.8212976019561394,
+      "eval_micro_f1": 0.8676422764227643,
+      "eval_micro_precision": 0.8562259306803595,
+      "eval_micro_recall": 0.8793671720500988,
+      "eval_runtime": 2.674,
+      "eval_samples_per_second": 361.259,
+      "eval_steps_per_second": 11.593,
       "step": 2272
-    },
-    {
-      "epoch": 17.12,
-      "eval_loss": 0.1411595642566681,
-      "eval_macro_f1": 0.6780053870985077,
-      "eval_macro_precision": 0.7825949490722317,
-      "eval_macro_recall": 0.6371295063080809,
-      "eval_micro_f1": 0.8134328358208954,
-      "eval_micro_precision": 0.8384615384615385,
-      "eval_micro_recall": 0.7898550724637681,
-      "eval_runtime": 2.6685,
-      "eval_samples_per_second": 363.87,
-      "eval_steps_per_second": 22.859,
-      "step": 2414
-    },
-    {
-      "epoch": 17.73,
-      "learning_rate": 4.06075806996096e-05,
-      "loss": 0.0173,
-      "step": 2500
-    },
-    {
-      "epoch": 18.13,
-      "eval_loss": 0.14683738350868225,
-      "eval_macro_f1": 0.6538188838769178,
-      "eval_macro_precision": 0.7058131112592007,
-      "eval_macro_recall": 0.628120629850237,
-      "eval_micro_f1": 0.8044280442804428,
-      "eval_micro_precision": 0.8195488721804511,
-      "eval_micro_recall": 0.7898550724637681,
-      "eval_runtime": 2.6704,
-      "eval_samples_per_second": 363.611,
-      "eval_steps_per_second": 22.843,
-      "step": 2556
-    },
-    {
-      "epoch": 19.13,
-      "eval_loss": 0.14477181434631348,
-      "eval_macro_f1": 0.7213223353389469,
-      "eval_macro_precision": 0.7931053666626622,
-      "eval_macro_recall": 0.6989155005450692,
-      "eval_micro_f1": 0.8122731771692511,
-      "eval_micro_precision": 0.813615333773959,
-      "eval_micro_recall": 0.810935441370224,
-      "eval_runtime": 2.6682,
-      "eval_samples_per_second": 363.916,
-      "eval_steps_per_second": 22.862,
-      "step": 2698
-    },
-    {
-      "epoch": 20.14,
-      "eval_loss": 0.1553182750940323,
-      "eval_macro_f1": 0.6767777822180807,
-      "eval_macro_precision": 0.7296284296772766,
-      "eval_macro_recall": 0.6640188299255232,
-      "eval_micro_f1": 0.8082867477803354,
-      "eval_micro_precision": 0.8069599474720945,
-      "eval_micro_recall": 0.8096179183135704,
-      "eval_runtime": 2.6706,
-      "eval_samples_per_second": 363.587,
-      "eval_steps_per_second": 22.841,
-      "step": 2840
-    },
-    {
-      "epoch": 21.15,
-      "eval_loss": 0.14831620454788208,
-      "eval_macro_f1": 0.6951923518777028,
-      "eval_macro_precision": 0.8479068478364982,
-      "eval_macro_recall": 0.6493756779822191,
-      "eval_micro_f1": 0.8177506775067751,
-      "eval_micro_precision": 0.8417015341701534,
-      "eval_micro_recall": 0.7951251646903821,
-      "eval_runtime": 2.6679,
-      "eval_samples_per_second": 363.953,
-      "eval_steps_per_second": 22.864,
-      "step": 2982
-    },
-    {
-      "epoch": 21.28,
-      "learning_rate": 3.885725394531609e-05,
-      "loss": 0.0121,
-      "step": 3000
-    },
-    {
-      "epoch": 22.16,
-      "eval_loss": 0.1529681533575058,
-      "eval_macro_f1": 0.7215214471733791,
-      "eval_macro_precision": 0.840496134606828,
-      "eval_macro_recall": 0.6777226713039917,
-      "eval_micro_f1": 0.8099395567494962,
-      "eval_micro_precision": 0.826027397260274,
-      "eval_micro_recall": 0.7944664031620553,
-      "eval_runtime": 2.6724,
-      "eval_samples_per_second": 363.343,
-      "eval_steps_per_second": 22.826,
-      "step": 3124
-    },
-    {
-      "epoch": 23.16,
-      "eval_loss": 0.15208803117275238,
-      "eval_macro_f1": 0.7282532116551124,
-      "eval_macro_precision": 0.8178655579947314,
-      "eval_macro_recall": 0.702450635375965,
-      "eval_micro_f1": 0.8099009900990098,
-      "eval_micro_precision": 0.8115079365079365,
-      "eval_micro_recall": 0.808300395256917,
-      "eval_runtime": 2.6735,
-      "eval_samples_per_second": 363.196,
-      "eval_steps_per_second": 22.817,
-      "step": 3266
-    },
-    {
-      "epoch": 24.17,
-      "eval_loss": 0.17097046971321106,
-      "eval_macro_f1": 0.6781929633024913,
-      "eval_macro_precision": 0.7664743620916477,
-      "eval_macro_recall": 0.647287649907105,
-      "eval_micro_f1": 0.8018836192398252,
-      "eval_micro_precision": 0.8192439862542955,
-      "eval_micro_recall": 0.7852437417654808,
-      "eval_runtime": 2.6679,
-      "eval_samples_per_second": 363.956,
-      "eval_steps_per_second": 22.864,
-      "step": 3408
-    },
-    {
-      "epoch": 24.82,
-      "learning_rate": 3.710692719102257e-05,
-      "loss": 0.0086,
-      "step": 3500
-    },
-    {
-      "epoch": 25.18,
-      "eval_loss": 0.16351111233234406,
-      "eval_macro_f1": 0.7009939961760294,
-      "eval_macro_precision": 0.8049793687947511,
-      "eval_macro_recall": 0.6631263784729529,
-      "eval_micro_f1": 0.8080672268907563,
-      "eval_micro_precision": 0.8249828414550446,
-      "eval_micro_recall": 0.7918313570487484,
-      "eval_runtime": 2.6758,
-      "eval_samples_per_second": 362.882,
-      "eval_steps_per_second": 22.797,
-      "step": 3550
-    },
-    {
-      "epoch": 26.18,
-      "eval_loss": 0.1703951209783554,
-      "eval_macro_f1": 0.728858993061112,
-      "eval_macro_precision": 0.8293166501604144,
-      "eval_macro_recall": 0.6848344575219967,
-      "eval_micro_f1": 0.8101945003353456,
-      "eval_micro_precision": 0.825136612021858,
-      "eval_micro_recall": 0.7957839262187089,
-      "eval_runtime": 2.6853,
-      "eval_samples_per_second": 361.599,
-      "eval_steps_per_second": 22.716,
-      "step": 3692
-    },
-    {
-      "epoch": 27.19,
-      "eval_loss": 0.1729104220867157,
-      "eval_macro_f1": 0.7246800474910258,
-      "eval_macro_precision": 0.8088443156400115,
-      "eval_macro_recall": 0.6938957592472167,
-      "eval_micro_f1": 0.8112827400940228,
-      "eval_micro_precision": 0.8273972602739726,
-      "eval_micro_recall": 0.7957839262187089,
-      "eval_runtime": 2.6704,
-      "eval_samples_per_second": 363.619,
-      "eval_steps_per_second": 22.843,
-      "step": 3834
-    },
-    {
-      "epoch": 28.2,
-      "eval_loss": 0.17421075701713562,
-      "eval_macro_f1": 0.721641920467399,
-      "eval_macro_precision": 0.7953793533738551,
-      "eval_macro_recall": 0.6939586213926427,
-      "eval_micro_f1": 0.8118745830553702,
-      "eval_micro_precision": 0.8222972972972973,
-      "eval_micro_recall": 0.8017127799736495,
-      "eval_runtime": 2.6674,
-      "eval_samples_per_second": 364.019,
-      "eval_steps_per_second": 22.868,
-      "step": 3976
-    },
-    {
-      "epoch": 28.37,
-      "learning_rate": 3.535660043672905e-05,
-      "loss": 0.0054,
-      "step": 4000
-    },
-    {
-      "epoch": 29.21,
-      "eval_loss": 0.17936377227306366,
-      "eval_macro_f1": 0.7165402537125084,
-      "eval_macro_precision": 0.7314012003015316,
-      "eval_macro_recall": 0.7193045979731636,
-      "eval_micro_f1": 0.8045826513911619,
-      "eval_micro_precision": 0.7996096291476903,
-      "eval_micro_recall": 0.8096179183135704,
-      "eval_runtime": 2.669,
-      "eval_samples_per_second": 363.809,
-      "eval_steps_per_second": 22.855,
-      "step": 4118
-    },
-    {
-      "epoch": 30.21,
-      "eval_loss": 0.184128999710083,
-      "eval_macro_f1": 0.7249069877656021,
-      "eval_macro_precision": 0.7920679958383108,
-      "eval_macro_recall": 0.6986794530201526,
-      "eval_micro_f1": 0.8046822742474916,
-      "eval_micro_precision": 0.8172554347826086,
-      "eval_micro_recall": 0.7924901185770751,
-      "eval_runtime": 2.6699,
-      "eval_samples_per_second": 363.687,
-      "eval_steps_per_second": 22.847,
-      "step": 4260
-    },
-    {
-      "epoch": 31.22,
-      "eval_loss": 0.18270088732242584,
-      "eval_macro_f1": 0.7341637485973148,
-      "eval_macro_precision": 0.8378115033399074,
-      "eval_macro_recall": 0.6892808840460984,
-      "eval_micro_f1": 0.8134680134680136,
-      "eval_micro_precision": 0.8319559228650137,
-      "eval_micro_recall": 0.7957839262187089,
-      "eval_runtime": 2.6693,
-      "eval_samples_per_second": 363.77,
-      "eval_steps_per_second": 22.853,
-      "step": 4402
-    },
-    {
-      "epoch": 31.91,
-      "learning_rate": 3.3606273682435536e-05,
-      "loss": 0.0035,
-      "step": 4500
-    },
-    {
-      "epoch": 32.23,
-      "eval_loss": 0.18414482474327087,
-      "eval_macro_f1": 0.7314425663595913,
-      "eval_macro_precision": 0.8099477622958757,
-      "eval_macro_recall": 0.6986944147838622,
-      "eval_micro_f1": 0.8134003350083752,
-      "eval_micro_precision": 0.8275391956373551,
-      "eval_micro_recall": 0.7997364953886693,
-      "eval_runtime": 2.6686,
-      "eval_samples_per_second": 363.857,
-      "eval_steps_per_second": 22.858,
-      "step": 4544
-    },
-    {
-      "epoch": 33.23,
-      "eval_loss": 0.18519891798496246,
-      "eval_macro_f1": 0.7352936873025266,
-      "eval_macro_precision": 0.8265688293622399,
-      "eval_macro_recall": 0.6958315897622773,
-      "eval_micro_f1": 0.8212722988892629,
-      "eval_micro_precision": 0.8396421197522368,
-      "eval_micro_recall": 0.8036890645586298,
-      "eval_runtime": 2.6687,
-      "eval_samples_per_second": 363.853,
-      "eval_steps_per_second": 22.858,
-      "step": 4686
-    },
-    {
-      "epoch": 34.24,
-      "eval_loss": 0.1865757256746292,
-      "eval_macro_f1": 0.7207303759640109,
-      "eval_macro_precision": 0.7610227795725935,
-      "eval_macro_recall": 0.6996289765073358,
-      "eval_micro_f1": 0.8130245048674052,
-      "eval_micro_precision": 0.8288843258042436,
-      "eval_micro_recall": 0.7977602108036891,
-      "eval_runtime": 2.6701,
-      "eval_samples_per_second": 363.659,
-      "eval_steps_per_second": 22.846,
-      "step": 4828
-    },
-    {
-      "epoch": 35.25,
-      "eval_loss": 0.19158615171909332,
-      "eval_macro_f1": 0.7231833700345036,
-      "eval_macro_precision": 0.759857266859788,
-      "eval_macro_recall": 0.7038379676451936,
-      "eval_micro_f1": 0.8071928071928073,
-      "eval_micro_precision": 0.8161616161616162,
-      "eval_micro_recall": 0.7984189723320159,
-      "eval_runtime": 2.6696,
-      "eval_samples_per_second": 363.724,
-      "eval_steps_per_second": 22.85,
-      "step": 4970
-    },
-    {
-      "epoch": 35.46,
-      "learning_rate": 3.185594692814201e-05,
-      "loss": 0.0025,
-      "step": 5000
-    },
-    {
-      "epoch": 36.26,
-      "eval_loss": 0.18590226769447327,
-      "eval_macro_f1": 0.7370102490601179,
-      "eval_macro_precision": 0.8141501549264045,
-      "eval_macro_recall": 0.7143003391573518,
-      "eval_micro_f1": 0.8030253206182177,
-      "eval_micro_precision": 0.8017071569271176,
-      "eval_micro_recall": 0.8043478260869565,
-      "eval_runtime": 2.6693,
-      "eval_samples_per_second": 363.761,
-      "eval_steps_per_second": 22.852,
-      "step": 5112
     }
   ],
-  "max_steps": 14100,
+  "max_steps": 7100,
   "num_train_epochs": 100,
-  "total_flos": 1.9741043739581184e+16,
+  "total_flos": 1.673289313217472e+16,
   "trial_name": null,
   "trial_params": {
-    "adam_epsilon": 2.7636948844125687e-08,
-    "learning_rate": 4.7699904708006934e-05,
-    "per_device_eval_batch_size": 16,
-    "per_device_train_batch_size": 16,
-    "seed": 320,
-    "warmup_steps": 474,
-    "weight_decay": 0.08343382340090989
+    "adam_epsilon": 1.724439344881123e-07,
+    "learning_rate": 4.540969453284462e-05,
+    "per_device_eval_batch_size": 32,
+    "per_device_train_batch_size": 32,
+    "seed": 324,
+    "warmup_steps": 300,
+    "weight_decay": 0.00598936569463419
   }
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ad2c884e186c33efca82a69390d8020a7e0d5a1fa537e67f0f88b9f3eb3c34c9
+oid sha256:ac58073aa21113afba4497d625fe9f04a3f41c55ebf5e736db1a27388505e1c5
 size 3439