Training in progress, epoch 1

config.json

@@ -1,13 +1,19 @@
 {
-  "_name_or_path": "ntu-spml/distilhubert",
-  "activation_dropout": 0.1,
-  "apply_spec_augment": false,
+  "_name_or_path": "facebook/wav2vec2-base",
+  "activation_dropout": 0.0,
+  "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,
+  "codevector_dim": 256,
+  "contrastive_logits_temperature": 0.1,
   "conv_bias": false,
   "conv_dim": [
     512,
@@ -38,95 +44,135 @@
   ],
   "ctc_loss_reduction": "sum",
   "ctc_zero_infinity": false,
+  "diversity_loss_weight": 0.1,
   "do_stable_layer_norm": false,
   "eos_token_id": 2,
   "feat_extract_activation": "gelu",
   "feat_extract_norm": "group",
-  "feat_proj_dropout": 0.0,
-  "feat_proj_layer_norm": false,
+  "feat_proj_dropout": 0.1,
+  "feat_quantizer_dropout": 0.0,
   "final_dropout": 0.0,
+  "freeze_feat_extract_train": true,
   "hidden_act": "gelu",
   "hidden_dropout": 0.1,
   "hidden_size": 768,
   "id2label": {
-    "0": "\u062e\u0627\u0621",
-    "1": "\u062b\u0627\u0621",
-    "10": "\u0643\u0627\u0641",
-    "11": "\u0627\u0644\u0641",
-    "12": "\u0642\u0627\u0641",
-    "13": "\u0641\u0627\u0621",
-    "14": "\u0647\u0627\u0621",
-    "15": "\u062f\u0627\u0644",
-    "16": "\u0628\u0627\u0621",
-    "17": "\u0637\u0627\u0621",
+    "0": "\u0627\u0644\u0641",
+    "1": "\u0628\u0627\u0621",
+    "10": "\u0632\u0627\u064a",
+    "11": "\u0633\u064a\u0646",
+    "12": "\u0634\u064a\u0646",
+    "13": "\u0635\u0627\u062f",
+    "14": "\u0636\u0627\u062f",
+    "15": "\u0637\u0627\u0621",
+    "16": "\u0638\u0627\u0621",
+    "17": "\u0639\u064a\u0646",
     "18": "\u063a\u064a\u0646",
-    "19": "\u0633\u064a\u0646",
-    "2": "\u0646\u0648\u0646",
-    "20": "\u062a\u0627\u0621",
-    "21": "\u0638\u0627\u0621",
-    "22": "\u064a\u0627\u0621",
-    "23": "\u0630\u0627\u0644",
-    "24": "\u0644\u0627\u0645",
-    "25": "\u062c\u064a\u0645",
-    "26": "\u0639\u064a\u0646",
-    "27": "\u0648\u0627\u0648",
-    "3": "\u0645\u064a\u0645",
-    "4": "\u0636\u0627\u062f",
-    "5": "\u0631\u0627\u0621",
-    "6": "\u0634\u064a\u0646",
-    "7": "\u0632\u0627\u064a",
-    "8": "\u0635\u0627\u062f",
-    "9": "\u062d\u0627\u0621"
+    "19": "\u0641\u0627\u0621",
+    "2": "\u062a\u0627\u0621",
+    "20": "\u0642\u0627\u0641",
+    "21": "\u0643\u0627\u0641",
+    "22": "\u0644\u0627\u0645",
+    "23": "\u0645\u064a\u0645",
+    "24": "\u0646\u0648\u0646",
+    "25": "\u0647\u0627\u0621",
+    "26": "\u0648\u0627\u0648",
+    "27": "\u064a\u0627\u0621",
+    "3": "\u062b\u0627\u0621",
+    "4": "\u062c\u064a\u0645",
+    "5": "\u062d\u0627\u0621",
+    "6": "\u062e\u0627\u0621",
+    "7": "\u062f\u0627\u0644",
+    "8": "\u0630\u0627\u0644",
+    "9": "\u0631\u0627\u0621"
   },
   "initializer_range": 0.02,
   "intermediate_size": 3072,
   "label2id": {
-    "\u0627\u0644\u0641": "11",
-    "\u0628\u0627\u0621": "16",
-    "\u062a\u0627\u0621": "20",
-    "\u062b\u0627\u0621": "1",
-    "\u062c\u064a\u0645": "25",
-    "\u062d\u0627\u0621": "9",
-    "\u062e\u0627\u0621": "0",
-    "\u062f\u0627\u0644": "15",
-    "\u0630\u0627\u0644": "23",
-    "\u0631\u0627\u0621": "5",
-    "\u0632\u0627\u064a": "7",
-    "\u0633\u064a\u0646": "19",
-    "\u0634\u064a\u0646": "6",
-    "\u0635\u0627\u062f": "8",
-    "\u0636\u0627\u062f": "4",
-    "\u0637\u0627\u0621": "17",
-    "\u0638\u0627\u0621": "21",
-    "\u0639\u064a\u0646": "26",
+    "\u0627\u0644\u0641": "0",
+    "\u0628\u0627\u0621": "1",
+    "\u062a\u0627\u0621": "2",
+    "\u062b\u0627\u0621": "3",
+    "\u062c\u064a\u0645": "4",
+    "\u062d\u0627\u0621": "5",
+    "\u062e\u0627\u0621": "6",
+    "\u062f\u0627\u0644": "7",
+    "\u0630\u0627\u0644": "8",
+    "\u0631\u0627\u0621": "9",
+    "\u0632\u0627\u064a": "10",
+    "\u0633\u064a\u0646": "11",
+    "\u0634\u064a\u0646": "12",
+    "\u0635\u0627\u062f": "13",
+    "\u0636\u0627\u062f": "14",
+    "\u0637\u0627\u0621": "15",
+    "\u0638\u0627\u0621": "16",
+    "\u0639\u064a\u0646": "17",
     "\u063a\u064a\u0646": "18",
-    "\u0641\u0627\u0621": "13",
-    "\u0642\u0627\u0641": "12",
-    "\u0643\u0627\u0641": "10",
-    "\u0644\u0627\u0645": "24",
-    "\u0645\u064a\u0645": "3",
-    "\u0646\u0648\u0646": "2",
-    "\u0647\u0627\u0621": "14",
-    "\u0648\u0627\u0648": "27",
-    "\u064a\u0627\u0621": "22"
+    "\u0641\u0627\u0621": "19",
+    "\u0642\u0627\u0641": "20",
+    "\u0643\u0627\u0641": "21",
+    "\u0644\u0627\u0645": "22",
+    "\u0645\u064a\u0645": "23",
+    "\u0646\u0648\u0646": "24",
+    "\u0647\u0627\u0621": "25",
+    "\u0648\u0627\u0648": "26",
+    "\u064a\u0627\u0621": "27"
   },
   "layer_norm_eps": 1e-05,
   "layerdrop": 0.0,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
   "mask_feature_length": 10,
   "mask_feature_min_masks": 0,
   "mask_feature_prob": 0.0,
   "mask_time_length": 10,
   "mask_time_min_masks": 2,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
   "mask_time_prob": 0.05,
-  "model_type": "hubert",
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "no_mask_channel_overlap": false,
+  "no_mask_time_overlap": false,
+  "num_adapter_layers": 3,
   "num_attention_heads": 12,
+  "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": 2,
+  "num_hidden_layers": 12,
+  "num_negatives": 100,
+  "output_hidden_size": 768,
   "pad_token_id": 0,
+  "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": "4.42.3",
   "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 sha256:149889981493cc6b6c86e1c1e0aca5000b11f3721278b924351f93c6ef8b5529
-size 94790240
+oid sha256:003114ba63e705676627cc1d192700bc80f14aeb5ffe8bd17ad7b28d4b1ae360
+size 378329104

preprocessor_config.json

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

runs/Aug03_20-00-25_e97ec11884da/events.out.tfevents.1722715290.e97ec11884da.34.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72367e039faeaadc78cc465f029e60775f563ef3e23f1c6b0ce1d5d4461b03c2
+size 16767

runs/Aug03_20-13-58_e97ec11884da/events.out.tfevents.1722716095.e97ec11884da.34.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59455cad4592ee4948830932a3ed726372b34239a1f9ba5beefcec496136c8eb
+size 32071

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dd8ff9d63785cfa9c21d5d3c102863736b3f94b93b5bddd53ff40dc74e745c2e
+oid sha256:0df6344fdd92af6ed3162b79944db931ab54a5b1cad53703c97e1ce48a3e5a2b
 size 5176