Copying full model from Akashpb13 including all logs and scripts

README.md

@@ -0,0 +1,137 @@
+---
+language:
+- ckb
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- mozilla-foundation/common_voice_8_0
+- generated_from_trainer
+- ckb
+- robust-speech-event
+- model_for_talk
+- hf-asr-leaderboard
+datasets:
+- mozilla-foundation/common_voice_8_0
+model-index:
+- name: Akashpb13/Central_kurdish_xlsr
+  results:
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: Common Voice 8
+      type: mozilla-foundation/common_voice_8_0
+      args: ckb
+    metrics:
+    - name: Test WER
+      type: wer
+      value: 0.36754389884276845
+    - name: Test CER
+      type: cer
+      value: 0.07827896768334217
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: Robust Speech Event - Dev Data
+      type: speech-recognition-community-v2/dev_data
+      args: ckb
+    metrics:
+    - name: Test WER
+      type: wer
+      value: 0.36754389884276845
+    - name: Test CER
+      type: cer
+      value: 0.07827896768334217
+---
+
+# Akashpb13/Central_kurdish_xlsr
+
+This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - hu dataset.
+It achieves the following results on evaluation set (which is 10 percent of train data set merged with invalidated data, reported, other and dev datasets):
+- Loss: 0.348580	
+- Wer: 0.401147
+
+## Model description
+"facebook/wav2vec2-xls-r-300m" was finetuned.
+
+## Intended uses & limitations
+More information needed
+## Training and evaluation data
+Training data - 
+Common voice Central Kurdish train.tsv, dev.tsv, invalidated.tsv, reported.tsv, and other.tsv 
+Only those points were considered where upvotes were greater than downvotes and duplicates were removed after concatenation of all the datasets given in common voice 7.0
+
+## Training procedure
+For creating the train dataset, all possible datasets were appended and 90-10 split was used. 
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+
+- learning_rate: 0.000095637994662983496
+- train_batch_size: 16
+- eval_batch_size: 16
+- seed: 13
+- gradient_accumulation_steps: 2
+- lr_scheduler_type: cosine_with_restarts
+- lr_scheduler_warmup_steps: 200
+- num_epochs: 100
+- mixed_precision_training: Native AMP
+
+
+### Training results
+
+| Step  | Training Loss | Validation Loss | Wer      |
+|-------|---------------|-----------------|----------|
+| 500   | 5.097800      | 2.190326        | 1.001207 |
+| 1000  | 0.797500      | 0.331392        | 0.576819 |
+| 1500  | 0.405100      | 0.262009        | 0.549049 |
+| 2000  | 0.322100      | 0.248178        | 0.479626 |
+| 2500  | 0.264600      | 0.258866        | 0.488983 |
+| 3000  | 0.228300      | 0.261523        | 0.469665 |
+| 3500  | 0.201000      | 0.270135        | 0.451856 |
+| 4000  | 0.180900      | 0.279302        | 0.448536 |
+| 4500  | 0.163800      | 0.280921        | 0.459704 |
+| 5000  | 0.147300      | 0.319249        | 0.471778 |
+| 5500  | 0.137600      | 0.289546        | 0.449140 |
+| 6000  | 0.132000      | 0.311350        | 0.458195 |
+| 6500  | 0.117100      | 0.316726        | 0.432840 |
+| 7000  | 0.109200      | 0.302210        | 0.439481 |
+| 7500  | 0.104900      | 0.325913        | 0.439481 |
+| 8000  | 0.097500      | 0.329446        | 0.431935 |
+| 8500  | 0.088600      | 0.345259        | 0.425898 |
+| 9000  | 0.084900      | 0.342891        | 0.428313 |
+| 9500  | 0.080900      | 0.353081        | 0.424389 |
+| 10000 | 0.075600      | 0.347063        | 0.424992 |
+| 10500 | 0.072800      | 0.330086        | 0.424691 |
+| 11000 | 0.068100      | 0.350658        | 0.421974 |
+| 11500 | 0.064700      | 0.342949        | 0.413522 |
+| 12000 | 0.061500      | 0.341704        | 0.415334 |
+| 12500 | 0.059500      | 0.346279        | 0.411410 |
+| 13000 | 0.057400      | 0.349901        | 0.407184 |
+| 13500 | 0.056400      | 0.347733        | 0.402656 |
+| 14000 | 0.053300      | 0.344899        | 0.405976 |
+| 14500 | 0.052900      | 0.346708        | 0.402656 |
+| 15000 | 0.050600      | 0.344118        | 0.400845 |
+| 15500 | 0.050200      | 0.348396        | 0.402958 |
+| 16000 | 0.049800      | 0.348312        | 0.401751 |
+| 16500 | 0.051900      | 0.348372        | 0.401147 |
+| 17000 | 0.049800      | 0.348580        | 0.401147 |
+
+     
+
+### Framework versions
+- Transformers 4.16.0.dev0
+- Pytorch 1.10.0+cu102
+- Datasets 1.18.1
+- Tokenizers 0.10.3
+
+#### Evaluation Commands
+
+1. To evaluate on `mozilla-foundation/common_voice_8_0` with split `test`
+
+```bash
+python eval.py --model_id Akashpb13/Central_kurdish_xlsr --dataset mozilla-foundation/common_voice_8_0 --config ckb --split test
+```
+

config.json

@@ -0,0 +1,107 @@
+{
+  "_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": 0.01,
+  "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": 0.0,
+  "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": 42,
+  "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": "4.16.1",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 43,
+  "xvector_output_dim": 512
+}

eval.py

@@ -0,0 +1,129 @@
+from datasets import load_dataset, load_metric, Audio, Dataset
+from transformers import pipeline, AutoFeatureExtractor
+import re
+import argparse
+import unicodedata
+from typing import Dict
+
+
+def log_results(result: Dataset, args: Dict[str, str]):
+    """ DO NOT CHANGE. This function computes and logs the result metrics. """
+
+    log_outputs = args.log_outputs
+    dataset_id = "_".join(args.dataset.split("/") + [args.config, args.split])
+
+    # load metric
+    wer = load_metric("wer")
+    cer = load_metric("cer")
+
+    # compute metrics
+    wer_result = wer.compute(references=result["target"], predictions=result["prediction"])
+    cer_result = cer.compute(references=result["target"], predictions=result["prediction"])
+
+    # print & log results
+    result_str = (
+        f"WER: {wer_result}\n"
+        f"CER: {cer_result}"
+    )
+    print(result_str)
+
+    with open(f"{dataset_id}_eval_results.txt", "w") as f:
+        f.write(result_str)
+
+    # log all results in text file. Possibly interesting for analysis
+    if log_outputs is not None:
+        pred_file = f"log_{dataset_id}_predictions.txt"
+        target_file = f"log_{dataset_id}_targets.txt"
+
+        with open(pred_file, "w") as p, open(target_file, "w") as t:
+
+            # mapping function to write output
+            def write_to_file(batch, i):
+                p.write(f"{i}" + "\n")
+                p.write(batch["prediction"] + "\n")
+                t.write(f"{i}" + "\n")
+                t.write(batch["target"] + "\n")
+
+            result.map(write_to_file, with_indices=True)
+
+
+def normalize_text(text: str) -> str:
+    """ DO ADAPT FOR YOUR USE CASE. this function normalizes the target text. """
+    
+    CHARS_TO_IGNORE = [",", "?", "¿", ".", "!", "¡", ";", "；", ":", '""', "%", '"', "�", "ʿ", "·", "჻", "~", "՞",
+                   "؟", "،", "।", "॥", "«", "»", "„", "“", "”", "「", "」", "‘", "’", "《", "》", "(", ")", "[", "]",
+                   "{", "}", "=", "`", "_", "+", "<", ">", "…", "–", "°", "´", "ʾ", "‹", "›", "©", "®", "—", "→", "。",
+                   "、", "﹂", "﹁", "‧", "～", "﹏", "，", "｛", "｝", "（", "）", "［", "］", "【", "】", "‥", "〽",
+                   "『", "』", "〝", "〟", "⟨", "⟩", "〜", "：", "！", "？", "♪", "؛", "/", "\\",]
+    chars_to_ignore_regex = f"[{re.escape(''.join(CHARS_TO_IGNORE))}]"
+
+    text = text.lower()
+    
+    text = re.sub(chars_to_ignore_regex, "", text)
+
+    text = " ".join(text.split())
+
+    return text
+
+
+def main(args):
+    # load dataset
+    dataset = load_dataset(args.dataset, args.config, split=args.split, use_auth_token=True)
+
+    # for testing: only process the first two examples as a test
+    # dataset = dataset.select(range(10))
+
+    # load processor
+    feature_extractor = AutoFeatureExtractor.from_pretrained(args.model_id)
+    sampling_rate = feature_extractor.sampling_rate
+
+    # resample audio
+    dataset = dataset.cast_column("audio", Audio(sampling_rate=sampling_rate))
+
+    # load eval pipeline
+    asr = pipeline("automatic-speech-recognition", model=args.model_id)
+
+    # map function to decode audio
+    def map_to_pred(batch):
+        prediction = asr(batch["audio"]["array"], chunk_length_s=args.chunk_length_s, stride_length_s=args.stride_length_s)
+
+        batch["prediction"] = prediction["text"]
+        batch["target"] = normalize_text(batch["sentence"])
+        return batch
+
+    # run inference on all examples
+    result = dataset.map(map_to_pred, remove_columns=dataset.column_names)
+
+    # compute and log_results
+    # do not change function below
+    log_results(result, args)
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+
+    parser.add_argument(
+        "--model_id", type=str, required=True, help="Model identifier. Should be loadable with 🤗 Transformers"
+    )
+    parser.add_argument(
+        "--dataset", type=str, required=True, help="Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets"
+    )
+    parser.add_argument(
+        "--config", type=str, required=True, help="Config of the dataset. *E.g.* `'en'`  for Common Voice"
+    )
+    parser.add_argument(
+        "--split", type=str, required=True, help="Split of the dataset. *E.g.* `'test'`"
+    )
+    parser.add_argument(
+        "--chunk_length_s", type=float, default=None, help="Chunk length in seconds. Defaults to None. For long audio files a good value would be 5.0 seconds."
+    )
+    parser.add_argument(
+        "--stride_length_s", type=float, default=None, help="Stride of the audio chunks. Defaults to None. For long audio files a good value would be 1.0 seconds."
+    )
+    parser.add_argument(
+        "--log_outputs", action='store_true', help="If defined, write outputs to log file for analysis."
+    )
+    args = parser.parse_args()
+
+    main(args)
+

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:896f0159864e1ed83c339360f394da16ad4ec4a6c8f297fb1b3ea5aa4318f1a7
+size 1261983732

mozilla-foundation_common_voice_8_0_ckb_test_eval_results.txt

@@ -0,0 +1,2 @@
+WER: 0.36754389884276845
+CER: 0.07827896768334217

preprocessor_config.json

@@ -0,0 +1,10 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "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 sha256:389a8c7465980807312bc135a917d9a93683a889e8be2b3158dd3c4f743d73fd
+size 1262099953

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": "|", "tokenizer_class": "Wav2Vec2CTCTokenizer", "processor_class": "Wav2Vec2Processor"}

vocab.json

@@ -0,0 +1 @@
+{"ێ": 0, "د": 1, "ي": 2, "ش": 3, "ڵ": 4, "ى": 5, "ر": 6, "غ": 7, "س": 8, "-": 9, "ک": 10, "خ": 11, "ث": 13, "ڕ": 14, "ە": 15, "ی": 16, "گ": 17, "ز": 18, "ژ": 19, "ؤ": 20, "ن": 21, "ڤ": 22, "ب": 23, "ج": 24, "ا": 25, "ف": 26, "ح": 27, "ت": 28, "پ": 29, "ھ": 30, "ل": 31, "ق": 32, "ع": 33, "ئ": 34, "ۆ": 35, "چ": 36, "ك": 37, "م": 38, "ه": 39, "و": 40, "|": 12, "[UNK]": 41, "[PAD]": 42}