commit 068d7ac5c141f68cf32ae649bde5cd9455fd8b26 Author: ModelHub XC Date: Fri May 8 11:39:09 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: TalTechNLP/xls-r-300m-et Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..6d34772 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,27 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bin.* filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.xz 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 diff --git a/README.md b/README.md new file mode 100644 index 0000000..45ec6ca --- /dev/null +++ b/README.md @@ -0,0 +1,93 @@ +--- +license: cc-by-4.0 +tags: +- audio +- automatic-speech-recognition +- hf-asr-leaderboard +language: et +model-index: +- name: xls-r-300m-et + results: + - task: + name: Automatic Speech Recognition + type: automatic-speech-recognition + dataset: + name: Common Voice + type: common_voice + args: et + metrics: + - name: Test WER + type: wer + value: 12.520395591222402 + - name: Test CER + type: cer + value: 2.7091152438624897 + - task: + name: Automatic Speech Recognition + type: automatic-speech-recognition + dataset: + name: Common Voice 8 + type: mozilla-foundation/common_voice_8_0 + args: et + metrics: + - name: Test WER + type: wer + value: 13.38447882323104 + - name: Test CER + type: cer + value: 2.9816686199500255 +--- + + +# XLS-R-300m-ET + +This is a XLS-R-300M model [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) finetuned on around 800 hours of diverse Estonian data. + +## Model description +This is a general-purpose Estonian ASR model trained in the Lab of Language Technology at TalTech. It consists of only the CTC-based end-to-end model, no language model is currently provided. + +## Intended uses & limitations + +This model is intended for general-purpose speech recognition, such as broadcast conversations, interviews, talks, etc. + +## How to use + + +TODO + +#### Limitations and bias + +Since this model was trained on mostly broadcast speech and texts from the web, it might have problems correctly decoding the following: + * Speech containing technical and other domain-specific terms + * Children's speech + * Non-native speech + * Speech recorded under very noisy conditions or with a microphone far from the speaker + * Very spontaneous and overlapping speech + +## Training data +Acoustic training data: + +| Type | Amount (h) | +|-----------------------|:------:| +| Broadcast speech | 591 | +| Spontaneous speech | 53 | +| Elderly speech corpus | 53 | +| Talks, lectures | 49 | +| Parliament speeches | 31 | +| *Total* | *761* | + + +## Training procedure + +Finetuned using Fairseq. + +## Evaluation results + +### WER + +|Dataset | WER | +|---|---| +| jutusaated.devset | 7.9 | +| jutusaated.testset | 6.1 | +| Common Voice 6.1 | 12.5 | +| Common Voice 8.0 | 13.4 | diff --git a/config.json b/config.json new file mode 100644 index 0000000..038a8ee --- /dev/null +++ b/config.json @@ -0,0 +1,107 @@ +{ + "_name_or_path": "facebook/wav2vec2-xls-r-300m", + "activation_dropout": 0.1, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForCTC" + ], + "attention_dropout": 0.0, + "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": 0.0, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "hidden_act": "gelu", + "hidden_dropout": 0.0, + "hidden_size": 1024, + "initializer_range": 0.02, + "intermediate_size": 4096, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_feature_length": 64, + "mask_feature_min_masks": 0, + "mask_feature_prob": 0.25, + "mask_time_length": 10, + "mask_time_min_masks": 2, + "mask_time_prob": 0.75, + "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": "4.15.0", + "use_weighted_layer_sum": false, + "vocab_size": 197, + "xvector_output_dim": 512 +} diff --git a/preprocessor_config.json b/preprocessor_config.json new file mode 100644 index 0000000..36ebe8b --- /dev/null +++ b/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0, + "return_attention_mask": true, + "sampling_rate": 16000 +} diff --git a/pytorch_model.bin b/pytorch_model.bin new file mode 100644 index 0000000..eaa0fd5 --- /dev/null +++ b/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3ebb8bbc9f56a42858c9b69ed4b7a1b308ae623d7f8f1fc38788965dc638b106 +size 1262725489 diff --git a/special_tokens_map.json b/special_tokens_map.json new file mode 100644 index 0000000..25bc396 --- /dev/null +++ b/special_tokens_map.json @@ -0,0 +1 @@ +{"bos_token": "", "eos_token": "", "unk_token": "", "pad_token": ""} \ No newline at end of file diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..94d9657 --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1 @@ +{"unk_token": "", "bos_token": "", "eos_token": "", "pad_token": "", "do_lower_case": false, "word_delimiter_token": "|", "tokenizer_class": "Wav2Vec2CTCTokenizer"} \ No newline at end of file diff --git a/vocab.json b/vocab.json new file mode 100644 index 0000000..307c436 --- /dev/null +++ b/vocab.json @@ -0,0 +1 @@ +{"": 1, "": 0, "": 2, "": 3, "|": 4, "e": 5, "a": 6, "i": 7, "s": 8, "t": 9, "l": 10, "n": 11, "u": 12, "k": 13, "o": 14, "m": 15, "d": 16, "r": 17, "v": 18, "g": 19, "j": 20, "h": 21, "p": 22, "ä": 23, "õ": 24, "ü": 25, "b": 26, "ö": 27, "-": 28, "E": 29, "f": 30, "M": 31, "T": 32, "S": 33, "K": 34, "A": 35, "L": 36, "R": 37, "V": 38, "P": 39, "N": 40, "I": 41, "H": 42, "J": 43, "c": 44, "O": 45, "B": 46, "U": 47, "y": 48, "C": 49, "G": 50, "š": 51, "Ü": 52, "D": 53, "'": 54, "w": 55, "F": 56, "ž": 57, "z": 58, "W": 59, "x": 60, "Y": 61, "Õ": 62, ".": 63, "Z": 64, "Ö": 65, "Š": 66, "Q": 67, "X": 68, "q": 69, "Ä": 70, "3": 71, "1": 72, "„": 73, "0": 74, "4": 75, "Ž": 76, "9": 77, "5": 78, "\"": 79, "2": 80, "O2": 81, "é": 82, "č": 83, "@": 84, "8": 85, "6": 86, "о": 87, "e2": 88, "а": 89, "l2": 90, "т": 91, "и": 92, "12": 93, "ˇ": 94, "л": 95, "н": 96, "ú": 97, "ē": 98, "V2": 99, "я": 100, "B2": 101, "с": 102, "S2": 103, "ç": 104, "ā": 105, "ć": 106, "ц": 107, "47": 108, "´": 109, "á": 110, "æ": 111, "ë": 112, "í": 113, "ø": 114, "ė": 115, "г": 116, "д": 117, "з": 118, "к": 119, "7": 120, "87": 121, "97": 122, "É": 123, "à": 124, "ð": 125, "ğ": 126, "́": 127, "К": 128, "е": 129, "р": 130, "у": 131, "+": 132, ",": 133, "02": 134, "37": 135, "U2": 136, "e´": 137, "g2": 138, "n´": 139, "è": 140, "ą": 141, "ī": 142, "ņ": 143, "ū": 144, "С": 145, "й": 146, "ф": 147, "ш": 148, "ы": 149, "э": 150, "/": 151, "07": 152, "72": 153, ">": 154, "A2": 155, "C2": 156, "F2": 157, "G7": 158, "K2": 159, "P2": 160, "a2": 161, "o2": 162, "r2": 163, "s2": 164, "y2": 165, "y´": 166, "§": 167, "Å": 168, "Ø": 169, "ß": 170, "å": 171, "ó": 172, "ô": 173, "þ": 174, "Ā": 175, "ķ": 176, "ł": 177, "Ś": 178, "ś": 179, "А": 180, "В": 181, "И": 182, "М": 183, "Р": 184, "У": 185, "б": 186, "в": 187, "м": 188, "п": 189, "х": 190, "ч": 191, "ь": 192, "ю": 193, "–": 194, "”": 195, "": 196} \ No newline at end of file