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: 0.4971
 - Wer: 0.3381

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 2.013162988772745,
-    "eval_loss": 0.4348967373371124,
-    "eval_runtime": 152.6747,
+    "epoch": 4.024767801857585,
+    "eval_loss": 0.49710649251937866,
+    "eval_runtime": 152.3497,
     "eval_samples": 5656,
-    "eval_samples_per_second": 37.046,
-    "eval_steps_per_second": 4.631,
-    "eval_wer": 0.3390733578340903,
-    "total_flos": 1.1633941226063049e+19,
-    "train_loss": 1.1519982103201059,
-    "train_runtime": 7259.2244,
+    "eval_samples_per_second": 37.125,
+    "eval_steps_per_second": 4.641,
+    "eval_wer": 0.3380783489271557,
+    "total_flos": 2.3576121775744893e+19,
+    "train_loss": 1.074923675243671,
+    "train_runtime": 10028.377,
     "train_samples": 41326,
-    "train_samples_per_second": 11.461,
-    "train_steps_per_second": 0.358
+    "train_samples_per_second": 16.593,
+    "train_steps_per_second": 0.259
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 2.013162988772745,
-    "eval_loss": 0.4348967373371124,
-    "eval_runtime": 152.6747,
+    "epoch": 4.024767801857585,
+    "eval_loss": 0.49710649251937866,
+    "eval_runtime": 152.3497,
     "eval_samples": 5656,
-    "eval_samples_per_second": 37.046,
-    "eval_steps_per_second": 4.631,
-    "eval_wer": 0.3390733578340903
+    "eval_samples_per_second": 37.125,
+    "eval_steps_per_second": 4.641,
+    "eval_wer": 0.3380783489271557
 }

runs/May23_22-05-49_7571a1a6d5fc/events.out.tfevents.1716508587.7571a1a6d5fc.916.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eacbee8724b4ae6317d572c5bcbc55aacdbd4eb30fc86b27a434fab49a21465a
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 2.013162988772745,
-    "total_flos": 1.1633941226063049e+19,
-    "train_loss": 1.1519982103201059,
-    "train_runtime": 7259.2244,
+    "epoch": 4.024767801857585,
+    "total_flos": 2.3576121775744893e+19,
+    "train_loss": 1.074923675243671,
+    "train_runtime": 10028.377,
     "train_samples": 41326,
-    "train_samples_per_second": 11.461,
-    "train_steps_per_second": 0.358
+    "train_samples_per_second": 16.593,
+    "train_steps_per_second": 0.259
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 2.013162988772745,
+  "epoch": 4.024767801857585,
   "eval_steps": 100,
   "global_step": 2600,
   "is_hyper_param_search": false,
@@ -9,291 +9,291 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.07742934572202866,
-      "eval_loss": 3.5345706939697266,
-      "eval_runtime": 152.1587,
-      "eval_samples_per_second": 37.172,
-      "eval_steps_per_second": 4.646,
+      "epoch": 0.15479876160990713,
+      "eval_loss": 3.556110143661499,
+      "eval_runtime": 152.2069,
+      "eval_samples_per_second": 37.16,
+      "eval_steps_per_second": 4.645,
       "eval_wer": 1.0,
       "step": 100
     },
     {
-      "epoch": 0.1548586914440573,
-      "eval_loss": 2.982929229736328,
-      "eval_runtime": 150.2931,
-      "eval_samples_per_second": 37.633,
+      "epoch": 0.30959752321981426,
+      "eval_loss": 3.3711678981781006,
+      "eval_runtime": 150.3074,
+      "eval_samples_per_second": 37.63,
       "eval_steps_per_second": 4.704,
       "eval_wer": 1.0,
       "step": 200
     },
     {
-      "epoch": 0.23228803716608595,
-      "eval_loss": 2.770493268966675,
-      "eval_runtime": 150.0834,
-      "eval_samples_per_second": 37.686,
-      "eval_steps_per_second": 4.711,
+      "epoch": 0.46439628482972134,
+      "eval_loss": 3.1735892295837402,
+      "eval_runtime": 149.399,
+      "eval_samples_per_second": 37.858,
+      "eval_steps_per_second": 4.732,
       "eval_wer": 1.0,
       "step": 300
     },
     {
-      "epoch": 0.3097173828881146,
-      "eval_loss": 1.3696156740188599,
-      "eval_runtime": 150.2358,
-      "eval_samples_per_second": 37.647,
-      "eval_steps_per_second": 4.706,
-      "eval_wer": 0.8535090112500201,
+      "epoch": 0.6191950464396285,
+      "eval_loss": 1.1719417572021484,
+      "eval_runtime": 149.6776,
+      "eval_samples_per_second": 37.788,
+      "eval_steps_per_second": 4.723,
+      "eval_wer": 0.7909839354207122,
       "step": 400
     },
     {
-      "epoch": 0.38714672861014326,
-      "grad_norm": 3.0665018558502197,
+      "epoch": 0.7739938080495357,
+      "grad_norm": 4.971697807312012,
       "learning_rate": 0.00029699999999999996,
-      "loss": 3.7305,
+      "loss": 3.7217,
       "step": 500
     },
     {
-      "epoch": 0.38714672861014326,
-      "eval_loss": 1.0936249494552612,
-      "eval_runtime": 153.8749,
-      "eval_samples_per_second": 36.757,
-      "eval_steps_per_second": 4.595,
-      "eval_wer": 0.7465295052237968,
+      "epoch": 0.7739938080495357,
+      "eval_loss": 0.9168550968170166,
+      "eval_runtime": 151.3653,
+      "eval_samples_per_second": 37.367,
+      "eval_steps_per_second": 4.671,
+      "eval_wer": 0.6775689685609283,
       "step": 500
     },
     {
-      "epoch": 0.4645760743321719,
-      "eval_loss": 0.8456823825836182,
-      "eval_runtime": 153.7133,
-      "eval_samples_per_second": 36.796,
-      "eval_steps_per_second": 4.599,
-      "eval_wer": 0.6413313861116015,
+      "epoch": 0.9287925696594427,
+      "eval_loss": 0.8096873760223389,
+      "eval_runtime": 152.1183,
+      "eval_samples_per_second": 37.182,
+      "eval_steps_per_second": 4.648,
+      "eval_wer": 0.6191362680746577,
       "step": 600
     },
     {
-      "epoch": 0.5420054200542005,
-      "eval_loss": 0.7860042452812195,
-      "eval_runtime": 152.7402,
-      "eval_samples_per_second": 37.03,
-      "eval_steps_per_second": 4.629,
-      "eval_wer": 0.5835566753863684,
+      "epoch": 1.08359133126935,
+      "eval_loss": 0.698630690574646,
+      "eval_runtime": 152.6154,
+      "eval_samples_per_second": 37.06,
+      "eval_steps_per_second": 4.633,
+      "eval_wer": 0.5080001925823691,
       "step": 700
     },
     {
-      "epoch": 0.6194347657762292,
-      "eval_loss": 0.7366299629211426,
-      "eval_runtime": 151.7301,
-      "eval_samples_per_second": 37.277,
-      "eval_steps_per_second": 4.66,
-      "eval_wer": 0.563736739901462,
+      "epoch": 1.238390092879257,
+      "eval_loss": 0.6122636795043945,
+      "eval_runtime": 152.7743,
+      "eval_samples_per_second": 37.022,
+      "eval_steps_per_second": 4.628,
+      "eval_wer": 0.4628556755629022,
       "step": 800
     },
     {
-      "epoch": 0.6968641114982579,
-      "eval_loss": 0.7318999171257019,
-      "eval_runtime": 152.5414,
-      "eval_samples_per_second": 37.078,
-      "eval_steps_per_second": 4.635,
-      "eval_wer": 0.5493572563431818,
+      "epoch": 1.3931888544891642,
+      "eval_loss": 0.581542432308197,
+      "eval_runtime": 152.3981,
+      "eval_samples_per_second": 37.113,
+      "eval_steps_per_second": 4.639,
+      "eval_wer": 0.44985636564972475,
       "step": 900
     },
     {
-      "epoch": 0.7742934572202865,
-      "grad_norm": 2.5090723037719727,
-      "learning_rate": 0.00022928571428571426,
-      "loss": 0.7504,
+      "epoch": 1.5479876160990713,
+      "grad_norm": 0.40550583600997925,
+      "learning_rate": 0.00022971428571428568,
+      "loss": 0.6148,
       "step": 1000
     },
     {
-      "epoch": 0.7742934572202865,
-      "eval_loss": 0.6438552737236023,
-      "eval_runtime": 151.1771,
-      "eval_samples_per_second": 37.413,
-      "eval_steps_per_second": 4.677,
-      "eval_wer": 0.5104074721959204,
+      "epoch": 1.5479876160990713,
+      "eval_loss": 0.5666411519050598,
+      "eval_runtime": 152.2007,
+      "eval_samples_per_second": 37.161,
+      "eval_steps_per_second": 4.645,
+      "eval_wer": 0.4218356309479867,
       "step": 1000
     },
     {
-      "epoch": 0.8517228029423152,
-      "eval_loss": 0.6213911175727844,
-      "eval_runtime": 151.6033,
-      "eval_samples_per_second": 37.308,
-      "eval_steps_per_second": 4.663,
-      "eval_wer": 0.4759191795991077,
+      "epoch": 1.7027863777089784,
+      "eval_loss": 0.5285393595695496,
+      "eval_runtime": 153.0291,
+      "eval_samples_per_second": 36.96,
+      "eval_steps_per_second": 4.62,
+      "eval_wer": 0.404872333937828,
       "step": 1100
     },
     {
-      "epoch": 0.9291521486643438,
-      "eval_loss": 0.5957211852073669,
-      "eval_runtime": 152.1457,
-      "eval_samples_per_second": 37.175,
-      "eval_steps_per_second": 4.647,
-      "eval_wer": 0.4628396270321452,
+      "epoch": 1.8575851393188856,
+      "eval_loss": 0.5309253334999084,
+      "eval_runtime": 151.5626,
+      "eval_samples_per_second": 37.318,
+      "eval_steps_per_second": 4.665,
+      "eval_wer": 0.39715299064370657,
       "step": 1200
     },
     {
-      "epoch": 1.0065814943863725,
-      "eval_loss": 0.5717456340789795,
-      "eval_runtime": 152.0809,
-      "eval_samples_per_second": 37.191,
-      "eval_steps_per_second": 4.649,
-      "eval_wer": 0.43531639678387446,
+      "epoch": 2.0123839009287927,
+      "eval_loss": 0.4992642104625702,
+      "eval_runtime": 152.6009,
+      "eval_samples_per_second": 37.064,
+      "eval_steps_per_second": 4.633,
+      "eval_wer": 0.38177849817849174,
       "step": 1300
     },
     {
-      "epoch": 1.084010840108401,
-      "eval_loss": 0.549981951713562,
-      "eval_runtime": 151.522,
-      "eval_samples_per_second": 37.328,
-      "eval_steps_per_second": 4.666,
-      "eval_wer": 0.4192197204345942,
+      "epoch": 2.1671826625387,
+      "eval_loss": 0.4779135286808014,
+      "eval_runtime": 152.6609,
+      "eval_samples_per_second": 37.049,
+      "eval_steps_per_second": 4.631,
+      "eval_wer": 0.37216542825504323,
       "step": 1400
     },
     {
-      "epoch": 1.1614401858304297,
-      "grad_norm": 0.4617447853088379,
-      "learning_rate": 0.00015799999999999996,
-      "loss": 0.5571,
+      "epoch": 2.321981424148607,
+      "grad_norm": 0.44798457622528076,
+      "learning_rate": 0.0001584285714285714,
+      "loss": 0.4634,
       "step": 1500
     },
     {
-      "epoch": 1.1614401858304297,
-      "eval_loss": 0.5342110395431519,
-      "eval_runtime": 152.7132,
-      "eval_samples_per_second": 37.037,
-      "eval_steps_per_second": 4.63,
-      "eval_wer": 0.4073277591436504,
+      "epoch": 2.321981424148607,
+      "eval_loss": 0.4679175317287445,
+      "eval_runtime": 152.7285,
+      "eval_samples_per_second": 37.033,
+      "eval_steps_per_second": 4.629,
+      "eval_wer": 0.36587440419829564,
       "step": 1500
     },
     {
-      "epoch": 1.2388695315524583,
-      "eval_loss": 0.5206533670425415,
-      "eval_runtime": 151.157,
-      "eval_samples_per_second": 37.418,
-      "eval_steps_per_second": 4.677,
-      "eval_wer": 0.4023687631397346,
+      "epoch": 2.476780185758514,
+      "eval_loss": 0.4679177403450012,
+      "eval_runtime": 152.8181,
+      "eval_samples_per_second": 37.011,
+      "eval_steps_per_second": 4.626,
+      "eval_wer": 0.3621350965319125,
       "step": 1600
     },
     {
-      "epoch": 1.316298877274487,
-      "eval_loss": 0.5142083168029785,
-      "eval_runtime": 151.9132,
-      "eval_samples_per_second": 37.232,
-      "eval_steps_per_second": 4.654,
-      "eval_wer": 0.3968641170900804,
+      "epoch": 2.6315789473684212,
+      "eval_loss": 0.45253580808639526,
+      "eval_runtime": 152.233,
+      "eval_samples_per_second": 37.154,
+      "eval_steps_per_second": 4.644,
+      "eval_wer": 0.3504196690792958,
       "step": 1700
     },
     {
-      "epoch": 1.3937282229965158,
-      "eval_loss": 0.5083270072937012,
-      "eval_runtime": 152.7068,
-      "eval_samples_per_second": 37.038,
-      "eval_steps_per_second": 4.63,
-      "eval_wer": 0.39583701112163183,
+      "epoch": 2.7863777089783284,
+      "eval_loss": 0.43527287244796753,
+      "eval_runtime": 151.4529,
+      "eval_samples_per_second": 37.345,
+      "eval_steps_per_second": 4.668,
+      "eval_wer": 0.33567106931360435,
       "step": 1800
     },
     {
-      "epoch": 1.4711575687185443,
-      "eval_loss": 0.4886321723461151,
-      "eval_runtime": 152.3343,
-      "eval_samples_per_second": 37.129,
-      "eval_steps_per_second": 4.641,
-      "eval_wer": 0.3825488276548282,
+      "epoch": 2.9411764705882355,
+      "eval_loss": 0.43578359484672546,
+      "eval_runtime": 151.2727,
+      "eval_samples_per_second": 37.389,
+      "eval_steps_per_second": 4.674,
+      "eval_wer": 0.34231516104700616,
       "step": 1900
     },
     {
-      "epoch": 1.5485869144405728,
-      "grad_norm": 0.441532164812088,
-      "learning_rate": 8.685714285714285e-05,
-      "loss": 0.4603,
+      "epoch": 3.0959752321981426,
+      "grad_norm": 0.39644455909729004,
+      "learning_rate": 8.742857142857142e-05,
+      "loss": 0.3708,
       "step": 2000
     },
     {
-      "epoch": 1.5485869144405728,
-      "eval_loss": 0.4732557237148285,
-      "eval_runtime": 155.1846,
-      "eval_samples_per_second": 36.447,
-      "eval_steps_per_second": 4.556,
-      "eval_wer": 0.3743480284379965,
+      "epoch": 3.0959752321981426,
+      "eval_loss": 0.4390716254711151,
+      "eval_runtime": 151.9868,
+      "eval_samples_per_second": 37.214,
+      "eval_steps_per_second": 4.652,
+      "eval_wer": 0.33554268106754825,
       "step": 2000
     },
     {
-      "epoch": 1.6260162601626016,
-      "eval_loss": 0.4615860879421234,
-      "eval_runtime": 155.1621,
-      "eval_samples_per_second": 36.452,
-      "eval_steps_per_second": 4.557,
-      "eval_wer": 0.3618622715090433,
+      "epoch": 3.2507739938080498,
+      "eval_loss": 0.43068408966064453,
+      "eval_runtime": 150.3939,
+      "eval_samples_per_second": 37.608,
+      "eval_steps_per_second": 4.701,
+      "eval_wer": 0.32817640545008103,
       "step": 2100
     },
     {
-      "epoch": 1.7034456058846303,
-      "eval_loss": 0.4535791277885437,
-      "eval_runtime": 152.4739,
-      "eval_samples_per_second": 37.095,
-      "eval_steps_per_second": 4.637,
-      "eval_wer": 0.3627288921699218,
+      "epoch": 3.405572755417957,
+      "eval_loss": 0.42743122577667236,
+      "eval_runtime": 150.4729,
+      "eval_samples_per_second": 37.588,
+      "eval_steps_per_second": 4.699,
+      "eval_wer": 0.3275344642198007,
       "step": 2200
     },
     {
-      "epoch": 1.7808749516066589,
-      "eval_loss": 0.44882732629776,
-      "eval_runtime": 153.2132,
-      "eval_samples_per_second": 36.916,
-      "eval_steps_per_second": 4.614,
-      "eval_wer": 0.3487185248190528,
+      "epoch": 3.560371517027864,
+      "eval_loss": 0.43879058957099915,
+      "eval_runtime": 151.6942,
+      "eval_samples_per_second": 37.286,
+      "eval_steps_per_second": 4.661,
+      "eval_wer": 0.324629680152782,
       "step": 2300
     },
     {
-      "epoch": 1.8583042973286876,
-      "eval_loss": 0.4429319202899933,
-      "eval_runtime": 154.0199,
-      "eval_samples_per_second": 36.723,
-      "eval_steps_per_second": 4.59,
-      "eval_wer": 0.34810868065028644,
+      "epoch": 3.715170278637771,
+      "eval_loss": 0.4521045684814453,
+      "eval_runtime": 151.7513,
+      "eval_samples_per_second": 37.272,
+      "eval_steps_per_second": 4.659,
+      "eval_wer": 0.32769494952737077,
       "step": 2400
     },
     {
-      "epoch": 1.9357336430507162,
-      "grad_norm": 0.48590919375419617,
-      "learning_rate": 1.557142857142857e-05,
-      "loss": 0.4163,
+      "epoch": 3.8699690402476783,
+      "grad_norm": 6.033559322357178,
+      "learning_rate": 1.642857142857143e-05,
+      "loss": 0.3391,
       "step": 2500
     },
     {
-      "epoch": 1.9357336430507162,
-      "eval_loss": 0.4377308487892151,
-      "eval_runtime": 154.6322,
-      "eval_samples_per_second": 36.577,
-      "eval_steps_per_second": 4.572,
-      "eval_wer": 0.3418978992473239,
+      "epoch": 3.8699690402476783,
+      "eval_loss": 0.4861215353012085,
+      "eval_runtime": 151.4972,
+      "eval_samples_per_second": 37.334,
+      "eval_steps_per_second": 4.667,
+      "eval_wer": 0.3390412607725763,
       "step": 2500
     },
     {
-      "epoch": 2.013162988772745,
-      "eval_loss": 0.4348967373371124,
-      "eval_runtime": 154.2249,
-      "eval_samples_per_second": 36.674,
-      "eval_steps_per_second": 4.584,
-      "eval_wer": 0.3390733578340903,
+      "epoch": 4.024767801857585,
+      "eval_loss": 0.49710649251937866,
+      "eval_runtime": 152.3447,
+      "eval_samples_per_second": 37.126,
+      "eval_steps_per_second": 4.641,
+      "eval_wer": 0.3380783489271557,
       "step": 2600
     },
     {
-      "epoch": 2.013162988772745,
+      "epoch": 4.024767801857585,
       "step": 2600,
-      "total_flos": 1.1633941226063049e+19,
-      "train_loss": 1.1519982103201059,
-      "train_runtime": 7259.2244,
-      "train_samples_per_second": 11.461,
-      "train_steps_per_second": 0.358
+      "total_flos": 2.3576121775744893e+19,
+      "train_loss": 1.074923675243671,
+      "train_runtime": 10028.377,
+      "train_samples_per_second": 16.593,
+      "train_steps_per_second": 0.259
     }
   ],
   "logging_steps": 500,
   "max_steps": 2600,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
+  "num_train_epochs": 5,
   "save_steps": 400,
-  "total_flos": 1.1633941226063049e+19,
-  "train_batch_size": 16,
+  "total_flos": 2.3576121775744893e+19,
+  "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null
 }