End of training

README.md

@@ -2,6 +2,8 @@
 license: apache-2.0
 base_model: facebook/wav2vec2-large-xlsr-53
 tags:
+- automatic-speech-recognition
+- ./data-configs/btb-cv-other.json
 - generated_from_trainer
 metrics:
 - wer

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 1.2075471698113207,
-    "eval_loss": 0.49337518215179443,
-    "eval_runtime": 195.3217,
-    "eval_samples": 7022,
-    "eval_samples_per_second": 35.951,
-    "eval_steps_per_second": 0.563,
-    "eval_wer": 0.37760267874938547,
-    "total_flos": 9.085819982279884e+18,
-    "train_loss": 0.9942695980072022,
-    "train_runtime": 6320.527,
-    "train_samples": 52993,
-    "train_samples_per_second": 10.126,
-    "train_steps_per_second": 1.266
+    "epoch": 1.0289389067524115,
+    "eval_loss": Infinity,
+    "eval_runtime": 105.9708,
+    "eval_samples": 3901,
+    "eval_samples_per_second": 36.812,
+    "eval_steps_per_second": 0.576,
+    "eval_wer": 0.37269960875235475,
+    "total_flos": 8.75110352693014e+18,
+    "train_loss": 1.0117860870361328,
+    "train_runtime": 5013.4693,
+    "train_samples": 62198,
+    "train_samples_per_second": 12.766,
+    "train_steps_per_second": 1.596
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 1.2075471698113207,
-    "eval_loss": 0.49337518215179443,
-    "eval_runtime": 195.3217,
-    "eval_samples": 7022,
-    "eval_samples_per_second": 35.951,
-    "eval_steps_per_second": 0.563,
-    "eval_wer": 0.37760267874938547
+    "epoch": 1.0289389067524115,
+    "eval_loss": Infinity,
+    "eval_runtime": 105.9708,
+    "eval_samples": 3901,
+    "eval_samples_per_second": 36.812,
+    "eval_steps_per_second": 0.576,
+    "eval_wer": 0.37269960875235475
 }

runs/Oct01_17-04-21_455c32bffda2/events.out.tfevents.1727805382.455c32bffda2.35.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d75edd8d5aaa92f813a59c6f063b36ea4902ae5ecd72e5dd6c227790733a643
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 1.2075471698113207,
-    "total_flos": 9.085819982279884e+18,
-    "train_loss": 0.9942695980072022,
-    "train_runtime": 6320.527,
-    "train_samples": 52993,
-    "train_samples_per_second": 10.126,
-    "train_steps_per_second": 1.266
+    "epoch": 1.0289389067524115,
+    "total_flos": 8.75110352693014e+18,
+    "train_loss": 1.0117860870361328,
+    "train_runtime": 5013.4693,
+    "train_samples": 62198,
+    "train_samples_per_second": 12.766,
+    "train_steps_per_second": 1.596
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.2075471698113207,
+  "epoch": 1.0289389067524115,
   "eval_steps": 500,
   "global_step": 8000,
   "is_hyper_param_search": false,
@@ -9,269 +9,269 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.07547169811320754,
-      "grad_norm": 9.590713500976562,
-      "learning_rate": 0.00018449999999999999,
-      "loss": 5.1054,
+      "epoch": 0.06430868167202572,
+      "grad_norm": 6.638989448547363,
+      "learning_rate": 0.00018487499999999998,
+      "loss": 4.9036,
       "step": 500
     },
     {
-      "epoch": 0.07547169811320754,
-      "eval_loss": 2.816082239151001,
-      "eval_runtime": 195.1014,
-      "eval_samples_per_second": 35.992,
-      "eval_steps_per_second": 0.564,
-      "eval_wer": 1.0,
+      "epoch": 0.06430868167202572,
+      "eval_loss": Infinity,
+      "eval_runtime": 109.3576,
+      "eval_samples_per_second": 35.672,
+      "eval_steps_per_second": 0.558,
+      "eval_wer": 0.9911607013476308,
       "step": 500
     },
     {
-      "epoch": 0.1509433962264151,
-      "grad_norm": 6.886263847351074,
-      "learning_rate": 0.00029204166666666663,
-      "loss": 1.4196,
+      "epoch": 0.12861736334405144,
+      "grad_norm": 5.176253795623779,
+      "learning_rate": 0.000292,
+      "loss": 1.3889,
       "step": 1000
     },
     {
-      "epoch": 0.1509433962264151,
-      "eval_loss": 1.168688178062439,
-      "eval_runtime": 193.3054,
-      "eval_samples_per_second": 36.326,
-      "eval_steps_per_second": 0.569,
-      "eval_wer": 0.7941110033351493,
+      "epoch": 0.12861736334405144,
+      "eval_loss": Infinity,
+      "eval_runtime": 104.5051,
+      "eval_samples_per_second": 37.328,
+      "eval_steps_per_second": 0.584,
+      "eval_wer": 0.793387431773173,
       "step": 1000
     },
     {
-      "epoch": 0.22641509433962265,
-      "grad_norm": 3.8018798828125,
-      "learning_rate": 0.00027124999999999995,
-      "loss": 1.0028,
+      "epoch": 0.19292604501607716,
+      "grad_norm": 8.991605758666992,
+      "learning_rate": 0.0002711666666666666,
+      "loss": 1.0207,
       "step": 1500
     },
     {
-      "epoch": 0.22641509433962265,
-      "eval_loss": 0.9889835119247437,
-      "eval_runtime": 194.0625,
-      "eval_samples_per_second": 36.184,
-      "eval_steps_per_second": 0.567,
-      "eval_wer": 0.6952523950623846,
+      "epoch": 0.19292604501607716,
+      "eval_loss": Infinity,
+      "eval_runtime": 104.6871,
+      "eval_samples_per_second": 37.263,
+      "eval_steps_per_second": 0.583,
+      "eval_wer": 0.6828961986185577,
       "step": 1500
     },
     {
-      "epoch": 0.3018867924528302,
-      "grad_norm": 9.146143913269043,
-      "learning_rate": 0.00025041666666666664,
-      "loss": 0.8942,
+      "epoch": 0.2572347266881029,
+      "grad_norm": 5.992569446563721,
+      "learning_rate": 0.000250375,
+      "loss": 0.8901,
       "step": 2000
     },
     {
-      "epoch": 0.3018867924528302,
-      "eval_loss": 0.8935320973396301,
-      "eval_runtime": 191.4777,
-      "eval_samples_per_second": 36.673,
-      "eval_steps_per_second": 0.574,
-      "eval_wer": 0.6194873702812953,
+      "epoch": 0.2572347266881029,
+      "eval_loss": Infinity,
+      "eval_runtime": 104.7623,
+      "eval_samples_per_second": 37.237,
+      "eval_steps_per_second": 0.582,
+      "eval_wer": 0.6450031396415978,
       "step": 2000
     },
     {
-      "epoch": 0.37735849056603776,
-      "grad_norm": 5.679136276245117,
-      "learning_rate": 0.0002295833333333333,
-      "loss": 0.8285,
+      "epoch": 0.3215434083601286,
+      "grad_norm": 5.538565635681152,
+      "learning_rate": 0.00022954166666666666,
+      "loss": 0.8467,
       "step": 2500
     },
     {
-      "epoch": 0.37735849056603776,
-      "eval_loss": 0.8221403360366821,
-      "eval_runtime": 191.0145,
-      "eval_samples_per_second": 36.762,
-      "eval_steps_per_second": 0.576,
-      "eval_wer": 0.6075419551150029,
+      "epoch": 0.3215434083601286,
+      "eval_loss": Infinity,
+      "eval_runtime": 105.5938,
+      "eval_samples_per_second": 36.943,
+      "eval_steps_per_second": 0.578,
+      "eval_wer": 0.5457904651499783,
       "step": 2500
     },
     {
-      "epoch": 0.4528301886792453,
-      "grad_norm": 7.3102006912231445,
-      "learning_rate": 0.00020879166666666664,
-      "loss": 0.763,
+      "epoch": 0.3858520900321543,
+      "grad_norm": 10.923360824584961,
+      "learning_rate": 0.00020874999999999998,
+      "loss": 0.7813,
       "step": 3000
     },
     {
-      "epoch": 0.4528301886792453,
-      "eval_loss": 0.7164860963821411,
-      "eval_runtime": 191.5603,
-      "eval_samples_per_second": 36.657,
-      "eval_steps_per_second": 0.574,
-      "eval_wer": 0.5306873596513374,
+      "epoch": 0.3858520900321543,
+      "eval_loss": Infinity,
+      "eval_runtime": 107.2937,
+      "eval_samples_per_second": 36.358,
+      "eval_steps_per_second": 0.569,
+      "eval_wer": 0.5366855045162536,
       "step": 3000
     },
     {
-      "epoch": 0.5283018867924528,
-      "grad_norm": 4.605156898498535,
-      "learning_rate": 0.00018795833333333333,
-      "loss": 0.7203,
+      "epoch": 0.45016077170418006,
+      "grad_norm": 8.696784973144531,
+      "learning_rate": 0.00018791666666666664,
+      "loss": 0.7456,
       "step": 3500
     },
     {
-      "epoch": 0.5283018867924528,
-      "eval_loss": 0.6891956925392151,
-      "eval_runtime": 191.0308,
-      "eval_samples_per_second": 36.758,
-      "eval_steps_per_second": 0.576,
-      "eval_wer": 0.5053747724524642,
+      "epoch": 0.45016077170418006,
+      "eval_loss": Infinity,
+      "eval_runtime": 105.1616,
+      "eval_samples_per_second": 37.095,
+      "eval_steps_per_second": 0.58,
+      "eval_wer": 0.5156016036323238,
       "step": 3500
     },
     {
-      "epoch": 0.6037735849056604,
-      "grad_norm": 4.3389763832092285,
-      "learning_rate": 0.000167125,
-      "loss": 0.7051,
+      "epoch": 0.5144694533762058,
+      "grad_norm": 25.500185012817383,
+      "learning_rate": 0.00016708333333333332,
+      "loss": 0.7312,
       "step": 4000
     },
     {
-      "epoch": 0.6037735849056604,
-      "eval_loss": 0.6848454475402832,
-      "eval_runtime": 192.1753,
-      "eval_samples_per_second": 36.54,
-      "eval_steps_per_second": 0.572,
-      "eval_wer": 0.5069825535816315,
+      "epoch": 0.5144694533762058,
+      "eval_loss": Infinity,
+      "eval_runtime": 106.9175,
+      "eval_samples_per_second": 36.486,
+      "eval_steps_per_second": 0.571,
+      "eval_wer": 0.48007535139834806,
       "step": 4000
     },
     {
-      "epoch": 0.6792452830188679,
-      "grad_norm": 5.629897117614746,
+      "epoch": 0.5787781350482315,
+      "grad_norm": 4.240171909332275,
       "learning_rate": 0.00014629166666666667,
-      "loss": 0.6568,
+      "loss": 0.6929,
       "step": 4500
     },
     {
-      "epoch": 0.6792452830188679,
-      "eval_loss": 0.6341728568077087,
-      "eval_runtime": 190.6714,
-      "eval_samples_per_second": 36.828,
-      "eval_steps_per_second": 0.577,
-      "eval_wer": 0.4925789606558684,
+      "epoch": 0.5787781350482315,
+      "eval_loss": Infinity,
+      "eval_runtime": 107.3751,
+      "eval_samples_per_second": 36.331,
+      "eval_steps_per_second": 0.568,
+      "eval_wer": 0.4659469642080858,
       "step": 4500
     },
     {
-      "epoch": 0.7547169811320755,
-      "grad_norm": 5.2013421058654785,
+      "epoch": 0.6430868167202572,
+      "grad_norm": 7.269958019256592,
       "learning_rate": 0.0001255,
-      "loss": 0.6315,
+      "loss": 0.6567,
       "step": 5000
     },
     {
-      "epoch": 0.7547169811320755,
-      "eval_loss": 0.59560626745224,
-      "eval_runtime": 192.113,
-      "eval_samples_per_second": 36.551,
-      "eval_steps_per_second": 0.573,
-      "eval_wer": 0.4493947567732763,
+      "epoch": 0.6430868167202572,
+      "eval_loss": Infinity,
+      "eval_runtime": 106.0885,
+      "eval_samples_per_second": 36.771,
+      "eval_steps_per_second": 0.575,
+      "eval_wer": 0.44049171617639954,
       "step": 5000
     },
     {
-      "epoch": 0.8301886792452831,
-      "grad_norm": 6.034324645996094,
+      "epoch": 0.707395498392283,
+      "grad_norm": 9.208508491516113,
       "learning_rate": 0.00010466666666666667,
-      "loss": 0.6171,
+      "loss": 0.6413,
       "step": 5500
     },
     {
-      "epoch": 0.8301886792452831,
-      "eval_loss": 0.5532522201538086,
-      "eval_runtime": 191.0439,
-      "eval_samples_per_second": 36.756,
-      "eval_steps_per_second": 0.576,
-      "eval_wer": 0.4304601443016782,
+      "epoch": 0.707395498392283,
+      "eval_loss": Infinity,
+      "eval_runtime": 107.3705,
+      "eval_samples_per_second": 36.332,
+      "eval_steps_per_second": 0.568,
+      "eval_wer": 0.4199391392551804,
       "step": 5500
     },
     {
-      "epoch": 0.9056603773584906,
-      "grad_norm": 11.971189498901367,
-      "learning_rate": 8.383333333333333e-05,
-      "loss": 0.5717,
+      "epoch": 0.7717041800643086,
+      "grad_norm": 12.485828399658203,
+      "learning_rate": 8.387499999999999e-05,
+      "loss": 0.6187,
       "step": 6000
     },
     {
-      "epoch": 0.9056603773584906,
-      "eval_loss": 0.5359871983528137,
-      "eval_runtime": 188.0002,
-      "eval_samples_per_second": 37.351,
-      "eval_steps_per_second": 0.585,
-      "eval_wer": 0.42121208094712925,
+      "epoch": 0.7717041800643086,
+      "eval_loss": Infinity,
+      "eval_runtime": 105.1802,
+      "eval_samples_per_second": 37.089,
+      "eval_steps_per_second": 0.58,
+      "eval_wer": 0.41023040139110273,
       "step": 6000
     },
     {
-      "epoch": 0.9811320754716981,
-      "grad_norm": 13.140325546264648,
+      "epoch": 0.8360128617363344,
+      "grad_norm": 2.41277813911438,
       "learning_rate": 6.304166666666666e-05,
-      "loss": 0.5699,
+      "loss": 0.6086,
       "step": 6500
     },
     {
-      "epoch": 0.9811320754716981,
-      "eval_loss": 0.518449068069458,
-      "eval_runtime": 186.6376,
-      "eval_samples_per_second": 37.624,
-      "eval_steps_per_second": 0.589,
-      "eval_wer": 0.4039516868414409,
+      "epoch": 0.8360128617363344,
+      "eval_loss": Infinity,
+      "eval_runtime": 105.1613,
+      "eval_samples_per_second": 37.095,
+      "eval_steps_per_second": 0.58,
+      "eval_wer": 0.40366130512486115,
       "step": 6500
     },
     {
-      "epoch": 1.0566037735849056,
-      "grad_norm": 1.923995018005371,
+      "epoch": 0.9003215434083601,
+      "grad_norm": 6.724209308624268,
       "learning_rate": 4.2208333333333334e-05,
-      "loss": 0.4905,
+      "loss": 0.585,
       "step": 7000
     },
     {
-      "epoch": 1.0566037735849056,
-      "eval_loss": 0.5081383585929871,
-      "eval_runtime": 187.1769,
-      "eval_samples_per_second": 37.515,
-      "eval_steps_per_second": 0.588,
-      "eval_wer": 0.39667016569446845,
+      "epoch": 0.9003215434083601,
+      "eval_loss": Infinity,
+      "eval_runtime": 107.3364,
+      "eval_samples_per_second": 36.344,
+      "eval_steps_per_second": 0.568,
+      "eval_wer": 0.38716611119161476,
       "step": 7000
     },
     {
-      "epoch": 1.1320754716981132,
-      "grad_norm": 4.82823371887207,
+      "epoch": 0.9646302250803859,
+      "grad_norm": 7.223489761352539,
       "learning_rate": 2.1374999999999998e-05,
-      "loss": 0.4706,
+      "loss": 0.5509,
       "step": 7500
     },
     {
-      "epoch": 1.1320754716981132,
-      "eval_loss": 0.49912941455841064,
-      "eval_runtime": 187.5036,
-      "eval_samples_per_second": 37.45,
-      "eval_steps_per_second": 0.587,
-      "eval_wer": 0.3825323217156752,
+      "epoch": 0.9646302250803859,
+      "eval_loss": Infinity,
+      "eval_runtime": 107.0445,
+      "eval_samples_per_second": 36.443,
+      "eval_steps_per_second": 0.57,
+      "eval_wer": 0.3831570303820702,
       "step": 7500
     },
     {
-      "epoch": 1.2075471698113207,
-      "grad_norm": 2.875157594680786,
-      "learning_rate": 6.249999999999999e-07,
-      "loss": 0.4614,
+      "epoch": 1.0289389067524115,
+      "grad_norm": 2.0498945713043213,
+      "learning_rate": 5.416666666666666e-07,
+      "loss": 0.5263,
       "step": 8000
     },
     {
-      "epoch": 1.2075471698113207,
-      "eval_loss": 0.49337518215179443,
-      "eval_runtime": 189.5607,
-      "eval_samples_per_second": 37.044,
-      "eval_steps_per_second": 0.58,
-      "eval_wer": 0.37760267874938547,
+      "epoch": 1.0289389067524115,
+      "eval_loss": Infinity,
+      "eval_runtime": 108.0404,
+      "eval_samples_per_second": 36.107,
+      "eval_steps_per_second": 0.565,
+      "eval_wer": 0.37269960875235475,
       "step": 8000
     },
     {
-      "epoch": 1.2075471698113207,
+      "epoch": 1.0289389067524115,
       "step": 8000,
-      "total_flos": 9.085819982279884e+18,
-      "train_loss": 0.9942695980072022,
-      "train_runtime": 6320.527,
-      "train_samples_per_second": 10.126,
-      "train_steps_per_second": 1.266
+      "total_flos": 8.75110352693014e+18,
+      "train_loss": 1.0117860870361328,
+      "train_runtime": 5013.4693,
+      "train_samples_per_second": 12.766,
+      "train_steps_per_second": 1.596
     }
   ],
   "logging_steps": 500,
@@ -291,7 +291,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 9.085819982279884e+18,
+  "total_flos": 8.75110352693014e+18,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null