.

Author: ZiyiTsang

verl/workers/fsdp_workers.py

@@ -479,7 +479,7 @@ def _process_multi_modal_inputs(self, data: DataProto):
                     else:
                         multi_modal_inputs = {}
 
-                    if multi_modal_inputs['pixel_values'].ndim() == 3:
+                    if multi_modal_inputs['pixel_values'].ndim == 3:
                         # Some image processor return with batch dim (such as glm4.1), we need to squeeze the pix_value.
                         # i.e. (1,patch,pix_per_patch) -> (patch,pix_per_patch)
                         multi_modal_inputs['pixel_values'] = multi_modal_inputs['pixel_values'].squeeze(0)