Training in progress, step 500

config.json

@@ -8,7 +8,7 @@
   "architectures": [
     "Wav2Vec2ForCTC"
   ],
-  "attention_dropout": 0.05,
+  "attention_dropout": 0.08,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
   "codevector_dim": 768,
@@ -51,15 +51,15 @@
   "feat_extract_norm": "layer",
   "feat_proj_dropout": 0.1,
   "feat_quantizer_dropout": 0.0,
-  "final_dropout": 0.0,
+  "final_dropout": 0.08,
   "gradient_checkpointing": false,
   "hidden_act": "gelu",
-  "hidden_dropout": 0.05,
+  "hidden_dropout": 0.1,
   "hidden_size": 1024,
   "initializer_range": 0.02,
   "intermediate_size": 4096,
   "layer_norm_eps": 1e-05,
-  "layerdrop": 0.2,
+  "layerdrop": 0.1,
   "mask_channel_length": 10,
   "mask_channel_min_space": 1,
   "mask_channel_other": 0.0,
@@ -72,7 +72,7 @@
   "mask_time_min_masks": 2,
   "mask_time_min_space": 1,
   "mask_time_other": 0.0,
-  "mask_time_prob": 0.05,
+  "mask_time_prob": 0.1,
   "mask_time_selection": "static",
   "model_type": "wav2vec2",
   "num_adapter_layers": 3,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9b642a0a4de70e423f04dcca15da2e27abd656501c3d387ae5eba20afc6e3bc1
+oid sha256:197b721d4fd9fc273a5e0da7b378ef9a861a5b65550ea6458c8382a514cb1c8b
 size 1262197037

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dae0c5d649664a68c1678f38b012848fe3f91e787a5560bc62a85fbb34ba5ac3
+oid sha256:9f077cbde20090100685bb0ec1b4565b5d3acec9d67b8a09f68517d951da80b1
 size 3579