[model] Support Intern-S1-mini (#8976)

README.md

@@ -118,6 +118,8 @@ Choose your path:
 
 ## Changelog
 
+[25/08/20] We supported fine-tuning the **[Intern-S1-mini](https://huggingface.co/internlm/Intern-S1-mini)** models. See [PR #8976](https://github.com/hiyouga/LLaMA-Factory/pull/8976) to get started.
+
 [25/08/06] We supported fine-tuning the **[GPT-OSS](https://github.com/openai/gpt-oss)** models. See [PR #8826](https://github.com/hiyouga/LLaMA-Factory/pull/8826) to get started.
 
 [25/07/02] We supported fine-tuning the **[GLM-4.1V-9B-Thinking](https://github.com/THUDM/GLM-4.1V-Thinking)** model.
@@ -277,6 +279,7 @@ Choose your path:
 | [Index](https://huggingface.co/IndexTeam)                         | 1.9B                             | index               |
 | [InternLM 2-3](https://huggingface.co/internlm)                   | 7B/8B/20B                        | intern2             |
 | [InternVL 2.5-3](https://huggingface.co/OpenGVLab)                | 1B/2B/8B/14B/38B/78B             | intern_vl           |
+| [InternLM/Intern-S1-mini](https://huggingface.co/internlm/)       | 8B                               | intern_s1           |
 | [Kimi-VL](https://huggingface.co/moonshotai)                      | 16B                              | kimi_vl             |
 | [Llama](https://github.com/facebookresearch/llama)                | 7B/13B/33B/65B                   | -                   |
 | [Llama 2](https://huggingface.co/meta-llama)                      | 7B/13B/70B                       | llama2              |

README_zh.md

@@ -120,6 +120,8 @@ https://github.com/user-attachments/assets/43b700c6-a178-41db-b1f8-8190a5d3fcfc
 
 ## 更新日志
 
+[25/08/20] 我们支持了 **[Intern-S1-mini](https://huggingface.co/internlm/Intern-S1-mini)** 模型的微调。查看 [PR #8976](https://github.com/hiyouga/LLaMA-Factory/pull/8976) 以使用。
+
 [25/08/06] 我们支持了 **[GPT-OSS](https://github.com/openai/gpt-oss)** 模型的微调。查看 [PR #8826](https://github.com/hiyouga/LLaMA-Factory/pull/8826) 以使用。
 
 [25/07/02] 我们支持了 **[GLM-4.1V-9B-Thinking](https://github.com/THUDM/GLM-4.1V-Thinking)** 模型的微调。
@@ -279,6 +281,7 @@ https://github.com/user-attachments/assets/43b700c6-a178-41db-b1f8-8190a5d3fcfc
 | [Index](https://huggingface.co/IndexTeam)                         | 1.9B                             | index               |
 | [InternLM 2-3](https://huggingface.co/internlm)                   | 7B/8B/20B                        | intern2             |
 | [InternVL 2.5-3](https://huggingface.co/OpenGVLab)                | 1B/2B/8B/14B/38B/78B             | intern_vl           |
+| [InternLM/Intern-S1-mini](https://huggingface.co/internlm/)       | 8B                               | intern_s1           |
 | [Kimi-VL](https://huggingface.co/moonshotai)                      | 16B                              | kimi_vl             |
 | [Llama](https://github.com/facebookresearch/llama)                | 7B/13B/33B/65B                   | -                   |
 | [Llama 2](https://huggingface.co/meta-llama)                      | 7B/13B/70B                       | llama2              |

src/llamafactory/data/template.py

@@ -1201,6 +1201,17 @@ register_template(
 )
 
 
+register_template(
+    name="intern_s1",
+    format_user=StringFormatter(slots=["<|im_start|>user\n{{content}}<|im_end|>\n<|im_start|>assistant\n"]),
+    format_assistant=StringFormatter(slots=["{{content}}<|im_end|>\n"]),
+    format_system=StringFormatter(slots=["<|im_start|>system\n{{content}}<|im_end|>\n"]),
+    format_prefix=EmptyFormatter(slots=[{"bos_token"}]),
+    stop_words=["<|im_end|>"],
+    mm_plugin=get_mm_plugin(name="intern_vl", image_token="<image>", video_token="<video>"),
+)
+
+
 # copied from qwen template
 register_template(
     name="keye_vl",

src/llamafactory/extras/constants.py

@@ -1264,6 +1264,16 @@ register_model_group(
     multimodal=True,
 )
 
+register_model_group(
+    models={
+        "Intern-S1-mini": {
+            DownloadSource.DEFAULT: "internlm/Intern-S1-mini",
+            DownloadSource.MODELSCOPE: "Shanghai_AI_Laboratory/Intern-S1-mini",
+        }
+    },
+    template="intern_s1",
+    multimodal=True,
+)
 
 register_model_group(
     models={

src/llamafactory/model/model_utils/liger_kernel.py

@@ -47,6 +47,8 @@ def apply_liger_kernel(
         from liger_kernel.transformers import apply_liger_kernel_to_gemma3_text as apply_liger_kernel
     elif model_type == "glm4":
         from liger_kernel.transformers import apply_liger_kernel_to_glm4 as apply_liger_kernel
+    elif model_type == "glm4v":
+        from liger_kernel.transformers import apply_liger_kernel_to_glm4v as apply_liger_kernel
     elif model_type == "granite":
         from liger_kernel.transformers import apply_liger_kernel_to_granite as apply_liger_kernel
     elif model_type == "llama":

src/llamafactory/model/model_utils/visual.py

@@ -234,6 +234,9 @@ _register_composite_model(
     model_type="internvl",
 )
 
+_register_composite_model(
+    model_type="interns1",
+)
 
 _register_composite_model(
     model_type="Keye",