End of training

README.md

@@ -3,6 +3,8 @@ library_name: transformers
 license: apache-2.0
 base_model: facebook/wav2vec2-xls-r-300m
 tags:
+- automatic-speech-recognition
+- bemgen
 - generated_from_trainer
 metrics:
 - wer
@@ -16,10 +18,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # xls-r-300m-bemgen-combined-hp-tuning-test-model
 
-This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on an unknown dataset.
+This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the BEMGEN - NA dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.3925
-- Wer: 0.4228
+- Wer: 0.4224
 
 ## Model description
 

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 22.228163992869874,
+    "eval_loss": 0.3924838602542877,
+    "eval_runtime": 40.0406,
+    "eval_samples": 541,
+    "eval_samples_per_second": 13.511,
+    "eval_steps_per_second": 3.397,
+    "eval_wer": 0.42236746550472043,
+    "total_flos": 3.4912812658192863e+19,
+    "train_loss": 20.112410697937012,
+    "train_runtime": 9256.0708,
+    "train_samples": 4482,
+    "train_samples_per_second": 14.527,
+    "train_steps_per_second": 0.113
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 22.228163992869874,
+    "eval_loss": 0.3924838602542877,
+    "eval_runtime": 40.0406,
+    "eval_samples": 541,
+    "eval_samples_per_second": 13.511,
+    "eval_steps_per_second": 3.397,
+    "eval_wer": 0.42236746550472043
+}

runs/Dec24_19-41-58_srvrocgpu011.uct.ac.za/events.out.tfevents.1735144860.srvrocgpu011.uct.ac.za

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1296dc71e152b2e0fc62022a6e52b4725eab6ef101943ae660a27612252d715e
+size 358

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 22.228163992869874,
+    "total_flos": 3.4912812658192863e+19,
+    "train_loss": 20.112410697937012,
+    "train_runtime": 9256.0708,
+    "train_samples": 4482,
+    "train_samples_per_second": 14.527,
+    "train_steps_per_second": 0.113
+}

trainer_state.json

@@ -0,0 +1,130 @@
+{
+  "best_metric": 0.3383650779724121,
+  "best_model_checkpoint": "/scratch/skscla001/speech/results/xls-r-300m-bemgen-combined-hp-tuning-test-model/checkpoint-500",
+  "epoch": 22.228163992869874,
+  "eval_steps": 100,
+  "global_step": 800,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 2.7985739750445635,
+      "eval_loss": 2.9111883640289307,
+      "eval_runtime": 39.716,
+      "eval_samples_per_second": 13.622,
+      "eval_steps_per_second": 3.424,
+      "eval_wer": 1.0,
+      "step": 100
+    },
+    {
+      "epoch": 5.570409982174688,
+      "eval_loss": 1.6968202590942383,
+      "eval_runtime": 39.8919,
+      "eval_samples_per_second": 13.562,
+      "eval_steps_per_second": 3.409,
+      "eval_wer": 0.9970951343500363,
+      "step": 200
+    },
+    {
+      "epoch": 8.342245989304812,
+      "eval_loss": 0.3893207609653473,
+      "eval_runtime": 40.2019,
+      "eval_samples_per_second": 13.457,
+      "eval_steps_per_second": 3.383,
+      "eval_wer": 0.615831517792302,
+      "step": 300
+    },
+    {
+      "epoch": 11.114081996434937,
+      "eval_loss": 0.34530752897262573,
+      "eval_runtime": 40.0609,
+      "eval_samples_per_second": 13.504,
+      "eval_steps_per_second": 3.395,
+      "eval_wer": 0.5106753812636166,
+      "step": 400
+    },
+    {
+      "epoch": 13.9126559714795,
+      "grad_norm": 7.798405170440674,
+      "learning_rate": 0.0005819322340474784,
+      "loss": 31.2339,
+      "step": 500
+    },
+    {
+      "epoch": 13.9126559714795,
+      "eval_loss": 0.3383650779724121,
+      "eval_runtime": 39.984,
+      "eval_samples_per_second": 13.53,
+      "eval_steps_per_second": 3.401,
+      "eval_wer": 0.49092229484386346,
+      "step": 500
+    },
+    {
+      "epoch": 16.684491978609625,
+      "eval_loss": 0.3592090606689453,
+      "eval_runtime": 39.9246,
+      "eval_samples_per_second": 13.551,
+      "eval_steps_per_second": 3.406,
+      "eval_wer": 0.46187363834422657,
+      "step": 600
+    },
+    {
+      "epoch": 19.45632798573975,
+      "eval_loss": 0.367597371339798,
+      "eval_runtime": 39.9675,
+      "eval_samples_per_second": 13.536,
+      "eval_steps_per_second": 3.403,
+      "eval_wer": 0.4335511982570806,
+      "step": 700
+    },
+    {
+      "epoch": 22.228163992869874,
+      "eval_loss": 0.3925014138221741,
+      "eval_runtime": 39.8278,
+      "eval_samples_per_second": 13.583,
+      "eval_steps_per_second": 3.415,
+      "eval_wer": 0.42280319535221494,
+      "step": 800
+    },
+    {
+      "epoch": 22.228163992869874,
+      "step": 800,
+      "total_flos": 3.4912812658192863e+19,
+      "train_loss": 20.112410697937012,
+      "train_runtime": 9256.0708,
+      "train_samples_per_second": 14.527,
+      "train_steps_per_second": 0.113
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 1050,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 400,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 3
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.4912812658192863e+19,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}