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: 0.5865
 - Wer: 0.4452

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 1.008233910267182,
-    "eval_loss": 0.5155353546142578,
-    "eval_runtime": 186.6856,
+    "epoch": 0.38476337052712584,
+    "eval_loss": 0.5865370631217957,
+    "eval_runtime": 179.3337,
     "eval_samples": 7022,
-    "eval_samples_per_second": 37.614,
-    "eval_steps_per_second": 2.352,
-    "eval_wer": 0.39834438406037814,
-    "total_flos": 6.665403914161851e+18,
-    "train_loss": 1.086754165649414,
-    "train_runtime": 4718.794,
-    "train_samples": 47607,
-    "train_samples_per_second": 10.172,
-    "train_steps_per_second": 1.272
+    "eval_samples_per_second": 39.156,
+    "eval_steps_per_second": 2.448,
+    "eval_wer": 0.44523578575319894,
+    "total_flos": 6.835515215409967e+18,
+    "train_loss": 0.9274428431193034,
+    "train_runtime": 4544.4955,
+    "train_samples": 124748,
+    "train_samples_per_second": 10.562,
+    "train_steps_per_second": 1.32
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 1.008233910267182,
-    "eval_loss": 0.5155353546142578,
-    "eval_runtime": 186.6856,
+    "epoch": 0.38476337052712584,
+    "eval_loss": 0.5865370631217957,
+    "eval_runtime": 179.3337,
     "eval_samples": 7022,
-    "eval_samples_per_second": 37.614,
-    "eval_steps_per_second": 2.352,
-    "eval_wer": 0.39834438406037814
+    "eval_samples_per_second": 39.156,
+    "eval_steps_per_second": 2.448,
+    "eval_wer": 0.44523578575319894
 }

runs/Aug30_15-47-58_4b35055fdbcb/events.out.tfevents.1725037630.4b35055fdbcb.30.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2b50ba638d1137f7ceced29f612cfc608a9c7ade1ffa24786817e12c05671ce
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 1.008233910267182,
-    "total_flos": 6.665403914161851e+18,
-    "train_loss": 1.086754165649414,
-    "train_runtime": 4718.794,
-    "train_samples": 47607,
-    "train_samples_per_second": 10.172,
-    "train_steps_per_second": 1.272
+    "epoch": 0.38476337052712584,
+    "total_flos": 6.835515215409967e+18,
+    "train_loss": 0.9274428431193034,
+    "train_runtime": 4544.4955,
+    "train_samples": 124748,
+    "train_samples_per_second": 10.562,
+    "train_steps_per_second": 1.32
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.008233910267182,
+  "epoch": 0.38476337052712584,
   "eval_steps": 500,
   "global_step": 6000,
   "is_hyper_param_search": false,
@@ -9,211 +9,211 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.08401949252226516,
-      "grad_norm": 19.443201065063477,
-      "learning_rate": 0.0002455,
-      "loss": 4.6277,
+      "epoch": 0.03206361421059382,
+      "grad_norm": 10.316516876220703,
+      "learning_rate": 0.0002465,
+      "loss": 4.7051,
       "step": 500
     },
     {
-      "epoch": 0.08401949252226516,
-      "eval_loss": 1.7780689001083374,
-      "eval_runtime": 184.9372,
-      "eval_samples_per_second": 37.97,
-      "eval_steps_per_second": 2.374,
-      "eval_wer": 0.9507301452317996,
+      "epoch": 0.03206361421059382,
+      "eval_loss": 1.6561740636825562,
+      "eval_runtime": 191.0416,
+      "eval_samples_per_second": 36.756,
+      "eval_steps_per_second": 2.298,
+      "eval_wer": 0.9354894431230816,
       "step": 500
     },
     {
-      "epoch": 0.16803898504453033,
-      "grad_norm": 10.621150970458984,
-      "learning_rate": 0.0002782777777777778,
-      "loss": 1.2514,
+      "epoch": 0.06412722842118763,
+      "grad_norm": 4.912766933441162,
+      "learning_rate": 0.00027816666666666663,
+      "loss": 1.0362,
       "step": 1000
     },
     {
-      "epoch": 0.16803898504453033,
-      "eval_loss": 1.046749472618103,
-      "eval_runtime": 183.9043,
-      "eval_samples_per_second": 38.183,
-      "eval_steps_per_second": 2.387,
-      "eval_wer": 0.7116623925377695,
+      "epoch": 0.06412722842118763,
+      "eval_loss": 1.1839842796325684,
+      "eval_runtime": 175.5788,
+      "eval_samples_per_second": 39.993,
+      "eval_steps_per_second": 2.5,
+      "eval_wer": 0.7977517639086356,
       "step": 1000
     },
     {
-      "epoch": 0.2520584775667955,
-      "grad_norm": 6.651599884033203,
-      "learning_rate": 0.00025049999999999996,
-      "loss": 0.986,
+      "epoch": 0.09619084263178146,
+      "grad_norm": 4.178534030914307,
+      "learning_rate": 0.00025038888888888886,
+      "loss": 0.811,
       "step": 1500
     },
     {
-      "epoch": 0.2520584775667955,
-      "eval_loss": 0.8940034508705139,
-      "eval_runtime": 185.3076,
-      "eval_samples_per_second": 37.894,
-      "eval_steps_per_second": 2.369,
-      "eval_wer": 0.6577817935396431,
+      "epoch": 0.09619084263178146,
+      "eval_loss": 1.0081175565719604,
+      "eval_runtime": 175.5534,
+      "eval_samples_per_second": 39.999,
+      "eval_steps_per_second": 2.501,
+      "eval_wer": 0.7174690070290597,
       "step": 1500
     },
     {
-      "epoch": 0.33607797008906065,
-      "grad_norm": 16.908123016357422,
-      "learning_rate": 0.00022277777777777774,
-      "loss": 0.8702,
+      "epoch": 0.12825445684237527,
+      "grad_norm": 7.957609176635742,
+      "learning_rate": 0.0002226111111111111,
+      "loss": 0.6903,
       "step": 2000
     },
     {
-      "epoch": 0.33607797008906065,
-      "eval_loss": 0.8157631158828735,
-      "eval_runtime": 185.5186,
-      "eval_samples_per_second": 37.851,
-      "eval_steps_per_second": 2.366,
-      "eval_wer": 0.5836644122297665,
+      "epoch": 0.12825445684237527,
+      "eval_loss": 0.8935067057609558,
+      "eval_runtime": 176.1028,
+      "eval_samples_per_second": 39.874,
+      "eval_steps_per_second": 2.493,
+      "eval_wer": 0.6401094885661516,
       "step": 2000
     },
     {
-      "epoch": 0.4200974626113258,
-      "grad_norm": 10.515896797180176,
-      "learning_rate": 0.00019505555555555555,
-      "loss": 0.8063,
+      "epoch": 0.16031807105296908,
+      "grad_norm": 6.541042327880859,
+      "learning_rate": 0.00019483333333333332,
+      "loss": 0.6238,
       "step": 2500
     },
     {
-      "epoch": 0.4200974626113258,
-      "eval_loss": 0.7131509184837341,
-      "eval_runtime": 185.8391,
-      "eval_samples_per_second": 37.785,
-      "eval_steps_per_second": 2.362,
-      "eval_wer": 0.5281095948657304,
+      "epoch": 0.16031807105296908,
+      "eval_loss": 0.8059775233268738,
+      "eval_runtime": 177.0587,
+      "eval_samples_per_second": 39.659,
+      "eval_steps_per_second": 2.479,
+      "eval_wer": 0.5848868573858276,
       "step": 2500
     },
     {
-      "epoch": 0.504116955133591,
-      "grad_norm": 3.7379024028778076,
-      "learning_rate": 0.00016727777777777778,
-      "loss": 0.7518,
+      "epoch": 0.19238168526356292,
+      "grad_norm": 5.036514759063721,
+      "learning_rate": 0.00016705555555555554,
+      "loss": 0.5649,
       "step": 3000
     },
     {
-      "epoch": 0.504116955133591,
-      "eval_loss": 0.7195897102355957,
-      "eval_runtime": 183.5821,
-      "eval_samples_per_second": 38.25,
-      "eval_steps_per_second": 2.391,
-      "eval_wer": 0.5187950942744389,
+      "epoch": 0.19238168526356292,
+      "eval_loss": 0.7770201563835144,
+      "eval_runtime": 178.1454,
+      "eval_samples_per_second": 39.417,
+      "eval_steps_per_second": 2.464,
+      "eval_wer": 0.5589231852668783,
       "step": 3000
     },
     {
-      "epoch": 0.5881364476558562,
-      "grad_norm": 5.922393321990967,
-      "learning_rate": 0.0001395,
-      "loss": 0.7045,
+      "epoch": 0.22444529947415673,
+      "grad_norm": 4.140881538391113,
+      "learning_rate": 0.00013927777777777777,
+      "loss": 0.5309,
       "step": 3500
     },
     {
-      "epoch": 0.5881364476558562,
-      "eval_loss": 0.6231045722961426,
-      "eval_runtime": 184.2635,
-      "eval_samples_per_second": 38.108,
-      "eval_steps_per_second": 2.382,
-      "eval_wer": 0.4771256593895747,
+      "epoch": 0.22444529947415673,
+      "eval_loss": 0.7264481782913208,
+      "eval_runtime": 176.7301,
+      "eval_samples_per_second": 39.733,
+      "eval_steps_per_second": 2.484,
+      "eval_wer": 0.5326539018589139,
       "step": 3500
     },
     {
-      "epoch": 0.6721559401781213,
-      "grad_norm": 8.91370964050293,
-      "learning_rate": 0.00011177777777777777,
-      "loss": 0.6726,
+      "epoch": 0.25650891368475054,
+      "grad_norm": 6.205782413482666,
+      "learning_rate": 0.00011155555555555555,
+      "loss": 0.4892,
       "step": 4000
     },
     {
-      "epoch": 0.6721559401781213,
-      "eval_loss": 0.6074743270874023,
-      "eval_runtime": 184.3908,
-      "eval_samples_per_second": 38.082,
-      "eval_steps_per_second": 2.381,
-      "eval_wer": 0.4517333475066105,
+      "epoch": 0.25650891368475054,
+      "eval_loss": 0.6865007877349854,
+      "eval_runtime": 178.7785,
+      "eval_samples_per_second": 39.278,
+      "eval_steps_per_second": 2.456,
+      "eval_wer": 0.5106366016024662,
       "step": 4000
     },
     {
-      "epoch": 0.7561754327003865,
-      "grad_norm": 12.337292671203613,
-      "learning_rate": 8.405555555555556e-05,
-      "loss": 0.6355,
+      "epoch": 0.2885725278953444,
+      "grad_norm": 9.618443489074707,
+      "learning_rate": 8.388888888888888e-05,
+      "loss": 0.4521,
       "step": 4500
     },
     {
-      "epoch": 0.7561754327003865,
-      "eval_loss": 0.5601897239685059,
-      "eval_runtime": 184.3021,
-      "eval_samples_per_second": 38.1,
-      "eval_steps_per_second": 2.382,
-      "eval_wer": 0.4328784597191034,
+      "epoch": 0.2885725278953444,
+      "eval_loss": 0.6477864980697632,
+      "eval_runtime": 177.7072,
+      "eval_samples_per_second": 39.514,
+      "eval_steps_per_second": 2.47,
+      "eval_wer": 0.4861345486918508,
       "step": 4500
     },
     {
-      "epoch": 0.8401949252226516,
-      "grad_norm": 18.518247604370117,
-      "learning_rate": 5.6333333333333325e-05,
-      "loss": 0.5938,
+      "epoch": 0.32063614210593816,
+      "grad_norm": 8.743291854858398,
+      "learning_rate": 5.61111111111111e-05,
+      "loss": 0.4309,
       "step": 5000
     },
     {
-      "epoch": 0.8401949252226516,
-      "eval_loss": 0.5438016653060913,
-      "eval_runtime": 185.1047,
-      "eval_samples_per_second": 37.935,
-      "eval_steps_per_second": 2.372,
-      "eval_wer": 0.4225939754713722,
+      "epoch": 0.32063614210593816,
+      "eval_loss": 0.6221807599067688,
+      "eval_runtime": 177.9115,
+      "eval_samples_per_second": 39.469,
+      "eval_steps_per_second": 2.468,
+      "eval_wer": 0.4763416999960138,
       "step": 5000
     },
     {
-      "epoch": 0.9242144177449169,
-      "grad_norm": 5.622241020202637,
-      "learning_rate": 2.8555555555555556e-05,
-      "loss": 0.5842,
+      "epoch": 0.352699756316532,
+      "grad_norm": 8.870036125183105,
+      "learning_rate": 2.833333333333333e-05,
+      "loss": 0.4055,
       "step": 5500
     },
     {
-      "epoch": 0.9242144177449169,
-      "eval_loss": 0.5245193839073181,
-      "eval_runtime": 184.9427,
-      "eval_samples_per_second": 37.969,
-      "eval_steps_per_second": 2.374,
-      "eval_wer": 0.4090939289653065,
+      "epoch": 0.352699756316532,
+      "eval_loss": 0.5988152027130127,
+      "eval_runtime": 178.6515,
+      "eval_samples_per_second": 39.306,
+      "eval_steps_per_second": 2.457,
+      "eval_wer": 0.45257045668956536,
       "step": 5500
     },
     {
-      "epoch": 1.008233910267182,
-      "grad_norm": 10.519143104553223,
-      "learning_rate": 7.777777777777777e-07,
-      "loss": 0.557,
+      "epoch": 0.38476337052712584,
+      "grad_norm": 14.018965721130371,
+      "learning_rate": 6.11111111111111e-07,
+      "loss": 0.3896,
       "step": 6000
     },
     {
-      "epoch": 1.008233910267182,
-      "eval_loss": 0.5155353546142578,
-      "eval_runtime": 185.543,
-      "eval_samples_per_second": 37.846,
-      "eval_steps_per_second": 2.366,
-      "eval_wer": 0.39834438406037814,
+      "epoch": 0.38476337052712584,
+      "eval_loss": 0.5865370631217957,
+      "eval_runtime": 176.8507,
+      "eval_samples_per_second": 39.706,
+      "eval_steps_per_second": 2.482,
+      "eval_wer": 0.44523578575319894,
       "step": 6000
     },
     {
-      "epoch": 1.008233910267182,
+      "epoch": 0.38476337052712584,
       "step": 6000,
-      "total_flos": 6.665403914161851e+18,
-      "train_loss": 1.086754165649414,
-      "train_runtime": 4718.794,
-      "train_samples_per_second": 10.172,
-      "train_steps_per_second": 1.272
+      "total_flos": 6.835515215409967e+18,
+      "train_loss": 0.9274428431193034,
+      "train_runtime": 4544.4955,
+      "train_samples_per_second": 10.562,
+      "train_steps_per_second": 1.32
     }
   ],
   "logging_steps": 500,
   "max_steps": 6000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 2,
+  "num_train_epochs": 1,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -227,7 +227,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 6.665403914161851e+18,
+  "total_flos": 6.835515215409967e+18,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null