jonatasgrosman/exp_w2v2r_de_xls-r_age_teens-2_sixties-8_s289

.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:
+- de
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- de
+datasets:
+- mozilla-foundation/common_voice_7_0
+---
+# exp_w2v2r_de_xls-r_age_teens-2_sixties-8_s289
+
+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 (de)](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": 38,
+  "xvector_output_dim": 512
+}

cv7_evaluation.json

@@ -0,0 +1,396 @@
+{
+  "validation": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.046023709302817434,
+      "wer": 0.18849217136022617
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.045290565257231366,
+      "wer": 0.18631099566234566
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.049948125092633765,
+      "wer": 0.19968635650810246
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.06044277360066834,
+      "wer": 0.2229137199434229
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.04911519075825025,
+      "wer": 0.2051667920742413
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.05196927602549436,
+      "wer": 0.21064384071455156
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.048939641109298535,
+      "wer": 0.18292682926829268
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.046894181193223666,
+      "wer": 0.20068027210884354
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.044074259909683895,
+      "wer": 0.17639802631578946
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.047309604686653864,
+      "wer": 0.19116655428186108
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.04529253457817156,
+      "wer": 0.18632919823495128
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.08029878618113913,
+      "wer": 0.31097560975609756
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.048923268585375775,
+      "wer": 0.1970894874022589
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.05035754610462928,
+      "wer": 0.1993260756868844
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.04597873418192126,
+      "wer": 0.1884255948315347
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.08716707021791767,
+      "wer": 0.2892561983471074
+    },
+    {
+      "subset_type": "accent",
+      "subset": "russia",
+      "cer": 0.06911262798634812,
+      "wer": 0.26737967914438504
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.07566999474513926,
+      "wer": 0.29328621908127206
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.055527826303968054,
+      "wer": 0.231433506044905
+    },
+    {
+      "subset_type": "accent",
+      "subset": "czechia",
+      "cer": 0.04195205479452055,
+      "wer": 0.18787878787878787
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.10751295336787564,
+      "wer": 0.39316239316239315
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.043243243243243246,
+      "wer": 0.19230769230769232
+    },
+    {
+      "subset_type": "accent",
+      "subset": "turkey",
+      "cer": 0.10810810810810811,
+      "wer": 0.3898305084745763
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.06826347305389222,
+      "wer": 0.256
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.04366934133073843,
+      "wer": 0.17907257180442884
+    },
+    {
+      "subset_type": "accent",
+      "subset": "netherlands",
+      "cer": 0.050505050505050504,
+      "wer": 0.1956521739130435
+    },
+    {
+      "subset_type": "accent",
+      "subset": "slovakia",
+      "cer": 0.075,
+      "wer": 0.2631578947368421
+    },
+    {
+      "subset_type": "accent",
+      "subset": "poland",
+      "cer": 0.03902439024390244,
+      "wer": 0.15151515151515152
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.07352941176470588,
+      "wer": 0.325
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.057432432432432436,
+      "wer": 0.22916666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "austria",
+      "cer": 0.04849646226415094,
+      "wer": 0.19796954314720813
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.05241010444002677,
+      "wer": 0.2032061320365944
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.0515633295707536,
+      "wer": 0.20088146246119648
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.07559269335406141,
+      "wer": 0.2561307901907357
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.07966115500529446,
+      "wer": 0.2703601108033241
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.056875656074052865,
+      "wer": 0.22056599697558868
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.06587354042740692,
+      "wer": 0.2374749498997996
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.060550458715596334,
+      "wer": 0.2054794520547945
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.06269821907782386,
+      "wer": 0.23181049069373943
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.05486830693385676,
+      "wer": 0.20437956204379562
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.04388669774112585,
+      "wer": 0.1900744416873449
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.05167061751411044,
+      "wer": 0.20116615841978638
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.1362876254180602,
+      "wer": 0.2897727272727273
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.05500623867033924,
+      "wer": 0.20945837993289662
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.0731367681782406,
+      "wer": 0.27554438860971525
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.05224197540695321,
+      "wer": 0.20303025697916507
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hungary",
+      "cer": 0.09230769230769231,
+      "wer": 0.39285714285714285
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bulgaria",
+      "cer": 0.0,
+      "wer": 0.0
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.0743651753325272,
+      "wer": 0.2697095435684647
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.08121827411167512,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.10815047021943573,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "denmark",
+      "cer": 0.04411764705882353,
+      "wer": 0.2222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "austria",
+      "cer": 0.05748519116855143,
+      "wer": 0.21975992613111728
+    },
+    {
+      "subset_type": "accent",
+      "subset": "greece",
+      "cer": 0.12179487179487179,
+      "wer": 0.2222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "czechia",
+      "cer": 0.14905149051490515,
+      "wer": 0.4528301886792453
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.04963787784500907,
+      "wer": 0.19244636740580495
+    },
+    {
+      "subset_type": "accent",
+      "subset": "netherlands",
+      "cer": 0.06656580937972768,
+      "wer": 0.26424870466321243
+    },
+    {
+      "subset_type": "accent",
+      "subset": "slovakia",
+      "cer": 0.047979797979797977,
+      "wer": 0.1896551724137931
+    },
+    {
+      "subset_type": "accent",
+      "subset": "poland",
+      "cer": 0.09364548494983277,
+      "wer": 0.31521739130434784
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.045774647887323945,
+      "wer": 0.23684210526315788
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.07953443258971872,
+      "wer": 0.29333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "turkey",
+      "cer": 0.21505376344086022,
+      "wer": 0.5333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "russia",
+      "cer": 0.16835016835016836,
+      "wer": 0.37777777777777777
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.0692104729044043,
+      "wer": 0.24089635854341737
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.021739130434782608,
+      "wer": 0.21428571428571427
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35ef8aa28fbf946a09a1a9af46d4461b3538f5417b6f92b19acdc7048c6c738d
+size 138730257

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:ffd00ece9ff90a1bdb88e4c791e6706d594d0e0c0085b2af6e69110129b5a216
+size 1262054513

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/tmp95ix38n8", "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, "ß": 33, "á": 34, "ä": 35, "ö": 36, "ü": 37}