Upload folder using huggingface_hub

added_tokens.json

@@ -0,0 +1,4 @@
+{
+  "</s>": 80,
+  "<s>": 79
+}

config.yaml

@@ -0,0 +1,71 @@
+seed: 42
+data:
+  hf_dataset: google/WaxalNLP
+  languages:
+  - lin
+  - sna
+  - lug
+  configs:
+    lin: lin_asr
+    sna: sna_asr
+    lug: lug_asr
+  sample_rate: 16000
+  max_duration_s: 40.0
+  min_duration_s: 0.5
+  holdout: speaker_disjoint
+  holdout_frac: 0.05
+  limit: null
+model:
+  type: w2v_bert
+  id: facebook/w2v-bert-2.0
+  language_adapter: null
+  freeze_encoder: false
+  gradient_checkpointing: true
+  add_adapter: true
+  freeze_base: false
+train:
+  output_dir: experiments/w2vbert_multi_s42
+  per_device_batch_size: 4
+  grad_accum: 8
+  lr: 0.0001
+  warmup_ratio: 0.1
+  num_epochs: 5
+  max_steps: -1
+  save_steps: 1000
+  eval_steps: 1000
+  eval_max_samples: 800
+  save_total_limit: 1
+  fp16: true
+  bf16: false
+  auto_find_batch_size: true
+  resume: true
+  hub_model_id: anyantudre/waxal-w2vbert_multi_s42
+augment:
+  enabled: false
+  noise_dir: null
+  rir_dir: null
+  noise_min_snr: 5.0
+  noise_max_snr: 25.0
+  p_noise: 0.5
+  p_rir: 0.3
+  p_gaussian: 0.2
+  p_gain: 0.4
+  p_pitch: 0.25
+  p_speed: 0.25
+  p_air: 0.15
+  p_mp3: 0.2
+decode:
+  num_beams: 1
+  max_new_tokens: 225
+  ctc_beam_width: 1
+  lm: null
+  lm_alpha: 0.5
+  lm_beta: 1.5
+normalize:
+  lowercase: true
+  nfc: true
+  strip_punct: true
+  collapse_ws: true
+infer:
+  language_from_id: false
+  audio_dir: null

preprocessor_config.json

@@ -0,0 +1,11 @@
+{
+  "feature_extractor_type": "SeamlessM4TFeatureExtractor",
+  "feature_size": 80,
+  "num_mel_bins": 80,
+  "padding_side": "right",
+  "padding_value": 1,
+  "processor_class": "Wav2Vec2BertProcessor",
+  "return_attention_mask": true,
+  "sampling_rate": 16000,
+  "stride": 2
+}

special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "[PAD]",
+  "unk_token": "[UNK]"
+}

tokenizer_config.json

@@ -0,0 +1,49 @@
+{
+  "added_tokens_decoder": {
+    "77": {
+      "content": "[UNK]",
+      "lstrip": true,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": false
+    },
+    "78": {
+      "content": "[PAD]",
+      "lstrip": true,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": false
+    },
+    "79": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "80": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "do_lower_case": false,
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "[PAD]",
+  "processor_class": "Wav2Vec2BertProcessor",
+  "replace_word_delimiter_char": " ",
+  "target_lang": null,
+  "tokenizer_class": "Wav2Vec2CTCTokenizer",
+  "unk_token": "[UNK]",
+  "word_delimiter_token": "|"
+}

vocab.json

@@ -0,0 +1,81 @@
+{
+  "&": 0,
+  "'": 1,
+  "-": 2,
+  "/": 3,
+  "0": 4,
+  "1": 5,
+  "2": 6,
+  "3": 7,
+  "4": 8,
+  "5": 9,
+  "6": 10,
+  "7": 11,
+  "8": 12,
+  "9": 13,
+  "<": 14,
+  "=": 15,
+  ">": 16,
+  "[PAD]": 78,
+  "[UNK]": 77,
+  "\\": 17,
+  "`": 18,
+  "a": 19,
+  "b": 20,
+  "c": 21,
+  "d": 22,
+  "e": 23,
+  "f": 24,
+  "g": 25,
+  "h": 26,
+  "i": 27,
+  "j": 28,
+  "k": 29,
+  "l": 30,
+  "m": 31,
+  "n": 32,
+  "o": 33,
+  "p": 34,
+  "q": 35,
+  "r": 36,
+  "s": 37,
+  "t": 38,
+  "u": 39,
+  "v": 40,
+  "w": 41,
+  "x": 42,
+  "y": 43,
+  "z": 44,
+  "|": 76,
+  "×": 45,
+  "à": 46,
+  "á": 47,
+  "â": 48,
+  "ç": 49,
+  "è": 50,
+  "é": 51,
+  "ê": 52,
+  "ë": 53,
+  "ì": 54,
+  "í": 55,
+  "î": 56,
+  "ï": 57,
+  "ñ": 58,
+  "ò": 59,
+  "ó": 60,
+  "ô": 61,
+  "ù": 62,
+  "ú": 63,
+  "û": 64,
+  "ü": 65,
+  "þ": 66,
+  "ā": 67,
+  "ĝ": 68,
+  "ķ": 69,
+  "ĺ": 70,
+  "ŋ": 71,
+  "œ": 72,
+  "ᵑ": 73,
+  "⭐": 74,
+  "️": 75
+}