base_epoch_2.0

README.md

@@ -15,13 +15,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [nguyenminhly/base](https://huggingface.co/nguyenminhly/base) on the None dataset.
 It achieves the following results on the evaluation set:
-- eval_loss: 3.3248
-- eval_wer: 1.0
-- eval_runtime: 16.6109
-- eval_samples_per_second: 22.576
-- eval_steps_per_second: 2.829
-- epoch: 0.56
-- step: 500
+- eval_loss: 0.2759
+- eval_wer: 0.1653
+- eval_runtime: 16.3191
+- eval_samples_per_second: 22.979
+- eval_steps_per_second: 2.88
+- epoch: 1.69
+- step: 1500
 
 ## Model description
 

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:774b1d208e3041efae123e53f06147b60df7c0860e22875afdeeaa7f9f383bc6
+oid sha256:c05d6cd0e09fdf1aecc0d66ff953f999274945b428871d610f0e25aedf9c8c3b
 size 377898657