README.md

@@ -283,7 +283,7 @@ Choose your path:
 | [GPT-OSS](https://huggingface.co/openai)                          | 20B/120B                         | gpt                  |
 | [Granite 3.0-3.3](https://huggingface.co/ibm-granite)             | 1B/2B/3B/8B                      | granite3             |
 | [Granite 4](https://huggingface.co/ibm-granite)                   | 7B                               | granite4             |
-| [Hunyuan (MT)](https://huggingface.co/tencent/)                   | 7B                               | hunyuan              |
+| [Hunyuan](https://huggingface.co/tencent/)                        | 7B                               | hunyuan              |
 | [Index](https://huggingface.co/IndexTeam)                         | 1.9B                             | index                |
 | [InternLM 2-3](https://huggingface.co/internlm)                   | 7B/8B/20B                        | intern2              |
 | [InternVL 2.5-3.5](https://huggingface.co/OpenGVLab)              | 1B/2B/4B/8B/14B/30B/38B/78B/241B | intern_vl            |

README_zh.md

@@ -285,7 +285,7 @@ https://github.com/user-attachments/assets/43b700c6-a178-41db-b1f8-8190a5d3fcfc
 | [GPT-OSS](https://huggingface.co/openai)                          | 20B/120B                         | gpt                  |
 | [Granite 3.0-3.3](https://huggingface.co/ibm-granite)             | 1B/2B/3B/8B                      | granite3             |
 | [Granite 4](https://huggingface.co/ibm-granite)                   | 7B                               | granite4             |
-| [Hunyuan (MT)](https://huggingface.co/tencent/)                   | 7B                               | hunyuan              |
+| [Hunyuan](https://huggingface.co/tencent/)                        | 7B                               | hunyuan              |
 | [Index](https://huggingface.co/IndexTeam)                         | 1.9B                             | index                |
 | [InternLM 2-3](https://huggingface.co/internlm)                   | 7B/8B/20B                        | intern2              |
 | [InternVL 2.5-3.5](https://huggingface.co/OpenGVLab)              | 1B/2B/4B/8B/14B/30B/38B/78B/241B | intern_vl            |

src/llamafactory/data/template.py

@@ -1201,10 +1201,10 @@ register_template(
 
 register_template(
     name="hunyuan",
-    format_user=StringFormatter(slots=["{{content}}<|extra_0|>"]),
-    format_assistant=StringFormatter(slots=["{{content}}<|eos|>"]),
-    format_system=StringFormatter(slots=["{{content}}<|extra_4|>"]),
-    format_prefix=EmptyFormatter(slots=["<|startoftext|>"]),
+    format_user=StringFormatter(slots=["<|bos|>user\n{{content}}<|eos|>\n<|bos|>assistant\n"]),
+    format_assistant=StringFormatter(slots=["{{content}}<|eos|>\n"]),
+    format_system=StringFormatter(slots=["<|bos|>system\n{{content}}<|eos|>\n"]),
+    format_prefix=EmptyFormatter(slots=["<|bos|>"]),
     stop_words=["<|eos|>"],
 )
 

src/llamafactory/extras/constants.py

@@ -1152,10 +1152,6 @@ register_model_group(
             DownloadSource.DEFAULT: "tencent/Hunyuan-7B-Instruct",
             DownloadSource.MODELSCOPE: "AI-ModelScope/Hunyuan-7B-Instruct",
         },
-        "Hunyuan-MT-7B-Instruct": {
-            DownloadSource.DEFAULT: "tencent/Hunyuan-MT-7B",
-            DownloadSource.MODELSCOPE: "Tencent-Hunyuan/Hunyuan-MT-7B",
-        },
     },
     template="hunyuan",
 )

src/llamafactory/model/model_utils/attention.py

@@ -14,6 +14,8 @@
 
 from typing import TYPE_CHECKING
 
+from transformers.utils import is_flash_attn_2_available, is_torch_sdpa_available
+
 from ...extras import logging
 from ...extras.constants import AttentionFunction
 
@@ -28,8 +30,6 @@ logger = logging.get_logger(__name__)
 
 
 def configure_attn_implementation(config: "PretrainedConfig", model_args: "ModelArguments") -> None:
-    from transformers.utils import is_flash_attn_2_available
-
     if getattr(config, "model_type", None) == "gemma2":
         if model_args.flash_attn == AttentionFunction.AUTO or model_args.flash_attn == AttentionFunction.FA2:
             if is_flash_attn_2_available():
@@ -51,8 +51,6 @@ def configure_attn_implementation(config: "PretrainedConfig", model_args: "Model
         requested_attn_implementation = "eager"
 
     elif model_args.flash_attn == AttentionFunction.SDPA:
-        from transformers.utils import is_torch_sdpa_available
-
         if not is_torch_sdpa_available():
             logger.warning_rank0("torch>=2.1.1 is required for SDPA attention.")
             return