Image Caption Model

app.py

@@ -1,17 +1,33 @@
 import gradio as gr
+from transformers import AutoProcessor, MusicgenForConditionalGeneration
+import numpy as np
 
 title = "Musicalization System of Painting Demo"
 description = "Pui Ching Middle School: Musicalization System of Painting Demo"
 
-def greet(name):
-    return "Hello " + name + "!!"
+def generate_music(text):
+    processor = AutoProcessor.from_pretrained("facebook/musicgen-small")
+    model = MusicgenForConditionalGeneration.from_pretrained("facebook/musicgen-small")
+
+    inputs = processor(
+        text=[text],
+        padding=True,
+        return_tensors="pt",
+    )
+
+    audio_values = model.generate(**inputs, max_new_tokens=256)
+    sampling_rate = model.audio_encoder.config.sampling_rate
+    target_dtype = np.int16
+    max_range = np.iinfo(target_dtype).max
+    audio_values = audio_values[0, 0].numpy()
+    return sampling_rate, (audio_values * max_range).astype(np.int16)
+
 
 iface = gr.Interface(
-    fn=greet,
+    fn=generate_music,
     title=title,
     description=description,
-    #fn=greet,
-    inputs=gr.Image(),
-    outputs=gr.Audio())
+    inputs=gr.Text(label="Content"),
+    outputs=gr.Audio(label='Generated Music'))
 
 iface.launch()

ram_swin_large_14m.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15c729c793af28b9d107c69f85836a1356d76ea830d4714699fb62e55fcc08ed
+size 5625634877