Training in progress, step 1000

config.json

@@ -54,12 +54,12 @@
   "final_dropout": 0.0,
   "gradient_checkpointing": false,
   "hidden_act": "gelu",
-  "hidden_dropout": 0.08,
+  "hidden_dropout": 0.05,
   "hidden_size": 1024,
   "initializer_range": 0.02,
   "intermediate_size": 4096,
   "layer_norm_eps": 1e-05,
-  "layerdrop": 0.3,
+  "layerdrop": 0.25,
   "mask_channel_length": 10,
   "mask_channel_min_space": 1,
   "mask_channel_other": 0.0,
@@ -82,7 +82,7 @@
   "num_conv_pos_embedding_groups": 16,
   "num_conv_pos_embeddings": 128,
   "num_feat_extract_layers": 7,
-  "num_hidden_layers": 32,
+  "num_hidden_layers": 24,
   "num_negatives": 100,
   "output_hidden_size": 1024,
   "pad_token_id": 69,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9d9e53a707a0c12bd3bb5f63cde3aa57ed1a345070dad78003250db99a94388c
-size 1665320813
+oid sha256:561a654b417c3744ec5b46c40fe86535fe717c3c86a483f91291e8a3dc0e8aa2
+size 1262197037

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cd38da3854fe738a469da4e3feb290654e3235a80d62e8093d9b52b8df640ae5
+oid sha256:3b887c852210af941b93e4bb319b1ec7b70c0ccd532c4f5bf434a6311fa295a0
 size 3579