[data] correct gpt_oss template format_assistant (#10269)

Co-authored-by: frozenleaves <frozen@Mac.local>

docker/docker-npu/Dockerfile

@@ -1,6 +1,6 @@
 # https://hub.docker.com/r/ascendai/cann/tags
 
-ARG BASE_IMAGE=quay.io/ascend/cann:8.3.rc2-910b-ubuntu22.04-py3.11
+ARG BASE_IMAGE=quay.io/ascend/cann:8.5.1-910b-ubuntu22.04-py3.11
 FROM ${BASE_IMAGE}
 
 # Installation arguments
@@ -33,9 +33,11 @@ RUN pip config set global.index-url "${PIP_INDEX}" && \
 COPY . /app
 
 # Install torch-npu
-RUN pip uninstall -y torch torchvision torchaudio && \
-    pip install --no-cache-dir "torch==2.7.1" "torch-npu==2.7.1" "torchvision==0.22.1" "torchaudio==2.7.1" --index-url "${PYTORCH_INDEX}" && \
-    pip install --no-cache-dir -e . --no-build-isolation && \
+RUN source /usr/local/Ascend/ascend-toolkit/set_env.sh
+RUN pip uninstall -y torch torchvision torchaudio
+RUN pip install --no-cache-dir -r requirements/npu.txt --index-url "${PYTORCH_INDEX}"
+RUN pip install --no-cache-dir -r requirements/deepspeed.txt
+RUN pip install --no-cache-dir -e . --no-build-isolation && \
     pip install --no-cache-dir -r requirements/metrics.txt --no-build-isolation
 
 # Set up volumes

docker/docker-npu/docker-compose.yml

@@ -33,7 +33,7 @@ services:
       dockerfile: ./docker/docker-npu/Dockerfile
       context: ../..
       args:
-        BASE_IMAGE: quay.io/ascend/cann:8.3.rc2-a3-ubuntu22.04-py3.11
+        BASE_IMAGE: quay.io/ascend/cann:8.5.1-a3-ubuntu22.04-py3.11
         PIP_INDEX: https://pypi.org/simple
     container_name: llamafactory-a3
     image: llamafactory:npu-a3

requirements/npu.txt

@@ -1,4 +1,4 @@
 torch==2.7.1
-torch-npu==2.7.1
+torch-npu==2.7.1.post2
 torchvision==0.22.1
 torchaudio==2.7.1

src/llamafactory/data/mm_plugin.py

@@ -161,7 +161,9 @@ class MMPluginMixin:
         video_processor: BaseImageProcessor = getattr(
             processor, "video_processor", getattr(processor, "image_processor", None)
         )
-        feature_extractor: SequenceFeatureExtractor = getattr(processor, "feature_extractor", None)
+        feature_extractor: SequenceFeatureExtractor = getattr(processor, "feature_extractor", None) or getattr(
+            processor, "audio_processor", None
+        )
         if len(images) != 0 and self.image_token is None:
             raise ValueError(
                 "This model does not support image input. Please check whether the correct `template` is used."
@@ -390,7 +392,9 @@ class MMPluginMixin:
                 mm_inputs.update(video_processor(videos, return_tensors="pt"))
 
         if len(audios) != 0:
-            feature_extractor: SequenceFeatureExtractor = getattr(processor, "feature_extractor", None)
+            feature_extractor: SequenceFeatureExtractor = getattr(processor, "feature_extractor", None) or getattr(
+                processor, "audio_processor", None
+            )
             audios = self._regularize_audios(
                 audios,
                 sampling_rate=getattr(processor, "audio_sampling_rate", 16000),
@@ -1876,7 +1880,9 @@ class Qwen2OmniPlugin(Qwen2VLPlugin):
     ) -> dict[str, "torch.Tensor"]:
         image_processor: BaseImageProcessor = getattr(processor, "image_processor", None)
         video_processor: BaseVideoProcessor = getattr(processor, "video_processor", None)
-        feature_extractor: SequenceFeatureExtractor = getattr(processor, "feature_extractor", None)
+        feature_extractor: SequenceFeatureExtractor = getattr(processor, "feature_extractor", None) or getattr(
+            processor, "audio_processor", None
+        )
         mm_inputs = {}
         if len(images) != 0:
             images = self._regularize_images(

src/llamafactory/data/template.py

@@ -1113,7 +1113,7 @@ register_template(
 register_template(
     name="gpt_oss",
     format_user=StringFormatter(slots=["<|start|>user<|message|>{{content}}<|end|><|start|>assistant"]),
-    format_assistant=StringFormatter(slots=["{{content}}<|end|>"]),
+    format_assistant=StringFormatter(slots=["{{content}}"]),
     format_system=StringFormatter(slots=["<|start|>system<|message|>{{content}}<|end|>"]),
     default_system="You are ChatGPT, a large language model trained by OpenAI.",
     thought_words=("<|channel|>analysis<|message|>", "<|end|><|start|>assistant<|channel|>final<|message|>"),