jonatasgrosman/exp_w2v2r_es_xls-r_gender_male-8_female-2_s471

.gitattributes

@@ -29,3 +29,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zstandard 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_xls-r_gender_male-8_female-2_s471
+
+Fine-tuned [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) 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 +1,108 @@
+{
+  "_name_or_path": "facebook/wav2vec2-xls-r-300m",
+  "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_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.05,
+  "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": 0.03690563418302418,
+      "wer": 0.13646866211595082
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.03693384728756873,
+      "wer": 0.136176253091824
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.046969247503213685,
+      "wer": 0.14719626168224298
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.025806451612903226,
+      "wer": 0.12735849056603774
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.029193503220386448,
+      "wer": 0.11458551941238194
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.03595734775765368,
+      "wer": 0.13677811550151975
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.038871473354231974,
+      "wer": 0.1535365152386429
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.033775633293124246,
+      "wer": 0.12479935794542536
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.03888847867286913,
+      "wer": 0.14305667867382196
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.036870686729268014,
+      "wer": 0.135988533335595
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.033913733544953785,
+      "wer": 0.12956027188236927
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.03758691975192633,
+      "wer": 0.13933124103994732
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.051328502415458936,
+      "wer": 0.19529837251356238
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.03686238106289162,
+      "wer": 0.13610704744943392
+    },
+    {
+      "subset_type": "accent",
+      "subset": "rioplatense",
+      "cer": 0.037995725480883404,
+      "wer": 0.13896848137535817
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mexicano",
+      "cer": 0.03608283122099869,
+      "wer": 0.13434692467173462
+    },
+    {
+      "subset_type": "accent",
+      "subset": "americacentral",
+      "cer": 0.04083579811086729,
+      "wer": 0.14491916859122403
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canario",
+      "cer": 0.032281731474688186,
+      "wer": 0.13302752293577982
+    },
+    {
+      "subset_type": "accent",
+      "subset": "andino",
+      "cer": 0.03429837292796578,
+      "wer": 0.12872413008869685
+    },
+    {
+      "subset_type": "accent",
+      "subset": "nortepeninsular",
+      "cer": 0.031941266209000764,
+      "wer": 0.11988795518207283
+    },
+    {
+      "subset_type": "accent",
+      "subset": "surpeninsular",
+      "cer": 0.04012616085509024,
+      "wer": 0.148380355276907
+    },
+    {
+      "subset_type": "accent",
+      "subset": "centrosurpeninsular",
+      "cer": 0.04435903595163744,
+      "wer": 0.15932521087160262
+    },
+    {
+      "subset_type": "accent",
+      "subset": "chileno",
+      "cer": 0.04279253160153079,
+      "wer": 0.1644962302947224
+    },
+    {
+      "subset_type": "accent",
+      "subset": "caribe",
+      "cer": 0.03541296625222025,
+      "wer": 0.13732041430003342
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.0421167068744239,
+      "wer": 0.14807180279996565
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.042423661750082865,
+      "wer": 0.1487352576039727
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.041002720559657986,
+      "wer": 0.14763552479815456
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.11226611226611227,
+      "wer": 0.3048780487804878
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.035294620347818655,
+      "wer": 0.12612612612612611
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.03739287498558856,
+      "wer": 0.13586206896551725
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.039912818417109384,
+      "wer": 0.1324182478805006
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.04690975264541349,
+      "wer": 0.17175417175417176
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.04173599278340182,
+      "wer": 0.15082793376529877
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.026845637583892617,
+      "wer": 0.07407407407407407
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.04245611631486908,
+      "wer": 0.14886332829867163
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.03811738180793349,
+      "wer": 0.13805496828752642
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.040364117777799254,
+      "wer": 0.1438071007285764
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.06527977044476327,
+      "wer": 0.23043478260869565
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.0426085646770653,
+      "wer": 0.1493802952384476
+    },
+    {
+      "subset_type": "accent",
+      "subset": "rioplatense",
+      "cer": 0.05138888888888889,
+      "wer": 0.15883668903803133
+    },
+    {
+      "subset_type": "accent",
+      "subset": "mexicano",
+      "cer": 0.03721461187214612,
+      "wer": 0.134023503077784
+    },
+    {
+      "subset_type": "accent",
+      "subset": "americacentral",
+      "cer": 0.03893039716869839,
+      "wer": 0.13479623824451412
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canario",
+      "cer": 0.03909465020576132,
+      "wer": 0.14012738853503184
+    },
+    {
+      "subset_type": "accent",
+      "subset": "andino",
+      "cer": 0.035129585437593654,
+      "wer": 0.1399933177413966
+    },
+    {
+      "subset_type": "accent",
+      "subset": "nortepeninsular",
+      "cer": 0.03587259045459231,
+      "wer": 0.13600485731633272
+    },
+    {
+      "subset_type": "accent",
+      "subset": "surpeninsular",
+      "cer": 0.03599412340842312,
+      "wer": 0.1348148148148148
+    },
+    {
+      "subset_type": "accent",
+      "subset": "centrosurpeninsular",
+      "cer": 0.029523664586820878,
+      "wer": 0.11071428571428571
+    },
+    {
+      "subset_type": "accent",
+      "subset": "chileno",
+      "cer": 0.03392967304133251,
+      "wer": 0.12259835315645014
+    },
+    {
+      "subset_type": "accent",
+      "subset": "caribe",
+      "cer": 0.04285297123667689,
+      "wer": 0.14798973481608213
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf4765d5c8c5c55150a883a5db92ecdcb855300bfadca19df45efbeac5d4aa5f
+size 125560874

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 sha256:6955b8350c35031567b430280788de1e5c4e416eb77d64e4297415089657b416
+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/tmpv5tbw9fs", "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}