Upload VNEncoder checkpoints

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57849770ad0347e1baa59969fa28a052532e32e80baed61aa34e0407e2e766a4
+size 540026404

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<mask>": 64000
+}

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:760f6f27b0d7ebb2325878289e1759ee21a91facd2f3083921bb212774ac75dd
+size 540026436

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eab00e1e4f8f85a0b064722047e04bb0bdb83c524080d13c01a833f122803b82
+size 1075446330

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5eeae9ec5f8da102a9dea275a2bbd37abe42d31853b4aab6a3cfd42b698fecc
+size 14244

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8236ff544476b41bfc1e33b4e8d7748d72c7404d194d626951f6e750d202a060
+size 1064

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64000": {
+      "content": "<mask>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "mask_token": "<mask>",
+  "model_max_length": 256,
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "tokenizer_class": "PhobertTokenizer",
+  "unk_token": "<unk>"
+}

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/trainer_state.json

@@ -0,0 +1,1631 @@
+{
+  "best_global_step": 5220,
+  "best_metric": 0.8675740894345955,
+  "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220",
+  "epoch": 12.0,
+  "eval_steps": 500,
+  "global_step": 5220,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05747126436781609,
+      "grad_norm": 16.450016021728516,
+      "learning_rate": 5.517241379310345e-07,
+      "loss": 4.464,
+      "step": 25
+    },
+    {
+      "epoch": 0.11494252873563218,
+      "grad_norm": 15.485475540161133,
+      "learning_rate": 1.1264367816091955e-06,
+      "loss": 4.2217,
+      "step": 50
+    },
+    {
+      "epoch": 0.1724137931034483,
+      "grad_norm": 13.874435424804688,
+      "learning_rate": 1.7011494252873565e-06,
+      "loss": 3.8906,
+      "step": 75
+    },
+    {
+      "epoch": 0.22988505747126436,
+      "grad_norm": 16.325193405151367,
+      "learning_rate": 2.2758620689655173e-06,
+      "loss": 3.5109,
+      "step": 100
+    },
+    {
+      "epoch": 0.28735632183908044,
+      "grad_norm": 12.38318157196045,
+      "learning_rate": 2.850574712643678e-06,
+      "loss": 3.0685,
+      "step": 125
+    },
+    {
+      "epoch": 0.3448275862068966,
+      "grad_norm": 17.294038772583008,
+      "learning_rate": 3.425287356321839e-06,
+      "loss": 2.541,
+      "step": 150
+    },
+    {
+      "epoch": 0.40229885057471265,
+      "grad_norm": 17.30904769897461,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 2.1361,
+      "step": 175
+    },
+    {
+      "epoch": 0.45977011494252873,
+      "grad_norm": 15.655984878540039,
+      "learning_rate": 4.5747126436781615e-06,
+      "loss": 1.8631,
+      "step": 200
+    },
+    {
+      "epoch": 0.5172413793103449,
+      "grad_norm": 30.74538230895996,
+      "learning_rate": 5.149425287356322e-06,
+      "loss": 1.4674,
+      "step": 225
+    },
+    {
+      "epoch": 0.5747126436781609,
+      "grad_norm": 26.51157569885254,
+      "learning_rate": 5.724137931034483e-06,
+      "loss": 1.4242,
+      "step": 250
+    },
+    {
+      "epoch": 0.632183908045977,
+      "grad_norm": 15.323351860046387,
+      "learning_rate": 6.298850574712644e-06,
+      "loss": 1.2337,
+      "step": 275
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 11.085951805114746,
+      "learning_rate": 6.873563218390805e-06,
+      "loss": 1.235,
+      "step": 300
+    },
+    {
+      "epoch": 0.7471264367816092,
+      "grad_norm": 17.022886276245117,
+      "learning_rate": 7.4482758620689665e-06,
+      "loss": 1.123,
+      "step": 325
+    },
+    {
+      "epoch": 0.8045977011494253,
+      "grad_norm": 7.172950267791748,
+      "learning_rate": 8.022988505747126e-06,
+      "loss": 1.1503,
+      "step": 350
+    },
+    {
+      "epoch": 0.8620689655172413,
+      "grad_norm": 19.81337547302246,
+      "learning_rate": 8.597701149425289e-06,
+      "loss": 1.0711,
+      "step": 375
+    },
+    {
+      "epoch": 0.9195402298850575,
+      "grad_norm": 23.27373695373535,
+      "learning_rate": 9.172413793103448e-06,
+      "loss": 0.9703,
+      "step": 400
+    },
+    {
+      "epoch": 0.9770114942528736,
+      "grad_norm": 21.556955337524414,
+      "learning_rate": 9.74712643678161e-06,
+      "loss": 0.9083,
+      "step": 425
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9372979961215255,
+      "eval_loss": 0.21047352254390717,
+      "eval_macro_f1": 0.7385325809144866,
+      "eval_runtime": 1.2627,
+      "eval_samples_per_second": 1225.158,
+      "eval_steps_per_second": 76.82,
+      "eval_weighted_f1": 0.926186545957586,
+      "step": 435
+    },
+    {
+      "epoch": 1.0344827586206897,
+      "grad_norm": 20.03977394104004,
+      "learning_rate": 1.032183908045977e-05,
+      "loss": 1.1259,
+      "step": 450
+    },
+    {
+      "epoch": 1.0919540229885056,
+      "grad_norm": 28.675647735595703,
+      "learning_rate": 1.0896551724137932e-05,
+      "loss": 0.9905,
+      "step": 475
+    },
+    {
+      "epoch": 1.1494252873563218,
+      "grad_norm": 28.670623779296875,
+      "learning_rate": 1.1471264367816093e-05,
+      "loss": 0.7334,
+      "step": 500
+    },
+    {
+      "epoch": 1.206896551724138,
+      "grad_norm": 28.721160888671875,
+      "learning_rate": 1.2045977011494254e-05,
+      "loss": 0.9009,
+      "step": 525
+    },
+    {
+      "epoch": 1.264367816091954,
+      "grad_norm": 44.211883544921875,
+      "learning_rate": 1.2620689655172415e-05,
+      "loss": 0.7161,
+      "step": 550
+    },
+    {
+      "epoch": 1.3218390804597702,
+      "grad_norm": 15.909297943115234,
+      "learning_rate": 1.3195402298850574e-05,
+      "loss": 0.9911,
+      "step": 575
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "grad_norm": 20.90113067626953,
+      "learning_rate": 1.3770114942528736e-05,
+      "loss": 0.8012,
+      "step": 600
+    },
+    {
+      "epoch": 1.4367816091954024,
+      "grad_norm": 27.909393310546875,
+      "learning_rate": 1.4344827586206897e-05,
+      "loss": 0.8335,
+      "step": 625
+    },
+    {
+      "epoch": 1.4942528735632183,
+      "grad_norm": 4.127943992614746,
+      "learning_rate": 1.4919540229885058e-05,
+      "loss": 0.887,
+      "step": 650
+    },
+    {
+      "epoch": 1.5517241379310345,
+      "grad_norm": 13.942008972167969,
+      "learning_rate": 1.549425287356322e-05,
+      "loss": 0.9997,
+      "step": 675
+    },
+    {
+      "epoch": 1.6091954022988506,
+      "grad_norm": 150.57318115234375,
+      "learning_rate": 1.6068965517241383e-05,
+      "loss": 0.8995,
+      "step": 700
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 84.19149017333984,
+      "learning_rate": 1.664367816091954e-05,
+      "loss": 0.7276,
+      "step": 725
+    },
+    {
+      "epoch": 1.7241379310344827,
+      "grad_norm": 0.8938015103340149,
+      "learning_rate": 1.72183908045977e-05,
+      "loss": 0.7947,
+      "step": 750
+    },
+    {
+      "epoch": 1.7816091954022988,
+      "grad_norm": 21.69007110595703,
+      "learning_rate": 1.7793103448275862e-05,
+      "loss": 0.896,
+      "step": 775
+    },
+    {
+      "epoch": 1.839080459770115,
+      "grad_norm": 9.371072769165039,
+      "learning_rate": 1.8367816091954027e-05,
+      "loss": 0.9257,
+      "step": 800
+    },
+    {
+      "epoch": 1.896551724137931,
+      "grad_norm": 9.058280944824219,
+      "learning_rate": 1.8942528735632187e-05,
+      "loss": 0.7923,
+      "step": 825
+    },
+    {
+      "epoch": 1.9540229885057472,
+      "grad_norm": 19.173526763916016,
+      "learning_rate": 1.9517241379310345e-05,
+      "loss": 0.829,
+      "step": 850
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9444085326438267,
+      "eval_loss": 0.1997242271900177,
+      "eval_macro_f1": 0.7984669829693747,
+      "eval_runtime": 1.2248,
+      "eval_samples_per_second": 1263.023,
+      "eval_steps_per_second": 79.194,
+      "eval_weighted_f1": 0.9372675430430231,
+      "step": 870
+    },
+    {
+      "epoch": 2.0114942528735633,
+      "grad_norm": 13.507299423217773,
+      "learning_rate": 1.9989782886334612e-05,
+      "loss": 1.0265,
+      "step": 875
+    },
+    {
+      "epoch": 2.0689655172413794,
+      "grad_norm": 29.802522659301758,
+      "learning_rate": 1.9925925925925928e-05,
+      "loss": 0.6546,
+      "step": 900
+    },
+    {
+      "epoch": 2.1264367816091956,
+      "grad_norm": 20.382816314697266,
+      "learning_rate": 1.9862068965517244e-05,
+      "loss": 0.6422,
+      "step": 925
+    },
+    {
+      "epoch": 2.1839080459770113,
+      "grad_norm": 16.437686920166016,
+      "learning_rate": 1.979821200510856e-05,
+      "loss": 0.6465,
+      "step": 950
+    },
+    {
+      "epoch": 2.2413793103448274,
+      "grad_norm": 14.521719932556152,
+      "learning_rate": 1.9734355044699872e-05,
+      "loss": 0.7665,
+      "step": 975
+    },
+    {
+      "epoch": 2.2988505747126435,
+      "grad_norm": 25.880773544311523,
+      "learning_rate": 1.9670498084291188e-05,
+      "loss": 0.8589,
+      "step": 1000
+    },
+    {
+      "epoch": 2.3563218390804597,
+      "grad_norm": 20.663999557495117,
+      "learning_rate": 1.9606641123882504e-05,
+      "loss": 0.7237,
+      "step": 1025
+    },
+    {
+      "epoch": 2.413793103448276,
+      "grad_norm": 15.50725269317627,
+      "learning_rate": 1.954278416347382e-05,
+      "loss": 0.737,
+      "step": 1050
+    },
+    {
+      "epoch": 2.471264367816092,
+      "grad_norm": 1.7924487590789795,
+      "learning_rate": 1.9478927203065136e-05,
+      "loss": 0.7695,
+      "step": 1075
+    },
+    {
+      "epoch": 2.528735632183908,
+      "grad_norm": 37.092926025390625,
+      "learning_rate": 1.941507024265645e-05,
+      "loss": 0.6687,
+      "step": 1100
+    },
+    {
+      "epoch": 2.586206896551724,
+      "grad_norm": 0.8467792868614197,
+      "learning_rate": 1.9351213282247767e-05,
+      "loss": 0.6213,
+      "step": 1125
+    },
+    {
+      "epoch": 2.6436781609195403,
+      "grad_norm": 17.73621940612793,
+      "learning_rate": 1.928735632183908e-05,
+      "loss": 0.8202,
+      "step": 1150
+    },
+    {
+      "epoch": 2.7011494252873565,
+      "grad_norm": 3.8395161628723145,
+      "learning_rate": 1.92234993614304e-05,
+      "loss": 0.6882,
+      "step": 1175
+    },
+    {
+      "epoch": 2.7586206896551726,
+      "grad_norm": 77.6451416015625,
+      "learning_rate": 1.915964240102171e-05,
+      "loss": 0.5662,
+      "step": 1200
+    },
+    {
+      "epoch": 2.8160919540229887,
+      "grad_norm": 9.742815017700195,
+      "learning_rate": 1.9095785440613027e-05,
+      "loss": 0.818,
+      "step": 1225
+    },
+    {
+      "epoch": 2.873563218390805,
+      "grad_norm": 41.80664825439453,
+      "learning_rate": 1.9031928480204343e-05,
+      "loss": 0.8144,
+      "step": 1250
+    },
+    {
+      "epoch": 2.9310344827586206,
+      "grad_norm": 13.355256080627441,
+      "learning_rate": 1.896807151979566e-05,
+      "loss": 0.7142,
+      "step": 1275
+    },
+    {
+      "epoch": 2.9885057471264367,
+      "grad_norm": 3.4613044261932373,
+      "learning_rate": 1.8904214559386975e-05,
+      "loss": 0.6519,
+      "step": 1300
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9489334195216548,
+      "eval_loss": 0.1842694729566574,
+      "eval_macro_f1": 0.8417608126834946,
+      "eval_runtime": 1.2612,
+      "eval_samples_per_second": 1226.651,
+      "eval_steps_per_second": 76.913,
+      "eval_weighted_f1": 0.9451926750188842,
+      "step": 1305
+    },
+    {
+      "epoch": 3.045977011494253,
+      "grad_norm": 9.822381973266602,
+      "learning_rate": 1.884035759897829e-05,
+      "loss": 0.5464,
+      "step": 1325
+    },
+    {
+      "epoch": 3.103448275862069,
+      "grad_norm": 14.551055908203125,
+      "learning_rate": 1.8776500638569606e-05,
+      "loss": 0.4742,
+      "step": 1350
+    },
+    {
+      "epoch": 3.160919540229885,
+      "grad_norm": 0.7838442921638489,
+      "learning_rate": 1.871264367816092e-05,
+      "loss": 0.5285,
+      "step": 1375
+    },
+    {
+      "epoch": 3.218390804597701,
+      "grad_norm": 32.79100799560547,
+      "learning_rate": 1.8648786717752235e-05,
+      "loss": 0.4901,
+      "step": 1400
+    },
+    {
+      "epoch": 3.2758620689655173,
+      "grad_norm": 17.9625244140625,
+      "learning_rate": 1.8584929757343554e-05,
+      "loss": 0.7674,
+      "step": 1425
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 38.27335739135742,
+      "learning_rate": 1.8521072796934866e-05,
+      "loss": 0.6257,
+      "step": 1450
+    },
+    {
+      "epoch": 3.3908045977011496,
+      "grad_norm": 0.9759963750839233,
+      "learning_rate": 1.8457215836526182e-05,
+      "loss": 0.5915,
+      "step": 1475
+    },
+    {
+      "epoch": 3.4482758620689653,
+      "grad_norm": 32.85970687866211,
+      "learning_rate": 1.8393358876117498e-05,
+      "loss": 0.6667,
+      "step": 1500
+    },
+    {
+      "epoch": 3.5057471264367814,
+      "grad_norm": 26.64856719970703,
+      "learning_rate": 1.8329501915708814e-05,
+      "loss": 0.7413,
+      "step": 1525
+    },
+    {
+      "epoch": 3.5632183908045976,
+      "grad_norm": 0.5582925081253052,
+      "learning_rate": 1.826564495530013e-05,
+      "loss": 0.5404,
+      "step": 1550
+    },
+    {
+      "epoch": 3.6206896551724137,
+      "grad_norm": 28.135393142700195,
+      "learning_rate": 1.8201787994891446e-05,
+      "loss": 0.8905,
+      "step": 1575
+    },
+    {
+      "epoch": 3.67816091954023,
+      "grad_norm": 0.7919555902481079,
+      "learning_rate": 1.813793103448276e-05,
+      "loss": 0.5488,
+      "step": 1600
+    },
+    {
+      "epoch": 3.735632183908046,
+      "grad_norm": 280.8408203125,
+      "learning_rate": 1.8074074074074074e-05,
+      "loss": 0.586,
+      "step": 1625
+    },
+    {
+      "epoch": 3.793103448275862,
+      "grad_norm": 21.92925453186035,
+      "learning_rate": 1.801021711366539e-05,
+      "loss": 0.5316,
+      "step": 1650
+    },
+    {
+      "epoch": 3.8505747126436782,
+      "grad_norm": 14.422163963317871,
+      "learning_rate": 1.7946360153256706e-05,
+      "loss": 0.5578,
+      "step": 1675
+    },
+    {
+      "epoch": 3.9080459770114944,
+      "grad_norm": 19.861310958862305,
+      "learning_rate": 1.788250319284802e-05,
+      "loss": 0.648,
+      "step": 1700
+    },
+    {
+      "epoch": 3.9655172413793105,
+      "grad_norm": 1.0063120126724243,
+      "learning_rate": 1.7818646232439337e-05,
+      "loss": 0.5442,
+      "step": 1725
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9469941822882999,
+      "eval_loss": 0.25301477313041687,
+      "eval_macro_f1": 0.8476622089963451,
+      "eval_runtime": 1.2449,
+      "eval_samples_per_second": 1242.637,
+      "eval_steps_per_second": 77.916,
+      "eval_weighted_f1": 0.9441704666909856,
+      "step": 1740
+    },
+    {
+      "epoch": 4.022988505747127,
+      "grad_norm": 5.532073974609375,
+      "learning_rate": 1.7754789272030653e-05,
+      "loss": 0.5918,
+      "step": 1750
+    },
+    {
+      "epoch": 4.080459770114943,
+      "grad_norm": 18.194129943847656,
+      "learning_rate": 1.769093231162197e-05,
+      "loss": 0.3519,
+      "step": 1775
+    },
+    {
+      "epoch": 4.137931034482759,
+      "grad_norm": 0.2660662531852722,
+      "learning_rate": 1.762707535121328e-05,
+      "loss": 0.3811,
+      "step": 1800
+    },
+    {
+      "epoch": 4.195402298850575,
+      "grad_norm": 16.584251403808594,
+      "learning_rate": 1.75632183908046e-05,
+      "loss": 0.5086,
+      "step": 1825
+    },
+    {
+      "epoch": 4.252873563218391,
+      "grad_norm": 19.36803436279297,
+      "learning_rate": 1.7499361430395913e-05,
+      "loss": 0.4792,
+      "step": 1850
+    },
+    {
+      "epoch": 4.310344827586207,
+      "grad_norm": 50.34165954589844,
+      "learning_rate": 1.743550446998723e-05,
+      "loss": 0.5772,
+      "step": 1875
+    },
+    {
+      "epoch": 4.3678160919540225,
+      "grad_norm": 13.773152351379395,
+      "learning_rate": 1.7371647509578545e-05,
+      "loss": 0.5231,
+      "step": 1900
+    },
+    {
+      "epoch": 4.425287356321839,
+      "grad_norm": 427.54718017578125,
+      "learning_rate": 1.730779054916986e-05,
+      "loss": 0.2846,
+      "step": 1925
+    },
+    {
+      "epoch": 4.482758620689655,
+      "grad_norm": 11.35075855255127,
+      "learning_rate": 1.7243933588761176e-05,
+      "loss": 0.5971,
+      "step": 1950
+    },
+    {
+      "epoch": 4.540229885057471,
+      "grad_norm": 0.7868359684944153,
+      "learning_rate": 1.7180076628352492e-05,
+      "loss": 0.3201,
+      "step": 1975
+    },
+    {
+      "epoch": 4.597701149425287,
+      "grad_norm": 1.2590872049331665,
+      "learning_rate": 1.7116219667943808e-05,
+      "loss": 0.4207,
+      "step": 2000
+    },
+    {
+      "epoch": 4.655172413793103,
+      "grad_norm": 19.67682456970215,
+      "learning_rate": 1.705236270753512e-05,
+      "loss": 0.6388,
+      "step": 2025
+    },
+    {
+      "epoch": 4.712643678160919,
+      "grad_norm": 21.207443237304688,
+      "learning_rate": 1.6988505747126436e-05,
+      "loss": 0.5218,
+      "step": 2050
+    },
+    {
+      "epoch": 4.7701149425287355,
+      "grad_norm": 25.38108253479004,
+      "learning_rate": 1.6924648786717756e-05,
+      "loss": 0.6049,
+      "step": 2075
+    },
+    {
+      "epoch": 4.827586206896552,
+      "grad_norm": 57.224735260009766,
+      "learning_rate": 1.6860791826309068e-05,
+      "loss": 0.6162,
+      "step": 2100
+    },
+    {
+      "epoch": 4.885057471264368,
+      "grad_norm": 3.48508882522583,
+      "learning_rate": 1.6796934865900384e-05,
+      "loss": 0.5359,
+      "step": 2125
+    },
+    {
+      "epoch": 4.942528735632184,
+      "grad_norm": 44.712677001953125,
+      "learning_rate": 1.67330779054917e-05,
+      "loss": 0.4515,
+      "step": 2150
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 57.54368209838867,
+      "learning_rate": 1.6669220945083016e-05,
+      "loss": 0.5796,
+      "step": 2175
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.2165171504020691,
+      "eval_macro_f1": 0.8580507118283248,
+      "eval_runtime": 1.2482,
+      "eval_samples_per_second": 1239.372,
+      "eval_steps_per_second": 77.711,
+      "eval_weighted_f1": 0.948388591648449,
+      "step": 2175
+    },
+    {
+      "epoch": 5.057471264367816,
+      "grad_norm": 12.27694320678711,
+      "learning_rate": 1.6605363984674328e-05,
+      "loss": 0.3551,
+      "step": 2200
+    },
+    {
+      "epoch": 5.114942528735632,
+      "grad_norm": 1.7980570793151855,
+      "learning_rate": 1.6541507024265647e-05,
+      "loss": 0.2956,
+      "step": 2225
+    },
+    {
+      "epoch": 5.172413793103448,
+      "grad_norm": 1.7773587703704834,
+      "learning_rate": 1.6477650063856963e-05,
+      "loss": 0.4865,
+      "step": 2250
+    },
+    {
+      "epoch": 5.2298850574712645,
+      "grad_norm": 9.462156295776367,
+      "learning_rate": 1.6413793103448276e-05,
+      "loss": 0.4026,
+      "step": 2275
+    },
+    {
+      "epoch": 5.287356321839081,
+      "grad_norm": 2.051429510116577,
+      "learning_rate": 1.634993614303959e-05,
+      "loss": 0.3801,
+      "step": 2300
+    },
+    {
+      "epoch": 5.344827586206897,
+      "grad_norm": 0.48756808042526245,
+      "learning_rate": 1.6286079182630907e-05,
+      "loss": 0.3938,
+      "step": 2325
+    },
+    {
+      "epoch": 5.402298850574713,
+      "grad_norm": 2.7134180068969727,
+      "learning_rate": 1.6222222222222223e-05,
+      "loss": 0.5045,
+      "step": 2350
+    },
+    {
+      "epoch": 5.459770114942529,
+      "grad_norm": 0.3124593496322632,
+      "learning_rate": 1.615836526181354e-05,
+      "loss": 0.4567,
+      "step": 2375
+    },
+    {
+      "epoch": 5.517241379310345,
+      "grad_norm": 0.5791522264480591,
+      "learning_rate": 1.6094508301404855e-05,
+      "loss": 0.3833,
+      "step": 2400
+    },
+    {
+      "epoch": 5.574712643678161,
+      "grad_norm": 64.2642593383789,
+      "learning_rate": 1.603065134099617e-05,
+      "loss": 0.4675,
+      "step": 2425
+    },
+    {
+      "epoch": 5.6321839080459775,
+      "grad_norm": 137.23037719726562,
+      "learning_rate": 1.5966794380587483e-05,
+      "loss": 0.4846,
+      "step": 2450
+    },
+    {
+      "epoch": 5.689655172413794,
+      "grad_norm": 15.930166244506836,
+      "learning_rate": 1.5902937420178802e-05,
+      "loss": 0.4571,
+      "step": 2475
+    },
+    {
+      "epoch": 5.747126436781609,
+      "grad_norm": 14.224031448364258,
+      "learning_rate": 1.5839080459770115e-05,
+      "loss": 0.3343,
+      "step": 2500
+    },
+    {
+      "epoch": 5.804597701149425,
+      "grad_norm": 19.296125411987305,
+      "learning_rate": 1.577522349936143e-05,
+      "loss": 0.5827,
+      "step": 2525
+    },
+    {
+      "epoch": 5.862068965517241,
+      "grad_norm": 17.853538513183594,
+      "learning_rate": 1.5711366538952746e-05,
+      "loss": 0.3024,
+      "step": 2550
+    },
+    {
+      "epoch": 5.919540229885057,
+      "grad_norm": 2.493351697921753,
+      "learning_rate": 1.5647509578544062e-05,
+      "loss": 0.559,
+      "step": 2575
+    },
+    {
+      "epoch": 5.977011494252873,
+      "grad_norm": 69.57341766357422,
+      "learning_rate": 1.5583652618135378e-05,
+      "loss": 0.4154,
+      "step": 2600
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9541047188106012,
+      "eval_loss": 0.2397681325674057,
+      "eval_macro_f1": 0.8600498112631202,
+      "eval_runtime": 1.2855,
+      "eval_samples_per_second": 1203.442,
+      "eval_steps_per_second": 75.458,
+      "eval_weighted_f1": 0.9516447037124957,
+      "step": 2610
+    },
+    {
+      "epoch": 6.0344827586206895,
+      "grad_norm": 0.20201376080513,
+      "learning_rate": 1.5519795657726694e-05,
+      "loss": 0.3892,
+      "step": 2625
+    },
+    {
+      "epoch": 6.091954022988506,
+      "grad_norm": 0.17590577900409698,
+      "learning_rate": 1.545593869731801e-05,
+      "loss": 0.2642,
+      "step": 2650
+    },
+    {
+      "epoch": 6.149425287356322,
+      "grad_norm": 14.469496726989746,
+      "learning_rate": 1.5392081736909322e-05,
+      "loss": 0.2353,
+      "step": 2675
+    },
+    {
+      "epoch": 6.206896551724138,
+      "grad_norm": 81.51103973388672,
+      "learning_rate": 1.5328224776500638e-05,
+      "loss": 0.4785,
+      "step": 2700
+    },
+    {
+      "epoch": 6.264367816091954,
+      "grad_norm": 0.19920621812343597,
+      "learning_rate": 1.5264367816091957e-05,
+      "loss": 0.2868,
+      "step": 2725
+    },
+    {
+      "epoch": 6.32183908045977,
+      "grad_norm": 12.891411781311035,
+      "learning_rate": 1.520051085568327e-05,
+      "loss": 0.2671,
+      "step": 2750
+    },
+    {
+      "epoch": 6.379310344827586,
+      "grad_norm": 16.071155548095703,
+      "learning_rate": 1.5136653895274587e-05,
+      "loss": 0.5203,
+      "step": 2775
+    },
+    {
+      "epoch": 6.436781609195402,
+      "grad_norm": 0.24430899322032928,
+      "learning_rate": 1.5072796934865901e-05,
+      "loss": 0.2203,
+      "step": 2800
+    },
+    {
+      "epoch": 6.494252873563219,
+      "grad_norm": 0.21658585965633392,
+      "learning_rate": 1.5008939974457217e-05,
+      "loss": 0.3588,
+      "step": 2825
+    },
+    {
+      "epoch": 6.551724137931035,
+      "grad_norm": 0.218729630112648,
+      "learning_rate": 1.4945083014048531e-05,
+      "loss": 0.3032,
+      "step": 2850
+    },
+    {
+      "epoch": 6.609195402298851,
+      "grad_norm": 0.19259417057037354,
+      "learning_rate": 1.4881226053639847e-05,
+      "loss": 0.1498,
+      "step": 2875
+    },
+    {
+      "epoch": 6.666666666666667,
+      "grad_norm": 12.594219207763672,
+      "learning_rate": 1.4817369093231165e-05,
+      "loss": 0.5802,
+      "step": 2900
+    },
+    {
+      "epoch": 6.724137931034483,
+      "grad_norm": 0.2996440529823303,
+      "learning_rate": 1.4753512132822479e-05,
+      "loss": 0.2958,
+      "step": 2925
+    },
+    {
+      "epoch": 6.781609195402299,
+      "grad_norm": 0.3330232799053192,
+      "learning_rate": 1.4689655172413795e-05,
+      "loss": 0.3348,
+      "step": 2950
+    },
+    {
+      "epoch": 6.8390804597701145,
+      "grad_norm": 55.417236328125,
+      "learning_rate": 1.4625798212005109e-05,
+      "loss": 0.4195,
+      "step": 2975
+    },
+    {
+      "epoch": 6.896551724137931,
+      "grad_norm": 2.850304365158081,
+      "learning_rate": 1.4561941251596425e-05,
+      "loss": 0.4105,
+      "step": 3000
+    },
+    {
+      "epoch": 6.954022988505747,
+      "grad_norm": 0.2381744235754013,
+      "learning_rate": 1.4498084291187742e-05,
+      "loss": 0.3145,
+      "step": 3025
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9495798319327731,
+      "eval_loss": 0.2417064607143402,
+      "eval_macro_f1": 0.8506231029028258,
+      "eval_runtime": 1.2843,
+      "eval_samples_per_second": 1204.53,
+      "eval_steps_per_second": 75.526,
+      "eval_weighted_f1": 0.9484848377448473,
+      "step": 3045
+    },
+    {
+      "epoch": 7.011494252873563,
+      "grad_norm": 0.9536013603210449,
+      "learning_rate": 1.4434227330779056e-05,
+      "loss": 0.4517,
+      "step": 3050
+    },
+    {
+      "epoch": 7.068965517241379,
+      "grad_norm": 0.2184518724679947,
+      "learning_rate": 1.4370370370370372e-05,
+      "loss": 0.3963,
+      "step": 3075
+    },
+    {
+      "epoch": 7.126436781609195,
+      "grad_norm": 0.23555926978588104,
+      "learning_rate": 1.4306513409961686e-05,
+      "loss": 0.1847,
+      "step": 3100
+    },
+    {
+      "epoch": 7.183908045977011,
+      "grad_norm": 0.1601639837026596,
+      "learning_rate": 1.4242656449553002e-05,
+      "loss": 0.0835,
+      "step": 3125
+    },
+    {
+      "epoch": 7.241379310344827,
+      "grad_norm": 2.4888522624969482,
+      "learning_rate": 1.4178799489144316e-05,
+      "loss": 0.436,
+      "step": 3150
+    },
+    {
+      "epoch": 7.2988505747126435,
+      "grad_norm": 13.896317481994629,
+      "learning_rate": 1.4114942528735634e-05,
+      "loss": 0.4248,
+      "step": 3175
+    },
+    {
+      "epoch": 7.35632183908046,
+      "grad_norm": 0.2253255695104599,
+      "learning_rate": 1.405108556832695e-05,
+      "loss": 0.1896,
+      "step": 3200
+    },
+    {
+      "epoch": 7.413793103448276,
+      "grad_norm": 0.5111873149871826,
+      "learning_rate": 1.3987228607918264e-05,
+      "loss": 0.3082,
+      "step": 3225
+    },
+    {
+      "epoch": 7.471264367816092,
+      "grad_norm": 0.3114178478717804,
+      "learning_rate": 1.392337164750958e-05,
+      "loss": 0.354,
+      "step": 3250
+    },
+    {
+      "epoch": 7.528735632183908,
+      "grad_norm": 308.2452392578125,
+      "learning_rate": 1.3859514687100894e-05,
+      "loss": 0.4526,
+      "step": 3275
+    },
+    {
+      "epoch": 7.586206896551724,
+      "grad_norm": 0.30504605174064636,
+      "learning_rate": 1.3795657726692211e-05,
+      "loss": 0.1799,
+      "step": 3300
+    },
+    {
+      "epoch": 7.64367816091954,
+      "grad_norm": 0.23172464966773987,
+      "learning_rate": 1.3731800766283526e-05,
+      "loss": 0.2987,
+      "step": 3325
+    },
+    {
+      "epoch": 7.7011494252873565,
+      "grad_norm": 118.74394226074219,
+      "learning_rate": 1.3667943805874841e-05,
+      "loss": 0.2358,
+      "step": 3350
+    },
+    {
+      "epoch": 7.758620689655173,
+      "grad_norm": 0.14811089634895325,
+      "learning_rate": 1.3604086845466157e-05,
+      "loss": 0.3334,
+      "step": 3375
+    },
+    {
+      "epoch": 7.816091954022989,
+      "grad_norm": 0.26182156801223755,
+      "learning_rate": 1.3540229885057471e-05,
+      "loss": 0.3044,
+      "step": 3400
+    },
+    {
+      "epoch": 7.873563218390805,
+      "grad_norm": 0.42604881525039673,
+      "learning_rate": 1.3476372924648789e-05,
+      "loss": 0.3334,
+      "step": 3425
+    },
+    {
+      "epoch": 7.931034482758621,
+      "grad_norm": 3.8690788745880127,
+      "learning_rate": 1.3412515964240103e-05,
+      "loss": 0.2318,
+      "step": 3450
+    },
+    {
+      "epoch": 7.988505747126437,
+      "grad_norm": 5.706087112426758,
+      "learning_rate": 1.3348659003831419e-05,
+      "loss": 0.2909,
+      "step": 3475
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9515190691661279,
+      "eval_loss": 0.2455778568983078,
+      "eval_macro_f1": 0.853849899937441,
+      "eval_runtime": 1.2822,
+      "eval_samples_per_second": 1206.499,
+      "eval_steps_per_second": 75.65,
+      "eval_weighted_f1": 0.9495985858714602,
+      "step": 3480
+    },
+    {
+      "epoch": 8.045977011494253,
+      "grad_norm": 1.9139503240585327,
+      "learning_rate": 1.3284802043422733e-05,
+      "loss": 0.1426,
+      "step": 3500
+    },
+    {
+      "epoch": 8.10344827586207,
+      "grad_norm": 0.10656487196683884,
+      "learning_rate": 1.3220945083014049e-05,
+      "loss": 0.1865,
+      "step": 3525
+    },
+    {
+      "epoch": 8.160919540229886,
+      "grad_norm": 0.9113854765892029,
+      "learning_rate": 1.3157088122605367e-05,
+      "loss": 0.2629,
+      "step": 3550
+    },
+    {
+      "epoch": 8.218390804597702,
+      "grad_norm": 12.390186309814453,
+      "learning_rate": 1.309323116219668e-05,
+      "loss": 0.3263,
+      "step": 3575
+    },
+    {
+      "epoch": 8.275862068965518,
+      "grad_norm": 16.489566802978516,
+      "learning_rate": 1.3029374201787996e-05,
+      "loss": 0.2176,
+      "step": 3600
+    },
+    {
+      "epoch": 8.333333333333334,
+      "grad_norm": 0.15907344222068787,
+      "learning_rate": 1.296551724137931e-05,
+      "loss": 0.2475,
+      "step": 3625
+    },
+    {
+      "epoch": 8.39080459770115,
+      "grad_norm": 0.163466215133667,
+      "learning_rate": 1.2901660280970626e-05,
+      "loss": 0.1819,
+      "step": 3650
+    },
+    {
+      "epoch": 8.448275862068966,
+      "grad_norm": 83.0460205078125,
+      "learning_rate": 1.283780332056194e-05,
+      "loss": 0.2247,
+      "step": 3675
+    },
+    {
+      "epoch": 8.505747126436782,
+      "grad_norm": 0.15427705645561218,
+      "learning_rate": 1.2773946360153258e-05,
+      "loss": 0.2735,
+      "step": 3700
+    },
+    {
+      "epoch": 8.563218390804598,
+      "grad_norm": 0.1480730175971985,
+      "learning_rate": 1.2710089399744574e-05,
+      "loss": 0.2022,
+      "step": 3725
+    },
+    {
+      "epoch": 8.620689655172415,
+      "grad_norm": 0.3764588236808777,
+      "learning_rate": 1.2646232439335888e-05,
+      "loss": 0.3027,
+      "step": 3750
+    },
+    {
+      "epoch": 8.678160919540229,
+      "grad_norm": 0.17370322346687317,
+      "learning_rate": 1.2582375478927204e-05,
+      "loss": 0.3577,
+      "step": 3775
+    },
+    {
+      "epoch": 8.735632183908045,
+      "grad_norm": 24.64053726196289,
+      "learning_rate": 1.2518518518518518e-05,
+      "loss": 0.2364,
+      "step": 3800
+    },
+    {
+      "epoch": 8.793103448275861,
+      "grad_norm": 5.574195861816406,
+      "learning_rate": 1.2454661558109836e-05,
+      "loss": 0.2676,
+      "step": 3825
+    },
+    {
+      "epoch": 8.850574712643677,
+      "grad_norm": 0.09678659588098526,
+      "learning_rate": 1.2390804597701151e-05,
+      "loss": 0.2759,
+      "step": 3850
+    },
+    {
+      "epoch": 8.908045977011493,
+      "grad_norm": 0.1522231101989746,
+      "learning_rate": 1.2326947637292466e-05,
+      "loss": 0.2181,
+      "step": 3875
+    },
+    {
+      "epoch": 8.96551724137931,
+      "grad_norm": 0.37765446305274963,
+      "learning_rate": 1.2263090676883781e-05,
+      "loss": 0.4644,
+      "step": 3900
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.26898178458213806,
+      "eval_macro_f1": 0.8663256210777838,
+      "eval_runtime": 1.2494,
+      "eval_samples_per_second": 1238.187,
+      "eval_steps_per_second": 77.637,
+      "eval_weighted_f1": 0.9512561247191414,
+      "step": 3915
+    },
+    {
+      "epoch": 9.022988505747126,
+      "grad_norm": 1.2754137516021729,
+      "learning_rate": 1.2199233716475096e-05,
+      "loss": 0.2067,
+      "step": 3925
+    },
+    {
+      "epoch": 9.080459770114942,
+      "grad_norm": 2.7214295864105225,
+      "learning_rate": 1.2135376756066413e-05,
+      "loss": 0.1737,
+      "step": 3950
+    },
+    {
+      "epoch": 9.137931034482758,
+      "grad_norm": 10.367477416992188,
+      "learning_rate": 1.2071519795657727e-05,
+      "loss": 0.3776,
+      "step": 3975
+    },
+    {
+      "epoch": 9.195402298850574,
+      "grad_norm": 0.20507702231407166,
+      "learning_rate": 1.2007662835249043e-05,
+      "loss": 0.1399,
+      "step": 4000
+    },
+    {
+      "epoch": 9.25287356321839,
+      "grad_norm": 3.616419553756714,
+      "learning_rate": 1.1943805874840359e-05,
+      "loss": 0.2167,
+      "step": 4025
+    },
+    {
+      "epoch": 9.310344827586206,
+      "grad_norm": 0.13054685294628143,
+      "learning_rate": 1.1879948914431673e-05,
+      "loss": 0.1431,
+      "step": 4050
+    },
+    {
+      "epoch": 9.367816091954023,
+      "grad_norm": 0.11281740665435791,
+      "learning_rate": 1.181609195402299e-05,
+      "loss": 0.2202,
+      "step": 4075
+    },
+    {
+      "epoch": 9.425287356321839,
+      "grad_norm": 0.6165698170661926,
+      "learning_rate": 1.1752234993614305e-05,
+      "loss": 0.1859,
+      "step": 4100
+    },
+    {
+      "epoch": 9.482758620689655,
+      "grad_norm": 1.5480984449386597,
+      "learning_rate": 1.168837803320562e-05,
+      "loss": 0.2284,
+      "step": 4125
+    },
+    {
+      "epoch": 9.540229885057471,
+      "grad_norm": 6.850103855133057,
+      "learning_rate": 1.1624521072796935e-05,
+      "loss": 0.1698,
+      "step": 4150
+    },
+    {
+      "epoch": 9.597701149425287,
+      "grad_norm": 3.5420010089874268,
+      "learning_rate": 1.156066411238825e-05,
+      "loss": 0.3109,
+      "step": 4175
+    },
+    {
+      "epoch": 9.655172413793103,
+      "grad_norm": 0.1575000286102295,
+      "learning_rate": 1.1496807151979568e-05,
+      "loss": 0.1249,
+      "step": 4200
+    },
+    {
+      "epoch": 9.71264367816092,
+      "grad_norm": 86.61784362792969,
+      "learning_rate": 1.1432950191570882e-05,
+      "loss": 0.1512,
+      "step": 4225
+    },
+    {
+      "epoch": 9.770114942528735,
+      "grad_norm": 6.673755645751953,
+      "learning_rate": 1.1369093231162198e-05,
+      "loss": 0.2927,
+      "step": 4250
+    },
+    {
+      "epoch": 9.827586206896552,
+      "grad_norm": 3.7119579315185547,
+      "learning_rate": 1.1305236270753512e-05,
+      "loss": 0.1886,
+      "step": 4275
+    },
+    {
+      "epoch": 9.885057471264368,
+      "grad_norm": 7.661352157592773,
+      "learning_rate": 1.1241379310344828e-05,
+      "loss": 0.2061,
+      "step": 4300
+    },
+    {
+      "epoch": 9.942528735632184,
+      "grad_norm": 4.866251468658447,
+      "learning_rate": 1.1177522349936142e-05,
+      "loss": 0.2876,
+      "step": 4325
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 17.001205444335938,
+      "learning_rate": 1.111366538952746e-05,
+      "loss": 0.1963,
+      "step": 4350
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.9495798319327731,
+      "eval_loss": 0.2947230339050293,
+      "eval_macro_f1": 0.8586701439701274,
+      "eval_runtime": 1.2427,
+      "eval_samples_per_second": 1244.9,
+      "eval_steps_per_second": 78.058,
+      "eval_weighted_f1": 0.94867080753628,
+      "step": 4350
+    },
+    {
+      "epoch": 10.057471264367816,
+      "grad_norm": 0.09062144160270691,
+      "learning_rate": 1.1049808429118776e-05,
+      "loss": 0.0144,
+      "step": 4375
+    },
+    {
+      "epoch": 10.114942528735632,
+      "grad_norm": 0.09079045802354813,
+      "learning_rate": 1.098595146871009e-05,
+      "loss": 0.2467,
+      "step": 4400
+    },
+    {
+      "epoch": 10.172413793103448,
+      "grad_norm": 16.911344528198242,
+      "learning_rate": 1.0922094508301406e-05,
+      "loss": 0.1387,
+      "step": 4425
+    },
+    {
+      "epoch": 10.229885057471265,
+      "grad_norm": 0.12063690274953842,
+      "learning_rate": 1.085823754789272e-05,
+      "loss": 0.1926,
+      "step": 4450
+    },
+    {
+      "epoch": 10.28735632183908,
+      "grad_norm": 0.3631066679954529,
+      "learning_rate": 1.0794380587484037e-05,
+      "loss": 0.1826,
+      "step": 4475
+    },
+    {
+      "epoch": 10.344827586206897,
+      "grad_norm": 0.36074990034103394,
+      "learning_rate": 1.0730523627075351e-05,
+      "loss": 0.161,
+      "step": 4500
+    },
+    {
+      "epoch": 10.402298850574713,
+      "grad_norm": 55.926822662353516,
+      "learning_rate": 1.0666666666666667e-05,
+      "loss": 0.2874,
+      "step": 4525
+    },
+    {
+      "epoch": 10.459770114942529,
+      "grad_norm": 0.10639549791812897,
+      "learning_rate": 1.0602809706257983e-05,
+      "loss": 0.1727,
+      "step": 4550
+    },
+    {
+      "epoch": 10.517241379310345,
+      "grad_norm": 0.7862879633903503,
+      "learning_rate": 1.0538952745849297e-05,
+      "loss": 0.274,
+      "step": 4575
+    },
+    {
+      "epoch": 10.574712643678161,
+      "grad_norm": 0.27984291315078735,
+      "learning_rate": 1.0475095785440615e-05,
+      "loss": 0.2045,
+      "step": 4600
+    },
+    {
+      "epoch": 10.632183908045977,
+      "grad_norm": 17.181734085083008,
+      "learning_rate": 1.0411238825031929e-05,
+      "loss": 0.3352,
+      "step": 4625
+    },
+    {
+      "epoch": 10.689655172413794,
+      "grad_norm": 0.11635159701108932,
+      "learning_rate": 1.0347381864623245e-05,
+      "loss": 0.134,
+      "step": 4650
+    },
+    {
+      "epoch": 10.74712643678161,
+      "grad_norm": 0.07812677323818207,
+      "learning_rate": 1.028352490421456e-05,
+      "loss": 0.0612,
+      "step": 4675
+    },
+    {
+      "epoch": 10.804597701149426,
+      "grad_norm": 0.09490698575973511,
+      "learning_rate": 1.0219667943805875e-05,
+      "loss": 0.2156,
+      "step": 4700
+    },
+    {
+      "epoch": 10.862068965517242,
+      "grad_norm": 7.355884075164795,
+      "learning_rate": 1.0155810983397192e-05,
+      "loss": 0.2202,
+      "step": 4725
+    },
+    {
+      "epoch": 10.919540229885058,
+      "grad_norm": 5.701615333557129,
+      "learning_rate": 1.0091954022988506e-05,
+      "loss": 0.1911,
+      "step": 4750
+    },
+    {
+      "epoch": 10.977011494252874,
+      "grad_norm": 5.884024620056152,
+      "learning_rate": 1.0028097062579822e-05,
+      "loss": 0.1841,
+      "step": 4775
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.9521654815772462,
+      "eval_loss": 0.3065274953842163,
+      "eval_macro_f1": 0.8665082853654282,
+      "eval_runtime": 1.3338,
+      "eval_samples_per_second": 1159.816,
+      "eval_steps_per_second": 72.723,
+      "eval_weighted_f1": 0.9509302446253226,
+      "step": 4785
+    },
+    {
+      "epoch": 11.03448275862069,
+      "grad_norm": 0.10210714489221573,
+      "learning_rate": 9.964240102171138e-06,
+      "loss": 0.1149,
+      "step": 4800
+    },
+    {
+      "epoch": 11.091954022988507,
+      "grad_norm": 0.07134879380464554,
+      "learning_rate": 9.900383141762452e-06,
+      "loss": 0.1125,
+      "step": 4825
+    },
+    {
+      "epoch": 11.149425287356323,
+      "grad_norm": 0.6779069900512695,
+      "learning_rate": 9.836526181353768e-06,
+      "loss": 0.134,
+      "step": 4850
+    },
+    {
+      "epoch": 11.206896551724139,
+      "grad_norm": 0.09422433376312256,
+      "learning_rate": 9.772669220945084e-06,
+      "loss": 0.233,
+      "step": 4875
+    },
+    {
+      "epoch": 11.264367816091955,
+      "grad_norm": 0.07436396181583405,
+      "learning_rate": 9.708812260536398e-06,
+      "loss": 0.1367,
+      "step": 4900
+    },
+    {
+      "epoch": 11.32183908045977,
+      "grad_norm": 15.747620582580566,
+      "learning_rate": 9.644955300127716e-06,
+      "loss": 0.1444,
+      "step": 4925
+    },
+    {
+      "epoch": 11.379310344827585,
+      "grad_norm": 0.3826606571674347,
+      "learning_rate": 9.58109833971903e-06,
+      "loss": 0.1271,
+      "step": 4950
+    },
+    {
+      "epoch": 11.436781609195402,
+      "grad_norm": 0.0642637312412262,
+      "learning_rate": 9.517241379310346e-06,
+      "loss": 0.1503,
+      "step": 4975
+    },
+    {
+      "epoch": 11.494252873563218,
+      "grad_norm": 0.16133825480937958,
+      "learning_rate": 9.453384418901661e-06,
+      "loss": 0.1837,
+      "step": 5000
+    },
+    {
+      "epoch": 11.551724137931034,
+      "grad_norm": 0.15818732976913452,
+      "learning_rate": 9.389527458492976e-06,
+      "loss": 0.0784,
+      "step": 5025
+    },
+    {
+      "epoch": 11.60919540229885,
+      "grad_norm": 0.06489365547895432,
+      "learning_rate": 9.325670498084291e-06,
+      "loss": 0.1532,
+      "step": 5050
+    },
+    {
+      "epoch": 11.666666666666666,
+      "grad_norm": 0.05736462399363518,
+      "learning_rate": 9.261813537675607e-06,
+      "loss": 0.1367,
+      "step": 5075
+    },
+    {
+      "epoch": 11.724137931034482,
+      "grad_norm": 0.9656385183334351,
+      "learning_rate": 9.197956577266923e-06,
+      "loss": 0.0903,
+      "step": 5100
+    },
+    {
+      "epoch": 11.781609195402298,
+      "grad_norm": 0.06174295395612717,
+      "learning_rate": 9.134099616858239e-06,
+      "loss": 0.0899,
+      "step": 5125
+    },
+    {
+      "epoch": 11.839080459770114,
+      "grad_norm": 0.060515888035297394,
+      "learning_rate": 9.070242656449553e-06,
+      "loss": 0.1605,
+      "step": 5150
+    },
+    {
+      "epoch": 11.89655172413793,
+      "grad_norm": 3.496469497680664,
+      "learning_rate": 9.006385696040869e-06,
+      "loss": 0.1872,
+      "step": 5175
+    },
+    {
+      "epoch": 11.954022988505747,
+      "grad_norm": 0.07404470443725586,
+      "learning_rate": 8.942528735632185e-06,
+      "loss": 0.1987,
+      "step": 5200
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.9508726567550096,
+      "eval_loss": 0.2994757890701294,
+      "eval_macro_f1": 0.8675740894345955,
+      "eval_runtime": 1.2571,
+      "eval_samples_per_second": 1230.643,
+      "eval_steps_per_second": 77.164,
+      "eval_weighted_f1": 0.9504132974662869,
+      "step": 5220
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 8700,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 0
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f0712802a8bbb778f406abcda3b020380ef5f43ee0e78335abc61a86f8ff84d
+size 5624

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<mask>": 64000
+}

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c64231604c8d2524f73e46c13a5682018d93c1d4639ce3114f161a3718c5bef2
+size 540026436

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9eb873326b0e1a1f699b6aa64ce8594adb59d849d8a564404109f28a922684aa
+size 1075446330

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b98915e397aab486973bd7a24fb28d450e10ec8c421d01a59c90d08cb7a12f23
+size 14244

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c16d9a42a51b51ff81efdb48e1615074f87f8c6018d865e3a0a45a7f8cbcf6c5
+size 1064

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64000": {
+      "content": "<mask>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "mask_token": "<mask>",
+  "model_max_length": 256,
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "tokenizer_class": "PhobertTokenizer",
+  "unk_token": "<unk>"
+}

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/trainer_state.json

@@ -0,0 +1,2035 @@
+{
+  "best_global_step": 5220,
+  "best_metric": 0.8675740894345955,
+  "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-5220",
+  "epoch": 15.0,
+  "eval_steps": 500,
+  "global_step": 6525,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05747126436781609,
+      "grad_norm": 16.450016021728516,
+      "learning_rate": 5.517241379310345e-07,
+      "loss": 4.464,
+      "step": 25
+    },
+    {
+      "epoch": 0.11494252873563218,
+      "grad_norm": 15.485475540161133,
+      "learning_rate": 1.1264367816091955e-06,
+      "loss": 4.2217,
+      "step": 50
+    },
+    {
+      "epoch": 0.1724137931034483,
+      "grad_norm": 13.874435424804688,
+      "learning_rate": 1.7011494252873565e-06,
+      "loss": 3.8906,
+      "step": 75
+    },
+    {
+      "epoch": 0.22988505747126436,
+      "grad_norm": 16.325193405151367,
+      "learning_rate": 2.2758620689655173e-06,
+      "loss": 3.5109,
+      "step": 100
+    },
+    {
+      "epoch": 0.28735632183908044,
+      "grad_norm": 12.38318157196045,
+      "learning_rate": 2.850574712643678e-06,
+      "loss": 3.0685,
+      "step": 125
+    },
+    {
+      "epoch": 0.3448275862068966,
+      "grad_norm": 17.294038772583008,
+      "learning_rate": 3.425287356321839e-06,
+      "loss": 2.541,
+      "step": 150
+    },
+    {
+      "epoch": 0.40229885057471265,
+      "grad_norm": 17.30904769897461,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 2.1361,
+      "step": 175
+    },
+    {
+      "epoch": 0.45977011494252873,
+      "grad_norm": 15.655984878540039,
+      "learning_rate": 4.5747126436781615e-06,
+      "loss": 1.8631,
+      "step": 200
+    },
+    {
+      "epoch": 0.5172413793103449,
+      "grad_norm": 30.74538230895996,
+      "learning_rate": 5.149425287356322e-06,
+      "loss": 1.4674,
+      "step": 225
+    },
+    {
+      "epoch": 0.5747126436781609,
+      "grad_norm": 26.51157569885254,
+      "learning_rate": 5.724137931034483e-06,
+      "loss": 1.4242,
+      "step": 250
+    },
+    {
+      "epoch": 0.632183908045977,
+      "grad_norm": 15.323351860046387,
+      "learning_rate": 6.298850574712644e-06,
+      "loss": 1.2337,
+      "step": 275
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 11.085951805114746,
+      "learning_rate": 6.873563218390805e-06,
+      "loss": 1.235,
+      "step": 300
+    },
+    {
+      "epoch": 0.7471264367816092,
+      "grad_norm": 17.022886276245117,
+      "learning_rate": 7.4482758620689665e-06,
+      "loss": 1.123,
+      "step": 325
+    },
+    {
+      "epoch": 0.8045977011494253,
+      "grad_norm": 7.172950267791748,
+      "learning_rate": 8.022988505747126e-06,
+      "loss": 1.1503,
+      "step": 350
+    },
+    {
+      "epoch": 0.8620689655172413,
+      "grad_norm": 19.81337547302246,
+      "learning_rate": 8.597701149425289e-06,
+      "loss": 1.0711,
+      "step": 375
+    },
+    {
+      "epoch": 0.9195402298850575,
+      "grad_norm": 23.27373695373535,
+      "learning_rate": 9.172413793103448e-06,
+      "loss": 0.9703,
+      "step": 400
+    },
+    {
+      "epoch": 0.9770114942528736,
+      "grad_norm": 21.556955337524414,
+      "learning_rate": 9.74712643678161e-06,
+      "loss": 0.9083,
+      "step": 425
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9372979961215255,
+      "eval_loss": 0.21047352254390717,
+      "eval_macro_f1": 0.7385325809144866,
+      "eval_runtime": 1.2627,
+      "eval_samples_per_second": 1225.158,
+      "eval_steps_per_second": 76.82,
+      "eval_weighted_f1": 0.926186545957586,
+      "step": 435
+    },
+    {
+      "epoch": 1.0344827586206897,
+      "grad_norm": 20.03977394104004,
+      "learning_rate": 1.032183908045977e-05,
+      "loss": 1.1259,
+      "step": 450
+    },
+    {
+      "epoch": 1.0919540229885056,
+      "grad_norm": 28.675647735595703,
+      "learning_rate": 1.0896551724137932e-05,
+      "loss": 0.9905,
+      "step": 475
+    },
+    {
+      "epoch": 1.1494252873563218,
+      "grad_norm": 28.670623779296875,
+      "learning_rate": 1.1471264367816093e-05,
+      "loss": 0.7334,
+      "step": 500
+    },
+    {
+      "epoch": 1.206896551724138,
+      "grad_norm": 28.721160888671875,
+      "learning_rate": 1.2045977011494254e-05,
+      "loss": 0.9009,
+      "step": 525
+    },
+    {
+      "epoch": 1.264367816091954,
+      "grad_norm": 44.211883544921875,
+      "learning_rate": 1.2620689655172415e-05,
+      "loss": 0.7161,
+      "step": 550
+    },
+    {
+      "epoch": 1.3218390804597702,
+      "grad_norm": 15.909297943115234,
+      "learning_rate": 1.3195402298850574e-05,
+      "loss": 0.9911,
+      "step": 575
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "grad_norm": 20.90113067626953,
+      "learning_rate": 1.3770114942528736e-05,
+      "loss": 0.8012,
+      "step": 600
+    },
+    {
+      "epoch": 1.4367816091954024,
+      "grad_norm": 27.909393310546875,
+      "learning_rate": 1.4344827586206897e-05,
+      "loss": 0.8335,
+      "step": 625
+    },
+    {
+      "epoch": 1.4942528735632183,
+      "grad_norm": 4.127943992614746,
+      "learning_rate": 1.4919540229885058e-05,
+      "loss": 0.887,
+      "step": 650
+    },
+    {
+      "epoch": 1.5517241379310345,
+      "grad_norm": 13.942008972167969,
+      "learning_rate": 1.549425287356322e-05,
+      "loss": 0.9997,
+      "step": 675
+    },
+    {
+      "epoch": 1.6091954022988506,
+      "grad_norm": 150.57318115234375,
+      "learning_rate": 1.6068965517241383e-05,
+      "loss": 0.8995,
+      "step": 700
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 84.19149017333984,
+      "learning_rate": 1.664367816091954e-05,
+      "loss": 0.7276,
+      "step": 725
+    },
+    {
+      "epoch": 1.7241379310344827,
+      "grad_norm": 0.8938015103340149,
+      "learning_rate": 1.72183908045977e-05,
+      "loss": 0.7947,
+      "step": 750
+    },
+    {
+      "epoch": 1.7816091954022988,
+      "grad_norm": 21.69007110595703,
+      "learning_rate": 1.7793103448275862e-05,
+      "loss": 0.896,
+      "step": 775
+    },
+    {
+      "epoch": 1.839080459770115,
+      "grad_norm": 9.371072769165039,
+      "learning_rate": 1.8367816091954027e-05,
+      "loss": 0.9257,
+      "step": 800
+    },
+    {
+      "epoch": 1.896551724137931,
+      "grad_norm": 9.058280944824219,
+      "learning_rate": 1.8942528735632187e-05,
+      "loss": 0.7923,
+      "step": 825
+    },
+    {
+      "epoch": 1.9540229885057472,
+      "grad_norm": 19.173526763916016,
+      "learning_rate": 1.9517241379310345e-05,
+      "loss": 0.829,
+      "step": 850
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9444085326438267,
+      "eval_loss": 0.1997242271900177,
+      "eval_macro_f1": 0.7984669829693747,
+      "eval_runtime": 1.2248,
+      "eval_samples_per_second": 1263.023,
+      "eval_steps_per_second": 79.194,
+      "eval_weighted_f1": 0.9372675430430231,
+      "step": 870
+    },
+    {
+      "epoch": 2.0114942528735633,
+      "grad_norm": 13.507299423217773,
+      "learning_rate": 1.9989782886334612e-05,
+      "loss": 1.0265,
+      "step": 875
+    },
+    {
+      "epoch": 2.0689655172413794,
+      "grad_norm": 29.802522659301758,
+      "learning_rate": 1.9925925925925928e-05,
+      "loss": 0.6546,
+      "step": 900
+    },
+    {
+      "epoch": 2.1264367816091956,
+      "grad_norm": 20.382816314697266,
+      "learning_rate": 1.9862068965517244e-05,
+      "loss": 0.6422,
+      "step": 925
+    },
+    {
+      "epoch": 2.1839080459770113,
+      "grad_norm": 16.437686920166016,
+      "learning_rate": 1.979821200510856e-05,
+      "loss": 0.6465,
+      "step": 950
+    },
+    {
+      "epoch": 2.2413793103448274,
+      "grad_norm": 14.521719932556152,
+      "learning_rate": 1.9734355044699872e-05,
+      "loss": 0.7665,
+      "step": 975
+    },
+    {
+      "epoch": 2.2988505747126435,
+      "grad_norm": 25.880773544311523,
+      "learning_rate": 1.9670498084291188e-05,
+      "loss": 0.8589,
+      "step": 1000
+    },
+    {
+      "epoch": 2.3563218390804597,
+      "grad_norm": 20.663999557495117,
+      "learning_rate": 1.9606641123882504e-05,
+      "loss": 0.7237,
+      "step": 1025
+    },
+    {
+      "epoch": 2.413793103448276,
+      "grad_norm": 15.50725269317627,
+      "learning_rate": 1.954278416347382e-05,
+      "loss": 0.737,
+      "step": 1050
+    },
+    {
+      "epoch": 2.471264367816092,
+      "grad_norm": 1.7924487590789795,
+      "learning_rate": 1.9478927203065136e-05,
+      "loss": 0.7695,
+      "step": 1075
+    },
+    {
+      "epoch": 2.528735632183908,
+      "grad_norm": 37.092926025390625,
+      "learning_rate": 1.941507024265645e-05,
+      "loss": 0.6687,
+      "step": 1100
+    },
+    {
+      "epoch": 2.586206896551724,
+      "grad_norm": 0.8467792868614197,
+      "learning_rate": 1.9351213282247767e-05,
+      "loss": 0.6213,
+      "step": 1125
+    },
+    {
+      "epoch": 2.6436781609195403,
+      "grad_norm": 17.73621940612793,
+      "learning_rate": 1.928735632183908e-05,
+      "loss": 0.8202,
+      "step": 1150
+    },
+    {
+      "epoch": 2.7011494252873565,
+      "grad_norm": 3.8395161628723145,
+      "learning_rate": 1.92234993614304e-05,
+      "loss": 0.6882,
+      "step": 1175
+    },
+    {
+      "epoch": 2.7586206896551726,
+      "grad_norm": 77.6451416015625,
+      "learning_rate": 1.915964240102171e-05,
+      "loss": 0.5662,
+      "step": 1200
+    },
+    {
+      "epoch": 2.8160919540229887,
+      "grad_norm": 9.742815017700195,
+      "learning_rate": 1.9095785440613027e-05,
+      "loss": 0.818,
+      "step": 1225
+    },
+    {
+      "epoch": 2.873563218390805,
+      "grad_norm": 41.80664825439453,
+      "learning_rate": 1.9031928480204343e-05,
+      "loss": 0.8144,
+      "step": 1250
+    },
+    {
+      "epoch": 2.9310344827586206,
+      "grad_norm": 13.355256080627441,
+      "learning_rate": 1.896807151979566e-05,
+      "loss": 0.7142,
+      "step": 1275
+    },
+    {
+      "epoch": 2.9885057471264367,
+      "grad_norm": 3.4613044261932373,
+      "learning_rate": 1.8904214559386975e-05,
+      "loss": 0.6519,
+      "step": 1300
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9489334195216548,
+      "eval_loss": 0.1842694729566574,
+      "eval_macro_f1": 0.8417608126834946,
+      "eval_runtime": 1.2612,
+      "eval_samples_per_second": 1226.651,
+      "eval_steps_per_second": 76.913,
+      "eval_weighted_f1": 0.9451926750188842,
+      "step": 1305
+    },
+    {
+      "epoch": 3.045977011494253,
+      "grad_norm": 9.822381973266602,
+      "learning_rate": 1.884035759897829e-05,
+      "loss": 0.5464,
+      "step": 1325
+    },
+    {
+      "epoch": 3.103448275862069,
+      "grad_norm": 14.551055908203125,
+      "learning_rate": 1.8776500638569606e-05,
+      "loss": 0.4742,
+      "step": 1350
+    },
+    {
+      "epoch": 3.160919540229885,
+      "grad_norm": 0.7838442921638489,
+      "learning_rate": 1.871264367816092e-05,
+      "loss": 0.5285,
+      "step": 1375
+    },
+    {
+      "epoch": 3.218390804597701,
+      "grad_norm": 32.79100799560547,
+      "learning_rate": 1.8648786717752235e-05,
+      "loss": 0.4901,
+      "step": 1400
+    },
+    {
+      "epoch": 3.2758620689655173,
+      "grad_norm": 17.9625244140625,
+      "learning_rate": 1.8584929757343554e-05,
+      "loss": 0.7674,
+      "step": 1425
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 38.27335739135742,
+      "learning_rate": 1.8521072796934866e-05,
+      "loss": 0.6257,
+      "step": 1450
+    },
+    {
+      "epoch": 3.3908045977011496,
+      "grad_norm": 0.9759963750839233,
+      "learning_rate": 1.8457215836526182e-05,
+      "loss": 0.5915,
+      "step": 1475
+    },
+    {
+      "epoch": 3.4482758620689653,
+      "grad_norm": 32.85970687866211,
+      "learning_rate": 1.8393358876117498e-05,
+      "loss": 0.6667,
+      "step": 1500
+    },
+    {
+      "epoch": 3.5057471264367814,
+      "grad_norm": 26.64856719970703,
+      "learning_rate": 1.8329501915708814e-05,
+      "loss": 0.7413,
+      "step": 1525
+    },
+    {
+      "epoch": 3.5632183908045976,
+      "grad_norm": 0.5582925081253052,
+      "learning_rate": 1.826564495530013e-05,
+      "loss": 0.5404,
+      "step": 1550
+    },
+    {
+      "epoch": 3.6206896551724137,
+      "grad_norm": 28.135393142700195,
+      "learning_rate": 1.8201787994891446e-05,
+      "loss": 0.8905,
+      "step": 1575
+    },
+    {
+      "epoch": 3.67816091954023,
+      "grad_norm": 0.7919555902481079,
+      "learning_rate": 1.813793103448276e-05,
+      "loss": 0.5488,
+      "step": 1600
+    },
+    {
+      "epoch": 3.735632183908046,
+      "grad_norm": 280.8408203125,
+      "learning_rate": 1.8074074074074074e-05,
+      "loss": 0.586,
+      "step": 1625
+    },
+    {
+      "epoch": 3.793103448275862,
+      "grad_norm": 21.92925453186035,
+      "learning_rate": 1.801021711366539e-05,
+      "loss": 0.5316,
+      "step": 1650
+    },
+    {
+      "epoch": 3.8505747126436782,
+      "grad_norm": 14.422163963317871,
+      "learning_rate": 1.7946360153256706e-05,
+      "loss": 0.5578,
+      "step": 1675
+    },
+    {
+      "epoch": 3.9080459770114944,
+      "grad_norm": 19.861310958862305,
+      "learning_rate": 1.788250319284802e-05,
+      "loss": 0.648,
+      "step": 1700
+    },
+    {
+      "epoch": 3.9655172413793105,
+      "grad_norm": 1.0063120126724243,
+      "learning_rate": 1.7818646232439337e-05,
+      "loss": 0.5442,
+      "step": 1725
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9469941822882999,
+      "eval_loss": 0.25301477313041687,
+      "eval_macro_f1": 0.8476622089963451,
+      "eval_runtime": 1.2449,
+      "eval_samples_per_second": 1242.637,
+      "eval_steps_per_second": 77.916,
+      "eval_weighted_f1": 0.9441704666909856,
+      "step": 1740
+    },
+    {
+      "epoch": 4.022988505747127,
+      "grad_norm": 5.532073974609375,
+      "learning_rate": 1.7754789272030653e-05,
+      "loss": 0.5918,
+      "step": 1750
+    },
+    {
+      "epoch": 4.080459770114943,
+      "grad_norm": 18.194129943847656,
+      "learning_rate": 1.769093231162197e-05,
+      "loss": 0.3519,
+      "step": 1775
+    },
+    {
+      "epoch": 4.137931034482759,
+      "grad_norm": 0.2660662531852722,
+      "learning_rate": 1.762707535121328e-05,
+      "loss": 0.3811,
+      "step": 1800
+    },
+    {
+      "epoch": 4.195402298850575,
+      "grad_norm": 16.584251403808594,
+      "learning_rate": 1.75632183908046e-05,
+      "loss": 0.5086,
+      "step": 1825
+    },
+    {
+      "epoch": 4.252873563218391,
+      "grad_norm": 19.36803436279297,
+      "learning_rate": 1.7499361430395913e-05,
+      "loss": 0.4792,
+      "step": 1850
+    },
+    {
+      "epoch": 4.310344827586207,
+      "grad_norm": 50.34165954589844,
+      "learning_rate": 1.743550446998723e-05,
+      "loss": 0.5772,
+      "step": 1875
+    },
+    {
+      "epoch": 4.3678160919540225,
+      "grad_norm": 13.773152351379395,
+      "learning_rate": 1.7371647509578545e-05,
+      "loss": 0.5231,
+      "step": 1900
+    },
+    {
+      "epoch": 4.425287356321839,
+      "grad_norm": 427.54718017578125,
+      "learning_rate": 1.730779054916986e-05,
+      "loss": 0.2846,
+      "step": 1925
+    },
+    {
+      "epoch": 4.482758620689655,
+      "grad_norm": 11.35075855255127,
+      "learning_rate": 1.7243933588761176e-05,
+      "loss": 0.5971,
+      "step": 1950
+    },
+    {
+      "epoch": 4.540229885057471,
+      "grad_norm": 0.7868359684944153,
+      "learning_rate": 1.7180076628352492e-05,
+      "loss": 0.3201,
+      "step": 1975
+    },
+    {
+      "epoch": 4.597701149425287,
+      "grad_norm": 1.2590872049331665,
+      "learning_rate": 1.7116219667943808e-05,
+      "loss": 0.4207,
+      "step": 2000
+    },
+    {
+      "epoch": 4.655172413793103,
+      "grad_norm": 19.67682456970215,
+      "learning_rate": 1.705236270753512e-05,
+      "loss": 0.6388,
+      "step": 2025
+    },
+    {
+      "epoch": 4.712643678160919,
+      "grad_norm": 21.207443237304688,
+      "learning_rate": 1.6988505747126436e-05,
+      "loss": 0.5218,
+      "step": 2050
+    },
+    {
+      "epoch": 4.7701149425287355,
+      "grad_norm": 25.38108253479004,
+      "learning_rate": 1.6924648786717756e-05,
+      "loss": 0.6049,
+      "step": 2075
+    },
+    {
+      "epoch": 4.827586206896552,
+      "grad_norm": 57.224735260009766,
+      "learning_rate": 1.6860791826309068e-05,
+      "loss": 0.6162,
+      "step": 2100
+    },
+    {
+      "epoch": 4.885057471264368,
+      "grad_norm": 3.48508882522583,
+      "learning_rate": 1.6796934865900384e-05,
+      "loss": 0.5359,
+      "step": 2125
+    },
+    {
+      "epoch": 4.942528735632184,
+      "grad_norm": 44.712677001953125,
+      "learning_rate": 1.67330779054917e-05,
+      "loss": 0.4515,
+      "step": 2150
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 57.54368209838867,
+      "learning_rate": 1.6669220945083016e-05,
+      "loss": 0.5796,
+      "step": 2175
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.2165171504020691,
+      "eval_macro_f1": 0.8580507118283248,
+      "eval_runtime": 1.2482,
+      "eval_samples_per_second": 1239.372,
+      "eval_steps_per_second": 77.711,
+      "eval_weighted_f1": 0.948388591648449,
+      "step": 2175
+    },
+    {
+      "epoch": 5.057471264367816,
+      "grad_norm": 12.27694320678711,
+      "learning_rate": 1.6605363984674328e-05,
+      "loss": 0.3551,
+      "step": 2200
+    },
+    {
+      "epoch": 5.114942528735632,
+      "grad_norm": 1.7980570793151855,
+      "learning_rate": 1.6541507024265647e-05,
+      "loss": 0.2956,
+      "step": 2225
+    },
+    {
+      "epoch": 5.172413793103448,
+      "grad_norm": 1.7773587703704834,
+      "learning_rate": 1.6477650063856963e-05,
+      "loss": 0.4865,
+      "step": 2250
+    },
+    {
+      "epoch": 5.2298850574712645,
+      "grad_norm": 9.462156295776367,
+      "learning_rate": 1.6413793103448276e-05,
+      "loss": 0.4026,
+      "step": 2275
+    },
+    {
+      "epoch": 5.287356321839081,
+      "grad_norm": 2.051429510116577,
+      "learning_rate": 1.634993614303959e-05,
+      "loss": 0.3801,
+      "step": 2300
+    },
+    {
+      "epoch": 5.344827586206897,
+      "grad_norm": 0.48756808042526245,
+      "learning_rate": 1.6286079182630907e-05,
+      "loss": 0.3938,
+      "step": 2325
+    },
+    {
+      "epoch": 5.402298850574713,
+      "grad_norm": 2.7134180068969727,
+      "learning_rate": 1.6222222222222223e-05,
+      "loss": 0.5045,
+      "step": 2350
+    },
+    {
+      "epoch": 5.459770114942529,
+      "grad_norm": 0.3124593496322632,
+      "learning_rate": 1.615836526181354e-05,
+      "loss": 0.4567,
+      "step": 2375
+    },
+    {
+      "epoch": 5.517241379310345,
+      "grad_norm": 0.5791522264480591,
+      "learning_rate": 1.6094508301404855e-05,
+      "loss": 0.3833,
+      "step": 2400
+    },
+    {
+      "epoch": 5.574712643678161,
+      "grad_norm": 64.2642593383789,
+      "learning_rate": 1.603065134099617e-05,
+      "loss": 0.4675,
+      "step": 2425
+    },
+    {
+      "epoch": 5.6321839080459775,
+      "grad_norm": 137.23037719726562,
+      "learning_rate": 1.5966794380587483e-05,
+      "loss": 0.4846,
+      "step": 2450
+    },
+    {
+      "epoch": 5.689655172413794,
+      "grad_norm": 15.930166244506836,
+      "learning_rate": 1.5902937420178802e-05,
+      "loss": 0.4571,
+      "step": 2475
+    },
+    {
+      "epoch": 5.747126436781609,
+      "grad_norm": 14.224031448364258,
+      "learning_rate": 1.5839080459770115e-05,
+      "loss": 0.3343,
+      "step": 2500
+    },
+    {
+      "epoch": 5.804597701149425,
+      "grad_norm": 19.296125411987305,
+      "learning_rate": 1.577522349936143e-05,
+      "loss": 0.5827,
+      "step": 2525
+    },
+    {
+      "epoch": 5.862068965517241,
+      "grad_norm": 17.853538513183594,
+      "learning_rate": 1.5711366538952746e-05,
+      "loss": 0.3024,
+      "step": 2550
+    },
+    {
+      "epoch": 5.919540229885057,
+      "grad_norm": 2.493351697921753,
+      "learning_rate": 1.5647509578544062e-05,
+      "loss": 0.559,
+      "step": 2575
+    },
+    {
+      "epoch": 5.977011494252873,
+      "grad_norm": 69.57341766357422,
+      "learning_rate": 1.5583652618135378e-05,
+      "loss": 0.4154,
+      "step": 2600
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9541047188106012,
+      "eval_loss": 0.2397681325674057,
+      "eval_macro_f1": 0.8600498112631202,
+      "eval_runtime": 1.2855,
+      "eval_samples_per_second": 1203.442,
+      "eval_steps_per_second": 75.458,
+      "eval_weighted_f1": 0.9516447037124957,
+      "step": 2610
+    },
+    {
+      "epoch": 6.0344827586206895,
+      "grad_norm": 0.20201376080513,
+      "learning_rate": 1.5519795657726694e-05,
+      "loss": 0.3892,
+      "step": 2625
+    },
+    {
+      "epoch": 6.091954022988506,
+      "grad_norm": 0.17590577900409698,
+      "learning_rate": 1.545593869731801e-05,
+      "loss": 0.2642,
+      "step": 2650
+    },
+    {
+      "epoch": 6.149425287356322,
+      "grad_norm": 14.469496726989746,
+      "learning_rate": 1.5392081736909322e-05,
+      "loss": 0.2353,
+      "step": 2675
+    },
+    {
+      "epoch": 6.206896551724138,
+      "grad_norm": 81.51103973388672,
+      "learning_rate": 1.5328224776500638e-05,
+      "loss": 0.4785,
+      "step": 2700
+    },
+    {
+      "epoch": 6.264367816091954,
+      "grad_norm": 0.19920621812343597,
+      "learning_rate": 1.5264367816091957e-05,
+      "loss": 0.2868,
+      "step": 2725
+    },
+    {
+      "epoch": 6.32183908045977,
+      "grad_norm": 12.891411781311035,
+      "learning_rate": 1.520051085568327e-05,
+      "loss": 0.2671,
+      "step": 2750
+    },
+    {
+      "epoch": 6.379310344827586,
+      "grad_norm": 16.071155548095703,
+      "learning_rate": 1.5136653895274587e-05,
+      "loss": 0.5203,
+      "step": 2775
+    },
+    {
+      "epoch": 6.436781609195402,
+      "grad_norm": 0.24430899322032928,
+      "learning_rate": 1.5072796934865901e-05,
+      "loss": 0.2203,
+      "step": 2800
+    },
+    {
+      "epoch": 6.494252873563219,
+      "grad_norm": 0.21658585965633392,
+      "learning_rate": 1.5008939974457217e-05,
+      "loss": 0.3588,
+      "step": 2825
+    },
+    {
+      "epoch": 6.551724137931035,
+      "grad_norm": 0.218729630112648,
+      "learning_rate": 1.4945083014048531e-05,
+      "loss": 0.3032,
+      "step": 2850
+    },
+    {
+      "epoch": 6.609195402298851,
+      "grad_norm": 0.19259417057037354,
+      "learning_rate": 1.4881226053639847e-05,
+      "loss": 0.1498,
+      "step": 2875
+    },
+    {
+      "epoch": 6.666666666666667,
+      "grad_norm": 12.594219207763672,
+      "learning_rate": 1.4817369093231165e-05,
+      "loss": 0.5802,
+      "step": 2900
+    },
+    {
+      "epoch": 6.724137931034483,
+      "grad_norm": 0.2996440529823303,
+      "learning_rate": 1.4753512132822479e-05,
+      "loss": 0.2958,
+      "step": 2925
+    },
+    {
+      "epoch": 6.781609195402299,
+      "grad_norm": 0.3330232799053192,
+      "learning_rate": 1.4689655172413795e-05,
+      "loss": 0.3348,
+      "step": 2950
+    },
+    {
+      "epoch": 6.8390804597701145,
+      "grad_norm": 55.417236328125,
+      "learning_rate": 1.4625798212005109e-05,
+      "loss": 0.4195,
+      "step": 2975
+    },
+    {
+      "epoch": 6.896551724137931,
+      "grad_norm": 2.850304365158081,
+      "learning_rate": 1.4561941251596425e-05,
+      "loss": 0.4105,
+      "step": 3000
+    },
+    {
+      "epoch": 6.954022988505747,
+      "grad_norm": 0.2381744235754013,
+      "learning_rate": 1.4498084291187742e-05,
+      "loss": 0.3145,
+      "step": 3025
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9495798319327731,
+      "eval_loss": 0.2417064607143402,
+      "eval_macro_f1": 0.8506231029028258,
+      "eval_runtime": 1.2843,
+      "eval_samples_per_second": 1204.53,
+      "eval_steps_per_second": 75.526,
+      "eval_weighted_f1": 0.9484848377448473,
+      "step": 3045
+    },
+    {
+      "epoch": 7.011494252873563,
+      "grad_norm": 0.9536013603210449,
+      "learning_rate": 1.4434227330779056e-05,
+      "loss": 0.4517,
+      "step": 3050
+    },
+    {
+      "epoch": 7.068965517241379,
+      "grad_norm": 0.2184518724679947,
+      "learning_rate": 1.4370370370370372e-05,
+      "loss": 0.3963,
+      "step": 3075
+    },
+    {
+      "epoch": 7.126436781609195,
+      "grad_norm": 0.23555926978588104,
+      "learning_rate": 1.4306513409961686e-05,
+      "loss": 0.1847,
+      "step": 3100
+    },
+    {
+      "epoch": 7.183908045977011,
+      "grad_norm": 0.1601639837026596,
+      "learning_rate": 1.4242656449553002e-05,
+      "loss": 0.0835,
+      "step": 3125
+    },
+    {
+      "epoch": 7.241379310344827,
+      "grad_norm": 2.4888522624969482,
+      "learning_rate": 1.4178799489144316e-05,
+      "loss": 0.436,
+      "step": 3150
+    },
+    {
+      "epoch": 7.2988505747126435,
+      "grad_norm": 13.896317481994629,
+      "learning_rate": 1.4114942528735634e-05,
+      "loss": 0.4248,
+      "step": 3175
+    },
+    {
+      "epoch": 7.35632183908046,
+      "grad_norm": 0.2253255695104599,
+      "learning_rate": 1.405108556832695e-05,
+      "loss": 0.1896,
+      "step": 3200
+    },
+    {
+      "epoch": 7.413793103448276,
+      "grad_norm": 0.5111873149871826,
+      "learning_rate": 1.3987228607918264e-05,
+      "loss": 0.3082,
+      "step": 3225
+    },
+    {
+      "epoch": 7.471264367816092,
+      "grad_norm": 0.3114178478717804,
+      "learning_rate": 1.392337164750958e-05,
+      "loss": 0.354,
+      "step": 3250
+    },
+    {
+      "epoch": 7.528735632183908,
+      "grad_norm": 308.2452392578125,
+      "learning_rate": 1.3859514687100894e-05,
+      "loss": 0.4526,
+      "step": 3275
+    },
+    {
+      "epoch": 7.586206896551724,
+      "grad_norm": 0.30504605174064636,
+      "learning_rate": 1.3795657726692211e-05,
+      "loss": 0.1799,
+      "step": 3300
+    },
+    {
+      "epoch": 7.64367816091954,
+      "grad_norm": 0.23172464966773987,
+      "learning_rate": 1.3731800766283526e-05,
+      "loss": 0.2987,
+      "step": 3325
+    },
+    {
+      "epoch": 7.7011494252873565,
+      "grad_norm": 118.74394226074219,
+      "learning_rate": 1.3667943805874841e-05,
+      "loss": 0.2358,
+      "step": 3350
+    },
+    {
+      "epoch": 7.758620689655173,
+      "grad_norm": 0.14811089634895325,
+      "learning_rate": 1.3604086845466157e-05,
+      "loss": 0.3334,
+      "step": 3375
+    },
+    {
+      "epoch": 7.816091954022989,
+      "grad_norm": 0.26182156801223755,
+      "learning_rate": 1.3540229885057471e-05,
+      "loss": 0.3044,
+      "step": 3400
+    },
+    {
+      "epoch": 7.873563218390805,
+      "grad_norm": 0.42604881525039673,
+      "learning_rate": 1.3476372924648789e-05,
+      "loss": 0.3334,
+      "step": 3425
+    },
+    {
+      "epoch": 7.931034482758621,
+      "grad_norm": 3.8690788745880127,
+      "learning_rate": 1.3412515964240103e-05,
+      "loss": 0.2318,
+      "step": 3450
+    },
+    {
+      "epoch": 7.988505747126437,
+      "grad_norm": 5.706087112426758,
+      "learning_rate": 1.3348659003831419e-05,
+      "loss": 0.2909,
+      "step": 3475
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9515190691661279,
+      "eval_loss": 0.2455778568983078,
+      "eval_macro_f1": 0.853849899937441,
+      "eval_runtime": 1.2822,
+      "eval_samples_per_second": 1206.499,
+      "eval_steps_per_second": 75.65,
+      "eval_weighted_f1": 0.9495985858714602,
+      "step": 3480
+    },
+    {
+      "epoch": 8.045977011494253,
+      "grad_norm": 1.9139503240585327,
+      "learning_rate": 1.3284802043422733e-05,
+      "loss": 0.1426,
+      "step": 3500
+    },
+    {
+      "epoch": 8.10344827586207,
+      "grad_norm": 0.10656487196683884,
+      "learning_rate": 1.3220945083014049e-05,
+      "loss": 0.1865,
+      "step": 3525
+    },
+    {
+      "epoch": 8.160919540229886,
+      "grad_norm": 0.9113854765892029,
+      "learning_rate": 1.3157088122605367e-05,
+      "loss": 0.2629,
+      "step": 3550
+    },
+    {
+      "epoch": 8.218390804597702,
+      "grad_norm": 12.390186309814453,
+      "learning_rate": 1.309323116219668e-05,
+      "loss": 0.3263,
+      "step": 3575
+    },
+    {
+      "epoch": 8.275862068965518,
+      "grad_norm": 16.489566802978516,
+      "learning_rate": 1.3029374201787996e-05,
+      "loss": 0.2176,
+      "step": 3600
+    },
+    {
+      "epoch": 8.333333333333334,
+      "grad_norm": 0.15907344222068787,
+      "learning_rate": 1.296551724137931e-05,
+      "loss": 0.2475,
+      "step": 3625
+    },
+    {
+      "epoch": 8.39080459770115,
+      "grad_norm": 0.163466215133667,
+      "learning_rate": 1.2901660280970626e-05,
+      "loss": 0.1819,
+      "step": 3650
+    },
+    {
+      "epoch": 8.448275862068966,
+      "grad_norm": 83.0460205078125,
+      "learning_rate": 1.283780332056194e-05,
+      "loss": 0.2247,
+      "step": 3675
+    },
+    {
+      "epoch": 8.505747126436782,
+      "grad_norm": 0.15427705645561218,
+      "learning_rate": 1.2773946360153258e-05,
+      "loss": 0.2735,
+      "step": 3700
+    },
+    {
+      "epoch": 8.563218390804598,
+      "grad_norm": 0.1480730175971985,
+      "learning_rate": 1.2710089399744574e-05,
+      "loss": 0.2022,
+      "step": 3725
+    },
+    {
+      "epoch": 8.620689655172415,
+      "grad_norm": 0.3764588236808777,
+      "learning_rate": 1.2646232439335888e-05,
+      "loss": 0.3027,
+      "step": 3750
+    },
+    {
+      "epoch": 8.678160919540229,
+      "grad_norm": 0.17370322346687317,
+      "learning_rate": 1.2582375478927204e-05,
+      "loss": 0.3577,
+      "step": 3775
+    },
+    {
+      "epoch": 8.735632183908045,
+      "grad_norm": 24.64053726196289,
+      "learning_rate": 1.2518518518518518e-05,
+      "loss": 0.2364,
+      "step": 3800
+    },
+    {
+      "epoch": 8.793103448275861,
+      "grad_norm": 5.574195861816406,
+      "learning_rate": 1.2454661558109836e-05,
+      "loss": 0.2676,
+      "step": 3825
+    },
+    {
+      "epoch": 8.850574712643677,
+      "grad_norm": 0.09678659588098526,
+      "learning_rate": 1.2390804597701151e-05,
+      "loss": 0.2759,
+      "step": 3850
+    },
+    {
+      "epoch": 8.908045977011493,
+      "grad_norm": 0.1522231101989746,
+      "learning_rate": 1.2326947637292466e-05,
+      "loss": 0.2181,
+      "step": 3875
+    },
+    {
+      "epoch": 8.96551724137931,
+      "grad_norm": 0.37765446305274963,
+      "learning_rate": 1.2263090676883781e-05,
+      "loss": 0.4644,
+      "step": 3900
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.26898178458213806,
+      "eval_macro_f1": 0.8663256210777838,
+      "eval_runtime": 1.2494,
+      "eval_samples_per_second": 1238.187,
+      "eval_steps_per_second": 77.637,
+      "eval_weighted_f1": 0.9512561247191414,
+      "step": 3915
+    },
+    {
+      "epoch": 9.022988505747126,
+      "grad_norm": 1.2754137516021729,
+      "learning_rate": 1.2199233716475096e-05,
+      "loss": 0.2067,
+      "step": 3925
+    },
+    {
+      "epoch": 9.080459770114942,
+      "grad_norm": 2.7214295864105225,
+      "learning_rate": 1.2135376756066413e-05,
+      "loss": 0.1737,
+      "step": 3950
+    },
+    {
+      "epoch": 9.137931034482758,
+      "grad_norm": 10.367477416992188,
+      "learning_rate": 1.2071519795657727e-05,
+      "loss": 0.3776,
+      "step": 3975
+    },
+    {
+      "epoch": 9.195402298850574,
+      "grad_norm": 0.20507702231407166,
+      "learning_rate": 1.2007662835249043e-05,
+      "loss": 0.1399,
+      "step": 4000
+    },
+    {
+      "epoch": 9.25287356321839,
+      "grad_norm": 3.616419553756714,
+      "learning_rate": 1.1943805874840359e-05,
+      "loss": 0.2167,
+      "step": 4025
+    },
+    {
+      "epoch": 9.310344827586206,
+      "grad_norm": 0.13054685294628143,
+      "learning_rate": 1.1879948914431673e-05,
+      "loss": 0.1431,
+      "step": 4050
+    },
+    {
+      "epoch": 9.367816091954023,
+      "grad_norm": 0.11281740665435791,
+      "learning_rate": 1.181609195402299e-05,
+      "loss": 0.2202,
+      "step": 4075
+    },
+    {
+      "epoch": 9.425287356321839,
+      "grad_norm": 0.6165698170661926,
+      "learning_rate": 1.1752234993614305e-05,
+      "loss": 0.1859,
+      "step": 4100
+    },
+    {
+      "epoch": 9.482758620689655,
+      "grad_norm": 1.5480984449386597,
+      "learning_rate": 1.168837803320562e-05,
+      "loss": 0.2284,
+      "step": 4125
+    },
+    {
+      "epoch": 9.540229885057471,
+      "grad_norm": 6.850103855133057,
+      "learning_rate": 1.1624521072796935e-05,
+      "loss": 0.1698,
+      "step": 4150
+    },
+    {
+      "epoch": 9.597701149425287,
+      "grad_norm": 3.5420010089874268,
+      "learning_rate": 1.156066411238825e-05,
+      "loss": 0.3109,
+      "step": 4175
+    },
+    {
+      "epoch": 9.655172413793103,
+      "grad_norm": 0.1575000286102295,
+      "learning_rate": 1.1496807151979568e-05,
+      "loss": 0.1249,
+      "step": 4200
+    },
+    {
+      "epoch": 9.71264367816092,
+      "grad_norm": 86.61784362792969,
+      "learning_rate": 1.1432950191570882e-05,
+      "loss": 0.1512,
+      "step": 4225
+    },
+    {
+      "epoch": 9.770114942528735,
+      "grad_norm": 6.673755645751953,
+      "learning_rate": 1.1369093231162198e-05,
+      "loss": 0.2927,
+      "step": 4250
+    },
+    {
+      "epoch": 9.827586206896552,
+      "grad_norm": 3.7119579315185547,
+      "learning_rate": 1.1305236270753512e-05,
+      "loss": 0.1886,
+      "step": 4275
+    },
+    {
+      "epoch": 9.885057471264368,
+      "grad_norm": 7.661352157592773,
+      "learning_rate": 1.1241379310344828e-05,
+      "loss": 0.2061,
+      "step": 4300
+    },
+    {
+      "epoch": 9.942528735632184,
+      "grad_norm": 4.866251468658447,
+      "learning_rate": 1.1177522349936142e-05,
+      "loss": 0.2876,
+      "step": 4325
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 17.001205444335938,
+      "learning_rate": 1.111366538952746e-05,
+      "loss": 0.1963,
+      "step": 4350
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.9495798319327731,
+      "eval_loss": 0.2947230339050293,
+      "eval_macro_f1": 0.8586701439701274,
+      "eval_runtime": 1.2427,
+      "eval_samples_per_second": 1244.9,
+      "eval_steps_per_second": 78.058,
+      "eval_weighted_f1": 0.94867080753628,
+      "step": 4350
+    },
+    {
+      "epoch": 10.057471264367816,
+      "grad_norm": 0.09062144160270691,
+      "learning_rate": 1.1049808429118776e-05,
+      "loss": 0.0144,
+      "step": 4375
+    },
+    {
+      "epoch": 10.114942528735632,
+      "grad_norm": 0.09079045802354813,
+      "learning_rate": 1.098595146871009e-05,
+      "loss": 0.2467,
+      "step": 4400
+    },
+    {
+      "epoch": 10.172413793103448,
+      "grad_norm": 16.911344528198242,
+      "learning_rate": 1.0922094508301406e-05,
+      "loss": 0.1387,
+      "step": 4425
+    },
+    {
+      "epoch": 10.229885057471265,
+      "grad_norm": 0.12063690274953842,
+      "learning_rate": 1.085823754789272e-05,
+      "loss": 0.1926,
+      "step": 4450
+    },
+    {
+      "epoch": 10.28735632183908,
+      "grad_norm": 0.3631066679954529,
+      "learning_rate": 1.0794380587484037e-05,
+      "loss": 0.1826,
+      "step": 4475
+    },
+    {
+      "epoch": 10.344827586206897,
+      "grad_norm": 0.36074990034103394,
+      "learning_rate": 1.0730523627075351e-05,
+      "loss": 0.161,
+      "step": 4500
+    },
+    {
+      "epoch": 10.402298850574713,
+      "grad_norm": 55.926822662353516,
+      "learning_rate": 1.0666666666666667e-05,
+      "loss": 0.2874,
+      "step": 4525
+    },
+    {
+      "epoch": 10.459770114942529,
+      "grad_norm": 0.10639549791812897,
+      "learning_rate": 1.0602809706257983e-05,
+      "loss": 0.1727,
+      "step": 4550
+    },
+    {
+      "epoch": 10.517241379310345,
+      "grad_norm": 0.7862879633903503,
+      "learning_rate": 1.0538952745849297e-05,
+      "loss": 0.274,
+      "step": 4575
+    },
+    {
+      "epoch": 10.574712643678161,
+      "grad_norm": 0.27984291315078735,
+      "learning_rate": 1.0475095785440615e-05,
+      "loss": 0.2045,
+      "step": 4600
+    },
+    {
+      "epoch": 10.632183908045977,
+      "grad_norm": 17.181734085083008,
+      "learning_rate": 1.0411238825031929e-05,
+      "loss": 0.3352,
+      "step": 4625
+    },
+    {
+      "epoch": 10.689655172413794,
+      "grad_norm": 0.11635159701108932,
+      "learning_rate": 1.0347381864623245e-05,
+      "loss": 0.134,
+      "step": 4650
+    },
+    {
+      "epoch": 10.74712643678161,
+      "grad_norm": 0.07812677323818207,
+      "learning_rate": 1.028352490421456e-05,
+      "loss": 0.0612,
+      "step": 4675
+    },
+    {
+      "epoch": 10.804597701149426,
+      "grad_norm": 0.09490698575973511,
+      "learning_rate": 1.0219667943805875e-05,
+      "loss": 0.2156,
+      "step": 4700
+    },
+    {
+      "epoch": 10.862068965517242,
+      "grad_norm": 7.355884075164795,
+      "learning_rate": 1.0155810983397192e-05,
+      "loss": 0.2202,
+      "step": 4725
+    },
+    {
+      "epoch": 10.919540229885058,
+      "grad_norm": 5.701615333557129,
+      "learning_rate": 1.0091954022988506e-05,
+      "loss": 0.1911,
+      "step": 4750
+    },
+    {
+      "epoch": 10.977011494252874,
+      "grad_norm": 5.884024620056152,
+      "learning_rate": 1.0028097062579822e-05,
+      "loss": 0.1841,
+      "step": 4775
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.9521654815772462,
+      "eval_loss": 0.3065274953842163,
+      "eval_macro_f1": 0.8665082853654282,
+      "eval_runtime": 1.3338,
+      "eval_samples_per_second": 1159.816,
+      "eval_steps_per_second": 72.723,
+      "eval_weighted_f1": 0.9509302446253226,
+      "step": 4785
+    },
+    {
+      "epoch": 11.03448275862069,
+      "grad_norm": 0.10210714489221573,
+      "learning_rate": 9.964240102171138e-06,
+      "loss": 0.1149,
+      "step": 4800
+    },
+    {
+      "epoch": 11.091954022988507,
+      "grad_norm": 0.07134879380464554,
+      "learning_rate": 9.900383141762452e-06,
+      "loss": 0.1125,
+      "step": 4825
+    },
+    {
+      "epoch": 11.149425287356323,
+      "grad_norm": 0.6779069900512695,
+      "learning_rate": 9.836526181353768e-06,
+      "loss": 0.134,
+      "step": 4850
+    },
+    {
+      "epoch": 11.206896551724139,
+      "grad_norm": 0.09422433376312256,
+      "learning_rate": 9.772669220945084e-06,
+      "loss": 0.233,
+      "step": 4875
+    },
+    {
+      "epoch": 11.264367816091955,
+      "grad_norm": 0.07436396181583405,
+      "learning_rate": 9.708812260536398e-06,
+      "loss": 0.1367,
+      "step": 4900
+    },
+    {
+      "epoch": 11.32183908045977,
+      "grad_norm": 15.747620582580566,
+      "learning_rate": 9.644955300127716e-06,
+      "loss": 0.1444,
+      "step": 4925
+    },
+    {
+      "epoch": 11.379310344827585,
+      "grad_norm": 0.3826606571674347,
+      "learning_rate": 9.58109833971903e-06,
+      "loss": 0.1271,
+      "step": 4950
+    },
+    {
+      "epoch": 11.436781609195402,
+      "grad_norm": 0.0642637312412262,
+      "learning_rate": 9.517241379310346e-06,
+      "loss": 0.1503,
+      "step": 4975
+    },
+    {
+      "epoch": 11.494252873563218,
+      "grad_norm": 0.16133825480937958,
+      "learning_rate": 9.453384418901661e-06,
+      "loss": 0.1837,
+      "step": 5000
+    },
+    {
+      "epoch": 11.551724137931034,
+      "grad_norm": 0.15818732976913452,
+      "learning_rate": 9.389527458492976e-06,
+      "loss": 0.0784,
+      "step": 5025
+    },
+    {
+      "epoch": 11.60919540229885,
+      "grad_norm": 0.06489365547895432,
+      "learning_rate": 9.325670498084291e-06,
+      "loss": 0.1532,
+      "step": 5050
+    },
+    {
+      "epoch": 11.666666666666666,
+      "grad_norm": 0.05736462399363518,
+      "learning_rate": 9.261813537675607e-06,
+      "loss": 0.1367,
+      "step": 5075
+    },
+    {
+      "epoch": 11.724137931034482,
+      "grad_norm": 0.9656385183334351,
+      "learning_rate": 9.197956577266923e-06,
+      "loss": 0.0903,
+      "step": 5100
+    },
+    {
+      "epoch": 11.781609195402298,
+      "grad_norm": 0.06174295395612717,
+      "learning_rate": 9.134099616858239e-06,
+      "loss": 0.0899,
+      "step": 5125
+    },
+    {
+      "epoch": 11.839080459770114,
+      "grad_norm": 0.060515888035297394,
+      "learning_rate": 9.070242656449553e-06,
+      "loss": 0.1605,
+      "step": 5150
+    },
+    {
+      "epoch": 11.89655172413793,
+      "grad_norm": 3.496469497680664,
+      "learning_rate": 9.006385696040869e-06,
+      "loss": 0.1872,
+      "step": 5175
+    },
+    {
+      "epoch": 11.954022988505747,
+      "grad_norm": 0.07404470443725586,
+      "learning_rate": 8.942528735632185e-06,
+      "loss": 0.1987,
+      "step": 5200
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.9508726567550096,
+      "eval_loss": 0.2994757890701294,
+      "eval_macro_f1": 0.8675740894345955,
+      "eval_runtime": 1.2571,
+      "eval_samples_per_second": 1230.643,
+      "eval_steps_per_second": 77.164,
+      "eval_weighted_f1": 0.9504132974662869,
+      "step": 5220
+    },
+    {
+      "epoch": 12.011494252873563,
+      "grad_norm": 20.24570083618164,
+      "learning_rate": 8.878671775223499e-06,
+      "loss": 0.1455,
+      "step": 5225
+    },
+    {
+      "epoch": 12.068965517241379,
+      "grad_norm": 0.06321834027767181,
+      "learning_rate": 8.814814814814817e-06,
+      "loss": 0.0736,
+      "step": 5250
+    },
+    {
+      "epoch": 12.126436781609195,
+      "grad_norm": 0.05353408306837082,
+      "learning_rate": 8.75095785440613e-06,
+      "loss": 0.0637,
+      "step": 5275
+    },
+    {
+      "epoch": 12.183908045977011,
+      "grad_norm": 0.053601209074258804,
+      "learning_rate": 8.687100893997446e-06,
+      "loss": 0.0976,
+      "step": 5300
+    },
+    {
+      "epoch": 12.241379310344827,
+      "grad_norm": 0.05540817975997925,
+      "learning_rate": 8.623243933588762e-06,
+      "loss": 0.0982,
+      "step": 5325
+    },
+    {
+      "epoch": 12.298850574712644,
+      "grad_norm": 25.066635131835938,
+      "learning_rate": 8.559386973180076e-06,
+      "loss": 0.2073,
+      "step": 5350
+    },
+    {
+      "epoch": 12.35632183908046,
+      "grad_norm": 0.21873502433300018,
+      "learning_rate": 8.495530012771392e-06,
+      "loss": 0.1598,
+      "step": 5375
+    },
+    {
+      "epoch": 12.413793103448276,
+      "grad_norm": 0.9713755249977112,
+      "learning_rate": 8.431673052362708e-06,
+      "loss": 0.061,
+      "step": 5400
+    },
+    {
+      "epoch": 12.471264367816092,
+      "grad_norm": 0.14838582277297974,
+      "learning_rate": 8.367816091954024e-06,
+      "loss": 0.1455,
+      "step": 5425
+    },
+    {
+      "epoch": 12.528735632183908,
+      "grad_norm": 0.07204707711935043,
+      "learning_rate": 8.30395913154534e-06,
+      "loss": 0.0782,
+      "step": 5450
+    },
+    {
+      "epoch": 12.586206896551724,
+      "grad_norm": 0.30839934945106506,
+      "learning_rate": 8.240102171136654e-06,
+      "loss": 0.1285,
+      "step": 5475
+    },
+    {
+      "epoch": 12.64367816091954,
+      "grad_norm": 243.3436279296875,
+      "learning_rate": 8.17624521072797e-06,
+      "loss": 0.1549,
+      "step": 5500
+    },
+    {
+      "epoch": 12.701149425287356,
+      "grad_norm": 1.1939635276794434,
+      "learning_rate": 8.112388250319286e-06,
+      "loss": 0.1488,
+      "step": 5525
+    },
+    {
+      "epoch": 12.758620689655173,
+      "grad_norm": 20.439424514770508,
+      "learning_rate": 8.0485312899106e-06,
+      "loss": 0.1271,
+      "step": 5550
+    },
+    {
+      "epoch": 12.816091954022989,
+      "grad_norm": 0.07725687325000763,
+      "learning_rate": 7.984674329501917e-06,
+      "loss": 0.1875,
+      "step": 5575
+    },
+    {
+      "epoch": 12.873563218390805,
+      "grad_norm": 0.04689912497997284,
+      "learning_rate": 7.920817369093231e-06,
+      "loss": 0.1545,
+      "step": 5600
+    },
+    {
+      "epoch": 12.931034482758621,
+      "grad_norm": 2.4595677852630615,
+      "learning_rate": 7.856960408684547e-06,
+      "loss": 0.0675,
+      "step": 5625
+    },
+    {
+      "epoch": 12.988505747126437,
+      "grad_norm": 0.042330410331487656,
+      "learning_rate": 7.793103448275863e-06,
+      "loss": 0.1067,
+      "step": 5650
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": 0.9508726567550096,
+      "eval_loss": 0.31408828496932983,
+      "eval_macro_f1": 0.8589171045025167,
+      "eval_runtime": 1.4008,
+      "eval_samples_per_second": 1104.359,
+      "eval_steps_per_second": 69.246,
+      "eval_weighted_f1": 0.9493507314299067,
+      "step": 5655
+    },
+    {
+      "epoch": 13.045977011494253,
+      "grad_norm": 0.06908591091632843,
+      "learning_rate": 7.729246487867177e-06,
+      "loss": 0.0377,
+      "step": 5675
+    },
+    {
+      "epoch": 13.10344827586207,
+      "grad_norm": 0.049803633242845535,
+      "learning_rate": 7.665389527458493e-06,
+      "loss": 0.1461,
+      "step": 5700
+    },
+    {
+      "epoch": 13.160919540229886,
+      "grad_norm": 185.27767944335938,
+      "learning_rate": 7.601532567049809e-06,
+      "loss": 0.0542,
+      "step": 5725
+    },
+    {
+      "epoch": 13.218390804597702,
+      "grad_norm": 0.04537845775485039,
+      "learning_rate": 7.537675606641125e-06,
+      "loss": 0.0976,
+      "step": 5750
+    },
+    {
+      "epoch": 13.275862068965518,
+      "grad_norm": 0.054079122841358185,
+      "learning_rate": 7.473818646232441e-06,
+      "loss": 0.0621,
+      "step": 5775
+    },
+    {
+      "epoch": 13.333333333333334,
+      "grad_norm": 0.061256758868694305,
+      "learning_rate": 7.409961685823756e-06,
+      "loss": 0.1405,
+      "step": 5800
+    },
+    {
+      "epoch": 13.39080459770115,
+      "grad_norm": 0.0523013100028038,
+      "learning_rate": 7.346104725415071e-06,
+      "loss": 0.1024,
+      "step": 5825
+    },
+    {
+      "epoch": 13.448275862068966,
+      "grad_norm": 72.62133026123047,
+      "learning_rate": 7.2822477650063865e-06,
+      "loss": 0.0354,
+      "step": 5850
+    },
+    {
+      "epoch": 13.505747126436782,
+      "grad_norm": 0.04726962000131607,
+      "learning_rate": 7.2183908045977015e-06,
+      "loss": 0.0762,
+      "step": 5875
+    },
+    {
+      "epoch": 13.563218390804598,
+      "grad_norm": 48.139495849609375,
+      "learning_rate": 7.1545338441890165e-06,
+      "loss": 0.1211,
+      "step": 5900
+    },
+    {
+      "epoch": 13.620689655172415,
+      "grad_norm": 0.045234814286231995,
+      "learning_rate": 7.090676883780333e-06,
+      "loss": 0.1277,
+      "step": 5925
+    },
+    {
+      "epoch": 13.678160919540229,
+      "grad_norm": 32.473392486572266,
+      "learning_rate": 7.026819923371648e-06,
+      "loss": 0.0727,
+      "step": 5950
+    },
+    {
+      "epoch": 13.735632183908045,
+      "grad_norm": 0.038368646055459976,
+      "learning_rate": 6.962962962962964e-06,
+      "loss": 0.1127,
+      "step": 5975
+    },
+    {
+      "epoch": 13.793103448275861,
+      "grad_norm": 0.35299959778785706,
+      "learning_rate": 6.899106002554279e-06,
+      "loss": 0.0471,
+      "step": 6000
+    },
+    {
+      "epoch": 13.850574712643677,
+      "grad_norm": 120.72334289550781,
+      "learning_rate": 6.835249042145594e-06,
+      "loss": 0.0659,
+      "step": 6025
+    },
+    {
+      "epoch": 13.908045977011493,
+      "grad_norm": 0.03927996754646301,
+      "learning_rate": 6.77139208173691e-06,
+      "loss": 0.0072,
+      "step": 6050
+    },
+    {
+      "epoch": 13.96551724137931,
+      "grad_norm": 0.03977316617965698,
+      "learning_rate": 6.707535121328226e-06,
+      "loss": 0.0747,
+      "step": 6075
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.9541047188106012,
+      "eval_loss": 0.3180047571659088,
+      "eval_macro_f1": 0.8659646333711333,
+      "eval_runtime": 1.3168,
+      "eval_samples_per_second": 1174.822,
+      "eval_steps_per_second": 73.664,
+      "eval_weighted_f1": 0.952362261289871,
+      "step": 6090
+    },
+    {
+      "epoch": 14.022988505747126,
+      "grad_norm": 0.034806590527296066,
+      "learning_rate": 6.6436781609195415e-06,
+      "loss": 0.0055,
+      "step": 6100
+    },
+    {
+      "epoch": 14.080459770114942,
+      "grad_norm": 0.11559601128101349,
+      "learning_rate": 6.5798212005108565e-06,
+      "loss": 0.0428,
+      "step": 6125
+    },
+    {
+      "epoch": 14.137931034482758,
+      "grad_norm": 0.03521871566772461,
+      "learning_rate": 6.5159642401021715e-06,
+      "loss": 0.0787,
+      "step": 6150
+    },
+    {
+      "epoch": 14.195402298850574,
+      "grad_norm": 0.22859551012516022,
+      "learning_rate": 6.452107279693487e-06,
+      "loss": 0.0643,
+      "step": 6175
+    },
+    {
+      "epoch": 14.25287356321839,
+      "grad_norm": 40.91361618041992,
+      "learning_rate": 6.388250319284802e-06,
+      "loss": 0.0426,
+      "step": 6200
+    },
+    {
+      "epoch": 14.310344827586206,
+      "grad_norm": 0.04818647727370262,
+      "learning_rate": 6.324393358876117e-06,
+      "loss": 0.0694,
+      "step": 6225
+    },
+    {
+      "epoch": 14.367816091954023,
+      "grad_norm": 0.041468918323516846,
+      "learning_rate": 6.260536398467434e-06,
+      "loss": 0.1062,
+      "step": 6250
+    },
+    {
+      "epoch": 14.425287356321839,
+      "grad_norm": 59.16422653198242,
+      "learning_rate": 6.196679438058749e-06,
+      "loss": 0.0675,
+      "step": 6275
+    },
+    {
+      "epoch": 14.482758620689655,
+      "grad_norm": 0.0676654800772667,
+      "learning_rate": 6.132822477650065e-06,
+      "loss": 0.0925,
+      "step": 6300
+    },
+    {
+      "epoch": 14.540229885057471,
+      "grad_norm": 0.03517609089612961,
+      "learning_rate": 6.06896551724138e-06,
+      "loss": 0.0093,
+      "step": 6325
+    },
+    {
+      "epoch": 14.597701149425287,
+      "grad_norm": 0.08453945815563202,
+      "learning_rate": 6.005108556832695e-06,
+      "loss": 0.0449,
+      "step": 6350
+    },
+    {
+      "epoch": 14.655172413793103,
+      "grad_norm": 402.1295166015625,
+      "learning_rate": 5.941251596424011e-06,
+      "loss": 0.0741,
+      "step": 6375
+    },
+    {
+      "epoch": 14.71264367816092,
+      "grad_norm": 1.8468683958053589,
+      "learning_rate": 5.8773946360153265e-06,
+      "loss": 0.0987,
+      "step": 6400
+    },
+    {
+      "epoch": 14.770114942528735,
+      "grad_norm": 65.16915130615234,
+      "learning_rate": 5.813537675606642e-06,
+      "loss": 0.0783,
+      "step": 6425
+    },
+    {
+      "epoch": 14.827586206896552,
+      "grad_norm": 0.031167171895503998,
+      "learning_rate": 5.749680715197957e-06,
+      "loss": 0.0945,
+      "step": 6450
+    },
+    {
+      "epoch": 14.885057471264368,
+      "grad_norm": 0.03129557892680168,
+      "learning_rate": 5.685823754789272e-06,
+      "loss": 0.0046,
+      "step": 6475
+    },
+    {
+      "epoch": 14.942528735632184,
+      "grad_norm": 0.031040074303746223,
+      "learning_rate": 5.621966794380588e-06,
+      "loss": 0.0735,
+      "step": 6500
+    },
+    {
+      "epoch": 15.0,
+      "grad_norm": 0.03432578593492508,
+      "learning_rate": 5.558109833971903e-06,
+      "loss": 0.1317,
+      "step": 6525
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.34154069423675537,
+      "eval_macro_f1": 0.8576239462902868,
+      "eval_runtime": 1.2511,
+      "eval_samples_per_second": 1236.469,
+      "eval_steps_per_second": 77.529,
+      "eval_weighted_f1": 0.9494076245238235,
+      "step": 6525
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 8700,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 3
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

ood-uit_vsfc-phobert-topic-domain-00-s42-20260811T150238Z-54bb115e/trainer/checkpoint-6525/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f0712802a8bbb778f406abcda3b020380ef5f43ee0e78335abc61a86f8ff84d
+size 5624

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:403cad5e777e6748a19db900f7b6dafbc1d25830d80d1ef209884e5db3289846
+size 540026404

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<mask>": 64000
+}

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ece1475d67efb2c0a48dc33b49970068dc45e5ed78b9ece6e89208fcca7fb5a9
+size 540026436

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:810b8f965a980596ac562ab44f6076506df789dc5c5503afff71d8d5b8520481
+size 1075446330

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2c3ddf936b099cdc668222de40fb42e5c06ea4095eea22a68a22387cb769c0d
+size 14244

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77f2d1d2a465dab7ef78796e23483ed57a47b031a4f3ed5e697f94c08f289dfe
+size 1064

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64000": {
+      "content": "<mask>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "mask_token": "<mask>",
+  "model_max_length": 256,
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "tokenizer_class": "PhobertTokenizer",
+  "unk_token": "<unk>"
+}

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/trainer_state.json

@@ -0,0 +1,1104 @@
+{
+  "best_global_step": 3480,
+  "best_metric": 0.8505330085886426,
+  "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480",
+  "epoch": 8.0,
+  "eval_steps": 500,
+  "global_step": 3480,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05747126436781609,
+      "grad_norm": 12.892321586608887,
+      "learning_rate": 5.517241379310345e-07,
+      "loss": 4.2257,
+      "step": 25
+    },
+    {
+      "epoch": 0.11494252873563218,
+      "grad_norm": 147.58221435546875,
+      "learning_rate": 1.1264367816091955e-06,
+      "loss": 4.04,
+      "step": 50
+    },
+    {
+      "epoch": 0.1724137931034483,
+      "grad_norm": 10.12353801727295,
+      "learning_rate": 1.7011494252873565e-06,
+      "loss": 3.7585,
+      "step": 75
+    },
+    {
+      "epoch": 0.22988505747126436,
+      "grad_norm": 9.198833465576172,
+      "learning_rate": 2.2758620689655173e-06,
+      "loss": 3.406,
+      "step": 100
+    },
+    {
+      "epoch": 0.28735632183908044,
+      "grad_norm": 12.207880973815918,
+      "learning_rate": 2.850574712643678e-06,
+      "loss": 3.0539,
+      "step": 125
+    },
+    {
+      "epoch": 0.3448275862068966,
+      "grad_norm": 11.848393440246582,
+      "learning_rate": 3.425287356321839e-06,
+      "loss": 2.4264,
+      "step": 150
+    },
+    {
+      "epoch": 0.40229885057471265,
+      "grad_norm": 8.870543479919434,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.9691,
+      "step": 175
+    },
+    {
+      "epoch": 0.45977011494252873,
+      "grad_norm": 19.7670955657959,
+      "learning_rate": 4.5747126436781615e-06,
+      "loss": 1.6845,
+      "step": 200
+    },
+    {
+      "epoch": 0.5172413793103449,
+      "grad_norm": 16.912090301513672,
+      "learning_rate": 5.149425287356322e-06,
+      "loss": 1.4159,
+      "step": 225
+    },
+    {
+      "epoch": 0.5747126436781609,
+      "grad_norm": 21.332857131958008,
+      "learning_rate": 5.724137931034483e-06,
+      "loss": 1.2909,
+      "step": 250
+    },
+    {
+      "epoch": 0.632183908045977,
+      "grad_norm": 13.889664649963379,
+      "learning_rate": 6.298850574712644e-06,
+      "loss": 1.3003,
+      "step": 275
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 6.493246078491211,
+      "learning_rate": 6.873563218390805e-06,
+      "loss": 1.2024,
+      "step": 300
+    },
+    {
+      "epoch": 0.7471264367816092,
+      "grad_norm": 19.06503677368164,
+      "learning_rate": 7.4482758620689665e-06,
+      "loss": 0.9701,
+      "step": 325
+    },
+    {
+      "epoch": 0.8045977011494253,
+      "grad_norm": 48.411109924316406,
+      "learning_rate": 8.022988505747126e-06,
+      "loss": 1.0011,
+      "step": 350
+    },
+    {
+      "epoch": 0.8620689655172413,
+      "grad_norm": 24.054780960083008,
+      "learning_rate": 8.597701149425289e-06,
+      "loss": 1.1521,
+      "step": 375
+    },
+    {
+      "epoch": 0.9195402298850575,
+      "grad_norm": 2.5997345447540283,
+      "learning_rate": 9.172413793103448e-06,
+      "loss": 0.8185,
+      "step": 400
+    },
+    {
+      "epoch": 0.9770114942528736,
+      "grad_norm": 8.2587251663208,
+      "learning_rate": 9.74712643678161e-06,
+      "loss": 0.9219,
+      "step": 425
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.940530058177117,
+      "eval_loss": 0.20286934077739716,
+      "eval_macro_f1": 0.8090601999154878,
+      "eval_runtime": 1.2107,
+      "eval_samples_per_second": 1277.73,
+      "eval_steps_per_second": 80.116,
+      "eval_weighted_f1": 0.9374981989040525,
+      "step": 435
+    },
+    {
+      "epoch": 1.0344827586206897,
+      "grad_norm": 74.17597961425781,
+      "learning_rate": 1.032183908045977e-05,
+      "loss": 0.9663,
+      "step": 450
+    },
+    {
+      "epoch": 1.0919540229885056,
+      "grad_norm": 5.443840980529785,
+      "learning_rate": 1.0896551724137932e-05,
+      "loss": 0.8417,
+      "step": 475
+    },
+    {
+      "epoch": 1.1494252873563218,
+      "grad_norm": 14.226733207702637,
+      "learning_rate": 1.1471264367816093e-05,
+      "loss": 0.9687,
+      "step": 500
+    },
+    {
+      "epoch": 1.206896551724138,
+      "grad_norm": 15.986963272094727,
+      "learning_rate": 1.2045977011494254e-05,
+      "loss": 0.7588,
+      "step": 525
+    },
+    {
+      "epoch": 1.264367816091954,
+      "grad_norm": 8.215805053710938,
+      "learning_rate": 1.2620689655172415e-05,
+      "loss": 0.9942,
+      "step": 550
+    },
+    {
+      "epoch": 1.3218390804597702,
+      "grad_norm": 12.830760955810547,
+      "learning_rate": 1.3195402298850574e-05,
+      "loss": 0.8837,
+      "step": 575
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "grad_norm": 3.7767229080200195,
+      "learning_rate": 1.3770114942528736e-05,
+      "loss": 0.7615,
+      "step": 600
+    },
+    {
+      "epoch": 1.4367816091954024,
+      "grad_norm": 12.500882148742676,
+      "learning_rate": 1.4344827586206897e-05,
+      "loss": 0.8383,
+      "step": 625
+    },
+    {
+      "epoch": 1.4942528735632183,
+      "grad_norm": 20.17972183227539,
+      "learning_rate": 1.4919540229885058e-05,
+      "loss": 0.8934,
+      "step": 650
+    },
+    {
+      "epoch": 1.5517241379310345,
+      "grad_norm": 17.473194122314453,
+      "learning_rate": 1.549425287356322e-05,
+      "loss": 0.8354,
+      "step": 675
+    },
+    {
+      "epoch": 1.6091954022988506,
+      "grad_norm": 39.43209457397461,
+      "learning_rate": 1.6068965517241383e-05,
+      "loss": 0.7423,
+      "step": 700
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 28.584741592407227,
+      "learning_rate": 1.664367816091954e-05,
+      "loss": 0.7329,
+      "step": 725
+    },
+    {
+      "epoch": 1.7241379310344827,
+      "grad_norm": 41.615875244140625,
+      "learning_rate": 1.72183908045977e-05,
+      "loss": 0.9602,
+      "step": 750
+    },
+    {
+      "epoch": 1.7816091954022988,
+      "grad_norm": 12.660417556762695,
+      "learning_rate": 1.7793103448275862e-05,
+      "loss": 0.7677,
+      "step": 775
+    },
+    {
+      "epoch": 1.839080459770115,
+      "grad_norm": 2.226778030395508,
+      "learning_rate": 1.8367816091954027e-05,
+      "loss": 0.7449,
+      "step": 800
+    },
+    {
+      "epoch": 1.896551724137931,
+      "grad_norm": 16.681472778320312,
+      "learning_rate": 1.8942528735632187e-05,
+      "loss": 0.896,
+      "step": 825
+    },
+    {
+      "epoch": 1.9540229885057472,
+      "grad_norm": 6.7044758796691895,
+      "learning_rate": 1.9517241379310345e-05,
+      "loss": 0.9774,
+      "step": 850
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9385908209437621,
+      "eval_loss": 0.22220541536808014,
+      "eval_macro_f1": 0.8067419009961782,
+      "eval_runtime": 1.258,
+      "eval_samples_per_second": 1229.692,
+      "eval_steps_per_second": 77.104,
+      "eval_weighted_f1": 0.9333166421022604,
+      "step": 870
+    },
+    {
+      "epoch": 2.0114942528735633,
+      "grad_norm": 3.5579495429992676,
+      "learning_rate": 1.9989782886334612e-05,
+      "loss": 0.7244,
+      "step": 875
+    },
+    {
+      "epoch": 2.0689655172413794,
+      "grad_norm": 11.459723472595215,
+      "learning_rate": 1.9925925925925928e-05,
+      "loss": 0.6527,
+      "step": 900
+    },
+    {
+      "epoch": 2.1264367816091956,
+      "grad_norm": 1.200738549232483,
+      "learning_rate": 1.9862068965517244e-05,
+      "loss": 0.6132,
+      "step": 925
+    },
+    {
+      "epoch": 2.1839080459770113,
+      "grad_norm": 8.861099243164062,
+      "learning_rate": 1.979821200510856e-05,
+      "loss": 0.7293,
+      "step": 950
+    },
+    {
+      "epoch": 2.2413793103448274,
+      "grad_norm": 35.11263656616211,
+      "learning_rate": 1.9734355044699872e-05,
+      "loss": 0.7609,
+      "step": 975
+    },
+    {
+      "epoch": 2.2988505747126435,
+      "grad_norm": 55.1410026550293,
+      "learning_rate": 1.9670498084291188e-05,
+      "loss": 0.7171,
+      "step": 1000
+    },
+    {
+      "epoch": 2.3563218390804597,
+      "grad_norm": 21.957231521606445,
+      "learning_rate": 1.9606641123882504e-05,
+      "loss": 0.8151,
+      "step": 1025
+    },
+    {
+      "epoch": 2.413793103448276,
+      "grad_norm": 17.54927635192871,
+      "learning_rate": 1.954278416347382e-05,
+      "loss": 0.4637,
+      "step": 1050
+    },
+    {
+      "epoch": 2.471264367816092,
+      "grad_norm": 1.489425539970398,
+      "learning_rate": 1.9478927203065136e-05,
+      "loss": 0.7887,
+      "step": 1075
+    },
+    {
+      "epoch": 2.528735632183908,
+      "grad_norm": 17.168004989624023,
+      "learning_rate": 1.941507024265645e-05,
+      "loss": 0.8487,
+      "step": 1100
+    },
+    {
+      "epoch": 2.586206896551724,
+      "grad_norm": 32.26480484008789,
+      "learning_rate": 1.9351213282247767e-05,
+      "loss": 0.6533,
+      "step": 1125
+    },
+    {
+      "epoch": 2.6436781609195403,
+      "grad_norm": 23.594446182250977,
+      "learning_rate": 1.928735632183908e-05,
+      "loss": 0.7716,
+      "step": 1150
+    },
+    {
+      "epoch": 2.7011494252873565,
+      "grad_norm": 74.25870513916016,
+      "learning_rate": 1.92234993614304e-05,
+      "loss": 0.7378,
+      "step": 1175
+    },
+    {
+      "epoch": 2.7586206896551726,
+      "grad_norm": 7.432177543640137,
+      "learning_rate": 1.915964240102171e-05,
+      "loss": 0.7095,
+      "step": 1200
+    },
+    {
+      "epoch": 2.8160919540229887,
+      "grad_norm": 0.4676114022731781,
+      "learning_rate": 1.9095785440613027e-05,
+      "loss": 0.6418,
+      "step": 1225
+    },
+    {
+      "epoch": 2.873563218390805,
+      "grad_norm": 175.59898376464844,
+      "learning_rate": 1.9031928480204343e-05,
+      "loss": 0.6508,
+      "step": 1250
+    },
+    {
+      "epoch": 2.9310344827586206,
+      "grad_norm": 35.65433120727539,
+      "learning_rate": 1.896807151979566e-05,
+      "loss": 0.7479,
+      "step": 1275
+    },
+    {
+      "epoch": 2.9885057471264367,
+      "grad_norm": 6.105052947998047,
+      "learning_rate": 1.8904214559386975e-05,
+      "loss": 0.647,
+      "step": 1300
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9463477698771816,
+      "eval_loss": 0.22952817380428314,
+      "eval_macro_f1": 0.8238374735381336,
+      "eval_runtime": 1.2279,
+      "eval_samples_per_second": 1259.889,
+      "eval_steps_per_second": 78.998,
+      "eval_weighted_f1": 0.9424683719153055,
+      "step": 1305
+    },
+    {
+      "epoch": 3.045977011494253,
+      "grad_norm": 20.142383575439453,
+      "learning_rate": 1.884035759897829e-05,
+      "loss": 0.7061,
+      "step": 1325
+    },
+    {
+      "epoch": 3.103448275862069,
+      "grad_norm": 3.260765790939331,
+      "learning_rate": 1.8776500638569606e-05,
+      "loss": 0.4698,
+      "step": 1350
+    },
+    {
+      "epoch": 3.160919540229885,
+      "grad_norm": 8.108859062194824,
+      "learning_rate": 1.871264367816092e-05,
+      "loss": 0.6163,
+      "step": 1375
+    },
+    {
+      "epoch": 3.218390804597701,
+      "grad_norm": 8.331541061401367,
+      "learning_rate": 1.8648786717752235e-05,
+      "loss": 0.603,
+      "step": 1400
+    },
+    {
+      "epoch": 3.2758620689655173,
+      "grad_norm": 33.89698791503906,
+      "learning_rate": 1.8584929757343554e-05,
+      "loss": 0.4861,
+      "step": 1425
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 8.176798820495605,
+      "learning_rate": 1.8521072796934866e-05,
+      "loss": 0.7145,
+      "step": 1450
+    },
+    {
+      "epoch": 3.3908045977011496,
+      "grad_norm": 28.097333908081055,
+      "learning_rate": 1.8457215836526182e-05,
+      "loss": 0.5092,
+      "step": 1475
+    },
+    {
+      "epoch": 3.4482758620689653,
+      "grad_norm": 152.21823120117188,
+      "learning_rate": 1.8393358876117498e-05,
+      "loss": 0.4571,
+      "step": 1500
+    },
+    {
+      "epoch": 3.5057471264367814,
+      "grad_norm": 0.46784135699272156,
+      "learning_rate": 1.8329501915708814e-05,
+      "loss": 0.677,
+      "step": 1525
+    },
+    {
+      "epoch": 3.5632183908045976,
+      "grad_norm": 0.6691511869430542,
+      "learning_rate": 1.826564495530013e-05,
+      "loss": 0.639,
+      "step": 1550
+    },
+    {
+      "epoch": 3.6206896551724137,
+      "grad_norm": 67.59940338134766,
+      "learning_rate": 1.8201787994891446e-05,
+      "loss": 0.7678,
+      "step": 1575
+    },
+    {
+      "epoch": 3.67816091954023,
+      "grad_norm": 30.84942054748535,
+      "learning_rate": 1.813793103448276e-05,
+      "loss": 0.6909,
+      "step": 1600
+    },
+    {
+      "epoch": 3.735632183908046,
+      "grad_norm": 8.947328567504883,
+      "learning_rate": 1.8074074074074074e-05,
+      "loss": 0.6153,
+      "step": 1625
+    },
+    {
+      "epoch": 3.793103448275862,
+      "grad_norm": 7.73529052734375,
+      "learning_rate": 1.801021711366539e-05,
+      "loss": 0.6799,
+      "step": 1650
+    },
+    {
+      "epoch": 3.8505747126436782,
+      "grad_norm": 7.296356678009033,
+      "learning_rate": 1.7946360153256706e-05,
+      "loss": 0.7399,
+      "step": 1675
+    },
+    {
+      "epoch": 3.9080459770114944,
+      "grad_norm": 14.251794815063477,
+      "learning_rate": 1.788250319284802e-05,
+      "loss": 0.5385,
+      "step": 1700
+    },
+    {
+      "epoch": 3.9655172413793105,
+      "grad_norm": 35.0350341796875,
+      "learning_rate": 1.7818646232439337e-05,
+      "loss": 0.5121,
+      "step": 1725
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9495798319327731,
+      "eval_loss": 0.2132098525762558,
+      "eval_macro_f1": 0.8422899931926481,
+      "eval_runtime": 1.254,
+      "eval_samples_per_second": 1233.676,
+      "eval_steps_per_second": 77.354,
+      "eval_weighted_f1": 0.9466516338744746,
+      "step": 1740
+    },
+    {
+      "epoch": 4.022988505747127,
+      "grad_norm": 5.14169979095459,
+      "learning_rate": 1.7754789272030653e-05,
+      "loss": 0.4637,
+      "step": 1750
+    },
+    {
+      "epoch": 4.080459770114943,
+      "grad_norm": 0.6433016657829285,
+      "learning_rate": 1.769093231162197e-05,
+      "loss": 0.4158,
+      "step": 1775
+    },
+    {
+      "epoch": 4.137931034482759,
+      "grad_norm": 21.9676456451416,
+      "learning_rate": 1.762707535121328e-05,
+      "loss": 0.4166,
+      "step": 1800
+    },
+    {
+      "epoch": 4.195402298850575,
+      "grad_norm": 22.019309997558594,
+      "learning_rate": 1.75632183908046e-05,
+      "loss": 0.6026,
+      "step": 1825
+    },
+    {
+      "epoch": 4.252873563218391,
+      "grad_norm": 7.289479732513428,
+      "learning_rate": 1.7499361430395913e-05,
+      "loss": 0.4516,
+      "step": 1850
+    },
+    {
+      "epoch": 4.310344827586207,
+      "grad_norm": 1.7169973850250244,
+      "learning_rate": 1.743550446998723e-05,
+      "loss": 0.3704,
+      "step": 1875
+    },
+    {
+      "epoch": 4.3678160919540225,
+      "grad_norm": 1.0054987668991089,
+      "learning_rate": 1.7371647509578545e-05,
+      "loss": 0.4166,
+      "step": 1900
+    },
+    {
+      "epoch": 4.425287356321839,
+      "grad_norm": 11.990793228149414,
+      "learning_rate": 1.730779054916986e-05,
+      "loss": 0.5573,
+      "step": 1925
+    },
+    {
+      "epoch": 4.482758620689655,
+      "grad_norm": 90.84062957763672,
+      "learning_rate": 1.7243933588761176e-05,
+      "loss": 0.4951,
+      "step": 1950
+    },
+    {
+      "epoch": 4.540229885057471,
+      "grad_norm": 8.116483688354492,
+      "learning_rate": 1.7180076628352492e-05,
+      "loss": 0.3999,
+      "step": 1975
+    },
+    {
+      "epoch": 4.597701149425287,
+      "grad_norm": 20.215063095092773,
+      "learning_rate": 1.7116219667943808e-05,
+      "loss": 0.583,
+      "step": 2000
+    },
+    {
+      "epoch": 4.655172413793103,
+      "grad_norm": 1.2730953693389893,
+      "learning_rate": 1.705236270753512e-05,
+      "loss": 0.5667,
+      "step": 2025
+    },
+    {
+      "epoch": 4.712643678160919,
+      "grad_norm": 15.416848182678223,
+      "learning_rate": 1.6988505747126436e-05,
+      "loss": 0.599,
+      "step": 2050
+    },
+    {
+      "epoch": 4.7701149425287355,
+      "grad_norm": 153.432373046875,
+      "learning_rate": 1.6924648786717756e-05,
+      "loss": 0.5637,
+      "step": 2075
+    },
+    {
+      "epoch": 4.827586206896552,
+      "grad_norm": 31.256765365600586,
+      "learning_rate": 1.6860791826309068e-05,
+      "loss": 0.4586,
+      "step": 2100
+    },
+    {
+      "epoch": 4.885057471264368,
+      "grad_norm": 11.178466796875,
+      "learning_rate": 1.6796934865900384e-05,
+      "loss": 0.4444,
+      "step": 2125
+    },
+    {
+      "epoch": 4.942528735632184,
+      "grad_norm": 30.921354293823242,
+      "learning_rate": 1.67330779054917e-05,
+      "loss": 0.596,
+      "step": 2150
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.1400842666625977,
+      "learning_rate": 1.6669220945083016e-05,
+      "loss": 0.462,
+      "step": 2175
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9489334195216548,
+      "eval_loss": 0.24173498153686523,
+      "eval_macro_f1": 0.8385216516337387,
+      "eval_runtime": 1.2088,
+      "eval_samples_per_second": 1279.755,
+      "eval_steps_per_second": 80.243,
+      "eval_weighted_f1": 0.9453540079211139,
+      "step": 2175
+    },
+    {
+      "epoch": 5.057471264367816,
+      "grad_norm": 0.6968454718589783,
+      "learning_rate": 1.6605363984674328e-05,
+      "loss": 0.3327,
+      "step": 2200
+    },
+    {
+      "epoch": 5.114942528735632,
+      "grad_norm": 7.195080757141113,
+      "learning_rate": 1.6541507024265647e-05,
+      "loss": 0.3164,
+      "step": 2225
+    },
+    {
+      "epoch": 5.172413793103448,
+      "grad_norm": 2.0190861225128174,
+      "learning_rate": 1.6477650063856963e-05,
+      "loss": 0.3164,
+      "step": 2250
+    },
+    {
+      "epoch": 5.2298850574712645,
+      "grad_norm": 10.99679946899414,
+      "learning_rate": 1.6413793103448276e-05,
+      "loss": 0.5999,
+      "step": 2275
+    },
+    {
+      "epoch": 5.287356321839081,
+      "grad_norm": 0.2958125174045563,
+      "learning_rate": 1.634993614303959e-05,
+      "loss": 0.2475,
+      "step": 2300
+    },
+    {
+      "epoch": 5.344827586206897,
+      "grad_norm": 0.3118029236793518,
+      "learning_rate": 1.6286079182630907e-05,
+      "loss": 0.3365,
+      "step": 2325
+    },
+    {
+      "epoch": 5.402298850574713,
+      "grad_norm": 0.6170334815979004,
+      "learning_rate": 1.6222222222222223e-05,
+      "loss": 0.53,
+      "step": 2350
+    },
+    {
+      "epoch": 5.459770114942529,
+      "grad_norm": 3.2743406295776367,
+      "learning_rate": 1.615836526181354e-05,
+      "loss": 0.3217,
+      "step": 2375
+    },
+    {
+      "epoch": 5.517241379310345,
+      "grad_norm": 64.94971466064453,
+      "learning_rate": 1.6094508301404855e-05,
+      "loss": 0.309,
+      "step": 2400
+    },
+    {
+      "epoch": 5.574712643678161,
+      "grad_norm": 19.9976749420166,
+      "learning_rate": 1.603065134099617e-05,
+      "loss": 0.3461,
+      "step": 2425
+    },
+    {
+      "epoch": 5.6321839080459775,
+      "grad_norm": 14.538490295410156,
+      "learning_rate": 1.5966794380587483e-05,
+      "loss": 0.464,
+      "step": 2450
+    },
+    {
+      "epoch": 5.689655172413794,
+      "grad_norm": 2.2718899250030518,
+      "learning_rate": 1.5902937420178802e-05,
+      "loss": 0.4085,
+      "step": 2475
+    },
+    {
+      "epoch": 5.747126436781609,
+      "grad_norm": 0.6248209476470947,
+      "learning_rate": 1.5839080459770115e-05,
+      "loss": 0.4028,
+      "step": 2500
+    },
+    {
+      "epoch": 5.804597701149425,
+      "grad_norm": 14.494895935058594,
+      "learning_rate": 1.577522349936143e-05,
+      "loss": 0.5133,
+      "step": 2525
+    },
+    {
+      "epoch": 5.862068965517241,
+      "grad_norm": 18.209218978881836,
+      "learning_rate": 1.5711366538952746e-05,
+      "loss": 0.3852,
+      "step": 2550
+    },
+    {
+      "epoch": 5.919540229885057,
+      "grad_norm": 18.12614631652832,
+      "learning_rate": 1.5647509578544062e-05,
+      "loss": 0.6005,
+      "step": 2575
+    },
+    {
+      "epoch": 5.977011494252873,
+      "grad_norm": 1.4146639108657837,
+      "learning_rate": 1.5583652618135378e-05,
+      "loss": 0.4763,
+      "step": 2600
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.24376273155212402,
+      "eval_macro_f1": 0.8487241418627351,
+      "eval_runtime": 1.2012,
+      "eval_samples_per_second": 1287.837,
+      "eval_steps_per_second": 80.75,
+      "eval_weighted_f1": 0.9504475602335294,
+      "step": 2610
+    },
+    {
+      "epoch": 6.0344827586206895,
+      "grad_norm": 77.5924072265625,
+      "learning_rate": 1.5519795657726694e-05,
+      "loss": 0.4138,
+      "step": 2625
+    },
+    {
+      "epoch": 6.091954022988506,
+      "grad_norm": 9.609664916992188,
+      "learning_rate": 1.545593869731801e-05,
+      "loss": 0.3479,
+      "step": 2650
+    },
+    {
+      "epoch": 6.149425287356322,
+      "grad_norm": 0.2075120359659195,
+      "learning_rate": 1.5392081736909322e-05,
+      "loss": 0.1707,
+      "step": 2675
+    },
+    {
+      "epoch": 6.206896551724138,
+      "grad_norm": 0.2037656158208847,
+      "learning_rate": 1.5328224776500638e-05,
+      "loss": 0.2967,
+      "step": 2700
+    },
+    {
+      "epoch": 6.264367816091954,
+      "grad_norm": 3.0877346992492676,
+      "learning_rate": 1.5264367816091957e-05,
+      "loss": 0.3796,
+      "step": 2725
+    },
+    {
+      "epoch": 6.32183908045977,
+      "grad_norm": 1.7352479696273804,
+      "learning_rate": 1.520051085568327e-05,
+      "loss": 0.4252,
+      "step": 2750
+    },
+    {
+      "epoch": 6.379310344827586,
+      "grad_norm": 0.3938375413417816,
+      "learning_rate": 1.5136653895274587e-05,
+      "loss": 0.3933,
+      "step": 2775
+    },
+    {
+      "epoch": 6.436781609195402,
+      "grad_norm": 194.0776824951172,
+      "learning_rate": 1.5072796934865901e-05,
+      "loss": 0.3817,
+      "step": 2800
+    },
+    {
+      "epoch": 6.494252873563219,
+      "grad_norm": 7.678138732910156,
+      "learning_rate": 1.5008939974457217e-05,
+      "loss": 0.4296,
+      "step": 2825
+    },
+    {
+      "epoch": 6.551724137931035,
+      "grad_norm": 62.08183288574219,
+      "learning_rate": 1.4945083014048531e-05,
+      "loss": 0.4575,
+      "step": 2850
+    },
+    {
+      "epoch": 6.609195402298851,
+      "grad_norm": 0.5328659415245056,
+      "learning_rate": 1.4881226053639847e-05,
+      "loss": 0.2983,
+      "step": 2875
+    },
+    {
+      "epoch": 6.666666666666667,
+      "grad_norm": 0.22880423069000244,
+      "learning_rate": 1.4817369093231165e-05,
+      "loss": 0.2895,
+      "step": 2900
+    },
+    {
+      "epoch": 6.724137931034483,
+      "grad_norm": 0.22284340858459473,
+      "learning_rate": 1.4753512132822479e-05,
+      "loss": 0.3552,
+      "step": 2925
+    },
+    {
+      "epoch": 6.781609195402299,
+      "grad_norm": 4.827826499938965,
+      "learning_rate": 1.4689655172413795e-05,
+      "loss": 0.3262,
+      "step": 2950
+    },
+    {
+      "epoch": 6.8390804597701145,
+      "grad_norm": 334.7279968261719,
+      "learning_rate": 1.4625798212005109e-05,
+      "loss": 0.3739,
+      "step": 2975
+    },
+    {
+      "epoch": 6.896551724137931,
+      "grad_norm": 25.783180236816406,
+      "learning_rate": 1.4561941251596425e-05,
+      "loss": 0.3972,
+      "step": 3000
+    },
+    {
+      "epoch": 6.954022988505747,
+      "grad_norm": 1.802674412727356,
+      "learning_rate": 1.4498084291187742e-05,
+      "loss": 0.2355,
+      "step": 3025
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.26368269324302673,
+      "eval_macro_f1": 0.8419003061028798,
+      "eval_runtime": 1.2579,
+      "eval_samples_per_second": 1229.876,
+      "eval_steps_per_second": 77.116,
+      "eval_weighted_f1": 0.9477420634574679,
+      "step": 3045
+    },
+    {
+      "epoch": 7.011494252873563,
+      "grad_norm": 0.8118590116500854,
+      "learning_rate": 1.4434227330779056e-05,
+      "loss": 0.3043,
+      "step": 3050
+    },
+    {
+      "epoch": 7.068965517241379,
+      "grad_norm": 0.23474839329719543,
+      "learning_rate": 1.4370370370370372e-05,
+      "loss": 0.2503,
+      "step": 3075
+    },
+    {
+      "epoch": 7.126436781609195,
+      "grad_norm": 8.772260665893555,
+      "learning_rate": 1.4306513409961686e-05,
+      "loss": 0.3619,
+      "step": 3100
+    },
+    {
+      "epoch": 7.183908045977011,
+      "grad_norm": 5.47619104385376,
+      "learning_rate": 1.4242656449553002e-05,
+      "loss": 0.4062,
+      "step": 3125
+    },
+    {
+      "epoch": 7.241379310344827,
+      "grad_norm": 18.230541229248047,
+      "learning_rate": 1.4178799489144316e-05,
+      "loss": 0.3913,
+      "step": 3150
+    },
+    {
+      "epoch": 7.2988505747126435,
+      "grad_norm": 1.6260206699371338,
+      "learning_rate": 1.4114942528735634e-05,
+      "loss": 0.2985,
+      "step": 3175
+    },
+    {
+      "epoch": 7.35632183908046,
+      "grad_norm": 0.2555771470069885,
+      "learning_rate": 1.405108556832695e-05,
+      "loss": 0.2745,
+      "step": 3200
+    },
+    {
+      "epoch": 7.413793103448276,
+      "grad_norm": 14.79897689819336,
+      "learning_rate": 1.3987228607918264e-05,
+      "loss": 0.29,
+      "step": 3225
+    },
+    {
+      "epoch": 7.471264367816092,
+      "grad_norm": 0.14732934534549713,
+      "learning_rate": 1.392337164750958e-05,
+      "loss": 0.2911,
+      "step": 3250
+    },
+    {
+      "epoch": 7.528735632183908,
+      "grad_norm": 0.40408188104629517,
+      "learning_rate": 1.3859514687100894e-05,
+      "loss": 0.2399,
+      "step": 3275
+    },
+    {
+      "epoch": 7.586206896551724,
+      "grad_norm": 27.339426040649414,
+      "learning_rate": 1.3795657726692211e-05,
+      "loss": 0.3214,
+      "step": 3300
+    },
+    {
+      "epoch": 7.64367816091954,
+      "grad_norm": 10.715410232543945,
+      "learning_rate": 1.3731800766283526e-05,
+      "loss": 0.5003,
+      "step": 3325
+    },
+    {
+      "epoch": 7.7011494252873565,
+      "grad_norm": 5.327493667602539,
+      "learning_rate": 1.3667943805874841e-05,
+      "loss": 0.398,
+      "step": 3350
+    },
+    {
+      "epoch": 7.758620689655173,
+      "grad_norm": 0.21260882914066315,
+      "learning_rate": 1.3604086845466157e-05,
+      "loss": 0.2106,
+      "step": 3375
+    },
+    {
+      "epoch": 7.816091954022989,
+      "grad_norm": 0.4254225492477417,
+      "learning_rate": 1.3540229885057471e-05,
+      "loss": 0.3463,
+      "step": 3400
+    },
+    {
+      "epoch": 7.873563218390805,
+      "grad_norm": 0.13743259012699127,
+      "learning_rate": 1.3476372924648789e-05,
+      "loss": 0.2527,
+      "step": 3425
+    },
+    {
+      "epoch": 7.931034482758621,
+      "grad_norm": 2.6045758724212646,
+      "learning_rate": 1.3412515964240103e-05,
+      "loss": 0.36,
+      "step": 3450
+    },
+    {
+      "epoch": 7.988505747126437,
+      "grad_norm": 4.317957401275635,
+      "learning_rate": 1.3348659003831419e-05,
+      "loss": 0.4838,
+      "step": 3475
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.25451889634132385,
+      "eval_macro_f1": 0.8505330085886426,
+      "eval_runtime": 1.2337,
+      "eval_samples_per_second": 1253.933,
+      "eval_steps_per_second": 78.624,
+      "eval_weighted_f1": 0.9487835254244087,
+      "step": 3480
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 8700,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 0
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86b5d49f4b1c774db0dd8115cb41191a0ada580948360e34c5c9da5ad5ea15d2
+size 5624

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<mask>": 64000
+}

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b519d654d7d6a175ea4fa50dfa7726fe015d9179074a99a556e144c1c544a3ab
+size 540026436

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b193ed0286acc027b78436c1f4fd9a8f63e3692cb3c02c09288a4bd2baad985a
+size 1075446330

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c357365334623ba54d7c0de37d9229fd16eab668edcf91ca6cad1497f5bcd99c
+size 14244

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a68f86e1a8ce6594d764bb1f0bc46f7fda81a56b0881d8477b91a1cfd6805d66
+size 1064

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "64000": {
+      "content": "<mask>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "mask_token": "<mask>",
+  "model_max_length": 256,
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "tokenizer_class": "PhobertTokenizer",
+  "unk_token": "<unk>"
+}

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/trainer_state.json

@@ -0,0 +1,1501 @@
+{
+  "best_global_step": 3480,
+  "best_metric": 0.8505330085886426,
+  "best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-3480",
+  "epoch": 11.0,
+  "eval_steps": 500,
+  "global_step": 4785,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05747126436781609,
+      "grad_norm": 12.892321586608887,
+      "learning_rate": 5.517241379310345e-07,
+      "loss": 4.2257,
+      "step": 25
+    },
+    {
+      "epoch": 0.11494252873563218,
+      "grad_norm": 147.58221435546875,
+      "learning_rate": 1.1264367816091955e-06,
+      "loss": 4.04,
+      "step": 50
+    },
+    {
+      "epoch": 0.1724137931034483,
+      "grad_norm": 10.12353801727295,
+      "learning_rate": 1.7011494252873565e-06,
+      "loss": 3.7585,
+      "step": 75
+    },
+    {
+      "epoch": 0.22988505747126436,
+      "grad_norm": 9.198833465576172,
+      "learning_rate": 2.2758620689655173e-06,
+      "loss": 3.406,
+      "step": 100
+    },
+    {
+      "epoch": 0.28735632183908044,
+      "grad_norm": 12.207880973815918,
+      "learning_rate": 2.850574712643678e-06,
+      "loss": 3.0539,
+      "step": 125
+    },
+    {
+      "epoch": 0.3448275862068966,
+      "grad_norm": 11.848393440246582,
+      "learning_rate": 3.425287356321839e-06,
+      "loss": 2.4264,
+      "step": 150
+    },
+    {
+      "epoch": 0.40229885057471265,
+      "grad_norm": 8.870543479919434,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.9691,
+      "step": 175
+    },
+    {
+      "epoch": 0.45977011494252873,
+      "grad_norm": 19.7670955657959,
+      "learning_rate": 4.5747126436781615e-06,
+      "loss": 1.6845,
+      "step": 200
+    },
+    {
+      "epoch": 0.5172413793103449,
+      "grad_norm": 16.912090301513672,
+      "learning_rate": 5.149425287356322e-06,
+      "loss": 1.4159,
+      "step": 225
+    },
+    {
+      "epoch": 0.5747126436781609,
+      "grad_norm": 21.332857131958008,
+      "learning_rate": 5.724137931034483e-06,
+      "loss": 1.2909,
+      "step": 250
+    },
+    {
+      "epoch": 0.632183908045977,
+      "grad_norm": 13.889664649963379,
+      "learning_rate": 6.298850574712644e-06,
+      "loss": 1.3003,
+      "step": 275
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 6.493246078491211,
+      "learning_rate": 6.873563218390805e-06,
+      "loss": 1.2024,
+      "step": 300
+    },
+    {
+      "epoch": 0.7471264367816092,
+      "grad_norm": 19.06503677368164,
+      "learning_rate": 7.4482758620689665e-06,
+      "loss": 0.9701,
+      "step": 325
+    },
+    {
+      "epoch": 0.8045977011494253,
+      "grad_norm": 48.411109924316406,
+      "learning_rate": 8.022988505747126e-06,
+      "loss": 1.0011,
+      "step": 350
+    },
+    {
+      "epoch": 0.8620689655172413,
+      "grad_norm": 24.054780960083008,
+      "learning_rate": 8.597701149425289e-06,
+      "loss": 1.1521,
+      "step": 375
+    },
+    {
+      "epoch": 0.9195402298850575,
+      "grad_norm": 2.5997345447540283,
+      "learning_rate": 9.172413793103448e-06,
+      "loss": 0.8185,
+      "step": 400
+    },
+    {
+      "epoch": 0.9770114942528736,
+      "grad_norm": 8.2587251663208,
+      "learning_rate": 9.74712643678161e-06,
+      "loss": 0.9219,
+      "step": 425
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.940530058177117,
+      "eval_loss": 0.20286934077739716,
+      "eval_macro_f1": 0.8090601999154878,
+      "eval_runtime": 1.2107,
+      "eval_samples_per_second": 1277.73,
+      "eval_steps_per_second": 80.116,
+      "eval_weighted_f1": 0.9374981989040525,
+      "step": 435
+    },
+    {
+      "epoch": 1.0344827586206897,
+      "grad_norm": 74.17597961425781,
+      "learning_rate": 1.032183908045977e-05,
+      "loss": 0.9663,
+      "step": 450
+    },
+    {
+      "epoch": 1.0919540229885056,
+      "grad_norm": 5.443840980529785,
+      "learning_rate": 1.0896551724137932e-05,
+      "loss": 0.8417,
+      "step": 475
+    },
+    {
+      "epoch": 1.1494252873563218,
+      "grad_norm": 14.226733207702637,
+      "learning_rate": 1.1471264367816093e-05,
+      "loss": 0.9687,
+      "step": 500
+    },
+    {
+      "epoch": 1.206896551724138,
+      "grad_norm": 15.986963272094727,
+      "learning_rate": 1.2045977011494254e-05,
+      "loss": 0.7588,
+      "step": 525
+    },
+    {
+      "epoch": 1.264367816091954,
+      "grad_norm": 8.215805053710938,
+      "learning_rate": 1.2620689655172415e-05,
+      "loss": 0.9942,
+      "step": 550
+    },
+    {
+      "epoch": 1.3218390804597702,
+      "grad_norm": 12.830760955810547,
+      "learning_rate": 1.3195402298850574e-05,
+      "loss": 0.8837,
+      "step": 575
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "grad_norm": 3.7767229080200195,
+      "learning_rate": 1.3770114942528736e-05,
+      "loss": 0.7615,
+      "step": 600
+    },
+    {
+      "epoch": 1.4367816091954024,
+      "grad_norm": 12.500882148742676,
+      "learning_rate": 1.4344827586206897e-05,
+      "loss": 0.8383,
+      "step": 625
+    },
+    {
+      "epoch": 1.4942528735632183,
+      "grad_norm": 20.17972183227539,
+      "learning_rate": 1.4919540229885058e-05,
+      "loss": 0.8934,
+      "step": 650
+    },
+    {
+      "epoch": 1.5517241379310345,
+      "grad_norm": 17.473194122314453,
+      "learning_rate": 1.549425287356322e-05,
+      "loss": 0.8354,
+      "step": 675
+    },
+    {
+      "epoch": 1.6091954022988506,
+      "grad_norm": 39.43209457397461,
+      "learning_rate": 1.6068965517241383e-05,
+      "loss": 0.7423,
+      "step": 700
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 28.584741592407227,
+      "learning_rate": 1.664367816091954e-05,
+      "loss": 0.7329,
+      "step": 725
+    },
+    {
+      "epoch": 1.7241379310344827,
+      "grad_norm": 41.615875244140625,
+      "learning_rate": 1.72183908045977e-05,
+      "loss": 0.9602,
+      "step": 750
+    },
+    {
+      "epoch": 1.7816091954022988,
+      "grad_norm": 12.660417556762695,
+      "learning_rate": 1.7793103448275862e-05,
+      "loss": 0.7677,
+      "step": 775
+    },
+    {
+      "epoch": 1.839080459770115,
+      "grad_norm": 2.226778030395508,
+      "learning_rate": 1.8367816091954027e-05,
+      "loss": 0.7449,
+      "step": 800
+    },
+    {
+      "epoch": 1.896551724137931,
+      "grad_norm": 16.681472778320312,
+      "learning_rate": 1.8942528735632187e-05,
+      "loss": 0.896,
+      "step": 825
+    },
+    {
+      "epoch": 1.9540229885057472,
+      "grad_norm": 6.7044758796691895,
+      "learning_rate": 1.9517241379310345e-05,
+      "loss": 0.9774,
+      "step": 850
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9385908209437621,
+      "eval_loss": 0.22220541536808014,
+      "eval_macro_f1": 0.8067419009961782,
+      "eval_runtime": 1.258,
+      "eval_samples_per_second": 1229.692,
+      "eval_steps_per_second": 77.104,
+      "eval_weighted_f1": 0.9333166421022604,
+      "step": 870
+    },
+    {
+      "epoch": 2.0114942528735633,
+      "grad_norm": 3.5579495429992676,
+      "learning_rate": 1.9989782886334612e-05,
+      "loss": 0.7244,
+      "step": 875
+    },
+    {
+      "epoch": 2.0689655172413794,
+      "grad_norm": 11.459723472595215,
+      "learning_rate": 1.9925925925925928e-05,
+      "loss": 0.6527,
+      "step": 900
+    },
+    {
+      "epoch": 2.1264367816091956,
+      "grad_norm": 1.200738549232483,
+      "learning_rate": 1.9862068965517244e-05,
+      "loss": 0.6132,
+      "step": 925
+    },
+    {
+      "epoch": 2.1839080459770113,
+      "grad_norm": 8.861099243164062,
+      "learning_rate": 1.979821200510856e-05,
+      "loss": 0.7293,
+      "step": 950
+    },
+    {
+      "epoch": 2.2413793103448274,
+      "grad_norm": 35.11263656616211,
+      "learning_rate": 1.9734355044699872e-05,
+      "loss": 0.7609,
+      "step": 975
+    },
+    {
+      "epoch": 2.2988505747126435,
+      "grad_norm": 55.1410026550293,
+      "learning_rate": 1.9670498084291188e-05,
+      "loss": 0.7171,
+      "step": 1000
+    },
+    {
+      "epoch": 2.3563218390804597,
+      "grad_norm": 21.957231521606445,
+      "learning_rate": 1.9606641123882504e-05,
+      "loss": 0.8151,
+      "step": 1025
+    },
+    {
+      "epoch": 2.413793103448276,
+      "grad_norm": 17.54927635192871,
+      "learning_rate": 1.954278416347382e-05,
+      "loss": 0.4637,
+      "step": 1050
+    },
+    {
+      "epoch": 2.471264367816092,
+      "grad_norm": 1.489425539970398,
+      "learning_rate": 1.9478927203065136e-05,
+      "loss": 0.7887,
+      "step": 1075
+    },
+    {
+      "epoch": 2.528735632183908,
+      "grad_norm": 17.168004989624023,
+      "learning_rate": 1.941507024265645e-05,
+      "loss": 0.8487,
+      "step": 1100
+    },
+    {
+      "epoch": 2.586206896551724,
+      "grad_norm": 32.26480484008789,
+      "learning_rate": 1.9351213282247767e-05,
+      "loss": 0.6533,
+      "step": 1125
+    },
+    {
+      "epoch": 2.6436781609195403,
+      "grad_norm": 23.594446182250977,
+      "learning_rate": 1.928735632183908e-05,
+      "loss": 0.7716,
+      "step": 1150
+    },
+    {
+      "epoch": 2.7011494252873565,
+      "grad_norm": 74.25870513916016,
+      "learning_rate": 1.92234993614304e-05,
+      "loss": 0.7378,
+      "step": 1175
+    },
+    {
+      "epoch": 2.7586206896551726,
+      "grad_norm": 7.432177543640137,
+      "learning_rate": 1.915964240102171e-05,
+      "loss": 0.7095,
+      "step": 1200
+    },
+    {
+      "epoch": 2.8160919540229887,
+      "grad_norm": 0.4676114022731781,
+      "learning_rate": 1.9095785440613027e-05,
+      "loss": 0.6418,
+      "step": 1225
+    },
+    {
+      "epoch": 2.873563218390805,
+      "grad_norm": 175.59898376464844,
+      "learning_rate": 1.9031928480204343e-05,
+      "loss": 0.6508,
+      "step": 1250
+    },
+    {
+      "epoch": 2.9310344827586206,
+      "grad_norm": 35.65433120727539,
+      "learning_rate": 1.896807151979566e-05,
+      "loss": 0.7479,
+      "step": 1275
+    },
+    {
+      "epoch": 2.9885057471264367,
+      "grad_norm": 6.105052947998047,
+      "learning_rate": 1.8904214559386975e-05,
+      "loss": 0.647,
+      "step": 1300
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9463477698771816,
+      "eval_loss": 0.22952817380428314,
+      "eval_macro_f1": 0.8238374735381336,
+      "eval_runtime": 1.2279,
+      "eval_samples_per_second": 1259.889,
+      "eval_steps_per_second": 78.998,
+      "eval_weighted_f1": 0.9424683719153055,
+      "step": 1305
+    },
+    {
+      "epoch": 3.045977011494253,
+      "grad_norm": 20.142383575439453,
+      "learning_rate": 1.884035759897829e-05,
+      "loss": 0.7061,
+      "step": 1325
+    },
+    {
+      "epoch": 3.103448275862069,
+      "grad_norm": 3.260765790939331,
+      "learning_rate": 1.8776500638569606e-05,
+      "loss": 0.4698,
+      "step": 1350
+    },
+    {
+      "epoch": 3.160919540229885,
+      "grad_norm": 8.108859062194824,
+      "learning_rate": 1.871264367816092e-05,
+      "loss": 0.6163,
+      "step": 1375
+    },
+    {
+      "epoch": 3.218390804597701,
+      "grad_norm": 8.331541061401367,
+      "learning_rate": 1.8648786717752235e-05,
+      "loss": 0.603,
+      "step": 1400
+    },
+    {
+      "epoch": 3.2758620689655173,
+      "grad_norm": 33.89698791503906,
+      "learning_rate": 1.8584929757343554e-05,
+      "loss": 0.4861,
+      "step": 1425
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 8.176798820495605,
+      "learning_rate": 1.8521072796934866e-05,
+      "loss": 0.7145,
+      "step": 1450
+    },
+    {
+      "epoch": 3.3908045977011496,
+      "grad_norm": 28.097333908081055,
+      "learning_rate": 1.8457215836526182e-05,
+      "loss": 0.5092,
+      "step": 1475
+    },
+    {
+      "epoch": 3.4482758620689653,
+      "grad_norm": 152.21823120117188,
+      "learning_rate": 1.8393358876117498e-05,
+      "loss": 0.4571,
+      "step": 1500
+    },
+    {
+      "epoch": 3.5057471264367814,
+      "grad_norm": 0.46784135699272156,
+      "learning_rate": 1.8329501915708814e-05,
+      "loss": 0.677,
+      "step": 1525
+    },
+    {
+      "epoch": 3.5632183908045976,
+      "grad_norm": 0.6691511869430542,
+      "learning_rate": 1.826564495530013e-05,
+      "loss": 0.639,
+      "step": 1550
+    },
+    {
+      "epoch": 3.6206896551724137,
+      "grad_norm": 67.59940338134766,
+      "learning_rate": 1.8201787994891446e-05,
+      "loss": 0.7678,
+      "step": 1575
+    },
+    {
+      "epoch": 3.67816091954023,
+      "grad_norm": 30.84942054748535,
+      "learning_rate": 1.813793103448276e-05,
+      "loss": 0.6909,
+      "step": 1600
+    },
+    {
+      "epoch": 3.735632183908046,
+      "grad_norm": 8.947328567504883,
+      "learning_rate": 1.8074074074074074e-05,
+      "loss": 0.6153,
+      "step": 1625
+    },
+    {
+      "epoch": 3.793103448275862,
+      "grad_norm": 7.73529052734375,
+      "learning_rate": 1.801021711366539e-05,
+      "loss": 0.6799,
+      "step": 1650
+    },
+    {
+      "epoch": 3.8505747126436782,
+      "grad_norm": 7.296356678009033,
+      "learning_rate": 1.7946360153256706e-05,
+      "loss": 0.7399,
+      "step": 1675
+    },
+    {
+      "epoch": 3.9080459770114944,
+      "grad_norm": 14.251794815063477,
+      "learning_rate": 1.788250319284802e-05,
+      "loss": 0.5385,
+      "step": 1700
+    },
+    {
+      "epoch": 3.9655172413793105,
+      "grad_norm": 35.0350341796875,
+      "learning_rate": 1.7818646232439337e-05,
+      "loss": 0.5121,
+      "step": 1725
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9495798319327731,
+      "eval_loss": 0.2132098525762558,
+      "eval_macro_f1": 0.8422899931926481,
+      "eval_runtime": 1.254,
+      "eval_samples_per_second": 1233.676,
+      "eval_steps_per_second": 77.354,
+      "eval_weighted_f1": 0.9466516338744746,
+      "step": 1740
+    },
+    {
+      "epoch": 4.022988505747127,
+      "grad_norm": 5.14169979095459,
+      "learning_rate": 1.7754789272030653e-05,
+      "loss": 0.4637,
+      "step": 1750
+    },
+    {
+      "epoch": 4.080459770114943,
+      "grad_norm": 0.6433016657829285,
+      "learning_rate": 1.769093231162197e-05,
+      "loss": 0.4158,
+      "step": 1775
+    },
+    {
+      "epoch": 4.137931034482759,
+      "grad_norm": 21.9676456451416,
+      "learning_rate": 1.762707535121328e-05,
+      "loss": 0.4166,
+      "step": 1800
+    },
+    {
+      "epoch": 4.195402298850575,
+      "grad_norm": 22.019309997558594,
+      "learning_rate": 1.75632183908046e-05,
+      "loss": 0.6026,
+      "step": 1825
+    },
+    {
+      "epoch": 4.252873563218391,
+      "grad_norm": 7.289479732513428,
+      "learning_rate": 1.7499361430395913e-05,
+      "loss": 0.4516,
+      "step": 1850
+    },
+    {
+      "epoch": 4.310344827586207,
+      "grad_norm": 1.7169973850250244,
+      "learning_rate": 1.743550446998723e-05,
+      "loss": 0.3704,
+      "step": 1875
+    },
+    {
+      "epoch": 4.3678160919540225,
+      "grad_norm": 1.0054987668991089,
+      "learning_rate": 1.7371647509578545e-05,
+      "loss": 0.4166,
+      "step": 1900
+    },
+    {
+      "epoch": 4.425287356321839,
+      "grad_norm": 11.990793228149414,
+      "learning_rate": 1.730779054916986e-05,
+      "loss": 0.5573,
+      "step": 1925
+    },
+    {
+      "epoch": 4.482758620689655,
+      "grad_norm": 90.84062957763672,
+      "learning_rate": 1.7243933588761176e-05,
+      "loss": 0.4951,
+      "step": 1950
+    },
+    {
+      "epoch": 4.540229885057471,
+      "grad_norm": 8.116483688354492,
+      "learning_rate": 1.7180076628352492e-05,
+      "loss": 0.3999,
+      "step": 1975
+    },
+    {
+      "epoch": 4.597701149425287,
+      "grad_norm": 20.215063095092773,
+      "learning_rate": 1.7116219667943808e-05,
+      "loss": 0.583,
+      "step": 2000
+    },
+    {
+      "epoch": 4.655172413793103,
+      "grad_norm": 1.2730953693389893,
+      "learning_rate": 1.705236270753512e-05,
+      "loss": 0.5667,
+      "step": 2025
+    },
+    {
+      "epoch": 4.712643678160919,
+      "grad_norm": 15.416848182678223,
+      "learning_rate": 1.6988505747126436e-05,
+      "loss": 0.599,
+      "step": 2050
+    },
+    {
+      "epoch": 4.7701149425287355,
+      "grad_norm": 153.432373046875,
+      "learning_rate": 1.6924648786717756e-05,
+      "loss": 0.5637,
+      "step": 2075
+    },
+    {
+      "epoch": 4.827586206896552,
+      "grad_norm": 31.256765365600586,
+      "learning_rate": 1.6860791826309068e-05,
+      "loss": 0.4586,
+      "step": 2100
+    },
+    {
+      "epoch": 4.885057471264368,
+      "grad_norm": 11.178466796875,
+      "learning_rate": 1.6796934865900384e-05,
+      "loss": 0.4444,
+      "step": 2125
+    },
+    {
+      "epoch": 4.942528735632184,
+      "grad_norm": 30.921354293823242,
+      "learning_rate": 1.67330779054917e-05,
+      "loss": 0.596,
+      "step": 2150
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.1400842666625977,
+      "learning_rate": 1.6669220945083016e-05,
+      "loss": 0.462,
+      "step": 2175
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9489334195216548,
+      "eval_loss": 0.24173498153686523,
+      "eval_macro_f1": 0.8385216516337387,
+      "eval_runtime": 1.2088,
+      "eval_samples_per_second": 1279.755,
+      "eval_steps_per_second": 80.243,
+      "eval_weighted_f1": 0.9453540079211139,
+      "step": 2175
+    },
+    {
+      "epoch": 5.057471264367816,
+      "grad_norm": 0.6968454718589783,
+      "learning_rate": 1.6605363984674328e-05,
+      "loss": 0.3327,
+      "step": 2200
+    },
+    {
+      "epoch": 5.114942528735632,
+      "grad_norm": 7.195080757141113,
+      "learning_rate": 1.6541507024265647e-05,
+      "loss": 0.3164,
+      "step": 2225
+    },
+    {
+      "epoch": 5.172413793103448,
+      "grad_norm": 2.0190861225128174,
+      "learning_rate": 1.6477650063856963e-05,
+      "loss": 0.3164,
+      "step": 2250
+    },
+    {
+      "epoch": 5.2298850574712645,
+      "grad_norm": 10.99679946899414,
+      "learning_rate": 1.6413793103448276e-05,
+      "loss": 0.5999,
+      "step": 2275
+    },
+    {
+      "epoch": 5.287356321839081,
+      "grad_norm": 0.2958125174045563,
+      "learning_rate": 1.634993614303959e-05,
+      "loss": 0.2475,
+      "step": 2300
+    },
+    {
+      "epoch": 5.344827586206897,
+      "grad_norm": 0.3118029236793518,
+      "learning_rate": 1.6286079182630907e-05,
+      "loss": 0.3365,
+      "step": 2325
+    },
+    {
+      "epoch": 5.402298850574713,
+      "grad_norm": 0.6170334815979004,
+      "learning_rate": 1.6222222222222223e-05,
+      "loss": 0.53,
+      "step": 2350
+    },
+    {
+      "epoch": 5.459770114942529,
+      "grad_norm": 3.2743406295776367,
+      "learning_rate": 1.615836526181354e-05,
+      "loss": 0.3217,
+      "step": 2375
+    },
+    {
+      "epoch": 5.517241379310345,
+      "grad_norm": 64.94971466064453,
+      "learning_rate": 1.6094508301404855e-05,
+      "loss": 0.309,
+      "step": 2400
+    },
+    {
+      "epoch": 5.574712643678161,
+      "grad_norm": 19.9976749420166,
+      "learning_rate": 1.603065134099617e-05,
+      "loss": 0.3461,
+      "step": 2425
+    },
+    {
+      "epoch": 5.6321839080459775,
+      "grad_norm": 14.538490295410156,
+      "learning_rate": 1.5966794380587483e-05,
+      "loss": 0.464,
+      "step": 2450
+    },
+    {
+      "epoch": 5.689655172413794,
+      "grad_norm": 2.2718899250030518,
+      "learning_rate": 1.5902937420178802e-05,
+      "loss": 0.4085,
+      "step": 2475
+    },
+    {
+      "epoch": 5.747126436781609,
+      "grad_norm": 0.6248209476470947,
+      "learning_rate": 1.5839080459770115e-05,
+      "loss": 0.4028,
+      "step": 2500
+    },
+    {
+      "epoch": 5.804597701149425,
+      "grad_norm": 14.494895935058594,
+      "learning_rate": 1.577522349936143e-05,
+      "loss": 0.5133,
+      "step": 2525
+    },
+    {
+      "epoch": 5.862068965517241,
+      "grad_norm": 18.209218978881836,
+      "learning_rate": 1.5711366538952746e-05,
+      "loss": 0.3852,
+      "step": 2550
+    },
+    {
+      "epoch": 5.919540229885057,
+      "grad_norm": 18.12614631652832,
+      "learning_rate": 1.5647509578544062e-05,
+      "loss": 0.6005,
+      "step": 2575
+    },
+    {
+      "epoch": 5.977011494252873,
+      "grad_norm": 1.4146639108657837,
+      "learning_rate": 1.5583652618135378e-05,
+      "loss": 0.4763,
+      "step": 2600
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.24376273155212402,
+      "eval_macro_f1": 0.8487241418627351,
+      "eval_runtime": 1.2012,
+      "eval_samples_per_second": 1287.837,
+      "eval_steps_per_second": 80.75,
+      "eval_weighted_f1": 0.9504475602335294,
+      "step": 2610
+    },
+    {
+      "epoch": 6.0344827586206895,
+      "grad_norm": 77.5924072265625,
+      "learning_rate": 1.5519795657726694e-05,
+      "loss": 0.4138,
+      "step": 2625
+    },
+    {
+      "epoch": 6.091954022988506,
+      "grad_norm": 9.609664916992188,
+      "learning_rate": 1.545593869731801e-05,
+      "loss": 0.3479,
+      "step": 2650
+    },
+    {
+      "epoch": 6.149425287356322,
+      "grad_norm": 0.2075120359659195,
+      "learning_rate": 1.5392081736909322e-05,
+      "loss": 0.1707,
+      "step": 2675
+    },
+    {
+      "epoch": 6.206896551724138,
+      "grad_norm": 0.2037656158208847,
+      "learning_rate": 1.5328224776500638e-05,
+      "loss": 0.2967,
+      "step": 2700
+    },
+    {
+      "epoch": 6.264367816091954,
+      "grad_norm": 3.0877346992492676,
+      "learning_rate": 1.5264367816091957e-05,
+      "loss": 0.3796,
+      "step": 2725
+    },
+    {
+      "epoch": 6.32183908045977,
+      "grad_norm": 1.7352479696273804,
+      "learning_rate": 1.520051085568327e-05,
+      "loss": 0.4252,
+      "step": 2750
+    },
+    {
+      "epoch": 6.379310344827586,
+      "grad_norm": 0.3938375413417816,
+      "learning_rate": 1.5136653895274587e-05,
+      "loss": 0.3933,
+      "step": 2775
+    },
+    {
+      "epoch": 6.436781609195402,
+      "grad_norm": 194.0776824951172,
+      "learning_rate": 1.5072796934865901e-05,
+      "loss": 0.3817,
+      "step": 2800
+    },
+    {
+      "epoch": 6.494252873563219,
+      "grad_norm": 7.678138732910156,
+      "learning_rate": 1.5008939974457217e-05,
+      "loss": 0.4296,
+      "step": 2825
+    },
+    {
+      "epoch": 6.551724137931035,
+      "grad_norm": 62.08183288574219,
+      "learning_rate": 1.4945083014048531e-05,
+      "loss": 0.4575,
+      "step": 2850
+    },
+    {
+      "epoch": 6.609195402298851,
+      "grad_norm": 0.5328659415245056,
+      "learning_rate": 1.4881226053639847e-05,
+      "loss": 0.2983,
+      "step": 2875
+    },
+    {
+      "epoch": 6.666666666666667,
+      "grad_norm": 0.22880423069000244,
+      "learning_rate": 1.4817369093231165e-05,
+      "loss": 0.2895,
+      "step": 2900
+    },
+    {
+      "epoch": 6.724137931034483,
+      "grad_norm": 0.22284340858459473,
+      "learning_rate": 1.4753512132822479e-05,
+      "loss": 0.3552,
+      "step": 2925
+    },
+    {
+      "epoch": 6.781609195402299,
+      "grad_norm": 4.827826499938965,
+      "learning_rate": 1.4689655172413795e-05,
+      "loss": 0.3262,
+      "step": 2950
+    },
+    {
+      "epoch": 6.8390804597701145,
+      "grad_norm": 334.7279968261719,
+      "learning_rate": 1.4625798212005109e-05,
+      "loss": 0.3739,
+      "step": 2975
+    },
+    {
+      "epoch": 6.896551724137931,
+      "grad_norm": 25.783180236816406,
+      "learning_rate": 1.4561941251596425e-05,
+      "loss": 0.3972,
+      "step": 3000
+    },
+    {
+      "epoch": 6.954022988505747,
+      "grad_norm": 1.802674412727356,
+      "learning_rate": 1.4498084291187742e-05,
+      "loss": 0.2355,
+      "step": 3025
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.26368269324302673,
+      "eval_macro_f1": 0.8419003061028798,
+      "eval_runtime": 1.2579,
+      "eval_samples_per_second": 1229.876,
+      "eval_steps_per_second": 77.116,
+      "eval_weighted_f1": 0.9477420634574679,
+      "step": 3045
+    },
+    {
+      "epoch": 7.011494252873563,
+      "grad_norm": 0.8118590116500854,
+      "learning_rate": 1.4434227330779056e-05,
+      "loss": 0.3043,
+      "step": 3050
+    },
+    {
+      "epoch": 7.068965517241379,
+      "grad_norm": 0.23474839329719543,
+      "learning_rate": 1.4370370370370372e-05,
+      "loss": 0.2503,
+      "step": 3075
+    },
+    {
+      "epoch": 7.126436781609195,
+      "grad_norm": 8.772260665893555,
+      "learning_rate": 1.4306513409961686e-05,
+      "loss": 0.3619,
+      "step": 3100
+    },
+    {
+      "epoch": 7.183908045977011,
+      "grad_norm": 5.47619104385376,
+      "learning_rate": 1.4242656449553002e-05,
+      "loss": 0.4062,
+      "step": 3125
+    },
+    {
+      "epoch": 7.241379310344827,
+      "grad_norm": 18.230541229248047,
+      "learning_rate": 1.4178799489144316e-05,
+      "loss": 0.3913,
+      "step": 3150
+    },
+    {
+      "epoch": 7.2988505747126435,
+      "grad_norm": 1.6260206699371338,
+      "learning_rate": 1.4114942528735634e-05,
+      "loss": 0.2985,
+      "step": 3175
+    },
+    {
+      "epoch": 7.35632183908046,
+      "grad_norm": 0.2555771470069885,
+      "learning_rate": 1.405108556832695e-05,
+      "loss": 0.2745,
+      "step": 3200
+    },
+    {
+      "epoch": 7.413793103448276,
+      "grad_norm": 14.79897689819336,
+      "learning_rate": 1.3987228607918264e-05,
+      "loss": 0.29,
+      "step": 3225
+    },
+    {
+      "epoch": 7.471264367816092,
+      "grad_norm": 0.14732934534549713,
+      "learning_rate": 1.392337164750958e-05,
+      "loss": 0.2911,
+      "step": 3250
+    },
+    {
+      "epoch": 7.528735632183908,
+      "grad_norm": 0.40408188104629517,
+      "learning_rate": 1.3859514687100894e-05,
+      "loss": 0.2399,
+      "step": 3275
+    },
+    {
+      "epoch": 7.586206896551724,
+      "grad_norm": 27.339426040649414,
+      "learning_rate": 1.3795657726692211e-05,
+      "loss": 0.3214,
+      "step": 3300
+    },
+    {
+      "epoch": 7.64367816091954,
+      "grad_norm": 10.715410232543945,
+      "learning_rate": 1.3731800766283526e-05,
+      "loss": 0.5003,
+      "step": 3325
+    },
+    {
+      "epoch": 7.7011494252873565,
+      "grad_norm": 5.327493667602539,
+      "learning_rate": 1.3667943805874841e-05,
+      "loss": 0.398,
+      "step": 3350
+    },
+    {
+      "epoch": 7.758620689655173,
+      "grad_norm": 0.21260882914066315,
+      "learning_rate": 1.3604086845466157e-05,
+      "loss": 0.2106,
+      "step": 3375
+    },
+    {
+      "epoch": 7.816091954022989,
+      "grad_norm": 0.4254225492477417,
+      "learning_rate": 1.3540229885057471e-05,
+      "loss": 0.3463,
+      "step": 3400
+    },
+    {
+      "epoch": 7.873563218390805,
+      "grad_norm": 0.13743259012699127,
+      "learning_rate": 1.3476372924648789e-05,
+      "loss": 0.2527,
+      "step": 3425
+    },
+    {
+      "epoch": 7.931034482758621,
+      "grad_norm": 2.6045758724212646,
+      "learning_rate": 1.3412515964240103e-05,
+      "loss": 0.36,
+      "step": 3450
+    },
+    {
+      "epoch": 7.988505747126437,
+      "grad_norm": 4.317957401275635,
+      "learning_rate": 1.3348659003831419e-05,
+      "loss": 0.4838,
+      "step": 3475
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9502262443438914,
+      "eval_loss": 0.25451889634132385,
+      "eval_macro_f1": 0.8505330085886426,
+      "eval_runtime": 1.2337,
+      "eval_samples_per_second": 1253.933,
+      "eval_steps_per_second": 78.624,
+      "eval_weighted_f1": 0.9487835254244087,
+      "step": 3480
+    },
+    {
+      "epoch": 8.045977011494253,
+      "grad_norm": 74.6180191040039,
+      "learning_rate": 1.3284802043422733e-05,
+      "loss": 0.2501,
+      "step": 3500
+    },
+    {
+      "epoch": 8.10344827586207,
+      "grad_norm": 5.052477836608887,
+      "learning_rate": 1.3220945083014049e-05,
+      "loss": 0.2418,
+      "step": 3525
+    },
+    {
+      "epoch": 8.160919540229886,
+      "grad_norm": 0.24052293598651886,
+      "learning_rate": 1.3157088122605367e-05,
+      "loss": 0.3269,
+      "step": 3550
+    },
+    {
+      "epoch": 8.218390804597702,
+      "grad_norm": 64.35535430908203,
+      "learning_rate": 1.309323116219668e-05,
+      "loss": 0.178,
+      "step": 3575
+    },
+    {
+      "epoch": 8.275862068965518,
+      "grad_norm": 0.38525983691215515,
+      "learning_rate": 1.3029374201787996e-05,
+      "loss": 0.1562,
+      "step": 3600
+    },
+    {
+      "epoch": 8.333333333333334,
+      "grad_norm": 0.12001211196184158,
+      "learning_rate": 1.296551724137931e-05,
+      "loss": 0.2553,
+      "step": 3625
+    },
+    {
+      "epoch": 8.39080459770115,
+      "grad_norm": 2.894472599029541,
+      "learning_rate": 1.2901660280970626e-05,
+      "loss": 0.2997,
+      "step": 3650
+    },
+    {
+      "epoch": 8.448275862068966,
+      "grad_norm": 60.778141021728516,
+      "learning_rate": 1.283780332056194e-05,
+      "loss": 0.2853,
+      "step": 3675
+    },
+    {
+      "epoch": 8.505747126436782,
+      "grad_norm": 0.13517223298549652,
+      "learning_rate": 1.2773946360153258e-05,
+      "loss": 0.2024,
+      "step": 3700
+    },
+    {
+      "epoch": 8.563218390804598,
+      "grad_norm": 0.14038611948490143,
+      "learning_rate": 1.2710089399744574e-05,
+      "loss": 0.2584,
+      "step": 3725
+    },
+    {
+      "epoch": 8.620689655172415,
+      "grad_norm": 1.6525335311889648,
+      "learning_rate": 1.2646232439335888e-05,
+      "loss": 0.2312,
+      "step": 3750
+    },
+    {
+      "epoch": 8.678160919540229,
+      "grad_norm": 103.14274597167969,
+      "learning_rate": 1.2582375478927204e-05,
+      "loss": 0.2753,
+      "step": 3775
+    },
+    {
+      "epoch": 8.735632183908045,
+      "grad_norm": 3.4510152339935303,
+      "learning_rate": 1.2518518518518518e-05,
+      "loss": 0.2269,
+      "step": 3800
+    },
+    {
+      "epoch": 8.793103448275861,
+      "grad_norm": 0.4322863519191742,
+      "learning_rate": 1.2454661558109836e-05,
+      "loss": 0.2714,
+      "step": 3825
+    },
+    {
+      "epoch": 8.850574712643677,
+      "grad_norm": 0.19279372692108154,
+      "learning_rate": 1.2390804597701151e-05,
+      "loss": 0.1976,
+      "step": 3850
+    },
+    {
+      "epoch": 8.908045977011493,
+      "grad_norm": 7.546080112457275,
+      "learning_rate": 1.2326947637292466e-05,
+      "loss": 0.2793,
+      "step": 3875
+    },
+    {
+      "epoch": 8.96551724137931,
+      "grad_norm": 4.6050591468811035,
+      "learning_rate": 1.2263090676883781e-05,
+      "loss": 0.3487,
+      "step": 3900
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9482870071105365,
+      "eval_loss": 0.294405072927475,
+      "eval_macro_f1": 0.8443464110600202,
+      "eval_runtime": 1.209,
+      "eval_samples_per_second": 1279.525,
+      "eval_steps_per_second": 80.229,
+      "eval_weighted_f1": 0.9467917248505122,
+      "step": 3915
+    },
+    {
+      "epoch": 9.022988505747126,
+      "grad_norm": 0.1954064965248108,
+      "learning_rate": 1.2199233716475096e-05,
+      "loss": 0.218,
+      "step": 3925
+    },
+    {
+      "epoch": 9.080459770114942,
+      "grad_norm": 56.001216888427734,
+      "learning_rate": 1.2135376756066413e-05,
+      "loss": 0.0972,
+      "step": 3950
+    },
+    {
+      "epoch": 9.137931034482758,
+      "grad_norm": 0.14573942124843597,
+      "learning_rate": 1.2071519795657727e-05,
+      "loss": 0.2953,
+      "step": 3975
+    },
+    {
+      "epoch": 9.195402298850574,
+      "grad_norm": 0.1059441864490509,
+      "learning_rate": 1.2007662835249043e-05,
+      "loss": 0.1414,
+      "step": 4000
+    },
+    {
+      "epoch": 9.25287356321839,
+      "grad_norm": 0.1078171357512474,
+      "learning_rate": 1.1943805874840359e-05,
+      "loss": 0.1682,
+      "step": 4025
+    },
+    {
+      "epoch": 9.310344827586206,
+      "grad_norm": 0.20947247743606567,
+      "learning_rate": 1.1879948914431673e-05,
+      "loss": 0.1891,
+      "step": 4050
+    },
+    {
+      "epoch": 9.367816091954023,
+      "grad_norm": 49.15815353393555,
+      "learning_rate": 1.181609195402299e-05,
+      "loss": 0.2722,
+      "step": 4075
+    },
+    {
+      "epoch": 9.425287356321839,
+      "grad_norm": 0.3581276834011078,
+      "learning_rate": 1.1752234993614305e-05,
+      "loss": 0.1491,
+      "step": 4100
+    },
+    {
+      "epoch": 9.482758620689655,
+      "grad_norm": 0.10994534939527512,
+      "learning_rate": 1.168837803320562e-05,
+      "loss": 0.2579,
+      "step": 4125
+    },
+    {
+      "epoch": 9.540229885057471,
+      "grad_norm": 81.43353271484375,
+      "learning_rate": 1.1624521072796935e-05,
+      "loss": 0.2124,
+      "step": 4150
+    },
+    {
+      "epoch": 9.597701149425287,
+      "grad_norm": 0.21784371137619019,
+      "learning_rate": 1.156066411238825e-05,
+      "loss": 0.3172,
+      "step": 4175
+    },
+    {
+      "epoch": 9.655172413793103,
+      "grad_norm": 10.71652603149414,
+      "learning_rate": 1.1496807151979568e-05,
+      "loss": 0.1866,
+      "step": 4200
+    },
+    {
+      "epoch": 9.71264367816092,
+      "grad_norm": 0.2584083378314972,
+      "learning_rate": 1.1432950191570882e-05,
+      "loss": 0.1141,
+      "step": 4225
+    },
+    {
+      "epoch": 9.770114942528735,
+      "grad_norm": 0.23807008564472198,
+      "learning_rate": 1.1369093231162198e-05,
+      "loss": 0.1708,
+      "step": 4250
+    },
+    {
+      "epoch": 9.827586206896552,
+      "grad_norm": 0.11672578752040863,
+      "learning_rate": 1.1305236270753512e-05,
+      "loss": 0.3457,
+      "step": 4275
+    },
+    {
+      "epoch": 9.885057471264368,
+      "grad_norm": 0.10359542071819305,
+      "learning_rate": 1.1241379310344828e-05,
+      "loss": 0.2364,
+      "step": 4300
+    },
+    {
+      "epoch": 9.942528735632184,
+      "grad_norm": 14.902913093566895,
+      "learning_rate": 1.1177522349936142e-05,
+      "loss": 0.293,
+      "step": 4325
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 2.314290761947632,
+      "learning_rate": 1.111366538952746e-05,
+      "loss": 0.2577,
+      "step": 4350
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.9476405946994182,
+      "eval_loss": 0.30137354135513306,
+      "eval_macro_f1": 0.847874177006584,
+      "eval_runtime": 1.2118,
+      "eval_samples_per_second": 1276.573,
+      "eval_steps_per_second": 80.044,
+      "eval_weighted_f1": 0.9466397068306324,
+      "step": 4350
+    },
+    {
+      "epoch": 10.057471264367816,
+      "grad_norm": 35.0774040222168,
+      "learning_rate": 1.1049808429118776e-05,
+      "loss": 0.1544,
+      "step": 4375
+    },
+    {
+      "epoch": 10.114942528735632,
+      "grad_norm": 0.11511561274528503,
+      "learning_rate": 1.098595146871009e-05,
+      "loss": 0.1373,
+      "step": 4400
+    },
+    {
+      "epoch": 10.172413793103448,
+      "grad_norm": 0.1211598664522171,
+      "learning_rate": 1.0922094508301406e-05,
+      "loss": 0.1435,
+      "step": 4425
+    },
+    {
+      "epoch": 10.229885057471265,
+      "grad_norm": 0.2477167546749115,
+      "learning_rate": 1.085823754789272e-05,
+      "loss": 0.3154,
+      "step": 4450
+    },
+    {
+      "epoch": 10.28735632183908,
+      "grad_norm": 0.11931533366441727,
+      "learning_rate": 1.0794380587484037e-05,
+      "loss": 0.1865,
+      "step": 4475
+    },
+    {
+      "epoch": 10.344827586206897,
+      "grad_norm": 0.09143740683794022,
+      "learning_rate": 1.0730523627075351e-05,
+      "loss": 0.1839,
+      "step": 4500
+    },
+    {
+      "epoch": 10.402298850574713,
+      "grad_norm": 0.13275441527366638,
+      "learning_rate": 1.0666666666666667e-05,
+      "loss": 0.2472,
+      "step": 4525
+    },
+    {
+      "epoch": 10.459770114942529,
+      "grad_norm": 0.1669142246246338,
+      "learning_rate": 1.0602809706257983e-05,
+      "loss": 0.1503,
+      "step": 4550
+    },
+    {
+      "epoch": 10.517241379310345,
+      "grad_norm": 0.2765091061592102,
+      "learning_rate": 1.0538952745849297e-05,
+      "loss": 0.0626,
+      "step": 4575
+    },
+    {
+      "epoch": 10.574712643678161,
+      "grad_norm": 0.06986454874277115,
+      "learning_rate": 1.0475095785440615e-05,
+      "loss": 0.1996,
+      "step": 4600
+    },
+    {
+      "epoch": 10.632183908045977,
+      "grad_norm": 0.07953900098800659,
+      "learning_rate": 1.0411238825031929e-05,
+      "loss": 0.1279,
+      "step": 4625
+    },
+    {
+      "epoch": 10.689655172413794,
+      "grad_norm": 1.7355632781982422,
+      "learning_rate": 1.0347381864623245e-05,
+      "loss": 0.1044,
+      "step": 4650
+    },
+    {
+      "epoch": 10.74712643678161,
+      "grad_norm": 52.49428939819336,
+      "learning_rate": 1.028352490421456e-05,
+      "loss": 0.1091,
+      "step": 4675
+    },
+    {
+      "epoch": 10.804597701149426,
+      "grad_norm": 0.06866724789142609,
+      "learning_rate": 1.0219667943805875e-05,
+      "loss": 0.0619,
+      "step": 4700
+    },
+    {
+      "epoch": 10.862068965517242,
+      "grad_norm": 0.07788178324699402,
+      "learning_rate": 1.0155810983397192e-05,
+      "loss": 0.1693,
+      "step": 4725
+    },
+    {
+      "epoch": 10.919540229885058,
+      "grad_norm": 0.19336405396461487,
+      "learning_rate": 1.0091954022988506e-05,
+      "loss": 0.1837,
+      "step": 4750
+    },
+    {
+      "epoch": 10.977011494252874,
+      "grad_norm": 0.10134702175855637,
+      "learning_rate": 1.0028097062579822e-05,
+      "loss": 0.2558,
+      "step": 4775
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.945054945054945,
+      "eval_loss": 0.3222886323928833,
+      "eval_macro_f1": 0.8414074519179547,
+      "eval_runtime": 1.2,
+      "eval_samples_per_second": 1289.13,
+      "eval_steps_per_second": 80.831,
+      "eval_weighted_f1": 0.9440240868022853,
+      "step": 4785
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 8700,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 3
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

ood-uit_vsfc-phobert-topic-domain-00-s43-20260811T155130Z-d606ec95/trainer/checkpoint-4785/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86b5d49f4b1c774db0dd8115cb41191a0ada580948360e34c5c9da5ad5ea15d2
+size 5624

ood-uit_vsfc-phobert-topic-domain-00-s44-20260811T164157Z-6beb231d/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f3a7f46190001809efc7afead5ba838db14c3c5eb01c7abc06f13e625def3ce
+size 540026404

ood-uit_vsfc-phobert-topic-domain-00-s44-20260811T164157Z-6beb231d/trainer/checkpoint-3480/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<mask>": 64000
+}

ood-uit_vsfc-phobert-topic-domain-00-s44-20260811T164157Z-6beb231d/trainer/checkpoint-3480/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a2ec5dc1337cb56095271b925e2ed11def943509144f2aa729ac9d0996149c3
+size 540026436