End of training

Browse files

Files changed (10) hide show

README.md +10 -10
config.json +2 -8
logs/events.out.tfevents.1740166055.1b9c09404399.787.0 +3 -0
logs/events.out.tfevents.1740166741.1b9c09404399.787.1 +3 -0
model.safetensors +2 -2
trial_4/checkpoint-3501/scheduler.pt +0 -3
trial_4/checkpoint-3501/trainer_state.json +0 -2717
trial_5/checkpoint-438/model.safetensors +0 -3
trial_5/checkpoint-438/rng_state.pth +1 -1
trial_5/checkpoint-438/trainer_state.json +2 -80

README.md CHANGED Viewed

@@ -17,9 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
 This model is a fine-tuned version of [dccuchile/bert-base-spanish-wwm-uncased](https://huggingface.co/dccuchile/bert-base-spanish-wwm-uncased) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.3731
-- Classification Report: {'ar': {'precision': 0.4898785425101215, 'recall': 0.32180851063829785, 'f1-score': 0.3884430176565008, 'support': 376.0}, 'cl': {'precision': 0.3626666666666667, 'recall': 0.4722222222222222, 'f1-score': 0.41025641025641024, 'support': 576.0}, 'co': {'precision': 0.34656084656084657, 'recall': 0.3808139534883721, 'f1-score': 0.3628808864265928, 'support': 344.0}, 'es': {'precision': 0.4630738522954092, 'recall': 0.427255985267035, 'f1-score': 0.4444444444444444, 'support': 543.0}, 'mx': {'precision': 0.43380855397148677, 'recall': 0.43917525773195876, 'f1-score': 0.4364754098360656, 'support': 485.0}, 'pe': {'precision': 0.3769968051118211, 'recall': 0.3390804597701149, 'f1-score': 0.35703479576399394, 'support': 348.0}, 'pr': {'precision': 0.5736434108527132, 'recall': 0.7326732673267327, 'f1-score': 0.6434782608695652, 'support': 101.0}, 'uy': {'precision': 0.35096153846153844, 'recall': 0.3201754385964912, 'f1-score': 0.3348623853211009, 'support': 228.0}, 've': {'precision': 0.16666666666666666, 'recall': 0.045454545454545456, 'f1-score': 0.07142857142857142, 'support': 22.0}, 'accuracy': 0.4085345683096262, 'macro avg': {'precision': 0.39602854256636333, 'recall': 0.38651773783286336, 'f1-score': 0.3832560202225828, 'support': 3023.0}, 'weighted avg': {'precision': 0.4124949665181113, 'recall': 0.4085345683096262, 'f1-score': 0.40601279016852304, 'support': 3023.0}}
-- F1: 0.3833
 ## Model description
@@ -49,13 +49,13 @@ The following hyperparameters were used during training:
 ### Training results
-| Training Loss | Epoch | Step | Validation Loss | Classification Report                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | F1     |
-|:-------------:|:-----:|:----:|:---------------:|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:------:|
-| 1.7603        | 1.0   | 882  | 1.7406          | {'ar': {'precision': 0.35516372795969775, 'recall': 0.375, 'f1-score': 0.3648124191461837, 'support': 376.0}, 'cl': {'precision': 0.3016759776536313, 'recall': 0.28125, 'f1-score': 0.29110512129380056, 'support': 576.0}, 'co': {'precision': 0.3670886075949367, 'recall': 0.25290697674418605, 'f1-score': 0.29948364888123924, 'support': 344.0}, 'es': {'precision': 0.3584905660377358, 'recall': 0.4548802946593002, 'f1-score': 0.400974025974026, 'support': 543.0}, 'mx': {'precision': 0.32465753424657534, 'recall': 0.488659793814433, 'f1-score': 0.39012345679012345, 'support': 485.0}, 'pe': {'precision': 0.3958333333333333, 'recall': 0.27298850574712646, 'f1-score': 0.3231292517006803, 'support': 348.0}, 'pr': {'precision': 0.5631067961165048, 'recall': 0.5742574257425742, 'f1-score': 0.5686274509803921, 'support': 101.0}, 'uy': {'precision': 0.4666666666666667, 'recall': 0.18421052631578946, 'f1-score': 0.2641509433962264, 'support': 228.0}, 've': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 22.0}, 'accuracy': 0.3536222295732716, 'macro avg': {'precision': 0.34807591217878686, 'recall': 0.3204615025581566, 'f1-score': 0.32248959090696355, 'support': 3023.0}, 'weighted avg': {'precision': 0.35948675942105285, 'recall': 0.3536222295732716, 'f1-score': 0.3456546260541325, 'support': 3023.0}}                                                                          | 0.3225 |
-| 1.4223        | 2.0   | 1764 | 1.6758          | {'ar': {'precision': 0.4349315068493151, 'recall': 0.3377659574468085, 'f1-score': 0.38023952095808383, 'support': 376.0}, 'cl': {'precision': 0.336996336996337, 'recall': 0.3194444444444444, 'f1-score': 0.32798573975044565, 'support': 576.0}, 'co': {'precision': 0.36333333333333334, 'recall': 0.3168604651162791, 'f1-score': 0.3385093167701863, 'support': 344.0}, 'es': {'precision': 0.38980716253443526, 'recall': 0.5211786372007366, 'f1-score': 0.44602048857368004, 'support': 543.0}, 'mx': {'precision': 0.35246995994659547, 'recall': 0.5443298969072164, 'f1-score': 0.42787682333873583, 'support': 485.0}, 'pe': {'precision': 0.44308943089430897, 'recall': 0.3132183908045977, 'f1-score': 0.367003367003367, 'support': 348.0}, 'pr': {'precision': 0.759493670886076, 'recall': 0.594059405940594, 'f1-score': 0.6666666666666666, 'support': 101.0}, 'uy': {'precision': 0.5542168674698795, 'recall': 0.20175438596491227, 'f1-score': 0.2958199356913183, 'support': 228.0}, 've': {'precision': 1.0, 'recall': 0.09090909090909091, 'f1-score': 0.16666666666666666, 'support': 22.0}, 'accuracy': 0.3916639100231558, 'macro avg': {'precision': 0.5149264743233645, 'recall': 0.35994674163718665, 'f1-score': 0.3796431694910167, 'support': 3023.0}, 'weighted avg': {'precision': 0.4116802685001794, 'recall': 0.3916639100231558, 'f1-score': 0.3851176158170783, 'support': 3023.0}}                     | 0.3796 |
-| 0.9068        | 3.0   | 2646 | 1.9523          | {'ar': {'precision': 0.39574468085106385, 'recall': 0.4946808510638298, 'f1-score': 0.4397163120567376, 'support': 376.0}, 'cl': {'precision': 0.35144927536231885, 'recall': 0.3368055555555556, 'f1-score': 0.34397163120567376, 'support': 576.0}, 'co': {'precision': 0.31555555555555553, 'recall': 0.4127906976744186, 'f1-score': 0.35768261964735515, 'support': 344.0}, 'es': {'precision': 0.47113163972286376, 'recall': 0.3756906077348066, 'f1-score': 0.4180327868852459, 'support': 543.0}, 'mx': {'precision': 0.43680709534368073, 'recall': 0.4061855670103093, 'f1-score': 0.42094017094017094, 'support': 485.0}, 'pe': {'precision': 0.38661710037174724, 'recall': 0.2988505747126437, 'f1-score': 0.3371150729335494, 'support': 348.0}, 'pr': {'precision': 0.64, 'recall': 0.6336633663366337, 'f1-score': 0.6368159203980099, 'support': 101.0}, 'uy': {'precision': 0.30662020905923343, 'recall': 0.38596491228070173, 'f1-score': 0.341747572815534, 'support': 228.0}, 've': {'precision': 0.18181818181818182, 'recall': 0.09090909090909091, 'f1-score': 0.12121212121212122, 'support': 22.0}, 'accuracy': 0.3906715183592458, 'macro avg': {'precision': 0.3873048597871828, 'recall': 0.3817268025864433, 'f1-score': 0.37969268978826637, 'support': 3023.0}, 'weighted avg': {'precision': 0.3971399185993649, 'recall': 0.3906715183592458, 'f1-score': 0.3902981034934984, 'support': 3023.0}}              | 0.3797 |
-| 0.4818        | 4.0   | 3528 | 2.3731          | {'ar': {'precision': 0.4898785425101215, 'recall': 0.32180851063829785, 'f1-score': 0.3884430176565008, 'support': 376.0}, 'cl': {'precision': 0.3626666666666667, 'recall': 0.4722222222222222, 'f1-score': 0.41025641025641024, 'support': 576.0}, 'co': {'precision': 0.34656084656084657, 'recall': 0.3808139534883721, 'f1-score': 0.3628808864265928, 'support': 344.0}, 'es': {'precision': 0.4630738522954092, 'recall': 0.427255985267035, 'f1-score': 0.4444444444444444, 'support': 543.0}, 'mx': {'precision': 0.43380855397148677, 'recall': 0.43917525773195876, 'f1-score': 0.4364754098360656, 'support': 485.0}, 'pe': {'precision': 0.3769968051118211, 'recall': 0.3390804597701149, 'f1-score': 0.35703479576399394, 'support': 348.0}, 'pr': {'precision': 0.5736434108527132, 'recall': 0.7326732673267327, 'f1-score': 0.6434782608695652, 'support': 101.0}, 'uy': {'precision': 0.35096153846153844, 'recall': 0.3201754385964912, 'f1-score': 0.3348623853211009, 'support': 228.0}, 've': {'precision': 0.16666666666666666, 'recall': 0.045454545454545456, 'f1-score': 0.07142857142857142, 'support': 22.0}, 'accuracy': 0.4085345683096262, 'macro avg': {'precision': 0.39602854256636333, 'recall': 0.38651773783286336, 'f1-score': 0.3832560202225828, 'support': 3023.0}, 'weighted avg': {'precision': 0.4124949665181113, 'recall': 0.4085345683096262, 'f1-score': 0.40601279016852304, 'support': 3023.0}} | 0.3833 |
-| 0.2357        | 5.0   | 4410 | 2.7721          | {'ar': {'precision': 0.42168674698795183, 'recall': 0.3723404255319149, 'f1-score': 0.3954802259887006, 'support': 376.0}, 'cl': {'precision': 0.38753799392097266, 'recall': 0.4427083333333333, 'f1-score': 0.413290113452188, 'support': 576.0}, 'co': {'precision': 0.35051546391752575, 'recall': 0.3953488372093023, 'f1-score': 0.37158469945355194, 'support': 344.0}, 'es': {'precision': 0.4642857142857143, 'recall': 0.40699815837937386, 'f1-score': 0.4337585868498528, 'support': 543.0}, 'mx': {'precision': 0.43089430894308944, 'recall': 0.43711340206185567, 'f1-score': 0.43398157625383826, 'support': 485.0}, 'pe': {'precision': 0.3407960199004975, 'recall': 0.3936781609195402, 'f1-score': 0.36533333333333334, 'support': 348.0}, 'pr': {'precision': 0.6601941747572816, 'recall': 0.6732673267326733, 'f1-score': 0.6666666666666666, 'support': 101.0}, 'uy': {'precision': 0.40853658536585363, 'recall': 0.29385964912280704, 'f1-score': 0.34183673469387754, 'support': 228.0}, 've': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 22.0}, 'accuracy': 0.40886536553092956, 'macro avg': {'precision': 0.3849385564532096, 'recall': 0.3794793659212001, 'f1-score': 0.38021465963244544, 'support': 3023.0}, 'weighted avg': {'precision': 0.41080624270175103, 'recall': 0.40886536553092956, 'f1-score': 0.4078732692419294, 'support': 3023.0}}                                            | 0.3802 |
 ### Framework versions

 This model is a fine-tuned version of [dccuchile/bert-base-spanish-wwm-uncased](https://huggingface.co/dccuchile/bert-base-spanish-wwm-uncased) on an unknown dataset.
 It achieves the following results on the evaluation set:
+- Loss: 2.6157
+- Classification Report: {'ar': {'precision': 0.46107784431137727, 'recall': 0.3938618925831202, 'f1-score': 0.42482758620689653, 'support': 391.0}, 'cl': {'precision': 0.4326241134751773, 'recall': 0.4236111111111111, 'f1-score': 0.4280701754385965, 'support': 576.0}, 'co': {'precision': 0.401840490797546, 'recall': 0.3819241982507289, 'f1-score': 0.39162929745889385, 'support': 343.0}, 'es': {'precision': 0.4861407249466951, 'recall': 0.4175824175824176, 'f1-score': 0.4492610837438424, 'support': 546.0}, 'mx': {'precision': 0.4580152671755725, 'recall': 0.5136986301369864, 'f1-score': 0.48426150121065376, 'support': 584.0}, 'pe': {'precision': 0.301255230125523, 'recall': 0.37305699481865284, 'f1-score': 0.3333333333333333, 'support': 386.0}, 'accuracy': 0.42498230714791224, 'macro avg': {'precision': 0.4234922784719819, 'recall': 0.4172892074138362, 'f1-score': 0.41856382956536947, 'support': 2826.0}, 'weighted avg': {'precision': 0.4304679708106477, 'recall': 0.42498230714791224, 'f1-score': 0.4259648943332467, 'support': 2826.0}}
+- F1: 0.4186
 ## Model description
 ### Training results
+| Training Loss | Epoch | Step | Validation Loss | Classification Report                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | F1     |
+|:-------------:|:-----:|:----:|:---------------:|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:------:|
+| 1.4573        | 1.0   | 825  | 1.5809          | {'ar': {'precision': 0.5341880341880342, 'recall': 0.319693094629156, 'f1-score': 0.4, 'support': 391.0}, 'cl': {'precision': 0.44039735099337746, 'recall': 0.2309027777777778, 'f1-score': 0.30296127562642367, 'support': 576.0}, 'co': {'precision': 0.45930232558139533, 'recall': 0.2303206997084548, 'f1-score': 0.3067961165048544, 'support': 343.0}, 'es': {'precision': 0.3184438040345821, 'recall': 0.8095238095238095, 'f1-score': 0.45708376421923474, 'support': 546.0}, 'mx': {'precision': 0.42574257425742573, 'recall': 0.3681506849315068, 'f1-score': 0.3948576675849403, 'support': 584.0}, 'pe': {'precision': 0.38222222222222224, 'recall': 0.22279792746113988, 'f1-score': 0.281505728314239, 'support': 386.0}, 'accuracy': 0.3821656050955414, 'macro avg': {'precision': 0.4267160518795062, 'recall': 0.36356483233864084, 'f1-score': 0.357200758708282, 'support': 2826.0}, 'weighted avg': {'precision': 0.4211319360796609, 'recall': 0.3821656050955414, 'f1-score': 0.3626902289400526, 'support': 2826.0}}                   | 0.3572 |
+| 1.6236        | 2.0   | 1650 | 1.5003          | {'ar': {'precision': 0.5714285714285714, 'recall': 0.3375959079283887, 'f1-score': 0.42443729903536975, 'support': 391.0}, 'cl': {'precision': 0.4013377926421405, 'recall': 0.4166666666666667, 'f1-score': 0.4088586030664395, 'support': 576.0}, 'co': {'precision': 0.5144508670520231, 'recall': 0.2594752186588921, 'f1-score': 0.3449612403100775, 'support': 343.0}, 'es': {'precision': 0.44165170556552963, 'recall': 0.45054945054945056, 'f1-score': 0.4460562103354488, 'support': 546.0}, 'mx': {'precision': 0.38311688311688313, 'recall': 0.6061643835616438, 'f1-score': 0.46949602122015915, 'support': 584.0}, 'pe': {'precision': 0.33819241982507287, 'recall': 0.3005181347150259, 'f1-score': 0.31824417009602196, 'support': 386.0}, 'accuracy': 0.4164897381457891, 'macro avg': {'precision': 0.4416963732717034, 'recall': 0.3951616270133447, 'f1-score': 0.4020089240105862, 'support': 2826.0}, 'weighted avg': {'precision': 0.4339986385070083, 'recall': 0.4164897381457891, 'f1-score': 0.41059938485783715, 'support': 2826.0}} | 0.4020 |
+| 0.7224        | 3.0   | 2475 | 1.7251          | {'ar': {'precision': 0.5352112676056338, 'recall': 0.3887468030690537, 'f1-score': 0.45037037037037037, 'support': 391.0}, 'cl': {'precision': 0.4388609715242881, 'recall': 0.4548611111111111, 'f1-score': 0.44671781756180734, 'support': 576.0}, 'co': {'precision': 0.291005291005291, 'recall': 0.48104956268221577, 'f1-score': 0.3626373626373626, 'support': 343.0}, 'es': {'precision': 0.4835164835164835, 'recall': 0.40293040293040294, 'f1-score': 0.43956043956043955, 'support': 546.0}, 'mx': {'precision': 0.4577922077922078, 'recall': 0.4828767123287671, 'f1-score': 0.47, 'support': 584.0}, 'pe': {'precision': 0.3289902280130293, 'recall': 0.2616580310880829, 'f1-score': 0.29148629148629146, 'support': 386.0}, 'accuracy': 0.4182590233545648, 'macro avg': {'precision': 0.42256274157615564, 'recall': 0.4120204372016056, 'f1-score': 0.4101287136027119, 'support': 2826.0}, 'weighted avg': {'precision': 0.43177891628106374, 'recall': 0.4182590233545648, 'f1-score': 0.4192436665352936, 'support': 2826.0}}                | 0.4101 |
+| 0.3648        | 4.0   | 3300 | 2.1768          | {'ar': {'precision': 0.3978260869565217, 'recall': 0.4680306905370844, 'f1-score': 0.4300822561692127, 'support': 391.0}, 'cl': {'precision': 0.4788732394366197, 'recall': 0.3541666666666667, 'f1-score': 0.40718562874251496, 'support': 576.0}, 'co': {'precision': 0.35443037974683544, 'recall': 0.40816326530612246, 'f1-score': 0.3794037940379404, 'support': 343.0}, 'es': {'precision': 0.4577702702702703, 'recall': 0.49633699633699635, 'f1-score': 0.47627416520210897, 'support': 546.0}, 'mx': {'precision': 0.47755834829443444, 'recall': 0.4554794520547945, 'f1-score': 0.4662576687116564, 'support': 584.0}, 'pe': {'precision': 0.3106060606060606, 'recall': 0.31865284974093266, 'f1-score': 0.3145780051150895, 'support': 386.0}, 'accuracy': 0.4200283085633404, 'macro avg': {'precision': 0.412844064218457, 'recall': 0.4168049867737662, 'f1-score': 0.4122969196630872, 'support': 2826.0}, 'weighted avg': {'precision': 0.4252233505074715, 'recall': 0.4200283085633404, 'f1-score': 0.41988813459845997, 'support': 2826.0}}  | 0.4123 |
+| 0.2228        | 5.0   | 4125 | 2.6157          | {'ar': {'precision': 0.46107784431137727, 'recall': 0.3938618925831202, 'f1-score': 0.42482758620689653, 'support': 391.0}, 'cl': {'precision': 0.4326241134751773, 'recall': 0.4236111111111111, 'f1-score': 0.4280701754385965, 'support': 576.0}, 'co': {'precision': 0.401840490797546, 'recall': 0.3819241982507289, 'f1-score': 0.39162929745889385, 'support': 343.0}, 'es': {'precision': 0.4861407249466951, 'recall': 0.4175824175824176, 'f1-score': 0.4492610837438424, 'support': 546.0}, 'mx': {'precision': 0.4580152671755725, 'recall': 0.5136986301369864, 'f1-score': 0.48426150121065376, 'support': 584.0}, 'pe': {'precision': 0.301255230125523, 'recall': 0.37305699481865284, 'f1-score': 0.3333333333333333, 'support': 386.0}, 'accuracy': 0.42498230714791224, 'macro avg': {'precision': 0.4234922784719819, 'recall': 0.4172892074138362, 'f1-score': 0.41856382956536947, 'support': 2826.0}, 'weighted avg': {'precision': 0.4304679708106477, 'recall': 0.42498230714791224, 'f1-score': 0.4259648943332467, 'support': 2826.0}}   | 0.4186 |
 ### Framework versions

config.json CHANGED Viewed

@@ -15,10 +15,7 @@
     "2": "co",
     "3": "es",
     "4": "mx",
-    "5": "pe",
-    "6": "pr",
-    "7": "uy",
-    "8": "ve"
   },
   "initializer_range": 0.02,
   "intermediate_size": 3072,
@@ -28,10 +25,7 @@
     "co": 2,
     "es": 3,
     "mx": 4,
-    "pe": 5,
-    "pr": 6,
-    "uy": 7,
-    "ve": 8
   },
   "layer_norm_eps": 1e-12,
   "max_position_embeddings": 512,

     "2": "co",
     "3": "es",
     "4": "mx",
+    "5": "pe"
   },
   "initializer_range": 0.02,
   "intermediate_size": 3072,
     "co": 2,
     "es": 3,
     "mx": 4,
+    "pe": 5
   },
   "layer_norm_eps": 1e-12,
   "max_position_embeddings": 512,

logs/events.out.tfevents.1740166055.1b9c09404399.787.0 ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:25dbc01e818ce46458394bc1819a49597b1a3882c48135947f65a6a85a092972
+size 94225

logs/events.out.tfevents.1740166741.1b9c09404399.787.1 ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:28788e08643cd89f215c1bd479bba25fc433866054f168dda0b6e3cfb8bb1bbe
+size 405

model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9640bf7ca7b6a541089109fcdb8dd4b103d14e514f0b0b70e6032897467855b5
-size 439454740

 version https://git-lfs.github.com/spec/v1
+oid sha256:e9f6de2ac7614e2d6d7deb9bea648290b20714d0437dc42ffa59cf791f6805af
+size 439445512

trial_4/checkpoint-3501/scheduler.pt CHANGED Viewed

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:463d1586a049aea444771846d142dae97553b631027abe884e5f7945608c0fd8
-size 1064

trial_4/checkpoint-3501/trainer_state.json CHANGED Viewed

@@ -1,2717 +0,0 @@
-{
-  "best_metric": 0.34262062899860507,
-  "best_model_checkpoint": "/content/drive/MyDrive/model_outputs/trial_4/checkpoint-3501",
-  "epoch": 3.0,
-  "eval_steps": 500,
-  "global_step": 3501,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.00856898029134533,
-      "grad_norm": 12.96240520477295,
-      "learning_rate": 6.233595156363815e-07,
-      "loss": 2.1941,
-      "step": 10
-    },
-    {
-      "epoch": 0.01713796058269066,
-      "grad_norm": 15.678441047668457,
-      "learning_rate": 1.246719031272763e-06,
-      "loss": 2.2004,
-      "step": 20
-    },
-    {
-      "epoch": 0.02570694087403599,
-      "grad_norm": 19.057016372680664,
-      "learning_rate": 1.8700785469091444e-06,
-      "loss": 2.2096,
-      "step": 30
-    },
-    {
-      "epoch": 0.03427592116538132,
-      "grad_norm": 16.71215057373047,
-      "learning_rate": 2.493438062545526e-06,
-      "loss": 2.2218,
-      "step": 40
-    },
-    {
-      "epoch": 0.04284490145672665,
-      "grad_norm": 19.787519454956055,
-      "learning_rate": 3.1167975781819074e-06,
-      "loss": 2.2402,
-      "step": 50
-    },
-    {
-      "epoch": 0.05141388174807198,
-      "grad_norm": 13.118367195129395,
-      "learning_rate": 3.7401570938182888e-06,
-      "loss": 2.1593,
-      "step": 60
-    },
-    {
-      "epoch": 0.05998286203941731,
-      "grad_norm": 14.878708839416504,
-      "learning_rate": 4.363516609454671e-06,
-      "loss": 2.1811,
-      "step": 70
-    },
-    {
-      "epoch": 0.06855184233076264,
-      "grad_norm": 26.236587524414062,
-      "learning_rate": 4.986876125091052e-06,
-      "loss": 2.0935,
-      "step": 80
-    },
-    {
-      "epoch": 0.07712082262210797,
-      "grad_norm": 14.972529411315918,
-      "learning_rate": 5.610235640727433e-06,
-      "loss": 2.1434,
-      "step": 90
-    },
-    {
-      "epoch": 0.0856898029134533,
-      "grad_norm": 14.05119514465332,
-      "learning_rate": 6.233595156363815e-06,
-      "loss": 2.0917,
-      "step": 100
-    },
-    {
-      "epoch": 0.09425878320479864,
-      "grad_norm": 13.932599067687988,
-      "learning_rate": 6.856954672000195e-06,
-      "loss": 2.0876,
-      "step": 110
-    },
-    {
-      "epoch": 0.10282776349614396,
-      "grad_norm": 11.802385330200195,
-      "learning_rate": 7.4803141876365775e-06,
-      "loss": 2.1551,
-      "step": 120
-    },
-    {
-      "epoch": 0.11139674378748929,
-      "grad_norm": 17.57957649230957,
-      "learning_rate": 8.103673703272958e-06,
-      "loss": 2.0485,
-      "step": 130
-    },
-    {
-      "epoch": 0.11996572407883462,
-      "grad_norm": 12.634247779846191,
-      "learning_rate": 8.727033218909341e-06,
-      "loss": 2.0586,
-      "step": 140
-    },
-    {
-      "epoch": 0.12853470437017994,
-      "grad_norm": 14.163841247558594,
-      "learning_rate": 9.350392734545721e-06,
-      "loss": 2.0832,
-      "step": 150
-    },
-    {
-      "epoch": 0.13710368466152528,
-      "grad_norm": 12.886134147644043,
-      "learning_rate": 9.973752250182104e-06,
-      "loss": 2.0785,
-      "step": 160
-    },
-    {
-      "epoch": 0.1456726649528706,
-      "grad_norm": 12.337018966674805,
-      "learning_rate": 1.0597111765818484e-05,
-      "loss": 2.1782,
-      "step": 170
-    },
-    {
-      "epoch": 0.15424164524421594,
-      "grad_norm": 15.206589698791504,
-      "learning_rate": 1.1220471281454867e-05,
-      "loss": 2.05,
-      "step": 180
-    },
-    {
-      "epoch": 0.16281062553556128,
-      "grad_norm": 15.637868881225586,
-      "learning_rate": 1.1843830797091246e-05,
-      "loss": 2.0865,
-      "step": 190
-    },
-    {
-      "epoch": 0.1713796058269066,
-      "grad_norm": 13.318354606628418,
-      "learning_rate": 1.246719031272763e-05,
-      "loss": 1.9896,
-      "step": 200
-    },
-    {
-      "epoch": 0.17994858611825193,
-      "grad_norm": 15.895234107971191,
-      "learning_rate": 1.3090549828364011e-05,
-      "loss": 2.2041,
-      "step": 210
-    },
-    {
-      "epoch": 0.18851756640959727,
-      "grad_norm": 11.681656837463379,
-      "learning_rate": 1.3580958892325523e-05,
-      "loss": 2.0332,
-      "step": 220
-    },
-    {
-      "epoch": 0.19708654670094258,
-      "grad_norm": 12.1074800491333,
-      "learning_rate": 1.353956614958756e-05,
-      "loss": 1.9992,
-      "step": 230
-    },
-    {
-      "epoch": 0.20565552699228792,
-      "grad_norm": 16.131893157958984,
-      "learning_rate": 1.3498173406849598e-05,
-      "loss": 2.0392,
-      "step": 240
-    },
-    {
-      "epoch": 0.21422450728363324,
-      "grad_norm": 13.486928939819336,
-      "learning_rate": 1.3456780664111635e-05,
-      "loss": 2.0591,
-      "step": 250
-    },
-    {
-      "epoch": 0.22279348757497858,
-      "grad_norm": 12.962348937988281,
-      "learning_rate": 1.3415387921373673e-05,
-      "loss": 2.1737,
-      "step": 260
-    },
-    {
-      "epoch": 0.23136246786632392,
-      "grad_norm": 15.706825256347656,
-      "learning_rate": 1.337399517863571e-05,
-      "loss": 2.0788,
-      "step": 270
-    },
-    {
-      "epoch": 0.23993144815766923,
-      "grad_norm": 15.435516357421875,
-      "learning_rate": 1.3332602435897748e-05,
-      "loss": 2.1001,
-      "step": 280
-    },
-    {
-      "epoch": 0.24850042844901457,
-      "grad_norm": 10.667174339294434,
-      "learning_rate": 1.3291209693159784e-05,
-      "loss": 1.9447,
-      "step": 290
-    },
-    {
-      "epoch": 0.2570694087403599,
-      "grad_norm": 14.559552192687988,
-      "learning_rate": 1.3249816950421821e-05,
-      "loss": 2.0502,
-      "step": 300
-    },
-    {
-      "epoch": 0.2656383890317052,
-      "grad_norm": 15.85730266571045,
-      "learning_rate": 1.320842420768386e-05,
-      "loss": 2.0924,
-      "step": 310
-    },
-    {
-      "epoch": 0.27420736932305056,
-      "grad_norm": 13.333989143371582,
-      "learning_rate": 1.3167031464945898e-05,
-      "loss": 2.1221,
-      "step": 320
-    },
-    {
-      "epoch": 0.2827763496143959,
-      "grad_norm": 13.683819770812988,
-      "learning_rate": 1.3125638722207935e-05,
-      "loss": 2.1562,
-      "step": 330
-    },
-    {
-      "epoch": 0.2913453299057412,
-      "grad_norm": 12.066128730773926,
-      "learning_rate": 1.3084245979469973e-05,
-      "loss": 2.0211,
-      "step": 340
-    },
-    {
-      "epoch": 0.29991431019708653,
-      "grad_norm": 14.573673248291016,
-      "learning_rate": 1.3042853236732009e-05,
-      "loss": 2.0365,
-      "step": 350
-    },
-    {
-      "epoch": 0.30848329048843187,
-      "grad_norm": 19.47612762451172,
-      "learning_rate": 1.3001460493994046e-05,
-      "loss": 1.9834,
-      "step": 360
-    },
-    {
-      "epoch": 0.3170522707797772,
-      "grad_norm": 13.694817543029785,
-      "learning_rate": 1.2960067751256084e-05,
-      "loss": 2.0116,
-      "step": 370
-    },
-    {
-      "epoch": 0.32562125107112255,
-      "grad_norm": 12.057513236999512,
-      "learning_rate": 1.2918675008518121e-05,
-      "loss": 2.0626,
-      "step": 380
-    },
-    {
-      "epoch": 0.3341902313624679,
-      "grad_norm": 12.8948974609375,
-      "learning_rate": 1.2877282265780159e-05,
-      "loss": 1.9787,
-      "step": 390
-    },
-    {
-      "epoch": 0.3427592116538132,
-      "grad_norm": 14.375242233276367,
-      "learning_rate": 1.2835889523042196e-05,
-      "loss": 2.0237,
-      "step": 400
-    },
-    {
-      "epoch": 0.3513281919451585,
-      "grad_norm": 13.976824760437012,
-      "learning_rate": 1.2794496780304234e-05,
-      "loss": 2.1159,
-      "step": 410
-    },
-    {
-      "epoch": 0.35989717223650386,
-      "grad_norm": 15.234496116638184,
-      "learning_rate": 1.2753104037566271e-05,
-      "loss": 1.9116,
-      "step": 420
-    },
-    {
-      "epoch": 0.3684661525278492,
-      "grad_norm": 12.984347343444824,
-      "learning_rate": 1.2711711294828309e-05,
-      "loss": 2.0484,
-      "step": 430
-    },
-    {
-      "epoch": 0.37703513281919454,
-      "grad_norm": 10.440543174743652,
-      "learning_rate": 1.2670318552090346e-05,
-      "loss": 2.0348,
-      "step": 440
-    },
-    {
-      "epoch": 0.3856041131105398,
-      "grad_norm": 11.408565521240234,
-      "learning_rate": 1.2628925809352384e-05,
-      "loss": 2.0866,
-      "step": 450
-    },
-    {
-      "epoch": 0.39417309340188517,
-      "grad_norm": 15.430132865905762,
-      "learning_rate": 1.2587533066614421e-05,
-      "loss": 2.0252,
-      "step": 460
-    },
-    {
-      "epoch": 0.4027420736932305,
-      "grad_norm": 15.553166389465332,
-      "learning_rate": 1.2546140323876457e-05,
-      "loss": 2.0838,
-      "step": 470
-    },
-    {
-      "epoch": 0.41131105398457585,
-      "grad_norm": 14.190583229064941,
-      "learning_rate": 1.2504747581138496e-05,
-      "loss": 2.0674,
-      "step": 480
-    },
-    {
-      "epoch": 0.4198800342759212,
-      "grad_norm": 19.912630081176758,
-      "learning_rate": 1.2463354838400534e-05,
-      "loss": 2.0175,
-      "step": 490
-    },
-    {
-      "epoch": 0.4284490145672665,
-      "grad_norm": 11.050692558288574,
-      "learning_rate": 1.2421962095662571e-05,
-      "loss": 2.0227,
-      "step": 500
-    },
-    {
-      "epoch": 0.4370179948586118,
-      "grad_norm": 13.648895263671875,
-      "learning_rate": 1.2380569352924609e-05,
-      "loss": 2.1153,
-      "step": 510
-    },
-    {
-      "epoch": 0.44558697514995715,
-      "grad_norm": 13.396595001220703,
-      "learning_rate": 1.2339176610186645e-05,
-      "loss": 2.0524,
-      "step": 520
-    },
-    {
-      "epoch": 0.4541559554413025,
-      "grad_norm": 12.05173110961914,
-      "learning_rate": 1.2297783867448682e-05,
-      "loss": 2.0801,
-      "step": 530
-    },
-    {
-      "epoch": 0.46272493573264784,
-      "grad_norm": 11.418408393859863,
-      "learning_rate": 1.225639112471072e-05,
-      "loss": 2.0219,
-      "step": 540
-    },
-    {
-      "epoch": 0.4712939160239931,
-      "grad_norm": 14.92882251739502,
-      "learning_rate": 1.2214998381972757e-05,
-      "loss": 1.9197,
-      "step": 550
-    },
-    {
-      "epoch": 0.47986289631533846,
-      "grad_norm": 13.67534351348877,
-      "learning_rate": 1.2173605639234796e-05,
-      "loss": 1.9579,
-      "step": 560
-    },
-    {
-      "epoch": 0.4884318766066838,
-      "grad_norm": 16.3277645111084,
-      "learning_rate": 1.2132212896496834e-05,
-      "loss": 2.0499,
-      "step": 570
-    },
-    {
-      "epoch": 0.49700085689802914,
-      "grad_norm": 12.686991691589355,
-      "learning_rate": 1.209082015375887e-05,
-      "loss": 1.8892,
-      "step": 580
-    },
-    {
-      "epoch": 0.5055698371893744,
-      "grad_norm": 14.13610553741455,
-      "learning_rate": 1.2049427411020907e-05,
-      "loss": 1.9821,
-      "step": 590
-    },
-    {
-      "epoch": 0.5141388174807198,
-      "grad_norm": 10.20384693145752,
-      "learning_rate": 1.2008034668282945e-05,
-      "loss": 1.8765,
-      "step": 600
-    },
-    {
-      "epoch": 0.5227077977720651,
-      "grad_norm": 11.100608825683594,
-      "learning_rate": 1.1966641925544982e-05,
-      "loss": 1.928,
-      "step": 610
-    },
-    {
-      "epoch": 0.5312767780634104,
-      "grad_norm": 13.737257957458496,
-      "learning_rate": 1.192524918280702e-05,
-      "loss": 2.0266,
-      "step": 620
-    },
-    {
-      "epoch": 0.5398457583547558,
-      "grad_norm": 13.313102722167969,
-      "learning_rate": 1.1883856440069057e-05,
-      "loss": 1.9792,
-      "step": 630
-    },
-    {
-      "epoch": 0.5484147386461011,
-      "grad_norm": 16.294010162353516,
-      "learning_rate": 1.1842463697331093e-05,
-      "loss": 2.0274,
-      "step": 640
-    },
-    {
-      "epoch": 0.5569837189374465,
-      "grad_norm": 14.80037784576416,
-      "learning_rate": 1.1801070954593132e-05,
-      "loss": 2.0204,
-      "step": 650
-    },
-    {
-      "epoch": 0.5655526992287918,
-      "grad_norm": 16.782167434692383,
-      "learning_rate": 1.175967821185517e-05,
-      "loss": 1.896,
-      "step": 660
-    },
-    {
-      "epoch": 0.5741216795201372,
-      "grad_norm": 14.986900329589844,
-      "learning_rate": 1.1718285469117207e-05,
-      "loss": 1.9126,
-      "step": 670
-    },
-    {
-      "epoch": 0.5826906598114824,
-      "grad_norm": 15.64176082611084,
-      "learning_rate": 1.1676892726379245e-05,
-      "loss": 1.8983,
-      "step": 680
-    },
-    {
-      "epoch": 0.5912596401028277,
-      "grad_norm": 14.483860969543457,
-      "learning_rate": 1.1635499983641282e-05,
-      "loss": 2.1211,
-      "step": 690
-    },
-    {
-      "epoch": 0.5998286203941731,
-      "grad_norm": 9.888971328735352,
-      "learning_rate": 1.1594107240903318e-05,
-      "loss": 1.8572,
-      "step": 700
-    },
-    {
-      "epoch": 0.6083976006855184,
-      "grad_norm": 12.11154556274414,
-      "learning_rate": 1.1552714498165355e-05,
-      "loss": 1.866,
-      "step": 710
-    },
-    {
-      "epoch": 0.6169665809768637,
-      "grad_norm": 14.539010047912598,
-      "learning_rate": 1.1511321755427393e-05,
-      "loss": 1.9029,
-      "step": 720
-    },
-    {
-      "epoch": 0.6255355612682091,
-      "grad_norm": 17.459091186523438,
-      "learning_rate": 1.1469929012689432e-05,
-      "loss": 1.9333,
-      "step": 730
-    },
-    {
-      "epoch": 0.6341045415595544,
-      "grad_norm": 14.461770057678223,
-      "learning_rate": 1.142853626995147e-05,
-      "loss": 2.0981,
-      "step": 740
-    },
-    {
-      "epoch": 0.6426735218508998,
-      "grad_norm": 15.937264442443848,
-      "learning_rate": 1.1387143527213507e-05,
-      "loss": 1.8607,
-      "step": 750
-    },
-    {
-      "epoch": 0.6512425021422451,
-      "grad_norm": 16.7382869720459,
-      "learning_rate": 1.1345750784475543e-05,
-      "loss": 2.1263,
-      "step": 760
-    },
-    {
-      "epoch": 0.6598114824335904,
-      "grad_norm": 15.768759727478027,
-      "learning_rate": 1.130435804173758e-05,
-      "loss": 1.928,
-      "step": 770
-    },
-    {
-      "epoch": 0.6683804627249358,
-      "grad_norm": 18.80556869506836,
-      "learning_rate": 1.1262965298999618e-05,
-      "loss": 1.98,
-      "step": 780
-    },
-    {
-      "epoch": 0.676949443016281,
-      "grad_norm": 14.333468437194824,
-      "learning_rate": 1.1221572556261655e-05,
-      "loss": 1.9067,
-      "step": 790
-    },
-    {
-      "epoch": 0.6855184233076264,
-      "grad_norm": 19.20866584777832,
-      "learning_rate": 1.1180179813523693e-05,
-      "loss": 2.0511,
-      "step": 800
-    },
-    {
-      "epoch": 0.6940874035989717,
-      "grad_norm": 18.691129684448242,
-      "learning_rate": 1.113878707078573e-05,
-      "loss": 1.872,
-      "step": 810
-    },
-    {
-      "epoch": 0.702656383890317,
-      "grad_norm": 14.753878593444824,
-      "learning_rate": 1.1097394328047768e-05,
-      "loss": 1.8993,
-      "step": 820
-    },
-    {
-      "epoch": 0.7112253641816624,
-      "grad_norm": 13.987009048461914,
-      "learning_rate": 1.1056001585309806e-05,
-      "loss": 1.8868,
-      "step": 830
-    },
-    {
-      "epoch": 0.7197943444730077,
-      "grad_norm": 16.61811637878418,
-      "learning_rate": 1.1014608842571843e-05,
-      "loss": 1.8531,
-      "step": 840
-    },
-    {
-      "epoch": 0.7283633247643531,
-      "grad_norm": 18.929941177368164,
-      "learning_rate": 1.097321609983388e-05,
-      "loss": 1.8008,
-      "step": 850
-    },
-    {
-      "epoch": 0.7369323050556984,
-      "grad_norm": 18.89458465576172,
-      "learning_rate": 1.0931823357095918e-05,
-      "loss": 1.935,
-      "step": 860
-    },
-    {
-      "epoch": 0.7455012853470437,
-      "grad_norm": 15.524568557739258,
-      "learning_rate": 1.0890430614357956e-05,
-      "loss": 2.0548,
-      "step": 870
-    },
-    {
-      "epoch": 0.7540702656383891,
-      "grad_norm": 17.038110733032227,
-      "learning_rate": 1.0849037871619991e-05,
-      "loss": 2.0176,
-      "step": 880
-    },
-    {
-      "epoch": 0.7626392459297343,
-      "grad_norm": 16.24259376525879,
-      "learning_rate": 1.0807645128882029e-05,
-      "loss": 2.0311,
-      "step": 890
-    },
-    {
-      "epoch": 0.7712082262210797,
-      "grad_norm": 12.702564239501953,
-      "learning_rate": 1.0766252386144068e-05,
-      "loss": 1.9613,
-      "step": 900
-    },
-    {
-      "epoch": 0.779777206512425,
-      "grad_norm": 13.47549057006836,
-      "learning_rate": 1.0724859643406106e-05,
-      "loss": 1.9554,
-      "step": 910
-    },
-    {
-      "epoch": 0.7883461868037703,
-      "grad_norm": 15.315031051635742,
-      "learning_rate": 1.0683466900668143e-05,
-      "loss": 1.8628,
-      "step": 920
-    },
-    {
-      "epoch": 0.7969151670951157,
-      "grad_norm": 12.436241149902344,
-      "learning_rate": 1.0642074157930179e-05,
-      "loss": 1.9767,
-      "step": 930
-    },
-    {
-      "epoch": 0.805484147386461,
-      "grad_norm": 17.100671768188477,
-      "learning_rate": 1.0600681415192216e-05,
-      "loss": 2.0268,
-      "step": 940
-    },
-    {
-      "epoch": 0.8140531276778064,
-      "grad_norm": 14.803923606872559,
-      "learning_rate": 1.0559288672454254e-05,
-      "loss": 2.0516,
-      "step": 950
-    },
-    {
-      "epoch": 0.8226221079691517,
-      "grad_norm": 17.308460235595703,
-      "learning_rate": 1.0517895929716291e-05,
-      "loss": 1.9354,
-      "step": 960
-    },
-    {
-      "epoch": 0.831191088260497,
-      "grad_norm": 38.664039611816406,
-      "learning_rate": 1.0476503186978329e-05,
-      "loss": 1.9623,
-      "step": 970
-    },
-    {
-      "epoch": 0.8397600685518424,
-      "grad_norm": 16.550312042236328,
-      "learning_rate": 1.0435110444240368e-05,
-      "loss": 1.888,
-      "step": 980
-    },
-    {
-      "epoch": 0.8483290488431876,
-      "grad_norm": 19.344846725463867,
-      "learning_rate": 1.0393717701502404e-05,
-      "loss": 1.9992,
-      "step": 990
-    },
-    {
-      "epoch": 0.856898029134533,
-      "grad_norm": 18.766752243041992,
-      "learning_rate": 1.0352324958764441e-05,
-      "loss": 1.8824,
-      "step": 1000
-    },
-    {
-      "epoch": 0.8654670094258783,
-      "grad_norm": 20.725662231445312,
-      "learning_rate": 1.0310932216026479e-05,
-      "loss": 1.7351,
-      "step": 1010
-    },
-    {
-      "epoch": 0.8740359897172236,
-      "grad_norm": 15.772370338439941,
-      "learning_rate": 1.0269539473288516e-05,
-      "loss": 1.9839,
-      "step": 1020
-    },
-    {
-      "epoch": 0.882604970008569,
-      "grad_norm": 15.904685974121094,
-      "learning_rate": 1.0228146730550554e-05,
-      "loss": 2.1096,
-      "step": 1030
-    },
-    {
-      "epoch": 0.8911739502999143,
-      "grad_norm": 19.53571891784668,
-      "learning_rate": 1.0186753987812591e-05,
-      "loss": 1.9122,
-      "step": 1040
-    },
-    {
-      "epoch": 0.8997429305912596,
-      "grad_norm": 14.403525352478027,
-      "learning_rate": 1.0145361245074627e-05,
-      "loss": 1.8495,
-      "step": 1050
-    },
-    {
-      "epoch": 0.908311910882605,
-      "grad_norm": 20.06512451171875,
-      "learning_rate": 1.0103968502336666e-05,
-      "loss": 1.9467,
-      "step": 1060
-    },
-    {
-      "epoch": 0.9168808911739503,
-      "grad_norm": 15.332779884338379,
-      "learning_rate": 1.0062575759598704e-05,
-      "loss": 1.7918,
-      "step": 1070
-    },
-    {
-      "epoch": 0.9254498714652957,
-      "grad_norm": 18.091779708862305,
-      "learning_rate": 1.0021183016860741e-05,
-      "loss": 1.8323,
-      "step": 1080
-    },
-    {
-      "epoch": 0.934018851756641,
-      "grad_norm": 17.92203712463379,
-      "learning_rate": 9.979790274122779e-06,
-      "loss": 1.9448,
-      "step": 1090
-    },
-    {
-      "epoch": 0.9425878320479862,
-      "grad_norm": 11.862146377563477,
-      "learning_rate": 9.938397531384816e-06,
-      "loss": 1.9453,
-      "step": 1100
-    },
-    {
-      "epoch": 0.9511568123393316,
-      "grad_norm": 16.67616844177246,
-      "learning_rate": 9.897004788646852e-06,
-      "loss": 1.9977,
-      "step": 1110
-    },
-    {
-      "epoch": 0.9597257926306769,
-      "grad_norm": 17.18949317932129,
-      "learning_rate": 9.85561204590889e-06,
-      "loss": 1.8888,
-      "step": 1120
-    },
-    {
-      "epoch": 0.9682947729220223,
-      "grad_norm": 19.521203994750977,
-      "learning_rate": 9.814219303170927e-06,
-      "loss": 1.948,
-      "step": 1130
-    },
-    {
-      "epoch": 0.9768637532133676,
-      "grad_norm": 21.371353149414062,
-      "learning_rate": 9.772826560432965e-06,
-      "loss": 1.9285,
-      "step": 1140
-    },
-    {
-      "epoch": 0.9854327335047129,
-      "grad_norm": 18.078819274902344,
-      "learning_rate": 9.731433817695004e-06,
-      "loss": 1.9144,
-      "step": 1150
-    },
-    {
-      "epoch": 0.9940017137960583,
-      "grad_norm": 26.718231201171875,
-      "learning_rate": 9.690041074957041e-06,
-      "loss": 1.8113,
-      "step": 1160
-    },
-    {
-      "epoch": 1.0,
-      "eval_classification_report": {
-        "accuracy": 0.3035,
-        "ar": {
-          "f1-score": 0.23214285714285715,
-          "precision": 0.3,
-          "recall": 0.18932038834951456,
-          "support": 206.0
-        },
-        "cl": {
-          "f1-score": 0.2225609756097561,
-          "precision": 0.1994535519125683,
-          "recall": 0.2517241379310345,
-          "support": 290.0
-        },
-        "co": {
-          "f1-score": 0.3453038674033149,
-          "precision": 0.28868360277136257,
-          "recall": 0.42955326460481097,
-          "support": 291.0
-        },
-        "es": {
-          "f1-score": 0.3566666666666667,
-          "precision": 0.3333333333333333,
-          "recall": 0.3835125448028674,
-          "support": 279.0
-        },
-        "macro avg": {
-          "f1-score": 0.28981047871583737,
-          "precision": 0.31679889037420644,
-          "recall": 0.28794417545744694,
-          "support": 2000.0
-        },
-        "mx": {
-          "f1-score": 0.2676767676767677,
-          "precision": 0.5047619047619047,
-          "recall": 0.18213058419243985,
-          "support": 291.0
-        },
-        "pe": {
-          "f1-score": 0.27666666666666667,
-          "precision": 0.2686084142394822,
-          "recall": 0.2852233676975945,
-          "support": 291.0
-        },
-        "pr": {
-          "f1-score": 0.6162162162162163,
-          "precision": 0.6785714285714286,
-          "recall": 0.5643564356435643,
-          "support": 101.0
-        },
-        "uy": {
-          "f1-score": 0.2910602910602911,
-          "precision": 0.2777777777777778,
-          "recall": 0.3056768558951965,
-          "support": 229.0
-        },
-        "ve": {
-          "f1-score": 0.0,
-          "precision": 0.0,
-          "recall": 0.0,
-          "support": 22.0
-        },
-        "weighted avg": {
-          "f1-score": 0.2998260603986032,
-          "precision": 0.3269230233436701,
-          "recall": 0.3035,
-          "support": 2000.0
-        }
-      },
-      "eval_f1": 0.28981047871583737,
-      "eval_loss": 1.8416967391967773,
-      "eval_runtime": 5.6259,
-      "eval_samples_per_second": 355.502,
-      "eval_steps_per_second": 88.875,
-      "step": 1167
-    },
-    {
-      "epoch": 1.0025706940874035,
-      "grad_norm": 22.602876663208008,
-      "learning_rate": 9.648648332219077e-06,
-      "loss": 1.722,
-      "step": 1170
-    },
-    {
-      "epoch": 1.0111396743787489,
-      "grad_norm": 19.420223236083984,
-      "learning_rate": 9.607255589481115e-06,
-      "loss": 1.7384,
-      "step": 1180
-    },
-    {
-      "epoch": 1.0197086546700942,
-      "grad_norm": 15.554736137390137,
-      "learning_rate": 9.565862846743152e-06,
-      "loss": 1.6844,
-      "step": 1190
-    },
-    {
-      "epoch": 1.0282776349614395,
-      "grad_norm": 24.70032501220703,
-      "learning_rate": 9.52447010400519e-06,
-      "loss": 1.7337,
-      "step": 1200
-    },
-    {
-      "epoch": 1.0368466152527849,
-      "grad_norm": 20.22475814819336,
-      "learning_rate": 9.483077361267227e-06,
-      "loss": 1.7448,
-      "step": 1210
-    },
-    {
-      "epoch": 1.0454155955441302,
-      "grad_norm": 14.43090534210205,
-      "learning_rate": 9.441684618529265e-06,
-      "loss": 1.7228,
-      "step": 1220
-    },
-    {
-      "epoch": 1.0539845758354756,
-      "grad_norm": 16.916563034057617,
-      "learning_rate": 9.400291875791302e-06,
-      "loss": 1.5953,
-      "step": 1230
-    },
-    {
-      "epoch": 1.062553556126821,
-      "grad_norm": 17.133316040039062,
-      "learning_rate": 9.35889913305334e-06,
-      "loss": 1.6757,
-      "step": 1240
-    },
-    {
-      "epoch": 1.0711225364181662,
-      "grad_norm": 21.03934669494629,
-      "learning_rate": 9.317506390315377e-06,
-      "loss": 1.6807,
-      "step": 1250
-    },
-    {
-      "epoch": 1.0796915167095116,
-      "grad_norm": 21.76130485534668,
-      "learning_rate": 9.276113647577415e-06,
-      "loss": 1.6364,
-      "step": 1260
-    },
-    {
-      "epoch": 1.088260497000857,
-      "grad_norm": 21.917848587036133,
-      "learning_rate": 9.234720904839452e-06,
-      "loss": 1.9007,
-      "step": 1270
-    },
-    {
-      "epoch": 1.0968294772922023,
-      "grad_norm": 19.40070915222168,
-      "learning_rate": 9.19332816210149e-06,
-      "loss": 1.6513,
-      "step": 1280
-    },
-    {
-      "epoch": 1.1053984575835476,
-      "grad_norm": 18.84430694580078,
-      "learning_rate": 9.151935419363526e-06,
-      "loss": 1.8647,
-      "step": 1290
-    },
-    {
-      "epoch": 1.113967437874893,
-      "grad_norm": 21.544525146484375,
-      "learning_rate": 9.110542676625563e-06,
-      "loss": 1.8551,
-      "step": 1300
-    },
-    {
-      "epoch": 1.1225364181662383,
-      "grad_norm": 21.512622833251953,
-      "learning_rate": 9.069149933887602e-06,
-      "loss": 1.7686,
-      "step": 1310
-    },
-    {
-      "epoch": 1.1311053984575836,
-      "grad_norm": 21.45429229736328,
-      "learning_rate": 9.02775719114964e-06,
-      "loss": 1.7454,
-      "step": 1320
-    },
-    {
-      "epoch": 1.139674378748929,
-      "grad_norm": 20.992483139038086,
-      "learning_rate": 8.986364448411677e-06,
-      "loss": 1.5629,
-      "step": 1330
-    },
-    {
-      "epoch": 1.1482433590402743,
-      "grad_norm": 26.23398780822754,
-      "learning_rate": 8.944971705673713e-06,
-      "loss": 1.4882,
-      "step": 1340
-    },
-    {
-      "epoch": 1.1568123393316196,
-      "grad_norm": 21.12709617614746,
-      "learning_rate": 8.90357896293575e-06,
-      "loss": 1.6059,
-      "step": 1350
-    },
-    {
-      "epoch": 1.165381319622965,
-      "grad_norm": 20.56809425354004,
-      "learning_rate": 8.862186220197788e-06,
-      "loss": 1.7381,
-      "step": 1360
-    },
-    {
-      "epoch": 1.17395029991431,
-      "grad_norm": 22.67593765258789,
-      "learning_rate": 8.820793477459826e-06,
-      "loss": 1.8096,
-      "step": 1370
-    },
-    {
-      "epoch": 1.1825192802056554,
-      "grad_norm": 22.51243782043457,
-      "learning_rate": 8.779400734721863e-06,
-      "loss": 1.6126,
-      "step": 1380
-    },
-    {
-      "epoch": 1.1910882604970008,
-      "grad_norm": 28.95170021057129,
-      "learning_rate": 8.7380079919839e-06,
-      "loss": 1.6991,
-      "step": 1390
-    },
-    {
-      "epoch": 1.1996572407883461,
-      "grad_norm": 22.659038543701172,
-      "learning_rate": 8.696615249245938e-06,
-      "loss": 1.6806,
-      "step": 1400
-    },
-    {
-      "epoch": 1.2082262210796915,
-      "grad_norm": 23.431997299194336,
-      "learning_rate": 8.655222506507976e-06,
-      "loss": 1.5507,
-      "step": 1410
-    },
-    {
-      "epoch": 1.2167952013710368,
-      "grad_norm": 30.707691192626953,
-      "learning_rate": 8.613829763770013e-06,
-      "loss": 1.6909,
-      "step": 1420
-    },
-    {
-      "epoch": 1.2253641816623821,
-      "grad_norm": 23.785497665405273,
-      "learning_rate": 8.57243702103205e-06,
-      "loss": 1.7508,
-      "step": 1430
-    },
-    {
-      "epoch": 1.2339331619537275,
-      "grad_norm": 30.96599769592285,
-      "learning_rate": 8.531044278294088e-06,
-      "loss": 1.6873,
-      "step": 1440
-    },
-    {
-      "epoch": 1.2425021422450728,
-      "grad_norm": 36.57966995239258,
-      "learning_rate": 8.489651535556126e-06,
-      "loss": 1.6249,
-      "step": 1450
-    },
-    {
-      "epoch": 1.2510711225364182,
-      "grad_norm": 22.85884666442871,
-      "learning_rate": 8.448258792818161e-06,
-      "loss": 1.7765,
-      "step": 1460
-    },
-    {
-      "epoch": 1.2596401028277635,
-      "grad_norm": 19.838592529296875,
-      "learning_rate": 8.406866050080199e-06,
-      "loss": 1.6322,
-      "step": 1470
-    },
-    {
-      "epoch": 1.2682090831191088,
-      "grad_norm": 27.4448299407959,
-      "learning_rate": 8.365473307342238e-06,
-      "loss": 1.784,
-      "step": 1480
-    },
-    {
-      "epoch": 1.2767780634104542,
-      "grad_norm": 26.21979331970215,
-      "learning_rate": 8.324080564604276e-06,
-      "loss": 1.6915,
-      "step": 1490
-    },
-    {
-      "epoch": 1.2853470437017995,
-      "grad_norm": 14.650322914123535,
-      "learning_rate": 8.282687821866313e-06,
-      "loss": 1.7377,
-      "step": 1500
-    },
-    {
-      "epoch": 1.2939160239931449,
-      "grad_norm": 18.169477462768555,
-      "learning_rate": 8.24129507912835e-06,
-      "loss": 1.8747,
-      "step": 1510
-    },
-    {
-      "epoch": 1.3024850042844902,
-      "grad_norm": 23.312740325927734,
-      "learning_rate": 8.199902336390387e-06,
-      "loss": 1.8179,
-      "step": 1520
-    },
-    {
-      "epoch": 1.3110539845758356,
-      "grad_norm": 26.74571418762207,
-      "learning_rate": 8.158509593652424e-06,
-      "loss": 1.7476,
-      "step": 1530
-    },
-    {
-      "epoch": 1.3196229648671807,
-      "grad_norm": 24.429569244384766,
-      "learning_rate": 8.117116850914462e-06,
-      "loss": 1.5958,
-      "step": 1540
-    },
-    {
-      "epoch": 1.328191945158526,
-      "grad_norm": 34.8058967590332,
-      "learning_rate": 8.075724108176499e-06,
-      "loss": 1.5784,
-      "step": 1550
-    },
-    {
-      "epoch": 1.3367609254498714,
-      "grad_norm": 21.165245056152344,
-      "learning_rate": 8.034331365438538e-06,
-      "loss": 1.6505,
-      "step": 1560
-    },
-    {
-      "epoch": 1.3453299057412167,
-      "grad_norm": 25.115171432495117,
-      "learning_rate": 7.992938622700576e-06,
-      "loss": 1.618,
-      "step": 1570
-    },
-    {
-      "epoch": 1.353898886032562,
-      "grad_norm": 22.172096252441406,
-      "learning_rate": 7.951545879962612e-06,
-      "loss": 1.6814,
-      "step": 1580
-    },
-    {
-      "epoch": 1.3624678663239074,
-      "grad_norm": 30.089229583740234,
-      "learning_rate": 7.910153137224649e-06,
-      "loss": 1.6126,
-      "step": 1590
-    },
-    {
-      "epoch": 1.3710368466152527,
-      "grad_norm": 32.49215316772461,
-      "learning_rate": 7.868760394486687e-06,
-      "loss": 1.7492,
-      "step": 1600
-    },
-    {
-      "epoch": 1.379605826906598,
-      "grad_norm": 18.02476692199707,
-      "learning_rate": 7.827367651748724e-06,
-      "loss": 1.605,
-      "step": 1610
-    },
-    {
-      "epoch": 1.3881748071979434,
-      "grad_norm": 60.369930267333984,
-      "learning_rate": 7.785974909010762e-06,
-      "loss": 1.7559,
-      "step": 1620
-    },
-    {
-      "epoch": 1.3967437874892887,
-      "grad_norm": 44.660247802734375,
-      "learning_rate": 7.744582166272799e-06,
-      "loss": 1.7434,
-      "step": 1630
-    },
-    {
-      "epoch": 1.405312767780634,
-      "grad_norm": 26.649761199951172,
-      "learning_rate": 7.703189423534835e-06,
-      "loss": 1.7789,
-      "step": 1640
-    },
-    {
-      "epoch": 1.4138817480719794,
-      "grad_norm": 26.30625343322754,
-      "learning_rate": 7.661796680796874e-06,
-      "loss": 1.7798,
-      "step": 1650
-    },
-    {
-      "epoch": 1.4224507283633248,
-      "grad_norm": 24.721965789794922,
-      "learning_rate": 7.620403938058912e-06,
-      "loss": 1.7247,
-      "step": 1660
-    },
-    {
-      "epoch": 1.43101970865467,
-      "grad_norm": 24.725996017456055,
-      "learning_rate": 7.579011195320949e-06,
-      "loss": 1.6941,
-      "step": 1670
-    },
-    {
-      "epoch": 1.4395886889460154,
-      "grad_norm": 28.72591209411621,
-      "learning_rate": 7.537618452582986e-06,
-      "loss": 1.6854,
-      "step": 1680
-    },
-    {
-      "epoch": 1.4481576692373608,
-      "grad_norm": 15.998064041137695,
-      "learning_rate": 7.496225709845023e-06,
-      "loss": 1.6259,
-      "step": 1690
-    },
-    {
-      "epoch": 1.4567266495287061,
-      "grad_norm": 28.62874984741211,
-      "learning_rate": 7.454832967107061e-06,
-      "loss": 1.7328,
-      "step": 1700
-    },
-    {
-      "epoch": 1.4652956298200515,
-      "grad_norm": 24.98537254333496,
-      "learning_rate": 7.413440224369097e-06,
-      "loss": 1.9542,
-      "step": 1710
-    },
-    {
-      "epoch": 1.4738646101113968,
-      "grad_norm": 31.268770217895508,
-      "learning_rate": 7.372047481631135e-06,
-      "loss": 1.4888,
-      "step": 1720
-    },
-    {
-      "epoch": 1.4824335904027421,
-      "grad_norm": 21.951004028320312,
-      "learning_rate": 7.330654738893174e-06,
-      "loss": 1.7112,
-      "step": 1730
-    },
-    {
-      "epoch": 1.4910025706940875,
-      "grad_norm": 16.30368423461914,
-      "learning_rate": 7.289261996155211e-06,
-      "loss": 1.7718,
-      "step": 1740
-    },
-    {
-      "epoch": 1.4995715509854328,
-      "grad_norm": 28.71227264404297,
-      "learning_rate": 7.247869253417248e-06,
-      "loss": 1.5179,
-      "step": 1750
-    },
-    {
-      "epoch": 1.5081405312767782,
-      "grad_norm": 25.384668350219727,
-      "learning_rate": 7.206476510679286e-06,
-      "loss": 1.6098,
-      "step": 1760
-    },
-    {
-      "epoch": 1.5167095115681235,
-      "grad_norm": 28.461658477783203,
-      "learning_rate": 7.1650837679413224e-06,
-      "loss": 1.5626,
-      "step": 1770
-    },
-    {
-      "epoch": 1.5252784918594688,
-      "grad_norm": 43.065731048583984,
-      "learning_rate": 7.12369102520336e-06,
-      "loss": 1.7604,
-      "step": 1780
-    },
-    {
-      "epoch": 1.5338474721508142,
-      "grad_norm": 28.194068908691406,
-      "learning_rate": 7.0822982824653974e-06,
-      "loss": 1.6964,
-      "step": 1790
-    },
-    {
-      "epoch": 1.5424164524421595,
-      "grad_norm": 37.095558166503906,
-      "learning_rate": 7.040905539727434e-06,
-      "loss": 1.7532,
-      "step": 1800
-    },
-    {
-      "epoch": 1.5509854327335049,
-      "grad_norm": 20.070463180541992,
-      "learning_rate": 6.999512796989473e-06,
-      "loss": 1.5389,
-      "step": 1810
-    },
-    {
-      "epoch": 1.5595544130248502,
-      "grad_norm": 29.111501693725586,
-      "learning_rate": 6.958120054251511e-06,
-      "loss": 1.6142,
-      "step": 1820
-    },
-    {
-      "epoch": 1.5681233933161953,
-      "grad_norm": 22.289817810058594,
-      "learning_rate": 6.9167273115135475e-06,
-      "loss": 1.8116,
-      "step": 1830
-    },
-    {
-      "epoch": 1.5766923736075407,
-      "grad_norm": 71.48851776123047,
-      "learning_rate": 6.875334568775585e-06,
-      "loss": 1.4829,
-      "step": 1840
-    },
-    {
-      "epoch": 1.585261353898886,
-      "grad_norm": 25.093791961669922,
-      "learning_rate": 6.8339418260376225e-06,
-      "loss": 1.5003,
-      "step": 1850
-    },
-    {
-      "epoch": 1.5938303341902313,
-      "grad_norm": 38.602638244628906,
-      "learning_rate": 6.792549083299659e-06,
-      "loss": 1.6257,
-      "step": 1860
-    },
-    {
-      "epoch": 1.6023993144815767,
-      "grad_norm": 30.006364822387695,
-      "learning_rate": 6.751156340561697e-06,
-      "loss": 1.6087,
-      "step": 1870
-    },
-    {
-      "epoch": 1.610968294772922,
-      "grad_norm": 17.102664947509766,
-      "learning_rate": 6.709763597823735e-06,
-      "loss": 1.6324,
-      "step": 1880
-    },
-    {
-      "epoch": 1.6195372750642674,
-      "grad_norm": 19.19342803955078,
-      "learning_rate": 6.668370855085772e-06,
-      "loss": 1.5725,
-      "step": 1890
-    },
-    {
-      "epoch": 1.6281062553556127,
-      "grad_norm": 22.23579978942871,
-      "learning_rate": 6.626978112347809e-06,
-      "loss": 1.6174,
-      "step": 1900
-    },
-    {
-      "epoch": 1.636675235646958,
-      "grad_norm": 18.06783676147461,
-      "learning_rate": 6.585585369609847e-06,
-      "loss": 1.6147,
-      "step": 1910
-    },
-    {
-      "epoch": 1.6452442159383034,
-      "grad_norm": 20.512718200683594,
-      "learning_rate": 6.544192626871884e-06,
-      "loss": 1.6959,
-      "step": 1920
-    },
-    {
-      "epoch": 1.6538131962296485,
-      "grad_norm": 19.70807647705078,
-      "learning_rate": 6.502799884133922e-06,
-      "loss": 1.8072,
-      "step": 1930
-    },
-    {
-      "epoch": 1.6623821765209938,
-      "grad_norm": 24.146228790283203,
-      "learning_rate": 6.461407141395959e-06,
-      "loss": 1.701,
-      "step": 1940
-    },
-    {
-      "epoch": 1.6709511568123392,
-      "grad_norm": 34.090030670166016,
-      "learning_rate": 6.420014398657996e-06,
-      "loss": 1.769,
-      "step": 1950
-    },
-    {
-      "epoch": 1.6795201371036845,
-      "grad_norm": 42.51077651977539,
-      "learning_rate": 6.378621655920034e-06,
-      "loss": 1.5787,
-      "step": 1960
-    },
-    {
-      "epoch": 1.6880891173950299,
-      "grad_norm": 24.358346939086914,
-      "learning_rate": 6.337228913182072e-06,
-      "loss": 1.5821,
-      "step": 1970
-    },
-    {
-      "epoch": 1.6966580976863752,
-      "grad_norm": 35.25168991088867,
-      "learning_rate": 6.295836170444108e-06,
-      "loss": 1.5565,
-      "step": 1980
-    },
-    {
-      "epoch": 1.7052270779777206,
-      "grad_norm": 21.4340763092041,
-      "learning_rate": 6.254443427706146e-06,
-      "loss": 1.6134,
-      "step": 1990
-    },
-    {
-      "epoch": 1.713796058269066,
-      "grad_norm": 22.839839935302734,
-      "learning_rate": 6.213050684968183e-06,
-      "loss": 1.6213,
-      "step": 2000
-    },
-    {
-      "epoch": 1.7223650385604112,
-      "grad_norm": 25.80986785888672,
-      "learning_rate": 6.171657942230221e-06,
-      "loss": 1.6262,
-      "step": 2010
-    },
-    {
-      "epoch": 1.7309340188517566,
-      "grad_norm": 40.22003173828125,
-      "learning_rate": 6.130265199492258e-06,
-      "loss": 1.6161,
-      "step": 2020
-    },
-    {
-      "epoch": 1.739502999143102,
-      "grad_norm": 37.73822784423828,
-      "learning_rate": 6.088872456754296e-06,
-      "loss": 1.622,
-      "step": 2030
-    },
-    {
-      "epoch": 1.7480719794344473,
-      "grad_norm": 23.955387115478516,
-      "learning_rate": 6.047479714016333e-06,
-      "loss": 1.549,
-      "step": 2040
-    },
-    {
-      "epoch": 1.7566409597257926,
-      "grad_norm": 27.26932144165039,
-      "learning_rate": 6.006086971278371e-06,
-      "loss": 1.4373,
-      "step": 2050
-    },
-    {
-      "epoch": 1.765209940017138,
-      "grad_norm": 28.924217224121094,
-      "learning_rate": 5.9646942285404075e-06,
-      "loss": 1.4752,
-      "step": 2060
-    },
-    {
-      "epoch": 1.7737789203084833,
-      "grad_norm": 26.8709659576416,
-      "learning_rate": 5.923301485802445e-06,
-      "loss": 1.7233,
-      "step": 2070
-    },
-    {
-      "epoch": 1.7823479005998286,
-      "grad_norm": 28.21649742126465,
-      "learning_rate": 5.8819087430644825e-06,
-      "loss": 1.7996,
-      "step": 2080
-    },
-    {
-      "epoch": 1.790916880891174,
-      "grad_norm": 24.994258880615234,
-      "learning_rate": 5.84051600032652e-06,
-      "loss": 1.5322,
-      "step": 2090
-    },
-    {
-      "epoch": 1.7994858611825193,
-      "grad_norm": 24.2746639251709,
-      "learning_rate": 5.7991232575885575e-06,
-      "loss": 1.8132,
-      "step": 2100
-    },
-    {
-      "epoch": 1.8080548414738646,
-      "grad_norm": 26.497304916381836,
-      "learning_rate": 5.757730514850595e-06,
-      "loss": 1.5681,
-      "step": 2110
-    },
-    {
-      "epoch": 1.81662382176521,
-      "grad_norm": 43.27742004394531,
-      "learning_rate": 5.716337772112632e-06,
-      "loss": 1.6066,
-      "step": 2120
-    },
-    {
-      "epoch": 1.8251928020565553,
-      "grad_norm": 20.854333877563477,
-      "learning_rate": 5.67494502937467e-06,
-      "loss": 1.6055,
-      "step": 2130
-    },
-    {
-      "epoch": 1.8337617823479007,
-      "grad_norm": 26.56975555419922,
-      "learning_rate": 5.6335522866367075e-06,
-      "loss": 1.8364,
-      "step": 2140
-    },
-    {
-      "epoch": 1.842330762639246,
-      "grad_norm": 25.438674926757812,
-      "learning_rate": 5.592159543898744e-06,
-      "loss": 1.5838,
-      "step": 2150
-    },
-    {
-      "epoch": 1.8508997429305913,
-      "grad_norm": 24.237918853759766,
-      "learning_rate": 5.550766801160782e-06,
-      "loss": 1.4289,
-      "step": 2160
-    },
-    {
-      "epoch": 1.8594687232219367,
-      "grad_norm": 16.972482681274414,
-      "learning_rate": 5.50937405842282e-06,
-      "loss": 1.3532,
-      "step": 2170
-    },
-    {
-      "epoch": 1.868037703513282,
-      "grad_norm": 28.264667510986328,
-      "learning_rate": 5.467981315684857e-06,
-      "loss": 1.6642,
-      "step": 2180
-    },
-    {
-      "epoch": 1.8766066838046274,
-      "grad_norm": 39.41172409057617,
-      "learning_rate": 5.426588572946894e-06,
-      "loss": 1.5638,
-      "step": 2190
-    },
-    {
-      "epoch": 1.8851756640959727,
-      "grad_norm": 25.234155654907227,
-      "learning_rate": 5.385195830208932e-06,
-      "loss": 1.7821,
-      "step": 2200
-    },
-    {
-      "epoch": 1.893744644387318,
-      "grad_norm": 33.262840270996094,
-      "learning_rate": 5.343803087470969e-06,
-      "loss": 1.6387,
-      "step": 2210
-    },
-    {
-      "epoch": 1.9023136246786634,
-      "grad_norm": 25.7631778717041,
-      "learning_rate": 5.302410344733007e-06,
-      "loss": 1.7162,
-      "step": 2220
-    },
-    {
-      "epoch": 1.9108826049700087,
-      "grad_norm": 27.602434158325195,
-      "learning_rate": 5.261017601995044e-06,
-      "loss": 1.712,
-      "step": 2230
-    },
-    {
-      "epoch": 1.919451585261354,
-      "grad_norm": 54.29318618774414,
-      "learning_rate": 5.219624859257081e-06,
-      "loss": 1.5472,
-      "step": 2240
-    },
-    {
-      "epoch": 1.9280205655526992,
-      "grad_norm": 25.333751678466797,
-      "learning_rate": 5.178232116519119e-06,
-      "loss": 1.3439,
-      "step": 2250
-    },
-    {
-      "epoch": 1.9365895458440445,
-      "grad_norm": 27.495887756347656,
-      "learning_rate": 5.136839373781157e-06,
-      "loss": 1.5618,
-      "step": 2260
-    },
-    {
-      "epoch": 1.9451585261353899,
-      "grad_norm": 18.012102127075195,
-      "learning_rate": 5.095446631043193e-06,
-      "loss": 1.3402,
-      "step": 2270
-    },
-    {
-      "epoch": 1.9537275064267352,
-      "grad_norm": 24.922513961791992,
-      "learning_rate": 5.054053888305231e-06,
-      "loss": 1.5323,
-      "step": 2280
-    },
-    {
-      "epoch": 1.9622964867180805,
-      "grad_norm": 29.22248649597168,
-      "learning_rate": 5.012661145567269e-06,
-      "loss": 1.7345,
-      "step": 2290
-    },
-    {
-      "epoch": 1.9708654670094259,
-      "grad_norm": 37.99812316894531,
-      "learning_rate": 4.971268402829306e-06,
-      "loss": 1.5922,
-      "step": 2300
-    },
-    {
-      "epoch": 1.9794344473007712,
-      "grad_norm": 18.794361114501953,
-      "learning_rate": 4.929875660091343e-06,
-      "loss": 1.3667,
-      "step": 2310
-    },
-    {
-      "epoch": 1.9880034275921166,
-      "grad_norm": 26.54824447631836,
-      "learning_rate": 4.888482917353381e-06,
-      "loss": 1.5468,
-      "step": 2320
-    },
-    {
-      "epoch": 1.996572407883462,
-      "grad_norm": 38.71345138549805,
-      "learning_rate": 4.847090174615418e-06,
-      "loss": 1.7227,
-      "step": 2330
-    },
-    {
-      "epoch": 2.0,
-      "eval_classification_report": {
-        "accuracy": 0.3395,
-        "ar": {
-          "f1-score": 0.2974683544303797,
-          "precision": 0.42727272727272725,
-          "recall": 0.22815533980582525,
-          "support": 206.0
-        },
-        "cl": {
-          "f1-score": 0.21897810218978103,
-          "precision": 0.23255813953488372,
-          "recall": 0.20689655172413793,
-          "support": 290.0
-        },
-        "co": {
-          "f1-score": 0.35368956743002544,
-          "precision": 0.2808080808080808,
-          "recall": 0.47766323024054985,
-          "support": 291.0
-        },
-        "es": {
-          "f1-score": 0.3333333333333333,
-          "precision": 0.4207650273224044,
-          "recall": 0.27598566308243727,
-          "support": 279.0
-        },
-        "macro avg": {
-          "f1-score": 0.32076116800689736,
-          "precision": 0.3395273288843494,
-          "recall": 0.32532248394277347,
-          "support": 2000.0
-        },
-        "mx": {
-          "f1-score": 0.39759036144578314,
-          "precision": 0.4782608695652174,
-          "recall": 0.3402061855670103,
-          "support": 291.0
-        },
-        "pe": {
-          "f1-score": 0.3117241379310345,
-          "precision": 0.26036866359447003,
-          "recall": 0.38831615120274915,
-          "support": 291.0
-        },
-        "pr": {
-          "f1-score": 0.6160714285714286,
-          "precision": 0.5609756097560976,
-          "recall": 0.6831683168316832,
-          "support": 101.0
-        },
-        "uy": {
-          "f1-score": 0.35799522673031026,
-          "precision": 0.39473684210526316,
-          "recall": 0.32751091703056767,
-          "support": 229.0
-        },
-        "ve": {
-          "f1-score": 0.0,
-          "precision": 0.0,
-          "recall": 0.0,
-          "support": 22.0
-        },
-        "weighted avg": {
-          "f1-score": 0.33566021764772075,
-          "precision": 0.3582815519991702,
-          "recall": 0.3395,
-          "support": 2000.0
-        }
-      },
-      "eval_f1": 0.32076116800689736,
-      "eval_loss": 1.803144097328186,
-      "eval_runtime": 5.5294,
-      "eval_samples_per_second": 361.705,
-      "eval_steps_per_second": 90.426,
-      "step": 2334
-    },
-    {
-      "epoch": 2.005141388174807,
-      "grad_norm": 22.46615219116211,
-      "learning_rate": 4.805697431877456e-06,
-      "loss": 1.3894,
-      "step": 2340
-    },
-    {
-      "epoch": 2.0137103684661524,
-      "grad_norm": 33.21921157836914,
-      "learning_rate": 4.7643046891394934e-06,
-      "loss": 1.3936,
-      "step": 2350
-    },
-    {
-      "epoch": 2.0222793487574977,
-      "grad_norm": 27.54530143737793,
-      "learning_rate": 4.72291194640153e-06,
-      "loss": 1.4118,
-      "step": 2360
-    },
-    {
-      "epoch": 2.030848329048843,
-      "grad_norm": 29.95819664001465,
-      "learning_rate": 4.681519203663568e-06,
-      "loss": 1.215,
-      "step": 2370
-    },
-    {
-      "epoch": 2.0394173093401884,
-      "grad_norm": 43.5531005859375,
-      "learning_rate": 4.640126460925606e-06,
-      "loss": 1.3518,
-      "step": 2380
-    },
-    {
-      "epoch": 2.0479862896315337,
-      "grad_norm": 33.676414489746094,
-      "learning_rate": 4.598733718187643e-06,
-      "loss": 1.4004,
-      "step": 2390
-    },
-    {
-      "epoch": 2.056555269922879,
-      "grad_norm": 23.472457885742188,
-      "learning_rate": 4.55734097544968e-06,
-      "loss": 1.2582,
-      "step": 2400
-    },
-    {
-      "epoch": 2.0651242502142244,
-      "grad_norm": 27.10287857055664,
-      "learning_rate": 4.515948232711718e-06,
-      "loss": 1.2922,
-      "step": 2410
-    },
-    {
-      "epoch": 2.0736932305055698,
-      "grad_norm": 20.592119216918945,
-      "learning_rate": 4.474555489973755e-06,
-      "loss": 1.3074,
-      "step": 2420
-    },
-    {
-      "epoch": 2.082262210796915,
-      "grad_norm": 22.98575782775879,
-      "learning_rate": 4.433162747235793e-06,
-      "loss": 1.1699,
-      "step": 2430
-    },
-    {
-      "epoch": 2.0908311910882604,
-      "grad_norm": 40.38471221923828,
-      "learning_rate": 4.39177000449783e-06,
-      "loss": 1.438,
-      "step": 2440
-    },
-    {
-      "epoch": 2.0994001713796058,
-      "grad_norm": 21.089099884033203,
-      "learning_rate": 4.350377261759867e-06,
-      "loss": 1.0525,
-      "step": 2450
-    },
-    {
-      "epoch": 2.107969151670951,
-      "grad_norm": 18.718284606933594,
-      "learning_rate": 4.308984519021905e-06,
-      "loss": 0.986,
-      "step": 2460
-    },
-    {
-      "epoch": 2.1165381319622965,
-      "grad_norm": 42.87383270263672,
-      "learning_rate": 4.267591776283942e-06,
-      "loss": 1.3242,
-      "step": 2470
-    },
-    {
-      "epoch": 2.125107112253642,
-      "grad_norm": 9.905497550964355,
-      "learning_rate": 4.226199033545979e-06,
-      "loss": 1.0894,
-      "step": 2480
-    },
-    {
-      "epoch": 2.133676092544987,
-      "grad_norm": 40.20561218261719,
-      "learning_rate": 4.184806290808017e-06,
-      "loss": 1.2448,
-      "step": 2490
-    },
-    {
-      "epoch": 2.1422450728363325,
-      "grad_norm": 25.991907119750977,
-      "learning_rate": 4.143413548070054e-06,
-      "loss": 1.2575,
-      "step": 2500
-    },
-    {
-      "epoch": 2.150814053127678,
-      "grad_norm": 32.69858932495117,
-      "learning_rate": 4.102020805332092e-06,
-      "loss": 1.3396,
-      "step": 2510
-    },
-    {
-      "epoch": 2.159383033419023,
-      "grad_norm": 36.805259704589844,
-      "learning_rate": 4.060628062594129e-06,
-      "loss": 1.4618,
-      "step": 2520
-    },
-    {
-      "epoch": 2.1679520137103685,
-      "grad_norm": 34.98189926147461,
-      "learning_rate": 4.019235319856166e-06,
-      "loss": 1.1907,
-      "step": 2530
-    },
-    {
-      "epoch": 2.176520994001714,
-      "grad_norm": 26.45909309387207,
-      "learning_rate": 3.977842577118204e-06,
-      "loss": 1.4029,
-      "step": 2540
-    },
-    {
-      "epoch": 2.185089974293059,
-      "grad_norm": 35.13296890258789,
-      "learning_rate": 3.936449834380242e-06,
-      "loss": 1.2296,
-      "step": 2550
-    },
-    {
-      "epoch": 2.1936589545844045,
-      "grad_norm": 35.67057800292969,
-      "learning_rate": 3.8950570916422785e-06,
-      "loss": 1.5262,
-      "step": 2560
-    },
-    {
-      "epoch": 2.20222793487575,
-      "grad_norm": 28.157564163208008,
-      "learning_rate": 3.853664348904316e-06,
-      "loss": 1.0519,
-      "step": 2570
-    },
-    {
-      "epoch": 2.210796915167095,
-      "grad_norm": 32.18168640136719,
-      "learning_rate": 3.8122716061663535e-06,
-      "loss": 1.3485,
-      "step": 2580
-    },
-    {
-      "epoch": 2.2193658954584405,
-      "grad_norm": 38.181419372558594,
-      "learning_rate": 3.7708788634283914e-06,
-      "loss": 1.2754,
-      "step": 2590
-    },
-    {
-      "epoch": 2.227934875749786,
-      "grad_norm": 51.16127014160156,
-      "learning_rate": 3.7294861206904285e-06,
-      "loss": 1.4616,
-      "step": 2600
-    },
-    {
-      "epoch": 2.236503856041131,
-      "grad_norm": 37.655372619628906,
-      "learning_rate": 3.688093377952466e-06,
-      "loss": 1.1636,
-      "step": 2610
-    },
-    {
-      "epoch": 2.2450728363324766,
-      "grad_norm": 37.25764083862305,
-      "learning_rate": 3.646700635214503e-06,
-      "loss": 1.1733,
-      "step": 2620
-    },
-    {
-      "epoch": 2.253641816623822,
-      "grad_norm": 20.390323638916016,
-      "learning_rate": 3.605307892476541e-06,
-      "loss": 1.2678,
-      "step": 2630
-    },
-    {
-      "epoch": 2.2622107969151672,
-      "grad_norm": 23.969755172729492,
-      "learning_rate": 3.563915149738578e-06,
-      "loss": 1.2794,
-      "step": 2640
-    },
-    {
-      "epoch": 2.2707797772065126,
-      "grad_norm": 41.76385498046875,
-      "learning_rate": 3.5225224070006156e-06,
-      "loss": 1.0864,
-      "step": 2650
-    },
-    {
-      "epoch": 2.279348757497858,
-      "grad_norm": 43.808013916015625,
-      "learning_rate": 3.4811296642626527e-06,
-      "loss": 1.116,
-      "step": 2660
-    },
-    {
-      "epoch": 2.2879177377892033,
-      "grad_norm": 23.652198791503906,
-      "learning_rate": 3.4397369215246906e-06,
-      "loss": 1.1137,
-      "step": 2670
-    },
-    {
-      "epoch": 2.2964867180805486,
-      "grad_norm": 73.83686828613281,
-      "learning_rate": 3.398344178786728e-06,
-      "loss": 1.3697,
-      "step": 2680
-    },
-    {
-      "epoch": 2.305055698371894,
-      "grad_norm": 24.451356887817383,
-      "learning_rate": 3.356951436048765e-06,
-      "loss": 1.1336,
-      "step": 2690
-    },
-    {
-      "epoch": 2.3136246786632393,
-      "grad_norm": 29.073118209838867,
-      "learning_rate": 3.3155586933108027e-06,
-      "loss": 1.5349,
-      "step": 2700
-    },
-    {
-      "epoch": 2.3221936589545846,
-      "grad_norm": 40.77454376220703,
-      "learning_rate": 3.27416595057284e-06,
-      "loss": 1.0589,
-      "step": 2710
-    },
-    {
-      "epoch": 2.33076263924593,
-      "grad_norm": 88.82485961914062,
-      "learning_rate": 3.2327732078348777e-06,
-      "loss": 1.1665,
-      "step": 2720
-    },
-    {
-      "epoch": 2.3393316195372753,
-      "grad_norm": 43.5887336730957,
-      "learning_rate": 3.1913804650969148e-06,
-      "loss": 1.2165,
-      "step": 2730
-    },
-    {
-      "epoch": 2.34790059982862,
-      "grad_norm": 27.64069175720215,
-      "learning_rate": 3.1499877223589527e-06,
-      "loss": 1.144,
-      "step": 2740
-    },
-    {
-      "epoch": 2.3564695801199655,
-      "grad_norm": 20.499149322509766,
-      "learning_rate": 3.10859497962099e-06,
-      "loss": 1.1935,
-      "step": 2750
-    },
-    {
-      "epoch": 2.365038560411311,
-      "grad_norm": 34.570640563964844,
-      "learning_rate": 3.0672022368830273e-06,
-      "loss": 1.0865,
-      "step": 2760
-    },
-    {
-      "epoch": 2.3736075407026562,
-      "grad_norm": 26.797300338745117,
-      "learning_rate": 3.025809494145065e-06,
-      "loss": 1.336,
-      "step": 2770
-    },
-    {
-      "epoch": 2.3821765209940016,
-      "grad_norm": 37.89352035522461,
-      "learning_rate": 2.9844167514071023e-06,
-      "loss": 1.2351,
-      "step": 2780
-    },
-    {
-      "epoch": 2.390745501285347,
-      "grad_norm": 34.16749572753906,
-      "learning_rate": 2.9430240086691394e-06,
-      "loss": 1.3755,
-      "step": 2790
-    },
-    {
-      "epoch": 2.3993144815766922,
-      "grad_norm": 10.907548904418945,
-      "learning_rate": 2.901631265931177e-06,
-      "loss": 1.0435,
-      "step": 2800
-    },
-    {
-      "epoch": 2.4078834618680376,
-      "grad_norm": 34.773712158203125,
-      "learning_rate": 2.8602385231932144e-06,
-      "loss": 1.5544,
-      "step": 2810
-    },
-    {
-      "epoch": 2.416452442159383,
-      "grad_norm": 23.92730140686035,
-      "learning_rate": 2.818845780455252e-06,
-      "loss": 1.1912,
-      "step": 2820
-    },
-    {
-      "epoch": 2.4250214224507283,
-      "grad_norm": 41.49195861816406,
-      "learning_rate": 2.777453037717289e-06,
-      "loss": 1.5421,
-      "step": 2830
-    },
-    {
-      "epoch": 2.4335904027420736,
-      "grad_norm": 26.17296600341797,
-      "learning_rate": 2.7360602949793265e-06,
-      "loss": 1.3269,
-      "step": 2840
-    },
-    {
-      "epoch": 2.442159383033419,
-      "grad_norm": 38.39194869995117,
-      "learning_rate": 2.694667552241364e-06,
-      "loss": 1.2148,
-      "step": 2850
-    },
-    {
-      "epoch": 2.4507283633247643,
-      "grad_norm": 16.456998825073242,
-      "learning_rate": 2.6532748095034015e-06,
-      "loss": 1.2613,
-      "step": 2860
-    },
-    {
-      "epoch": 2.4592973436161096,
-      "grad_norm": 20.516918182373047,
-      "learning_rate": 2.611882066765439e-06,
-      "loss": 0.8369,
-      "step": 2870
-    },
-    {
-      "epoch": 2.467866323907455,
-      "grad_norm": 39.00393295288086,
-      "learning_rate": 2.570489324027476e-06,
-      "loss": 1.0123,
-      "step": 2880
-    },
-    {
-      "epoch": 2.4764353041988003,
-      "grad_norm": 55.170372009277344,
-      "learning_rate": 2.5290965812895136e-06,
-      "loss": 1.2133,
-      "step": 2890
-    },
-    {
-      "epoch": 2.4850042844901457,
-      "grad_norm": 33.54553985595703,
-      "learning_rate": 2.487703838551551e-06,
-      "loss": 1.1492,
-      "step": 2900
-    },
-    {
-      "epoch": 2.493573264781491,
-      "grad_norm": 36.33115768432617,
-      "learning_rate": 2.4463110958135886e-06,
-      "loss": 1.5454,
-      "step": 2910
-    },
-    {
-      "epoch": 2.5021422450728363,
-      "grad_norm": 22.1439266204834,
-      "learning_rate": 2.4049183530756257e-06,
-      "loss": 1.0734,
-      "step": 2920
-    },
-    {
-      "epoch": 2.5107112253641817,
-      "grad_norm": 13.999841690063477,
-      "learning_rate": 2.3635256103376636e-06,
-      "loss": 1.3985,
-      "step": 2930
-    },
-    {
-      "epoch": 2.519280205655527,
-      "grad_norm": 32.42293930053711,
-      "learning_rate": 2.3221328675997007e-06,
-      "loss": 1.0033,
-      "step": 2940
-    },
-    {
-      "epoch": 2.5278491859468724,
-      "grad_norm": 43.845096588134766,
-      "learning_rate": 2.280740124861738e-06,
-      "loss": 1.4406,
-      "step": 2950
-    },
-    {
-      "epoch": 2.5364181662382177,
-      "grad_norm": 35.83656692504883,
-      "learning_rate": 2.2393473821237757e-06,
-      "loss": 1.2356,
-      "step": 2960
-    },
-    {
-      "epoch": 2.544987146529563,
-      "grad_norm": 39.58152770996094,
-      "learning_rate": 2.197954639385813e-06,
-      "loss": 1.234,
-      "step": 2970
-    },
-    {
-      "epoch": 2.5535561268209084,
-      "grad_norm": 25.64226722717285,
-      "learning_rate": 2.1565618966478503e-06,
-      "loss": 1.2667,
-      "step": 2980
-    },
-    {
-      "epoch": 2.5621251071122537,
-      "grad_norm": 33.893375396728516,
-      "learning_rate": 2.115169153909888e-06,
-      "loss": 1.3049,
-      "step": 2990
-    },
-    {
-      "epoch": 2.570694087403599,
-      "grad_norm": 45.953590393066406,
-      "learning_rate": 2.0737764111719253e-06,
-      "loss": 1.4963,
-      "step": 3000
-    },
-    {
-      "epoch": 2.5792630676949444,
-      "grad_norm": 41.743202209472656,
-      "learning_rate": 2.0323836684339628e-06,
-      "loss": 1.2745,
-      "step": 3010
-    },
-    {
-      "epoch": 2.5878320479862897,
-      "grad_norm": 42.73529052734375,
-      "learning_rate": 1.9909909256960003e-06,
-      "loss": 1.4276,
-      "step": 3020
-    },
-    {
-      "epoch": 2.596401028277635,
-      "grad_norm": 17.132654190063477,
-      "learning_rate": 1.949598182958038e-06,
-      "loss": 0.9052,
-      "step": 3030
-    },
-    {
-      "epoch": 2.6049700085689804,
-      "grad_norm": 20.87359046936035,
-      "learning_rate": 1.908205440220075e-06,
-      "loss": 1.0161,
-      "step": 3040
-    },
-    {
-      "epoch": 2.6135389888603258,
-      "grad_norm": 20.73636245727539,
-      "learning_rate": 1.8668126974821126e-06,
-      "loss": 1.1655,
-      "step": 3050
-    },
-    {
-      "epoch": 2.622107969151671,
-      "grad_norm": 30.01190948486328,
-      "learning_rate": 1.8254199547441499e-06,
-      "loss": 1.3549,
-      "step": 3060
-    },
-    {
-      "epoch": 2.6306769494430164,
-      "grad_norm": 29.046810150146484,
-      "learning_rate": 1.7840272120061874e-06,
-      "loss": 1.2511,
-      "step": 3070
-    },
-    {
-      "epoch": 2.6392459297343613,
-      "grad_norm": 28.32902717590332,
-      "learning_rate": 1.7426344692682247e-06,
-      "loss": 1.3624,
-      "step": 3080
-    },
-    {
-      "epoch": 2.6478149100257067,
-      "grad_norm": 36.059173583984375,
-      "learning_rate": 1.701241726530262e-06,
-      "loss": 1.4086,
-      "step": 3090
-    },
-    {
-      "epoch": 2.656383890317052,
-      "grad_norm": 46.75105667114258,
-      "learning_rate": 1.6598489837922995e-06,
-      "loss": 1.3701,
-      "step": 3100
-    },
-    {
-      "epoch": 2.6649528706083974,
-      "grad_norm": 37.93562316894531,
-      "learning_rate": 1.618456241054337e-06,
-      "loss": 1.3166,
-      "step": 3110
-    },
-    {
-      "epoch": 2.6735218508997427,
-      "grad_norm": 47.81494140625,
-      "learning_rate": 1.5770634983163743e-06,
-      "loss": 1.0946,
-      "step": 3120
-    },
-    {
-      "epoch": 2.682090831191088,
-      "grad_norm": 36.22248077392578,
-      "learning_rate": 1.5356707555784118e-06,
-      "loss": 1.3498,
-      "step": 3130
-    },
-    {
-      "epoch": 2.6906598114824334,
-      "grad_norm": 38.44594955444336,
-      "learning_rate": 1.4942780128404493e-06,
-      "loss": 1.4325,
-      "step": 3140
-    },
-    {
-      "epoch": 2.6992287917737787,
-      "grad_norm": 50.71770477294922,
-      "learning_rate": 1.4528852701024866e-06,
-      "loss": 1.3387,
-      "step": 3150
-    },
-    {
-      "epoch": 2.707797772065124,
-      "grad_norm": 44.393768310546875,
-      "learning_rate": 1.411492527364524e-06,
-      "loss": 1.1321,
-      "step": 3160
-    },
-    {
-      "epoch": 2.7163667523564694,
-      "grad_norm": 29.496095657348633,
-      "learning_rate": 1.3700997846265616e-06,
-      "loss": 1.1442,
-      "step": 3170
-    },
-    {
-      "epoch": 2.7249357326478147,
-      "grad_norm": 21.904565811157227,
-      "learning_rate": 1.3287070418885989e-06,
-      "loss": 1.1811,
-      "step": 3180
-    },
-    {
-      "epoch": 2.73350471293916,
-      "grad_norm": 42.967010498046875,
-      "learning_rate": 1.2873142991506364e-06,
-      "loss": 1.195,
-      "step": 3190
-    },
-    {
-      "epoch": 2.7420736932305054,
-      "grad_norm": 41.519500732421875,
-      "learning_rate": 1.2459215564126739e-06,
-      "loss": 1.2147,
-      "step": 3200
-    },
-    {
-      "epoch": 2.7506426735218508,
-      "grad_norm": 57.64628219604492,
-      "learning_rate": 1.204528813674711e-06,
-      "loss": 1.2644,
-      "step": 3210
-    },
-    {
-      "epoch": 2.759211653813196,
-      "grad_norm": 75.4446792602539,
-      "learning_rate": 1.1631360709367485e-06,
-      "loss": 1.198,
-      "step": 3220
-    },
-    {
-      "epoch": 2.7677806341045414,
-      "grad_norm": 33.584678649902344,
-      "learning_rate": 1.121743328198786e-06,
-      "loss": 1.2919,
-      "step": 3230
-    },
-    {
-      "epoch": 2.776349614395887,
-      "grad_norm": 42.54429244995117,
-      "learning_rate": 1.0803505854608233e-06,
-      "loss": 1.3359,
-      "step": 3240
-    },
-    {
-      "epoch": 2.784918594687232,
-      "grad_norm": 39.373809814453125,
-      "learning_rate": 1.0389578427228608e-06,
-      "loss": 1.1131,
-      "step": 3250
-    },
-    {
-      "epoch": 2.7934875749785775,
-      "grad_norm": 20.756141662597656,
-      "learning_rate": 9.97565099984898e-07,
-      "loss": 1.0463,
-      "step": 3260
-    },
-    {
-      "epoch": 2.802056555269923,
-      "grad_norm": 37.37822341918945,
-      "learning_rate": 9.561723572469356e-07,
-      "loss": 1.2405,
-      "step": 3270
-    },
-    {
-      "epoch": 2.810625535561268,
-      "grad_norm": 41.48339080810547,
-      "learning_rate": 9.14779614508973e-07,
-      "loss": 1.3686,
-      "step": 3280
-    },
-    {
-      "epoch": 2.8191945158526135,
-      "grad_norm": 43.064998626708984,
-      "learning_rate": 8.733868717710105e-07,
-      "loss": 1.6151,
-      "step": 3290
-    },
-    {
-      "epoch": 2.827763496143959,
-      "grad_norm": 32.76227569580078,
-      "learning_rate": 8.319941290330479e-07,
-      "loss": 1.2218,
-      "step": 3300
-    },
-    {
-      "epoch": 2.836332476435304,
-      "grad_norm": 48.318965911865234,
-      "learning_rate": 7.906013862950853e-07,
-      "loss": 1.1332,
-      "step": 3310
-    },
-    {
-      "epoch": 2.8449014567266495,
-      "grad_norm": 76.0499496459961,
-      "learning_rate": 7.492086435571228e-07,
-      "loss": 1.388,
-      "step": 3320
-    },
-    {
-      "epoch": 2.853470437017995,
-      "grad_norm": 29.409503936767578,
-      "learning_rate": 7.078159008191602e-07,
-      "loss": 1.4353,
-      "step": 3330
-    },
-    {
-      "epoch": 2.86203941730934,
-      "grad_norm": 25.97176170349121,
-      "learning_rate": 6.664231580811976e-07,
-      "loss": 1.1936,
-      "step": 3340
-    },
-    {
-      "epoch": 2.8706083976006855,
-      "grad_norm": 31.24247932434082,
-      "learning_rate": 6.25030415343235e-07,
-      "loss": 0.9297,
-      "step": 3350
-    },
-    {
-      "epoch": 2.879177377892031,
-      "grad_norm": 44.379268646240234,
-      "learning_rate": 5.836376726052725e-07,
-      "loss": 1.1583,
-      "step": 3360
-    },
-    {
-      "epoch": 2.887746358183376,
-      "grad_norm": 35.18986511230469,
-      "learning_rate": 5.422449298673099e-07,
-      "loss": 1.3518,
-      "step": 3370
-    },
-    {
-      "epoch": 2.8963153384747216,
-      "grad_norm": 30.07046890258789,
-      "learning_rate": 5.008521871293473e-07,
-      "loss": 1.1271,
-      "step": 3380
-    },
-    {
-      "epoch": 2.904884318766067,
-      "grad_norm": 26.008453369140625,
-      "learning_rate": 4.594594443913846e-07,
-      "loss": 1.0143,
-      "step": 3390
-    },
-    {
-      "epoch": 2.9134532990574122,
-      "grad_norm": 32.31906509399414,
-      "learning_rate": 4.1806670165342206e-07,
-      "loss": 1.452,
-      "step": 3400
-    },
-    {
-      "epoch": 2.9220222793487576,
-      "grad_norm": 36.26763153076172,
-      "learning_rate": 3.766739589154595e-07,
-      "loss": 1.2661,
-      "step": 3410
-    },
-    {
-      "epoch": 2.930591259640103,
-      "grad_norm": 33.175926208496094,
-      "learning_rate": 3.352812161774969e-07,
-      "loss": 1.3136,
-      "step": 3420
-    },
-    {
-      "epoch": 2.9391602399314483,
-      "grad_norm": 37.862266540527344,
-      "learning_rate": 2.938884734395343e-07,
-      "loss": 1.1489,
-      "step": 3430
-    },
-    {
-      "epoch": 2.9477292202227936,
-      "grad_norm": 45.077796936035156,
-      "learning_rate": 2.5249573070157175e-07,
-      "loss": 1.1297,
-      "step": 3440
-    },
-    {
-      "epoch": 2.956298200514139,
-      "grad_norm": 32.95222854614258,
-      "learning_rate": 2.1110298796360915e-07,
-      "loss": 1.2034,
-      "step": 3450
-    },
-    {
-      "epoch": 2.9648671808054843,
-      "grad_norm": 43.23259735107422,
-      "learning_rate": 1.6971024522564658e-07,
-      "loss": 0.9911,
-      "step": 3460
-    },
-    {
-      "epoch": 2.9734361610968296,
-      "grad_norm": 30.152484893798828,
-      "learning_rate": 1.28317502487684e-07,
-      "loss": 1.0821,
-      "step": 3470
-    },
-    {
-      "epoch": 2.982005141388175,
-      "grad_norm": 30.485458374023438,
-      "learning_rate": 8.692475974972141e-08,
-      "loss": 0.9814,
-      "step": 3480
-    },
-    {
-      "epoch": 2.9905741216795203,
-      "grad_norm": 30.098020553588867,
-      "learning_rate": 4.553201701175884e-08,
-      "loss": 1.0793,
-      "step": 3490
-    },
-    {
-      "epoch": 2.9991431019708656,
-      "grad_norm": 70.42772674560547,
-      "learning_rate": 4.139274273796258e-09,
-      "loss": 1.4007,
-      "step": 3500
-    },
-    {
-      "epoch": 3.0,
-      "eval_classification_report": {
-        "accuracy": 0.3645,
-        "ar": {
-          "f1-score": 0.3172043010752688,
-          "precision": 0.35542168674698793,
-          "recall": 0.28640776699029125,
-          "support": 206.0
-        },
-        "cl": {
-          "f1-score": 0.23825503355704697,
-          "precision": 0.23202614379084968,
-          "recall": 0.24482758620689654,
-          "support": 290.0
-        },
-        "co": {
-          "f1-score": 0.39937106918238996,
-          "precision": 0.3681159420289855,
-          "recall": 0.436426116838488,
-          "support": 291.0
-        },
-        "es": {
-          "f1-score": 0.35080645161290325,
-          "precision": 0.4009216589861751,
-          "recall": 0.3118279569892473,
-          "support": 279.0
-        },
-        "macro avg": {
-          "f1-score": 0.34262062899860507,
-          "precision": 0.3462253190782673,
-          "recall": 0.34320174371834444,
-          "support": 2000.0
-        },
-        "mx": {
-          "f1-score": 0.4259567387687188,
-          "precision": 0.4129032258064516,
-          "recall": 0.43986254295532645,
-          "support": 291.0
-        },
-        "pe": {
-          "f1-score": 0.3543913713405239,
-          "precision": 0.32122905027932963,
-          "recall": 0.3951890034364261,
-          "support": 291.0
-        },
-        "pr": {
-          "f1-score": 0.6305418719211823,
-          "precision": 0.6274509803921569,
-          "recall": 0.6336633663366337,
-          "support": 101.0
-        },
-        "uy": {
-          "f1-score": 0.36705882352941177,
-          "precision": 0.3979591836734694,
-          "recall": 0.3406113537117904,
-          "support": 229.0
-        },
-        "ve": {
-          "f1-score": 0.0,
-          "precision": 0.0,
-          "recall": 0.0,
-          "support": 22.0
-        },
-        "weighted avg": {
-          "f1-score": 0.36167626328959446,
-          "precision": 0.3638105127892991,
-          "recall": 0.3645,
-          "support": 2000.0
-        }
-      },
-      "eval_f1": 0.34262062899860507,
-      "eval_loss": 1.8317162990570068,
-      "eval_runtime": 5.6289,
-      "eval_samples_per_second": 355.31,
-      "eval_steps_per_second": 88.828,
-      "step": 3501
-    }
-  ],
-  "logging_steps": 10,
-  "max_steps": 3501,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
-  "save_steps": 500,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": true
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 460407503990016.0,
-  "train_batch_size": 4,
-  "trial_name": null,
-  "trial_params": null
-}

trial_5/checkpoint-438/model.safetensors CHANGED Viewed

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:455a1836d4bc83ba7b52be8311c689a7970d706e2bf7247a151c27698f8cd7b1
-size 439454740

trial_5/checkpoint-438/rng_state.pth CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fd04b482a544fd0ce40d1f3fc9b79e64e3aac20c06edc3861fb06060a0bfbc6b
 size 14244

 version https://git-lfs.github.com/spec/v1
+oid sha256:9266ef30f9c03ba585bd4567285194480552bc80c185cae7eee533c54caf9ebb
 size 14244

trial_5/checkpoint-438/trainer_state.json CHANGED Viewed

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.24538914841155643,
-  "best_model_checkpoint": "/content/drive/MyDrive/model_outputs/trial_5/checkpoint-438",
   "epoch": 3.0,
   "eval_steps": 500,
   "global_step": 438,
@@ -464,84 +464,6 @@
       "learning_rate": 2.0422444096166686e-07,
       "loss": 1.8367,
       "step": 430
-    },
-    {
-      "epoch": 3.0,
-      "eval_classification_report": {
-        "accuracy": 0.2665,
-        "ar": {
-          "f1-score": 0.10077519379844961,
-          "precision": 0.25,
-          "recall": 0.06310679611650485,
-          "support": 206.0
-        },
-        "cl": {
-          "f1-score": 0.19504132231404958,
-          "precision": 0.1873015873015873,
-          "recall": 0.20344827586206896,
-          "support": 290.0
-        },
-        "co": {
-          "f1-score": 0.3190348525469169,
-          "precision": 0.26153846153846155,
-          "recall": 0.40893470790378006,
-          "support": 291.0
-        },
-        "es": {
-          "f1-score": 0.3266563944530046,
-          "precision": 0.2864864864864865,
-          "recall": 0.37992831541218636,
-          "support": 279.0
-        },
-        "macro avg": {
-          "f1-score": 0.24538914841155643,
-          "precision": 0.2739309736044076,
-          "recall": 0.24660703914377965,
-          "support": 2000.0
-        },
-        "mx": {
-          "f1-score": 0.2955223880597015,
-          "precision": 0.2612137203166227,
-          "recall": 0.3402061855670103,
-          "support": 291.0
-        },
-        "pe": {
-          "f1-score": 0.18181818181818182,
-          "precision": 0.1975806451612903,
-          "recall": 0.16838487972508592,
-          "support": 291.0
-        },
-        "pr": {
-          "f1-score": 0.5568181818181818,
-          "precision": 0.6533333333333333,
-          "recall": 0.48514851485148514,
-          "support": 101.0
-        },
-        "uy": {
-          "f1-score": 0.23283582089552238,
-          "precision": 0.36792452830188677,
-          "recall": 0.1703056768558952,
-          "support": 229.0
-        },
-        "ve": {
-          "f1-score": 0.0,
-          "precision": 0.0,
-          "recall": 0.0,
-          "support": 22.0
-        },
-        "weighted avg": {
-          "f1-score": 0.25488104736013556,
-          "precision": 0.27280271317837684,
-          "recall": 0.2665,
-          "support": 2000.0
-        }
-      },
-      "eval_f1": 0.24538914841155643,
-      "eval_loss": 1.9040703773498535,
-      "eval_runtime": 3.6694,
-      "eval_samples_per_second": 545.044,
-      "eval_steps_per_second": 17.169,
-      "step": 438
     }
   ],
   "logging_steps": 10,

 {
+  "best_metric": 0.23586615753847973,
+  "best_model_checkpoint": "/content/drive/MyDrive/model_outputs/trial_5/checkpoint-292",
   "epoch": 3.0,
   "eval_steps": 500,
   "global_step": 438,
       "learning_rate": 2.0422444096166686e-07,
       "loss": 1.8367,
       "step": 430
     }
   ],
   "logging_steps": 10,