tiny fix

Former-commit-id: bee1bd43b946501690d70e4980205f9d82404296

src/llamafactory/__init__.py

@@ -20,7 +20,7 @@ Level:
 
 Dependency graph:
   main:
-    transformers>=4.41.2,<=4.44.3
+    transformers>=4.41.2,<=4.45.0
     datasets>=2.16.0,<=2.21.0
     accelerate>=0.30.1,<=0.33.0
     peft>=0.11.1,<=0.12.0

src/llamafactory/chat/vllm_engine.py

@@ -16,6 +16,7 @@ import uuid
 from typing import TYPE_CHECKING, Any, AsyncGenerator, AsyncIterator, Dict, List, Optional, Sequence, Union
 
 from ..data import get_template_and_fix_tokenizer
+from ..extras.constants import IMAGE_PLACEHOLDER
 from ..extras.logging import get_logger
 from ..extras.misc import get_device_count
 from ..extras.packages import is_vllm_available, is_vllm_version_greater_than_0_5, is_vllm_version_greater_than_0_5_1
@@ -115,13 +116,11 @@ class VllmEngine(BaseEngine):
     ) -> AsyncIterator["RequestOutput"]:
         request_id = "chatcmpl-{}".format(uuid.uuid4().hex)
 
-        if (
-            self.processor is not None
-            and image is not None
-            and not hasattr(self.processor, "image_seq_length")
-            and self.template.image_token not in messages[0]["content"]
-        ):  # llava-like models (TODO: paligemma models)
-            messages[0]["content"] = self.template.image_token * self.image_feature_size + messages[0]["content"]
+        if image is not None:
+            if IMAGE_PLACEHOLDER not in messages[0]["content"]:
+                messages[0]["content"] = IMAGE_PLACEHOLDER + messages[0]["content"]
+
+            messages = self.template.mm_plugin.process_messages(messages, [image], self.processor)
 
         paired_messages = messages + [{"role": "assistant", "content": ""}]
         system = system or self.generating_args["default_system"]

src/llamafactory/data/collator.py

@@ -68,7 +68,7 @@ class CustomDataCollatorForSeq2Seq(DataCollatorForSeq2Seq):
     """
 
     def __call__(self, features: Sequence[Dict[str, Any]]) -> Dict[str, "torch.Tensor"]:
-        image_grid_thw = None
+        image_grid_thw = None  # TODO: better handle various VLMs
         if "image_grid_thw" in features[0]:
             image_grid_thw_list = [
                 torch.Tensor(feature["image_grid_thw"]).long()

src/llamafactory/data/mm_plugin.py

@@ -74,6 +74,9 @@ class BasePlugin:
         images: Sequence["ImageObject"],
         processor: Optional["ProcessorMixin"],
     ) -> List[Dict[str, str]]:
+        r"""
+        Pre-processes input messages before tokenization for VLMs.
+        """
         return messages
 
     def process_token_ids(
@@ -83,6 +86,9 @@ class BasePlugin:
         tokenizer: "PreTrainedTokenizer",
         processor: Optional["ProcessorMixin"],
     ) -> Tuple[List[int], Optional[List[int]]]:
+        r"""
+        Pre-processes token ids after tokenization for VLMs.
+        """
         return input_ids, labels
 
     def get_mm_inputs(
@@ -91,6 +97,9 @@ class BasePlugin:
         feature_seqlens: Dict[str, int],
         processor: Optional["ProcessorMixin"],
     ) -> Dict[str, Any]:
+        r"""
+        Builds batched multimodal inputs for VLMs.
+        """
         return {}
 
     def process_model_inputs(
@@ -100,6 +109,9 @@ class BasePlugin:
         feature_seqlens: Dict[str, int],
         processor: Optional["ProcessorMixin"],
     ) -> None:
+        r"""
+        Appends multimodal inputs to model inputs for VLMs.
+        """
         return
 
 

src/llamafactory/data/processors/feedback.py

@@ -84,7 +84,7 @@ def preprocess_feedback_dataset(
     tokenizer: "PreTrainedTokenizer",
     processor: Optional["ProcessorMixin"],
     data_args: "DataArguments",
-) -> Dict[str, List[List[int]]]:
+) -> Dict[str, List[Any]]:
     # create unrelated input-output pairs for estimating the KL term by flipping the matched pairs
     kl_response = examples["response"][::-1]
     model_inputs = defaultdict(list)

src/llamafactory/data/processors/pairwise.py

@@ -70,7 +70,7 @@ def preprocess_pairwise_dataset(
     tokenizer: "PreTrainedTokenizer",
     processor: Optional["ProcessorMixin"],
     data_args: "DataArguments",
-) -> Dict[str, List[List[int]]]:
+) -> Dict[str, List[Any]]:
     # build input pairs with format `<bos> X`, `Y1 <eos>` and `Y2 <eos>`
     model_inputs = defaultdict(list)
     for i in range(len(examples["prompt"])):

src/llamafactory/data/processors/pretrain.py

@@ -27,7 +27,7 @@ if TYPE_CHECKING:
 
 def preprocess_pretrain_dataset(
     examples: Dict[str, List[Any]], tokenizer: "PreTrainedTokenizer", data_args: "DataArguments"
-) -> Dict[str, List[List[int]]]:
+) -> Dict[str, List[Any]]:
     # build grouped texts with format `X1 X2 X3 ...` if packing is enabled
     eos_token = "<|end_of_text|>" if data_args.template == "llama3" else tokenizer.eos_token
     text_examples = [messages[0]["content"] + eos_token for messages in examples["prompt"]]

src/llamafactory/data/processors/unsupervised.py

@@ -62,7 +62,7 @@ def preprocess_unsupervised_dataset(
     tokenizer: "PreTrainedTokenizer",
     processor: Optional["ProcessorMixin"],
     data_args: "DataArguments",
-) -> Dict[str, List[List[int]]]:
+) -> Dict[str, List[Any]]:
     # build inputs with format `<bos> X` and labels with format `Y <eos>`
     model_inputs = defaultdict(list)
     for i in range(len(examples["prompt"])):