Training in progress, epoch 0

config.json

@@ -1,14 +1,20 @@
 {
-  "_name_or_path": "facebook/hubert-base-ls960",
+  "_name_or_path": "facebook/wav2vec2-large-960h-lv60-self",
   "activation_dropout": 0.1,
+  "adapter_attn_dim": null,
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": false,
   "apply_spec_augment": true,
   "architectures": [
-    "HubertForSequenceClassification"
+    "Wav2Vec2ForSequenceClassification"
   ],
   "attention_dropout": 0.1,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
-  "conv_bias": false,
+  "codevector_dim": 256,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": true,
   "conv_dim": [
     512,
     512,
@@ -38,19 +44,20 @@
   ],
   "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,
+  "feat_quantizer_dropout": 0.0,
   "final_dropout": 0.1,
   "gradient_checkpointing": false,
   "hidden_act": "gelu",
   "hidden_dropout": 0.1,
   "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
+  "hidden_size": 1024,
   "id2label": {
     "0": "abroad",
     "1": "address",
@@ -68,7 +75,7 @@
     "9": "freeze"
   },
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "label2id": {
     "abroad": "0",
     "address": "1",
@@ -93,16 +100,43 @@
   "mask_time_length": 10,
   "mask_time_min_masks": 2,
   "mask_time_prob": 0.05,
-  "model_type": "hubert",
-  "num_attention_heads": 12,
+  "model_type": "wav2vec2",
+  "num_adapter_layers": 3,
+  "num_attention_heads": 16,
+  "num_codevector_groups": 2,
+  "num_codevectors_per_group": 320,
   "num_conv_pos_embedding_groups": 16,
   "num_conv_pos_embeddings": 128,
   "num_feat_extract_layers": 7,
-  "num_hidden_layers": 12,
+  "num_hidden_layers": 24,
+  "num_negatives": 100,
+  "output_hidden_size": 1024,
   "pad_token_id": 0,
-  "tokenizer_class": "Wav2Vec2CTCTokenizer",
+  "proj_codevector_dim": 256,
+  "tdnn_dilation": [
+    1,
+    2,
+    3,
+    1,
+    1
+  ],
+  "tdnn_dim": [
+    512,
+    512,
+    512,
+    512,
+    1500
+  ],
+  "tdnn_kernel": [
+    5,
+    3,
+    3,
+    1,
+    1
+  ],
   "torch_dtype": "float32",
   "transformers_version": "4.35.2",
   "use_weighted_layer_sum": false,
-  "vocab_size": 32
+  "vocab_size": 32,
+  "xvector_output_dim": 512
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2dcc44d5f4873c4fe6d7013a4b457ae797639fc87ae6555ff608ef0408675d4e
-size 378314288
+oid sha256:2326b4a24978acd3b20dfccfeee077120e7e19dcfb9ccea2e2c595fc0afee455
+size 1262871640

preprocessor_config.json

@@ -3,7 +3,7 @@
   "feature_extractor_type": "Wav2Vec2FeatureExtractor",
   "feature_size": 1,
   "padding_side": "right",
-  "padding_value": 0,
-  "return_attention_mask": false,
+  "padding_value": 0.0,
+  "return_attention_mask": true,
   "sampling_rate": 16000
 }

runs/Dec20_13-48-08_6c50f0686240/events.out.tfevents.1703080089.6c50f0686240.1077.1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:568e4763396dac1d1a6a27cb67c08c3209dd9dba3c586a63bcdcc4ea72d6d373
-size 11400
+oid sha256:f2e090a6152a42742bfe83dbb63b03bfe3eb2e0e134c7e7a0d1e5b2908b5bf4d
+size 12219

runs/Dec20_14-04-38_6c50f0686240/events.out.tfevents.1703081079.6c50f0686240.1077.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60ecd2e1f29ddde167b66ca89b83f8400d51883c2bc91be57ae1f5caa222a006
+size 6755

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:97a42cd734085a69fa5e52de72f42af40ccb5a2a2ccaf80de445575de0f4ef25
+oid sha256:21bcb2db7592df1e01a7ac9f190d324674f824db1e8b42d6cd05443833cafa44
 size 4600