Training in progress, step 4770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3af5b47ba404b3ef2f87defb7e11fd5376d55cadf07c6b8c59e1c82fd72748ba
+oid sha256:4384a128a0ec6df3743bda704006b555752b73cff6bd957895b0f805429a080f
 size 268290900

run-2/checkpoint-4770/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f122a6f5a1be1a5d1da16d264e660ab340bff88781ea5d2351bd9f81b035bc37
+oid sha256:4384a128a0ec6df3743bda704006b555752b73cff6bd957895b0f805429a080f
 size 268290900

run-2/checkpoint-4770/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0094912f5709ce2ffb621e0aaf759f22a99a7101d37fa97db05de19146f77601
+oid sha256:7755b025a2bbd38e78c29ceb5f7677ae0af302bd43caf53266da731e06ac1891
 size 536643898

run-2/checkpoint-4770/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3aab55f18bed58889b499312e3ac62c2767f96b1e152abc648dff58c68381976
+oid sha256:45a2e1c2d46ee5968aba751a8ac45b77d27d443ec720f1b98069f4f0186816d2
 size 1064

run-2/checkpoint-4770/trainer_state.json

@@ -10,233 +10,233 @@
   "log_history": [
     {
       "epoch": 0.9968553459119497,
-      "grad_norm": 1.04072105884552,
-      "learning_rate": 0.00041128900119572176,
-      "loss": 0.2584,
+      "grad_norm": 0.5192151069641113,
+      "learning_rate": 5.2799658937698636e-05,
+      "loss": 0.263,
       "step": 317
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.7422580645161291,
-      "eval_loss": 0.1071159690618515,
-      "eval_runtime": 5.4205,
-      "eval_samples_per_second": 571.899,
-      "eval_steps_per_second": 11.991,
+      "eval_accuracy": 0.8558064516129033,
+      "eval_loss": 0.07722163945436478,
+      "eval_runtime": 5.3713,
+      "eval_samples_per_second": 577.147,
+      "eval_steps_per_second": 12.101,
       "step": 318
     },
     {
       "epoch": 1.9937106918238994,
-      "grad_norm": 0.5266364216804504,
-      "learning_rate": 0.0003820101749260061,
-      "loss": 0.0823,
+      "grad_norm": 0.3630720376968384,
+      "learning_rate": 5.1086081394216154e-05,
+      "loss": 0.0714,
       "step": 634
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8493548387096774,
-      "eval_loss": 0.06963900476694107,
-      "eval_runtime": 5.4463,
-      "eval_samples_per_second": 569.198,
-      "eval_steps_per_second": 11.935,
+      "eval_accuracy": 0.9187096774193548,
+      "eval_loss": 0.03628459945321083,
+      "eval_runtime": 5.4431,
+      "eval_samples_per_second": 569.53,
+      "eval_steps_per_second": 11.942,
       "step": 636
     },
     {
       "epoch": 2.990566037735849,
-      "grad_norm": 0.6715505123138428,
-      "learning_rate": 0.0003527313486562904,
-      "loss": 0.0581,
+      "grad_norm": 0.2893732786178589,
+      "learning_rate": 4.8312932511839625e-05,
+      "loss": 0.0429,
       "step": 951
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8670967741935484,
-      "eval_loss": 0.0621778629720211,
-      "eval_runtime": 5.5165,
-      "eval_samples_per_second": 561.951,
-      "eval_steps_per_second": 11.783,
+      "eval_accuracy": 0.932258064516129,
+      "eval_loss": 0.025179537013173103,
+      "eval_runtime": 5.4106,
+      "eval_samples_per_second": 572.946,
+      "eval_steps_per_second": 12.013,
       "step": 954
     },
     {
       "epoch": 3.9874213836477987,
-      "grad_norm": 0.15568938851356506,
-      "learning_rate": 0.00032345252238657476,
-      "loss": 0.0477,
+      "grad_norm": 0.19381985068321228,
+      "learning_rate": 4.4600653905279676e-05,
+      "loss": 0.0333,
       "step": 1268
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8770967741935484,
-      "eval_loss": 0.06203094497323036,
-      "eval_runtime": 5.4775,
-      "eval_samples_per_second": 565.955,
-      "eval_steps_per_second": 11.867,
+      "eval_accuracy": 0.9409677419354838,
+      "eval_loss": 0.02168211154639721,
+      "eval_runtime": 5.42,
+      "eval_samples_per_second": 571.958,
+      "eval_steps_per_second": 11.993,
       "step": 1272
     },
     {
       "epoch": 4.984276729559748,
-      "grad_norm": 0.11771262437105179,
-      "learning_rate": 0.00029417369611685915,
-      "loss": 0.0359,
+      "grad_norm": 0.12074659019708633,
+      "learning_rate": 4.011047486607245e-05,
+      "loss": 0.0286,
       "step": 1585
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.9083870967741936,
-      "eval_loss": 0.04885668307542801,
-      "eval_runtime": 5.4322,
-      "eval_samples_per_second": 570.673,
-      "eval_steps_per_second": 11.966,
+      "eval_accuracy": 0.94,
+      "eval_loss": 0.019550619646906853,
+      "eval_runtime": 5.4026,
+      "eval_samples_per_second": 573.8,
+      "eval_steps_per_second": 12.031,
       "step": 1590
     },
     {
       "epoch": 5.981132075471698,
-      "grad_norm": 0.19639770686626434,
-      "learning_rate": 0.0002648948698471435,
-      "loss": 0.0304,
+      "grad_norm": 0.1383192092180252,
+      "learning_rate": 3.5037409955916656e-05,
+      "loss": 0.0255,
       "step": 1902
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.9012903225806451,
-      "eval_loss": 0.05035402998328209,
-      "eval_runtime": 5.4681,
-      "eval_samples_per_second": 566.923,
-      "eval_steps_per_second": 11.887,
+      "eval_accuracy": 0.9370967741935484,
+      "eval_loss": 0.018356727436184883,
+      "eval_runtime": 5.4519,
+      "eval_samples_per_second": 568.612,
+      "eval_steps_per_second": 11.923,
       "step": 1908
     },
     {
       "epoch": 6.977987421383648,
-      "grad_norm": 0.09110367298126221,
-      "learning_rate": 0.00023561604357742784,
-      "loss": 0.0274,
+      "grad_norm": 0.12834681570529938,
+      "learning_rate": 2.9601789259934643e-05,
+      "loss": 0.0234,
       "step": 2219
     },
     {
       "epoch": 7.0,
-      "eval_accuracy": 0.9058064516129032,
-      "eval_loss": 0.04763167351484299,
-      "eval_runtime": 5.4566,
-      "eval_samples_per_second": 568.117,
-      "eval_steps_per_second": 11.912,
+      "eval_accuracy": 0.9409677419354838,
+      "eval_loss": 0.017373906448483467,
+      "eval_runtime": 5.4536,
+      "eval_samples_per_second": 568.43,
+      "eval_steps_per_second": 11.919,
       "step": 2226
     },
     {
       "epoch": 7.9748427672955975,
-      "grad_norm": 0.07694189995527267,
-      "learning_rate": 0.00020633721730771218,
-      "loss": 0.025,
+      "grad_norm": 0.10002776980400085,
+      "learning_rate": 2.403968915243833e-05,
+      "loss": 0.0221,
       "step": 2536
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.9103225806451613,
-      "eval_loss": 0.04542699456214905,
-      "eval_runtime": 5.4199,
-      "eval_samples_per_second": 571.971,
-      "eval_steps_per_second": 11.993,
+      "eval_accuracy": 0.9383870967741935,
+      "eval_loss": 0.016409363597631454,
+      "eval_runtime": 5.3751,
+      "eval_samples_per_second": 576.738,
+      "eval_steps_per_second": 12.093,
       "step": 2544
     },
     {
       "epoch": 8.971698113207546,
-      "grad_norm": 0.09620998799800873,
-      "learning_rate": 0.00017705839103799654,
-      "loss": 0.0231,
+      "grad_norm": 0.10179214179515839,
+      "learning_rate": 1.8592679180012948e-05,
+      "loss": 0.0209,
       "step": 2853
     },
     {
       "epoch": 9.0,
-      "eval_accuracy": 0.9174193548387096,
-      "eval_loss": 0.04218007251620293,
-      "eval_runtime": 5.4278,
-      "eval_samples_per_second": 571.134,
-      "eval_steps_per_second": 11.975,
+      "eval_accuracy": 0.9409677419354838,
+      "eval_loss": 0.015916811302304268,
+      "eval_runtime": 5.4101,
+      "eval_samples_per_second": 573.001,
+      "eval_steps_per_second": 12.015,
       "step": 2862
     },
     {
       "epoch": 9.968553459119496,
-      "grad_norm": 0.06652583926916122,
-      "learning_rate": 0.0001477795647682809,
-      "loss": 0.0205,
+      "grad_norm": 0.09086760133504868,
+      "learning_rate": 1.3497330368683434e-05,
+      "loss": 0.0201,
       "step": 3170
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.9167741935483871,
-      "eval_loss": 0.040490638464689255,
-      "eval_runtime": 5.4856,
-      "eval_samples_per_second": 565.113,
-      "eval_steps_per_second": 11.849,
+      "eval_accuracy": 0.9441935483870968,
+      "eval_loss": 0.015536032617092133,
+      "eval_runtime": 5.4274,
+      "eval_samples_per_second": 571.174,
+      "eval_steps_per_second": 11.976,
       "step": 3180
     },
     {
       "epoch": 10.965408805031446,
-      "grad_norm": 0.1630864441394806,
-      "learning_rate": 0.00011850073849856523,
-      "loss": 0.0191,
+      "grad_norm": 0.11637966334819794,
+      "learning_rate": 8.974940623603928e-06,
+      "loss": 0.0194,
       "step": 3487
     },
     {
       "epoch": 11.0,
-      "eval_accuracy": 0.9241935483870968,
-      "eval_loss": 0.03858407959342003,
-      "eval_runtime": 5.4468,
-      "eval_samples_per_second": 569.139,
-      "eval_steps_per_second": 11.934,
+      "eval_accuracy": 0.9435483870967742,
+      "eval_loss": 0.015223735943436623,
+      "eval_runtime": 5.4108,
+      "eval_samples_per_second": 572.927,
+      "eval_steps_per_second": 12.013,
       "step": 3498
     },
     {
       "epoch": 11.962264150943396,
-      "grad_norm": 0.053046438843011856,
-      "learning_rate": 8.922191222884958e-05,
-      "loss": 0.0178,
+      "grad_norm": 0.08573295176029205,
+      "learning_rate": 5.221923461087825e-06,
+      "loss": 0.0189,
       "step": 3804
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.9225806451612903,
-      "eval_loss": 0.03706495836377144,
-      "eval_runtime": 5.4181,
-      "eval_samples_per_second": 572.151,
-      "eval_steps_per_second": 11.997,
+      "eval_accuracy": 0.9416129032258065,
+      "eval_loss": 0.01511952094733715,
+      "eval_runtime": 5.3754,
+      "eval_samples_per_second": 576.706,
+      "eval_steps_per_second": 12.092,
       "step": 3816
     },
     {
       "epoch": 12.959119496855346,
-      "grad_norm": 0.06040259450674057,
-      "learning_rate": 5.9943085959133934e-05,
-      "loss": 0.0169,
+      "grad_norm": 0.11835107207298279,
+      "learning_rate": 2.4012775033738117e-06,
+      "loss": 0.0185,
       "step": 4121
     },
     {
       "epoch": 13.0,
-      "eval_accuracy": 0.9241935483870968,
-      "eval_loss": 0.03619401156902313,
-      "eval_runtime": 5.4534,
-      "eval_samples_per_second": 568.455,
-      "eval_steps_per_second": 11.919,
+      "eval_accuracy": 0.9416129032258065,
+      "eval_loss": 0.014925230294466019,
+      "eval_runtime": 5.4176,
+      "eval_samples_per_second": 572.21,
+      "eval_steps_per_second": 11.998,
       "step": 4134
     },
     {
       "epoch": 13.955974842767295,
-      "grad_norm": 0.05235395208001137,
-      "learning_rate": 3.066425968941828e-05,
-      "loss": 0.0159,
+      "grad_norm": 0.07725111395120621,
+      "learning_rate": 6.355072275323512e-07,
+      "loss": 0.0184,
       "step": 4438
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.9241935483870968,
-      "eval_loss": 0.0354730561375618,
-      "eval_runtime": 5.4401,
-      "eval_samples_per_second": 569.842,
-      "eval_steps_per_second": 11.948,
+      "eval_accuracy": 0.9412903225806452,
+      "eval_loss": 0.014896313659846783,
+      "eval_runtime": 5.4395,
+      "eval_samples_per_second": 569.905,
+      "eval_steps_per_second": 11.95,
       "step": 4452
     },
     {
       "epoch": 14.952830188679245,
-      "grad_norm": 0.051196809858083725,
-      "learning_rate": 1.3854334197026332e-06,
-      "loss": 0.0154,
+      "grad_norm": 0.07756518572568893,
+      "learning_rate": 1.302429982933581e-09,
+      "loss": 0.0182,
       "step": 4755
     }
   ],
@@ -261,11 +261,11 @@
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.015638508087186143,
-    "learning_rate": 0.00044056782746543737,
-    "lr_scheduler_type": "linear",
+    "alpha": 0.0990358007931369,
+    "learning_rate": 5.337924214697641e-05,
+    "lr_scheduler_type": "cosine",
     "num_train_epochs": 15,
-    "temperature": 5.955400780245458,
-    "weight_decay": 0.08329448851063012
+    "temperature": 13.220869405817513,
+    "weight_decay": 0.027535903658101313
   }
 }

run-2/checkpoint-4770/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:91a9eb1b39a6a3bda2d5fe5efe49cd9a7c6230d2da261f5095044d7586a45bdc
+oid sha256:d5589e749649dba3af878a3715400f3ce342a88ed8dcc73e44b015459960a4ca
 size 5240

runs/Oct20_13-24-54_87443764e281/events.out.tfevents.1729432230.87443764e281.307.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:025075d61fc4f18d6130d33e00fa89f3faddb9375024fcdd90262849b93120b1
+size 20826

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c8997ba64c78b786621c16831c31cf39c03bf94132fb6b567c38b51563d00871
+oid sha256:d5589e749649dba3af878a3715400f3ce342a88ed8dcc73e44b015459960a4ca
 size 5240