Training in progress, step 250

config.json

@@ -17,7 +17,7 @@
   "decoder_layerdrop": 0.0,
   "decoder_layers": 12,
   "decoder_start_token_id": 50258,
-  "dropout": 0.2,
+  "dropout": 0.0,
   "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:a31797292a625078330a665bf71535ab8abc64c8396796e12ef1e43ddd38399e
+oid sha256:8e417bc82fa3fef7087c8b7372242f89b201be8d5a582a6a638d082d2de002f0
 size 967102601

run.sh

@@ -7,7 +7,7 @@ python run_speech_recognition_seq2seq_streaming.py \
        --model_index_name="Whisper Small Odia" \
        --max_steps="1000" \
        --output_dir="./" \
-       --per_device_train_batch_size="64" \
+       --per_device_train_batch_size="32" \
        --per_device_eval_batch_size="32" \
        --logging_steps="25" \
        --learning_rate="5e-6" \

run_speech_recognition_seq2seq_streaming.py

@@ -398,7 +398,7 @@ def main():
     )
 
     config.update({"forced_decoder_ids": model_args.forced_decoder_ids, "suppress_tokens": model_args.suppress_tokens})
-    config.update({"dropout": 0.2})
+    # config.update({"dropout": 0.02})
     
 
     if training_args.gradient_checkpointing:

runs/Dec22_00-11-50_129-146-20-112/1671667926.096101/events.out.tfevents.1671667926.129-146-20-112.1012225.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb1ab8cca2c2f5eede088d366ccb24ea2687a430228a9cbe169c821248017c33
+size 5870

runs/Dec22_00-11-50_129-146-20-112/events.out.tfevents.1671667926.129-146-20-112.1012225.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a92c062eadd1a67bcbc5961d7d9815d46f3ac384f7a7a083adc2e1292575e42c
+size 4270

runs/Dec22_00-13-49_129-146-20-112/1671668045.0982327/events.out.tfevents.1671668045.129-146-20-112.1029389.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f74315acddba8f6c3d4393b32e322f79fcadd4085fa8875d05cfdbec75c6880
+size 5870

runs/Dec22_00-13-49_129-146-20-112/events.out.tfevents.1671668045.129-146-20-112.1029389.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a521faaa49749e05b43aba517fa8add4161b1425ce80ce7d2ddd8015275211b
+size 6143

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5781f69bbe126ffff57509622a45f9506396b4128236c211b2c1d3ddfb93257d
+oid sha256:1e6c47b77ae95a383a4202b2562ac77d158360afc349d411422fc8691bda8d10
 size 3579