End of training

README.md

@@ -2,6 +2,8 @@
 license: apache-2.0
 base_model: facebook/wav2vec2-large-xlsr-53
 tags:
+- automatic-speech-recognition
+- DewiBrynJones/banc-trawsgrifiadau-bangor-clean-with-ccv
 - generated_from_trainer
 metrics:
 - wer
@@ -15,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # wav2vec2-xlsr-53-ft-btb-ccv-cy
 
-This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on an unknown dataset.
+This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-CLEAN-WITH-CCV - DEFAULT dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.5355
 - Wer: 0.4186

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 6.191950464396285,
-    "eval_loss": 1.5562535524368286,
-    "eval_runtime": 151.7345,
+    "epoch": 0.5807200929152149,
+    "eval_loss": 0.5355480313301086,
+    "eval_runtime": 163.5057,
     "eval_samples": 5656,
-    "eval_samples_per_second": 37.276,
-    "eval_steps_per_second": 4.659,
-    "eval_wer": 0.8702315802988236,
-    "total_flos": 3.641759707440775e+19,
-    "train_loss": 1.7097147216796875,
-    "train_runtime": 12239.4176,
+    "eval_samples_per_second": 34.592,
+    "eval_steps_per_second": 4.324,
+    "eval_wer": 0.41856173067355684,
+    "total_flos": 3.3874766991231493e+18,
+    "train_loss": 1.2740035095214843,
+    "train_runtime": 6100.931,
     "train_samples": 41326,
-    "train_samples_per_second": 20.916,
-    "train_steps_per_second": 0.163
+    "train_samples_per_second": 3.934,
+    "train_steps_per_second": 0.492
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 6.191950464396285,
-    "eval_loss": 1.5562535524368286,
-    "eval_runtime": 151.7345,
+    "epoch": 0.5807200929152149,
+    "eval_loss": 0.5355480313301086,
+    "eval_runtime": 163.5057,
     "eval_samples": 5656,
-    "eval_samples_per_second": 37.276,
-    "eval_steps_per_second": 4.659,
-    "eval_wer": 0.8702315802988236
+    "eval_samples_per_second": 34.592,
+    "eval_steps_per_second": 4.324,
+    "eval_wer": 0.41856173067355684
 }

runs/May31_06-27-23_af3b5d38d2a2/events.out.tfevents.1717139582.af3b5d38d2a2.810.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a40d231ed4ffd1a98c35204e818acdb811ea420b06bab0d1a24549c859a734ac
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 6.191950464396285,
-    "total_flos": 3.641759707440775e+19,
-    "train_loss": 1.7097147216796875,
-    "train_runtime": 12239.4176,
+    "epoch": 0.5807200929152149,
+    "total_flos": 3.3874766991231493e+18,
+    "train_loss": 1.2740035095214843,
+    "train_runtime": 6100.931,
     "train_samples": 41326,
-    "train_samples_per_second": 20.916,
-    "train_steps_per_second": 0.163
+    "train_samples_per_second": 3.934,
+    "train_steps_per_second": 0.492
 }

trainer_state.json

@@ -1,238 +1,342 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 6.191950464396285,
+  "epoch": 0.5807200929152149,
   "eval_steps": 100,
-  "global_step": 2000,
+  "global_step": 3000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.30959752321981426,
-      "eval_loss": 3.562331438064575,
-      "eval_runtime": 151.987,
-      "eval_samples_per_second": 37.214,
-      "eval_steps_per_second": 4.652,
+      "epoch": 0.019357336430507164,
+      "eval_loss": 3.5493686199188232,
+      "eval_runtime": 162.4972,
+      "eval_samples_per_second": 34.807,
+      "eval_steps_per_second": 4.351,
       "eval_wer": 1.0,
       "step": 100
     },
     {
-      "epoch": 0.6191950464396285,
-      "eval_loss": 3.1175434589385986,
-      "eval_runtime": 149.1497,
-      "eval_samples_per_second": 37.922,
-      "eval_steps_per_second": 4.74,
+      "epoch": 0.03871467286101433,
+      "eval_loss": 3.0426220893859863,
+      "eval_runtime": 160.0022,
+      "eval_samples_per_second": 35.35,
+      "eval_steps_per_second": 4.419,
       "eval_wer": 1.0,
       "step": 200
     },
     {
-      "epoch": 0.9287925696594427,
-      "eval_loss": 3.293273687362671,
-      "eval_runtime": 149.8183,
-      "eval_samples_per_second": 37.752,
-      "eval_steps_per_second": 4.719,
+      "epoch": 0.05807200929152149,
+      "eval_loss": 2.8965415954589844,
+      "eval_runtime": 160.2187,
+      "eval_samples_per_second": 35.302,
+      "eval_steps_per_second": 4.413,
       "eval_wer": 1.0,
       "step": 300
     },
     {
-      "epoch": 1.238390092879257,
-      "eval_loss": 0.9764726161956787,
-      "eval_runtime": 149.5566,
-      "eval_samples_per_second": 37.818,
-      "eval_steps_per_second": 4.727,
-      "eval_wer": 0.6866203399078814,
+      "epoch": 0.07742934572202866,
+      "eval_loss": 1.826250433921814,
+      "eval_runtime": 161.1204,
+      "eval_samples_per_second": 35.104,
+      "eval_steps_per_second": 4.388,
+      "eval_wer": 0.9828762176822712,
       "step": 400
     },
     {
-      "epoch": 1.5479876160990713,
-      "grad_norm": 0.48750218749046326,
-      "learning_rate": 0.00029519999999999997,
-      "loss": 3.7398,
+      "epoch": 0.09678668215253582,
+      "grad_norm": 4.890473365783691,
+      "learning_rate": 0.0002982,
+      "loss": 3.9715,
       "step": 500
     },
     {
-      "epoch": 1.5479876160990713,
-      "eval_loss": 0.6787042617797852,
-      "eval_runtime": 150.878,
-      "eval_samples_per_second": 37.487,
-      "eval_steps_per_second": 4.686,
-      "eval_wer": 0.492176341255958,
+      "epoch": 0.09678668215253582,
+      "eval_loss": 1.3860080242156982,
+      "eval_runtime": 161.4467,
+      "eval_samples_per_second": 35.033,
+      "eval_steps_per_second": 4.379,
+      "eval_wer": 0.8748856542183563,
       "step": 500
     },
     {
-      "epoch": 1.8575851393188856,
-      "eval_loss": 0.6015087366104126,
-      "eval_runtime": 150.6376,
-      "eval_samples_per_second": 37.547,
-      "eval_steps_per_second": 4.693,
-      "eval_wer": 0.43465840702283703,
+      "epoch": 0.11614401858304298,
+      "eval_loss": 1.308407187461853,
+      "eval_runtime": 161.0068,
+      "eval_samples_per_second": 35.129,
+      "eval_steps_per_second": 4.391,
+      "eval_wer": 0.8153456051098522,
       "step": 600
     },
     {
-      "epoch": 2.1671826625387,
-      "eval_loss": 0.5698955059051514,
-      "eval_runtime": 150.4703,
-      "eval_samples_per_second": 37.589,
-      "eval_steps_per_second": 4.699,
-      "eval_wer": 0.42732422846688384,
+      "epoch": 0.13550135501355012,
+      "eval_loss": 1.0549893379211426,
+      "eval_runtime": 161.8736,
+      "eval_samples_per_second": 34.941,
+      "eval_steps_per_second": 4.368,
+      "eval_wer": 0.7336906806181894,
       "step": 700
     },
     {
-      "epoch": 2.476780185758514,
-      "eval_loss": 0.5419376492500305,
-      "eval_runtime": 150.5345,
-      "eval_samples_per_second": 37.573,
-      "eval_steps_per_second": 4.697,
-      "eval_wer": 0.39578886552936077,
+      "epoch": 0.1548586914440573,
+      "eval_loss": 1.0011754035949707,
+      "eval_runtime": 161.7881,
+      "eval_samples_per_second": 34.959,
+      "eval_steps_per_second": 4.37,
+      "eval_wer": 0.7190062749755259,
       "step": 800
     },
     {
-      "epoch": 2.7863777089783284,
-      "eval_loss": 0.49708282947540283,
-      "eval_runtime": 150.3649,
-      "eval_samples_per_second": 37.615,
-      "eval_steps_per_second": 4.702,
-      "eval_wer": 0.3729839033236507,
+      "epoch": 0.17421602787456447,
+      "eval_loss": 0.913667619228363,
+      "eval_runtime": 162.0369,
+      "eval_samples_per_second": 34.906,
+      "eval_steps_per_second": 4.363,
+      "eval_wer": 0.6752258830704049,
       "step": 900
     },
     {
-      "epoch": 3.0959752321981426,
-      "grad_norm": 0.4975910186767578,
-      "learning_rate": 0.0002024,
-      "loss": 0.5228,
+      "epoch": 0.19357336430507163,
+      "grad_norm": 3.201047658920288,
+      "learning_rate": 0.0002406,
+      "loss": 1.0155,
       "step": 1000
     },
     {
-      "epoch": 3.0959752321981426,
-      "eval_loss": 0.49601131677627563,
-      "eval_runtime": 150.6812,
-      "eval_samples_per_second": 37.536,
-      "eval_steps_per_second": 4.692,
-      "eval_wer": 0.354303413522492,
+      "epoch": 0.19357336430507163,
+      "eval_loss": 0.8486206531524658,
+      "eval_runtime": 161.6295,
+      "eval_samples_per_second": 34.994,
+      "eval_steps_per_second": 4.374,
+      "eval_wer": 0.6469483718765547,
       "step": 1000
     },
     {
-      "epoch": 3.405572755417957,
-      "eval_loss": 0.6976510286331177,
-      "eval_runtime": 149.5565,
-      "eval_samples_per_second": 37.818,
-      "eval_steps_per_second": 4.727,
-      "eval_wer": 0.4655036831378087,
+      "epoch": 0.2129307007355788,
+      "eval_loss": 0.8534524440765381,
+      "eval_runtime": 161.4191,
+      "eval_samples_per_second": 35.039,
+      "eval_steps_per_second": 4.38,
+      "eval_wer": 0.6111761968191812,
       "step": 1100
     },
     {
-      "epoch": 3.715170278637771,
-      "eval_loss": 0.799861490726471,
-      "eval_runtime": 151.2369,
-      "eval_samples_per_second": 37.398,
-      "eval_steps_per_second": 4.675,
-      "eval_wer": 0.5316396783874436,
+      "epoch": 0.23228803716608595,
+      "eval_loss": 0.8349705934524536,
+      "eval_runtime": 162.6145,
+      "eval_samples_per_second": 34.782,
+      "eval_steps_per_second": 4.348,
+      "eval_wer": 0.6192807048514708,
       "step": 1200
     },
     {
-      "epoch": 4.024767801857585,
-      "eval_loss": 1.0526387691497803,
-      "eval_runtime": 150.3745,
-      "eval_samples_per_second": 37.613,
-      "eval_steps_per_second": 4.702,
-      "eval_wer": 0.6214633050344242,
+      "epoch": 0.2516453735965931,
+      "eval_loss": 0.7680675983428955,
+      "eval_runtime": 161.7995,
+      "eval_samples_per_second": 34.957,
+      "eval_steps_per_second": 4.37,
+      "eval_wer": 0.5669785431143779,
       "step": 1300
     },
     {
-      "epoch": 4.3343653250774,
-      "eval_loss": 1.1285134553909302,
-      "eval_runtime": 150.4438,
-      "eval_samples_per_second": 37.595,
-      "eval_steps_per_second": 4.699,
-      "eval_wer": 0.7571857296464509,
+      "epoch": 0.27100271002710025,
+      "eval_loss": 0.7377049326896667,
+      "eval_runtime": 162.5136,
+      "eval_samples_per_second": 34.803,
+      "eval_steps_per_second": 4.35,
+      "eval_wer": 0.5559211054227985,
       "step": 1400
     },
     {
-      "epoch": 4.643962848297214,
-      "grad_norm": 3.2222559452056885,
-      "learning_rate": 0.00010359999999999998,
-      "loss": 0.9047,
+      "epoch": 0.29036004645760743,
+      "grad_norm": 5.448112487792969,
+      "learning_rate": 0.00018059999999999997,
+      "loss": 0.7987,
       "step": 1500
     },
     {
-      "epoch": 4.643962848297214,
-      "eval_loss": 1.1592659950256348,
-      "eval_runtime": 150.3382,
-      "eval_samples_per_second": 37.622,
-      "eval_steps_per_second": 4.703,
-      "eval_wer": 0.7047070340710307,
+      "epoch": 0.29036004645760743,
+      "eval_loss": 0.7129804491996765,
+      "eval_runtime": 162.7944,
+      "eval_samples_per_second": 34.743,
+      "eval_steps_per_second": 4.343,
+      "eval_wer": 0.5437242220474715,
       "step": 1500
     },
     {
-      "epoch": 4.953560371517028,
-      "eval_loss": 2.0400760173797607,
-      "eval_runtime": 150.5385,
-      "eval_samples_per_second": 37.572,
-      "eval_steps_per_second": 4.696,
-      "eval_wer": 0.966747444271477,
+      "epoch": 0.3097173828881146,
+      "eval_loss": 0.7039781808853149,
+      "eval_runtime": 162.4512,
+      "eval_samples_per_second": 34.817,
+      "eval_steps_per_second": 4.352,
+      "eval_wer": 0.5451846383463594,
       "step": 1600
     },
     {
-      "epoch": 5.2631578947368425,
-      "eval_loss": 1.6264142990112305,
-      "eval_runtime": 151.1274,
-      "eval_samples_per_second": 37.425,
-      "eval_steps_per_second": 4.678,
-      "eval_wer": 0.8680489801158704,
+      "epoch": 0.32907471931862176,
+      "eval_loss": 0.6728500127792358,
+      "eval_runtime": 166.2937,
+      "eval_samples_per_second": 34.012,
+      "eval_steps_per_second": 4.252,
+      "eval_wer": 0.5050954085153504,
       "step": 1700
     },
     {
-      "epoch": 5.572755417956657,
-      "eval_loss": 1.5916314125061035,
-      "eval_runtime": 151.1725,
-      "eval_samples_per_second": 37.414,
-      "eval_steps_per_second": 4.677,
-      "eval_wer": 0.8627369164353004,
+      "epoch": 0.34843205574912894,
+      "eval_loss": 0.6646420359611511,
+      "eval_runtime": 164.3307,
+      "eval_samples_per_second": 34.418,
+      "eval_steps_per_second": 4.302,
+      "eval_wer": 0.511338286979827,
       "step": 1800
     },
     {
-      "epoch": 5.882352941176471,
-      "eval_loss": 1.5764083862304688,
-      "eval_runtime": 150.5972,
-      "eval_samples_per_second": 37.557,
-      "eval_steps_per_second": 4.695,
-      "eval_wer": 0.8720129672128517,
+      "epoch": 0.3677893921796361,
+      "eval_loss": 0.6530969142913818,
+      "eval_runtime": 163.5666,
+      "eval_samples_per_second": 34.579,
+      "eval_steps_per_second": 4.322,
+      "eval_wer": 0.49691065782927574,
       "step": 1900
     },
     {
-      "epoch": 6.191950464396285,
-      "grad_norm": 0.7321383953094482,
-      "learning_rate": 4.399999999999999e-06,
-      "loss": 1.6715,
+      "epoch": 0.38714672861014326,
+      "grad_norm": 2.6550886631011963,
+      "learning_rate": 0.00012059999999999999,
+      "loss": 0.6851,
       "step": 2000
     },
     {
-      "epoch": 6.191950464396285,
-      "eval_loss": 1.5562535524368286,
-      "eval_runtime": 151.1236,
-      "eval_samples_per_second": 37.426,
-      "eval_steps_per_second": 4.678,
-      "eval_wer": 0.8702315802988236,
+      "epoch": 0.38714672861014326,
+      "eval_loss": 0.6413969397544861,
+      "eval_runtime": 162.7473,
+      "eval_samples_per_second": 34.753,
+      "eval_steps_per_second": 4.344,
+      "eval_wer": 0.5037954775240326,
       "step": 2000
     },
     {
-      "epoch": 6.191950464396285,
-      "step": 2000,
-      "total_flos": 3.641759707440775e+19,
-      "train_loss": 1.7097147216796875,
-      "train_runtime": 12239.4176,
-      "train_samples_per_second": 20.916,
-      "train_steps_per_second": 0.163
+      "epoch": 0.4065040650406504,
+      "eval_loss": 0.6108531355857849,
+      "eval_runtime": 163.3073,
+      "eval_samples_per_second": 34.634,
+      "eval_steps_per_second": 4.329,
+      "eval_wer": 0.467654186259248,
+      "step": 2100
+    },
+    {
+      "epoch": 0.4258614014711576,
+      "eval_loss": 0.6034538745880127,
+      "eval_runtime": 164.2548,
+      "eval_samples_per_second": 34.434,
+      "eval_steps_per_second": 4.304,
+      "eval_wer": 0.46924299080419185,
+      "step": 2200
+    },
+    {
+      "epoch": 0.4452187379016647,
+      "eval_loss": 0.5801683664321899,
+      "eval_runtime": 163.2622,
+      "eval_samples_per_second": 34.644,
+      "eval_steps_per_second": 4.33,
+      "eval_wer": 0.458955882588949,
+      "step": 2300
+    },
+    {
+      "epoch": 0.4645760743321719,
+      "eval_loss": 0.571967363357544,
+      "eval_runtime": 164.105,
+      "eval_samples_per_second": 34.466,
+      "eval_steps_per_second": 4.308,
+      "eval_wer": 0.44545906822230424,
+      "step": 2400
+    },
+    {
+      "epoch": 0.48393341076267904,
+      "grad_norm": 6.736985683441162,
+      "learning_rate": 6.0599999999999996e-05,
+      "loss": 0.5979,
+      "step": 2500
+    },
+    {
+      "epoch": 0.48393341076267904,
+      "eval_loss": 0.569513201713562,
+      "eval_runtime": 163.6068,
+      "eval_samples_per_second": 34.571,
+      "eval_steps_per_second": 4.321,
+      "eval_wer": 0.4425542841552856,
+      "step": 2500
+    },
+    {
+      "epoch": 0.5032907471931862,
+      "eval_loss": 0.5556703209877014,
+      "eval_runtime": 163.3416,
+      "eval_samples_per_second": 34.627,
+      "eval_steps_per_second": 4.328,
+      "eval_wer": 0.43513986294554735,
+      "step": 2600
+    },
+    {
+      "epoch": 0.5226480836236934,
+      "eval_loss": 0.5499459505081177,
+      "eval_runtime": 163.0587,
+      "eval_samples_per_second": 34.687,
+      "eval_steps_per_second": 4.336,
+      "eval_wer": 0.4269711607902297,
+      "step": 2700
+    },
+    {
+      "epoch": 0.5420054200542005,
+      "eval_loss": 0.5451160073280334,
+      "eval_runtime": 163.5278,
+      "eval_samples_per_second": 34.587,
+      "eval_steps_per_second": 4.323,
+      "eval_wer": 0.425815666575725,
+      "step": 2800
+    },
+    {
+      "epoch": 0.5613627564847077,
+      "eval_loss": 0.5382982492446899,
+      "eval_runtime": 163.1776,
+      "eval_samples_per_second": 34.662,
+      "eval_steps_per_second": 4.333,
+      "eval_wer": 0.42172329123268765,
+      "step": 2900
+    },
+    {
+      "epoch": 0.5807200929152149,
+      "grad_norm": 4.4167070388793945,
+      "learning_rate": 6e-07,
+      "loss": 0.5753,
+      "step": 3000
+    },
+    {
+      "epoch": 0.5807200929152149,
+      "eval_loss": 0.5355480313301086,
+      "eval_runtime": 164.5272,
+      "eval_samples_per_second": 34.377,
+      "eval_steps_per_second": 4.297,
+      "eval_wer": 0.41856173067355684,
+      "step": 3000
+    },
+    {
+      "epoch": 0.5807200929152149,
+      "step": 3000,
+      "total_flos": 3.3874766991231493e+18,
+      "train_loss": 1.2740035095214843,
+      "train_runtime": 6100.931,
+      "train_samples_per_second": 3.934,
+      "train_steps_per_second": 0.492
     }
   ],
   "logging_steps": 500,
-  "max_steps": 2000,
+  "max_steps": 3000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 7,
+  "num_train_epochs": 1,
   "save_steps": 400,
-  "total_flos": 3.641759707440775e+19,
-  "train_batch_size": 128,
+  "total_flos": 3.3874766991231493e+18,
+  "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null
 }