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: 0.3501

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 3.088803088803089,
+    "epoch": 1.5444015444015444,
     "eval_loss": Infinity,
-    "eval_runtime": 177.1303,
+    "eval_runtime": 179.4638,
     "eval_samples": 7022,
-    "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,
+    "eval_samples_per_second": 39.128,
+    "eval_steps_per_second": 4.892,
+    "eval_wer": 0.3500967369145469,
+    "total_flos": 8.411238439350073e+18,
+    "train_loss": 1.0689988899230958,
+    "train_runtime": 5961.6242,
     "train_samples": 41435,
-    "train_samples_per_second": 16.17,
-    "train_steps_per_second": 0.505
+    "train_samples_per_second": 10.735,
+    "train_steps_per_second": 0.671
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 3.088803088803089,
+    "epoch": 1.5444015444015444,
     "eval_loss": Infinity,
-    "eval_runtime": 177.1303,
+    "eval_runtime": 179.4638,
     "eval_samples": 7022,
-    "eval_samples_per_second": 39.643,
-    "eval_steps_per_second": 4.957,
-    "eval_wer": 0.33521613234129305
+    "eval_samples_per_second": 39.128,
+    "eval_steps_per_second": 4.892,
+    "eval_wer": 0.3500967369145469
 }

runs/Aug12_11-48-44_940cfb8ec62d/events.out.tfevents.1723465934.940cfb8ec62d.1296.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d439a837838a85e2783ed7f6d518c621e539df3044655c51c336568e11dfaf9
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 3.088803088803089,
-    "total_flos": 1.6887033849357677e+19,
-    "train_loss": 0.9425676002502441,
-    "train_runtime": 7915.8947,
+    "epoch": 1.5444015444015444,
+    "total_flos": 8.411238439350073e+18,
+    "train_loss": 1.0689988899230958,
+    "train_runtime": 5961.6242,
     "train_samples": 41435,
-    "train_samples_per_second": 16.17,
-    "train_steps_per_second": 0.505
+    "train_samples_per_second": 10.735,
+    "train_steps_per_second": 0.671
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.088803088803089,
+  "epoch": 1.5444015444015444,
   "eval_steps": 200,
   "global_step": 4000,
   "is_hyper_param_search": false,
@@ -9,255 +9,255 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.15444015444015444,
+      "epoch": 0.07722007722007722,
       "eval_loss": Infinity,
-      "eval_runtime": 182.0736,
-      "eval_samples_per_second": 38.567,
-      "eval_steps_per_second": 4.822,
+      "eval_runtime": 178.8025,
+      "eval_samples_per_second": 39.272,
+      "eval_steps_per_second": 4.91,
       "eval_wer": 1.0,
       "step": 200
     },
     {
-      "epoch": 0.3088803088803089,
+      "epoch": 0.15444015444015444,
       "eval_loss": Infinity,
-      "eval_runtime": 177.405,
-      "eval_samples_per_second": 39.582,
-      "eval_steps_per_second": 4.949,
-      "eval_wer": 0.8236661342890161,
+      "eval_runtime": 176.9522,
+      "eval_samples_per_second": 39.683,
+      "eval_steps_per_second": 4.962,
+      "eval_wer": 0.9261812940672354,
       "step": 400
     },
     {
-      "epoch": 0.3861003861003861,
-      "grad_norm": 3.3234875202178955,
+      "epoch": 0.19305019305019305,
+      "grad_norm": 3.265190362930298,
       "learning_rate": 0.00029699999999999996,
-      "loss": 3.6679,
+      "loss": 3.9094,
       "step": 500
     },
     {
-      "epoch": 0.46332046332046334,
+      "epoch": 0.23166023166023167,
       "eval_loss": Infinity,
-      "eval_runtime": 177.3768,
-      "eval_samples_per_second": 39.588,
-      "eval_steps_per_second": 4.95,
-      "eval_wer": 0.7189305701634789,
+      "eval_runtime": 177.1009,
+      "eval_samples_per_second": 39.65,
+      "eval_steps_per_second": 4.958,
+      "eval_wer": 0.7767000376559802,
       "step": 600
     },
     {
-      "epoch": 0.6177606177606177,
+      "epoch": 0.3088803088803089,
       "eval_loss": Infinity,
-      "eval_runtime": 177.3999,
-      "eval_samples_per_second": 39.583,
-      "eval_steps_per_second": 4.949,
-      "eval_wer": 0.5541142404529106,
+      "eval_runtime": 177.4675,
+      "eval_samples_per_second": 39.568,
+      "eval_steps_per_second": 4.947,
+      "eval_wer": 0.7365379870930882,
       "step": 800
     },
     {
-      "epoch": 0.7722007722007722,
-      "grad_norm": 3.4127838611602783,
-      "learning_rate": 0.0002577428571428571,
-      "loss": 0.8341,
+      "epoch": 0.3861003861003861,
+      "grad_norm": 3.0774288177490234,
+      "learning_rate": 0.00025765714285714284,
+      "loss": 0.9885,
       "step": 1000
     },
     {
-      "epoch": 0.7722007722007722,
+      "epoch": 0.3861003861003861,
       "eval_loss": Infinity,
-      "eval_runtime": 178.223,
-      "eval_samples_per_second": 39.4,
-      "eval_steps_per_second": 4.926,
-      "eval_wer": 0.512381026579928,
+      "eval_runtime": 177.9576,
+      "eval_samples_per_second": 39.459,
+      "eval_steps_per_second": 4.934,
+      "eval_wer": 0.6104813473050005,
       "step": 1000
     },
     {
-      "epoch": 0.9266409266409267,
+      "epoch": 0.46332046332046334,
       "eval_loss": Infinity,
-      "eval_runtime": 177.4709,
-      "eval_samples_per_second": 39.567,
-      "eval_steps_per_second": 4.947,
-      "eval_wer": 0.489462817965798,
+      "eval_runtime": 176.7839,
+      "eval_samples_per_second": 39.721,
+      "eval_steps_per_second": 4.967,
+      "eval_wer": 0.5430511731785542,
       "step": 1200
     },
     {
-      "epoch": 1.0810810810810811,
+      "epoch": 0.5405405405405406,
       "eval_loss": Infinity,
-      "eval_runtime": 178.0824,
-      "eval_samples_per_second": 39.431,
-      "eval_steps_per_second": 4.93,
-      "eval_wer": 0.45271577525872253,
+      "eval_runtime": 177.8714,
+      "eval_samples_per_second": 39.478,
+      "eval_steps_per_second": 4.936,
+      "eval_wer": 0.5144845675405451,
       "step": 1400
     },
     {
-      "epoch": 1.1583011583011582,
-      "grad_norm": 0.5421157479286194,
+      "epoch": 0.5791505791505791,
+      "grad_norm": 3.033381700515747,
       "learning_rate": 0.00021488571428571426,
-      "loss": 0.6524,
+      "loss": 0.7974,
       "step": 1500
     },
     {
-      "epoch": 1.2355212355212355,
+      "epoch": 0.6177606177606177,
       "eval_loss": Infinity,
-      "eval_runtime": 176.6316,
-      "eval_samples_per_second": 39.755,
-      "eval_steps_per_second": 4.971,
-      "eval_wer": 0.4340436030280602,
+      "eval_runtime": 178.2105,
+      "eval_samples_per_second": 39.403,
+      "eval_steps_per_second": 4.927,
+      "eval_wer": 0.5230675340526924,
       "step": 1600
     },
     {
-      "epoch": 1.3899613899613898,
+      "epoch": 0.694980694980695,
       "eval_loss": Infinity,
-      "eval_runtime": 177.018,
-      "eval_samples_per_second": 39.668,
-      "eval_steps_per_second": 4.96,
-      "eval_wer": 0.40710009998311975,
+      "eval_runtime": 178.6698,
+      "eval_samples_per_second": 39.302,
+      "eval_steps_per_second": 4.914,
+      "eval_wer": 0.46214275511926556,
       "step": 1800
     },
     {
-      "epoch": 1.5444015444015444,
-      "grad_norm": 0.6360776424407959,
+      "epoch": 0.7722007722007722,
+      "grad_norm": 3.2283124923706055,
       "learning_rate": 0.0001721142857142857,
-      "loss": 0.5566,
+      "loss": 0.7068,
       "step": 2000
     },
     {
-      "epoch": 1.5444015444015444,
+      "epoch": 0.7722007722007722,
       "eval_loss": Infinity,
-      "eval_runtime": 176.5522,
-      "eval_samples_per_second": 39.773,
-      "eval_steps_per_second": 4.973,
-      "eval_wer": 0.3941931881630374,
+      "eval_runtime": 179.2886,
+      "eval_samples_per_second": 39.166,
+      "eval_steps_per_second": 4.897,
+      "eval_wer": 0.4532481529092491,
       "step": 2000
     },
     {
-      "epoch": 1.698841698841699,
+      "epoch": 0.8494208494208494,
       "eval_loss": Infinity,
-      "eval_runtime": 177.2873,
-      "eval_samples_per_second": 39.608,
-      "eval_steps_per_second": 4.952,
-      "eval_wer": 0.39070027138275354,
+      "eval_runtime": 179.5146,
+      "eval_samples_per_second": 39.117,
+      "eval_steps_per_second": 4.891,
+      "eval_wer": 0.4404061651928895,
       "step": 2200
     },
     {
-      "epoch": 1.8532818532818531,
+      "epoch": 0.9266409266409267,
       "eval_loss": Infinity,
-      "eval_runtime": 177.1067,
-      "eval_samples_per_second": 39.648,
-      "eval_steps_per_second": 4.957,
-      "eval_wer": 0.37781932920416034,
+      "eval_runtime": 178.4125,
+      "eval_samples_per_second": 39.358,
+      "eval_steps_per_second": 4.921,
+      "eval_wer": 0.4195006037941646,
       "step": 2400
     },
     {
-      "epoch": 1.9305019305019306,
-      "grad_norm": 3.162639617919922,
-      "learning_rate": 0.0001295142857142857,
-      "loss": 0.5324,
+      "epoch": 0.9652509652509652,
+      "grad_norm": 1.9969470500946045,
+      "learning_rate": 0.00012925714285714286,
+      "loss": 0.6447,
       "step": 2500
     },
     {
-      "epoch": 2.0077220077220077,
+      "epoch": 1.0038610038610039,
       "eval_loss": Infinity,
-      "eval_runtime": 177.9344,
-      "eval_samples_per_second": 39.464,
-      "eval_steps_per_second": 4.934,
-      "eval_wer": 0.3624453014426136,
+      "eval_runtime": 179.4629,
+      "eval_samples_per_second": 39.128,
+      "eval_steps_per_second": 4.892,
+      "eval_wer": 0.4115928479607339,
       "step": 2600
     },
     {
-      "epoch": 2.1621621621621623,
+      "epoch": 1.0810810810810811,
       "eval_loss": Infinity,
-      "eval_runtime": 178.1586,
-      "eval_samples_per_second": 39.414,
-      "eval_steps_per_second": 4.928,
-      "eval_wer": 0.35340786620440706,
+      "eval_runtime": 182.2946,
+      "eval_samples_per_second": 38.52,
+      "eval_steps_per_second": 4.816,
+      "eval_wer": 0.38897329022372845,
       "step": 2800
     },
     {
-      "epoch": 2.3166023166023164,
-      "grad_norm": 0.8761777877807617,
-      "learning_rate": 8.674285714285714e-05,
-      "loss": 0.4444,
+      "epoch": 1.1583011583011582,
+      "grad_norm": 0.5247506499290466,
+      "learning_rate": 8.648571428571429e-05,
+      "loss": 0.5318,
       "step": 3000
     },
     {
-      "epoch": 2.3166023166023164,
+      "epoch": 1.1583011583011582,
       "eval_loss": Infinity,
-      "eval_runtime": 177.4921,
-      "eval_samples_per_second": 39.562,
-      "eval_steps_per_second": 4.947,
-      "eval_wer": 0.34959032890550945,
+      "eval_runtime": 180.6608,
+      "eval_samples_per_second": 38.868,
+      "eval_steps_per_second": 4.86,
+      "eval_wer": 0.38040330853232573,
       "step": 3000
     },
     {
-      "epoch": 2.471042471042471,
+      "epoch": 1.2355212355212355,
       "eval_loss": Infinity,
-      "eval_runtime": 177.8631,
-      "eval_samples_per_second": 39.48,
-      "eval_steps_per_second": 4.936,
-      "eval_wer": 0.3415916793268669,
+      "eval_runtime": 178.267,
+      "eval_samples_per_second": 39.39,
+      "eval_steps_per_second": 4.925,
+      "eval_wer": 0.37341747497175803,
       "step": 3200
     },
     {
-      "epoch": 2.6254826254826256,
+      "epoch": 1.3127413127413128,
       "eval_loss": Infinity,
-      "eval_runtime": 177.4214,
-      "eval_samples_per_second": 39.578,
-      "eval_steps_per_second": 4.949,
-      "eval_wer": 0.3406697570540039,
+      "eval_runtime": 179.0534,
+      "eval_samples_per_second": 39.217,
+      "eval_steps_per_second": 4.904,
+      "eval_wer": 0.36753535117447705,
       "step": 3400
     },
     {
-      "epoch": 2.7027027027027026,
-      "grad_norm": 3.2265968322753906,
-      "learning_rate": 4.405714285714285e-05,
-      "loss": 0.4254,
+      "epoch": 1.3513513513513513,
+      "grad_norm": 0.6272704005241394,
+      "learning_rate": 4.3799999999999994e-05,
+      "loss": 0.494,
       "step": 3500
     },
     {
-      "epoch": 2.7799227799227797,
+      "epoch": 1.3899613899613898,
       "eval_loss": Infinity,
-      "eval_runtime": 178.3037,
-      "eval_samples_per_second": 39.382,
-      "eval_steps_per_second": 4.924,
-      "eval_wer": 0.3346448002285328,
+      "eval_runtime": 180.0451,
+      "eval_samples_per_second": 39.001,
+      "eval_steps_per_second": 4.877,
+      "eval_wer": 0.3590173087660525,
       "step": 3600
     },
     {
-      "epoch": 2.9343629343629343,
+      "epoch": 1.4671814671814671,
       "eval_loss": Infinity,
-      "eval_runtime": 177.742,
-      "eval_samples_per_second": 39.507,
-      "eval_steps_per_second": 4.94,
-      "eval_wer": 0.33778712684871387,
+      "eval_runtime": 179.7417,
+      "eval_samples_per_second": 39.067,
+      "eval_steps_per_second": 4.885,
+      "eval_wer": 0.35345980548738526,
       "step": 3800
     },
     {
-      "epoch": 3.088803088803089,
-      "grad_norm": 3.3560454845428467,
-      "learning_rate": 1.2857142857142856e-06,
-      "loss": 0.4273,
+      "epoch": 1.5444015444015444,
+      "grad_norm": 0.7731852531433105,
+      "learning_rate": 1.0285714285714284e-06,
+      "loss": 0.4794,
       "step": 4000
     },
     {
-      "epoch": 3.088803088803089,
+      "epoch": 1.5444015444015444,
       "eval_loss": Infinity,
-      "eval_runtime": 177.2694,
-      "eval_samples_per_second": 39.612,
-      "eval_steps_per_second": 4.953,
-      "eval_wer": 0.33521613234129305,
+      "eval_runtime": 179.8748,
+      "eval_samples_per_second": 39.038,
+      "eval_steps_per_second": 4.881,
+      "eval_wer": 0.3500967369145469,
       "step": 4000
     },
     {
-      "epoch": 3.088803088803089,
+      "epoch": 1.5444015444015444,
       "step": 4000,
-      "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
+      "total_flos": 8.411238439350073e+18,
+      "train_loss": 1.0689988899230958,
+      "train_runtime": 5961.6242,
+      "train_samples_per_second": 10.735,
+      "train_steps_per_second": 0.671
     }
   ],
   "logging_steps": 500,
   "max_steps": 4000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 2,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -271,8 +271,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.6887033849357677e+19,
-  "train_batch_size": 32,
+  "total_flos": 8.411238439350073e+18,
+  "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null
 }