ihanif
/

whisper_small_ps_augmented

Automatic Speech Recognition

Generated from Trainer

Inference Endpoints

Model card Files Files and versions Metrics Training metrics Community

ihanif commited on Dec 20, 2022

Commit

e8768b1

•

1 Parent(s): 0deb8f5

fix: use correct lanauge iso code

Files changed (2) hide show

README.md +1 -1
whisper_small_ps_augmented.py +1 -1

README.md CHANGED Viewed

@@ -1,6 +1,6 @@
 ---
 language:
-- ps_af
 license: apache-2.0
 tags:
 - whisper-event

 ---
 language:
+- ps
 license: apache-2.0
 tags:
 - whisper-event

whisper_small_ps_augmented.py CHANGED Viewed

@@ -256,7 +256,7 @@ training_args = Seq2SeqTrainingArguments(
     gradient_accumulation_steps=16,
     learning_rate=1e-5,
     warmup_steps=30,
-    max_steps=300,
     gradient_checkpointing=True,
     fp16=True,
     evaluation_strategy="steps",

     gradient_accumulation_steps=16,
     learning_rate=1e-5,
     warmup_steps=30,
+    max_steps=500,
     gradient_checkpointing=True,
     fp16=True,
     evaluation_strategy="steps",