jonatasgrosman/exp_w2v2r_es_vp-100k_age_teens-10_sixties-0_s613

.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-10_sixties-0_s613
+
+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": 0.053127806210988504,
+      "wer": 0.1875123780350942
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.05295618162235701,
+      "wer": 0.18738790809951636
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.05329313440776785,
+      "wer": 0.19069644509052464
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.0782161574211411,
+      "wer": 0.20794392523364486
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.04596774193548387,
+      "wer": 0.19339622641509435
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.05434928293794398,
+      "wer": 0.19662921348314608
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.05385315907356814,
+      "wer": 0.18745864882746452
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.05097660959729925,
+      "wer": 0.18890166762507188
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.0473956874824979,
+      "wer": 0.17229800629590766
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.05287928498698948,
+      "wer": 0.18717294138600762
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.05470192924586382,
+      "wer": 0.18993374404277577
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.06340579710144928,
+      "wer": 0.244122965641953
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.050765568107552984,
+      "wer": 0.18005271188791788
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.05280907071577761,
+      "wer": 0.18754770385447145
+    },
+    {
+      "subset_type": "accent",
+      "subset": "chileno",
+      "cer": 0.06389887510147281,
+      "wer": 0.2213845099383139
+    },
+    {
+      "subset_type": "accent",
+      "subset": "nortepeninsular",
+      "cer": 0.05034324942791762,
+      "wer": 0.1781512605042017
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canario",
+      "cer": 0.046221570066030816,
+      "wer": 0.1834862385321101
+    },
+    {
+      "subset_type": "accent",
+      "subset": "surpeninsular",
+      "cer": 0.0639565445943578,
+      "wer": 0.21525600835945663
+    },
+    {
+      "subset_type": "accent",
+      "subset": "caribe",
+      "cer": 0.05278641207815275,
+      "wer": 0.18843969261610424
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mexicano",
+      "cer": 0.05145785719232634,
+      "wer": 0.17816171389080857
+    },
+    {
+      "subset_type": "accent",
+      "subset": "rioplatense",
+      "cer": 0.051947280930895276,
+      "wer": 0.18302292263610315
+    },
+    {
+      "subset_type": "accent",
+      "subset": "andino",
+      "cer": 0.0523642196929188,
+      "wer": 0.18148737775756196
+    },
+    {
+      "subset_type": "accent",
+      "subset": "centrosurpeninsular",
+      "cer": 0.07126271118584354,
+      "wer": 0.21227741330834116
+    },
+    {
+      "subset_type": "accent",
+      "subset": "americacentral",
+      "cer": 0.05467035588207232,
+      "wer": 0.18071593533487298
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.06017050441256246,
+      "wer": 0.20354917778262277
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.06061355444149818,
+      "wer": 0.20427529484792054
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.05583951423849967,
+      "wer": 0.1967816091954023
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.0575204041974349,
+      "wer": 0.1914648212226067
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.14137214137214138,
+      "wer": 0.34146341463414637
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.0551695954229669,
+      "wer": 0.18813080339119903
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.05877518292071765,
+      "wer": 0.20242380609551236
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.10067114093959731,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.06524229965626474,
+      "wer": 0.2254782254782255
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.0519164209716703,
+      "wer": 0.1842985842985843
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.06067393656140097,
+      "wer": 0.20448472173812013
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.05639573649778225,
+      "wer": 0.19706256505146294
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.1054519368723099,
+      "wer": 0.3130434782608696
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.05798633129007257,
+      "wer": 0.19640591966173362
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.060862146873124946,
+      "wer": 0.20509822782936846
+    },
+    {
+      "subset_type": "accent",
+      "subset": "chileno",
+      "cer": 0.04395434916718075,
+      "wer": 0.16193961573650503
+    },
+    {
+      "subset_type": "accent",
+      "subset": "nortepeninsular",
+      "cer": 0.051128749613441916,
+      "wer": 0.1742562234365513
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canario",
+      "cer": 0.060699588477366256,
+      "wer": 0.21656050955414013
+    },
+    {
+      "subset_type": "accent",
+      "subset": "surpeninsular",
+      "cer": 0.05215475024485798,
+      "wer": 0.20444444444444446
+    },
+    {
+      "subset_type": "accent",
+      "subset": "caribe",
+      "cer": 0.06730909621842605,
+      "wer": 0.21984602224123181
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mexicano",
+      "cer": 0.05488584474885845,
+      "wer": 0.19250139899272523
+    },
+    {
+      "subset_type": "accent",
+      "subset": "rioplatense",
+      "cer": 0.06893939393939394,
+      "wer": 0.2244593586875466
+    },
+    {
+      "subset_type": "accent",
+      "subset": "andino",
+      "cer": 0.049281314168377825,
+      "wer": 0.1814233210825259
+    },
+    {
+      "subset_type": "accent",
+      "subset": "centrosurpeninsular",
+      "cer": 0.04169837163293258,
+      "wer": 0.16160714285714287
+    },
+    {
+      "subset_type": "accent",
+      "subset": "americacentral",
+      "cer": 0.051120723554856466,
+      "wer": 0.1786833855799373
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a475664524ac433e5cc6682acc093f12147d2efa5df6a0ffc12dfe65a286a606
+size 125311689

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 sha256:002ac8218a53d046aa9c5fcb9eb36637eb6a3b4c44156bf42ef0c6b12162458f
+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/tmplxbpbu_b", "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}