End of training

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 2.0,
-    "train_loss": 0.6001299892563418,
-    "train_runtime": 234.7434,
+    "epoch": 3.0,
+    "train_loss": 0.5543336676785267,
+    "train_runtime": 347.758,
     "train_samples": 7964,
-    "train_samples_per_second": 67.853,
-    "train_steps_per_second": 2.829
+    "train_samples_per_second": 68.703,
+    "train_steps_per_second": 2.864
 }

predict_results_spanish-offensive-language.txt

@@ -1,16 +1,16 @@
 index	prediction
-0	1
+0	0
 1	1
 2	0
 3	1
 4	0
 5	1
-6	1
+6	0
 7	1
 8	1
 9	1
 10	1
-11	1
+11	0
 12	1
 13	0
 14	1
@@ -24,10 +24,10 @@ index	prediction
 22	1
 23	1
 24	0
-25	1
+25	0
 26	1
 27	1
-28	1
+28	0
 29	0
 30	0
 31	1
@@ -60,7 +60,7 @@ index	prediction
 58	1
 59	1
 60	1
-61	1
+61	0
 62	1
 63	0
 64	1
@@ -72,7 +72,7 @@ index	prediction
 70	1
 71	1
 72	1
-73	1
+73	0
 74	0
 75	1
 76	1
@@ -91,7 +91,7 @@ index	prediction
 89	1
 90	1
 91	1
-92	1
+92	0
 93	1
 94	1
 95	1
@@ -120,11 +120,11 @@ index	prediction
 118	1
 119	0
 120	1
-121	0
+121	1
 122	0
 123	1
 124	1
-125	1
+125	0
 126	1
 127	1
 128	1
@@ -159,7 +159,7 @@ index	prediction
 157	1
 158	1
 159	1
-160	0
+160	1
 161	1
 162	1
 163	1
@@ -175,11 +175,11 @@ index	prediction
 173	1
 174	1
 175	1
-176	1
+176	0
 177	0
 178	1
-179	0
-180	1
+179	1
+180	0
 181	1
 182	1
 183	1
@@ -212,9 +212,9 @@ index	prediction
 210	1
 211	1
 212	1
-213	1
+213	0
 214	1
-215	0
+215	1
 216	0
 217	0
 218	1
@@ -227,11 +227,11 @@ index	prediction
 225	1
 226	1
 227	1
-228	1
+228	0
 229	0
 230	1
 231	1
-232	1
+232	0
 233	1
 234	0
 235	0
@@ -251,7 +251,7 @@ index	prediction
 249	1
 250	1
 251	1
-252	1
+252	0
 253	1
 254	1
 255	0
@@ -259,7 +259,7 @@ index	prediction
 257	1
 258	1
 259	0
-260	1
+260	0
 261	1
 262	0
 263	1
@@ -271,7 +271,7 @@ index	prediction
 269	1
 270	1
 271	1
-272	1
+272	0
 273	1
 274	1
 275	1
@@ -297,8 +297,8 @@ index	prediction
 295	0
 296	1
 297	1
-298	1
-299	1
+298	0
+299	0
 300	1
 301	1
 302	0
@@ -314,7 +314,7 @@ index	prediction
 312	1
 313	1
 314	1
-315	1
+315	0
 316	1
 317	1
 318	1
@@ -330,7 +330,7 @@ index	prediction
 328	1
 329	0
 330	1
-331	1
+331	0
 332	1
 333	1
 334	1
@@ -340,7 +340,7 @@ index	prediction
 338	0
 339	0
 340	1
-341	1
+341	0
 342	0
 343	0
 344	1
@@ -358,16 +358,16 @@ index	prediction
 356	0
 357	1
 358	1
-359	1
+359	0
 360	1
 361	1
 362	1
 363	1
 364	1
-365	1
-366	1
+365	0
+366	0
 367	1
-368	1
+368	0
 369	1
 370	1
 371	1
@@ -376,7 +376,7 @@ index	prediction
 374	0
 375	1
 376	1
-377	1
+377	0
 378	1
 379	1
 380	0
@@ -388,7 +388,7 @@ index	prediction
 386	1
 387	1
 388	1
-389	0
+389	1
 390	1
 391	0
 392	1
@@ -396,7 +396,7 @@ index	prediction
 394	1
 395	1
 396	1
-397	0
+397	1
 398	1
 399	1
 400	1
@@ -406,18 +406,18 @@ index	prediction
 404	0
 405	1
 406	0
-407	1
+407	0
 408	1
 409	1
 410	0
 411	0
 412	1
 413	1
-414	1
+414	0
 415	1
 416	0
-417	1
-418	0
+417	0
+418	1
 419	0
 420	1
 421	1
@@ -427,7 +427,7 @@ index	prediction
 425	1
 426	1
 427	1
-428	1
+428	0
 429	0
 430	1
 431	1
@@ -446,7 +446,7 @@ index	prediction
 444	1
 445	0
 446	1
-447	1
+447	0
 448	1
 449	1
 450	1
@@ -480,14 +480,14 @@ index	prediction
 478	1
 479	1
 480	0
-481	1
+481	0
 482	1
 483	1
 484	1
 485	1
-486	1
+486	0
 487	1
-488	1
+488	0
 489	1
 490	1
 491	1
@@ -507,13 +507,13 @@ index	prediction
 505	1
 506	1
 507	1
-508	1
+508	0
 509	1
 510	1
 511	0
 512	1
 513	1
-514	1
+514	0
 515	1
 516	1
 517	1
@@ -541,7 +541,7 @@ index	prediction
 539	1
 540	1
 541	1
-542	1
+542	0
 543	1
 544	0
 545	0
@@ -558,7 +558,7 @@ index	prediction
 556	0
 557	1
 558	1
-559	1
+559	0
 560	1
 561	1
 562	0
@@ -567,22 +567,22 @@ index	prediction
 565	1
 566	1
 567	1
-568	0
+568	1
 569	1
 570	0
 571	1
 572	1
-573	0
+573	1
 574	1
 575	1
 576	1
 577	1
 578	1
-579	1
+579	0
 580	1
 581	1
 582	0
-583	1
+583	0
 584	1
 585	1
 586	1
@@ -611,14 +611,14 @@ index	prediction
 609	1
 610	0
 611	1
-612	1
+612	0
 613	1
 614	1
 615	1
 616	1
 617	1
 618	1
-619	1
+619	0
 620	1
 621	1
 622	1
@@ -633,7 +633,7 @@ index	prediction
 631	1
 632	1
 633	1
-634	1
+634	0
 635	1
 636	1
 637	1
@@ -644,7 +644,7 @@ index	prediction
 642	1
 643	1
 644	1
-645	1
+645	0
 646	1
 647	1
 648	0
@@ -657,12 +657,12 @@ index	prediction
 655	1
 656	1
 657	1
-658	1
-659	1
+658	0
+659	0
 660	1
 661	1
 662	1
-663	1
+663	0
 664	1
 665	1
 666	1
@@ -697,13 +697,13 @@ index	prediction
 695	1
 696	1
 697	1
-698	1
+698	0
 699	1
 700	1
 701	1
 702	1
 703	0
-704	1
+704	0
 705	1
 706	0
 707	1
@@ -715,7 +715,7 @@ index	prediction
 713	1
 714	1
 715	1
-716	1
+716	0
 717	1
 718	1
 719	1
@@ -754,11 +754,11 @@ index	prediction
 752	1
 753	1
 754	1
-755	1
+755	0
 756	1
 757	1
 758	1
-759	1
+759	0
 760	1
 761	1
 762	0
@@ -772,7 +772,7 @@ index	prediction
 770	0
 771	1
 772	1
-773	1
+773	0
 774	0
 775	1
 776	0
@@ -780,13 +780,13 @@ index	prediction
 778	1
 779	1
 780	1
-781	0
+781	1
 782	1
-783	1
+783	0
 784	1
 785	1
 786	1
-787	1
+787	0
 788	1
 789	1
 790	1
@@ -829,9 +829,9 @@ index	prediction
 827	1
 828	1
 829	1
-830	0
+830	1
 831	1
-832	1
+832	0
 833	1
 834	1
 835	1
@@ -865,7 +865,7 @@ index	prediction
 863	1
 864	1
 865	1
-866	1
+866	0
 867	0
 868	0
 869	1
@@ -877,7 +877,7 @@ index	prediction
 875	1
 876	1
 877	1
-878	1
+878	0
 879	0
 880	0
 881	1

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 2.0,
-    "train_loss": 0.6001299892563418,
-    "train_runtime": 234.7434,
+    "epoch": 3.0,
+    "train_loss": 0.5543336676785267,
+    "train_runtime": 347.758,
     "train_samples": 7964,
-    "train_samples_per_second": 67.853,
-    "train_steps_per_second": 2.829
+    "train_samples_per_second": 68.703,
+    "train_steps_per_second": 2.864
 }

trainer_state.json

@@ -1,55 +1,70 @@
 {
-  "best_metric": 0.5963656236116517,
-  "best_model_checkpoint": "models/spanish-offensive-language-bert-base-spanish-wwm-cased/checkpoint-664",
-  "epoch": 2.0,
-  "global_step": 664,
+  "best_metric": 0.6014237329035705,
+  "best_model_checkpoint": "models/spanish-offensive-language-bert-base-spanish-wwm-cased/checkpoint-996",
+  "epoch": 3.0,
+  "global_step": 996,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 1.0,
-      "learning_rate": 1e-05,
-      "loss": 0.6339,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 0.6335,
       "step": 332
     },
     {
       "epoch": 1.0,
-      "eval_f1": 0.502414198094741,
-      "eval_loss": 0.6030057668685913,
-      "eval_runtime": 5.4442,
-      "eval_samples_per_second": 180.743,
-      "eval_steps_per_second": 7.531,
+      "eval_f1": 0.5017770187437626,
+      "eval_loss": 0.6012123823165894,
+      "eval_runtime": 5.4247,
+      "eval_samples_per_second": 181.393,
+      "eval_steps_per_second": 7.558,
       "step": 332
     },
     {
       "epoch": 2.0,
-      "learning_rate": 0.0,
-      "loss": 0.5663,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 0.5659,
       "step": 664
     },
     {
       "epoch": 2.0,
-      "eval_f1": 0.5963656236116517,
-      "eval_loss": 0.5984072685241699,
-      "eval_runtime": 5.4203,
-      "eval_samples_per_second": 181.541,
-      "eval_steps_per_second": 7.564,
+      "eval_f1": 0.5776452050829993,
+      "eval_loss": 0.5975989103317261,
+      "eval_runtime": 5.271,
+      "eval_samples_per_second": 186.682,
+      "eval_steps_per_second": 7.778,
       "step": 664
     },
     {
-      "epoch": 2.0,
-      "step": 664,
-      "total_flos": 4190832889774080.0,
-      "train_loss": 0.6001299892563418,
-      "train_runtime": 234.7434,
-      "train_samples_per_second": 67.853,
-      "train_steps_per_second": 2.829
+      "epoch": 3.0,
+      "learning_rate": 0.0,
+      "loss": 0.4636,
+      "step": 996
+    },
+    {
+      "epoch": 3.0,
+      "eval_f1": 0.6014237329035705,
+      "eval_loss": 0.6521763801574707,
+      "eval_runtime": 5.4331,
+      "eval_samples_per_second": 181.113,
+      "eval_steps_per_second": 7.546,
+      "step": 996
+    },
+    {
+      "epoch": 3.0,
+      "step": 996,
+      "total_flos": 6286249334661120.0,
+      "train_loss": 0.5543336676785267,
+      "train_runtime": 347.758,
+      "train_samples_per_second": 68.703,
+      "train_steps_per_second": 2.864
     }
   ],
-  "max_steps": 664,
-  "num_train_epochs": 2,
-  "total_flos": 4190832889774080.0,
+  "max_steps": 996,
+  "num_train_epochs": 3,
+  "total_flos": 6286249334661120.0,
   "trial_name": null,
   "trial_params": null
 }