wav2vec2-large-xls-r-300m-hungarian-colab

README.md

@@ -19,9 +19,9 @@ This model is a fine-tuned version of [facebook/wav2vec2-lv-60-espeak-cv-ft](htt
 It achieves the following results on the evaluation set:
 - eval_loss: 17417.8594
 - eval_wer: 0.9967
-- eval_runtime: 120.1605
-- eval_samples_per_second: 4.161
-- eval_steps_per_second: 0.524
+- eval_runtime: 122.9984
+- eval_samples_per_second: 4.065
+- eval_steps_per_second: 0.512
 - step: 0
 
 ## Model description
@@ -42,11 +42,11 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 0.0003
-- train_batch_size: 8
+- train_batch_size: 4
 - eval_batch_size: 8
 - seed: 42
 - gradient_accumulation_steps: 2
-- total_train_batch_size: 16
+- total_train_batch_size: 8
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 500

runs/Nov29_06-37-04_a33e6f86237e/events.out.tfevents.1701239952.a33e6f86237e.233.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e05bebaae93513cd01548dbc5316c223be7df4065c8c79bbcae02606f482495b
+size 341

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:621edd08f4c5b95f8c8180a71a40b193357f674c3ea04abab1addcdc784ac4f1
+oid sha256:8faea839b3103ac4931f0ddfafa5ab0b82f2c2a3547e238be819d577af82edc4
 size 4600