jonatasgrosman/exp_w2v2r_de_vp-100k_age_teens-0_sixties-10_s278

.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_vp-100k_age_teens-0_sixties-10_s278
+
+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 (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 +1,116 @@
+{
+  "_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": 38,
+  "xvector_output_dim": 512
+}

cv7_evaluation.json

@@ -0,0 +1,396 @@
+{
+  "validation": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.06742130099167488,
+      "wer": 0.2605265303903447
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.06634671579367618,
+      "wer": 0.257145473770096
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.06635223281485199,
+      "wer": 0.25082236842105265
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.07129094412331406,
+      "wer": 0.276006272869838
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.06776331942057451,
+      "wer": 0.27721088435374147
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.06961059420075813,
+      "wer": 0.2663519892110587
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.07271623560575402,
+      "wer": 0.2840481565086531
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.0636215334420881,
+      "wer": 0.24390243902439024
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.07430408018739446,
+      "wer": 0.28321548195013024
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.08567251461988304,
+      "wer": 0.3094766619519095
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.06635088757254141,
+      "wer": 0.25713928898509697
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.07239016576833113,
+      "wer": 0.27654213727193744
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.1111111111111111,
+      "wer": 0.43902439024390244
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.06974030861874295,
+      "wer": 0.26283048211508553
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.06691924472885617,
+      "wer": 0.25890539747844465
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.11139896373056994,
+      "wer": 0.4188034188034188
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.037162162162162164,
+      "wer": 0.22916666666666666
+    },
+    {
+      "subset_type": "accent",
+      "subset": "poland",
+      "cer": 0.04390243902439024,
+      "wer": 0.18181818181818182
+    },
+    {
+      "subset_type": "accent",
+      "subset": "turkey",
+      "cer": 0.14496314496314497,
+      "wer": 0.4915254237288136
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.06825976677780082,
+      "wer": 0.2622780092085069
+    },
+    {
+      "subset_type": "accent",
+      "subset": "austria",
+      "cer": 0.06972287735849056,
+      "wer": 0.26649746192893403
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.043243243243243246,
+      "wer": 0.23076923076923078
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.09927360774818401,
+      "wer": 0.32231404958677684
+    },
+    {
+      "subset_type": "accent",
+      "subset": "russia",
+      "cer": 0.09897610921501707,
+      "wer": 0.3422459893048128
+    },
+    {
+      "subset_type": "accent",
+      "subset": "czechia",
+      "cer": 0.0547945205479452,
+      "wer": 0.22424242424242424
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.09580838323353294,
+      "wer": 0.32
+    },
+    {
+      "subset_type": "accent",
+      "subset": "slovakia",
+      "cer": 0.06111111111111111,
+      "wer": 0.2631578947368421
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.10982658959537572,
+      "wer": 0.4098939929328622
+    },
+    {
+      "subset_type": "accent",
+      "subset": "netherlands",
+      "cer": 0.07744107744107744,
+      "wer": 0.21739130434782608
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.07848764661841777,
+      "wer": 0.3126079447322971
+    },
+    {
+      "subset_type": "accent",
+      "subset": "canada",
+      "cer": 0.10661764705882353,
+      "wer": 0.425
+    }
+  ],
+  "test": [
+    {
+      "subset_type": "all",
+      "subset": "all",
+      "cer": 0.07679478470187219,
+      "wer": 0.2818758757108712
+    },
+    {
+      "subset_type": "age",
+      "subset": "",
+      "cer": 0.0758498150684626,
+      "wer": 0.2796152224734603
+    },
+    {
+      "subset_type": "age",
+      "subset": "thirties",
+      "cer": 0.07463367873282513,
+      "wer": 0.2688564476885645
+    },
+    {
+      "subset_type": "age",
+      "subset": "seventies",
+      "cer": 0.097096852890949,
+      "wer": 0.3197969543147208
+    },
+    {
+      "subset_type": "age",
+      "subset": "sixties",
+      "cer": 0.10182666148464826,
+      "wer": 0.33923705722070846
+    },
+    {
+      "subset_type": "age",
+      "subset": "fifties",
+      "cer": 0.0666905700968089,
+      "wer": 0.26104218362282877
+    },
+    {
+      "subset_type": "age",
+      "subset": "twenties",
+      "cer": 0.0811146101727264,
+      "wer": 0.30157701447396845
+    },
+    {
+      "subset_type": "age",
+      "subset": "eighties",
+      "cer": 0.1706422018348624,
+      "wer": 0.4931506849315068
+    },
+    {
+      "subset_type": "age",
+      "subset": "teens",
+      "cer": 0.09201733127708012,
+      "wer": 0.3251503006012024
+    },
+    {
+      "subset_type": "age",
+      "subset": "fourties",
+      "cer": 0.11175368575384866,
+      "wer": 0.34958448753462606
+    },
+    {
+      "subset_type": "gender",
+      "subset": "",
+      "cer": 0.07597557684974365,
+      "wer": 0.2798550347089201
+    },
+    {
+      "subset_type": "gender",
+      "subset": "male",
+      "cer": 0.08009040186453846,
+      "wer": 0.28926346061671193
+    },
+    {
+      "subset_type": "gender",
+      "subset": "other",
+      "cer": 0.11872909698996656,
+      "wer": 0.3068181818181818
+    },
+    {
+      "subset_type": "gender",
+      "subset": "female",
+      "cer": 0.1012895804322819,
+      "wer": 0.35175879396984927
+    },
+    {
+      "subset_type": "accent",
+      "subset": "",
+      "cer": 0.07672884606451527,
+      "wer": 0.2818792456270991
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_kingdom",
+      "cer": 0.14576802507836992,
+      "wer": 0.49019607843137253
+    },
+    {
+      "subset_type": "accent",
+      "subset": "russia",
+      "cer": 0.26936026936026936,
+      "wer": 0.6888888888888889
+    },
+    {
+      "subset_type": "accent",
+      "subset": "bulgaria",
+      "cer": 0.0,
+      "wer": 0.0
+    },
+    {
+      "subset_type": "accent",
+      "subset": "czechia",
+      "cer": 0.21951219512195122,
+      "wer": 0.7547169811320755
+    },
+    {
+      "subset_type": "accent",
+      "subset": "other",
+      "cer": 0.021739130434782608,
+      "wer": 0.07142857142857142
+    },
+    {
+      "subset_type": "accent",
+      "subset": "slovakia",
+      "cer": 0.05303030303030303,
+      "wer": 0.25862068965517243
+    },
+    {
+      "subset_type": "accent",
+      "subset": "united_states",
+      "cer": 0.07376058041112454,
+      "wer": 0.26556016597510373
+    },
+    {
+      "subset_type": "accent",
+      "subset": "switzerland",
+      "cer": 0.0909275421148772,
+      "wer": 0.3487394957983193
+    },
+    {
+      "subset_type": "accent",
+      "subset": "turkey",
+      "cer": 0.21505376344086022,
+      "wer": 0.6
+    },
+    {
+      "subset_type": "accent",
+      "subset": "poland",
+      "cer": 0.13879598662207357,
+      "wer": 0.43478260869565216
+    },
+    {
+      "subset_type": "accent",
+      "subset": "austria",
+      "cer": 0.08104469574582661,
+      "wer": 0.2871652816251154
+    },
+    {
+      "subset_type": "accent",
+      "subset": "belgium",
+      "cer": 0.06598984771573604,
+      "wer": 0.3333333333333333
+    },
+    {
+      "subset_type": "accent",
+      "subset": "luxembourg",
+      "cer": 0.04929577464788732,
+      "wer": 0.2631578947368421
+    },
+    {
+      "subset_type": "accent",
+      "subset": "hungary",
+      "cer": 0.11282051282051282,
+      "wer": 0.42857142857142855
+    },
+    {
+      "subset_type": "accent",
+      "subset": "france",
+      "cer": 0.0979631425800194,
+      "wer": 0.37333333333333335
+    },
+    {
+      "subset_type": "accent",
+      "subset": "germany",
+      "cer": 0.07328505037933454,
+      "wer": 0.2677122769064359
+    },
+    {
+      "subset_type": "accent",
+      "subset": "denmark",
+      "cer": 0.058823529411764705,
+      "wer": 0.2222222222222222
+    },
+    {
+      "subset_type": "accent",
+      "subset": "greece",
+      "cer": 0.09615384615384616,
+      "wer": 0.25925925925925924
+    },
+    {
+      "subset_type": "accent",
+      "subset": "netherlands",
+      "cer": 0.09606656580937972,
+      "wer": 0.35751295336787564
+    }
+  ]
+}

cv7_transcriptions.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98a734dca949edce8fcfc671ed7204aeea83edae8ffafc95ab31d5d45d8745ae
+size 137502576

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 sha256:8ae6e05cb9f5f9338268776c0e589f2a80ac72c7396b0a55c37c9c63c8ac634a
+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/tmpsedmsc3k", "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}