fix audio output

README.md

@@ -15,74 +15,5 @@ Check out the configuration reference at https://huggingface.co/docs/hub/spaces-
 
 1. Create an account on Coqui and get an api key from [account settings](https://app.coqui.ai/account).
 2. [Clone a voice](https://docs.coqui.ai/reference/voices_clone_from_file_create) using a recording of length between 3 and 5 seconds. I used a `.wav` file.
-3. [Create a sample of the cloned voice]()
-
-## Creating a voice with [Resemble.ai](https://www.resemble.ai/)
-
-### [API Reference for creating a voice](https://docs.app.resemble.ai/docs/resource_voice/create)
-
-1. Get your Resemble API key after creating an account with [Resemble.ai](https://www.resemble.ai/).
-2. Make a request to the endpoint with the `Authorization` header set to `Token <your_resemble_api_key>` and a JSON body with `name` and the `dataset_url`:
-
-```
-{
- "name": "chatty_vader",
- "dataset_url": "https://huggingface.co/spaces/aar2dee2/chatty_vader/resolve/main/data.zip"
-}
-```
-
-This will return a response like below.
-
-```
-{
-  "success": true,
-  "item": {
-    "uuid": "10f91c43",
-    "name": "chatty_vader",
-    "status": "initializing",
-    "dataset_url": "https://huggingface.co/spaces/aar2dee2/chatty_vader/resolve/main/data.zip",
-    "created_at": "2023-04-07T06:38:29.307Z",
-    "updated_at": "2023-04-07T06:38:29.323Z"
-  }
-}
-```
-
-If you specify a `callback_url`, you get a notification from Resemble when the voice has been created.
-
-I temporarily modified the `app.py` so I could provide my huggingface space url as the `callback_url` in the `create a voice` request.
-
-```
-def receive_data_from_resemble(data):
-    print("data from resemble", data)
-    return data
-
-
-iface = gr.Interface(fn=receive_data_from_resemble,
-                     inputs="json", outputs="json")
-iface.launch()
-```
-
-Test creating a clip
-[api endpoint](https://app.resemble.ai/api/v2/projects/e2da3585/clips)
-json body;
-
-```json
-{
-  "title": "testing_vader",
-  "body": "There is no need to panic. It will all be over soon",
-  "voice_uuid": "f00d917f",
-  "is_public": true,
-  "callback_uri": "https://aar2dee2-chatty-vader.hf.space/run/predict"
-}
-```
-
-response received:
-
-```
-{
-  "success": false,
-  "message": "This voice is still building and cannot be used at this time."
-}
-```
-
-Same for voice id "10f91c43"
+3. [Create a sample of the cloned voice](https://docs.coqui.ai/reference/samples_create)
+4. Store the sample and save the `voice_id` in your env variables as `COQUI_VOICE_ID`.

app.py

@@ -142,7 +142,7 @@ description = "Darth Vader resurrected with all the knowledge of humanity"
 
 mic_translate = gr.Interface(
     fn=main,
-    inputs=gr.Audio(source="microphone",  format="wav"),
+    inputs=gr.Audio(source="microphone"),
     outputs=gr.Audio(label="Generated Speech", type="numpy"),
     title=title,
     description=description,
@@ -150,8 +150,8 @@ mic_translate = gr.Interface(
 
 file_translate = gr.Interface(
     fn=main,
-    inputs=gr.Audio(source="upload", type="filepath", format="wav"),
-    outputs=gr.Audio(label="Generated Speech", type="filepath", format="wav"),
+    inputs=gr.Audio(source="upload", type="filepath"),
+    outputs=gr.Audio(label="Generated Speech", type="numpy"),
     title=title,
     description=description,
 )