Bring the Space up to the pyharp example standard (taxonomy tags, __main__ guard, show_error, v0.4.0)

README.md

@@ -4,7 +4,7 @@ emoji: ✏️
 colorFrom: indigo
 colorTo: purple
 sdk: gradio
-sdk_version: 5.28.0
+sdk_version: 6.24.0
 python_version: "3.10.13"
 app_file: app.py
 pinned: false

app.py

@@ -9,7 +9,7 @@ import gradio as gr
 import torch
 import torchaudio
 from huggingface_hub import hf_hub_download
-from pyharp import ModelCard, build_endpoint
+from pyharp import ModelCard, build_endpoint, Subcategory
 
 from data.tokenizer import AudioTokenizer, TextTokenizer
 from inference_speech_editing_scale import inference_one_sample
@@ -69,33 +69,33 @@ def edit_speech(audio_path: str, target_text: str, start_time: float, end_time:
     return str(output_path)
 
 
-with gr.Blocks(title="VoiceCraft") as demo:
-    build_endpoint(
-        model_card=ModelCard(
-            name="VoiceCraft",
-            description="Edit English speech while retaining the surrounding recording.",
-            author="VoiceCraft",
-            tags=["speech-editing", "voicecraft", "audio"],
-        ),
-        input_components=[
-            gr.Audio(type="filepath", label="Speech Recording")
-            .harp_required(True)
-            .set_info("English speech to edit."),
-            gr.Textbox(label="Target Transcript", lines=4)
-            .harp_required(True)
-            .set_info("Full transcript of the recording, with the replacement words in place."),
-            gr.Number(label="Edit Start (seconds)", value=0.5, precision=2)
-            .set_info("Start of the words to replace."),
-            gr.Number(label="Edit End (seconds)", value=1.5, precision=2)
-            .set_info("End of the words to replace."),
-        ],
-        output_components=[
-            gr.Audio(type="filepath", label="Edited Speech")
-            .set_info("16 kHz WAV recording."),
-        ],
-        process_fn=edit_speech,
-    )
+if __name__ == "__main__":
+    with gr.Blocks(title="VoiceCraft") as demo:
+        build_endpoint(
+            model_card=ModelCard(
+                name="VoiceCraft",
+                description="Edit English speech while retaining the surrounding recording.",
+                author="VoiceCraft",
+                tags=[Subcategory.EDITING, 'speech-editing', 'voicecraft', 'audio'],
+            ),
+            input_components=[
+                gr.Audio(type="filepath", label="Speech Recording")
+                .harp_required(True)
+                .set_info("English speech to edit."),
+                gr.Textbox(label="Target Transcript", lines=4)
+                .harp_required(True)
+                .set_info("Full transcript of the recording, with the replacement words in place."),
+                gr.Number(label="Edit Start (seconds)", value=0.5, precision=2)
+                .set_info("Start of the words to replace."),
+                gr.Number(label="Edit End (seconds)", value=1.5, precision=2)
+                .set_info("End of the words to replace."),
+            ],
+            output_components=[
+                gr.Audio(type="filepath", label="Edited Speech")
+                .set_info("16 kHz WAV recording."),
+            ],
+            process_fn=edit_speech,
+        )
 
 
-if __name__ == "__main__":
-    demo.queue(default_concurrency_limit=1).launch(show_error=True, pwa=True)
+    demo.queue(default_concurrency_limit=1).launch(share=True, show_error=True, pwa=True)

requirements.txt

@@ -2,9 +2,8 @@ torch==2.8.0
 torchaudio==2.8.0
 xformers==0.0.32.post2
 numpy==2.0.2
-gradio==5.28.0
 spaces>=0.51.1
-huggingface-hub>=0.33.5,<1.0
+huggingface-hub
 phonemizer==3.2.1
 num2words==0.5.13
 av
@@ -26,4 +25,4 @@ torchmetrics
 encodec
 protobuf
 safetensors
-git+https://github.com/TEAMuP-dev/pyharp.git@v0.3.0
+git+https://github.com/TEAMuP-dev/pyharp.git@v0.4.0