faster-whisper-medium-nyanja

README.md

@@ -16,14 +16,14 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- eval_loss: 0.4645
-- eval_wer: 0.2780
-- eval_cer: 0.0891
-- eval_runtime: 480.529
-- eval_samples_per_second: 1.199
-- eval_steps_per_second: 0.15
-- epoch: 1.3879
-- step: 1600
+- eval_loss: 0.4364
+- eval_wer: 0.2745
+- eval_cer: 0.1100
+- eval_runtime: 659.4612
+- eval_samples_per_second: 0.943
+- eval_steps_per_second: 0.118
+- epoch: 1.4780
+- step: 3000
 
 ## Model description
 
@@ -57,6 +57,6 @@ The following hyperparameters were used during training:
 ### Framework versions
 
 - Transformers 4.48.0
-- Pytorch 2.8.0+cu126
+- Pytorch 2.6.0+cu124
 - Datasets 3.6.0
-- Tokenizers 0.21.4
+- Tokenizers 0.21.2

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2accc30dbb7087dedb0e6da93ec0df02ea66fe1f7cff43b9c73fcb5a3307b1b0
+oid sha256:a1a0409513223152ae7cf2110c45fba5a884365aaa531519ac3901db49603cb9
 size 3055544304

runs/Oct27_21-52-24_6d7a776b9d70/events.out.tfevents.1761601956.6d7a776b9d70.37.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68eea34178698f41ffee8d8e6bbf90372373ab69c312f1edbfc2b7ddf5cafde3
+size 21679

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b42fcdd4228464785684848906472a811a7029daa8bfd596e123baf04d54a07b
-size 5905
+oid sha256:4b4d978bf267859d0fcf9950d539a4c739cd929dc00027e40a0d76e7c8677c1f
+size 5496