Training in progress, step 1000

config.json

@@ -0,0 +1,163 @@
+{
+  "_name_or_path": "facebook/mms-lid-126",
+  "activation_dropout": 0.0,
+  "adapter_attn_dim": 16,
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": false,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "bos_token_id": 1,
+  "classifier_proj_size": 1024,
+  "codevector_dim": 1024,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": true,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": 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": "layer",
+  "feat_proj_dropout": 0.1,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.1,
+  "finetuning_task": "audio-classification",
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 1280,
+  "id2label": {
+    "0": "Hungarian",
+    "1": "Czech",
+    "10": "Spanish",
+    "11": "Dutch",
+    "12": "German",
+    "13": "South african",
+    "14": "Estonian",
+    "15": "Eastern european",
+    "16": "Finnish",
+    "17": "African",
+    "18": "Scottish",
+    "19": "Polish",
+    "2": "English",
+    "20": "Chinese",
+    "21": "Indian",
+    "22": "Latin american",
+    "23": "Lithuanian",
+    "24": "Egyptian",
+    "3": "Italian",
+    "4": "Irish",
+    "5": "Vietnamese",
+    "6": "French",
+    "7": "Romanian",
+    "8": "Slovak",
+    "9": "American"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 5120,
+  "label2id": {
+    "African": "17",
+    "American": "9",
+    "Chinese": "20",
+    "Czech": "1",
+    "Dutch": "11",
+    "Eastern european": "15",
+    "Egyptian": "24",
+    "English": "2",
+    "Estonian": "14",
+    "Finnish": "16",
+    "French": "6",
+    "German": "12",
+    "Hungarian": "0",
+    "Indian": "21",
+    "Irish": "4",
+    "Italian": "3",
+    "Latin american": "22",
+    "Lithuanian": "23",
+    "Polish": "19",
+    "Romanian": "7",
+    "Scottish": "18",
+    "Slovak": "8",
+    "South african": "13",
+    "Spanish": "10",
+    "Vietnamese": "5"
+  },
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.0,
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.1,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.1,
+  "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": 48,
+  "num_negatives": 100,
+  "output_hidden_size": 1280,
+  "pad_token_id": 0,
+  "proj_codevector_dim": 1024,
+  "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": "4.38.0.dev0",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 154,
+  "xvector_output_dim": 512
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43bc1290d03f49a235eee306a9ac1bebc0d52484f1f2889e2b8a856a3184d06a
+size 3864081756

preprocessor_config.json

@@ -0,0 +1,9 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

runs/Feb16_19-15-14_hf-dgx-01/events.out.tfevents.1708107337.hf-dgx-01.3660307.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b559134414ab1b6e10cee4a9cb2d7ecf03ac308ebf0d8b8d9e87cdcaab46dcb
+size 23188

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abbe5e9d4e0b0f02edf7dcf0ba6b1b6a44761dd5d5840f39cf055bcfd3085f58
+size 4920