Upload checkpoint and eval results at step 3500

checkpoint-3500/eval_results.json

@@ -1,9 +1,9 @@
 {
   "step": 3500,
   "epoch": 2.2082018927444795,
-  "time": "2026-07-08 23:52:15",
+  "time": "2026-07-09 00:13:11",
   "metrics": {
-    "eval_loss": 0.07981470227241516,
+    "eval_loss": 0.0798133835196495,
     "eval_accuracy": 0.9820189274447949,
     "eval_precision": 0.9769038701622972,
     "eval_recall": 0.9873817034700315,

checkpoint-3500/eval_step_003500.json

@@ -1,16 +1,16 @@
 {
   "step": 3500,
   "epoch": 2.2082018927444795,
-  "time": "2026-07-08 23:52:15",
+  "time": "2026-07-09 00:13:11",
   "metrics": {
-    "eval_loss": 0.07981470227241516,
+    "eval_loss": 0.0798133835196495,
     "eval_accuracy": 0.9820189274447949,
     "eval_precision": 0.9769038701622972,
     "eval_recall": 0.9873817034700315,
     "eval_f1": 0.9821148415437716,
-    "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,
     "epoch": 2.2082018927444795
   }
 }

checkpoint-3500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1ef6e603e11cd1762c832332d3b03f99318d158ed5c3d9cce9247d7b2f613a5e
+oid sha256:bece9ef69e7f7f30c669bbe1017080e460d330a34071716e0c4d9d32a4a83a80
 size 498612824

checkpoint-3500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:26ef273ef171262b629827651041420b2b4528b4185f3b12f20122b7c382ab38
+oid sha256:618125457c5a643ae3713cf6dadc3b9e3a5d4569d43ce1216088d951e371ac58
 size 997348747

checkpoint-3500/trainer_state.json

@@ -930,154 +930,154 @@
     },
     {
       "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
     }
   ],