Support InternLM3 Dense 8B Model (#6640)

* support internlm3

* update

* update

* update

* add hint

Former-commit-id: 24ab7ae0944c5f373e9cac60f0332e704824a057

src/llamafactory/data/template.py

@@ -732,6 +732,14 @@ _register_template(
     stop_words=["<|im_end|>"],
 )
 
+_register_template(
+    name="intern3",
+    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|>"]
+)
 
 _register_template(
     name="llama2",

src/llamafactory/extras/constants.py

@@ -817,6 +817,15 @@ register_model_group(
     template="intern2",
 )
 
+register_model_group(
+    models={
+        "InternLM3-8B-Chat": {
+            DownloadSource.DEFAULT: "internlm/internlm3-8b-instruct",
+            DownloadSource.MODELSCOPE: "Shanghai_AI_Laboratory/internlm3-8b-instruct",
+        },
+    },
+    template="intern3",
+)
 
 register_model_group(
     models={

src/llamafactory/model/patcher.py

@@ -24,6 +24,7 @@ from transformers.modeling_utils import is_fsdp_enabled
 
 from ..extras import logging
 from ..extras.misc import infer_optim_dtype
+from ..extras.packages import is_transformers_version_greater_than
 from .model_utils.attention import configure_attn_implementation, print_attn_implementation
 from .model_utils.checkpointing import prepare_model_for_training
 from .model_utils.embedding import resize_embedding_layer
@@ -117,6 +118,9 @@ def patch_config(
     if "LlavaLlamaForCausalLM" in getattr(config, "architectures", []):
         raise ValueError("Please download llava models with hf-compatible format: https://huggingface.co/llava-hf")
 
+    if getattr(config, "model_type", None) == "internlm3" and not is_transformers_version_greater_than("4.47.1"):
+        raise RuntimeError("InternLM3 model requires transformers >= 4.47.1, please upgrade it.")
+
     # deepspeed zero3 is not compatible with low_cpu_mem_usage
     init_kwargs["low_cpu_mem_usage"] = model_args.low_cpu_mem_usage and (not is_deepspeed_zero3_enabled())