fix numpy

app.py

@@ -3,6 +3,8 @@ import time
 import os
 os.environ['PYTORCH_JIT'] = "0"
 os.system('mim install mmcv==2.0.1')
+os.system("pip install 'numpy<2.0'")
+
 # import spaces
 import sys
 import argparse
@@ -12,7 +14,7 @@ from functools import partial
 
 # import spaces
 
-from mmengine.runner import Runner
+# from mmengine.runner import Runner
 from mmengine.dataset import Compose
 from mmengine.runner.amp import autocast
 from mmengine.config import Config, DictAction, ConfigDict
@@ -60,9 +62,10 @@ def generate_image_embeddings(prompt_image,
                               projector,
                               device='cuda:0'):
     prompt_image = prompt_image.convert('RGB')
-    inputs = vision_processor(images=prompt_image,
+    inputs = vision_processor(images=[prompt_image],
                               return_tensors="pt",
-                              padding=True)
+                              padding=False)
+
     inputs = inputs.to(device)
     image_outputs = vision_encoder(**inputs)
     img_feats = image_outputs.image_embeds.view(1, -1)
@@ -283,10 +286,11 @@ if __name__ == '__main__':
 
     cfg.load_from = checkpoint
 
-    if 'runner_type' not in cfg:
-        runner = Runner.from_cfg(cfg)
-    else:
-        runner = RUNNERS.build(cfg)
+    # if 'runner_type' not in cfg:
+    #     runner = Runner.from_cfg(cfg)
+    # else:
+    #     
+    runner = RUNNERS.build(cfg)
     # runner.test()
     runner.call_hook('before_run')
     runner.load_or_resume()