Update README.md

README.md

@@ -40,8 +40,8 @@ from transformers import AutoModel, AutoTokenizer
 import requests
 from io import BytesIO
 
-IMAGENET_MEAN = (0.485, 0.456, 0.406)
-IMAGENET_STD = (0.229, 0.224, 0.225)
+IMAGENET_MEAN = (0.5, 0.5, 0.5)
+IMAGENET_STD = (0.5, 0.5, 0.5)
 
 def build_transform(input_size):
     MEAN, STD = IMAGENET_MEAN, IMAGENET_STD
@@ -128,7 +128,8 @@ response = requests.get(image_url)
 image_data = BytesIO(response.content)
 pixel_values = load_image(image_data, max_num=10).to(torch.bfloat16).cuda()
 generation_config = dict(max_new_tokens=1024, do_sample=True)
-# optimal alpha ranges from 64 to 576
+
+# The recommended range for `num_image_token` is 64 to 576, and the value can be adjusted based on task requirements.
 num_image_token = 256
 
 # pure-text conversation