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: 0.4159
 - Wer: 0.3171

all_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 3.6775106082036775,
-    "eval_loss": 0.41045042872428894,
-    "eval_runtime": 151.6211,
+    "eval_loss": 0.4158582091331482,
+    "eval_runtime": 162.0151,
     "eval_samples": 5656,
-    "eval_samples_per_second": 37.304,
-    "eval_steps_per_second": 4.663,
-    "eval_wer": 0.3136203880534737,
-    "total_flos": 9.93123920141323e+18,
-    "train_loss": 1.040142478942871,
-    "train_runtime": 6732.1803,
+    "eval_samples_per_second": 34.91,
+    "eval_steps_per_second": 4.364,
+    "eval_wer": 0.3171350162892587,
+    "total_flos": 9.931497341662648e+18,
+    "train_loss": 1.0819015429570125,
+    "train_runtime": 7138.5232,
     "train_samples": 22621,
-    "train_samples_per_second": 12.359,
-    "train_steps_per_second": 0.386
+    "train_samples_per_second": 11.655,
+    "train_steps_per_second": 0.364
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 3.6775106082036775,
-    "eval_loss": 0.41045042872428894,
-    "eval_runtime": 151.6211,
+    "eval_loss": 0.4158582091331482,
+    "eval_runtime": 162.0151,
     "eval_samples": 5656,
-    "eval_samples_per_second": 37.304,
-    "eval_steps_per_second": 4.663,
-    "eval_wer": 0.3136203880534737
+    "eval_samples_per_second": 34.91,
+    "eval_steps_per_second": 4.364,
+    "eval_wer": 0.3171350162892587
 }

runs/May17_07-56-01_2e698a9c250f/events.out.tfevents.1715936676.2e698a9c250f.3965.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1760110344ee9b88297ea2ffcbe874107a3a12cb3b4993d80c29cbe3f5fd8177
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 3.6775106082036775,
-    "total_flos": 9.93123920141323e+18,
-    "train_loss": 1.040142478942871,
-    "train_runtime": 6732.1803,
+    "total_flos": 9.931497341662648e+18,
+    "train_loss": 1.0819015429570125,
+    "train_runtime": 7138.5232,
     "train_samples": 22621,
-    "train_samples_per_second": 12.359,
-    "train_steps_per_second": 0.386
+    "train_samples_per_second": 11.655,
+    "train_steps_per_second": 0.364
 }

trainer_state.json

@@ -10,281 +10,281 @@
   "log_history": [
     {
       "epoch": 0.14144271570014144,
-      "eval_loss": 3.2194931507110596,
-      "eval_runtime": 152.4732,
-      "eval_samples_per_second": 37.095,
-      "eval_steps_per_second": 4.637,
+      "eval_loss": 3.3999907970428467,
+      "eval_runtime": 162.0918,
+      "eval_samples_per_second": 34.894,
+      "eval_steps_per_second": 4.362,
       "eval_wer": 1.0,
       "step": 100
     },
     {
       "epoch": 0.2828854314002829,
-      "eval_loss": 3.1180293560028076,
-      "eval_runtime": 149.651,
-      "eval_samples_per_second": 37.795,
-      "eval_steps_per_second": 4.724,
+      "eval_loss": 2.9512617588043213,
+      "eval_runtime": 159.7591,
+      "eval_samples_per_second": 35.403,
+      "eval_steps_per_second": 4.425,
       "eval_wer": 1.0,
       "step": 200
     },
     {
       "epoch": 0.4243281471004243,
-      "eval_loss": 1.7235759496688843,
-      "eval_runtime": 150.6233,
-      "eval_samples_per_second": 37.551,
-      "eval_steps_per_second": 4.694,
-      "eval_wer": 0.9364317696714866,
+      "eval_loss": 1.9453805685043335,
+      "eval_runtime": 160.8922,
+      "eval_samples_per_second": 35.154,
+      "eval_steps_per_second": 4.394,
+      "eval_wer": 0.9325480252282904,
       "step": 300
     },
     {
       "epoch": 0.5657708628005658,
-      "eval_loss": 1.157871127128601,
-      "eval_runtime": 150.7722,
-      "eval_samples_per_second": 37.514,
-      "eval_steps_per_second": 4.689,
-      "eval_wer": 0.76277061834989,
+      "eval_loss": 1.1412426233291626,
+      "eval_runtime": 161.9235,
+      "eval_samples_per_second": 34.93,
+      "eval_steps_per_second": 4.366,
+      "eval_wer": 0.773025629503619,
       "step": 400
     },
     {
       "epoch": 0.7072135785007072,
-      "grad_norm": 2.786440372467041,
-      "learning_rate": 0.0002964,
-      "loss": 3.0279,
+      "grad_norm": 2.79933500289917,
+      "learning_rate": 0.00029699999999999996,
+      "loss": 3.2243,
       "step": 500
     },
     {
       "epoch": 0.7072135785007072,
-      "eval_loss": 0.9558188319206238,
-      "eval_runtime": 149.998,
-      "eval_samples_per_second": 37.707,
-      "eval_steps_per_second": 4.713,
-      "eval_wer": 0.6721285166343022,
+      "eval_loss": 0.9250076413154602,
+      "eval_runtime": 160.568,
+      "eval_samples_per_second": 35.225,
+      "eval_steps_per_second": 4.403,
+      "eval_wer": 0.67890099661376,
       "step": 500
     },
     {
       "epoch": 0.8486562942008486,
-      "eval_loss": 0.8188507556915283,
-      "eval_runtime": 150.6554,
-      "eval_samples_per_second": 37.543,
-      "eval_steps_per_second": 4.693,
-      "eval_wer": 0.6423905891415641,
+      "eval_loss": 0.8018025159835815,
+      "eval_runtime": 160.6715,
+      "eval_samples_per_second": 35.202,
+      "eval_steps_per_second": 4.4,
+      "eval_wer": 0.5982892266213028,
       "step": 600
     },
     {
       "epoch": 0.9900990099009901,
-      "eval_loss": 0.6972773671150208,
-      "eval_runtime": 152.1964,
-      "eval_samples_per_second": 37.163,
-      "eval_steps_per_second": 4.645,
-      "eval_wer": 0.5163775256375279,
+      "eval_loss": 0.7181887030601501,
+      "eval_runtime": 160.5511,
+      "eval_samples_per_second": 35.229,
+      "eval_steps_per_second": 4.404,
+      "eval_wer": 0.5408675835727239,
       "step": 700
     },
     {
       "epoch": 1.1315417256011315,
-      "eval_loss": 0.6183159947395325,
-      "eval_runtime": 151.4826,
-      "eval_samples_per_second": 37.338,
-      "eval_steps_per_second": 4.667,
-      "eval_wer": 0.4751809471842853,
+      "eval_loss": 0.6198393106460571,
+      "eval_runtime": 161.1454,
+      "eval_samples_per_second": 35.099,
+      "eval_steps_per_second": 4.387,
+      "eval_wer": 0.4769944311598273,
       "step": 800
     },
     {
       "epoch": 1.272984441301273,
-      "eval_loss": 0.5935864448547363,
-      "eval_runtime": 150.8934,
-      "eval_samples_per_second": 37.483,
-      "eval_steps_per_second": 4.685,
-      "eval_wer": 0.47028614530339746,
+      "eval_loss": 0.6102356910705566,
+      "eval_runtime": 160.7517,
+      "eval_samples_per_second": 35.185,
+      "eval_steps_per_second": 4.398,
+      "eval_wer": 0.47124905714881804,
       "step": 900
     },
     {
       "epoch": 1.4144271570014144,
-      "grad_norm": 1.0183278322219849,
-      "learning_rate": 0.0002294285714285714,
-      "loss": 0.7925,
+      "grad_norm": 1.0946769714355469,
+      "learning_rate": 0.00022928571428571426,
+      "loss": 0.7983,
       "step": 1000
     },
     {
       "epoch": 1.4144271570014144,
-      "eval_loss": 0.5497952103614807,
-      "eval_runtime": 151.0928,
-      "eval_samples_per_second": 37.434,
-      "eval_steps_per_second": 4.679,
-      "eval_wer": 0.4304055463722296,
+      "eval_loss": 0.5605342984199524,
+      "eval_runtime": 160.7985,
+      "eval_samples_per_second": 35.174,
+      "eval_steps_per_second": 4.397,
+      "eval_wer": 0.4425863812167996,
       "step": 1000
     },
     {
       "epoch": 1.5558698727015559,
-      "eval_loss": 0.5285686254501343,
-      "eval_runtime": 151.3525,
-      "eval_samples_per_second": 37.37,
-      "eval_steps_per_second": 4.671,
-      "eval_wer": 0.4148224230071737,
+      "eval_loss": 0.5336170196533203,
+      "eval_runtime": 160.9878,
+      "eval_samples_per_second": 35.133,
+      "eval_steps_per_second": 4.392,
+      "eval_wer": 0.4162186451830335,
       "step": 1100
     },
     {
       "epoch": 1.6973125884016973,
-      "eval_loss": 0.5129849314689636,
-      "eval_runtime": 151.9884,
-      "eval_samples_per_second": 37.213,
-      "eval_steps_per_second": 4.652,
-      "eval_wer": 0.39739371860506173,
+      "eval_loss": 0.5258753299713135,
+      "eval_runtime": 161.623,
+      "eval_samples_per_second": 34.995,
+      "eval_steps_per_second": 4.374,
+      "eval_wer": 0.41161271685577183,
       "step": 1200
     },
     {
       "epoch": 1.8387553041018387,
-      "eval_loss": 0.48775196075439453,
-      "eval_runtime": 152.7263,
-      "eval_samples_per_second": 37.034,
-      "eval_steps_per_second": 4.629,
-      "eval_wer": 0.3863844265057534,
+      "eval_loss": 0.4960057735443115,
+      "eval_runtime": 161.2649,
+      "eval_samples_per_second": 35.073,
+      "eval_steps_per_second": 4.384,
+      "eval_wer": 0.3872349986358749,
       "step": 1300
     },
     {
       "epoch": 1.9801980198019802,
-      "eval_loss": 0.4740111827850342,
-      "eval_runtime": 152.3303,
-      "eval_samples_per_second": 37.13,
-      "eval_steps_per_second": 4.641,
-      "eval_wer": 0.3733209224695479,
+      "eval_loss": 0.48569169640541077,
+      "eval_runtime": 161.7331,
+      "eval_samples_per_second": 34.971,
+      "eval_steps_per_second": 4.371,
+      "eval_wer": 0.38676959124392163,
       "step": 1400
     },
     {
       "epoch": 2.1216407355021216,
-      "grad_norm": 0.966443657875061,
-      "learning_rate": 0.0001584285714285714,
-      "loss": 0.62,
+      "grad_norm": 1.4740066528320312,
+      "learning_rate": 0.0001582857142857143,
+      "loss": 0.6274,
       "step": 1500
     },
     {
       "epoch": 2.1216407355021216,
-      "eval_loss": 0.45776596665382385,
-      "eval_runtime": 152.151,
-      "eval_samples_per_second": 37.174,
-      "eval_steps_per_second": 4.647,
-      "eval_wer": 0.3593266036494359,
+      "eval_loss": 0.46894121170043945,
+      "eval_runtime": 161.6698,
+      "eval_samples_per_second": 34.985,
+      "eval_steps_per_second": 4.373,
+      "eval_wer": 0.36555343358315545,
       "step": 1500
     },
     {
       "epoch": 2.263083451202263,
-      "eval_loss": 0.45351675152778625,
-      "eval_runtime": 151.9696,
-      "eval_samples_per_second": 37.218,
-      "eval_steps_per_second": 4.652,
-      "eval_wer": 0.35000240727961357,
+      "eval_loss": 0.4679875373840332,
+      "eval_runtime": 161.7805,
+      "eval_samples_per_second": 34.961,
+      "eval_steps_per_second": 4.37,
+      "eval_wer": 0.3562131886825761,
       "step": 1600
     },
     {
       "epoch": 2.4045261669024045,
-      "eval_loss": 0.44846034049987793,
-      "eval_runtime": 152.2898,
-      "eval_samples_per_second": 37.14,
-      "eval_steps_per_second": 4.642,
-      "eval_wer": 0.3496814366644734,
+      "eval_loss": 0.4536493718624115,
+      "eval_runtime": 162.2799,
+      "eval_samples_per_second": 34.853,
+      "eval_steps_per_second": 4.357,
+      "eval_wer": 0.3535330840461556,
       "step": 1700
     },
     {
       "epoch": 2.545968882602546,
-      "eval_loss": 0.4373452067375183,
-      "eval_runtime": 152.3551,
-      "eval_samples_per_second": 37.124,
-      "eval_steps_per_second": 4.64,
-      "eval_wer": 0.34305339346182856,
+      "eval_loss": 0.44862428307533264,
+      "eval_runtime": 161.7828,
+      "eval_samples_per_second": 34.96,
+      "eval_steps_per_second": 4.37,
+      "eval_wer": 0.3500826499333986,
       "step": 1800
     },
     {
       "epoch": 2.6874115983026874,
-      "eval_loss": 0.4362298250198364,
-      "eval_runtime": 151.9091,
-      "eval_samples_per_second": 37.233,
-      "eval_steps_per_second": 4.654,
-      "eval_wer": 0.34286081109274447,
+      "eval_loss": 0.43955981731414795,
+      "eval_runtime": 162.5762,
+      "eval_samples_per_second": 34.79,
+      "eval_steps_per_second": 4.349,
+      "eval_wer": 0.35051596026383786,
       "step": 1900
     },
     {
       "epoch": 2.828854314002829,
-      "grad_norm": 0.6778371334075928,
-      "learning_rate": 8.742857142857142e-05,
-      "loss": 0.4879,
+      "grad_norm": 0.927208662033081,
+      "learning_rate": 8.728571428571428e-05,
+      "loss": 0.4939,
       "step": 2000
     },
     {
       "epoch": 2.828854314002829,
-      "eval_loss": 0.4235607981681824,
-      "eval_runtime": 152.5551,
-      "eval_samples_per_second": 37.075,
-      "eval_steps_per_second": 4.634,
-      "eval_wer": 0.33270209112355764,
+      "eval_loss": 0.4298574924468994,
+      "eval_runtime": 162.5099,
+      "eval_samples_per_second": 34.804,
+      "eval_steps_per_second": 4.351,
+      "eval_wer": 0.3340020221148754,
       "step": 2000
     },
     {
       "epoch": 2.9702970297029703,
-      "eval_loss": 0.4171510338783264,
-      "eval_runtime": 152.8349,
-      "eval_samples_per_second": 37.007,
-      "eval_steps_per_second": 4.626,
-      "eval_wer": 0.3256728346519876,
+      "eval_loss": 0.4291674494743347,
+      "eval_runtime": 162.0566,
+      "eval_samples_per_second": 34.901,
+      "eval_steps_per_second": 4.363,
+      "eval_wer": 0.33114538364012774,
       "step": 2100
     },
     {
       "epoch": 3.1117397454031117,
-      "eval_loss": 0.4206392467021942,
-      "eval_runtime": 152.8774,
-      "eval_samples_per_second": 36.997,
-      "eval_steps_per_second": 4.625,
-      "eval_wer": 0.32169279902424935,
+      "eval_loss": 0.42761147022247314,
+      "eval_runtime": 162.7807,
+      "eval_samples_per_second": 34.746,
+      "eval_steps_per_second": 4.343,
+      "eval_wer": 0.3270690568278474,
       "step": 2200
     },
     {
       "epoch": 3.253182461103253,
-      "eval_loss": 0.41660308837890625,
-      "eval_runtime": 152.074,
-      "eval_samples_per_second": 37.192,
-      "eval_steps_per_second": 4.649,
-      "eval_wer": 0.3198632665179503,
+      "eval_loss": 0.4232628345489502,
+      "eval_runtime": 161.9234,
+      "eval_samples_per_second": 34.93,
+      "eval_steps_per_second": 4.366,
+      "eval_wer": 0.3260419508593988,
       "step": 2300
     },
     {
       "epoch": 3.3946251768033946,
-      "eval_loss": 0.41343268752098083,
-      "eval_runtime": 152.5935,
-      "eval_samples_per_second": 37.066,
-      "eval_steps_per_second": 4.633,
-      "eval_wer": 0.3172955015968288,
+      "eval_loss": 0.4192351996898651,
+      "eval_runtime": 162.4093,
+      "eval_samples_per_second": 34.826,
+      "eval_steps_per_second": 4.353,
+      "eval_wer": 0.32230264319301566,
       "step": 2400
     },
     {
       "epoch": 3.536067892503536,
-      "grad_norm": 0.7615213990211487,
-      "learning_rate": 1.6285714285714283e-05,
-      "loss": 0.4036,
+      "grad_norm": 0.7725916504859924,
+      "learning_rate": 1.614285714285714e-05,
+      "loss": 0.4072,
       "step": 2500
     },
     {
       "epoch": 3.536067892503536,
-      "eval_loss": 0.41102421283721924,
-      "eval_runtime": 151.5019,
-      "eval_samples_per_second": 37.333,
-      "eval_steps_per_second": 4.667,
-      "eval_wer": 0.31594742501324,
+      "eval_loss": 0.4178767800331116,
+      "eval_runtime": 162.2884,
+      "eval_samples_per_second": 34.852,
+      "eval_steps_per_second": 4.356,
+      "eval_wer": 0.31946205324902505,
       "step": 2500
     },
     {
       "epoch": 3.6775106082036775,
-      "eval_loss": 0.41045042872428894,
-      "eval_runtime": 152.2158,
-      "eval_samples_per_second": 37.158,
-      "eval_steps_per_second": 4.645,
-      "eval_wer": 0.3136203880534737,
+      "eval_loss": 0.4158582091331482,
+      "eval_runtime": 162.2061,
+      "eval_samples_per_second": 34.869,
+      "eval_steps_per_second": 4.359,
+      "eval_wer": 0.3171350162892587,
       "step": 2600
     },
     {
       "epoch": 3.6775106082036775,
       "step": 2600,
-      "total_flos": 9.93123920141323e+18,
-      "train_loss": 1.040142478942871,
-      "train_runtime": 6732.1803,
-      "train_samples_per_second": 12.359,
-      "train_steps_per_second": 0.386
+      "total_flos": 9.931497341662648e+18,
+      "train_loss": 1.0819015429570125,
+      "train_runtime": 7138.5232,
+      "train_samples_per_second": 11.655,
+      "train_steps_per_second": 0.364
     }
   ],
   "logging_steps": 500,
@@ -292,7 +292,7 @@
   "num_input_tokens_seen": 0,
   "num_train_epochs": 4,
   "save_steps": 400,
-  "total_flos": 9.93123920141323e+18,
+  "total_flos": 9.931497341662648e+18,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null