Training in progress, step 400

config.json

@@ -8,7 +8,7 @@
   "architectures": [
     "Wav2Vec2ForCTC"
   ],
-  "attention_dropout": 0.1,
+  "attention_dropout": 0.05,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
   "codevector_dim": 768,
@@ -52,14 +52,13 @@
   "feat_proj_dropout": 0.1,
   "feat_quantizer_dropout": 0.0,
   "final_dropout": 0.0,
-  "gradient_checkpointing": false,
   "hidden_act": "gelu",
-  "hidden_dropout": 0.1,
+  "hidden_dropout": 0.05,
   "hidden_size": 1024,
   "initializer_range": 0.02,
   "intermediate_size": 4096,
   "layer_norm_eps": 1e-05,
-  "layerdrop": 0.1,
+  "layerdrop": 0.12,
   "mask_channel_length": 10,
   "mask_channel_min_space": 1,
   "mask_channel_other": 0.0,
@@ -72,11 +71,11 @@
   "mask_time_min_masks": 2,
   "mask_time_min_space": 1,
   "mask_time_other": 0.0,
-  "mask_time_prob": 0.075,
+  "mask_time_prob": 0.05,
   "mask_time_selection": "static",
   "model_type": "wav2vec2",
   "num_adapter_layers": 3,
-  "num_attention_heads": 16,
+  "num_attention_heads": 8,
   "num_codevector_groups": 2,
   "num_codevectors_per_group": 320,
   "num_conv_pos_embedding_groups": 16,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2ebfddb0530761bfcbf61046f8cbcc619fbed78e873799bf86a22e740c132273
+oid sha256:78aeda35d59f625c75b860797782753fe7ed0c65cabb60bc2cd7e3c7df8aa5b9
 size 1262188845

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ce2c0c66a21281345382990dca0beda63eb3d9666127009f53b51a84eaeef0a7
+oid sha256:b9d7a0c25ff473c03d6320028d8353afebd2f006cddb30dbd3c022e085c67636
 size 3579