Update hawk/conversation/conversation_video.py

hawk/conversation/conversation_video.py

@@ -8,7 +8,6 @@ from PIL import Image
 import sys
 import os
 import torch
-import spaces
 
 from transformers import AutoTokenizer, AutoModelForCausalLM, LlamaTokenizer
 from transformers import StoppingCriteria, StoppingCriteriaList
@@ -22,6 +21,7 @@ from hawk.processors.video_processor import ToTHWC,ToUint8,load_video,load_video
 from hawk.processors import Blip2ImageEvalProcessor
             
 from hawk.models.ImageBind.data import load_and_transform_audio_data
+
 class SeparatorStyle(Enum):
     """Different separator style."""
     SINGLE = auto()
@@ -178,7 +178,6 @@ class Chat:
         #                   torch.tensor([2277, 29937]).to(self.device)]  # '###' can be encoded in two different ways.
         # self.stopping_criteria = StoppingCriteriaList([StoppingCriteriaSub(stops=stop_words_ids)])
 
-    @spaces.GPU
     def ask(self, text, conv):
         if len(conv.messages) > 0 and conv.messages[-1][0] == conv.roles[0] \
                 and ('</Video>' in conv.messages[-1][1] or '</Image>' in conv.messages[-1][1]):  # last message is image.
@@ -186,7 +185,6 @@ class Chat:
         else:
             conv.append_message(conv.roles[0], text)
     
-    @spaces.GPU
     def answer(self, conv, img_list, max_new_tokens=300, num_beams=1, min_length=1, top_p=0.9,
                repetition_penalty=1.0, length_penalty=1, temperature=1.0, max_length=2000):
         conv.append_message(conv.roles[1], None)
@@ -237,7 +235,6 @@ class Chat:
         conv.messages[-1][1] = output_text
         return output_text, output_token.cpu().numpy()
 
-    @spaces.GPU
     def upload_video(self, video_path, conv, img_list):
 
         msg = ""
@@ -285,7 +282,6 @@ class Chat:
                 conv.append_message(conv.roles[0], "<Video><ImageHere></Video> "+ msg)
             return "Received."
     
-    @spaces.GPU
     def upload_video_without_audio(self, video_path, conv, img_list):
         msg = ""
         if isinstance(video_path, str):  # is a video path
@@ -326,7 +322,6 @@ class Chat:
         conv.append_message(conv.roles[0], "<Video><ImageHere></Video> ")
         return "Received."
 
-    @spaces.GPU
     def upload_img(self, image, conv, img_list):
 
         msg = ""
@@ -350,7 +345,6 @@ class Chat:
 
         return "Received."
     
-    @spaces.GPU
     def get_context_emb(self, conv, img_list):
         prompt = conv.get_prompt()
         prompt_segs = prompt.split('<ImageHere>')