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
 - generated_from_trainer
 metrics:
 - wer
@@ -15,10 +17,10 @@ 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-CLEAN - DEFAULT dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.8068
-- Wer: 0.4827
+- Loss: 0.8192
+- Wer: 0.4756
 
 ## Model description
 

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 2.277904328018223,
-    "eval_loss": 0.5399738550186157,
-    "eval_runtime": 177.5085,
+    "epoch": 2.847380410022779,
+    "eval_loss": 0.8192350268363953,
+    "eval_runtime": 179.8741,
     "eval_samples": 7022,
-    "eval_samples_per_second": 39.559,
-    "eval_steps_per_second": 4.946,
-    "eval_wer": 0.40660917631113885,
-    "total_flos": 7.50335828249239e+18,
-    "train_loss": 1.4263343353271485,
-    "train_runtime": 3812.7792,
+    "eval_samples_per_second": 39.038,
+    "eval_steps_per_second": 4.881,
+    "eval_wer": 0.47557102804980134,
+    "total_flos": 9.363138171711945e+18,
+    "train_loss": 1.33308916015625,
+    "train_runtime": 4617.7601,
     "train_samples": 28086,
-    "train_samples_per_second": 16.786,
-    "train_steps_per_second": 0.525
+    "train_samples_per_second": 17.324,
+    "train_steps_per_second": 0.541
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 2.277904328018223,
-    "eval_loss": 0.5399738550186157,
-    "eval_runtime": 177.5085,
+    "epoch": 2.847380410022779,
+    "eval_loss": 0.8192350268363953,
+    "eval_runtime": 179.8741,
     "eval_samples": 7022,
-    "eval_samples_per_second": 39.559,
-    "eval_steps_per_second": 4.946,
-    "eval_wer": 0.40660917631113885
+    "eval_samples_per_second": 39.038,
+    "eval_steps_per_second": 4.881,
+    "eval_wer": 0.47557102804980134
 }

runs/Aug15_13-23-48_a142c9bbb9ba/events.out.tfevents.1723729494.a142c9bbb9ba.1181.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aec307f5be8dffd1b49677041356609849b6531fe0b3647e8cb07bb4dfa2ec07
+size 406

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 2.277904328018223,
-    "total_flos": 7.50335828249239e+18,
-    "train_loss": 1.4263343353271485,
-    "train_runtime": 3812.7792,
+    "epoch": 2.847380410022779,
+    "total_flos": 9.363138171711945e+18,
+    "train_loss": 1.33308916015625,
+    "train_runtime": 4617.7601,
     "train_samples": 28086,
-    "train_samples_per_second": 16.786,
-    "train_steps_per_second": 0.525
+    "train_samples_per_second": 17.324,
+    "train_steps_per_second": 0.541
 }

trainer_state.json

@@ -1,143 +1,168 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 2.277904328018223,
+  "epoch": 2.847380410022779,
   "eval_steps": 200,
-  "global_step": 2000,
+  "global_step": 2500,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.22779043280182232,
-      "eval_loss": 2.9681296348571777,
-      "eval_runtime": 179.5569,
-      "eval_samples_per_second": 39.107,
-      "eval_steps_per_second": 4.89,
+      "eval_loss": 2.9656879901885986,
+      "eval_runtime": 180.6727,
+      "eval_samples_per_second": 38.866,
+      "eval_steps_per_second": 4.86,
       "eval_wer": 1.0,
       "step": 200
     },
     {
       "epoch": 0.45558086560364464,
-      "eval_loss": 1.3578158617019653,
-      "eval_runtime": 178.971,
-      "eval_samples_per_second": 39.235,
-      "eval_steps_per_second": 4.906,
-      "eval_wer": 0.8299738237287235,
+      "eval_loss": 1.256484031677246,
+      "eval_runtime": 177.1703,
+      "eval_samples_per_second": 39.634,
+      "eval_steps_per_second": 4.956,
+      "eval_wer": 0.8238084481590242,
       "step": 400
     },
     {
       "epoch": 0.5694760820045558,
-      "grad_norm": 2.982609748840332,
+      "grad_norm": 2.394413948059082,
       "learning_rate": 0.00029699999999999996,
-      "loss": 3.3875,
+      "loss": 3.3708,
       "step": 500
     },
     {
       "epoch": 0.683371298405467,
-      "eval_loss": 0.9213220477104187,
-      "eval_runtime": 178.9628,
-      "eval_samples_per_second": 39.237,
-      "eval_steps_per_second": 4.906,
-      "eval_wer": 0.6513373815756255,
+      "eval_loss": 0.8798568248748779,
+      "eval_runtime": 177.6107,
+      "eval_samples_per_second": 39.536,
+      "eval_steps_per_second": 4.943,
+      "eval_wer": 0.6336517891547855,
       "step": 600
     },
     {
       "epoch": 0.9111617312072893,
-      "eval_loss": 0.7288344502449036,
-      "eval_runtime": 180.4166,
-      "eval_samples_per_second": 38.921,
-      "eval_steps_per_second": 4.867,
-      "eval_wer": 0.5341155210672478,
+      "eval_loss": 0.7572081685066223,
+      "eval_runtime": 179.3544,
+      "eval_samples_per_second": 39.152,
+      "eval_steps_per_second": 4.895,
+      "eval_wer": 0.5467651709430101,
       "step": 800
     },
     {
       "epoch": 1.1389521640091116,
-      "grad_norm": 1.1159461736679077,
-      "learning_rate": 0.0002022,
-      "loss": 0.8947,
+      "grad_norm": 0.749163806438446,
+      "learning_rate": 0.00022664999999999995,
+      "loss": 0.9128,
       "step": 1000
     },
     {
       "epoch": 1.1389521640091116,
-      "eval_loss": 0.6245301961898804,
-      "eval_runtime": 180.7702,
-      "eval_samples_per_second": 38.845,
-      "eval_steps_per_second": 4.857,
-      "eval_wer": 0.46335986393653916,
+      "eval_loss": 0.6133933663368225,
+      "eval_runtime": 178.9414,
+      "eval_samples_per_second": 39.242,
+      "eval_steps_per_second": 4.907,
+      "eval_wer": 0.4571280511300974,
       "step": 1000
     },
     {
       "epoch": 1.366742596810934,
-      "eval_loss": 0.5951615571975708,
-      "eval_runtime": 177.8647,
-      "eval_samples_per_second": 39.479,
-      "eval_steps_per_second": 4.936,
-      "eval_wer": 0.445833720883881,
+      "eval_loss": 0.5878995060920715,
+      "eval_runtime": 179.0164,
+      "eval_samples_per_second": 39.225,
+      "eval_steps_per_second": 4.905,
+      "eval_wer": 0.44289719501986474,
       "step": 1200
     },
     {
       "epoch": 1.5945330296127562,
-      "eval_loss": 0.578143835067749,
-      "eval_runtime": 178.6987,
-      "eval_samples_per_second": 39.295,
-      "eval_steps_per_second": 4.913,
-      "eval_wer": 0.4384857625001661,
+      "eval_loss": 0.6282891631126404,
+      "eval_runtime": 178.8026,
+      "eval_samples_per_second": 39.272,
+      "eval_steps_per_second": 4.91,
+      "eval_wer": 0.46545928061760056,
       "step": 1400
     },
     {
       "epoch": 1.7084282460136673,
-      "grad_norm": 1.4777103662490845,
-      "learning_rate": 0.00010279999999999999,
-      "loss": 0.7523,
+      "grad_norm": 2.0273807048797607,
+      "learning_rate": 0.00015209999999999998,
+      "loss": 0.7779,
       "step": 1500
     },
     {
       "epoch": 1.8223234624145785,
-      "eval_loss": 0.562013566493988,
-      "eval_runtime": 178.104,
-      "eval_samples_per_second": 39.426,
-      "eval_steps_per_second": 4.93,
-      "eval_wer": 0.42830757783122286,
+      "eval_loss": 0.6165733337402344,
+      "eval_runtime": 178.9936,
+      "eval_samples_per_second": 39.23,
+      "eval_steps_per_second": 4.905,
+      "eval_wer": 0.4566364155782033,
       "step": 1600
     },
     {
       "epoch": 2.050113895216401,
-      "eval_loss": 0.5437431931495667,
-      "eval_runtime": 177.6652,
-      "eval_samples_per_second": 39.524,
-      "eval_steps_per_second": 4.942,
-      "eval_wer": 0.4140368593789447,
+      "eval_loss": 0.6130892634391785,
+      "eval_runtime": 180.5144,
+      "eval_samples_per_second": 38.9,
+      "eval_steps_per_second": 4.864,
+      "eval_wer": 0.46482148314487304,
       "step": 1800
     },
     {
       "epoch": 2.277904328018223,
-      "grad_norm": 3.888875961303711,
-      "learning_rate": 2.9999999999999997e-06,
-      "loss": 0.6707,
+      "grad_norm": 2.559767246246338,
+      "learning_rate": 7.725e-05,
+      "loss": 0.7465,
       "step": 2000
     },
     {
       "epoch": 2.277904328018223,
-      "eval_loss": 0.5399738550186157,
-      "eval_runtime": 178.027,
-      "eval_samples_per_second": 39.443,
-      "eval_steps_per_second": 4.932,
-      "eval_wer": 0.40660917631113885,
+      "eval_loss": 0.6311540603637695,
+      "eval_runtime": 180.4386,
+      "eval_samples_per_second": 38.916,
+      "eval_steps_per_second": 4.866,
+      "eval_wer": 0.4721694415285879,
       "step": 2000
     },
     {
-      "epoch": 2.277904328018223,
-      "step": 2000,
-      "total_flos": 7.50335828249239e+18,
-      "train_loss": 1.4263343353271485,
-      "train_runtime": 3812.7792,
-      "train_samples_per_second": 16.786,
-      "train_steps_per_second": 0.525
+      "epoch": 2.5056947608200457,
+      "eval_loss": 0.7629979252815247,
+      "eval_runtime": 179.3715,
+      "eval_samples_per_second": 39.148,
+      "eval_steps_per_second": 4.895,
+      "eval_wer": 0.4721694415285879,
+      "step": 2200
+    },
+    {
+      "epoch": 2.733485193621868,
+      "eval_loss": 0.8067795038223267,
+      "eval_runtime": 179.4035,
+      "eval_samples_per_second": 39.141,
+      "eval_steps_per_second": 4.894,
+      "eval_wer": 0.48269309982859193,
+      "step": 2400
+    },
+    {
+      "epoch": 2.847380410022779,
+      "grad_norm": 5.402973175048828,
+      "learning_rate": 2.9999999999999997e-06,
+      "loss": 0.8574,
+      "step": 2500
+    },
+    {
+      "epoch": 2.847380410022779,
+      "step": 2500,
+      "total_flos": 9.363138171711945e+18,
+      "train_loss": 1.33308916015625,
+      "train_runtime": 4617.7601,
+      "train_samples_per_second": 17.324,
+      "train_steps_per_second": 0.541
     }
   ],
   "logging_steps": 500,
-  "max_steps": 2000,
+  "max_steps": 2500,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 500,
@@ -153,7 +178,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 7.50335828249239e+18,
+  "total_flos": 9.363138171711945e+18,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null