Training in progress, step 700

config.json

@@ -17,7 +17,7 @@
   "decoder_layerdrop": 0.0,
   "decoder_layers": 12,
   "decoder_start_token_id": 50258,
-  "dropout": 0.3,
+  "dropout": 0.1,
   "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 sha256:5cf458b90e3f790664a080179b0fb5349b651c5b2b960ff48ec721d68ab069d1
+oid sha256:6c7e79a8eb751ca7749426ae327fd153ec6bed1198758206d569ccd4a7cba534
 size 967102601

run.sh

@@ -6,10 +6,11 @@ python run_speech_recognition_seq2seq_streaming.py \
       --train_split_name="train+validation" \
       --eval_split_name="test" \
       --model_index_name="Whisper Small Pashto" \
-      --max_steps="600" \
+      --max_steps="700" \
       --output_dir="./" \
-      --per_device_train_batch_size="16" \
-      --per_device_eval_batch_size="8" \
+      --per_device_train_batch_size="32" \
+      --per_device_eval_batch_size="16" \
+      --gradient_accumulation_steps="2" \
       --logging_steps="10" \
       --learning_rate="1e-5" \
       --warmup_steps="500" \

runs/Dec16_18-23-51_129-146-104-29/1671215049.2554681/events.out.tfevents.1671215049.129-146-104-29.139583.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb6788e49bf1d8cdd0065b404d37dd7b2d58f658a66af2d0f9c0b9d462bbb015
+size 5870

runs/Dec16_18-23-51_129-146-104-29/events.out.tfevents.1671215049.129-146-104-29.139583.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a5419632071b8435e2c708c04d6dfb7d438473db53a03ee83c4dc98be84ec1c
+size 6153

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cb9de4b5b20cc9259c3b118b92455d62825571fa95245c161ad8a861c2dc5b3a
+oid sha256:013d703d0044ede131afa657620baa61d0afd0e3beb92cdb70877eb99dea606d
 size 3579