Update README.md

README.md

@@ -25,7 +25,7 @@ pipeline_tag: text-to-speech
 
 ## Model Summary
 
-YarnGPT2 is a text-to-speech (TTS) model designed to synthesize Nigerian-accented Languages (yoruba, igbo, hausa and english) leveraging pure language modelling without external adapters or complex architectures, offering high-quality, natural, and culturally relevant speech synthesis for diverse applications.
+YarnGPT2 is a text-to-speech (TTS) model designed to synthesise Nigerian-accented Languages (yoruba, igbo, hausa and english) leveraging pure language modelling without external adapters or complex architectures, offering high-quality, natural, and culturally relevant speech synthesis for diverse applications.
 
 <video controls width="600">
   <source src="https://huggingface.co/saheedniyi/YarnGPT/resolve/main/audio/YearnGPT.mp4" type="video/mp4">
@@ -33,11 +33,11 @@ YarnGPT2 is a text-to-speech (TTS) model designed to synthesize Nigerian-accente
 </video>
 
 #### How to use (Colab)
-The model can generate audio on its own but its better to use a voice to prompt the model:
+The model can generate audio on its own, but it's better to use a voice to prompt the model:
 
-##### Voices (arranged in order of perfomance and stability)
-- English: idera, chinenye, jude, emma,umar,,joke,zainab ,osagie, remi, tayo
-- Yoruba: yoruba_male2, yoruba_female2, yoruba_feamle1
+##### Voices (arranged in order of performance and stability)
+- English: Idera, Chinenye, Jude, Emma, Umar, Joke, Zainab, Osagie, Remi, Tayo
+- Yoruba: yoruba_male2, yoruba_female2, yoruba_female1
 - Igbo: igbo_female2, igbo_male2,igbo_female1,
 - Hausa: hausa_feamle1,hausa_female2, hausa_male2,hausa_male1
 
@@ -93,7 +93,7 @@ output  = model.generate(
             temperature=0.1,
             repetition_penalty=1.1,
             max_length=4000,
-            #num_beams=5,# using a beam size helps for the local languages but not english
+            #num_beams=5,# using a beam size helps for the local languages but not English
         )
 
 codes=audio_tokenizer.get_codes(output)
@@ -242,7 +242,7 @@ Trained on a dataset of publicly available Nigerian movies, podcasts ( using the
 
 #### Preprocessing 
 
-Audio files were preprocessed and resampled to 24Khz and tokenized using [wavtokenizer](https://huggingface.co/novateur/WavTokenizer).
+Audio files were preprocessed and resampled to 24KHz and tokenized using [wavtokenizer](https://huggingface.co/novateur/WavTokenizer).
 
 #### Training Hyperparameters
 - **Number of epochs:** 5
@@ -253,14 +253,14 @@ Audio files were preprocessed and resampled to 24Khz and tokenized using [wavtok
 
 #### Hardware
 
-- **GPUs:** 1 A100 (google colab: 50 hours)
+- **GPUs:** 1 A100 (Google Colab: 50 hours)
 
 #### Software
 
 - **Training Framework:** Pytorch
 
 ## Future Improvements?
-- Scaling up model size and human-annotaed/ reviewed training data
+- Scaling up model size and human-annotated/ reviewed training data
 - Wrap the model around an API endpoint 
 - Voice cloning.
 - Potential expansion into speech-to-speech assistant models