checkpoint-2000

qwen3.5-2b-50pct-masked-weighted/best_checkpoint_wer/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5c355b9758d25da25cd198145fefdcbabf96c8381dc3e00789d628ff12d935b2
+oid sha256:80b628f7e9c55852ca9b5bf5bdb6613db51cba9f9c635b749fc7b686b56ebfe1
 size 134604152

qwen3.5-2b-50pct-masked-weighted/best_checkpoint_wer/best_metrics.json

@@ -1,4 +1,4 @@
 {
-  "step": 1000,
-  "wer": 0.6495641332727022
+  "step": 2000,
+  "wer": 0.2231321104545041
 }

qwen3.5-2b-50pct-masked-weighted/best_checkpoint_wer/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:87a42ad7610cdce89a745a148c3b9f478ca80c2d6be4bbec707af20112091abe
+oid sha256:754de55d9e90d6cba75eea0f9ee8bb59b2f2f25dc5f7e568c2a0a5b8269ceb06
 size 269426431

qwen3.5-2b-50pct-masked-weighted/best_checkpoint_wer/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8372e63a426c4c2855a22732d06e8275ff2b5b5b4accbf12ba6a522510c9ee7d
+oid sha256:98f4a968e01d607786538f37082784acf14d994ddede3a162e85abf8496645a6
 size 14645

qwen3.5-2b-50pct-masked-weighted/best_checkpoint_wer/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:34532159d3083a204b1ddf71a34daca30b5d985bb7d15860a5e34b3afa0bfec2
+oid sha256:be32b28b4401079e06a80019d25bd8c958c4d5a19216b70fd633be2006252ebe
 size 1465

qwen3.5-2b-50pct-masked-weighted/best_checkpoint_wer/trainer_state.json

@@ -2,9 +2,9 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.17946474639388024,
+  "epoch": 0.3589294927877605,
   "eval_steps": 500,
-  "global_step": 1000,
+  "global_step": 2000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -708,6 +708,741 @@
       "learning_rate": 9.583490744238761e-05,
       "loss": 1.1706391334533692,
       "step": 1000
+    },
+    {
+      "epoch": 0.17946474639388024,
+      "step": 1000,
+      "test_entity_exact_match": 0.0019801980198019802,
+      "test_entity_fix_rate": 0.09490559539092588,
+      "test_entity_fix_rate_lenient": 0.09133004926108375,
+      "test_entity_fix_rate_recoverable": 0.09265880308751295,
+      "test_entity_hallucination_rate": 0.00891089108910891,
+      "test_entity_preservation_rate": 0.34338748592568763,
+      "test_entity_preservation_rate_lenient": 0.3451304158166866,
+      "test_entity_targeted_score": 0.21914654065830674,
+      "test_entity_targeted_score_lenient": 0.21823023253888515,
+      "test_entity_wer": 0.7114996637525218,
+      "test_entity_wer_zwnj_normalized": 0.7087723230964657
+    },
+    {
+      "epoch": 0.17946474639388024,
+      "step": 1000,
+      "test_typo_exact_match": 0.00558659217877095,
+      "test_typo_fix_rate": 0.15878247382706476,
+      "test_typo_fix_rate_lenient": 0.1542123783895674,
+      "test_typo_fix_rate_recoverable": 0.15657839428331233,
+      "test_typo_hallucination_rate": 0.00558659217877095,
+      "test_typo_preservation_rate": 0.459294975450563,
+      "test_typo_preservation_rate_lenient": 0.46072255266781914,
+      "test_typo_targeted_score": 0.3090387246388139,
+      "test_typo_targeted_score_lenient": 0.30746746552869325,
+      "test_typo_wer": 0.5876286027928825,
+      "test_typo_wer_zwnj_normalized": 0.5829312501595568
+    },
+    {
+      "epoch": 0.17946474639388024,
+      "step": 1000,
+      "test_best_wer": 0.6495641332727022
+    },
+    {
+      "epoch": 0.18125939385781906,
+      "grad_norm": 0.845633864402771,
+      "learning_rate": 9.57404608991311e-05,
+      "loss": 0.8403111457824707,
+      "step": 1010
+    },
+    {
+      "epoch": 0.18305404132175787,
+      "grad_norm": 1.0008710622787476,
+      "learning_rate": 9.564601435587457e-05,
+      "loss": 0.8566953659057617,
+      "step": 1020
+    },
+    {
+      "epoch": 0.18484868878569666,
+      "grad_norm": 1.3020329475402832,
+      "learning_rate": 9.555156781261806e-05,
+      "loss": 0.8532472610473633,
+      "step": 1030
+    },
+    {
+      "epoch": 0.18664333624963547,
+      "grad_norm": 1.4005179405212402,
+      "learning_rate": 9.545712126936155e-05,
+      "loss": 0.8881989479064941,
+      "step": 1040
+    },
+    {
+      "epoch": 0.18843798371357426,
+      "grad_norm": 4.764884948730469,
+      "learning_rate": 9.536267472610502e-05,
+      "loss": 1.096081829071045,
+      "step": 1050
+    },
+    {
+      "epoch": 0.19023263117751307,
+      "grad_norm": 0.8018057346343994,
+      "learning_rate": 9.526822818284851e-05,
+      "loss": 0.8475461006164551,
+      "step": 1060
+    },
+    {
+      "epoch": 0.19202727864145186,
+      "grad_norm": 0.9145928621292114,
+      "learning_rate": 9.517378163959199e-05,
+      "loss": 0.8785548210144043,
+      "step": 1070
+    },
+    {
+      "epoch": 0.19382192610539067,
+      "grad_norm": 1.1015050411224365,
+      "learning_rate": 9.507933509633547e-05,
+      "loss": 0.8983614921569825,
+      "step": 1080
+    },
+    {
+      "epoch": 0.1956165735693295,
+      "grad_norm": 1.5052039623260498,
+      "learning_rate": 9.498488855307897e-05,
+      "loss": 0.991839599609375,
+      "step": 1090
+    },
+    {
+      "epoch": 0.19741122103326827,
+      "grad_norm": 4.894501686096191,
+      "learning_rate": 9.489044200982244e-05,
+      "loss": 1.215499210357666,
+      "step": 1100
+    },
+    {
+      "epoch": 0.1992058684972071,
+      "grad_norm": 0.8113166689872742,
+      "learning_rate": 9.479599546656593e-05,
+      "loss": 0.9022952079772949,
+      "step": 1110
+    },
+    {
+      "epoch": 0.20100051596114588,
+      "grad_norm": 0.9681263566017151,
+      "learning_rate": 9.470154892330941e-05,
+      "loss": 0.8785144805908203,
+      "step": 1120
+    },
+    {
+      "epoch": 0.2027951634250847,
+      "grad_norm": 1.053642988204956,
+      "learning_rate": 9.460710238005289e-05,
+      "loss": 0.8290045738220215,
+      "step": 1130
+    },
+    {
+      "epoch": 0.20458981088902348,
+      "grad_norm": 1.5265758037567139,
+      "learning_rate": 9.451265583679639e-05,
+      "loss": 0.9883374214172364,
+      "step": 1140
+    },
+    {
+      "epoch": 0.2063844583529623,
+      "grad_norm": 3.2777318954467773,
+      "learning_rate": 9.441820929353986e-05,
+      "loss": 1.108902931213379,
+      "step": 1150
+    },
+    {
+      "epoch": 0.2081791058169011,
+      "grad_norm": 0.7957113981246948,
+      "learning_rate": 9.432376275028334e-05,
+      "loss": 0.8370280265808105,
+      "step": 1160
+    },
+    {
+      "epoch": 0.2099737532808399,
+      "grad_norm": 0.9634260535240173,
+      "learning_rate": 9.422931620702683e-05,
+      "loss": 0.828273868560791,
+      "step": 1170
+    },
+    {
+      "epoch": 0.2117684007447787,
+      "grad_norm": 1.2643507719039917,
+      "learning_rate": 9.413486966377031e-05,
+      "loss": 0.9653972625732422,
+      "step": 1180
+    },
+    {
+      "epoch": 0.2135630482087175,
+      "grad_norm": 1.4930179119110107,
+      "learning_rate": 9.404042312051379e-05,
+      "loss": 1.0028544425964356,
+      "step": 1190
+    },
+    {
+      "epoch": 0.2153576956726563,
+      "grad_norm": 4.001398086547852,
+      "learning_rate": 9.394597657725728e-05,
+      "loss": 1.049579906463623,
+      "step": 1200
+    },
+    {
+      "epoch": 0.2171523431365951,
+      "grad_norm": 0.7326865792274475,
+      "learning_rate": 9.385153003400076e-05,
+      "loss": 0.8290506362915039,
+      "step": 1210
+    },
+    {
+      "epoch": 0.2189469906005339,
+      "grad_norm": 0.9118326902389526,
+      "learning_rate": 9.375708349074424e-05,
+      "loss": 0.7963441371917724,
+      "step": 1220
+    },
+    {
+      "epoch": 0.22074163806447272,
+      "grad_norm": 1.1683183908462524,
+      "learning_rate": 9.366263694748773e-05,
+      "loss": 0.8069793701171875,
+      "step": 1230
+    },
+    {
+      "epoch": 0.2225362855284115,
+      "grad_norm": 1.5301281213760376,
+      "learning_rate": 9.356819040423121e-05,
+      "loss": 0.9929439544677734,
+      "step": 1240
+    },
+    {
+      "epoch": 0.22433093299235032,
+      "grad_norm": 3.8319618701934814,
+      "learning_rate": 9.347374386097469e-05,
+      "loss": 1.085425853729248,
+      "step": 1250
+    },
+    {
+      "epoch": 0.2261255804562891,
+      "grad_norm": 0.8658831119537354,
+      "learning_rate": 9.337929731771818e-05,
+      "loss": 0.8651467323303222,
+      "step": 1260
+    },
+    {
+      "epoch": 0.22792022792022792,
+      "grad_norm": 0.9741929769515991,
+      "learning_rate": 9.328485077446166e-05,
+      "loss": 0.7978952884674072,
+      "step": 1270
+    },
+    {
+      "epoch": 0.2297148753841667,
+      "grad_norm": 0.9907529354095459,
+      "learning_rate": 9.319040423120514e-05,
+      "loss": 0.8779685974121094,
+      "step": 1280
+    },
+    {
+      "epoch": 0.23150952284810553,
+      "grad_norm": 1.577710747718811,
+      "learning_rate": 9.309595768794863e-05,
+      "loss": 0.9846627235412597,
+      "step": 1290
+    },
+    {
+      "epoch": 0.23330417031204434,
+      "grad_norm": 4.1397552490234375,
+      "learning_rate": 9.300151114469211e-05,
+      "loss": 1.1263571739196778,
+      "step": 1300
+    },
+    {
+      "epoch": 0.23509881777598313,
+      "grad_norm": 0.898070752620697,
+      "learning_rate": 9.29070646014356e-05,
+      "loss": 0.8741483688354492,
+      "step": 1310
+    },
+    {
+      "epoch": 0.23689346523992194,
+      "grad_norm": 0.951369047164917,
+      "learning_rate": 9.281261805817908e-05,
+      "loss": 0.9030444145202636,
+      "step": 1320
+    },
+    {
+      "epoch": 0.23868811270386073,
+      "grad_norm": 1.1681045293807983,
+      "learning_rate": 9.271817151492256e-05,
+      "loss": 0.8972801208496094,
+      "step": 1330
+    },
+    {
+      "epoch": 0.24048276016779954,
+      "grad_norm": 1.5000301599502563,
+      "learning_rate": 9.262372497166604e-05,
+      "loss": 0.9021479606628418,
+      "step": 1340
+    },
+    {
+      "epoch": 0.24227740763173833,
+      "grad_norm": 4.603233337402344,
+      "learning_rate": 9.252927842840953e-05,
+      "loss": 0.9616855621337891,
+      "step": 1350
+    },
+    {
+      "epoch": 0.24407205509567714,
+      "grad_norm": 0.7789506912231445,
+      "learning_rate": 9.243483188515301e-05,
+      "loss": 0.8278372764587403,
+      "step": 1360
+    },
+    {
+      "epoch": 0.24586670255961596,
+      "grad_norm": 0.9593271017074585,
+      "learning_rate": 9.23403853418965e-05,
+      "loss": 0.8673165321350098,
+      "step": 1370
+    },
+    {
+      "epoch": 0.24766135002355474,
+      "grad_norm": 1.2493950128555298,
+      "learning_rate": 9.224593879863998e-05,
+      "loss": 0.8802228927612304,
+      "step": 1380
+    },
+    {
+      "epoch": 0.24945599748749356,
+      "grad_norm": 1.4462804794311523,
+      "learning_rate": 9.215149225538345e-05,
+      "loss": 0.925480842590332,
+      "step": 1390
+    },
+    {
+      "epoch": 0.25125064495143234,
+      "grad_norm": 3.787916660308838,
+      "learning_rate": 9.205704571212694e-05,
+      "loss": 1.1019389152526855,
+      "step": 1400
+    },
+    {
+      "epoch": 0.25304529241537116,
+      "grad_norm": 0.8386228084564209,
+      "learning_rate": 9.196259916887043e-05,
+      "loss": 0.8958368301391602,
+      "step": 1410
+    },
+    {
+      "epoch": 0.25483993987931,
+      "grad_norm": 0.9733511805534363,
+      "learning_rate": 9.186815262561391e-05,
+      "loss": 0.7509948253631592,
+      "step": 1420
+    },
+    {
+      "epoch": 0.2566345873432488,
+      "grad_norm": 1.0639692544937134,
+      "learning_rate": 9.17737060823574e-05,
+      "loss": 0.8474295616149903,
+      "step": 1430
+    },
+    {
+      "epoch": 0.25842923480718755,
+      "grad_norm": 1.7223182916641235,
+      "learning_rate": 9.167925953910086e-05,
+      "loss": 0.9051410675048828,
+      "step": 1440
+    },
+    {
+      "epoch": 0.26022388227112636,
+      "grad_norm": 3.473262310028076,
+      "learning_rate": 9.158481299584436e-05,
+      "loss": 0.9940505027770996,
+      "step": 1450
+    },
+    {
+      "epoch": 0.2620185297350652,
+      "grad_norm": 0.8055341243743896,
+      "learning_rate": 9.149036645258784e-05,
+      "loss": 0.8780458450317383,
+      "step": 1460
+    },
+    {
+      "epoch": 0.263813177199004,
+      "grad_norm": 1.0771476030349731,
+      "learning_rate": 9.139591990933131e-05,
+      "loss": 0.8636789321899414,
+      "step": 1470
+    },
+    {
+      "epoch": 0.26560782466294275,
+      "grad_norm": 1.1147541999816895,
+      "learning_rate": 9.130147336607481e-05,
+      "loss": 0.8464085578918457,
+      "step": 1480
+    },
+    {
+      "epoch": 0.26740247212688156,
+      "grad_norm": 1.4540315866470337,
+      "learning_rate": 9.120702682281828e-05,
+      "loss": 0.8976149559020996,
+      "step": 1490
+    },
+    {
+      "epoch": 0.2691971195908204,
+      "grad_norm": 4.394834995269775,
+      "learning_rate": 9.111258027956178e-05,
+      "loss": 1.1389759063720704,
+      "step": 1500
+    },
+    {
+      "epoch": 0.2709917670547592,
+      "grad_norm": 0.8100571036338806,
+      "learning_rate": 9.101813373630526e-05,
+      "loss": 0.7912180900573731,
+      "step": 1510
+    },
+    {
+      "epoch": 0.272786414518698,
+      "grad_norm": 0.9471364617347717,
+      "learning_rate": 9.092368719304873e-05,
+      "loss": 0.8603667259216309,
+      "step": 1520
+    },
+    {
+      "epoch": 0.27458106198263676,
+      "grad_norm": 1.1160653829574585,
+      "learning_rate": 9.082924064979223e-05,
+      "loss": 0.8775869369506836,
+      "step": 1530
+    },
+    {
+      "epoch": 0.2763757094465756,
+      "grad_norm": 1.5980632305145264,
+      "learning_rate": 9.07347941065357e-05,
+      "loss": 0.9348011970520019,
+      "step": 1540
+    },
+    {
+      "epoch": 0.2781703569105144,
+      "grad_norm": 4.296586036682129,
+      "learning_rate": 9.06403475632792e-05,
+      "loss": 1.0446084976196288,
+      "step": 1550
+    },
+    {
+      "epoch": 0.2799650043744532,
+      "grad_norm": 0.8044270873069763,
+      "learning_rate": 9.054590102002267e-05,
+      "loss": 0.8284408569335937,
+      "step": 1560
+    },
+    {
+      "epoch": 0.281759651838392,
+      "grad_norm": 0.9065347909927368,
+      "learning_rate": 9.045145447676615e-05,
+      "loss": 0.8259025573730469,
+      "step": 1570
+    },
+    {
+      "epoch": 0.2835542993023308,
+      "grad_norm": 1.1605452299118042,
+      "learning_rate": 9.035700793350965e-05,
+      "loss": 0.8801176071166992,
+      "step": 1580
+    },
+    {
+      "epoch": 0.2853489467662696,
+      "grad_norm": 1.5614631175994873,
+      "learning_rate": 9.026256139025312e-05,
+      "loss": 0.9447234153747559,
+      "step": 1590
+    },
+    {
+      "epoch": 0.2871435942302084,
+      "grad_norm": 3.539119005203247,
+      "learning_rate": 9.016811484699661e-05,
+      "loss": 1.0134998321533204,
+      "step": 1600
+    },
+    {
+      "epoch": 0.2889382416941472,
+      "grad_norm": 0.8091688752174377,
+      "learning_rate": 9.007366830374008e-05,
+      "loss": 0.8058991432189941,
+      "step": 1610
+    },
+    {
+      "epoch": 0.290732889158086,
+      "grad_norm": 0.9974891543388367,
+      "learning_rate": 8.997922176048357e-05,
+      "loss": 0.7382112979888916,
+      "step": 1620
+    },
+    {
+      "epoch": 0.2925275366220248,
+      "grad_norm": 1.2397563457489014,
+      "learning_rate": 8.988477521722706e-05,
+      "loss": 0.8194489479064941,
+      "step": 1630
+    },
+    {
+      "epoch": 0.2943221840859636,
+      "grad_norm": 1.3829917907714844,
+      "learning_rate": 8.979032867397053e-05,
+      "loss": 0.9316317558288574,
+      "step": 1640
+    },
+    {
+      "epoch": 0.2961168315499024,
+      "grad_norm": 4.598084926605225,
+      "learning_rate": 8.969588213071402e-05,
+      "loss": 1.0294858932495117,
+      "step": 1650
+    },
+    {
+      "epoch": 0.29791147901384124,
+      "grad_norm": 0.8185928463935852,
+      "learning_rate": 8.96014355874575e-05,
+      "loss": 0.8526276588439942,
+      "step": 1660
+    },
+    {
+      "epoch": 0.29970612647778,
+      "grad_norm": 0.8803004622459412,
+      "learning_rate": 8.950698904420098e-05,
+      "loss": 0.8622500419616699,
+      "step": 1670
+    },
+    {
+      "epoch": 0.3015007739417188,
+      "grad_norm": 1.0694444179534912,
+      "learning_rate": 8.941254250094448e-05,
+      "loss": 0.7721203327178955,
+      "step": 1680
+    },
+    {
+      "epoch": 0.30329542140565763,
+      "grad_norm": 1.434861421585083,
+      "learning_rate": 8.931809595768795e-05,
+      "loss": 0.9630851745605469,
+      "step": 1690
+    },
+    {
+      "epoch": 0.30509006886959644,
+      "grad_norm": 4.44917106628418,
+      "learning_rate": 8.922364941443143e-05,
+      "loss": 1.0820582389831543,
+      "step": 1700
+    },
+    {
+      "epoch": 0.30688471633353526,
+      "grad_norm": 0.8227750658988953,
+      "learning_rate": 8.912920287117492e-05,
+      "loss": 0.8339776992797852,
+      "step": 1710
+    },
+    {
+      "epoch": 0.308679363797474,
+      "grad_norm": 0.9008784294128418,
+      "learning_rate": 8.90347563279184e-05,
+      "loss": 0.7638604164123535,
+      "step": 1720
+    },
+    {
+      "epoch": 0.31047401126141283,
+      "grad_norm": 1.1473792791366577,
+      "learning_rate": 8.894030978466188e-05,
+      "loss": 0.9122580528259278,
+      "step": 1730
+    },
+    {
+      "epoch": 0.31226865872535164,
+      "grad_norm": 1.4027907848358154,
+      "learning_rate": 8.884586324140537e-05,
+      "loss": 0.9376218795776368,
+      "step": 1740
+    },
+    {
+      "epoch": 0.31406330618929046,
+      "grad_norm": 3.44396710395813,
+      "learning_rate": 8.875141669814885e-05,
+      "loss": 0.9627013206481934,
+      "step": 1750
+    },
+    {
+      "epoch": 0.3158579536532292,
+      "grad_norm": 0.761950671672821,
+      "learning_rate": 8.865697015489233e-05,
+      "loss": 0.850135612487793,
+      "step": 1760
+    },
+    {
+      "epoch": 0.31765260111716803,
+      "grad_norm": 0.9319304823875427,
+      "learning_rate": 8.856252361163582e-05,
+      "loss": 0.7397173881530762,
+      "step": 1770
+    },
+    {
+      "epoch": 0.31944724858110685,
+      "grad_norm": 1.1564598083496094,
+      "learning_rate": 8.84680770683793e-05,
+      "loss": 0.8239191055297852,
+      "step": 1780
+    },
+    {
+      "epoch": 0.32124189604504566,
+      "grad_norm": 1.4574592113494873,
+      "learning_rate": 8.837363052512278e-05,
+      "loss": 0.9266434669494629,
+      "step": 1790
+    },
+    {
+      "epoch": 0.3230365435089845,
+      "grad_norm": 4.408364295959473,
+      "learning_rate": 8.827918398186627e-05,
+      "loss": 1.0574697494506835,
+      "step": 1800
+    },
+    {
+      "epoch": 0.32483119097292323,
+      "grad_norm": 0.808427095413208,
+      "learning_rate": 8.818473743860975e-05,
+      "loss": 0.8238012313842773,
+      "step": 1810
+    },
+    {
+      "epoch": 0.32662583843686205,
+      "grad_norm": 0.8963003754615784,
+      "learning_rate": 8.809029089535323e-05,
+      "loss": 0.8102021217346191,
+      "step": 1820
+    },
+    {
+      "epoch": 0.32842048590080086,
+      "grad_norm": 1.1847169399261475,
+      "learning_rate": 8.799584435209672e-05,
+      "loss": 0.847196388244629,
+      "step": 1830
+    },
+    {
+      "epoch": 0.3302151333647397,
+      "grad_norm": 1.355234980583191,
+      "learning_rate": 8.79013978088402e-05,
+      "loss": 0.9291248321533203,
+      "step": 1840
+    },
+    {
+      "epoch": 0.3320097808286785,
+      "grad_norm": 4.019794464111328,
+      "learning_rate": 8.780695126558368e-05,
+      "loss": 0.9507856369018555,
+      "step": 1850
+    },
+    {
+      "epoch": 0.33380442829261725,
+      "grad_norm": 0.8047809600830078,
+      "learning_rate": 8.771250472232717e-05,
+      "loss": 0.8097662925720215,
+      "step": 1860
+    },
+    {
+      "epoch": 0.33559907575655606,
+      "grad_norm": 0.9057943820953369,
+      "learning_rate": 8.761805817907065e-05,
+      "loss": 0.8277406692504883,
+      "step": 1870
+    },
+    {
+      "epoch": 0.3373937232204949,
+      "grad_norm": 1.1131069660186768,
+      "learning_rate": 8.752361163581414e-05,
+      "loss": 0.875433349609375,
+      "step": 1880
+    },
+    {
+      "epoch": 0.3391883706844337,
+      "grad_norm": 1.5567286014556885,
+      "learning_rate": 8.742916509255762e-05,
+      "loss": 0.9548139572143555,
+      "step": 1890
+    },
+    {
+      "epoch": 0.34098301814837245,
+      "grad_norm": 3.943831205368042,
+      "learning_rate": 8.73347185493011e-05,
+      "loss": 1.0330249786376953,
+      "step": 1900
+    },
+    {
+      "epoch": 0.34277766561231127,
+      "grad_norm": 0.7638143301010132,
+      "learning_rate": 8.724027200604459e-05,
+      "loss": 0.804904842376709,
+      "step": 1910
+    },
+    {
+      "epoch": 0.3445723130762501,
+      "grad_norm": 0.9346854090690613,
+      "learning_rate": 8.714582546278807e-05,
+      "loss": 0.7564315795898438,
+      "step": 1920
+    },
+    {
+      "epoch": 0.3463669605401889,
+      "grad_norm": 1.1116750240325928,
+      "learning_rate": 8.705137891953155e-05,
+      "loss": 0.7835733413696289,
+      "step": 1930
+    },
+    {
+      "epoch": 0.3481616080041277,
+      "grad_norm": 1.437160611152649,
+      "learning_rate": 8.695693237627504e-05,
+      "loss": 0.9441628456115723,
+      "step": 1940
+    },
+    {
+      "epoch": 0.34995625546806647,
+      "grad_norm": 2.9362335205078125,
+      "learning_rate": 8.686248583301852e-05,
+      "loss": 1.0385236740112305,
+      "step": 1950
+    },
+    {
+      "epoch": 0.3517509029320053,
+      "grad_norm": 0.7919660210609436,
+      "learning_rate": 8.6768039289762e-05,
+      "loss": 0.7995383739471436,
+      "step": 1960
+    },
+    {
+      "epoch": 0.3535455503959441,
+      "grad_norm": 0.8578867316246033,
+      "learning_rate": 8.667359274650549e-05,
+      "loss": 0.7939890384674072,
+      "step": 1970
+    },
+    {
+      "epoch": 0.3553401978598829,
+      "grad_norm": 1.1968882083892822,
+      "learning_rate": 8.657914620324896e-05,
+      "loss": 0.817835807800293,
+      "step": 1980
+    },
+    {
+      "epoch": 0.3571348453238217,
+      "grad_norm": 1.4205067157745361,
+      "learning_rate": 8.648469965999245e-05,
+      "loss": 0.9530954360961914,
+      "step": 1990
+    },
+    {
+      "epoch": 0.3589294927877605,
+      "grad_norm": 4.1284332275390625,
+      "learning_rate": 8.639025311673594e-05,
+      "loss": 1.0853967666625977,
+      "step": 2000
     }
   ],
   "logging_steps": 10,
@@ -727,7 +1462,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.014956564093929e+17,
+  "total_flos": 2.035264485005284e+17,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null

qwen3.5-2b-50pct-masked-weighted/checkpoint-2000/README.md

@@ -0,0 +1,207 @@
+---
+base_model: Qwen/Qwen3.5-2B
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3.5-2B
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.21.0

qwen3.5-2b-50pct-masked-weighted/checkpoint-2000/adapter_config.json

@@ -0,0 +1,56 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3.5-2B",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "kasa_config": null,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "monteclora_config": null,
+  "peft_type": "LORA",
+  "peft_version": "0.21.0",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "o_proj",
+    "out_proj",
+    "v_proj",
+    "up_proj",
+    "in_proj_qkv",
+    "in_proj_b",
+    "in_proj_z",
+    "k_proj",
+    "q_proj",
+    "in_proj_a",
+    "gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false,
+  "velora_config": null
+}

qwen3.5-2b-50pct-masked-weighted/checkpoint-2000/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80b628f7e9c55852ca9b5bf5bdb6613db51cba9f9c635b749fc7b686b56ebfe1
+size 134604152

qwen3.5-2b-50pct-masked-weighted/checkpoint-2000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:754de55d9e90d6cba75eea0f9ee8bb59b2f2f25dc5f7e568c2a0a5b8269ceb06
+size 269426431

qwen3.5-2b-50pct-masked-weighted/checkpoint-2000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98f4a968e01d607786538f37082784acf14d994ddede3a162e85abf8496645a6
+size 14645

qwen3.5-2b-50pct-masked-weighted/checkpoint-2000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be32b28b4401079e06a80019d25bd8c958c4d5a19216b70fd633be2006252ebe
+size 1465

qwen3.5-2b-50pct-masked-weighted/checkpoint-2000/trainer_state.json

@@ -0,0 +1,1469 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3589294927877605,
+  "eval_steps": 500,
+  "global_step": 2000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0017946474639388024,
+      "grad_norm": 3.4698801040649414,
+      "learning_rate": 1.6129032258064516e-06,
+      "loss": 1.4609528541564942,
+      "step": 10
+    },
+    {
+      "epoch": 0.003589294927877605,
+      "grad_norm": 2.894972085952759,
+      "learning_rate": 3.405017921146954e-06,
+      "loss": 1.5060235977172851,
+      "step": 20
+    },
+    {
+      "epoch": 0.0053839423918164075,
+      "grad_norm": 2.3209266662597656,
+      "learning_rate": 5.197132616487455e-06,
+      "loss": 1.4531757354736328,
+      "step": 30
+    },
+    {
+      "epoch": 0.00717858985575521,
+      "grad_norm": 2.2354202270507812,
+      "learning_rate": 6.989247311827957e-06,
+      "loss": 1.5934924125671386,
+      "step": 40
+    },
+    {
+      "epoch": 0.008973237319694013,
+      "grad_norm": 6.718564033508301,
+      "learning_rate": 8.781362007168458e-06,
+      "loss": 1.6818513870239258,
+      "step": 50
+    },
+    {
+      "epoch": 0.010767884783632815,
+      "grad_norm": 1.1820789575576782,
+      "learning_rate": 1.0573476702508961e-05,
+      "loss": 1.2959606170654296,
+      "step": 60
+    },
+    {
+      "epoch": 0.012562532247571617,
+      "grad_norm": 1.3072574138641357,
+      "learning_rate": 1.2365591397849464e-05,
+      "loss": 1.1890168190002441,
+      "step": 70
+    },
+    {
+      "epoch": 0.01435717971151042,
+      "grad_norm": 1.6472913026809692,
+      "learning_rate": 1.4157706093189965e-05,
+      "loss": 1.2888777732849122,
+      "step": 80
+    },
+    {
+      "epoch": 0.01615182717544922,
+      "grad_norm": 2.2778408527374268,
+      "learning_rate": 1.5949820788530468e-05,
+      "loss": 1.382105255126953,
+      "step": 90
+    },
+    {
+      "epoch": 0.017946474639388026,
+      "grad_norm": 6.462679386138916,
+      "learning_rate": 1.774193548387097e-05,
+      "loss": 1.3651474952697753,
+      "step": 100
+    },
+    {
+      "epoch": 0.019741122103326826,
+      "grad_norm": 1.3327974081039429,
+      "learning_rate": 1.9534050179211473e-05,
+      "loss": 1.2019264221191406,
+      "step": 110
+    },
+    {
+      "epoch": 0.02153576956726563,
+      "grad_norm": 1.625113606452942,
+      "learning_rate": 2.132616487455197e-05,
+      "loss": 1.1140929222106934,
+      "step": 120
+    },
+    {
+      "epoch": 0.023330417031204434,
+      "grad_norm": 2.114356279373169,
+      "learning_rate": 2.3118279569892472e-05,
+      "loss": 1.194014072418213,
+      "step": 130
+    },
+    {
+      "epoch": 0.025125064495143234,
+      "grad_norm": 2.738748073577881,
+      "learning_rate": 2.4910394265232977e-05,
+      "loss": 1.310903263092041,
+      "step": 140
+    },
+    {
+      "epoch": 0.02691971195908204,
+      "grad_norm": 7.28310489654541,
+      "learning_rate": 2.670250896057348e-05,
+      "loss": 1.341808032989502,
+      "step": 150
+    },
+    {
+      "epoch": 0.02871435942302084,
+      "grad_norm": 1.4109766483306885,
+      "learning_rate": 2.8494623655913982e-05,
+      "loss": 1.0624680519104004,
+      "step": 160
+    },
+    {
+      "epoch": 0.030509006886959643,
+      "grad_norm": 1.8665562868118286,
+      "learning_rate": 3.0286738351254483e-05,
+      "loss": 1.1256374359130858,
+      "step": 170
+    },
+    {
+      "epoch": 0.03230365435089844,
+      "grad_norm": 2.103581190109253,
+      "learning_rate": 3.207885304659498e-05,
+      "loss": 1.1144988059997558,
+      "step": 180
+    },
+    {
+      "epoch": 0.03409830181483725,
+      "grad_norm": 2.7976624965667725,
+      "learning_rate": 3.387096774193548e-05,
+      "loss": 1.2549245834350586,
+      "step": 190
+    },
+    {
+      "epoch": 0.03589294927877605,
+      "grad_norm": 6.750956058502197,
+      "learning_rate": 3.566308243727599e-05,
+      "loss": 1.2815749168395996,
+      "step": 200
+    },
+    {
+      "epoch": 0.03768759674271485,
+      "grad_norm": 1.6264805793762207,
+      "learning_rate": 3.7455197132616484e-05,
+      "loss": 1.0297048568725586,
+      "step": 210
+    },
+    {
+      "epoch": 0.03948224420665365,
+      "grad_norm": 1.6889760494232178,
+      "learning_rate": 3.924731182795699e-05,
+      "loss": 1.0584090232849122,
+      "step": 220
+    },
+    {
+      "epoch": 0.04127689167059246,
+      "grad_norm": 2.0039634704589844,
+      "learning_rate": 4.1039426523297493e-05,
+      "loss": 1.085252571105957,
+      "step": 230
+    },
+    {
+      "epoch": 0.04307153913453126,
+      "grad_norm": 2.6522042751312256,
+      "learning_rate": 4.283154121863799e-05,
+      "loss": 1.2651546478271485,
+      "step": 240
+    },
+    {
+      "epoch": 0.04486618659847006,
+      "grad_norm": 8.020312309265137,
+      "learning_rate": 4.4623655913978496e-05,
+      "loss": 1.324000644683838,
+      "step": 250
+    },
+    {
+      "epoch": 0.04666083406240887,
+      "grad_norm": 1.7270288467407227,
+      "learning_rate": 4.6415770609319e-05,
+      "loss": 1.0575682640075683,
+      "step": 260
+    },
+    {
+      "epoch": 0.04845548152634767,
+      "grad_norm": 1.5692397356033325,
+      "learning_rate": 4.82078853046595e-05,
+      "loss": 1.0243738174438477,
+      "step": 270
+    },
+    {
+      "epoch": 0.05025012899028647,
+      "grad_norm": 2.029719591140747,
+      "learning_rate": 5e-05,
+      "loss": 1.0656013488769531,
+      "step": 280
+    },
+    {
+      "epoch": 0.052044776454225276,
+      "grad_norm": 2.4044599533081055,
+      "learning_rate": 5.17921146953405e-05,
+      "loss": 1.1308915138244628,
+      "step": 290
+    },
+    {
+      "epoch": 0.05383942391816408,
+      "grad_norm": 7.324768543243408,
+      "learning_rate": 5.3584229390681005e-05,
+      "loss": 1.2482977867126466,
+      "step": 300
+    },
+    {
+      "epoch": 0.05563407138210288,
+      "grad_norm": 1.4643646478652954,
+      "learning_rate": 5.53763440860215e-05,
+      "loss": 0.984095287322998,
+      "step": 310
+    },
+    {
+      "epoch": 0.05742871884604168,
+      "grad_norm": 1.6681671142578125,
+      "learning_rate": 5.7168458781362014e-05,
+      "loss": 1.011151123046875,
+      "step": 320
+    },
+    {
+      "epoch": 0.059223366309980485,
+      "grad_norm": 1.8012951612472534,
+      "learning_rate": 5.8960573476702505e-05,
+      "loss": 1.0619875907897949,
+      "step": 330
+    },
+    {
+      "epoch": 0.061018013773919286,
+      "grad_norm": 2.7174854278564453,
+      "learning_rate": 6.0752688172043016e-05,
+      "loss": 1.1800737380981445,
+      "step": 340
+    },
+    {
+      "epoch": 0.06281266123785809,
+      "grad_norm": 7.283389568328857,
+      "learning_rate": 6.254480286738351e-05,
+      "loss": 1.2749577522277833,
+      "step": 350
+    },
+    {
+      "epoch": 0.06460730870179689,
+      "grad_norm": 1.4635604619979858,
+      "learning_rate": 6.433691756272402e-05,
+      "loss": 1.0051876068115235,
+      "step": 360
+    },
+    {
+      "epoch": 0.06640195616573569,
+      "grad_norm": 1.4372541904449463,
+      "learning_rate": 6.612903225806452e-05,
+      "loss": 1.0501286506652832,
+      "step": 370
+    },
+    {
+      "epoch": 0.0681966036296745,
+      "grad_norm": 1.5717954635620117,
+      "learning_rate": 6.792114695340503e-05,
+      "loss": 1.0269421577453612,
+      "step": 380
+    },
+    {
+      "epoch": 0.0699912510936133,
+      "grad_norm": 2.4193785190582275,
+      "learning_rate": 6.971326164874552e-05,
+      "loss": 1.018425464630127,
+      "step": 390
+    },
+    {
+      "epoch": 0.0717858985575521,
+      "grad_norm": 5.167453289031982,
+      "learning_rate": 7.150537634408602e-05,
+      "loss": 1.2043580055236816,
+      "step": 400
+    },
+    {
+      "epoch": 0.0735805460214909,
+      "grad_norm": 1.1322799921035767,
+      "learning_rate": 7.329749103942653e-05,
+      "loss": 0.954345417022705,
+      "step": 410
+    },
+    {
+      "epoch": 0.0753751934854297,
+      "grad_norm": 1.3257076740264893,
+      "learning_rate": 7.508960573476703e-05,
+      "loss": 0.9895898818969726,
+      "step": 420
+    },
+    {
+      "epoch": 0.0771698409493685,
+      "grad_norm": 1.317622184753418,
+      "learning_rate": 7.688172043010752e-05,
+      "loss": 0.9647544860839844,
+      "step": 430
+    },
+    {
+      "epoch": 0.0789644884133073,
+      "grad_norm": 1.8667112588882446,
+      "learning_rate": 7.867383512544804e-05,
+      "loss": 1.139955997467041,
+      "step": 440
+    },
+    {
+      "epoch": 0.08075913587724612,
+      "grad_norm": 5.289592266082764,
+      "learning_rate": 8.046594982078853e-05,
+      "loss": 1.1954397201538085,
+      "step": 450
+    },
+    {
+      "epoch": 0.08255378334118492,
+      "grad_norm": 1.0380669832229614,
+      "learning_rate": 8.225806451612904e-05,
+      "loss": 0.9445084571838379,
+      "step": 460
+    },
+    {
+      "epoch": 0.08434843080512372,
+      "grad_norm": 1.3310823440551758,
+      "learning_rate": 8.405017921146954e-05,
+      "loss": 0.9562747955322266,
+      "step": 470
+    },
+    {
+      "epoch": 0.08614307826906252,
+      "grad_norm": 1.417316198348999,
+      "learning_rate": 8.584229390681004e-05,
+      "loss": 0.9948525428771973,
+      "step": 480
+    },
+    {
+      "epoch": 0.08793772573300132,
+      "grad_norm": 1.8197181224822998,
+      "learning_rate": 8.763440860215054e-05,
+      "loss": 1.0975610733032226,
+      "step": 490
+    },
+    {
+      "epoch": 0.08973237319694012,
+      "grad_norm": 4.941095352172852,
+      "learning_rate": 8.942652329749104e-05,
+      "loss": 1.2107452392578124,
+      "step": 500
+    },
+    {
+      "epoch": 0.09152702066087894,
+      "grad_norm": 0.9819939732551575,
+      "learning_rate": 9.121863799283154e-05,
+      "loss": 0.9593785285949707,
+      "step": 510
+    },
+    {
+      "epoch": 0.09332166812481774,
+      "grad_norm": 1.1892472505569458,
+      "learning_rate": 9.301075268817204e-05,
+      "loss": 0.9651047706604003,
+      "step": 520
+    },
+    {
+      "epoch": 0.09511631558875654,
+      "grad_norm": 1.4324865341186523,
+      "learning_rate": 9.480286738351255e-05,
+      "loss": 0.9062895774841309,
+      "step": 530
+    },
+    {
+      "epoch": 0.09691096305269534,
+      "grad_norm": 1.6410809755325317,
+      "learning_rate": 9.659498207885304e-05,
+      "loss": 1.0246500968933105,
+      "step": 540
+    },
+    {
+      "epoch": 0.09870561051663414,
+      "grad_norm": 5.0956501960754395,
+      "learning_rate": 9.838709677419355e-05,
+      "loss": 1.1241219520568848,
+      "step": 550
+    },
+    {
+      "epoch": 0.10050025798057294,
+      "grad_norm": 0.951851487159729,
+      "learning_rate": 9.999055534567435e-05,
+      "loss": 0.9464516639709473,
+      "step": 560
+    },
+    {
+      "epoch": 0.10229490544451174,
+      "grad_norm": 1.1737598180770874,
+      "learning_rate": 9.989610880241783e-05,
+      "loss": 0.907046890258789,
+      "step": 570
+    },
+    {
+      "epoch": 0.10408955290845055,
+      "grad_norm": 1.3924814462661743,
+      "learning_rate": 9.980166225916132e-05,
+      "loss": 0.9977164268493652,
+      "step": 580
+    },
+    {
+      "epoch": 0.10588420037238935,
+      "grad_norm": 1.6580984592437744,
+      "learning_rate": 9.97072157159048e-05,
+      "loss": 1.0950872421264648,
+      "step": 590
+    },
+    {
+      "epoch": 0.10767884783632815,
+      "grad_norm": 4.493084907531738,
+      "learning_rate": 9.961276917264828e-05,
+      "loss": 1.211144256591797,
+      "step": 600
+    },
+    {
+      "epoch": 0.10947349530026695,
+      "grad_norm": 0.8613011240959167,
+      "learning_rate": 9.951832262939177e-05,
+      "loss": 0.9501323699951172,
+      "step": 610
+    },
+    {
+      "epoch": 0.11126814276420575,
+      "grad_norm": 1.1255444288253784,
+      "learning_rate": 9.942387608613525e-05,
+      "loss": 0.903736686706543,
+      "step": 620
+    },
+    {
+      "epoch": 0.11306279022814456,
+      "grad_norm": 1.4082947969436646,
+      "learning_rate": 9.932942954287873e-05,
+      "loss": 1.0084752082824706,
+      "step": 630
+    },
+    {
+      "epoch": 0.11485743769208336,
+      "grad_norm": 1.652286171913147,
+      "learning_rate": 9.923498299962222e-05,
+      "loss": 0.930430793762207,
+      "step": 640
+    },
+    {
+      "epoch": 0.11665208515602217,
+      "grad_norm": 5.117873668670654,
+      "learning_rate": 9.91405364563657e-05,
+      "loss": 1.091403865814209,
+      "step": 650
+    },
+    {
+      "epoch": 0.11844673261996097,
+      "grad_norm": 1.0188069343566895,
+      "learning_rate": 9.904608991310918e-05,
+      "loss": 0.883120059967041,
+      "step": 660
+    },
+    {
+      "epoch": 0.12024138008389977,
+      "grad_norm": 1.1442457437515259,
+      "learning_rate": 9.895164336985267e-05,
+      "loss": 0.9172245979309082,
+      "step": 670
+    },
+    {
+      "epoch": 0.12203602754783857,
+      "grad_norm": 1.3536146879196167,
+      "learning_rate": 9.885719682659615e-05,
+      "loss": 0.9197479248046875,
+      "step": 680
+    },
+    {
+      "epoch": 0.12383067501177737,
+      "grad_norm": 1.8869231939315796,
+      "learning_rate": 9.876275028333963e-05,
+      "loss": 1.0764457702636718,
+      "step": 690
+    },
+    {
+      "epoch": 0.12562532247571617,
+      "grad_norm": 3.6568493843078613,
+      "learning_rate": 9.866830374008312e-05,
+      "loss": 1.0616437911987304,
+      "step": 700
+    },
+    {
+      "epoch": 0.127419969939655,
+      "grad_norm": 0.8427063822746277,
+      "learning_rate": 9.85738571968266e-05,
+      "loss": 0.8856773376464844,
+      "step": 710
+    },
+    {
+      "epoch": 0.12921461740359377,
+      "grad_norm": 1.0265027284622192,
+      "learning_rate": 9.847941065357009e-05,
+      "loss": 0.907376766204834,
+      "step": 720
+    },
+    {
+      "epoch": 0.1310092648675326,
+      "grad_norm": 1.3209590911865234,
+      "learning_rate": 9.838496411031357e-05,
+      "loss": 1.0000056266784667,
+      "step": 730
+    },
+    {
+      "epoch": 0.13280391233147137,
+      "grad_norm": 1.7275080680847168,
+      "learning_rate": 9.829051756705705e-05,
+      "loss": 1.1050853729248047,
+      "step": 740
+    },
+    {
+      "epoch": 0.1345985597954102,
+      "grad_norm": 4.58080530166626,
+      "learning_rate": 9.819607102380054e-05,
+      "loss": 1.1597484588623046,
+      "step": 750
+    },
+    {
+      "epoch": 0.136393207259349,
+      "grad_norm": 0.8872320055961609,
+      "learning_rate": 9.810162448054402e-05,
+      "loss": 0.917822265625,
+      "step": 760
+    },
+    {
+      "epoch": 0.1381878547232878,
+      "grad_norm": 0.9287213087081909,
+      "learning_rate": 9.80071779372875e-05,
+      "loss": 0.8946282386779785,
+      "step": 770
+    },
+    {
+      "epoch": 0.1399825021872266,
+      "grad_norm": 1.183867335319519,
+      "learning_rate": 9.791273139403099e-05,
+      "loss": 0.9173077583312989,
+      "step": 780
+    },
+    {
+      "epoch": 0.1417771496511654,
+      "grad_norm": 1.6546437740325928,
+      "learning_rate": 9.781828485077447e-05,
+      "loss": 0.962445068359375,
+      "step": 790
+    },
+    {
+      "epoch": 0.1435717971151042,
+      "grad_norm": 4.307088375091553,
+      "learning_rate": 9.772383830751794e-05,
+      "loss": 1.1235834121704102,
+      "step": 800
+    },
+    {
+      "epoch": 0.145366444579043,
+      "grad_norm": 0.8552832007408142,
+      "learning_rate": 9.762939176426144e-05,
+      "loss": 0.8700121879577637,
+      "step": 810
+    },
+    {
+      "epoch": 0.1471610920429818,
+      "grad_norm": 0.9782463908195496,
+      "learning_rate": 9.753494522100492e-05,
+      "loss": 0.8827217102050782,
+      "step": 820
+    },
+    {
+      "epoch": 0.14895573950692062,
+      "grad_norm": 1.2572635412216187,
+      "learning_rate": 9.74404986777484e-05,
+      "loss": 0.9147042274475098,
+      "step": 830
+    },
+    {
+      "epoch": 0.1507503869708594,
+      "grad_norm": 1.6980787515640259,
+      "learning_rate": 9.734605213449189e-05,
+      "loss": 1.0121471405029296,
+      "step": 840
+    },
+    {
+      "epoch": 0.15254503443479822,
+      "grad_norm": 4.741994380950928,
+      "learning_rate": 9.725160559123536e-05,
+      "loss": 1.070129680633545,
+      "step": 850
+    },
+    {
+      "epoch": 0.154339681898737,
+      "grad_norm": 0.7885052561759949,
+      "learning_rate": 9.715715904797885e-05,
+      "loss": 0.9009537696838379,
+      "step": 860
+    },
+    {
+      "epoch": 0.15613432936267582,
+      "grad_norm": 0.9653502106666565,
+      "learning_rate": 9.706271250472234e-05,
+      "loss": 0.8701554298400879,
+      "step": 870
+    },
+    {
+      "epoch": 0.1579289768266146,
+      "grad_norm": 1.2959074974060059,
+      "learning_rate": 9.69682659614658e-05,
+      "loss": 0.9180045127868652,
+      "step": 880
+    },
+    {
+      "epoch": 0.15972362429055342,
+      "grad_norm": 1.5041567087173462,
+      "learning_rate": 9.68738194182093e-05,
+      "loss": 0.9027626037597656,
+      "step": 890
+    },
+    {
+      "epoch": 0.16151827175449224,
+      "grad_norm": 5.474041938781738,
+      "learning_rate": 9.677937287495277e-05,
+      "loss": 1.0739646911621095,
+      "step": 900
+    },
+    {
+      "epoch": 0.16331291921843102,
+      "grad_norm": 0.933229923248291,
+      "learning_rate": 9.668492633169627e-05,
+      "loss": 0.881598949432373,
+      "step": 910
+    },
+    {
+      "epoch": 0.16510756668236984,
+      "grad_norm": 0.997329592704773,
+      "learning_rate": 9.659047978843975e-05,
+      "loss": 0.8712461471557618,
+      "step": 920
+    },
+    {
+      "epoch": 0.16690221414630863,
+      "grad_norm": 1.319835901260376,
+      "learning_rate": 9.649603324518322e-05,
+      "loss": 0.8690865516662598,
+      "step": 930
+    },
+    {
+      "epoch": 0.16869686161024744,
+      "grad_norm": 1.8019057512283325,
+      "learning_rate": 9.640158670192672e-05,
+      "loss": 1.1501221656799316,
+      "step": 940
+    },
+    {
+      "epoch": 0.17049150907418623,
+      "grad_norm": 3.8484811782836914,
+      "learning_rate": 9.630714015867019e-05,
+      "loss": 1.1916112899780273,
+      "step": 950
+    },
+    {
+      "epoch": 0.17228615653812504,
+      "grad_norm": 0.8143914937973022,
+      "learning_rate": 9.621269361541369e-05,
+      "loss": 0.91885404586792,
+      "step": 960
+    },
+    {
+      "epoch": 0.17408080400206385,
+      "grad_norm": 1.0034821033477783,
+      "learning_rate": 9.611824707215716e-05,
+      "loss": 0.8618650436401367,
+      "step": 970
+    },
+    {
+      "epoch": 0.17587545146600264,
+      "grad_norm": 1.3009910583496094,
+      "learning_rate": 9.602380052890064e-05,
+      "loss": 0.8944579124450683,
+      "step": 980
+    },
+    {
+      "epoch": 0.17767009892994146,
+      "grad_norm": 1.4715969562530518,
+      "learning_rate": 9.592935398564414e-05,
+      "loss": 1.0547027587890625,
+      "step": 990
+    },
+    {
+      "epoch": 0.17946474639388024,
+      "grad_norm": 4.630854606628418,
+      "learning_rate": 9.583490744238761e-05,
+      "loss": 1.1706391334533692,
+      "step": 1000
+    },
+    {
+      "epoch": 0.17946474639388024,
+      "step": 1000,
+      "test_entity_exact_match": 0.0019801980198019802,
+      "test_entity_fix_rate": 0.09490559539092588,
+      "test_entity_fix_rate_lenient": 0.09133004926108375,
+      "test_entity_fix_rate_recoverable": 0.09265880308751295,
+      "test_entity_hallucination_rate": 0.00891089108910891,
+      "test_entity_preservation_rate": 0.34338748592568763,
+      "test_entity_preservation_rate_lenient": 0.3451304158166866,
+      "test_entity_targeted_score": 0.21914654065830674,
+      "test_entity_targeted_score_lenient": 0.21823023253888515,
+      "test_entity_wer": 0.7114996637525218,
+      "test_entity_wer_zwnj_normalized": 0.7087723230964657
+    },
+    {
+      "epoch": 0.17946474639388024,
+      "step": 1000,
+      "test_typo_exact_match": 0.00558659217877095,
+      "test_typo_fix_rate": 0.15878247382706476,
+      "test_typo_fix_rate_lenient": 0.1542123783895674,
+      "test_typo_fix_rate_recoverable": 0.15657839428331233,
+      "test_typo_hallucination_rate": 0.00558659217877095,
+      "test_typo_preservation_rate": 0.459294975450563,
+      "test_typo_preservation_rate_lenient": 0.46072255266781914,
+      "test_typo_targeted_score": 0.3090387246388139,
+      "test_typo_targeted_score_lenient": 0.30746746552869325,
+      "test_typo_wer": 0.5876286027928825,
+      "test_typo_wer_zwnj_normalized": 0.5829312501595568
+    },
+    {
+      "epoch": 0.17946474639388024,
+      "step": 1000,
+      "test_best_wer": 0.6495641332727022
+    },
+    {
+      "epoch": 0.18125939385781906,
+      "grad_norm": 0.845633864402771,
+      "learning_rate": 9.57404608991311e-05,
+      "loss": 0.8403111457824707,
+      "step": 1010
+    },
+    {
+      "epoch": 0.18305404132175787,
+      "grad_norm": 1.0008710622787476,
+      "learning_rate": 9.564601435587457e-05,
+      "loss": 0.8566953659057617,
+      "step": 1020
+    },
+    {
+      "epoch": 0.18484868878569666,
+      "grad_norm": 1.3020329475402832,
+      "learning_rate": 9.555156781261806e-05,
+      "loss": 0.8532472610473633,
+      "step": 1030
+    },
+    {
+      "epoch": 0.18664333624963547,
+      "grad_norm": 1.4005179405212402,
+      "learning_rate": 9.545712126936155e-05,
+      "loss": 0.8881989479064941,
+      "step": 1040
+    },
+    {
+      "epoch": 0.18843798371357426,
+      "grad_norm": 4.764884948730469,
+      "learning_rate": 9.536267472610502e-05,
+      "loss": 1.096081829071045,
+      "step": 1050
+    },
+    {
+      "epoch": 0.19023263117751307,
+      "grad_norm": 0.8018057346343994,
+      "learning_rate": 9.526822818284851e-05,
+      "loss": 0.8475461006164551,
+      "step": 1060
+    },
+    {
+      "epoch": 0.19202727864145186,
+      "grad_norm": 0.9145928621292114,
+      "learning_rate": 9.517378163959199e-05,
+      "loss": 0.8785548210144043,
+      "step": 1070
+    },
+    {
+      "epoch": 0.19382192610539067,
+      "grad_norm": 1.1015050411224365,
+      "learning_rate": 9.507933509633547e-05,
+      "loss": 0.8983614921569825,
+      "step": 1080
+    },
+    {
+      "epoch": 0.1956165735693295,
+      "grad_norm": 1.5052039623260498,
+      "learning_rate": 9.498488855307897e-05,
+      "loss": 0.991839599609375,
+      "step": 1090
+    },
+    {
+      "epoch": 0.19741122103326827,
+      "grad_norm": 4.894501686096191,
+      "learning_rate": 9.489044200982244e-05,
+      "loss": 1.215499210357666,
+      "step": 1100
+    },
+    {
+      "epoch": 0.1992058684972071,
+      "grad_norm": 0.8113166689872742,
+      "learning_rate": 9.479599546656593e-05,
+      "loss": 0.9022952079772949,
+      "step": 1110
+    },
+    {
+      "epoch": 0.20100051596114588,
+      "grad_norm": 0.9681263566017151,
+      "learning_rate": 9.470154892330941e-05,
+      "loss": 0.8785144805908203,
+      "step": 1120
+    },
+    {
+      "epoch": 0.2027951634250847,
+      "grad_norm": 1.053642988204956,
+      "learning_rate": 9.460710238005289e-05,
+      "loss": 0.8290045738220215,
+      "step": 1130
+    },
+    {
+      "epoch": 0.20458981088902348,
+      "grad_norm": 1.5265758037567139,
+      "learning_rate": 9.451265583679639e-05,
+      "loss": 0.9883374214172364,
+      "step": 1140
+    },
+    {
+      "epoch": 0.2063844583529623,
+      "grad_norm": 3.2777318954467773,
+      "learning_rate": 9.441820929353986e-05,
+      "loss": 1.108902931213379,
+      "step": 1150
+    },
+    {
+      "epoch": 0.2081791058169011,
+      "grad_norm": 0.7957113981246948,
+      "learning_rate": 9.432376275028334e-05,
+      "loss": 0.8370280265808105,
+      "step": 1160
+    },
+    {
+      "epoch": 0.2099737532808399,
+      "grad_norm": 0.9634260535240173,
+      "learning_rate": 9.422931620702683e-05,
+      "loss": 0.828273868560791,
+      "step": 1170
+    },
+    {
+      "epoch": 0.2117684007447787,
+      "grad_norm": 1.2643507719039917,
+      "learning_rate": 9.413486966377031e-05,
+      "loss": 0.9653972625732422,
+      "step": 1180
+    },
+    {
+      "epoch": 0.2135630482087175,
+      "grad_norm": 1.4930179119110107,
+      "learning_rate": 9.404042312051379e-05,
+      "loss": 1.0028544425964356,
+      "step": 1190
+    },
+    {
+      "epoch": 0.2153576956726563,
+      "grad_norm": 4.001398086547852,
+      "learning_rate": 9.394597657725728e-05,
+      "loss": 1.049579906463623,
+      "step": 1200
+    },
+    {
+      "epoch": 0.2171523431365951,
+      "grad_norm": 0.7326865792274475,
+      "learning_rate": 9.385153003400076e-05,
+      "loss": 0.8290506362915039,
+      "step": 1210
+    },
+    {
+      "epoch": 0.2189469906005339,
+      "grad_norm": 0.9118326902389526,
+      "learning_rate": 9.375708349074424e-05,
+      "loss": 0.7963441371917724,
+      "step": 1220
+    },
+    {
+      "epoch": 0.22074163806447272,
+      "grad_norm": 1.1683183908462524,
+      "learning_rate": 9.366263694748773e-05,
+      "loss": 0.8069793701171875,
+      "step": 1230
+    },
+    {
+      "epoch": 0.2225362855284115,
+      "grad_norm": 1.5301281213760376,
+      "learning_rate": 9.356819040423121e-05,
+      "loss": 0.9929439544677734,
+      "step": 1240
+    },
+    {
+      "epoch": 0.22433093299235032,
+      "grad_norm": 3.8319618701934814,
+      "learning_rate": 9.347374386097469e-05,
+      "loss": 1.085425853729248,
+      "step": 1250
+    },
+    {
+      "epoch": 0.2261255804562891,
+      "grad_norm": 0.8658831119537354,
+      "learning_rate": 9.337929731771818e-05,
+      "loss": 0.8651467323303222,
+      "step": 1260
+    },
+    {
+      "epoch": 0.22792022792022792,
+      "grad_norm": 0.9741929769515991,
+      "learning_rate": 9.328485077446166e-05,
+      "loss": 0.7978952884674072,
+      "step": 1270
+    },
+    {
+      "epoch": 0.2297148753841667,
+      "grad_norm": 0.9907529354095459,
+      "learning_rate": 9.319040423120514e-05,
+      "loss": 0.8779685974121094,
+      "step": 1280
+    },
+    {
+      "epoch": 0.23150952284810553,
+      "grad_norm": 1.577710747718811,
+      "learning_rate": 9.309595768794863e-05,
+      "loss": 0.9846627235412597,
+      "step": 1290
+    },
+    {
+      "epoch": 0.23330417031204434,
+      "grad_norm": 4.1397552490234375,
+      "learning_rate": 9.300151114469211e-05,
+      "loss": 1.1263571739196778,
+      "step": 1300
+    },
+    {
+      "epoch": 0.23509881777598313,
+      "grad_norm": 0.898070752620697,
+      "learning_rate": 9.29070646014356e-05,
+      "loss": 0.8741483688354492,
+      "step": 1310
+    },
+    {
+      "epoch": 0.23689346523992194,
+      "grad_norm": 0.951369047164917,
+      "learning_rate": 9.281261805817908e-05,
+      "loss": 0.9030444145202636,
+      "step": 1320
+    },
+    {
+      "epoch": 0.23868811270386073,
+      "grad_norm": 1.1681045293807983,
+      "learning_rate": 9.271817151492256e-05,
+      "loss": 0.8972801208496094,
+      "step": 1330
+    },
+    {
+      "epoch": 0.24048276016779954,
+      "grad_norm": 1.5000301599502563,
+      "learning_rate": 9.262372497166604e-05,
+      "loss": 0.9021479606628418,
+      "step": 1340
+    },
+    {
+      "epoch": 0.24227740763173833,
+      "grad_norm": 4.603233337402344,
+      "learning_rate": 9.252927842840953e-05,
+      "loss": 0.9616855621337891,
+      "step": 1350
+    },
+    {
+      "epoch": 0.24407205509567714,
+      "grad_norm": 0.7789506912231445,
+      "learning_rate": 9.243483188515301e-05,
+      "loss": 0.8278372764587403,
+      "step": 1360
+    },
+    {
+      "epoch": 0.24586670255961596,
+      "grad_norm": 0.9593271017074585,
+      "learning_rate": 9.23403853418965e-05,
+      "loss": 0.8673165321350098,
+      "step": 1370
+    },
+    {
+      "epoch": 0.24766135002355474,
+      "grad_norm": 1.2493950128555298,
+      "learning_rate": 9.224593879863998e-05,
+      "loss": 0.8802228927612304,
+      "step": 1380
+    },
+    {
+      "epoch": 0.24945599748749356,
+      "grad_norm": 1.4462804794311523,
+      "learning_rate": 9.215149225538345e-05,
+      "loss": 0.925480842590332,
+      "step": 1390
+    },
+    {
+      "epoch": 0.25125064495143234,
+      "grad_norm": 3.787916660308838,
+      "learning_rate": 9.205704571212694e-05,
+      "loss": 1.1019389152526855,
+      "step": 1400
+    },
+    {
+      "epoch": 0.25304529241537116,
+      "grad_norm": 0.8386228084564209,
+      "learning_rate": 9.196259916887043e-05,
+      "loss": 0.8958368301391602,
+      "step": 1410
+    },
+    {
+      "epoch": 0.25483993987931,
+      "grad_norm": 0.9733511805534363,
+      "learning_rate": 9.186815262561391e-05,
+      "loss": 0.7509948253631592,
+      "step": 1420
+    },
+    {
+      "epoch": 0.2566345873432488,
+      "grad_norm": 1.0639692544937134,
+      "learning_rate": 9.17737060823574e-05,
+      "loss": 0.8474295616149903,
+      "step": 1430
+    },
+    {
+      "epoch": 0.25842923480718755,
+      "grad_norm": 1.7223182916641235,
+      "learning_rate": 9.167925953910086e-05,
+      "loss": 0.9051410675048828,
+      "step": 1440
+    },
+    {
+      "epoch": 0.26022388227112636,
+      "grad_norm": 3.473262310028076,
+      "learning_rate": 9.158481299584436e-05,
+      "loss": 0.9940505027770996,
+      "step": 1450
+    },
+    {
+      "epoch": 0.2620185297350652,
+      "grad_norm": 0.8055341243743896,
+      "learning_rate": 9.149036645258784e-05,
+      "loss": 0.8780458450317383,
+      "step": 1460
+    },
+    {
+      "epoch": 0.263813177199004,
+      "grad_norm": 1.0771476030349731,
+      "learning_rate": 9.139591990933131e-05,
+      "loss": 0.8636789321899414,
+      "step": 1470
+    },
+    {
+      "epoch": 0.26560782466294275,
+      "grad_norm": 1.1147541999816895,
+      "learning_rate": 9.130147336607481e-05,
+      "loss": 0.8464085578918457,
+      "step": 1480
+    },
+    {
+      "epoch": 0.26740247212688156,
+      "grad_norm": 1.4540315866470337,
+      "learning_rate": 9.120702682281828e-05,
+      "loss": 0.8976149559020996,
+      "step": 1490
+    },
+    {
+      "epoch": 0.2691971195908204,
+      "grad_norm": 4.394834995269775,
+      "learning_rate": 9.111258027956178e-05,
+      "loss": 1.1389759063720704,
+      "step": 1500
+    },
+    {
+      "epoch": 0.2709917670547592,
+      "grad_norm": 0.8100571036338806,
+      "learning_rate": 9.101813373630526e-05,
+      "loss": 0.7912180900573731,
+      "step": 1510
+    },
+    {
+      "epoch": 0.272786414518698,
+      "grad_norm": 0.9471364617347717,
+      "learning_rate": 9.092368719304873e-05,
+      "loss": 0.8603667259216309,
+      "step": 1520
+    },
+    {
+      "epoch": 0.27458106198263676,
+      "grad_norm": 1.1160653829574585,
+      "learning_rate": 9.082924064979223e-05,
+      "loss": 0.8775869369506836,
+      "step": 1530
+    },
+    {
+      "epoch": 0.2763757094465756,
+      "grad_norm": 1.5980632305145264,
+      "learning_rate": 9.07347941065357e-05,
+      "loss": 0.9348011970520019,
+      "step": 1540
+    },
+    {
+      "epoch": 0.2781703569105144,
+      "grad_norm": 4.296586036682129,
+      "learning_rate": 9.06403475632792e-05,
+      "loss": 1.0446084976196288,
+      "step": 1550
+    },
+    {
+      "epoch": 0.2799650043744532,
+      "grad_norm": 0.8044270873069763,
+      "learning_rate": 9.054590102002267e-05,
+      "loss": 0.8284408569335937,
+      "step": 1560
+    },
+    {
+      "epoch": 0.281759651838392,
+      "grad_norm": 0.9065347909927368,
+      "learning_rate": 9.045145447676615e-05,
+      "loss": 0.8259025573730469,
+      "step": 1570
+    },
+    {
+      "epoch": 0.2835542993023308,
+      "grad_norm": 1.1605452299118042,
+      "learning_rate": 9.035700793350965e-05,
+      "loss": 0.8801176071166992,
+      "step": 1580
+    },
+    {
+      "epoch": 0.2853489467662696,
+      "grad_norm": 1.5614631175994873,
+      "learning_rate": 9.026256139025312e-05,
+      "loss": 0.9447234153747559,
+      "step": 1590
+    },
+    {
+      "epoch": 0.2871435942302084,
+      "grad_norm": 3.539119005203247,
+      "learning_rate": 9.016811484699661e-05,
+      "loss": 1.0134998321533204,
+      "step": 1600
+    },
+    {
+      "epoch": 0.2889382416941472,
+      "grad_norm": 0.8091688752174377,
+      "learning_rate": 9.007366830374008e-05,
+      "loss": 0.8058991432189941,
+      "step": 1610
+    },
+    {
+      "epoch": 0.290732889158086,
+      "grad_norm": 0.9974891543388367,
+      "learning_rate": 8.997922176048357e-05,
+      "loss": 0.7382112979888916,
+      "step": 1620
+    },
+    {
+      "epoch": 0.2925275366220248,
+      "grad_norm": 1.2397563457489014,
+      "learning_rate": 8.988477521722706e-05,
+      "loss": 0.8194489479064941,
+      "step": 1630
+    },
+    {
+      "epoch": 0.2943221840859636,
+      "grad_norm": 1.3829917907714844,
+      "learning_rate": 8.979032867397053e-05,
+      "loss": 0.9316317558288574,
+      "step": 1640
+    },
+    {
+      "epoch": 0.2961168315499024,
+      "grad_norm": 4.598084926605225,
+      "learning_rate": 8.969588213071402e-05,
+      "loss": 1.0294858932495117,
+      "step": 1650
+    },
+    {
+      "epoch": 0.29791147901384124,
+      "grad_norm": 0.8185928463935852,
+      "learning_rate": 8.96014355874575e-05,
+      "loss": 0.8526276588439942,
+      "step": 1660
+    },
+    {
+      "epoch": 0.29970612647778,
+      "grad_norm": 0.8803004622459412,
+      "learning_rate": 8.950698904420098e-05,
+      "loss": 0.8622500419616699,
+      "step": 1670
+    },
+    {
+      "epoch": 0.3015007739417188,
+      "grad_norm": 1.0694444179534912,
+      "learning_rate": 8.941254250094448e-05,
+      "loss": 0.7721203327178955,
+      "step": 1680
+    },
+    {
+      "epoch": 0.30329542140565763,
+      "grad_norm": 1.434861421585083,
+      "learning_rate": 8.931809595768795e-05,
+      "loss": 0.9630851745605469,
+      "step": 1690
+    },
+    {
+      "epoch": 0.30509006886959644,
+      "grad_norm": 4.44917106628418,
+      "learning_rate": 8.922364941443143e-05,
+      "loss": 1.0820582389831543,
+      "step": 1700
+    },
+    {
+      "epoch": 0.30688471633353526,
+      "grad_norm": 0.8227750658988953,
+      "learning_rate": 8.912920287117492e-05,
+      "loss": 0.8339776992797852,
+      "step": 1710
+    },
+    {
+      "epoch": 0.308679363797474,
+      "grad_norm": 0.9008784294128418,
+      "learning_rate": 8.90347563279184e-05,
+      "loss": 0.7638604164123535,
+      "step": 1720
+    },
+    {
+      "epoch": 0.31047401126141283,
+      "grad_norm": 1.1473792791366577,
+      "learning_rate": 8.894030978466188e-05,
+      "loss": 0.9122580528259278,
+      "step": 1730
+    },
+    {
+      "epoch": 0.31226865872535164,
+      "grad_norm": 1.4027907848358154,
+      "learning_rate": 8.884586324140537e-05,
+      "loss": 0.9376218795776368,
+      "step": 1740
+    },
+    {
+      "epoch": 0.31406330618929046,
+      "grad_norm": 3.44396710395813,
+      "learning_rate": 8.875141669814885e-05,
+      "loss": 0.9627013206481934,
+      "step": 1750
+    },
+    {
+      "epoch": 0.3158579536532292,
+      "grad_norm": 0.761950671672821,
+      "learning_rate": 8.865697015489233e-05,
+      "loss": 0.850135612487793,
+      "step": 1760
+    },
+    {
+      "epoch": 0.31765260111716803,
+      "grad_norm": 0.9319304823875427,
+      "learning_rate": 8.856252361163582e-05,
+      "loss": 0.7397173881530762,
+      "step": 1770
+    },
+    {
+      "epoch": 0.31944724858110685,
+      "grad_norm": 1.1564598083496094,
+      "learning_rate": 8.84680770683793e-05,
+      "loss": 0.8239191055297852,
+      "step": 1780
+    },
+    {
+      "epoch": 0.32124189604504566,
+      "grad_norm": 1.4574592113494873,
+      "learning_rate": 8.837363052512278e-05,
+      "loss": 0.9266434669494629,
+      "step": 1790
+    },
+    {
+      "epoch": 0.3230365435089845,
+      "grad_norm": 4.408364295959473,
+      "learning_rate": 8.827918398186627e-05,
+      "loss": 1.0574697494506835,
+      "step": 1800
+    },
+    {
+      "epoch": 0.32483119097292323,
+      "grad_norm": 0.808427095413208,
+      "learning_rate": 8.818473743860975e-05,
+      "loss": 0.8238012313842773,
+      "step": 1810
+    },
+    {
+      "epoch": 0.32662583843686205,
+      "grad_norm": 0.8963003754615784,
+      "learning_rate": 8.809029089535323e-05,
+      "loss": 0.8102021217346191,
+      "step": 1820
+    },
+    {
+      "epoch": 0.32842048590080086,
+      "grad_norm": 1.1847169399261475,
+      "learning_rate": 8.799584435209672e-05,
+      "loss": 0.847196388244629,
+      "step": 1830
+    },
+    {
+      "epoch": 0.3302151333647397,
+      "grad_norm": 1.355234980583191,
+      "learning_rate": 8.79013978088402e-05,
+      "loss": 0.9291248321533203,
+      "step": 1840
+    },
+    {
+      "epoch": 0.3320097808286785,
+      "grad_norm": 4.019794464111328,
+      "learning_rate": 8.780695126558368e-05,
+      "loss": 0.9507856369018555,
+      "step": 1850
+    },
+    {
+      "epoch": 0.33380442829261725,
+      "grad_norm": 0.8047809600830078,
+      "learning_rate": 8.771250472232717e-05,
+      "loss": 0.8097662925720215,
+      "step": 1860
+    },
+    {
+      "epoch": 0.33559907575655606,
+      "grad_norm": 0.9057943820953369,
+      "learning_rate": 8.761805817907065e-05,
+      "loss": 0.8277406692504883,
+      "step": 1870
+    },
+    {
+      "epoch": 0.3373937232204949,
+      "grad_norm": 1.1131069660186768,
+      "learning_rate": 8.752361163581414e-05,
+      "loss": 0.875433349609375,
+      "step": 1880
+    },
+    {
+      "epoch": 0.3391883706844337,
+      "grad_norm": 1.5567286014556885,
+      "learning_rate": 8.742916509255762e-05,
+      "loss": 0.9548139572143555,
+      "step": 1890
+    },
+    {
+      "epoch": 0.34098301814837245,
+      "grad_norm": 3.943831205368042,
+      "learning_rate": 8.73347185493011e-05,
+      "loss": 1.0330249786376953,
+      "step": 1900
+    },
+    {
+      "epoch": 0.34277766561231127,
+      "grad_norm": 0.7638143301010132,
+      "learning_rate": 8.724027200604459e-05,
+      "loss": 0.804904842376709,
+      "step": 1910
+    },
+    {
+      "epoch": 0.3445723130762501,
+      "grad_norm": 0.9346854090690613,
+      "learning_rate": 8.714582546278807e-05,
+      "loss": 0.7564315795898438,
+      "step": 1920
+    },
+    {
+      "epoch": 0.3463669605401889,
+      "grad_norm": 1.1116750240325928,
+      "learning_rate": 8.705137891953155e-05,
+      "loss": 0.7835733413696289,
+      "step": 1930
+    },
+    {
+      "epoch": 0.3481616080041277,
+      "grad_norm": 1.437160611152649,
+      "learning_rate": 8.695693237627504e-05,
+      "loss": 0.9441628456115723,
+      "step": 1940
+    },
+    {
+      "epoch": 0.34995625546806647,
+      "grad_norm": 2.9362335205078125,
+      "learning_rate": 8.686248583301852e-05,
+      "loss": 1.0385236740112305,
+      "step": 1950
+    },
+    {
+      "epoch": 0.3517509029320053,
+      "grad_norm": 0.7919660210609436,
+      "learning_rate": 8.6768039289762e-05,
+      "loss": 0.7995383739471436,
+      "step": 1960
+    },
+    {
+      "epoch": 0.3535455503959441,
+      "grad_norm": 0.8578867316246033,
+      "learning_rate": 8.667359274650549e-05,
+      "loss": 0.7939890384674072,
+      "step": 1970
+    },
+    {
+      "epoch": 0.3553401978598829,
+      "grad_norm": 1.1968882083892822,
+      "learning_rate": 8.657914620324896e-05,
+      "loss": 0.817835807800293,
+      "step": 1980
+    },
+    {
+      "epoch": 0.3571348453238217,
+      "grad_norm": 1.4205067157745361,
+      "learning_rate": 8.648469965999245e-05,
+      "loss": 0.9530954360961914,
+      "step": 1990
+    },
+    {
+      "epoch": 0.3589294927877605,
+      "grad_norm": 4.1284332275390625,
+      "learning_rate": 8.639025311673594e-05,
+      "loss": 1.0853967666625977,
+      "step": 2000
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 11146,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 1000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.035264485005284e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

qwen3.5-2b-50pct-masked-weighted/checkpoint-2000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62edea7eb1e0274b5104b0e4cbcf850d6b18e77a951e843e93b693eaf58caa36
+size 5265

qwen3.5-2b-50pct-masked-weighted/tb/events.out.tfevents.1790426135.b0aacb224392.4711.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:85f106682ee3b33940d0fc6758d0dcb3a97c07d7e3b123ff12be14b8095c3acf
-size 30303
+oid sha256:91c4c6a13a168fc035b97566a3a72eba1351b492812466afcc13a8fa637d539b
+size 53051

qwen3.5-2b-50pct-masked-weighted/test_eval_entity/metrics.json

@@ -1,15 +1,15 @@
 {
-  "wer": 0.7114996637525218,
-  "wer_zwnj_normalized": 0.7087723230964657,
-  "exact_match": 0.0019801980198019802,
-  "fix_rate": 0.09490559539092588,
-  "preservation_rate": 0.34338748592568763,
-  "targeted_score": 0.21914654065830674,
-  "fix_rate_lenient": 0.09133004926108375,
-  "preservation_rate_lenient": 0.3451304158166866,
-  "targeted_score_lenient": 0.21823023253888515,
-  "fix_rate_recoverable": 0.09265880308751295,
+  "wer": 0.27376415708628216,
+  "wer_zwnj_normalized": 0.2673380941706429,
+  "exact_match": 0.0049504950495049506,
+  "fix_rate": 0.27269311456930834,
+  "preservation_rate": 0.8482998230657873,
+  "targeted_score": 0.5604964688175478,
+  "fix_rate_lenient": 0.26022988505747124,
+  "preservation_rate_lenient": 0.8518981820047984,
+  "targeted_score_lenient": 0.5560640335311349,
+  "fix_rate_recoverable": 0.26397567414040835,
   "unrecoverable_targets": 523,
-  "hallucination_rate": 0.00891089108910891,
+  "hallucination_rate": 0.006930693069306931,
   "n_examples": 1010
 }

qwen3.5-2b-50pct-masked-weighted/test_eval_typo/metrics.json

@@ -1,15 +1,15 @@
 {
-  "wer": 0.5876286027928825,
-  "wer_zwnj_normalized": 0.5829312501595568,
-  "exact_match": 0.00558659217877095,
-  "fix_rate": 0.15878247382706476,
-  "preservation_rate": 0.459294975450563,
-  "targeted_score": 0.3090387246388139,
-  "fix_rate_lenient": 0.1542123783895674,
-  "preservation_rate_lenient": 0.46072255266781914,
-  "targeted_score_lenient": 0.30746746552869325,
-  "fix_rate_recoverable": 0.15657839428331233,
+  "wer": 0.172500063822726,
+  "wer_zwnj_normalized": 0.16473922034157923,
+  "exact_match": 0.008379888268156424,
+  "fix_rate": 0.3557580457541683,
+  "preservation_rate": 0.9167083174080498,
+  "targeted_score": 0.6362331815811091,
+  "fix_rate_lenient": 0.3446491409646036,
+  "preservation_rate_lenient": 0.919745670837118,
+  "targeted_score_lenient": 0.6321974059008608,
+  "fix_rate_recoverable": 0.3495166036149643,
   "unrecoverable_targets": 73,
-  "hallucination_rate": 0.00558659217877095,
+  "hallucination_rate": 0.008379888268156424,
   "n_examples": 358
 }