add input image component

app.py

@@ -19,20 +19,26 @@ model = Magma.from_checkpoint(
     device = 'cuda:0'
 )
 
-def generate(context, length, temperature, top_k):
-  context = context.strip()
+def generate(image,context, length, temperature, top_k):
+  # context = context.strip()
 
-  url_regex = r'https?:\/\/(www\.)?[-a-zA-Z0-9@:%._\+~#=]{1,256}\.[a-zA-Z0-9()]{1,6}\b([-a-zA-Z0-9()@:%_\+.~#?&//=]*)'
-  lines = context.split('\n')
-  inputs = []
-  for line in lines:
-    if re.match(url_regex, line):
-      try:
-        inputs.append(ImageInput(line))
-      except Exception as e:
-        return str(e)
-    else:
-      inputs.append(line)
+  # url_regex = r'https?:\/\/(www\.)?[-a-zA-Z0-9@:%._\+~#=]{1,256}\.[a-zA-Z0-9()]{1,6}\b([-a-zA-Z0-9()@:%_\+.~#?&//=]*)'
+  # lines = context.split('\n')
+  # inputs = []
+  # for line in lines:
+  #   if re.match(url_regex, line):
+  #     try:
+  #       inputs.append(ImageInput(line))
+  #     except Exception as e:
+  #       return str(e)
+  #   else:
+  #     inputs.append(line)
+
+  inputs =[
+    ## supports urls and path/to/image
+    ImageInput(image),
+    context
+  ]
 
   ## returns a tensor of shape: (1, 149, 4096)
   embeddings = model.preprocess_inputs(inputs)  
@@ -50,10 +56,10 @@ def generate(context, length, temperature, top_k):
 iface = gr.Interface(
   fn=generate,
   inputs=[
-    gr.inputs.Textbox(
-      label="Prompt (image URLs need to be on their own lines):",
-      default="https://www.art-prints-on-demand.com/kunst/thomas_cole/woods_hi.jpg\nDescribe the painting:",
-      lines=7),
+    [gr.inputs.Image(type="filepath"),gr.inputs.Textbox(
+      label="Prompt:",
+      default="Describe the painting:",
+      lines=7)],
     gr.inputs.Slider(minimum=1, maximum=100, default=15, step=1, label="Output tokens:"),
     gr.inputs.Slider(minimum=0.0, maximum=1.0, default=0.7, label='Temperature'),
     gr.inputs.Slider(minimum=0, maximum=100, default=0, step=1, label='Top K')