[Model][VLM] Add Qwen2-VL model support (#7905)

Co-authored-by: Roger Wang <136131678+ywang96@users.noreply.github.com>
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>

vllm/transformers_utils/processor.py

@@ -0,0 +1,37 @@
+from typing import cast
+
+
+def get_processor(
+    processor_name: str,
+    *args,
+    trust_remote_code: bool = False,
+    **kwargs,
+):
+    """Gets a processor for the given model name via HuggingFace."""
+    # don't put this import at the top level
+    # it will call torch.cuda.device_count()
+    from transformers import AutoProcessor
+    from transformers.processing_utils import ProcessorMixin
+
+    try:
+        processor = AutoProcessor.from_pretrained(
+            processor_name,
+            *args,
+            trust_remote_code=trust_remote_code,
+            **kwargs)
+    except ValueError as e:
+        # If the error pertains to the processor class not existing or not
+        # currently being imported, suggest using the --trust-remote-code flag.
+        # Unlike AutoTokenizer, AutoProcessor does not separate such errors
+        if not trust_remote_code:
+            err_msg = (
+                "Failed to load the processor. If the processor is "
+                "a custom processor not yet available in the HuggingFace "
+                "transformers library, consider setting "
+                "`trust_remote_code=True` in LLM or using the "
+                "`--trust-remote-code` flag in the CLI.")
+            raise RuntimeError(err_msg) from e
+        else:
+            raise e
+
+    return cast(ProcessorMixin, processor)