Remove need for trust_remote_code on HF version

I opened for this one, but IMO should be valid for all! Also `AutoProcessor` will work directly with your path :)

README.md

@@ -369,10 +369,8 @@ from transformers import AutoTokenizer, AutoModel
 path = "OpenGVLab/InternVL3_5-8B-HF"
 model = AutoModel.from_pretrained(
     path,
-    torch_dtype=torch.bfloat16,
-    low_cpu_mem_usage=True,
-    use_flash_attn=True,
-    trust_remote_code=True).eval().cuda()
+    dtype=torch.bfloat16,
+    low_cpu_mem_usage=True).eval()
 ```
 
 #### BNB 8-bit Quantization
@@ -383,11 +381,10 @@ from transformers import AutoTokenizer, AutoModel
 path = "OpenGVLab/InternVL3_5-8B-HF"
 model = AutoModel.from_pretrained(
     path,
-    torch_dtype=torch.bfloat16,
+    dtype=torch.bfloat16,
     load_in_8bit=True,
-    low_cpu_mem_usage=True,
-    use_flash_attn=True,
-    trust_remote_code=True).eval()
+    low_cpu_mem_usage=True
+    device_map="auto").eval()
 ```
 
 #### Multiple GPUs
@@ -400,10 +397,8 @@ from transformers import AutoTokenizer, AutoModel
 path = "OpenGVLab/InternVL3_5-8B-HF"
 model = AutoModel.from_pretrained(
     path,
-    torch_dtype=torch.bfloat16,
+    dtype=torch.bfloat16,
     low_cpu_mem_usage=True,
-    use_flash_attn=True,
-    trust_remote_code=True,
     device_map="auto").eval()
 ```