[Bugfix] `embed_is_patch` for Idefics3 (#15696)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

vllm/model_executor/models/commandr.py

@@ -24,7 +24,6 @@
 from typing import Iterable, Optional, Set, Tuple, Union
 
 import torch
-import torch.utils.checkpoint
 from torch import nn
 from transformers import CohereConfig
 

vllm/model_executor/models/mllama.py

@@ -21,7 +21,6 @@ from typing import List, Literal, Optional, Set, Tuple, TypedDict, Union
 import numpy as np
 import torch
 import torch.nn.functional as F
-import torch.utils.checkpoint
 import transformers.models.mllama.configuration_mllama as config_mllama
 from PIL.Image import Image
 from torch import nn

vllm/model_executor/models/qwen2_audio.py

@@ -160,7 +160,7 @@ class Qwen2AudioMultiModalProcessor(
         mm_kwargs: Mapping[str, Any],
     ) -> BatchFeature:
         # Text-only input not supported in composite processor
-        if not mm_data or not mm_data.get("audios", []):
+        if not mm_data.get("audios", []):
             prompt_ids = self.info.get_tokenizer().encode(prompt)
             prompt_ids = self._apply_hf_processor_tokens_only(prompt_ids)
             return BatchFeature(dict(input_ids=[prompt_ids]), tensor_type="pt")

vllm/model_executor/models/ultravox.py

@@ -8,7 +8,6 @@ from functools import cached_property
 from typing import Any, Literal, Optional, Set, Tuple, TypedDict, Union
 
 import torch
-import torch.utils.checkpoint
 from torch import nn
 from torch.nn import functional as F
 from transformers import BatchFeature, ProcessorMixin
@@ -160,7 +159,7 @@ class UltravoxMultiModalProcessor(
         mm_kwargs: Mapping[str, object],
     ) -> BatchFeature:
         # Text-only input not supported in composite processor
-        if not mm_data or not mm_data.get("audios", []):
+        if not mm_data.get("audios", []):
             prompt_ids = self.info.get_tokenizer().encode(
                 prompt, add_special_tokens=False)
             prompt_ids = self._apply_hf_processor_tokens_only(prompt_ids)