transformers-4.49

config.json

@@ -36,7 +36,7 @@
     "vocab_size": 32207
   },
   "torch_dtype": "float32",
-  "transformers_version": "4.48.3",
+  "transformers_version": "4.49.0",
   "vision_config": {
     "apply_layernorm": true,
     "architectures": [

modeling_maira2.py

@@ -105,7 +105,7 @@ class Maira2ForConditionalGeneration(LlavaForConditionalGeneration):
         output_hidden_states: Optional[bool] = None,
         return_dict: Optional[bool] = None,
         cache_position: Optional[torch.LongTensor] = None,
-        num_logits_to_keep: int = 0,
+        logits_to_keep: int = 0,
     ) -> Union[Tuple, LlavaCausalLMOutputWithPast]:
         r"""
         Args:
@@ -246,7 +246,7 @@ class Maira2ForConditionalGeneration(LlavaForConditionalGeneration):
             output_hidden_states=output_hidden_states,
             return_dict=return_dict,
             cache_position=cache_position,
-            num_logits_to_keep=num_logits_to_keep,
+            logits_to_keep=logits_to_keep,
         )
 
         logits = outputs[0]
@@ -356,4 +356,4 @@ class Maira2ForConditionalGeneration(LlavaForConditionalGeneration):
         if labels is None:
             final_labels = None
 
-        return final_embedding, final_attention_mask, final_labels, position_ids
+        return final_embedding, final_attention_mask, final_labels, position_ids