Support new features of MiniCPM-V (#6626)

* fix template name

* tiny fix

* support minicpm-o-2.6

Former-commit-id: 53034a61c7654358f46916cbc370910fb2aeff3b

src/llamafactory/data/collator.py

@@ -153,9 +153,8 @@ class MultiModalDataCollatorForSeq2Seq(DataCollatorForSeq2Seq):
             features = features.data  # use default_collate() instead of BatchEncoding.to()
 
         if "image_bound" in features:  # for minicpmv inputs
-            features["position_ids"] = (
-                torch.arange(features["input_ids"].size(1)).long().unsqueeze(0).expand_as(features["input_ids"])
-            )
+            bsz, seq_length = features["input_ids"].shape
+            features["position_ids"] = torch.arange(seq_length).long().repeat(bsz, 1)
             return {"data": features, "labels": features["labels"]}
 
         return features