End of training

Signed-off-by: jjyaoao <jjyaoao@126.com>

.gitignore

@@ -0,0 +1 @@
+checkpoint-*/

README.md

@@ -1,3 +1,63 @@
 ---
 license: mit
+tags:
+- generated_from_trainer
+datasets:
+- common_voice_13_0
+model-index:
+- name: speecht5_finetuned_voxpopuli_nl
+  results: []
 ---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# speecht5_finetuned_voxpopuli_nl
+
+This model is a fine-tuned version of [arham061/speecht5_finetuned_voxpopuli_nl](https://huggingface.co/arham061/speecht5_finetuned_voxpopuli_nl) on the common_voice_13_0 dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0.5508
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 1e-05
+- train_batch_size: 4
+- eval_batch_size: 2
+- seed: 42
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 32
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_steps: 500
+- training_steps: 3000
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 0.5058        | 7.74  | 1000 | 0.5431          |
+| 0.4938        | 15.49 | 2000 | 0.5487          |
+| 0.4909        | 23.23 | 3000 | 0.5508          |
+
+
+### Framework versions
+
+- Transformers 4.32.0.dev0
+- Pytorch 2.0.0+cu117
+- Datasets 2.13.1
+- Tokenizers 0.13.3

added_tokens.json

@@ -0,0 +1,4 @@
+{
+  "<ctc_blank>": 80,
+  "<mask>": 79
+}

app.py

@@ -0,0 +1,114 @@
+import torch
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Tokenizer
+from urllib.request import urlopen
+from io import BytesIO
+import soundfile as sf
+import numpy as np
+
+# Load the TTS model from the Hugging Face Hub
+model_name = "arham061/speecht5_finetuned_voxpopuli_nl"  # Replace with your actual model name
+model = Wav2Vec2ForCTC.from_pretrained(model_name)
+tokenizer = Wav2Vec2Tokenizer.from_pretrained(model_name)
+
+# Buckwalter to Unicode mapping
+buck2uni = {
+    u"\u0627": "A",
+    u"\u0675": "A",
+    u"\u0673": "A",
+    u"\u0630": "A",
+    u"\u0622": "AA",
+    u"\u0628": "B",
+    u"\u067E": "P",
+    u"\u062A": "T",
+    u"\u0637": "T",
+    u"\u0679": "T",
+    u"\u062C": "J",
+    u"\u0633": "S",
+    u"\u062B": "S",
+    u"\u0635": "S",
+    u"\u0686": "CH",
+    u"\u062D": "H",
+    u"\u0647": "H",
+    u"\u0629": "H",
+    u"\u06DF": "H",
+    u"\u062E": "KH",
+    u"\u062F": "D",
+    u"\u0688": "D",
+    u"\u0630": "Z",
+    u"\u0632": "Z",
+    u"\u0636": "Z",
+    u"\u0638": "Z",
+    u"\u068E": "Z",
+    u"\u0631": "R",
+    u"\u0691": "R",
+    u"\u0634": "SH",
+    u"\u063A": "GH",
+    u"\u0641": "F",
+    u"\u06A9": "K",
+    u"\u0642": "K",
+    u"\u06AF": "G",
+    u"\u0644": "L",
+    u"\u0645": "M",
+    u"\u0646": "N",
+    u"\u06BA": "N",
+    u"\u0648": "O",
+    u"\u0649": "Y",
+    u"\u0626": "Y",
+    u"\u06CC": "Y",
+    u"\u06D2": "E",
+    u"\u06C1": "H",
+    u"\u064A": "E",
+    u"\u06C2": "AH",
+    u"\u06BE": "H",
+    u"\u0639": "A",
+    u"\u0643": "K",
+    u"\u0621": "A",
+    u"\u0624": "O",
+    u"\u060C": "",  # separator ulta comma
+}
+
+def transString(string, reverse=0):
+    """Given a Unicode string, transliterate into Buckwalter. To go from
+    Buckwalter back to Unicode, set reverse=1"""
+    for k, v in buck2uni.items():
+        if not reverse:
+            string = string.replace(k, v)
+        else:
+            string = string.replace(v, k)
+    return string
+
+
+def generate_audio(text):
+    # Convert input text to Roman Urdu
+    roman_urdu = transString(text)
+
+    # Tokenize the input text
+    inputs = tokenizer(roman_urdu, return_tensors="pt").input_values
+
+    # Generate speech from the model
+    with torch.no_grad():
+        logits = model(inputs).logits
+
+    # Convert logits to audio waveform
+    predicted_ids = torch.argmax(logits, dim=-1)
+    audio = tokenizer.decode(predicted_ids[0], skip_special_tokens=True)
+
+    return audio
+
+
+# Example usage
+def main():
+    # Get input text in Urdu
+    input_text_urdu = input("Enter text in Urdu: ")
+
+    # Generate audio
+    audio_output = generate_audio(input_text_urdu)
+
+    # Save audio as a .wav file
+    sf.write("output.wav", audio_output, samplerate=22050)
+
+    print("Audio generated and saved as 'output.wav'")
+
+
+if __name__ == "__main__":
+    main()

config.json

@@ -0,0 +1,92 @@
+{
+  "_name_or_path": "arham061/speecht5_finetuned_voxpopuli_nl",
+  "activation_dropout": 0.1,
+  "apply_spec_augment": true,
+  "architectures": [
+    "SpeechT5ForTextToSpeech"
+  ],
+  "attention_dropout": 0.1,
+  "bos_token_id": 0,
+  "conv_bias": false,
+  "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
+  ],
+  "decoder_attention_heads": 12,
+  "decoder_ffn_dim": 3072,
+  "decoder_layerdrop": 0.1,
+  "decoder_layers": 6,
+  "decoder_start_token_id": 2,
+  "encoder_attention_heads": 12,
+  "encoder_ffn_dim": 3072,
+  "encoder_layerdrop": 0.1,
+  "encoder_layers": 12,
+  "encoder_max_relative_position": 160,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_norm": "group",
+  "feat_proj_dropout": 0.0,
+  "guided_attention_loss_num_heads": 2,
+  "guided_attention_loss_scale": 10.0,
+  "guided_attention_loss_sigma": 0.4,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "is_encoder_decoder": true,
+  "layer_norm_eps": 1e-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,
+  "max_length": 1876,
+  "max_speech_positions": 1876,
+  "max_text_positions": 600,
+  "model_type": "speecht5",
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_mel_bins": 80,
+  "pad_token_id": 1,
+  "positional_dropout": 0.1,
+  "reduction_factor": 2,
+  "scale_embedding": false,
+  "speaker_embedding_dim": 512,
+  "speech_decoder_postnet_dropout": 0.5,
+  "speech_decoder_postnet_kernel": 5,
+  "speech_decoder_postnet_layers": 5,
+  "speech_decoder_postnet_units": 256,
+  "speech_decoder_prenet_dropout": 0.5,
+  "speech_decoder_prenet_layers": 2,
+  "speech_decoder_prenet_units": 256,
+  "torch_dtype": "float32",
+  "transformers_version": "4.30.2",
+  "use_cache": false,
+  "use_guided_attention_loss": true,
+  "vocab_size": 81
+}

preprocessor_config.json

@@ -0,0 +1,19 @@
+{
+  "do_normalize": false,
+  "feature_extractor_type": "SpeechT5FeatureExtractor",
+  "feature_size": 1,
+  "fmax": 7600,
+  "fmin": 80,
+  "frame_signal_scale": 1.0,
+  "hop_length": 16,
+  "mel_floor": 1e-10,
+  "num_mel_bins": 80,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "processor_class": "SpeechT5Processor",
+  "reduction_factor": 2,
+  "return_attention_mask": true,
+  "sampling_rate": 16000,
+  "win_function": "hann_window",
+  "win_length": 64
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:554b3a098956923d4719fd3445dcda4896dc343fd5ebaea12277e0971d530795
+size 585485317

requirements.txt

@@ -0,0 +1,3 @@
+torch==1.9.0
+transformers==4.11.3
+soundfile==0.11.0

runs/Jul14_07-06-38_64fa198a07c3/events.out.tfevents.1689318514.64fa198a07c3.314.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb112cc6bc1f0058fd498ebe98dde9e48cb080b27e20659871723b3e15bfc882
+size 32233

runs/Jul16_12-21-55_48e7dd533140/events.out.tfevents.1689510259.48e7dd533140.399.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9de4c790d9ba5a969003b6e7a3ecef192a10c0c00dfcce13d3b352674246ce0c
+size 25700

special_tokens_map.json

@@ -0,0 +1,13 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<pad>",
+  "unk_token": "<unk>"
+}

spm_char.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7fcc48f3e225f627b1641db410ceb0c8649bd2b0c982e150b03f8be3728ab560
+size 238473

tokenizer_config.json

@@ -0,0 +1,11 @@
+{
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "</s>",
+  "model_max_length": 600,
+  "pad_token": "<pad>",
+  "processor_class": "SpeechT5Processor",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "SpeechT5Tokenizer",
+  "unk_token": "<unk>"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:efa9e2e07460ea4899c70de1b01cd01ed09a7e9b687abead7dfcf18a2149fc47
+size 4155