Update

app.py

@@ -245,7 +245,7 @@ if __name__ == '__main__':
 
   gradio_helpers.set_warmup_function(warmup)
   for name, (repo, filename) in models.MODELS.items():
-    gradio_helpers.register_download(name, repo, filename + "-text-model-q4_k_m.gguf")
-    gradio_helpers.register_download(name, repo, filename + "-mmproj-f16.gguf")
+    gradio_helpers.register_download(name + "-text-model-q4_k_m.gguf", repo, filename + "-text-model-q4_k_m.gguf")
+    gradio_helpers.register_download(name + "-mmproj-f16.gguf", repo, filename + "-mmproj-f16.gguf")
 
   create_app().queue().launch()

models.py

@@ -78,7 +78,8 @@ def generate(
   # with gradio_helpers.timed('computation', start_message=True):
   #   tokens = model.predict(params, batch, sampler=sampler)
 
-  model_path = gradio_helpers.get_paths()[model_name]
+  model_path = gradio_helpers.get_paths()[model_name + "-text-model-q4_k_m.gguf"]
+  clip_path = gradio_helpers.get_paths()[model_name + "-mmproj-f16.gguf"]
   print(model_path)
   print(gradio_helpers.get_paths())
   model = llama_cpp.Llama(