Upload checkpoint and eval results at step 4000

checkpoint-4000/eval_results.json

@@ -1,9 +1,9 @@
 {
   "step": 4000,
   "epoch": 2.5236593059936907,
-  "time": "2026-07-09 00:01:30",
+  "time": "2026-07-09 00:22:28",
   "metrics": {
-    "eval_loss": 0.09398859739303589,
+    "eval_loss": 0.09398876130580902,
     "eval_accuracy": 0.9817034700315458,
     "eval_precision": 0.9698461538461538,
     "eval_recall": 0.9943217665615142,

checkpoint-4000/eval_step_004000.json

@@ -1,16 +1,16 @@
 {
   "step": 4000,
   "epoch": 2.5236593059936907,
-  "time": "2026-07-09 00:01:30",
+  "time": "2026-07-09 00:22:28",
   "metrics": {
-    "eval_loss": 0.09398859739303589,
+    "eval_loss": 0.09398876130580902,
     "eval_accuracy": 0.9817034700315458,
     "eval_precision": 0.9698461538461538,
     "eval_recall": 0.9943217665615142,
     "eval_f1": 0.9819314641744549,
-    "eval_runtime": 59.5237,
-    "eval_samples_per_second": 53.256,
-    "eval_steps_per_second": 3.343,
+    "eval_runtime": 59.2692,
+    "eval_samples_per_second": 53.485,
+    "eval_steps_per_second": 3.358,
     "epoch": 2.5236593059936907
   }
 }

checkpoint-4000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:921ef470f0115b54424e4f264029062578c19df0749eb2008d01c2c4ad065ad8
+oid sha256:1327ea1d328294f2a7ffa45826e5d08f6638cefad7e21ba4924ede37bf91731d
 size 498612824

checkpoint-4000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:11a67db49a1c6625690b21a322438c3422b37b8a71747a1fed870a41593954cc
+oid sha256:c99b4d2f0badc824d471e0c37782e135a46a13cb64dc23a0457f94cac42e65da
 size 997348747

checkpoint-4000/trainer_state.json

@@ -930,306 +930,306 @@
     },
     {
       "epoch": 1.9085173501577288,
-      "grad_norm": 32.33955001831055,
+      "grad_norm": 32.33957290649414,
       "learning_rate": 1.2521652607281511e-05,
-      "loss": 0.05197489261627197,
+      "loss": 0.0519749116897583,
       "step": 3025
     },
     {
       "epoch": 1.9242902208201893,
-      "grad_norm": 5.458424091339111,
+      "grad_norm": 5.458426475524902,
       "learning_rate": 1.2388221765257922e-05,
       "loss": 0.08039166450500489,
       "step": 3050
     },
     {
       "epoch": 1.9400630914826498,
-      "grad_norm": 0.032594233751297,
+      "grad_norm": 0.032594215124845505,
       "learning_rate": 1.2254338459241878e-05,
-      "loss": 0.03573692560195923,
+      "loss": 0.03573695421218872,
       "step": 3075
     },
     {
       "epoch": 1.9558359621451105,
-      "grad_norm": 13.031876564025879,
+      "grad_norm": 13.032034873962402,
       "learning_rate": 1.2120028054287853e-05,
-      "loss": 0.07872509479522705,
+      "loss": 0.07872512817382812,
       "step": 3100
     },
     {
       "epoch": 1.971608832807571,
-      "grad_norm": 3.845304012298584,
+      "grad_norm": 3.8453073501586914,
       "learning_rate": 1.198531599636696e-05,
-      "loss": 0.07527520656585693,
+      "loss": 0.07527520179748536,
       "step": 3125
     },
     {
       "epoch": 1.9873817034700316,
-      "grad_norm": 0.061868082731962204,
+      "grad_norm": 0.0618828721344471,
       "learning_rate": 1.1850227807546056e-05,
-      "loss": 0.08369281768798828,
+      "loss": 0.08369292259216309,
       "step": 3150
     },
     {
       "epoch": 2.003154574132492,
-      "grad_norm": 0.2873733639717102,
+      "grad_norm": 0.2873757481575012,
       "learning_rate": 1.1714789081152414e-05,
-      "loss": 0.06914769172668457,
+      "loss": 0.06914762020111084,
       "step": 3175
     },
     {
       "epoch": 2.0189274447949526,
-      "grad_norm": 0.2408512681722641,
+      "grad_norm": 0.24084864556789398,
       "learning_rate": 1.1579025476924912e-05,
-      "loss": 0.04823586940765381,
+      "loss": 0.04823594570159912,
       "step": 3200
     },
     {
       "epoch": 2.034700315457413,
-      "grad_norm": 0.06181686371564865,
+      "grad_norm": 0.06181693077087402,
       "learning_rate": 1.1442962716152626e-05,
-      "loss": 0.052933053970336916,
+      "loss": 0.05293301582336426,
       "step": 3225
     },
     {
       "epoch": 2.050473186119874,
-      "grad_norm": 0.23159997165203094,
+      "grad_norm": 0.23160123825073242,
       "learning_rate": 1.1306626576801762e-05,
-      "loss": 0.042997312545776364,
+      "loss": 0.042997422218322756,
       "step": 3250
     },
     {
       "epoch": 2.0662460567823344,
-      "grad_norm": 0.11183202266693115,
+      "grad_norm": 0.11183146387338638,
       "learning_rate": 1.1170042888631865e-05,
-      "loss": 0.030156188011169434,
+      "loss": 0.03015615463256836,
       "step": 3275
     },
     {
       "epoch": 2.082018927444795,
-      "grad_norm": 0.020436838269233704,
+      "grad_norm": 0.020436899736523628,
       "learning_rate": 1.103323752830219e-05,
-      "loss": 0.06210995674133301,
+      "loss": 0.06211026668548584,
       "step": 3300
     },
     {
       "epoch": 2.0977917981072554,
-      "grad_norm": 0.09942680597305298,
+      "grad_norm": 0.09942758828401566,
       "learning_rate": 1.0896236414469217e-05,
-      "loss": 0.06024267673492432,
+      "loss": 0.06024351596832275,
       "step": 3325
     },
     {
       "epoch": 2.1135646687697163,
-      "grad_norm": 0.06319867074489594,
+      "grad_norm": 0.06320168823003769,
       "learning_rate": 1.075906550287618e-05,
-      "loss": 0.09793594360351562,
+      "loss": 0.09793644905090332,
       "step": 3350
     },
     {
       "epoch": 2.1293375394321767,
-      "grad_norm": 0.14570729434490204,
+      "grad_norm": 0.14571192860603333,
       "learning_rate": 1.0621750781435597e-05,
-      "loss": 0.039036962985992435,
+      "loss": 0.03903674602508545,
       "step": 3375
     },
     {
       "epoch": 2.145110410094637,
-      "grad_norm": 30.197425842285156,
+      "grad_norm": 30.196697235107422,
       "learning_rate": 1.048431826530569e-05,
-      "loss": 0.024112117290496827,
+      "loss": 0.024111902713775633,
       "step": 3400
     },
     {
       "epoch": 2.1608832807570977,
-      "grad_norm": 0.0806874930858612,
+      "grad_norm": 0.08068069070577621,
       "learning_rate": 1.0346793991961636e-05,
-      "loss": 0.09378963470458984,
+      "loss": 0.09379005432128906,
       "step": 3425
     },
     {
       "epoch": 2.176656151419558,
-      "grad_norm": 0.041268810629844666,
+      "grad_norm": 0.041268039494752884,
       "learning_rate": 1.0209204016262617e-05,
-      "loss": 0.028721425533294678,
+      "loss": 0.028721418380737305,
       "step": 3450
     },
     {
       "epoch": 2.192429022082019,
-      "grad_norm": 0.027517961338162422,
+      "grad_norm": 0.02751922979950905,
       "learning_rate": 1.0071574405515545e-05,
-      "loss": 0.019525480270385743,
+      "loss": 0.019525556564331054,
       "step": 3475
     },
     {
       "epoch": 2.2082018927444795,
-      "grad_norm": 0.06598785519599915,
+      "grad_norm": 0.06599944829940796,
       "learning_rate": 9.93393123453644e-06,
-      "loss": 0.07426059722900391,
+      "loss": 0.07425961017608643,
       "step": 3500
     },
     {
       "epoch": 2.2082018927444795,
       "eval_accuracy": 0.9820189274447949,
       "eval_f1": 0.9821148415437716,
-      "eval_loss": 0.07981470227241516,
+      "eval_loss": 0.0798133835196495,
       "eval_precision": 0.9769038701622972,
       "eval_recall": 0.9873817034700315,
-      "eval_runtime": 59.5362,
-      "eval_samples_per_second": 53.245,
-      "eval_steps_per_second": 3.343,
+      "eval_runtime": 59.7613,
+      "eval_samples_per_second": 53.044,
+      "eval_steps_per_second": 3.33,
       "step": 3500
     },
     {
       "epoch": 2.22397476340694,
-      "grad_norm": 0.22194409370422363,
+      "grad_norm": 0.22195865213871002,
       "learning_rate": 9.796300580710401e-06,
-      "loss": 0.06443918228149415,
+      "loss": 0.06443671703338623,
       "step": 3525
     },
     {
       "epoch": 2.2397476340694005,
-      "grad_norm": 0.19404132664203644,
+      "grad_norm": 0.19403228163719177,
       "learning_rate": 9.658708519051078e-06,
-      "loss": 0.01147290587425232,
+      "loss": 0.011473581790924073,
       "step": 3550
     },
     {
       "epoch": 2.2555205047318614,
-      "grad_norm": 0.3892022967338562,
+      "grad_norm": 0.3892849087715149,
       "learning_rate": 9.521181117260576e-06,
-      "loss": 0.03980299949645996,
+      "loss": 0.03980313539505005,
       "step": 3575
     },
     {
       "epoch": 2.271293375394322,
-      "grad_norm": 0.09898478537797928,
+      "grad_norm": 0.09898291528224945,
       "learning_rate": 9.38374443079078e-06,
-      "loss": 0.05841335296630859,
+      "loss": 0.05841365337371826,
       "step": 3600
     },
     {
       "epoch": 2.2870662460567823,
-      "grad_norm": 0.0590379536151886,
+      "grad_norm": 0.059039290994405746,
       "learning_rate": 9.246424497906963e-06,
-      "loss": 0.019510600566864014,
+      "loss": 0.019511160850524904,
       "step": 3625
     },
     {
       "epoch": 2.302839116719243,
-      "grad_norm": 0.6535356640815735,
+      "grad_norm": 0.653461217880249,
       "learning_rate": 9.10924733475469e-06,
-      "loss": 0.08494356155395508,
+      "loss": 0.08494296073913574,
       "step": 3650
     },
     {
       "epoch": 2.3186119873817033,
-      "grad_norm": 3.5332858562469482,
+      "grad_norm": 3.5332391262054443,
       "learning_rate": 8.972238930430877e-06,
-      "loss": 0.04381181716918945,
+      "loss": 0.043811373710632324,
       "step": 3675
     },
     {
       "epoch": 2.334384858044164,
-      "grad_norm": 0.237799733877182,
+      "grad_norm": 0.237837016582489,
       "learning_rate": 8.83542524205999e-06,
-      "loss": 0.03998601913452148,
+      "loss": 0.03998267650604248,
       "step": 3700
     },
     {
       "epoch": 2.3501577287066246,
-      "grad_norm": 0.04083641991019249,
+      "grad_norm": 0.04083343222737312,
       "learning_rate": 8.698832189876313e-06,
-      "loss": 0.039485976696014405,
+      "loss": 0.03948731899261475,
       "step": 3725
     },
     {
       "epoch": 2.365930599369085,
-      "grad_norm": 0.033618684858083725,
+      "grad_norm": 0.03361862152814865,
       "learning_rate": 8.562485652313168e-06,
-      "loss": 0.017637826204299926,
+      "loss": 0.017638446092605592,
       "step": 3750
     },
     {
       "epoch": 2.3817034700315456,
-      "grad_norm": 4.868691444396973,
+      "grad_norm": 4.868694305419922,
       "learning_rate": 8.42641146110011e-06,
-      "loss": 0.048912029266357425,
+      "loss": 0.04891120433807373,
       "step": 3775
     },
     {
       "epoch": 2.3974763406940065,
-      "grad_norm": 0.06035134196281433,
+      "grad_norm": 0.060350608080625534,
       "learning_rate": 8.290635396368916e-06,
-      "loss": 0.019228267669677734,
+      "loss": 0.019228975772857666,
       "step": 3800
     },
     {
       "epoch": 2.413249211356467,
-      "grad_norm": 2.0009303092956543,
+      "grad_norm": 2.000920295715332,
       "learning_rate": 8.15518318176938e-06,
-      "loss": 0.04954765319824219,
+      "loss": 0.04954701423645019,
       "step": 3825
     },
     {
       "epoch": 2.4290220820189274,
-      "grad_norm": 0.032643403857946396,
+      "grad_norm": 0.03264353796839714,
       "learning_rate": 8.020080479595818e-06,
-      "loss": 0.048542871475219726,
+      "loss": 0.048543753623962405,
       "step": 3850
     },
     {
       "epoch": 2.444794952681388,
-      "grad_norm": 0.16234640777111053,
+      "grad_norm": 0.16232901811599731,
       "learning_rate": 7.885352885925139e-06,
-      "loss": 0.030491657257080078,
+      "loss": 0.03049187421798706,
       "step": 3875
     },
     {
       "epoch": 2.4605678233438484,
-      "grad_norm": 0.28245651721954346,
+      "grad_norm": 0.2823891341686249,
       "learning_rate": 7.75102592576756e-06,
-      "loss": 0.03745485067367554,
+      "loss": 0.03745495319366455,
       "step": 3900
     },
     {
       "epoch": 2.4763406940063093,
-      "grad_norm": 0.06313358247280121,
+      "grad_norm": 0.06313452869653702,
       "learning_rate": 7.617125048230678e-06,
-      "loss": 0.027678358554840087,
+      "loss": 0.027678728103637695,
       "step": 3925
     },
     {
       "epoch": 2.4921135646687698,
-      "grad_norm": 0.034601207822561264,
+      "grad_norm": 0.03460308164358139,
       "learning_rate": 7.483675621697999e-06,
-      "loss": 0.03676470041275024,
+      "loss": 0.036765682697296145,
       "step": 3950
     },
     {
       "epoch": 2.5078864353312302,
-      "grad_norm": 0.04799091815948486,
+      "grad_norm": 0.04799109697341919,
       "learning_rate": 7.350702929022739e-06,
-      "loss": 0.04815303325653076,
+      "loss": 0.048152894973754884,
       "step": 3975
     },
     {
       "epoch": 2.5236593059936907,
-      "grad_norm": 0.6451889276504517,
+      "grad_norm": 0.6452091336250305,
       "learning_rate": 7.2182321627378e-06,
-      "loss": 0.03499504089355469,
+      "loss": 0.03499542236328125,
       "step": 4000
     },
     {
       "epoch": 2.5236593059936907,
       "eval_accuracy": 0.9817034700315458,
       "eval_f1": 0.9819314641744549,
-      "eval_loss": 0.09398859739303589,
+      "eval_loss": 0.09398876130580902,
       "eval_precision": 0.9698461538461538,
       "eval_recall": 0.9943217665615142,
-      "eval_runtime": 59.5237,
-      "eval_samples_per_second": 53.256,
-      "eval_steps_per_second": 3.343,
+      "eval_runtime": 59.2692,
+      "eval_samples_per_second": 53.485,
+      "eval_steps_per_second": 3.358,
       "step": 4000
     }
   ],