format

README.md

@@ -19,9 +19,9 @@ tags:
 # Affective TTS / SoundScapes
 
   - [SHIFT TTS tool](https://github.com/audeering/shift) 
-  - Analysis of emotionality [#1](https://huggingface.co/dkounadis/artificial-styletts2/discussions/2)
-  - Soundscapes `trees, water, castles` via [AudioGen](https://huggingface.co/dkounadis/artificial-styletts2/discussions/3)
-  - `landscape2soundscape.py` shows how to overlay TTS & sound to image and create video
+  - Analysis of TTS emotionality [#1](https://huggingface.co/dkounadis/artificial-styletts2/discussions/2)
+  - Soundscapes `trees, water, ..` via [AudioGen](https://huggingface.co/dkounadis/artificial-styletts2/discussions/3)
+  - `landscape2soundscape.py` - overlays TTS & sound to still image and create video
 
 ## Available Voices
 
@@ -29,16 +29,16 @@ tags:
 
 ## Flask API
 
-
-```
-git clone https://huggingface.co/dkounadis/artificial-styletts2
-```
-
 <details>
 <summary>
 Create virtualenv
 </summary>
 
+
+```
+git clone https://huggingface.co/dkounadis/artificial-styletts2
+```
+
 ```
 virtualenv --python=python3 ~/.envs/.my_env
 source ~/.envs/.my_env/bin/activate
@@ -46,7 +46,6 @@ cd artificial-styletts2/
 pip install -r requirements.txt
 ```
 
-
 </details>
 
 Start Flask
@@ -57,7 +56,7 @@ CUDA_DEVICE_ORDER=PCI_BUS_ID HF_HOME=./hf_home CUDA_VISIBLE_DEVICES=2 python api
 
 ## Landscape 2 Soundscape
 
-The following needs `api.py` to be already running `on a tmux session`. 
+The following needs `api.py` to be already running on a tmux session. 
 
 ```python
 # TTS & soundscape - overlay to .mp4

demo.py

@@ -4,7 +4,7 @@ import numpy as np
 
 print('\n\n\n\n___________________')
 
-txt = 'australian music'
+txt = 'car'
 
 sound_generator = AudioGen.get_pretrained('facebook/audiogen-medium')
 sound_generator.set_generation_params(duration=1)   # why is generating so long at 14 seconds
@@ -12,4 +12,4 @@ sound_generator.set_generation_params(duration=1)   # why is generating so long
 x = sound_generator.generate([txt])[0].detach().cpu().numpy()[0, :]
 x /= np.abs(x).max() + 1e-7
 
-audiofile.write('_audio3_.wav', x, 16000)
+audiofile.write('_audio1_.wav', x, 16000)