Training in progress, step 500

config.json

@@ -17,7 +17,7 @@
   "decoder_layerdrop": 0.0,
   "decoder_layers": 12,
   "decoder_start_token_id": 50258,
-  "dropout": 0.1,
+  "dropout": 0.3,
   "encoder_attention_heads": 12,
   "encoder_ffn_dim": 3072,
   "encoder_layerdrop": 0.0,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7070c083059cf90be1a5185187847064565258a25781914bb16b1fd5e9ca11f0
+oid sha256:66bdb32f6748b24d1bc6561813d1ac5fe46c19c84986194087f457087f967899
 size 967102601

run.sh

@@ -6,7 +6,7 @@ python run_speech_recognition_seq2seq_streaming.py \
       --train_split_name="train+validation" \
       --eval_split_name="test" \
       --model_index_name="Whisper Small Pashto" \
-      --max_steps="400" \
+      --max_steps="500" \
       --output_dir="./" \
       --per_device_train_batch_size="16" \
       --per_device_eval_batch_size="8" \

run_speech_recognition_seq2seq_streaming.py

@@ -556,7 +556,7 @@ def main():
             elif isinstance(train_dataloader.dataset, IterableDataset):
                 train_dataloader.dataset.set_epoch(train_dataloader.dataset._epoch + 1)
 
-    model.config.dropout=0.1            
+    model.config.dropout=0.3            
 
     # Initialize Trainer
     trainer = Seq2SeqTrainer(

runs/Dec16_17-39-22_129-146-104-29/1671212380.5544906/events.out.tfevents.1671212380.129-146-104-29.136366.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cef281faa170f3723ff01b6f97198d3f90c9905775189de3c46e59a7d1496f94
+size 5870

runs/Dec16_17-39-22_129-146-104-29/events.out.tfevents.1671212380.129-146-104-29.136366.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa9c408a294e7c7e6a7d1feeee57cc0df9c6cd1d8e4d85705629d3e69dedfd42
+size 6152

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4c1898cd38db80433212b8ad15a044c342bdececf727b934c93897a580643187
+oid sha256:d7c6a47bb7da518d5ca6005b4c205f6bc7b8400020d318860e2865f05476053e
 size 3579