Training in progress, epoch 1

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "facebook/hubert-base-ls960",
+  "_name_or_path": "facebook/hubert-large-ls960-ft",
   "activation_dropout": 0.1,
   "apply_spec_augment": true,
   "architectures": [
@@ -8,7 +8,7 @@
   "attention_dropout": 0.1,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
-  "conv_bias": false,
+  "conv_bias": true,
   "conv_dim": [
     512,
     512,
@@ -38,11 +38,12 @@
   ],
   "ctc_loss_reduction": "sum",
   "ctc_zero_infinity": false,
-  "do_stable_layer_norm": false,
+  "diversity_loss_weight": 0.1,
+  "do_stable_layer_norm": true,
   "eos_token_id": 2,
   "feat_extract_activation": "gelu",
   "feat_extract_dropout": 0.0,
-  "feat_extract_norm": "group",
+  "feat_extract_norm": "layer",
   "feat_proj_dropout": 0.1,
   "feat_proj_layer_norm": true,
   "final_dropout": 0.1,
@@ -50,7 +51,7 @@
   "hidden_act": "gelu",
   "hidden_dropout": 0.1,
   "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
+  "hidden_size": 1024,
   "id2label": {
     "0": "Brazil",
     "1": "Canada",
@@ -65,7 +66,7 @@
     "9": "United States"
   },
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "label2id": {
     "Brazil": "0",
     "Canada": "1",
@@ -88,13 +89,12 @@
   "mask_time_min_masks": 2,
   "mask_time_prob": 0.05,
   "model_type": "hubert",
-  "num_attention_heads": 12,
+  "num_attention_heads": 16,
   "num_conv_pos_embedding_groups": 16,
   "num_conv_pos_embeddings": 128,
   "num_feat_extract_layers": 7,
-  "num_hidden_layers": 12,
+  "num_hidden_layers": 24,
   "pad_token_id": 0,
-  "tokenizer_class": "Wav2Vec2CTCTokenizer",
   "torch_dtype": "float32",
   "transformers_version": "4.46.2",
   "use_weighted_layer_sum": false,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e75e07742fb66b1543f3f827e8cb8bbf0515150a412c031b2644ea284cc8c887
-size 378311204
+oid sha256:8272495ca4409cb3989449b601ad24a3a51811ca17ac63a032cd23ded98e9352
+size 1262867716

preprocessor_config.json

@@ -4,6 +4,6 @@
   "feature_size": 1,
   "padding_side": "right",
   "padding_value": 0,
-  "return_attention_mask": false,
+  "return_attention_mask": true,
   "sampling_rate": 16000
 }

runs/Jun28_10-01-34_gl1703.arc-ts.umich.edu/events.out.tfevents.1751119295.gl1703.arc-ts.umich.edu.640729.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:273bb7a6723b3550a6433b4363a4157544aa6721a9217fa6b32a8abdda8b74ed
-size 174621
+oid sha256:25f9d93e5b559035a5873d17ee97677fa546809db48eb52a817b85b005db7e6a
+size 178320

runs/Jun28_13-16-37_gl1703.arc-ts.umich.edu/events.out.tfevents.1751131001.gl1703.arc-ts.umich.edu.710979.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa3616cafa23f6ecade4159a8d24c739b464f8270b4d5e54e9e25d3eb1bddb60
+size 10056

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3ffafda35c1c9f49fe249c520a1bcdb43a956d2644824cbfd575e24e2aec71bc
+oid sha256:8d9bffae36006adacfc86e0c417a7dadb510228d476c113f3ceb018c54806ce5
 size 5304