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: 1.5563
 - Wer: 0.8702

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 3.0959752321981426,
-    "eval_loss": 0.42000359296798706,
-    "eval_runtime": 155.5667,
+    "epoch": 6.191950464396285,
+    "eval_loss": 1.5562535524368286,
+    "eval_runtime": 151.7345,
     "eval_samples": 5656,
-    "eval_samples_per_second": 36.357,
-    "eval_steps_per_second": 4.545,
-    "eval_wer": 0.3226557108696699,
-    "total_flos": 1.8144283373471887e+19,
-    "train_loss": 1.2569797668457032,
-    "train_runtime": 7798.4407,
+    "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,
     "train_samples": 41326,
-    "train_samples_per_second": 16.414,
-    "train_steps_per_second": 0.256
+    "train_samples_per_second": 20.916,
+    "train_steps_per_second": 0.163
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 3.0959752321981426,
-    "eval_loss": 0.42000359296798706,
-    "eval_runtime": 155.5667,
+    "epoch": 6.191950464396285,
+    "eval_loss": 1.5562535524368286,
+    "eval_runtime": 151.7345,
     "eval_samples": 5656,
-    "eval_samples_per_second": 36.357,
-    "eval_steps_per_second": 4.545,
-    "eval_wer": 0.3226557108696699
+    "eval_samples_per_second": 37.276,
+    "eval_steps_per_second": 4.659,
+    "eval_wer": 0.8702315802988236
 }

runs/May30_13-12-01_af3b5d38d2a2/events.out.tfevents.1717083752.af3b5d38d2a2.605.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77be8106872d56df106b021b0ca194d837d4e0258f7a9e6cd406157be4d7f025
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 3.0959752321981426,
-    "total_flos": 1.8144283373471887e+19,
-    "train_loss": 1.2569797668457032,
-    "train_runtime": 7798.4407,
+    "epoch": 6.191950464396285,
+    "total_flos": 3.641759707440775e+19,
+    "train_loss": 1.7097147216796875,
+    "train_runtime": 12239.4176,
     "train_samples": 41326,
-    "train_samples_per_second": 16.414,
-    "train_steps_per_second": 0.256
+    "train_samples_per_second": 20.916,
+    "train_steps_per_second": 0.163
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.0959752321981426,
+  "epoch": 6.191950464396285,
   "eval_steps": 100,
   "global_step": 2000,
   "is_hyper_param_search": false,
@@ -9,230 +9,230 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.15479876160990713,
-      "eval_loss": 3.5622992515563965,
-      "eval_runtime": 158.441,
-      "eval_samples_per_second": 35.698,
-      "eval_steps_per_second": 4.462,
+      "epoch": 0.30959752321981426,
+      "eval_loss": 3.562331438064575,
+      "eval_runtime": 151.987,
+      "eval_samples_per_second": 37.214,
+      "eval_steps_per_second": 4.652,
       "eval_wer": 1.0,
       "step": 100
     },
     {
-      "epoch": 0.30959752321981426,
-      "eval_loss": 3.2967281341552734,
-      "eval_runtime": 153.6671,
-      "eval_samples_per_second": 36.807,
-      "eval_steps_per_second": 4.601,
+      "epoch": 0.6191950464396285,
+      "eval_loss": 3.1175434589385986,
+      "eval_runtime": 149.1497,
+      "eval_samples_per_second": 37.922,
+      "eval_steps_per_second": 4.74,
       "eval_wer": 1.0,
       "step": 200
     },
     {
-      "epoch": 0.46439628482972134,
-      "eval_loss": 2.648355722427368,
-      "eval_runtime": 153.3096,
-      "eval_samples_per_second": 36.893,
-      "eval_steps_per_second": 4.612,
-      "eval_wer": 0.9999518544077289,
+      "epoch": 0.9287925696594427,
+      "eval_loss": 3.293273687362671,
+      "eval_runtime": 149.8183,
+      "eval_samples_per_second": 37.752,
+      "eval_steps_per_second": 4.719,
+      "eval_wer": 1.0,
       "step": 300
     },
     {
-      "epoch": 0.6191950464396285,
-      "eval_loss": 1.060172200202942,
-      "eval_runtime": 153.5019,
-      "eval_samples_per_second": 36.846,
-      "eval_steps_per_second": 4.606,
-      "eval_wer": 0.7315080804352362,
+      "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,
       "step": 400
     },
     {
-      "epoch": 0.7739938080495357,
-      "grad_norm": 2.8772997856140137,
-      "learning_rate": 0.00029699999999999996,
-      "loss": 3.6398,
+      "epoch": 1.5479876160990713,
+      "grad_norm": 0.48750218749046326,
+      "learning_rate": 0.00029519999999999997,
+      "loss": 3.7398,
       "step": 500
     },
     {
-      "epoch": 0.7739938080495357,
-      "eval_loss": 0.8942204713821411,
-      "eval_runtime": 154.5421,
-      "eval_samples_per_second": 36.598,
-      "eval_steps_per_second": 4.575,
-      "eval_wer": 0.6695607517131806,
+      "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,
       "step": 500
     },
     {
-      "epoch": 0.9287925696594427,
-      "eval_loss": 0.7115893959999084,
-      "eval_runtime": 155.1607,
-      "eval_samples_per_second": 36.453,
-      "eval_steps_per_second": 4.557,
-      "eval_wer": 0.5361172184686492,
+      "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,
       "step": 600
     },
     {
-      "epoch": 1.08359133126935,
-      "eval_loss": 0.6647589802742004,
-      "eval_runtime": 154.7086,
-      "eval_samples_per_second": 36.559,
-      "eval_steps_per_second": 4.57,
-      "eval_wer": 0.5100704530500233,
+      "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,
       "step": 700
     },
     {
-      "epoch": 1.238390092879257,
-      "eval_loss": 0.5869073867797852,
-      "eval_runtime": 154.314,
-      "eval_samples_per_second": 36.653,
-      "eval_steps_per_second": 4.582,
-      "eval_wer": 0.45284139237052845,
+      "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,
       "step": 800
     },
     {
-      "epoch": 1.3931888544891642,
-      "eval_loss": 0.5697966814041138,
-      "eval_runtime": 153.6252,
-      "eval_samples_per_second": 36.817,
-      "eval_steps_per_second": 4.602,
-      "eval_wer": 0.43589414389112674,
+      "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,
       "step": 900
     },
     {
-      "epoch": 1.5479876160990713,
-      "grad_norm": 0.41244009137153625,
-      "learning_rate": 0.0002016,
-      "loss": 0.5976,
+      "epoch": 3.0959752321981426,
+      "grad_norm": 0.4975910186767578,
+      "learning_rate": 0.0002024,
+      "loss": 0.5228,
       "step": 1000
     },
     {
-      "epoch": 1.5479876160990713,
-      "eval_loss": 0.54075688123703,
-      "eval_runtime": 154.0436,
-      "eval_samples_per_second": 36.717,
-      "eval_steps_per_second": 4.59,
-      "eval_wer": 0.41122755211760365,
+      "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,
       "step": 1000
     },
     {
-      "epoch": 1.7027863777089784,
-      "eval_loss": 0.5229234099388123,
-      "eval_runtime": 154.4025,
-      "eval_samples_per_second": 36.632,
-      "eval_steps_per_second": 4.579,
-      "eval_wer": 0.41960488517276245,
+      "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,
       "step": 1100
     },
     {
-      "epoch": 1.8575851393188856,
-      "eval_loss": 0.5054619312286377,
-      "eval_runtime": 154.4414,
-      "eval_samples_per_second": 36.622,
-      "eval_steps_per_second": 4.578,
-      "eval_wer": 0.39546789491422063,
+      "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,
       "step": 1200
     },
     {
-      "epoch": 2.0123839009287927,
-      "eval_loss": 0.48077496886253357,
-      "eval_runtime": 154.8241,
-      "eval_samples_per_second": 36.532,
-      "eval_steps_per_second": 4.566,
-      "eval_wer": 0.3708815457944825,
+      "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,
       "step": 1300
     },
     {
-      "epoch": 2.1671826625387,
-      "eval_loss": 0.46668991446495056,
-      "eval_runtime": 154.4804,
-      "eval_samples_per_second": 36.613,
-      "eval_steps_per_second": 4.577,
-      "eval_wer": 0.3579785270658471,
+      "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,
       "step": 1400
     },
     {
-      "epoch": 2.321981424148607,
-      "grad_norm": 0.45199868083000183,
-      "learning_rate": 0.00010179999999999998,
-      "loss": 0.443,
+      "epoch": 4.643962848297214,
+      "grad_norm": 3.2222559452056885,
+      "learning_rate": 0.00010359999999999998,
+      "loss": 0.9047,
       "step": 1500
     },
     {
-      "epoch": 2.321981424148607,
-      "eval_loss": 0.45730867981910706,
-      "eval_runtime": 154.0073,
-      "eval_samples_per_second": 36.726,
-      "eval_steps_per_second": 4.591,
-      "eval_wer": 0.35817110943493125,
+      "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,
       "step": 1500
     },
     {
-      "epoch": 2.476780185758514,
-      "eval_loss": 0.44747394323349,
-      "eval_runtime": 156.828,
-      "eval_samples_per_second": 36.065,
-      "eval_steps_per_second": 4.508,
-      "eval_wer": 0.3452199451140248,
+      "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,
       "step": 1600
     },
     {
-      "epoch": 2.6315789473684212,
-      "eval_loss": 0.43689531087875366,
-      "eval_runtime": 155.5732,
-      "eval_samples_per_second": 36.356,
-      "eval_steps_per_second": 4.544,
-      "eval_wer": 0.3477877100351463,
+      "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,
       "step": 1700
     },
     {
-      "epoch": 2.7863777089783284,
-      "eval_loss": 0.4227137863636017,
-      "eval_runtime": 156.702,
-      "eval_samples_per_second": 36.094,
-      "eval_steps_per_second": 4.512,
-      "eval_wer": 0.32978125852578194,
+      "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,
       "step": 1800
     },
     {
-      "epoch": 2.9411764705882355,
-      "eval_loss": 0.4168964922428131,
-      "eval_runtime": 153.7827,
-      "eval_samples_per_second": 36.779,
-      "eval_steps_per_second": 4.597,
-      "eval_wer": 0.3270690568278474,
+      "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,
       "step": 1900
     },
     {
-      "epoch": 3.0959752321981426,
-      "grad_norm": 0.4056684672832489,
-      "learning_rate": 2.4e-06,
-      "loss": 0.3475,
+      "epoch": 6.191950464396285,
+      "grad_norm": 0.7321383953094482,
+      "learning_rate": 4.399999999999999e-06,
+      "loss": 1.6715,
       "step": 2000
     },
     {
-      "epoch": 3.0959752321981426,
-      "eval_loss": 0.42000359296798706,
-      "eval_runtime": 154.6879,
-      "eval_samples_per_second": 36.564,
-      "eval_steps_per_second": 4.57,
-      "eval_wer": 0.3226557108696699,
+      "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,
       "step": 2000
     },
     {
-      "epoch": 3.0959752321981426,
+      "epoch": 6.191950464396285,
       "step": 2000,
-      "total_flos": 1.8144283373471887e+19,
-      "train_loss": 1.2569797668457032,
-      "train_runtime": 7798.4407,
-      "train_samples_per_second": 16.414,
-      "train_steps_per_second": 0.256
+      "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
     }
   ],
   "logging_steps": 500,
   "max_steps": 2000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 7,
   "save_steps": 400,
-  "total_flos": 1.8144283373471887e+19,
-  "train_batch_size": 64,
+  "total_flos": 3.641759707440775e+19,
+  "train_batch_size": 128,
   "trial_name": null,
   "trial_params": null
 }