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: inf
 - Wer: 0.3352

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 6.172839506172839,
+    "epoch": 3.088803088803089,
     "eval_loss": Infinity,
-    "eval_runtime": 176.2004,
+    "eval_runtime": 177.1303,
     "eval_samples": 7022,
-    "eval_samples_per_second": 39.852,
-    "eval_steps_per_second": 4.983,
-    "eval_wer": 1.0,
-    "total_flos": 3.3889468875697558e+19,
-    "train_loss": 2.1502342529296876,
-    "train_runtime": 11441.7295,
+    "eval_samples_per_second": 39.643,
+    "eval_steps_per_second": 4.957,
+    "eval_wer": 0.33521613234129305,
+    "total_flos": 1.6887033849357677e+19,
+    "train_loss": 0.9425676002502441,
+    "train_runtime": 7915.8947,
     "train_samples": 41435,
-    "train_samples_per_second": 22.374,
-    "train_steps_per_second": 0.35
+    "train_samples_per_second": 16.17,
+    "train_steps_per_second": 0.505
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 6.172839506172839,
+    "epoch": 3.088803088803089,
     "eval_loss": Infinity,
-    "eval_runtime": 176.2004,
+    "eval_runtime": 177.1303,
     "eval_samples": 7022,
-    "eval_samples_per_second": 39.852,
-    "eval_steps_per_second": 4.983,
-    "eval_wer": 1.0
+    "eval_samples_per_second": 39.643,
+    "eval_steps_per_second": 4.957,
+    "eval_wer": 0.33521613234129305
 }

runs/Aug11_16-31-10_940cfb8ec62d/events.out.tfevents.1723398440.940cfb8ec62d.1010.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9a1643da227de462ff6a98e07e636712816083bccf5a16192e24a9afbcd085f
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 6.172839506172839,
-    "total_flos": 3.3889468875697558e+19,
-    "train_loss": 2.1502342529296876,
-    "train_runtime": 11441.7295,
+    "epoch": 3.088803088803089,
+    "total_flos": 1.6887033849357677e+19,
+    "train_loss": 0.9425676002502441,
+    "train_runtime": 7915.8947,
     "train_samples": 41435,
-    "train_samples_per_second": 22.374,
-    "train_steps_per_second": 0.35
+    "train_samples_per_second": 16.17,
+    "train_steps_per_second": 0.505
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 6.172839506172839,
+  "epoch": 3.088803088803089,
   "eval_steps": 200,
   "global_step": 4000,
   "is_hyper_param_search": false,
@@ -9,255 +9,255 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.30864197530864196,
+      "epoch": 0.15444015444015444,
       "eval_loss": Infinity,
-      "eval_runtime": 176.8693,
-      "eval_samples_per_second": 39.702,
-      "eval_steps_per_second": 4.964,
+      "eval_runtime": 182.0736,
+      "eval_samples_per_second": 38.567,
+      "eval_steps_per_second": 4.822,
       "eval_wer": 1.0,
       "step": 200
     },
     {
-      "epoch": 0.6172839506172839,
+      "epoch": 0.3088803088803089,
       "eval_loss": Infinity,
-      "eval_runtime": 175.9503,
-      "eval_samples_per_second": 39.909,
-      "eval_steps_per_second": 4.99,
-      "eval_wer": 0.8138755794476257,
+      "eval_runtime": 177.405,
+      "eval_samples_per_second": 39.582,
+      "eval_steps_per_second": 4.949,
+      "eval_wer": 0.8236661342890161,
       "step": 400
     },
     {
-      "epoch": 0.7716049382716049,
-      "grad_norm": 2.980449914932251,
-      "learning_rate": 0.0002958,
-      "loss": 3.6854,
+      "epoch": 0.3861003861003861,
+      "grad_norm": 3.3234875202178955,
+      "learning_rate": 0.00029699999999999996,
+      "loss": 3.6679,
       "step": 500
     },
     {
-      "epoch": 0.9259259259259259,
+      "epoch": 0.46332046332046334,
       "eval_loss": Infinity,
-      "eval_runtime": 174.7168,
-      "eval_samples_per_second": 40.191,
-      "eval_steps_per_second": 5.025,
-      "eval_wer": 0.6042746029891057,
+      "eval_runtime": 177.3768,
+      "eval_samples_per_second": 39.588,
+      "eval_steps_per_second": 4.95,
+      "eval_wer": 0.7189305701634789,
       "step": 600
     },
     {
-      "epoch": 1.2345679012345678,
+      "epoch": 0.6177606177606177,
       "eval_loss": Infinity,
-      "eval_runtime": 176.3896,
-      "eval_samples_per_second": 39.81,
-      "eval_steps_per_second": 4.978,
-      "eval_wer": 0.503447469907678,
+      "eval_runtime": 177.3999,
+      "eval_samples_per_second": 39.583,
+      "eval_steps_per_second": 4.949,
+      "eval_wer": 0.5541142404529106,
       "step": 800
     },
     {
-      "epoch": 1.5432098765432098,
-      "grad_norm": 0.7878520488739014,
-      "learning_rate": 0.0002578285714285714,
-      "loss": 0.7236,
+      "epoch": 0.7722007722007722,
+      "grad_norm": 3.4127838611602783,
+      "learning_rate": 0.0002577428571428571,
+      "loss": 0.8341,
       "step": 1000
     },
     {
-      "epoch": 1.5432098765432098,
+      "epoch": 0.7722007722007722,
       "eval_loss": Infinity,
-      "eval_runtime": 175.5785,
-      "eval_samples_per_second": 39.994,
-      "eval_steps_per_second": 5.001,
-      "eval_wer": 0.44518457922688376,
+      "eval_runtime": 178.223,
+      "eval_samples_per_second": 39.4,
+      "eval_steps_per_second": 4.926,
+      "eval_wer": 0.512381026579928,
       "step": 1000
     },
     {
-      "epoch": 1.8518518518518519,
+      "epoch": 0.9266409266409267,
       "eval_loss": Infinity,
-      "eval_runtime": 175.3263,
-      "eval_samples_per_second": 40.051,
-      "eval_steps_per_second": 5.008,
-      "eval_wer": 0.4584940204900471,
+      "eval_runtime": 177.4709,
+      "eval_samples_per_second": 39.567,
+      "eval_steps_per_second": 4.947,
+      "eval_wer": 0.489462817965798,
       "step": 1200
     },
     {
-      "epoch": 2.1604938271604937,
+      "epoch": 1.0810810810810811,
       "eval_loss": Infinity,
-      "eval_runtime": 174.0761,
-      "eval_samples_per_second": 40.339,
-      "eval_steps_per_second": 5.044,
-      "eval_wer": 0.4735564125537247,
+      "eval_runtime": 178.0824,
+      "eval_samples_per_second": 39.431,
+      "eval_steps_per_second": 4.93,
+      "eval_wer": 0.45271577525872253,
       "step": 1400
     },
     {
-      "epoch": 2.314814814814815,
-      "grad_norm": 1.6092010736465454,
-      "learning_rate": 0.0002153142857142857,
-      "loss": 0.6244,
+      "epoch": 1.1583011583011582,
+      "grad_norm": 0.5421157479286194,
+      "learning_rate": 0.00021488571428571426,
+      "loss": 0.6524,
       "step": 1500
     },
     {
-      "epoch": 2.4691358024691357,
+      "epoch": 1.2355212355212355,
       "eval_loss": Infinity,
-      "eval_runtime": 174.8676,
-      "eval_samples_per_second": 40.156,
-      "eval_steps_per_second": 5.021,
-      "eval_wer": 0.4445742926518899,
+      "eval_runtime": 176.6316,
+      "eval_samples_per_second": 39.755,
+      "eval_steps_per_second": 4.971,
+      "eval_wer": 0.4340436030280602,
       "step": 1600
     },
     {
-      "epoch": 2.7777777777777777,
+      "epoch": 1.3899613899613898,
       "eval_loss": Infinity,
-      "eval_runtime": 174.4183,
-      "eval_samples_per_second": 40.26,
-      "eval_steps_per_second": 5.034,
-      "eval_wer": 0.4992274031656993,
+      "eval_runtime": 177.018,
+      "eval_samples_per_second": 39.668,
+      "eval_steps_per_second": 4.96,
+      "eval_wer": 0.40710009998311975,
       "step": 1800
     },
     {
-      "epoch": 3.0864197530864197,
-      "grad_norm": 13.045821189880371,
-      "learning_rate": 0.0001727142857142857,
-      "loss": 0.8045,
+      "epoch": 1.5444015444015444,
+      "grad_norm": 0.6360776424407959,
+      "learning_rate": 0.0001721142857142857,
+      "loss": 0.5566,
       "step": 2000
     },
     {
-      "epoch": 3.0864197530864197,
+      "epoch": 1.5444015444015444,
       "eval_loss": Infinity,
-      "eval_runtime": 176.0054,
-      "eval_samples_per_second": 39.897,
-      "eval_steps_per_second": 4.988,
-      "eval_wer": 0.699323490839209,
+      "eval_runtime": 176.5522,
+      "eval_samples_per_second": 39.773,
+      "eval_steps_per_second": 4.973,
+      "eval_wer": 0.3941931881630374,
       "step": 2000
     },
     {
-      "epoch": 3.3950617283950617,
+      "epoch": 1.698841698841699,
       "eval_loss": Infinity,
-      "eval_runtime": 175.9958,
-      "eval_samples_per_second": 39.899,
-      "eval_steps_per_second": 4.989,
-      "eval_wer": 0.9154558321322375,
+      "eval_runtime": 177.2873,
+      "eval_samples_per_second": 39.608,
+      "eval_steps_per_second": 4.952,
+      "eval_wer": 0.39070027138275354,
       "step": 2200
     },
     {
-      "epoch": 3.7037037037037037,
+      "epoch": 1.8532818532818531,
       "eval_loss": Infinity,
-      "eval_runtime": 175.4175,
-      "eval_samples_per_second": 40.03,
-      "eval_steps_per_second": 5.005,
-      "eval_wer": 1.0,
+      "eval_runtime": 177.1067,
+      "eval_samples_per_second": 39.648,
+      "eval_steps_per_second": 4.957,
+      "eval_wer": 0.37781932920416034,
       "step": 2400
     },
     {
-      "epoch": 3.8580246913580245,
-      "grad_norm": 1.601144552230835,
-      "learning_rate": 0.0001302,
-      "loss": 2.3067,
+      "epoch": 1.9305019305019306,
+      "grad_norm": 3.162639617919922,
+      "learning_rate": 0.0001295142857142857,
+      "loss": 0.5324,
       "step": 2500
     },
     {
-      "epoch": 4.012345679012346,
+      "epoch": 2.0077220077220077,
       "eval_loss": Infinity,
-      "eval_runtime": 175.9931,
-      "eval_samples_per_second": 39.899,
-      "eval_steps_per_second": 4.989,
-      "eval_wer": 0.998558684897355,
+      "eval_runtime": 177.9344,
+      "eval_samples_per_second": 39.464,
+      "eval_steps_per_second": 4.934,
+      "eval_wer": 0.3624453014426136,
       "step": 2600
     },
     {
-      "epoch": 4.320987654320987,
+      "epoch": 2.1621621621621623,
       "eval_loss": Infinity,
-      "eval_runtime": 175.0274,
-      "eval_samples_per_second": 40.119,
-      "eval_steps_per_second": 5.016,
-      "eval_wer": 1.0,
+      "eval_runtime": 178.1586,
+      "eval_samples_per_second": 39.414,
+      "eval_steps_per_second": 4.928,
+      "eval_wer": 0.35340786620440706,
       "step": 2800
     },
     {
-      "epoch": 4.62962962962963,
-      "grad_norm": 0.7471032738685608,
-      "learning_rate": 8.759999999999999e-05,
-      "loss": 3.008,
+      "epoch": 2.3166023166023164,
+      "grad_norm": 0.8761777877807617,
+      "learning_rate": 8.674285714285714e-05,
+      "loss": 0.4444,
       "step": 3000
     },
     {
-      "epoch": 4.62962962962963,
+      "epoch": 2.3166023166023164,
       "eval_loss": Infinity,
-      "eval_runtime": 175.5031,
-      "eval_samples_per_second": 40.011,
-      "eval_steps_per_second": 5.003,
-      "eval_wer": 1.0,
+      "eval_runtime": 177.4921,
+      "eval_samples_per_second": 39.562,
+      "eval_steps_per_second": 4.947,
+      "eval_wer": 0.34959032890550945,
       "step": 3000
     },
     {
-      "epoch": 4.938271604938271,
+      "epoch": 2.471042471042471,
       "eval_loss": Infinity,
-      "eval_runtime": 175.2017,
-      "eval_samples_per_second": 40.08,
-      "eval_steps_per_second": 5.011,
-      "eval_wer": 1.0,
+      "eval_runtime": 177.8631,
+      "eval_samples_per_second": 39.48,
+      "eval_steps_per_second": 4.936,
+      "eval_wer": 0.3415916793268669,
       "step": 3200
     },
     {
-      "epoch": 5.246913580246914,
+      "epoch": 2.6254826254826256,
       "eval_loss": Infinity,
-      "eval_runtime": 176.0971,
-      "eval_samples_per_second": 39.876,
-      "eval_steps_per_second": 4.986,
-      "eval_wer": 1.0,
+      "eval_runtime": 177.4214,
+      "eval_samples_per_second": 39.578,
+      "eval_steps_per_second": 4.949,
+      "eval_wer": 0.3406697570540039,
       "step": 3400
     },
     {
-      "epoch": 5.401234567901234,
-      "grad_norm": 1.3811966180801392,
-      "learning_rate": 4.4999999999999996e-05,
-      "loss": 3.032,
+      "epoch": 2.7027027027027026,
+      "grad_norm": 3.2265968322753906,
+      "learning_rate": 4.405714285714285e-05,
+      "loss": 0.4254,
       "step": 3500
     },
     {
-      "epoch": 5.555555555555555,
+      "epoch": 2.7799227799227797,
       "eval_loss": Infinity,
-      "eval_runtime": 176.4078,
-      "eval_samples_per_second": 39.805,
-      "eval_steps_per_second": 4.977,
-      "eval_wer": 1.0,
+      "eval_runtime": 178.3037,
+      "eval_samples_per_second": 39.382,
+      "eval_steps_per_second": 4.924,
+      "eval_wer": 0.3346448002285328,
       "step": 3600
     },
     {
-      "epoch": 5.864197530864198,
+      "epoch": 2.9343629343629343,
       "eval_loss": Infinity,
-      "eval_runtime": 175.4996,
-      "eval_samples_per_second": 40.011,
-      "eval_steps_per_second": 5.003,
-      "eval_wer": 1.0,
+      "eval_runtime": 177.742,
+      "eval_samples_per_second": 39.507,
+      "eval_steps_per_second": 4.94,
+      "eval_wer": 0.33778712684871387,
       "step": 3800
     },
     {
-      "epoch": 6.172839506172839,
-      "grad_norm": 0.011152578517794609,
-      "learning_rate": 2.314285714285714e-06,
-      "loss": 3.0173,
+      "epoch": 3.088803088803089,
+      "grad_norm": 3.3560454845428467,
+      "learning_rate": 1.2857142857142856e-06,
+      "loss": 0.4273,
       "step": 4000
     },
     {
-      "epoch": 6.172839506172839,
+      "epoch": 3.088803088803089,
       "eval_loss": Infinity,
-      "eval_runtime": 175.8629,
-      "eval_samples_per_second": 39.929,
-      "eval_steps_per_second": 4.993,
-      "eval_wer": 1.0,
+      "eval_runtime": 177.2694,
+      "eval_samples_per_second": 39.612,
+      "eval_steps_per_second": 4.953,
+      "eval_wer": 0.33521613234129305,
       "step": 4000
     },
     {
-      "epoch": 6.172839506172839,
+      "epoch": 3.088803088803089,
       "step": 4000,
-      "total_flos": 3.3889468875697558e+19,
-      "train_loss": 2.1502342529296876,
-      "train_runtime": 11441.7295,
-      "train_samples_per_second": 22.374,
-      "train_steps_per_second": 0.35
+      "total_flos": 1.6887033849357677e+19,
+      "train_loss": 0.9425676002502441,
+      "train_runtime": 7915.8947,
+      "train_samples_per_second": 16.17,
+      "train_steps_per_second": 0.505
     }
   ],
   "logging_steps": 500,
   "max_steps": 4000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 7,
+  "num_train_epochs": 4,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -271,8 +271,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 3.3889468875697558e+19,
-  "train_batch_size": 64,
+  "total_flos": 1.6887033849357677e+19,
+  "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null
 }