test import path

misc.py

@@ -24,7 +24,7 @@ import time
 from contextlib import ContextDecorator
 from typing import Any, Callable, TypeVar
 
-from AutoregressiveVideo2WorldGeneration import log
+from .log import log
 import numpy as np
 import termcolor
 import torch

video2world_hf.py

@@ -79,14 +79,14 @@ class ARVideo2World(PreTrainedModel):
             offload_text_encoder_model=args.offload_text_encoder_model,
         )
 
-    def forward(self, prompt, input_image_or_video_path):
+    def forward(self):
         args = self.args
 
          # Load input image(s) or video(s)
         input_videos = load_vision_input(
             input_type=args.input_type,
             batch_input_path=args.batch_input_path,
-            input_image_or_video_path=input_image_or_video_path,
+            input_image_or_video_path=args.input_image_or_video_path,
             data_resolution=args.data_resolution,
             num_input_frames=args.num_input_frames,
         )
@@ -95,7 +95,7 @@ class ARVideo2World(PreTrainedModel):
         if args.batch_input_path:
             prompts_list = read_prompts_from_file(args.batch_input_path)
         else:
-            prompts_list = [{"visual_input": input_image_or_video_path, "prompt": prompt}]
+            prompts_list = [{"visual_input": args.input_image_or_video_path, "prompt": args.prompt}]
 
         # Iterate through prompts
         for idx, prompt_entry in enumerate(prompts_list):