jonatasgrosman/exp_w2v2r_en_xls-r_gender_male-10_female-0_s287

.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:
+- en
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- en
+datasets:
+- mozilla-foundation/common_voice_7_0
+---
+# exp_w2v2r_en_xls-r_gender_male-10_female-0_s287
+
+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 (en)](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": 33,
+  "xvector_output_dim": 512
+}

cv7_evaluation.json

@@ -0,0 +1,390 @@
+{
+  "validation": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.10189106096641337,
+      "wer": 0.29067266816704174
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.10056559649327608,
+      "wer": 0.2879053625405829
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.08635666836995401,
+      "wer": 0.23976023976023977
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.0893760539629005,
+      "wer": 0.25598086124401914
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.0950189155107188,
+      "wer": 0.27912087912087913
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.10508090411131399,
+      "wer": 0.2955583988301956
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.07771181867242309,
+      "wer": 0.25059288537549407
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.10758466842425471,
+      "wer": 0.3009994121105232
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.11990009549695144,
+      "wer": 0.32977900077392724
+    },
+    {
+      "subset_type": "age",
+      "subset": "nineties",
+      "cer": 0.038461538461538464,
+      "wer": 0.14285714285714285
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.11827956989247312,
+      "wer": 0.34615384615384615
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.10063755548143777,
+      "wer": 0.2880609602280784
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.10790631876768311,
+      "wer": 0.3021500457456542
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.1085830051010525,
+      "wer": 0.30463202445332704
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.09940652818991098,
+      "wer": 0.2994011976047904
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.10267686286403385,
+      "wer": 0.29226956759140643
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.11517813567593949,
+      "wer": 0.3294460641399417
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.07407407407407407,
+      "wer": 0.125
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.0711743772241993,
+      "wer": 0.22727272727272727
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.11794871794871795,
+      "wer": 0.3194444444444444
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.06779661016949153,
+      "wer": 0.2708333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.07270875763747454,
+      "wer": 0.22541966426858512
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.08500626747661749,
+      "wer": 0.2519170049616599
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.10681399631675875,
+      "wer": 0.25
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.054187192118226604,
+      "wer": 0.15
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.10408432147562582,
+      "wer": 0.33070866141732286
+    },
+    {
+      "subset_type": "accent",
+      "subset": "southatlandtic",
+      "cer": 0.12637362637362637,
+      "wer": 0.30303030303030304
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.09851301115241635,
+      "wer": 0.3
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.12844792095512556,
+      "wer": 0.34636199639206255
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.0912057712126417,
+      "wer": 0.2830188679245283
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.06501315295001879,
+      "wer": 0.23076923076923078
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.14874141876430205,
+      "wer": 0.3181818181818182
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.13087447947650208,
+      "wer": 0.3460207612456747
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.1303661965272159,
+      "wer": 0.3324069835872622
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.1292713821974886,
+      "wer": 0.32985448155726793
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.10367691380349608,
+      "wer": 0.28523489932885904
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.13873295910184444,
+      "wer": 0.2753303964757709
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.13050460654213508,
+      "wer": 0.3341165413533835
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.12503226916788573,
+      "wer": 0.3336585365853659
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.18073742246726396,
+      "wer": 0.37123420796890183
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.14236338716405458,
+      "wer": 0.3631722260040844
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.13798411122144985,
+      "wer": 0.3546727636467562
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.12,
+      "wer": 0.44
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.12919229709723076,
+      "wer": 0.3296936424535977
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.1338792114321477,
+      "wer": 0.34374070770145704
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.13748262365548994,
+      "wer": 0.3496656739899983
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.18295218295218296,
+      "wer": 0.377431906614786
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.13077961936691115,
+      "wer": 0.3329422941286543
+    },
+    {
+      "subset_type": "accent",
+      "subset": "african",
+      "cer": 0.1765295887662989,
+      "wer": 0.3930635838150289
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bermuda",
+      "cer": 0.15056179775280898,
+      "wer": 0.36904761904761907
+    },
+    {
+      "subset_type": "accent",
+      "subset": "scotland",
+      "cer": 0.12622720897615708,
+      "wer": 0.3191489361702128
+    },
+    {
+      "subset_type": "accent",
+      "subset": "malaysia",
+      "cer": 0.12206047032474804,
+      "wer": 0.34394904458598724
+    },
+    {
+      "subset_type": "accent",
+      "subset": "wales",
+      "cer": 0.19863013698630136,
+      "wer": 0.5
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.1643495531281033,
+      "wer": 0.324438202247191
+    },
+    {
+      "subset_type": "accent",
+      "subset": "us",
+      "cer": 0.11602120294244915,
+      "wer": 0.30460033623720006
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hongkong",
+      "cer": 0.2153179190751445,
+      "wer": 0.5333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "newzealand",
+      "cer": 0.08221024258760108,
+      "wer": 0.3161764705882353
+    },
+    {
+      "subset_type": "accent",
+      "subset": "ireland",
+      "cer": 0.10740203193033382,
+      "wer": 0.2911392405063291
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.19110764430577223,
+      "wer": 0.3938223938223938
+    },
+    {
+      "subset_type": "accent",
+      "subset": "indian",
+      "cer": 0.13791208791208792,
+      "wer": 0.3609802073515551
+    },
+    {
+      "subset_type": "accent",
+      "subset": "england",
+      "cer": 0.11592054550844945,
+      "wer": 0.3233465234595817
+    },
+    {
+      "subset_type": "accent",
+      "subset": "australia",
+      "cer": 0.10438972162740899,
+      "wer": 0.2929936305732484
+    },
+    {
+      "subset_type": "accent",
+      "subset": "singapore",
+      "cer": 0.15508021390374332,
+      "wer": 0.42424242424242425
+    },
+    {
+      "subset_type": "accent",
+      "subset": "philippines",
+      "cer": 0.13043478260869565,
+      "wer": 0.35353535353535354
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:91b7dc3ef806c8896f46c814bc5e82481ce9f2115d7cea35ac0556d10b1c1209
+size 123268039

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 sha256:2f41adba854dda63b32c8ea3ee9c7bda607f43586f85ce5a82ec318671343d9e
+size 1262034033

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/tmp1ksa008g", "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}