ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[model] add qwen3 next (#9130)

Browse Source
This commit is contained in:
Yaowei Zheng
2025-09-14 03:16:25 +08:00
committed by GitHub
parent 260b5625c3
commit 812720909e
6 changed files with 87 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
src/llamafactory/data/template.py
View File

@@ -917,6 +917,18 @@ register_template(
)
# copied from chatml template
register_template(
name="ernie",
format_user=StringFormatter(slots=["<|im_start|>user\n{{content}}<|im_end|>\n\n<|im_start|>assistant\n"]),
format_assistant=StringFormatter(slots=["{{content}}<|im_end|>\n\n"]),
format_system=StringFormatter(slots=["<|im_start|>system\n{{content}}<|im_end|>\n\n"]),
format_observation=StringFormatter(slots=["<|im_start|>tool\n{{content}}<|im_end|>\n\n<|im_start|>assistant\n"]),
default_system="<global_setting>\nthink_mode=True\n</global_setting>",
stop_words=["<|im_end|>"],
)
register_template(
name="exaone",
format_user=StringFormatter(slots=["[|user|]{{content}}\n[|assistant|]"]),

1
src/llamafactory/data/tool_utils.py
View File

@@ -78,7 +78,6 @@ SEED_TOOL_PROMPT = (
"lines</parameter>\n</function>\n</seed:tool_call>\n"
)
LING_TOOL_PROMPT = (
"# Tools\n\nYou may call one or more functions to assist with the user query.\n\n"
"You are provided with function signatures within <tools></tools> XML tags:\n<tools>{tool_text}"

60
src/llamafactory/extras/constants.py
View File

@@ -601,6 +601,17 @@ register_model_group(
)
register_model_group(
models={
"ERNIE-4.5-21B-A3B-Thinking": {
DownloadSource.DEFAULT: "baidu/ERNIE-4.5-21B-A3B-Thinking",
DownloadSource.MODELSCOPE: "PaddlePaddle/ERNIE-4.5-21B-A3B-Thinking",
},
},
template="ernie",
)
register_model_group(
models={
"EXAONE-3.0-7.8B-Instruct": {
@@ -1783,6 +1794,10 @@ register_model_group(
DownloadSource.DEFAULT: "openbmb/MiniCPM4-8B",
DownloadSource.MODELSCOPE: "OpenBMB/MiniCPM4-8B",
},
"MiniCPM4.1-8B-Chat": {
DownloadSource.DEFAULT: "openbmb/MiniCPM4.1-8B",
DownloadSource.MODELSCOPE: "OpenBMB/MiniCPM4.1-8B",
},
},
template="cpm4",
)
@@ -1790,7 +1805,7 @@ register_model_group(
register_model_group(
models={
"MiniCPM-o-2_6": {
"MiniCPM-o-2.6": {
DownloadSource.DEFAULT: "openbmb/MiniCPM-o-2_6",
DownloadSource.MODELSCOPE: "OpenBMB/MiniCPM-o-2_6",
},
@@ -1802,7 +1817,7 @@ register_model_group(
register_model_group(
models={
"MiniCPM-V-2_6": {
"MiniCPM-V-2.6": {
DownloadSource.DEFAULT: "openbmb/MiniCPM-V-2_6",
DownloadSource.MODELSCOPE: "OpenBMB/MiniCPM-V-2_6",
},
@@ -1826,7 +1841,7 @@ register_model_group(
register_model_group(
models={
"MiniCPM-V-4_5": {
"MiniCPM-V-4.5": {
DownloadSource.DEFAULT: "openbmb/MiniCPM-V-4_5",
DownloadSource.MODELSCOPE: "OpenBMB/MiniCPM-V-4_5",
},
@@ -1944,6 +1959,37 @@ register_model_group(
)
register_model_group(
models={
"MobileLLM-R1-140M-Base": {
DownloadSource.DEFAULT: "facebook/MobileLLM-R1-140M-base",
DownloadSource.MODELSCOPE: "facebook/MobileLLM-R1-140M-base",
},
"MobileLLM-R1-360M-Base": {
DownloadSource.DEFAULT: "facebook/MobileLLM-R1-360M-base",
DownloadSource.MODELSCOPE: "facebook/MobileLLM-R1-360M-base",
},
"MobileLLM-R1-950M-Base": {
DownloadSource.DEFAULT: "facebook/MobileLLM-R1-950M-base",
DownloadSource.MODELSCOPE: "facebook/MobileLLM-R1-950M-base",
},
"MobileLLM-R1-140M-Instruct": {
DownloadSource.DEFAULT: "facebook/MobileLLM-R1-140M",
DownloadSource.MODELSCOPE: "facebook/MobileLLM-R1-140M",
},
"MobileLLM-R1-360M-Instruct": {
DownloadSource.DEFAULT: "facebook/MobileLLM-R1-360M",
DownloadSource.MODELSCOPE: "facebook/MobileLLM-R1-360M",
},
"MobileLLM-R1-950M-Instruct": {
DownloadSource.DEFAULT: "facebook/MobileLLM-R1-950M",
DownloadSource.MODELSCOPE: "facebook/MobileLLM-R1-950M",
},
},
template="llama3",
)
register_model_group(
models={
"Moonlight-16B-A3B": {
@@ -2912,6 +2958,10 @@ register_model_group(
DownloadSource.DEFAULT: "Qwen/Qwen3-235B-A22B-GPTQ-Int4",
DownloadSource.MODELSCOPE: "Qwen/Qwen3-235B-A22B-GPTQ-Int4",
},
"Qwen/Qwen3-Next-80B-A3B-Thinking": {
DownloadSource.DEFAULT: "Qwen/Qwen3-Next-80B-A3B-Thinking",
DownloadSource.MODELSCOPE: "Qwen/Qwen3-Next-80B-A3B-Thinking",
},
},
template="qwen3",
)
@@ -2931,6 +2981,10 @@ register_model_group(
DownloadSource.DEFAULT: "Qwen/Qwen3-235B-A22B-Instruct-2507",
DownloadSource.MODELSCOPE: "Qwen/Qwen3-235B-A22B-Instruct-2507",
},
"Qwen3-Next-80B-A3B-Instruct": {
DownloadSource.DEFAULT: "Qwen/Qwen3-Next-80B-A3B-Instruct",
DownloadSource.MODELSCOPE: "Qwen/Qwen3-Next-80B-A3B-Instruct",
},
},
template="qwen3_nothink",
)

15
src/llamafactory/webui/locales.py
View File

@@ -34,31 +34,36 @@ LOCALES = {
"en": {
"value": (
"<h3><center>Visit <a href='https://github.com/hiyouga/LLaMA-Factory' target='_blank'>"
"GitHub Page</a></center></h3>"
"GitHub Page</a> <a href='https://llamafactory.readthedocs.io/en/latest/' target='_blank'>"
"Documentation</a></center></h3>"
),
},
"ru": {
"value": (
"<h3><center>Посетить <a href='https://github.com/hiyouga/LLaMA-Factory' target='_blank'>"
"страницу GitHub</a></center></h3>"
"страницу GitHub</a> <a href='https://llamafactory.readthedocs.io/en/latest/' target='_blank'>"
"Документацию</a></center></h3>"
),
},
"zh": {
"value": (
"<h3><center>访问 <a href='https://github.com/hiyouga/LLaMA-Factory' target='_blank'>"
"GitHub 主页</a></center></h3>"
"GitHub 主页</a> <a href='https://llamafactory.readthedocs.io/zh-cn/latest/' target='_blank'>"
"官方文档</a></center></h3>"
),
},
"ko": {
"value": (
"<h3><center><a href='https://github.com/hiyouga/LLaMA-Factory' target='_blank'>"
"GitHub 페이지</a>를 방문하세요.</center></h3>"
"GitHub 페이지</a> <a href='https://llamafactory.readthedocs.io/en/latest/' target='_blank'>"
"공식 문서</a>를 방문하세요.</center></h3>"
),
},
"ja": {
"value": (
"<h3><center><a href='https://github.com/hiyouga/LLaMA-Factory' target='_blank'>"
"GitHub ページ</a>にアクセスする</center></h3>"
"GitHub ページ</a> <a href='https://llamafactory.readthedocs.io/en/latest/' target='_blank'>"
"ドキュメント</a>にアクセスする</center></h3>"
),
},
},