jonatasgrosman/exp_w2v2r_es_vp-100k_age_teens-5_sixties-5_s12

.gitattributes

@@ -32,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+cv7_transcriptions.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,17 @@
+---
+language:
+- es
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- es
+datasets:
+- mozilla-foundation/common_voice_7_0
+---
+# exp_w2v2r_es_vp-100k_age_teens-5_sixties-5_s12
+
+Fine-tuned [facebook/wav2vec2-large-100k-voxpopuli](https://huggingface.co/facebook/wav2vec2-large-100k-voxpopuli) for speech recognition using the train split of [Common Voice 7.0 (es)](https://huggingface.co/datasets/mozilla-foundation/common_voice_7_0).
+When using this model, make sure that your speech input is sampled at 16kHz.
+
+This model has been fine-tuned by the [HuggingSound](https://github.com/jonatasgrosman/huggingsound) tool.
+    

config.json

@@ -0,0 +1,116 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-100k-voxpopuli",
+  "activation_dropout": 0.05,
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": false,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.05,
+  "bos_token_id": 1,
+  "classifier_proj_size": 256,
+  "codevector_dim": 768,
+  "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": 0.05,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.05,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.05,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.05,
+  "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,
+  "mask_time_selection": "static",
+  "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": 24,
+  "num_negatives": 100,
+  "output_hidden_size": 1024,
+  "pad_token_id": 0,
+  "proj_codevector_dim": 768,
+  "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": "4.18.0",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 41,
+  "xvector_output_dim": 512
+}

cv7_evaluation.json

@@ -0,0 +1,300 @@
+{
+  "validation": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.05367210784664797,
+      "wer": 0.1902652530400454
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.05355847174196474,
+      "wer": 0.18975086911065966
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.053226628832383785,
+      "wer": 0.19148936170212766
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.07089884307327203,
+      "wer": 0.21378504672897197
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.050806451612903224,
+      "wer": 0.19339622641509435
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.05240584372068088,
+      "wer": 0.18980738362760835
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.055711437643062835,
+      "wer": 0.1954715871498934
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.05102483723173378,
+      "wer": 0.1923519263944796
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.04799075889106693,
+      "wer": 0.17628541448058763
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.053490213825093336,
+      "wer": 0.18960706319387313
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.05512316196722161,
+      "wer": 0.19357588438141735
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.06521739130434782,
+      "wer": 0.24593128390596744
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.05055550368779759,
+      "wer": 0.18490775419614372
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.053423067997215135,
+      "wer": 0.18995674850527922
+    },
+    {
+      "subset_type": "accent",
+      "subset": "chileno",
+      "cer": 0.0636669372608141,
+      "wer": 0.22823851953392735
+    },
+    {
+      "subset_type": "accent",
+      "subset": "nortepeninsular",
+      "cer": 0.04948512585812357,
+      "wer": 0.17703081232492998
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canario",
+      "cer": 0.049889948642699924,
+      "wer": 0.2018348623853211
+    },
+    {
+      "subset_type": "accent",
+      "subset": "surpeninsular",
+      "cer": 0.06448221482390047,
+      "wer": 0.22152560083594566
+    },
+    {
+      "subset_type": "accent",
+      "subset": "caribe",
+      "cer": 0.05323046181172291,
+      "wer": 0.18910791847644504
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mexicano",
+      "cer": 0.05182722718563151,
+      "wer": 0.18424326192121632
+    },
+    {
+      "subset_type": "accent",
+      "subset": "rioplatense",
+      "cer": 0.05325338399430064,
+      "wer": 0.18982808022922637
+    },
+    {
+      "subset_type": "accent",
+      "subset": "andino",
+      "cer": 0.05324268581468184,
+      "wer": 0.18512622242438026
+    },
+    {
+      "subset_type": "accent",
+      "subset": "centrosurpeninsular",
+      "cer": 0.06685883577548242,
+      "wer": 0.21274601686972822
+    },
+    {
+      "subset_type": "accent",
+      "subset": "americacentral",
+      "cer": 0.05657857074706612,
+      "wer": 0.18879907621247113
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.061463841685814966,
+      "wer": 0.20691204355208478
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.06191477461054027,
+      "wer": 0.20782898820608317
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.05683870719803236,
+      "wer": 0.1960919540229885
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.05946366109599689,
+      "wer": 0.18915801614763553
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.15384615384615385,
+      "wer": 0.35365853658536583
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.05918812150933116,
+      "wer": 0.19418651594670974
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.05961711937456149,
+      "wer": 0.2056635469162467
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.026845637583892617,
+      "wer": 0.1111111111111111
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.06564669407562175,
+      "wer": 0.2323972323972324
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.053198307909242404,
+      "wer": 0.1848133848133848
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.061966438884673895,
+      "wer": 0.20807869563533557
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.058299430824386614,
+      "wer": 0.1978142708453799
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.10616929698708752,
+      "wer": 0.3217391304347826
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.05710561544423307,
+      "wer": 0.20274841437632135
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.06211773653111566,
+      "wer": 0.20864996855229567
+    },
+    {
+      "subset_type": "accent",
+      "subset": "chileno",
+      "cer": 0.04657618753855645,
+      "wer": 0.17200365965233302
+    },
+    {
+      "subset_type": "accent",
+      "subset": "nortepeninsular",
+      "cer": 0.0525718998041439,
+      "wer": 0.17729204614450517
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canario",
+      "cer": 0.059670781893004114,
+      "wer": 0.21019108280254778
+    },
+    {
+      "subset_type": "accent",
+      "subset": "surpeninsular",
+      "cer": 0.05827619980411362,
+      "wer": 0.22518518518518518
+    },
+    {
+      "subset_type": "accent",
+      "subset": "caribe",
+      "cer": 0.070375237260914,
+      "wer": 0.21941830624465355
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mexicano",
+      "cer": 0.055159817351598174,
+      "wer": 0.189423614997202
+    },
+    {
+      "subset_type": "accent",
+      "subset": "rioplatense",
+      "cer": 0.07146464646464647,
+      "wer": 0.22595078299776286
+    },
+    {
+      "subset_type": "accent",
+      "subset": "andino",
+      "cer": 0.05039125367667462,
+      "wer": 0.1827597728032075
+    },
+    {
+      "subset_type": "accent",
+      "subset": "centrosurpeninsular",
+      "cer": 0.04458986455638411,
+      "wer": 0.17410714285714285
+    },
+    {
+      "subset_type": "accent",
+      "subset": "americacentral",
+      "cer": 0.05033425088478175,
+      "wer": 0.1724137931034483
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:189eb6195d5076ad2db47e3ba37e10ead762a1fddadf39f20d59fac5cbf19a87
+size 125208611

preprocessor_config.json

@@ -0,0 +1,10 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "processor_class": "Wav2Vec2Processor",
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b8134ee60fed7f5db2a8de812762ffec9f6806659833672186aec615b2f7167
+size 1262066801

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "<pad>", "do_lower_case": false, "word_delimiter_token": "|", "replace_word_delimiter_char": " ", "special_tokens_map_file": null, "name_or_path": "/tmp/tmpdol17pab", "tokenizer_class": "Wav2Vec2CTCTokenizer", "processor_class": "Wav2Vec2Processor"}

vocab.json

@@ -0,0 +1 @@
+{"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "'": 5, "-": 6, "a": 7, "b": 8, "c": 9, "d": 10, "e": 11, "f": 12, "g": 13, "h": 14, "i": 15, "j": 16, "k": 17, "l": 18, "m": 19, "n": 20, "o": 21, "p": 22, "q": 23, "r": 24, "s": 25, "t": 26, "u": 27, "v": 28, "w": 29, "x": 30, "y": 31, "z": 32, "á": 33, "é": 34, "í": 35, "ñ": 36, "ó": 37, "ö": 38, "ú": 39, "ü": 40}