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: 0.4593
 - Wer: 0.3553

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 4.024767801857585,
-    "eval_loss": 0.49710649251937866,
-    "eval_runtime": 152.3497,
+    "epoch": 1.4722536806342017,
+    "eval_loss": 0.4593363106250763,
+    "eval_runtime": 154.8211,
     "eval_samples": 5656,
-    "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": 16.593,
-    "train_steps_per_second": 0.259
+    "eval_samples_per_second": 36.532,
+    "eval_steps_per_second": 4.567,
+    "eval_wer": 0.3553465680216976,
+    "total_flos": 2.3804122051094954e+19,
+    "train_loss": 0.9595056893275334,
+    "train_runtime": 10902.6713,
+    "train_samples": 112985,
+    "train_samples_per_second": 15.262,
+    "train_steps_per_second": 0.238
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 4.024767801857585,
-    "eval_loss": 0.49710649251937866,
-    "eval_runtime": 152.3497,
+    "epoch": 1.4722536806342017,
+    "eval_loss": 0.4593363106250763,
+    "eval_runtime": 154.8211,
     "eval_samples": 5656,
-    "eval_samples_per_second": 37.125,
-    "eval_steps_per_second": 4.641,
-    "eval_wer": 0.3380783489271557
+    "eval_samples_per_second": 36.532,
+    "eval_steps_per_second": 4.567,
+    "eval_wer": 0.3553465680216976
 }

runs/May27_10-13-44_5cce808c63ba/events.out.tfevents.1716813579.5cce808c63ba.31.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62117fda0463daeb87a269bfcca40e3e57857043da6d8f8c6636d4d11486f139
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 4.024767801857585,
-    "total_flos": 2.3576121775744893e+19,
-    "train_loss": 1.074923675243671,
-    "train_runtime": 10028.377,
-    "train_samples": 41326,
-    "train_samples_per_second": 16.593,
-    "train_steps_per_second": 0.259
+    "epoch": 1.4722536806342017,
+    "total_flos": 2.3804122051094954e+19,
+    "train_loss": 0.9595056893275334,
+    "train_runtime": 10902.6713,
+    "train_samples": 112985,
+    "train_samples_per_second": 15.262,
+    "train_steps_per_second": 0.238
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 4.024767801857585,
+  "epoch": 1.4722536806342017,
   "eval_steps": 100,
   "global_step": 2600,
   "is_hyper_param_search": false,
@@ -9,290 +9,290 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.15479876160990713,
-      "eval_loss": 3.556110143661499,
-      "eval_runtime": 152.2069,
-      "eval_samples_per_second": 37.16,
-      "eval_steps_per_second": 4.645,
+      "epoch": 0.056625141562853906,
+      "eval_loss": 3.599447011947632,
+      "eval_runtime": 153.6415,
+      "eval_samples_per_second": 36.813,
+      "eval_steps_per_second": 4.602,
       "eval_wer": 1.0,
       "step": 100
     },
     {
-      "epoch": 0.30959752321981426,
-      "eval_loss": 3.3711678981781006,
-      "eval_runtime": 150.3074,
-      "eval_samples_per_second": 37.63,
-      "eval_steps_per_second": 4.704,
+      "epoch": 0.11325028312570781,
+      "eval_loss": 3.0223705768585205,
+      "eval_runtime": 151.1357,
+      "eval_samples_per_second": 37.423,
+      "eval_steps_per_second": 4.678,
       "eval_wer": 1.0,
       "step": 200
     },
     {
-      "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,
+      "epoch": 0.16987542468856173,
+      "eval_loss": 1.9626648426055908,
+      "eval_runtime": 153.3369,
+      "eval_samples_per_second": 36.886,
+      "eval_steps_per_second": 4.611,
+      "eval_wer": 0.9060519009484682,
       "step": 300
     },
     {
-      "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,
+      "epoch": 0.22650056625141562,
+      "eval_loss": 0.9921229481697083,
+      "eval_runtime": 154.0126,
+      "eval_samples_per_second": 36.724,
+      "eval_steps_per_second": 4.591,
+      "eval_wer": 0.6951581582706103,
       "step": 400
     },
     {
-      "epoch": 0.7739938080495357,
-      "grad_norm": 4.971697807312012,
-      "learning_rate": 0.00029699999999999996,
-      "loss": 3.7217,
+      "epoch": 0.28312570781426954,
+      "grad_norm": 1.7500211000442505,
+      "learning_rate": 0.0002982,
+      "loss": 3.6184,
       "step": 500
     },
     {
-      "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,
+      "epoch": 0.28312570781426954,
+      "eval_loss": 0.8504024147987366,
+      "eval_runtime": 153.4805,
+      "eval_samples_per_second": 36.852,
+      "eval_steps_per_second": 4.606,
+      "eval_wer": 0.6157981736771998,
       "step": 500
     },
     {
-      "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,
+      "epoch": 0.33975084937712347,
+      "eval_loss": 0.7935531735420227,
+      "eval_runtime": 152.4296,
+      "eval_samples_per_second": 37.106,
+      "eval_steps_per_second": 4.638,
+      "eval_wer": 0.5963634029304618,
       "step": 600
     },
     {
-      "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,
+      "epoch": 0.39637599093997733,
+      "eval_loss": 0.7600815892219543,
+      "eval_runtime": 156.4822,
+      "eval_samples_per_second": 36.145,
+      "eval_steps_per_second": 4.518,
+      "eval_wer": 0.5606072764038452,
       "step": 700
     },
     {
-      "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,
+      "epoch": 0.45300113250283125,
+      "eval_loss": 0.6909334063529968,
+      "eval_runtime": 152.429,
+      "eval_samples_per_second": 37.106,
+      "eval_steps_per_second": 4.638,
+      "eval_wer": 0.5127666062172008,
       "step": 800
     },
     {
-      "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,
+      "epoch": 0.5096262740656852,
+      "eval_loss": 0.6616554856300354,
+      "eval_runtime": 152.3205,
+      "eval_samples_per_second": 37.132,
+      "eval_steps_per_second": 4.642,
+      "eval_wer": 0.4916146426794627,
       "step": 900
     },
     {
-      "epoch": 1.5479876160990713,
-      "grad_norm": 0.40550583600997925,
-      "learning_rate": 0.00022971428571428568,
-      "loss": 0.6148,
+      "epoch": 0.5662514156285391,
+      "grad_norm": 3.9567878246307373,
+      "learning_rate": 0.00022914285714285712,
+      "loss": 0.4662,
       "step": 1000
     },
     {
-      "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,
+      "epoch": 0.5662514156285391,
+      "eval_loss": 0.6466770172119141,
+      "eval_runtime": 152.3467,
+      "eval_samples_per_second": 37.126,
+      "eval_steps_per_second": 4.641,
+      "eval_wer": 0.4811510006258927,
       "step": 1000
     },
     {
-      "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,
+      "epoch": 0.622876557191393,
+      "eval_loss": 0.6144490838050842,
+      "eval_runtime": 152.8109,
+      "eval_samples_per_second": 37.013,
+      "eval_steps_per_second": 4.627,
+      "eval_wer": 0.46394697565437887,
       "step": 1100
     },
     {
-      "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,
+      "epoch": 0.6795016987542469,
+      "eval_loss": 0.5942133069038391,
+      "eval_runtime": 157.9769,
+      "eval_samples_per_second": 35.803,
+      "eval_steps_per_second": 4.475,
+      "eval_wer": 0.45369196450064997,
       "step": 1200
     },
     {
-      "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,
+      "epoch": 0.7361268403171007,
+      "eval_loss": 0.5675162076950073,
+      "eval_runtime": 165.0294,
+      "eval_samples_per_second": 34.273,
+      "eval_steps_per_second": 4.284,
+      "eval_wer": 0.4351077658840333,
       "step": 1300
     },
     {
-      "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,
+      "epoch": 0.7927519818799547,
+      "eval_loss": 0.5538555383682251,
+      "eval_runtime": 154.4584,
+      "eval_samples_per_second": 36.618,
+      "eval_steps_per_second": 4.577,
+      "eval_wer": 0.42295902810097735,
       "step": 1400
     },
     {
-      "epoch": 2.321981424148607,
-      "grad_norm": 0.44798457622528076,
-      "learning_rate": 0.0001584285714285714,
-      "loss": 0.4634,
+      "epoch": 0.8493771234428086,
+      "grad_norm": 1.9575612545013428,
+      "learning_rate": 0.00015785714285714285,
+      "loss": 0.3508,
       "step": 1500
     },
     {
-      "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,
+      "epoch": 0.8493771234428086,
+      "eval_loss": 0.5447892546653748,
+      "eval_runtime": 152.2149,
+      "eval_samples_per_second": 37.158,
+      "eval_steps_per_second": 4.645,
+      "eval_wer": 0.41440516120749143,
       "step": 1500
     },
     {
-      "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,
+      "epoch": 0.9060022650056625,
+      "eval_loss": 0.5326293110847473,
+      "eval_runtime": 157.3458,
+      "eval_samples_per_second": 35.946,
+      "eval_steps_per_second": 4.493,
+      "eval_wer": 0.406621623790342,
       "step": 1600
     },
     {
-      "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,
+      "epoch": 0.9626274065685164,
+      "eval_loss": 0.5154709219932556,
+      "eval_runtime": 154.1323,
+      "eval_samples_per_second": 36.696,
+      "eval_steps_per_second": 4.587,
+      "eval_wer": 0.3989343775577346,
       "step": 1700
     },
     {
-      "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,
+      "epoch": 1.0192525481313703,
+      "eval_loss": 0.5067195296287537,
+      "eval_runtime": 153.6675,
+      "eval_samples_per_second": 36.807,
+      "eval_steps_per_second": 4.601,
+      "eval_wer": 0.38559804849866,
       "step": 1800
     },
     {
-      "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,
+      "epoch": 1.0758776896942241,
+      "eval_loss": 0.49161842465400696,
+      "eval_runtime": 153.2473,
+      "eval_samples_per_second": 36.908,
+      "eval_steps_per_second": 4.613,
+      "eval_wer": 0.3724061562163984,
       "step": 1900
     },
     {
-      "epoch": 3.0959752321981426,
-      "grad_norm": 0.39644455909729004,
-      "learning_rate": 8.742857142857142e-05,
-      "loss": 0.3708,
+      "epoch": 1.1325028312570782,
+      "grad_norm": 0.9259862303733826,
+      "learning_rate": 8.671428571428571e-05,
+      "loss": 0.2774,
       "step": 2000
     },
     {
-      "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,
+      "epoch": 1.1325028312570782,
+      "eval_loss": 0.48551619052886963,
+      "eval_runtime": 153.8121,
+      "eval_samples_per_second": 36.772,
+      "eval_steps_per_second": 4.597,
+      "eval_wer": 0.36971000304922086,
       "step": 2000
     },
     {
-      "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,
+      "epoch": 1.189127972819932,
+      "eval_loss": 0.4797590672969818,
+      "eval_runtime": 152.9797,
+      "eval_samples_per_second": 36.972,
+      "eval_steps_per_second": 4.622,
+      "eval_wer": 0.3661311806904078,
       "step": 2100
     },
     {
-      "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,
+      "epoch": 1.245753114382786,
+      "eval_loss": 0.47735241055488586,
+      "eval_runtime": 152.7086,
+      "eval_samples_per_second": 37.038,
+      "eval_steps_per_second": 4.63,
+      "eval_wer": 0.36462261879924895,
       "step": 2200
     },
     {
-      "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,
+      "epoch": 1.3023782559456398,
+      "eval_loss": 0.46992629766464233,
+      "eval_runtime": 152.5925,
+      "eval_samples_per_second": 37.066,
+      "eval_steps_per_second": 4.633,
+      "eval_wer": 0.3584760315193144,
       "step": 2300
     },
     {
-      "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,
+      "epoch": 1.3590033975084936,
+      "eval_loss": 0.4651219844818115,
+      "eval_runtime": 153.5802,
+      "eval_samples_per_second": 36.828,
+      "eval_steps_per_second": 4.603,
+      "eval_wer": 0.3550255974065574,
       "step": 2400
     },
     {
-      "epoch": 3.8699690402476783,
-      "grad_norm": 6.033559322357178,
-      "learning_rate": 1.642857142857143e-05,
-      "loss": 0.3391,
+      "epoch": 1.4156285390713477,
+      "grad_norm": 0.43387308716773987,
+      "learning_rate": 1.5428571428571428e-05,
+      "loss": 0.2328,
       "step": 2500
     },
     {
-      "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,
+      "epoch": 1.4156285390713477,
+      "eval_loss": 0.4611109495162964,
+      "eval_runtime": 153.4779,
+      "eval_samples_per_second": 36.852,
+      "eval_steps_per_second": 4.607,
+      "eval_wer": 0.3570156152204266,
       "step": 2500
     },
     {
-      "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,
+      "epoch": 1.4722536806342017,
+      "eval_loss": 0.4593363106250763,
+      "eval_runtime": 153.5126,
+      "eval_samples_per_second": 36.844,
+      "eval_steps_per_second": 4.605,
+      "eval_wer": 0.3553465680216976,
       "step": 2600
     },
     {
-      "epoch": 4.024767801857585,
+      "epoch": 1.4722536806342017,
       "step": 2600,
-      "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
+      "total_flos": 2.3804122051094954e+19,
+      "train_loss": 0.9595056893275334,
+      "train_runtime": 10902.6713,
+      "train_samples_per_second": 15.262,
+      "train_steps_per_second": 0.238
     }
   ],
   "logging_steps": 500,
   "max_steps": 2600,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 5,
+  "num_train_epochs": 2,
   "save_steps": 400,
-  "total_flos": 2.3576121775744893e+19,
+  "total_flos": 2.3804122051094954e+19,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null