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-normalized
 - generated_from_trainer
 metrics:
 - wer
@@ -15,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # wav2vec2-xlsr-53-ft-btb-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-NORMALIZED - DEFAULT dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.4589
 - Wer: 0.3743

all_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 3.536067892503536,
-    "eval_loss": 0.4583967626094818,
-    "eval_runtime": 161.895,
+    "eval_loss": 0.4588949382305145,
+    "eval_runtime": 153.3999,
     "eval_samples": 5656,
-    "eval_samples_per_second": 34.936,
-    "eval_steps_per_second": 4.367,
-    "eval_wer": 0.37831418116453747,
+    "eval_samples_per_second": 36.871,
+    "eval_steps_per_second": 4.609,
+    "eval_wer": 0.3742697566925595,
     "total_flos": 9.55169606524761e+18,
-    "train_loss": 1.2462444946289062,
-    "train_runtime": 6962.0027,
+    "train_loss": 1.2622116821289062,
+    "train_runtime": 6522.8506,
     "train_samples": 22621,
-    "train_samples_per_second": 11.491,
-    "train_steps_per_second": 0.359
+    "train_samples_per_second": 12.265,
+    "train_steps_per_second": 0.383
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 3.536067892503536,
-    "eval_loss": 0.4583967626094818,
-    "eval_runtime": 161.895,
+    "eval_loss": 0.4588949382305145,
+    "eval_runtime": 153.3999,
     "eval_samples": 5656,
-    "eval_samples_per_second": 34.936,
-    "eval_steps_per_second": 4.367,
-    "eval_wer": 0.37831418116453747
+    "eval_samples_per_second": 36.871,
+    "eval_steps_per_second": 4.609,
+    "eval_wer": 0.3742697566925595
 }

runs/May16_08-14-26_2e698a9c250f/events.out.tfevents.1715850411.2e698a9c250f.233.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f32c40b2b9131b58fbffd90fb39b2a35a1ffd6bb07bca4c083c735df411d2c55
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 3.536067892503536,
     "total_flos": 9.55169606524761e+18,
-    "train_loss": 1.2462444946289062,
-    "train_runtime": 6962.0027,
+    "train_loss": 1.2622116821289062,
+    "train_runtime": 6522.8506,
     "train_samples": 22621,
-    "train_samples_per_second": 11.491,
-    "train_steps_per_second": 0.359
+    "train_samples_per_second": 12.265,
+    "train_steps_per_second": 0.383
 }

trainer_state.json

@@ -10,272 +10,272 @@
   "log_history": [
     {
       "epoch": 0.14144271570014144,
-      "eval_loss": 4.031618595123291,
-      "eval_runtime": 161.4756,
-      "eval_samples_per_second": 35.027,
-      "eval_steps_per_second": 4.378,
+      "eval_loss": 4.035430908203125,
+      "eval_runtime": 153.5895,
+      "eval_samples_per_second": 36.825,
+      "eval_steps_per_second": 4.603,
       "eval_wer": 1.0,
       "step": 100
     },
     {
       "epoch": 0.2828854314002829,
-      "eval_loss": 3.0548782348632812,
-      "eval_runtime": 158.3539,
-      "eval_samples_per_second": 35.717,
-      "eval_steps_per_second": 4.465,
+      "eval_loss": 3.097731590270996,
+      "eval_runtime": 149.7934,
+      "eval_samples_per_second": 37.759,
+      "eval_steps_per_second": 4.72,
       "eval_wer": 1.0,
       "step": 200
     },
     {
       "epoch": 0.4243281471004243,
-      "eval_loss": 2.5533947944641113,
-      "eval_runtime": 158.9465,
-      "eval_samples_per_second": 35.584,
-      "eval_steps_per_second": 4.448,
-      "eval_wer": 0.9862296976311228,
+      "eval_loss": 3.0768725872039795,
+      "eval_runtime": 151.0662,
+      "eval_samples_per_second": 37.441,
+      "eval_steps_per_second": 4.68,
+      "eval_wer": 1.0,
       "step": 300
     },
     {
       "epoch": 0.5657708628005658,
-      "eval_loss": 1.4279608726501465,
-      "eval_runtime": 159.2587,
-      "eval_samples_per_second": 35.515,
-      "eval_steps_per_second": 4.439,
-      "eval_wer": 0.8846536560313283,
+      "eval_loss": 1.373844027519226,
+      "eval_runtime": 150.8632,
+      "eval_samples_per_second": 37.491,
+      "eval_steps_per_second": 4.686,
+      "eval_wer": 0.8913622648777043,
       "step": 400
     },
     {
       "epoch": 0.7072135785007072,
-      "grad_norm": 1.9099421501159668,
-      "learning_rate": 0.00029699999999999996,
-      "loss": 3.6818,
+      "grad_norm": 2.2486917972564697,
+      "learning_rate": 0.0002964,
+      "loss": 3.7586,
       "step": 500
     },
     {
       "epoch": 0.7072135785007072,
-      "eval_loss": 1.1378353834152222,
-      "eval_runtime": 160.4319,
-      "eval_samples_per_second": 35.255,
-      "eval_steps_per_second": 4.407,
-      "eval_wer": 0.7958368106824164,
+      "eval_loss": 1.0915191173553467,
+      "eval_runtime": 151.7555,
+      "eval_samples_per_second": 37.27,
+      "eval_steps_per_second": 4.659,
+      "eval_wer": 0.7692431148488156,
       "step": 500
     },
     {
       "epoch": 0.8486562942008486,
-      "eval_loss": 0.9263126254081726,
-      "eval_runtime": 159.8222,
-      "eval_samples_per_second": 35.389,
-      "eval_steps_per_second": 4.424,
-      "eval_wer": 0.6777299865185851,
+      "eval_loss": 0.9360549449920654,
+      "eval_runtime": 152.9408,
+      "eval_samples_per_second": 36.982,
+      "eval_steps_per_second": 4.623,
+      "eval_wer": 0.6854818000898761,
       "step": 600
     },
     {
       "epoch": 0.9900990099009901,
-      "eval_loss": 0.8501134514808655,
-      "eval_runtime": 160.338,
-      "eval_samples_per_second": 35.275,
-      "eval_steps_per_second": 4.409,
-      "eval_wer": 0.6388104256275278,
+      "eval_loss": 0.8494995832443237,
+      "eval_runtime": 151.5832,
+      "eval_samples_per_second": 37.313,
+      "eval_steps_per_second": 4.664,
+      "eval_wer": 0.6247191371894459,
       "step": 700
     },
     {
       "epoch": 1.1315417256011315,
-      "eval_loss": 0.6984566450119019,
-      "eval_runtime": 159.8325,
-      "eval_samples_per_second": 35.387,
-      "eval_steps_per_second": 4.423,
-      "eval_wer": 0.5563972523592476,
+      "eval_loss": 0.6885886788368225,
+      "eval_runtime": 151.913,
+      "eval_samples_per_second": 37.232,
+      "eval_steps_per_second": 4.654,
+      "eval_wer": 0.5397059767606086,
       "step": 800
     },
     {
       "epoch": 1.272984441301273,
-      "eval_loss": 0.6664860248565674,
-      "eval_runtime": 162.9316,
-      "eval_samples_per_second": 34.714,
-      "eval_steps_per_second": 4.339,
-      "eval_wer": 0.5401393079540348,
+      "eval_loss": 0.6703779101371765,
+      "eval_runtime": 153.2384,
+      "eval_samples_per_second": 36.91,
+      "eval_steps_per_second": 4.614,
+      "eval_wer": 0.531215895230147,
       "step": 900
     },
     {
       "epoch": 1.4144271570014144,
-      "grad_norm": 0.8423302173614502,
-      "learning_rate": 0.00022574999999999996,
-      "loss": 0.895,
+      "grad_norm": 0.8719882369041443,
+      "learning_rate": 0.0002259,
+      "loss": 0.8877,
       "step": 1000
     },
     {
       "epoch": 1.4144271570014144,
-      "eval_loss": 0.6228350400924683,
-      "eval_runtime": 161.5875,
-      "eval_samples_per_second": 35.003,
-      "eval_steps_per_second": 4.375,
-      "eval_wer": 0.5000802465173011,
+      "eval_loss": 0.62369304895401,
+      "eval_runtime": 153.3779,
+      "eval_samples_per_second": 36.876,
+      "eval_steps_per_second": 4.61,
+      "eval_wer": 0.495040765230789,
       "step": 1000
     },
     {
       "epoch": 1.5558698727015559,
-      "eval_loss": 0.5974757075309753,
-      "eval_runtime": 160.961,
-      "eval_samples_per_second": 35.139,
-      "eval_steps_per_second": 4.392,
-      "eval_wer": 0.48359761186364514,
+      "eval_loss": 0.5992260575294495,
+      "eval_runtime": 152.5492,
+      "eval_samples_per_second": 37.077,
+      "eval_steps_per_second": 4.635,
+      "eval_wer": 0.4767606085895872,
       "step": 1100
     },
     {
       "epoch": 1.6973125884016973,
-      "eval_loss": 0.5826218128204346,
-      "eval_runtime": 162.2219,
-      "eval_samples_per_second": 34.866,
-      "eval_steps_per_second": 4.358,
-      "eval_wer": 0.4677248507414778,
+      "eval_loss": 0.5729629397392273,
+      "eval_runtime": 152.035,
+      "eval_samples_per_second": 37.202,
+      "eval_steps_per_second": 4.65,
+      "eval_wer": 0.4521730756885151,
       "step": 1200
     },
     {
       "epoch": 1.8387553041018387,
-      "eval_loss": 0.5473496317863464,
-      "eval_runtime": 164.7294,
-      "eval_samples_per_second": 34.335,
-      "eval_steps_per_second": 4.292,
-      "eval_wer": 0.4512903639982025,
+      "eval_loss": 0.5504249930381775,
+      "eval_runtime": 152.762,
+      "eval_samples_per_second": 37.025,
+      "eval_steps_per_second": 4.628,
+      "eval_wer": 0.4417891763497464,
       "step": 1300
     },
     {
       "epoch": 1.9801980198019802,
-      "eval_loss": 0.5284178256988525,
-      "eval_runtime": 168.0547,
-      "eval_samples_per_second": 33.656,
-      "eval_steps_per_second": 4.207,
-      "eval_wer": 0.4376966039673878,
+      "eval_loss": 0.5288310050964355,
+      "eval_runtime": 152.1876,
+      "eval_samples_per_second": 37.165,
+      "eval_steps_per_second": 4.646,
+      "eval_wer": 0.42586826731719846,
       "step": 1400
     },
     {
       "epoch": 2.1216407355021216,
-      "grad_norm": 1.0099021196365356,
-      "learning_rate": 0.0001512,
-      "loss": 0.687,
+      "grad_norm": 0.7933566570281982,
+      "learning_rate": 0.00015134999999999997,
+      "loss": 0.6844,
       "step": 1500
     },
     {
       "epoch": 2.1216407355021216,
-      "eval_loss": 0.5137470960617065,
-      "eval_runtime": 160.3266,
-      "eval_samples_per_second": 35.278,
-      "eval_steps_per_second": 4.41,
-      "eval_wer": 0.42363741413622646,
+      "eval_loss": 0.5165240168571472,
+      "eval_runtime": 152.8573,
+      "eval_samples_per_second": 37.002,
+      "eval_steps_per_second": 4.625,
+      "eval_wer": 0.4216633498106182,
       "step": 1500
     },
     {
       "epoch": 2.263083451202263,
-      "eval_loss": 0.5110819339752197,
-      "eval_runtime": 161.254,
-      "eval_samples_per_second": 35.075,
-      "eval_steps_per_second": 4.384,
-      "eval_wer": 0.4103004429607755,
+      "eval_loss": 0.5071681141853333,
+      "eval_runtime": 152.7102,
+      "eval_samples_per_second": 37.037,
+      "eval_steps_per_second": 4.63,
+      "eval_wer": 0.41930410220196446,
       "step": 1600
     },
     {
       "epoch": 2.4045261669024045,
-      "eval_loss": 0.49603915214538574,
-      "eval_runtime": 160.5657,
-      "eval_samples_per_second": 35.225,
-      "eval_steps_per_second": 4.403,
-      "eval_wer": 0.4084226744559286,
+      "eval_loss": 0.49842193722724915,
+      "eval_runtime": 153.4952,
+      "eval_samples_per_second": 36.848,
+      "eval_steps_per_second": 4.606,
+      "eval_wer": 0.41546831867496953,
       "step": 1700
     },
     {
       "epoch": 2.545968882602546,
-      "eval_loss": 0.4876905679702759,
-      "eval_runtime": 161.8676,
-      "eval_samples_per_second": 34.942,
-      "eval_steps_per_second": 4.368,
-      "eval_wer": 0.40148937536110935,
+      "eval_loss": 0.4882013499736786,
+      "eval_runtime": 152.4926,
+      "eval_samples_per_second": 37.09,
+      "eval_steps_per_second": 4.636,
+      "eval_wer": 0.4096905694292868,
       "step": 1800
     },
     {
       "epoch": 2.6874115983026874,
-      "eval_loss": 0.47859108448028564,
-      "eval_runtime": 161.5548,
-      "eval_samples_per_second": 35.01,
-      "eval_steps_per_second": 4.376,
-      "eval_wer": 0.39929062078705785,
+      "eval_loss": 0.4803565740585327,
+      "eval_runtime": 152.0827,
+      "eval_samples_per_second": 37.19,
+      "eval_steps_per_second": 4.649,
+      "eval_wer": 0.40800539256596263,
       "step": 1900
     },
     {
       "epoch": 2.828854314002829,
-      "grad_norm": 0.5755736231803894,
-      "learning_rate": 7.664999999999999e-05,
-      "loss": 0.5319,
+      "grad_norm": 0.7038583159446716,
+      "learning_rate": 7.68e-05,
+      "loss": 0.537,
       "step": 2000
     },
     {
       "epoch": 2.828854314002829,
-      "eval_loss": 0.47312092781066895,
-      "eval_runtime": 162.5051,
-      "eval_samples_per_second": 34.805,
-      "eval_steps_per_second": 4.351,
-      "eval_wer": 0.39299929383064774,
+      "eval_loss": 0.470017671585083,
+      "eval_runtime": 152.5003,
+      "eval_samples_per_second": 37.088,
+      "eval_steps_per_second": 4.636,
+      "eval_wer": 0.39272645567182385,
       "step": 2000
     },
     {
       "epoch": 2.9702970297029703,
-      "eval_loss": 0.4668172001838684,
-      "eval_runtime": 162.21,
-      "eval_samples_per_second": 34.868,
-      "eval_steps_per_second": 4.359,
-      "eval_wer": 0.3877190729922321,
+      "eval_loss": 0.46773043274879456,
+      "eval_runtime": 152.3208,
+      "eval_samples_per_second": 37.132,
+      "eval_steps_per_second": 4.642,
+      "eval_wer": 0.3884573409514027,
       "step": 2100
     },
     {
       "epoch": 3.1117397454031117,
-      "eval_loss": 0.46727854013442993,
-      "eval_runtime": 161.4684,
-      "eval_samples_per_second": 35.029,
-      "eval_steps_per_second": 4.379,
-      "eval_wer": 0.3849585927970726,
+      "eval_loss": 0.4683248996734619,
+      "eval_runtime": 152.8021,
+      "eval_samples_per_second": 37.015,
+      "eval_steps_per_second": 4.627,
+      "eval_wer": 0.38572895936316365,
       "step": 2200
     },
     {
       "epoch": 3.253182461103253,
-      "eval_loss": 0.46300554275512695,
-      "eval_runtime": 161.1936,
-      "eval_samples_per_second": 35.088,
-      "eval_steps_per_second": 4.386,
-      "eval_wer": 0.3804326892212878,
+      "eval_loss": 0.4618384838104248,
+      "eval_runtime": 152.4829,
+      "eval_samples_per_second": 37.093,
+      "eval_steps_per_second": 4.637,
+      "eval_wer": 0.3792289914617706,
       "step": 2300
     },
     {
       "epoch": 3.3946251768033946,
-      "eval_loss": 0.4593857228755951,
-      "eval_runtime": 161.5089,
-      "eval_samples_per_second": 35.02,
-      "eval_steps_per_second": 4.377,
-      "eval_wer": 0.3768697438531168,
+      "eval_loss": 0.46039119362831116,
+      "eval_runtime": 152.4542,
+      "eval_samples_per_second": 37.1,
+      "eval_steps_per_second": 4.637,
+      "eval_wer": 0.3762759196250883,
       "step": 2400
     },
     {
       "epoch": 3.536067892503536,
-      "grad_norm": 0.887208104133606,
-      "learning_rate": 1.9499999999999995e-06,
-      "loss": 0.4355,
+      "grad_norm": 0.6644078493118286,
+      "learning_rate": 2.1e-06,
+      "loss": 0.4434,
       "step": 2500
     },
     {
       "epoch": 3.536067892503536,
-      "eval_loss": 0.4583967626094818,
-      "eval_runtime": 162.1219,
-      "eval_samples_per_second": 34.887,
-      "eval_steps_per_second": 4.361,
-      "eval_wer": 0.37831418116453747,
+      "eval_loss": 0.4588949382305145,
+      "eval_runtime": 152.8982,
+      "eval_samples_per_second": 36.992,
+      "eval_steps_per_second": 4.624,
+      "eval_wer": 0.3742697566925595,
       "step": 2500
     },
     {
       "epoch": 3.536067892503536,
       "step": 2500,
       "total_flos": 9.55169606524761e+18,
-      "train_loss": 1.2462444946289062,
-      "train_runtime": 6962.0027,
-      "train_samples_per_second": 11.491,
-      "train_steps_per_second": 0.359
+      "train_loss": 1.2622116821289062,
+      "train_runtime": 6522.8506,
+      "train_samples_per_second": 12.265,
+      "train_steps_per_second": 0.383
     }
   ],
   "logging_steps": 500,