Update README.md
Browse files
README.md
CHANGED
@@ -29,18 +29,18 @@ datasets:
|
|
29 |
<img src="https://huggingface.co/datasets/parler-tts/images/resolve/main/thumbnail.png" alt="Parler Logo" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/>
|
30 |
|
31 |
|
32 |
-
# Parler-TTS Mini
|
33 |
|
34 |
-
<a target="_blank" href="https://huggingface.co/spaces/
|
35 |
<img src="https://huggingface.co/datasets/huggingface/badges/raw/main/open-in-hf-spaces-sm.svg" alt="Open in HuggingFace"/>
|
36 |
</a>
|
37 |
|
38 |
-
**Parler-TTS Mini v1.1
|
39 |
|
40 |
It is a fine-tuned version, trained on a [cleaned version](https://huggingface.co/datasets/PHBJT/cml-tts-cleaned-levenshtein) of [CML-TTS](https://huggingface.co/datasets/ylacombe/cml-tts) and on the non-English version of [Multilingual LibriSpeech](https://huggingface.co/datasets/facebook/multilingual_librispeech).
|
41 |
In all, this represents some 9,200 hours of non-English data. To retain English capabilities, we also added back the [LibriTTS-R English dataset](https://huggingface.co/datasets/parler-tts/libritts_r_filtered), some 580h of high-quality English data.
|
42 |
|
43 |
-
**Parler-TTS Mini
|
44 |
|
45 |
Thanks to its **better prompt tokenizer**, it can easily be extended to other languages. This tokenizer has a larger vocabulary and handles byte fallback, which simplifies multilingual training.
|
46 |
|
@@ -79,8 +79,8 @@ import soundfile as sf
|
|
79 |
|
80 |
device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
81 |
|
82 |
-
model = ParlerTTSForConditionalGeneration.from_pretrained("parler-tts/parler-tts-mini-
|
83 |
-
tokenizer = AutoTokenizer.from_pretrained("parler-tts/parler-tts-mini-
|
84 |
description_tokenizer = AutoTokenizer.from_pretrained(model.config.text_encoder._name_or_path)
|
85 |
|
86 |
prompt = "Hey, how are you doing today?"
|
@@ -108,8 +108,8 @@ import soundfile as sf
|
|
108 |
|
109 |
device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
110 |
|
111 |
-
model = ParlerTTSForConditionalGeneration.from_pretrained("parler-tts/parler-tts-mini-
|
112 |
-
tokenizer = AutoTokenizer.from_pretrained("parler-tts/parler-tts-mini-
|
113 |
description_tokenizer = AutoTokenizer.from_pretrained(model.config.text_encoder._name_or_path)
|
114 |
|
115 |
prompt = "Hey, how are you doing today?"
|
|
|
29 |
<img src="https://huggingface.co/datasets/parler-tts/images/resolve/main/thumbnail.png" alt="Parler Logo" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/>
|
30 |
|
31 |
|
32 |
+
# Parler-TTS Mini Multilingual
|
33 |
|
34 |
+
<a target="_blank" href="https://huggingface.co/spaces/PHBJT/multi_parler_tts">
|
35 |
<img src="https://huggingface.co/datasets/huggingface/badges/raw/main/open-in-hf-spaces-sm.svg" alt="Open in HuggingFace"/>
|
36 |
</a>
|
37 |
|
38 |
+
**Parler-TTS Mini Multilingual v1.1** is a multilingual extension of [Parler-TTS Mini](https://huggingface.co/parler-tts/parler-tts-mini-v1.1).
|
39 |
|
40 |
It is a fine-tuned version, trained on a [cleaned version](https://huggingface.co/datasets/PHBJT/cml-tts-cleaned-levenshtein) of [CML-TTS](https://huggingface.co/datasets/ylacombe/cml-tts) and on the non-English version of [Multilingual LibriSpeech](https://huggingface.co/datasets/facebook/multilingual_librispeech).
|
41 |
In all, this represents some 9,200 hours of non-English data. To retain English capabilities, we also added back the [LibriTTS-R English dataset](https://huggingface.co/datasets/parler-tts/libritts_r_filtered), some 580h of high-quality English data.
|
42 |
|
43 |
+
**Parler-TTS Mini Multilingual** can speak in 7 European languages: English, French, Spanish, Portuguese, Polish, German, Italian and Dutch.
|
44 |
|
45 |
Thanks to its **better prompt tokenizer**, it can easily be extended to other languages. This tokenizer has a larger vocabulary and handles byte fallback, which simplifies multilingual training.
|
46 |
|
|
|
79 |
|
80 |
device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
81 |
|
82 |
+
model = ParlerTTSForConditionalGeneration.from_pretrained("parler-tts/parler-tts-mini-multilingual").to(device)
|
83 |
+
tokenizer = AutoTokenizer.from_pretrained("parler-tts/parler-tts-mini-multilingual")
|
84 |
description_tokenizer = AutoTokenizer.from_pretrained(model.config.text_encoder._name_or_path)
|
85 |
|
86 |
prompt = "Hey, how are you doing today?"
|
|
|
108 |
|
109 |
device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
110 |
|
111 |
+
model = ParlerTTSForConditionalGeneration.from_pretrained("parler-tts/parler-tts-mini-multilingual").to(device)
|
112 |
+
tokenizer = AutoTokenizer.from_pretrained("parler-tts/parler-tts-mini-multilingual")
|
113 |
description_tokenizer = AutoTokenizer.from_pretrained(model.config.text_encoder._name_or_path)
|
114 |
|
115 |
prompt = "Hey, how are you doing today?"
|