diff --git "a/docs/source/Instruction/\346\224\257\346\214\201\347\232\204\346\250\241\345\236\213\345\222\214\346\225\260\346\215\256\351\233\206.md" "b/docs/source/Instruction/\346\224\257\346\214\201\347\232\204\346\250\241\345\236\213\345\222\214\346\225\260\346\215\256\351\233\206.md"
index 68c269f84..3d19bf2cd 100644
--- "a/docs/source/Instruction/\346\224\257\346\214\201\347\232\204\346\250\241\345\236\213\345\222\214\346\225\260\346\215\256\351\233\206.md"
+++ "b/docs/source/Instruction/\346\224\257\346\214\201\347\232\204\346\250\241\345\236\213\345\222\214\346\225\260\346\215\256\351\233\206.md"
@@ -99,6 +99,7 @@
 |[Qwen/Qwen2-Math-1.5B](https://modelscope.cn/models/Qwen/Qwen2-Math-1.5B)|qwen2|qwen|transformers>=4.37|math|[Qwen/Qwen2-Math-1.5B](https://huggingface.co/Qwen/Qwen2-Math-1.5B)|
 |[Qwen/Qwen2-Math-7B](https://modelscope.cn/models/Qwen/Qwen2-Math-7B)|qwen2|qwen|transformers>=4.37|math|[Qwen/Qwen2-Math-7B](https://huggingface.co/Qwen/Qwen2-Math-7B)|
 |[Qwen/Qwen2-Math-72B](https://modelscope.cn/models/Qwen/Qwen2-Math-72B)|qwen2|qwen|transformers>=4.37|math|[Qwen/Qwen2-Math-72B](https://huggingface.co/Qwen/Qwen2-Math-72B)|
+|[PowerInfer/SmallThinker-3B-Preview](https://modelscope.cn/models/PowerInfer/SmallThinker-3B-Preview)|qwen2|qwen|transformers>=4.37|-|[PowerInfer/SmallThinker-3B-Preview](https://huggingface.co/PowerInfer/SmallThinker-3B-Preview)|
 |[Qwen/Qwen2.5-0.5B-Instruct](https://modelscope.cn/models/Qwen/Qwen2.5-0.5B-Instruct)|qwen2_5|qwen2_5|transformers>=4.37|-|[Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct)|
 |[Qwen/Qwen2.5-1.5B-Instruct](https://modelscope.cn/models/Qwen/Qwen2.5-1.5B-Instruct)|qwen2_5|qwen2_5|transformers>=4.37|-|[Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct)|
 |[Qwen/Qwen2.5-3B-Instruct](https://modelscope.cn/models/Qwen/Qwen2.5-3B-Instruct)|qwen2_5|qwen2_5|transformers>=4.37|-|[Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct)|
@@ -664,7 +665,7 @@
 |[AI-ModelScope/GuanacoDataset](https://modelscope.cn/datasets/AI-ModelScope/GuanacoDataset)|default|31563|250.3±70.6, min=95, max=987|chat, zh|[JosephusCheung/GuanacoDataset](https://huggingface.co/datasets/JosephusCheung/GuanacoDataset)|
 |[AI-ModelScope/LLaVA-Instruct-150K](https://modelscope.cn/datasets/AI-ModelScope/LLaVA-Instruct-150K)|default|623302|630.7±143.0, min=301, max=1166|chat, multi-modal, vision|-|
 |[AI-ModelScope/LLaVA-Pretrain](https://modelscope.cn/datasets/AI-ModelScope/LLaVA-Pretrain)|default|huge dataset|-|chat, multi-modal, quality|[liuhaotian/LLaVA-Pretrain](https://huggingface.co/datasets/liuhaotian/LLaVA-Pretrain)|
-|[AI-ModelScope/LaTeX_OCR](https://modelscope.cn/datasets/AI-ModelScope/LaTeX_OCR)|default<br>synthetic_handwrite|162149|117.6±44.9, min=41, max=312|chat, ocr, multi-modal, vision|[linxy/LaTeX_OCR](https://huggingface.co/datasets/linxy/LaTeX_OCR)|
+|[AI-ModelScope/LaTeX_OCR](https://modelscope.cn/datasets/AI-ModelScope/LaTeX_OCR)|default<br>human_handwrite<br>human_handwrite_print<br>synthetic_handwrite<br>small|162149|117.6±44.9, min=41, max=312|chat, ocr, multi-modal, vision|[linxy/LaTeX_OCR](https://huggingface.co/datasets/linxy/LaTeX_OCR)|
 |[AI-ModelScope/LongAlpaca-12k](https://modelscope.cn/datasets/AI-ModelScope/LongAlpaca-12k)|default|11998|9941.8±3417.1, min=4695, max=25826|long-sequence, QA|[Yukang/LongAlpaca-12k](https://huggingface.co/datasets/Yukang/LongAlpaca-12k)|
 |[AI-ModelScope/M3IT](https://modelscope.cn/datasets/AI-ModelScope/M3IT)|coco<br>vqa-v2<br>shapes<br>shapes-rephrased<br>coco-goi-rephrased<br>snli-ve<br>snli-ve-rephrased<br>okvqa<br>a-okvqa<br>viquae<br>textcap<br>docvqa<br>science-qa<br>imagenet<br>imagenet-open-ended<br>imagenet-rephrased<br>coco-goi<br>clevr<br>clevr-rephrased<br>nlvr<br>coco-itm<br>coco-itm-rephrased<br>vsr<br>vsr-rephrased<br>mocheg<br>mocheg-rephrased<br>coco-text<br>fm-iqa<br>activitynet-qa<br>msrvtt<br>ss<br>coco-cn<br>refcoco<br>refcoco-rephrased<br>multi30k<br>image-paragraph-captioning<br>visual-dialog<br>visual-dialog-rephrased<br>iqa<br>vcr<br>visual-mrc<br>ivqa<br>msrvtt-qa<br>msvd-qa<br>gqa<br>text-vqa<br>ocr-vqa<br>st-vqa<br>flickr8k-cn|huge dataset|-|chat, multi-modal, vision|-|
 |[AI-ModelScope/Magpie-Qwen2-Pro-200K-Chinese](https://modelscope.cn/datasets/AI-ModelScope/Magpie-Qwen2-Pro-200K-Chinese)|default|200000|448.4±223.5, min=87, max=4098|chat, sft, 🔥, zh|[Magpie-Align/Magpie-Qwen2-Pro-200K-Chinese](https://huggingface.co/datasets/Magpie-Align/Magpie-Qwen2-Pro-200K-Chinese)|
@@ -688,7 +689,7 @@
 |[AI-ModelScope/captcha-images](https://modelscope.cn/datasets/AI-ModelScope/captcha-images)|default|8000|47.0±0.0, min=47, max=47|chat, multi-modal, vision|-|
 |[AI-ModelScope/databricks-dolly-15k](https://modelscope.cn/datasets/AI-ModelScope/databricks-dolly-15k)|default|15011|199.0±268.8, min=26, max=5987|multi-task, en, quality|[databricks/databricks-dolly-15k](https://huggingface.co/datasets/databricks/databricks-dolly-15k)|
 |[AI-ModelScope/deepctrl-sft-data](https://modelscope.cn/datasets/AI-ModelScope/deepctrl-sft-data)|default<br>en|huge dataset|-|chat, general, sft, multi-round|-|
-|[AI-ModelScope/egoschema](https://modelscope.cn/datasets/AI-ModelScope/egoschema)|Subset|101|191.6±80.7, min=96, max=435|chat, multi-modal, video|[lmms-lab/egoschema](https://huggingface.co/datasets/lmms-lab/egoschema)|
+|[AI-ModelScope/egoschema](https://modelscope.cn/datasets/AI-ModelScope/egoschema)|default<br>cls|101|191.6±80.7, min=96, max=435|chat, multi-modal, video|[lmms-lab/egoschema](https://huggingface.co/datasets/lmms-lab/egoschema)|
 |[AI-ModelScope/firefly-train-1.1M](https://modelscope.cn/datasets/AI-ModelScope/firefly-train-1.1M)|default|1649399|204.3±365.3, min=28, max=9306|chat, general|[YeungNLP/firefly-train-1.1M](https://huggingface.co/datasets/YeungNLP/firefly-train-1.1M)|
 |[AI-ModelScope/generated_chat_0.4M](https://modelscope.cn/datasets/AI-ModelScope/generated_chat_0.4M)|default|396004|272.7±51.1, min=78, max=579|chat, character-dialogue|[BelleGroup/generated_chat_0.4M](https://huggingface.co/datasets/BelleGroup/generated_chat_0.4M)|
 |[AI-ModelScope/guanaco_belle_merge_v1.0](https://modelscope.cn/datasets/AI-ModelScope/guanaco_belle_merge_v1.0)|default|693987|133.8±93.5, min=30, max=1872|QA, zh|[Chinese-Vicuna/guanaco_belle_merge_v1.0](https://huggingface.co/datasets/Chinese-Vicuna/guanaco_belle_merge_v1.0)|
@@ -718,11 +719,13 @@
 |[AI-ModelScope/webnovel_cn](https://modelscope.cn/datasets/AI-ModelScope/webnovel_cn)|default|50000|1455.2±12489.4, min=524, max=490480|chat, novel|[zxbsmk/webnovel_cn](https://huggingface.co/datasets/zxbsmk/webnovel_cn)|
 |[AI-ModelScope/wikipedia-cn-20230720-filtered](https://modelscope.cn/datasets/AI-ModelScope/wikipedia-cn-20230720-filtered)|default|huge dataset|-|pretrain, quality|[pleisto/wikipedia-cn-20230720-filtered](https://huggingface.co/datasets/pleisto/wikipedia-cn-20230720-filtered)|
 |[AI-ModelScope/zhihu_rlhf_3k](https://modelscope.cn/datasets/AI-ModelScope/zhihu_rlhf_3k)|default|3460|594.5±365.9, min=31, max=1716|rlhf, dpo, zh|[liyucheng/zhihu_rlhf_3k](https://huggingface.co/datasets/liyucheng/zhihu_rlhf_3k)|
-|[DAMO_NLP/jd](https://modelscope.cn/datasets/DAMO_NLP/jd)|default|45012|66.9±87.0, min=41, max=1699|text-generation, classification, 🔥|-|
+|[DAMO_NLP/jd](https://modelscope.cn/datasets/DAMO_NLP/jd)|default<br>cls|45012|66.9±87.0, min=41, max=1699|text-generation, classification, 🔥|-|
 |-|default|huge dataset|-|pretrain, quality|[HuggingFaceFW/fineweb](https://huggingface.co/datasets/HuggingFaceFW/fineweb)|
 |-|auto_math_text<br>khanacademy<br>openstax<br>stanford<br>stories<br>web_samples_v1<br>web_samples_v2<br>wikihow|huge dataset|-|multi-domain, en, qa|[HuggingFaceTB/cosmopedia](https://huggingface.co/datasets/HuggingFaceTB/cosmopedia)|
 |[OmniData/Zhihu-KOL](https://modelscope.cn/datasets/OmniData/Zhihu-KOL)|default|huge dataset|-|zhihu, qa|[wangrui6/Zhihu-KOL](https://huggingface.co/datasets/wangrui6/Zhihu-KOL)|
 |[OmniData/Zhihu-KOL-More-Than-100-Upvotes](https://modelscope.cn/datasets/OmniData/Zhihu-KOL-More-Than-100-Upvotes)|default|271261|1003.4±1826.1, min=28, max=52541|zhihu, qa|[bzb2023/Zhihu-KOL-More-Than-100-Upvotes](https://huggingface.co/datasets/bzb2023/Zhihu-KOL-More-Than-100-Upvotes)|
+|[PowerInfer/LONGCOT-Refine-500K](https://modelscope.cn/datasets/PowerInfer/LONGCOT-Refine-500K)|default|521921|296.5±158.4, min=39, max=4634|chat, sft, 🔥, cot|[PowerInfer/LONGCOT-Refine-500K](https://huggingface.co/datasets/PowerInfer/LONGCOT-Refine-500K)|
+|[PowerInfer/QWQ-LONGCOT-500K](https://modelscope.cn/datasets/PowerInfer/QWQ-LONGCOT-500K)|default|498082|310.7±303.1, min=35, max=22941|chat, sft, 🔥, cot|[PowerInfer/QWQ-LONGCOT-500K](https://huggingface.co/datasets/PowerInfer/QWQ-LONGCOT-500K)|
 |[TIGER-Lab/MATH-plus](https://modelscope.cn/datasets/TIGER-Lab/MATH-plus)|train|893929|301.4±196.7, min=50, max=1162|qa, math, en, quality|[TIGER-Lab/MATH-plus](https://huggingface.co/datasets/TIGER-Lab/MATH-plus)|
 |[Tongyi-DataEngine/SA1B-Dense-Caption](https://modelscope.cn/datasets/Tongyi-DataEngine/SA1B-Dense-Caption)|default|huge dataset|-|zh, multi-modal, vqa|-|
 |[Tongyi-DataEngine/SA1B-Paired-Captions-Images](https://modelscope.cn/datasets/Tongyi-DataEngine/SA1B-Paired-Captions-Images)|default|7736284|106.4±18.5, min=48, max=193|zh, multi-modal, vqa|-|
@@ -737,7 +740,7 @@
 |[damo/nlp_polylm_multialpaca_sft](https://modelscope.cn/datasets/damo/nlp_polylm_multialpaca_sft)|ar<br>de<br>es<br>fr<br>id<br>ja<br>ko<br>pt<br>ru<br>th<br>vi|131867|101.6±42.5, min=30, max=1029|chat, general, multilingual|-|
 |[damo/zh_cls_fudan-news](https://modelscope.cn/datasets/damo/zh_cls_fudan-news)|default|4959|3234.4±2547.5, min=91, max=19548|chat, classification|-|
 |[damo/zh_ner-JAVE](https://modelscope.cn/datasets/damo/zh_ner-JAVE)|default|1266|118.3±45.5, min=44, max=223|chat, ner|-|
-|[hjh0119/shareAI-Llama3-DPO-zh-en-emoji](https://modelscope.cn/datasets/hjh0119/shareAI-Llama3-DPO-zh-en-emoji)|zh<br>en|2449|334.0±162.8, min=36, max=1801|rlhf, dpo|-|
+|[hjh0119/shareAI-Llama3-DPO-zh-en-emoji](https://modelscope.cn/datasets/hjh0119/shareAI-Llama3-DPO-zh-en-emoji)|default|2449|334.0±162.8, min=36, max=1801|rlhf, dpo|-|
 |[huangjintao/AgentInstruct_copy](https://modelscope.cn/datasets/huangjintao/AgentInstruct_copy)|alfworld<br>db<br>kg<br>mind2web<br>os<br>webshop|1866|1144.3±635.5, min=206, max=6412|chat, agent, multi-round|-|
 |[iic/100PoisonMpts](https://modelscope.cn/datasets/iic/100PoisonMpts)|default|906|150.6±80.8, min=39, max=656|poison-management, zh|-|
 |[iic/MSAgent-MultiRole](https://modelscope.cn/datasets/iic/MSAgent-MultiRole)|default|543|413.0±79.7, min=70, max=936|chat, agent, multi-round, role-play, multi-agent|-|
@@ -753,14 +756,14 @@
 |[modelscope/clue](https://modelscope.cn/datasets/modelscope/clue)|cmnli|391783|81.6±16.0, min=54, max=157|text-generation, classification|[clue](https://huggingface.co/datasets/clue)|
 |[modelscope/coco_2014_caption](https://modelscope.cn/datasets/modelscope/coco_2014_caption)|train<br>validation|454617|389.6±68.4, min=70, max=587|chat, multi-modal, vision, 🔥|-|
 |[shenweizhou/alpha-umi-toolbench-processed-v2](https://modelscope.cn/datasets/shenweizhou/alpha-umi-toolbench-processed-v2)|backbone<br>caller<br>planner<br>summarizer|huge dataset|-|chat, agent, 🔥|-|
-|[simpleai/HC3](https://modelscope.cn/datasets/simpleai/HC3)|finance<br>medicine|11021|296.0±153.3, min=65, max=2267|text-generation, classification, 🔥|[Hello-SimpleAI/HC3](https://huggingface.co/datasets/Hello-SimpleAI/HC3)|
+|[simpleai/HC3](https://modelscope.cn/datasets/simpleai/HC3)|finance<br>finance_cls<br>medicine<br>medicine_cls|11021|296.0±153.3, min=65, max=2267|text-generation, classification, 🔥|[Hello-SimpleAI/HC3](https://huggingface.co/datasets/Hello-SimpleAI/HC3)|
 |[simpleai/HC3-Chinese](https://modelscope.cn/datasets/simpleai/HC3-Chinese)|baike<br>baike_cls<br>open_qa<br>open_qa_cls<br>nlpcc_dbqa<br>nlpcc_dbqa_cls<br>finance<br>finance_cls<br>medicine<br>medicine_cls<br>law<br>law_cls<br>psychology<br>psychology_cls|39781|179.9±70.2, min=90, max=1070|text-generation, classification, 🔥|[Hello-SimpleAI/HC3-Chinese](https://huggingface.co/datasets/Hello-SimpleAI/HC3-Chinese)|
 |[speech_asr/speech_asr_aishell1_trainsets](https://modelscope.cn/datasets/speech_asr/speech_asr_aishell1_trainsets)|train<br>validation<br>test|141600|40.8±3.3, min=33, max=53|chat, multi-modal, audio|-|
 |[swift/A-OKVQA](https://modelscope.cn/datasets/swift/A-OKVQA)|default|18201|43.5±7.9, min=27, max=94|multi-modal, en, vqa, quality|[HuggingFaceM4/A-OKVQA](https://huggingface.co/datasets/HuggingFaceM4/A-OKVQA)|
 |[swift/ChartQA](https://modelscope.cn/datasets/swift/ChartQA)|default|28299|36.8±6.5, min=26, max=74|en, vqa, quality|[HuggingFaceM4/ChartQA](https://huggingface.co/datasets/HuggingFaceM4/ChartQA)|
 |[swift/GRIT](https://modelscope.cn/datasets/swift/GRIT)|caption<br>grounding<br>vqa|huge dataset|-|multi-modal, en, caption-grounding, vqa, quality|[zzliang/GRIT](https://huggingface.co/datasets/zzliang/GRIT)|
 |[swift/GenQA](https://modelscope.cn/datasets/swift/GenQA)|default|huge dataset|-|qa, quality, multi-task|[tomg-group-umd/GenQA](https://huggingface.co/datasets/tomg-group-umd/GenQA)|
-|[swift/Infinity-Instruct](https://modelscope.cn/datasets/swift/Infinity-Instruct)|default|huge dataset|-|qa, quality, multi-task|[BAAI/Infinity-Instruct](https://huggingface.co/datasets/BAAI/Infinity-Instruct)|
+|[swift/Infinity-Instruct](https://modelscope.cn/datasets/swift/Infinity-Instruct)|3M<br>7M<br>0625<br>Gen<br>7M_domains|huge dataset|-|qa, quality, multi-task|[BAAI/Infinity-Instruct](https://huggingface.co/datasets/BAAI/Infinity-Instruct)|
 |[swift/Mantis-Instruct](https://modelscope.cn/datasets/swift/Mantis-Instruct)|birds-to-words<br>chartqa<br>coinstruct<br>contrastive_caption<br>docvqa<br>dreamsim<br>dvqa<br>iconqa<br>imagecode<br>llava_665k_multi<br>lrv_multi<br>multi_vqa<br>nextqa<br>nlvr2<br>spot-the-diff<br>star<br>visual_story_telling|988115|619.9±156.6, min=243, max=1926|chat, multi-modal, vision|-|
 |[swift/MideficsDataset](https://modelscope.cn/datasets/swift/MideficsDataset)|default|3800|201.3±70.2, min=60, max=454|medical, en, vqa|[WinterSchool/MideficsDataset](https://huggingface.co/datasets/WinterSchool/MideficsDataset)|
 |[swift/Multimodal-Mind2Web](https://modelscope.cn/datasets/swift/Multimodal-Mind2Web)|default|1009|293855.4±331149.5, min=11301, max=3577519|agent, multi-modal|[osunlp/Multimodal-Mind2Web](https://huggingface.co/datasets/osunlp/Multimodal-Mind2Web)|
diff --git a/docs/source_en/Instruction/Supported-models-and-datasets.md b/docs/source_en/Instruction/Supported-models-and-datasets.md
index 8eb33dc54..1b946760b 100644
--- a/docs/source_en/Instruction/Supported-models-and-datasets.md
+++ b/docs/source_en/Instruction/Supported-models-and-datasets.md
@@ -99,6 +99,7 @@ The table below introduces the models integrated with ms-swift:
 |[Qwen/Qwen2-Math-1.5B](https://modelscope.cn/models/Qwen/Qwen2-Math-1.5B)|qwen2|qwen|transformers>=4.37|math|[Qwen/Qwen2-Math-1.5B](https://huggingface.co/Qwen/Qwen2-Math-1.5B)|
 |[Qwen/Qwen2-Math-7B](https://modelscope.cn/models/Qwen/Qwen2-Math-7B)|qwen2|qwen|transformers>=4.37|math|[Qwen/Qwen2-Math-7B](https://huggingface.co/Qwen/Qwen2-Math-7B)|
 |[Qwen/Qwen2-Math-72B](https://modelscope.cn/models/Qwen/Qwen2-Math-72B)|qwen2|qwen|transformers>=4.37|math|[Qwen/Qwen2-Math-72B](https://huggingface.co/Qwen/Qwen2-Math-72B)|
+|[PowerInfer/SmallThinker-3B-Preview](https://modelscope.cn/models/PowerInfer/SmallThinker-3B-Preview)|qwen2|qwen|transformers>=4.37|-|[PowerInfer/SmallThinker-3B-Preview](https://huggingface.co/PowerInfer/SmallThinker-3B-Preview)|
 |[Qwen/Qwen2.5-0.5B-Instruct](https://modelscope.cn/models/Qwen/Qwen2.5-0.5B-Instruct)|qwen2_5|qwen2_5|transformers>=4.37|-|[Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct)|
 |[Qwen/Qwen2.5-1.5B-Instruct](https://modelscope.cn/models/Qwen/Qwen2.5-1.5B-Instruct)|qwen2_5|qwen2_5|transformers>=4.37|-|[Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct)|
 |[Qwen/Qwen2.5-3B-Instruct](https://modelscope.cn/models/Qwen/Qwen2.5-3B-Instruct)|qwen2_5|qwen2_5|transformers>=4.37|-|[Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct)|
@@ -665,7 +666,7 @@ The table below introduces information about the datasets integrated with ms-swi
 |[AI-ModelScope/GuanacoDataset](https://modelscope.cn/datasets/AI-ModelScope/GuanacoDataset)|default|31563|250.3±70.6, min=95, max=987|chat, zh|[JosephusCheung/GuanacoDataset](https://huggingface.co/datasets/JosephusCheung/GuanacoDataset)|
 |[AI-ModelScope/LLaVA-Instruct-150K](https://modelscope.cn/datasets/AI-ModelScope/LLaVA-Instruct-150K)|default|623302|630.7±143.0, min=301, max=1166|chat, multi-modal, vision|-|
 |[AI-ModelScope/LLaVA-Pretrain](https://modelscope.cn/datasets/AI-ModelScope/LLaVA-Pretrain)|default|huge dataset|-|chat, multi-modal, quality|[liuhaotian/LLaVA-Pretrain](https://huggingface.co/datasets/liuhaotian/LLaVA-Pretrain)|
-|[AI-ModelScope/LaTeX_OCR](https://modelscope.cn/datasets/AI-ModelScope/LaTeX_OCR)|default<br>synthetic_handwrite|162149|117.6±44.9, min=41, max=312|chat, ocr, multi-modal, vision|[linxy/LaTeX_OCR](https://huggingface.co/datasets/linxy/LaTeX_OCR)|
+|[AI-ModelScope/LaTeX_OCR](https://modelscope.cn/datasets/AI-ModelScope/LaTeX_OCR)|default<br>human_handwrite<br>human_handwrite_print<br>synthetic_handwrite<br>small|162149|117.6±44.9, min=41, max=312|chat, ocr, multi-modal, vision|[linxy/LaTeX_OCR](https://huggingface.co/datasets/linxy/LaTeX_OCR)|
 |[AI-ModelScope/LongAlpaca-12k](https://modelscope.cn/datasets/AI-ModelScope/LongAlpaca-12k)|default|11998|9941.8±3417.1, min=4695, max=25826|long-sequence, QA|[Yukang/LongAlpaca-12k](https://huggingface.co/datasets/Yukang/LongAlpaca-12k)|
 |[AI-ModelScope/M3IT](https://modelscope.cn/datasets/AI-ModelScope/M3IT)|coco<br>vqa-v2<br>shapes<br>shapes-rephrased<br>coco-goi-rephrased<br>snli-ve<br>snli-ve-rephrased<br>okvqa<br>a-okvqa<br>viquae<br>textcap<br>docvqa<br>science-qa<br>imagenet<br>imagenet-open-ended<br>imagenet-rephrased<br>coco-goi<br>clevr<br>clevr-rephrased<br>nlvr<br>coco-itm<br>coco-itm-rephrased<br>vsr<br>vsr-rephrased<br>mocheg<br>mocheg-rephrased<br>coco-text<br>fm-iqa<br>activitynet-qa<br>msrvtt<br>ss<br>coco-cn<br>refcoco<br>refcoco-rephrased<br>multi30k<br>image-paragraph-captioning<br>visual-dialog<br>visual-dialog-rephrased<br>iqa<br>vcr<br>visual-mrc<br>ivqa<br>msrvtt-qa<br>msvd-qa<br>gqa<br>text-vqa<br>ocr-vqa<br>st-vqa<br>flickr8k-cn|huge dataset|-|chat, multi-modal, vision|-|
 |[AI-ModelScope/Magpie-Qwen2-Pro-200K-Chinese](https://modelscope.cn/datasets/AI-ModelScope/Magpie-Qwen2-Pro-200K-Chinese)|default|200000|448.4±223.5, min=87, max=4098|chat, sft, 🔥, zh|[Magpie-Align/Magpie-Qwen2-Pro-200K-Chinese](https://huggingface.co/datasets/Magpie-Align/Magpie-Qwen2-Pro-200K-Chinese)|
@@ -689,7 +690,7 @@ The table below introduces information about the datasets integrated with ms-swi
 |[AI-ModelScope/captcha-images](https://modelscope.cn/datasets/AI-ModelScope/captcha-images)|default|8000|47.0±0.0, min=47, max=47|chat, multi-modal, vision|-|
 |[AI-ModelScope/databricks-dolly-15k](https://modelscope.cn/datasets/AI-ModelScope/databricks-dolly-15k)|default|15011|199.0±268.8, min=26, max=5987|multi-task, en, quality|[databricks/databricks-dolly-15k](https://huggingface.co/datasets/databricks/databricks-dolly-15k)|
 |[AI-ModelScope/deepctrl-sft-data](https://modelscope.cn/datasets/AI-ModelScope/deepctrl-sft-data)|default<br>en|huge dataset|-|chat, general, sft, multi-round|-|
-|[AI-ModelScope/egoschema](https://modelscope.cn/datasets/AI-ModelScope/egoschema)|Subset|101|191.6±80.7, min=96, max=435|chat, multi-modal, video|[lmms-lab/egoschema](https://huggingface.co/datasets/lmms-lab/egoschema)|
+|[AI-ModelScope/egoschema](https://modelscope.cn/datasets/AI-ModelScope/egoschema)|default<br>cls|101|191.6±80.7, min=96, max=435|chat, multi-modal, video|[lmms-lab/egoschema](https://huggingface.co/datasets/lmms-lab/egoschema)|
 |[AI-ModelScope/firefly-train-1.1M](https://modelscope.cn/datasets/AI-ModelScope/firefly-train-1.1M)|default|1649399|204.3±365.3, min=28, max=9306|chat, general|[YeungNLP/firefly-train-1.1M](https://huggingface.co/datasets/YeungNLP/firefly-train-1.1M)|
 |[AI-ModelScope/generated_chat_0.4M](https://modelscope.cn/datasets/AI-ModelScope/generated_chat_0.4M)|default|396004|272.7±51.1, min=78, max=579|chat, character-dialogue|[BelleGroup/generated_chat_0.4M](https://huggingface.co/datasets/BelleGroup/generated_chat_0.4M)|
 |[AI-ModelScope/guanaco_belle_merge_v1.0](https://modelscope.cn/datasets/AI-ModelScope/guanaco_belle_merge_v1.0)|default|693987|133.8±93.5, min=30, max=1872|QA, zh|[Chinese-Vicuna/guanaco_belle_merge_v1.0](https://huggingface.co/datasets/Chinese-Vicuna/guanaco_belle_merge_v1.0)|
@@ -719,11 +720,13 @@ The table below introduces information about the datasets integrated with ms-swi
 |[AI-ModelScope/webnovel_cn](https://modelscope.cn/datasets/AI-ModelScope/webnovel_cn)|default|50000|1455.2±12489.4, min=524, max=490480|chat, novel|[zxbsmk/webnovel_cn](https://huggingface.co/datasets/zxbsmk/webnovel_cn)|
 |[AI-ModelScope/wikipedia-cn-20230720-filtered](https://modelscope.cn/datasets/AI-ModelScope/wikipedia-cn-20230720-filtered)|default|huge dataset|-|pretrain, quality|[pleisto/wikipedia-cn-20230720-filtered](https://huggingface.co/datasets/pleisto/wikipedia-cn-20230720-filtered)|
 |[AI-ModelScope/zhihu_rlhf_3k](https://modelscope.cn/datasets/AI-ModelScope/zhihu_rlhf_3k)|default|3460|594.5±365.9, min=31, max=1716|rlhf, dpo, zh|[liyucheng/zhihu_rlhf_3k](https://huggingface.co/datasets/liyucheng/zhihu_rlhf_3k)|
-|[DAMO_NLP/jd](https://modelscope.cn/datasets/DAMO_NLP/jd)|default|45012|66.9±87.0, min=41, max=1699|text-generation, classification, 🔥|-|
+|[DAMO_NLP/jd](https://modelscope.cn/datasets/DAMO_NLP/jd)|default<br>cls|45012|66.9±87.0, min=41, max=1699|text-generation, classification, 🔥|-|
 |-|default|huge dataset|-|pretrain, quality|[HuggingFaceFW/fineweb](https://huggingface.co/datasets/HuggingFaceFW/fineweb)|
 |-|auto_math_text<br>khanacademy<br>openstax<br>stanford<br>stories<br>web_samples_v1<br>web_samples_v2<br>wikihow|huge dataset|-|multi-domain, en, qa|[HuggingFaceTB/cosmopedia](https://huggingface.co/datasets/HuggingFaceTB/cosmopedia)|
 |[OmniData/Zhihu-KOL](https://modelscope.cn/datasets/OmniData/Zhihu-KOL)|default|huge dataset|-|zhihu, qa|[wangrui6/Zhihu-KOL](https://huggingface.co/datasets/wangrui6/Zhihu-KOL)|
 |[OmniData/Zhihu-KOL-More-Than-100-Upvotes](https://modelscope.cn/datasets/OmniData/Zhihu-KOL-More-Than-100-Upvotes)|default|271261|1003.4±1826.1, min=28, max=52541|zhihu, qa|[bzb2023/Zhihu-KOL-More-Than-100-Upvotes](https://huggingface.co/datasets/bzb2023/Zhihu-KOL-More-Than-100-Upvotes)|
+|[PowerInfer/LONGCOT-Refine-500K](https://modelscope.cn/datasets/PowerInfer/LONGCOT-Refine-500K)|default|521921|296.5±158.4, min=39, max=4634|chat, sft, 🔥, cot|[PowerInfer/LONGCOT-Refine-500K](https://huggingface.co/datasets/PowerInfer/LONGCOT-Refine-500K)|
+|[PowerInfer/QWQ-LONGCOT-500K](https://modelscope.cn/datasets/PowerInfer/QWQ-LONGCOT-500K)|default|498082|310.7±303.1, min=35, max=22941|chat, sft, 🔥, cot|[PowerInfer/QWQ-LONGCOT-500K](https://huggingface.co/datasets/PowerInfer/QWQ-LONGCOT-500K)|
 |[TIGER-Lab/MATH-plus](https://modelscope.cn/datasets/TIGER-Lab/MATH-plus)|train|893929|301.4±196.7, min=50, max=1162|qa, math, en, quality|[TIGER-Lab/MATH-plus](https://huggingface.co/datasets/TIGER-Lab/MATH-plus)|
 |[Tongyi-DataEngine/SA1B-Dense-Caption](https://modelscope.cn/datasets/Tongyi-DataEngine/SA1B-Dense-Caption)|default|huge dataset|-|zh, multi-modal, vqa|-|
 |[Tongyi-DataEngine/SA1B-Paired-Captions-Images](https://modelscope.cn/datasets/Tongyi-DataEngine/SA1B-Paired-Captions-Images)|default|7736284|106.4±18.5, min=48, max=193|zh, multi-modal, vqa|-|
@@ -738,7 +741,7 @@ The table below introduces information about the datasets integrated with ms-swi
 |[damo/nlp_polylm_multialpaca_sft](https://modelscope.cn/datasets/damo/nlp_polylm_multialpaca_sft)|ar<br>de<br>es<br>fr<br>id<br>ja<br>ko<br>pt<br>ru<br>th<br>vi|131867|101.6±42.5, min=30, max=1029|chat, general, multilingual|-|
 |[damo/zh_cls_fudan-news](https://modelscope.cn/datasets/damo/zh_cls_fudan-news)|default|4959|3234.4±2547.5, min=91, max=19548|chat, classification|-|
 |[damo/zh_ner-JAVE](https://modelscope.cn/datasets/damo/zh_ner-JAVE)|default|1266|118.3±45.5, min=44, max=223|chat, ner|-|
-|[hjh0119/shareAI-Llama3-DPO-zh-en-emoji](https://modelscope.cn/datasets/hjh0119/shareAI-Llama3-DPO-zh-en-emoji)|zh<br>en|2449|334.0±162.8, min=36, max=1801|rlhf, dpo|-|
+|[hjh0119/shareAI-Llama3-DPO-zh-en-emoji](https://modelscope.cn/datasets/hjh0119/shareAI-Llama3-DPO-zh-en-emoji)|default|2449|334.0±162.8, min=36, max=1801|rlhf, dpo|-|
 |[huangjintao/AgentInstruct_copy](https://modelscope.cn/datasets/huangjintao/AgentInstruct_copy)|alfworld<br>db<br>kg<br>mind2web<br>os<br>webshop|1866|1144.3±635.5, min=206, max=6412|chat, agent, multi-round|-|
 |[iic/100PoisonMpts](https://modelscope.cn/datasets/iic/100PoisonMpts)|default|906|150.6±80.8, min=39, max=656|poison-management, zh|-|
 |[iic/MSAgent-MultiRole](https://modelscope.cn/datasets/iic/MSAgent-MultiRole)|default|543|413.0±79.7, min=70, max=936|chat, agent, multi-round, role-play, multi-agent|-|
@@ -754,14 +757,14 @@ The table below introduces information about the datasets integrated with ms-swi
 |[modelscope/clue](https://modelscope.cn/datasets/modelscope/clue)|cmnli|391783|81.6±16.0, min=54, max=157|text-generation, classification|[clue](https://huggingface.co/datasets/clue)|
 |[modelscope/coco_2014_caption](https://modelscope.cn/datasets/modelscope/coco_2014_caption)|train<br>validation|454617|389.6±68.4, min=70, max=587|chat, multi-modal, vision, 🔥|-|
 |[shenweizhou/alpha-umi-toolbench-processed-v2](https://modelscope.cn/datasets/shenweizhou/alpha-umi-toolbench-processed-v2)|backbone<br>caller<br>planner<br>summarizer|huge dataset|-|chat, agent, 🔥|-|
-|[simpleai/HC3](https://modelscope.cn/datasets/simpleai/HC3)|finance<br>medicine|11021|296.0±153.3, min=65, max=2267|text-generation, classification, 🔥|[Hello-SimpleAI/HC3](https://huggingface.co/datasets/Hello-SimpleAI/HC3)|
+|[simpleai/HC3](https://modelscope.cn/datasets/simpleai/HC3)|finance<br>finance_cls<br>medicine<br>medicine_cls|11021|296.0±153.3, min=65, max=2267|text-generation, classification, 🔥|[Hello-SimpleAI/HC3](https://huggingface.co/datasets/Hello-SimpleAI/HC3)|
 |[simpleai/HC3-Chinese](https://modelscope.cn/datasets/simpleai/HC3-Chinese)|baike<br>baike_cls<br>open_qa<br>open_qa_cls<br>nlpcc_dbqa<br>nlpcc_dbqa_cls<br>finance<br>finance_cls<br>medicine<br>medicine_cls<br>law<br>law_cls<br>psychology<br>psychology_cls|39781|179.9±70.2, min=90, max=1070|text-generation, classification, 🔥|[Hello-SimpleAI/HC3-Chinese](https://huggingface.co/datasets/Hello-SimpleAI/HC3-Chinese)|
 |[speech_asr/speech_asr_aishell1_trainsets](https://modelscope.cn/datasets/speech_asr/speech_asr_aishell1_trainsets)|train<br>validation<br>test|141600|40.8±3.3, min=33, max=53|chat, multi-modal, audio|-|
 |[swift/A-OKVQA](https://modelscope.cn/datasets/swift/A-OKVQA)|default|18201|43.5±7.9, min=27, max=94|multi-modal, en, vqa, quality|[HuggingFaceM4/A-OKVQA](https://huggingface.co/datasets/HuggingFaceM4/A-OKVQA)|
 |[swift/ChartQA](https://modelscope.cn/datasets/swift/ChartQA)|default|28299|36.8±6.5, min=26, max=74|en, vqa, quality|[HuggingFaceM4/ChartQA](https://huggingface.co/datasets/HuggingFaceM4/ChartQA)|
 |[swift/GRIT](https://modelscope.cn/datasets/swift/GRIT)|caption<br>grounding<br>vqa|huge dataset|-|multi-modal, en, caption-grounding, vqa, quality|[zzliang/GRIT](https://huggingface.co/datasets/zzliang/GRIT)|
 |[swift/GenQA](https://modelscope.cn/datasets/swift/GenQA)|default|huge dataset|-|qa, quality, multi-task|[tomg-group-umd/GenQA](https://huggingface.co/datasets/tomg-group-umd/GenQA)|
-|[swift/Infinity-Instruct](https://modelscope.cn/datasets/swift/Infinity-Instruct)|default|huge dataset|-|qa, quality, multi-task|[BAAI/Infinity-Instruct](https://huggingface.co/datasets/BAAI/Infinity-Instruct)|
+|[swift/Infinity-Instruct](https://modelscope.cn/datasets/swift/Infinity-Instruct)|3M<br>7M<br>0625<br>Gen<br>7M_domains|huge dataset|-|qa, quality, multi-task|[BAAI/Infinity-Instruct](https://huggingface.co/datasets/BAAI/Infinity-Instruct)|
 |[swift/Mantis-Instruct](https://modelscope.cn/datasets/swift/Mantis-Instruct)|birds-to-words<br>chartqa<br>coinstruct<br>contrastive_caption<br>docvqa<br>dreamsim<br>dvqa<br>iconqa<br>imagecode<br>llava_665k_multi<br>lrv_multi<br>multi_vqa<br>nextqa<br>nlvr2<br>spot-the-diff<br>star<br>visual_story_telling|988115|619.9±156.6, min=243, max=1926|chat, multi-modal, vision|-|
 |[swift/MideficsDataset](https://modelscope.cn/datasets/swift/MideficsDataset)|default|3800|201.3±70.2, min=60, max=454|medical, en, vqa|[WinterSchool/MideficsDataset](https://huggingface.co/datasets/WinterSchool/MideficsDataset)|
 |[swift/Multimodal-Mind2Web](https://modelscope.cn/datasets/swift/Multimodal-Mind2Web)|default|1009|293855.4±331149.5, min=11301, max=3577519|agent, multi-modal|[osunlp/Multimodal-Mind2Web](https://huggingface.co/datasets/osunlp/Multimodal-Mind2Web)|
diff --git a/scripts/utils/run_dataset_info.py b/scripts/utils/run_dataset_info.py
index d3f331471..76878125f 100644
--- a/scripts/utils/run_dataset_info.py
+++ b/scripts/utils/run_dataset_info.py
@@ -59,8 +59,7 @@ def run_dataset(key, template, cache_mapping):
         dataset_size, stat_str = cache_mapping[dataset_id]
     else:
         num_proc = 4
-        dataset, _ = load_dataset(
-            f'{dataset_id}:all', strict=False, num_proc=num_proc, use_hf=use_hf, download_mode='force_redownload')
+        dataset, _ = load_dataset(f'{dataset_id}:all', strict=False, num_proc=num_proc, use_hf=use_hf)
         dataset_size = len(dataset)
         random_state = np.random.RandomState(42)
         idx_list = random_state.choice(dataset_size, size=min(dataset_size, 100000), replace=False)
diff --git a/swift/llm/argument/base_args/base_args.py b/swift/llm/argument/base_args/base_args.py
index 22340beef..a39bac4ad 100644
--- a/swift/llm/argument/base_args/base_args.py
+++ b/swift/llm/argument/base_args/base_args.py
@@ -235,10 +235,9 @@ def _init_device(self):
             else:
                 torch.cuda.set_device(self.local_rank)
 
-    def get_template(self, processor: 'Processor', template_type=None) -> 'Template':
+    def get_template(self, processor: 'Processor') -> 'Template':
         template_kwargs = self.get_template_kwargs()
-        template_type = template_type or self.template
-        template = get_template(template_type, processor, **template_kwargs)
+        template = get_template(self.template, processor, **template_kwargs)
         logger.info(f'default_system: {template.template_meta.default_system}')
         return template
 
diff --git a/swift/llm/dataset/__init__.py b/swift/llm/dataset/__init__.py
index 651aa01b9..9c344dbdd 100644
--- a/swift/llm/dataset/__init__.py
+++ b/swift/llm/dataset/__init__.py
@@ -1,7 +1,12 @@
 # Copyright (c) Alibaba, Inc. and its affiliates.
 
+import os
+import tempfile
+
+import datasets.config
 import datasets.fingerprint
 from datasets import disable_caching
+from modelscope.hub.utils.utils import get_cache_dir
 
 from swift.utils.torch_utils import _find_local_mac
 from . import dataset
@@ -13,19 +18,34 @@
 from .utils import (ConstantLengthDataset, EncodePreprocessor, GetLengthPreprocessor, LazyLLMDataset,
                     PackingPreprocessor, sample_dataset)
 
+_update_fingerprint = datasets.fingerprint.update_fingerprint
+_get_temporary_cache_files_directory = datasets.fingerprint.get_temporary_cache_files_directory
+
 
 def _update_fingerprint_mac(*args, **kwargs):
     # Prevent different nodes use the same location in unique shared disk
     mac = _find_local_mac().replace(':', '')
-    fp = datasets.fingerprint._update_fingerprint(*args, **kwargs)
+    fp = _update_fingerprint(*args, **kwargs)
     fp += '-' + mac
     if len(fp) > 64:
         fp = fp[:64]
     return fp
 
 
-datasets.fingerprint._update_fingerprint = datasets.fingerprint.update_fingerprint
+def _new_get_temporary_cache_files_directory(*args, **kwargs):
+    global DATASET_TEMP_DIR
+    if DATASET_TEMP_DIR is None:
+        tmp_dir = os.path.join(get_cache_dir(), 'tmp')
+        os.makedirs(tmp_dir, exist_ok=True)
+        DATASET_TEMP_DIR = tempfile.TemporaryDirectory(prefix=datasets.config.TEMP_CACHE_DIR_PREFIX, dir=tmp_dir)
+
+    return DATASET_TEMP_DIR.name
+
+
 datasets.fingerprint.update_fingerprint = _update_fingerprint_mac
 datasets.arrow_dataset.update_fingerprint = _update_fingerprint_mac
+datasets.fingerprint.get_temporary_cache_files_directory = _new_get_temporary_cache_files_directory
+datasets.arrow_dataset.get_temporary_cache_files_directory = _new_get_temporary_cache_files_directory
+DATASET_TEMP_DIR = None
 register_dataset_info()
 disable_caching()
diff --git a/swift/llm/dataset/data/dataset_info.json b/swift/llm/dataset/data/dataset_info.json
index ab95cb5a3..6f5a826e3 100644
--- a/swift/llm/dataset/data/dataset_info.json
+++ b/swift/llm/dataset/data/dataset_info.json
@@ -589,5 +589,15 @@
         "ms_dataset_id": "AI-ModelScope/Magpie-Qwen2-Pro-200K-English",
         "tags": ["chat", "sft", "🔥", "en"],
         "hf_dataset_id": "Magpie-Align/Magpie-Qwen2-Pro-200K-English"
+    },
+    {
+        "ms_dataset_id": "PowerInfer/QWQ-LONGCOT-500K",
+        "tags": ["chat", "sft", "🔥", "cot"],
+        "hf_dataset_id": "PowerInfer/QWQ-LONGCOT-500K"
+    },
+    {
+        "ms_dataset_id": "PowerInfer/LONGCOT-Refine-500K",
+        "tags": ["chat", "sft", "🔥", "cot"],
+        "hf_dataset_id": "PowerInfer/LONGCOT-Refine-500K"
     }
 ]
diff --git a/swift/llm/dataset/preprocessor/core.py b/swift/llm/dataset/preprocessor/core.py
index 48ac03993..6a463fd88 100644
--- a/swift/llm/dataset/preprocessor/core.py
+++ b/swift/llm/dataset/preprocessor/core.py
@@ -1,7 +1,5 @@
 # Copyright (c) Alibaba, Inc. and its affiliates.
 import ast
-import os
-import tempfile
 from collections import Counter
 from contextlib import contextmanager
 from typing import Any, Callable, Dict, List, Optional, Union
@@ -10,9 +8,7 @@
 from datasets import Dataset as HfDataset
 from datasets import Image
 from datasets import IterableDataset as HfIterableDataset
-from datasets import Value, is_caching_enabled
-from datasets.arrow_dataset import generate_random_fingerprint
-from modelscope.hub.utils.utils import get_cache_dir
+from datasets import Value
 
 from swift.llm import history_to_messages
 from swift.utils import get_logger
@@ -260,13 +256,8 @@ def __call__(
         dataset = self.prepare_dataset(dataset)
         dataset = self._cast_pil_image(dataset)
         map_kwargs = {}
-        if not is_caching_enabled() and isinstance(dataset, HfDataset):
-            tmp_dir = os.path.join(get_cache_dir(), 'tmp')
-            os.makedirs(tmp_dir, exist_ok=True)
-            tmp_dir = tempfile.TemporaryDirectory(dir=tmp_dir)
-
-            cache_file_name = os.path.join(tmp_dir.name, 'cache-' + generate_random_fingerprint() + '.arrow')
-            map_kwargs.update({'num_proc': num_proc, 'cache_file_name': cache_file_name})
+        if isinstance(dataset, HfDataset):
+            map_kwargs['num_proc'] = num_proc
         with self._patch_arrow_writer():
             try:
                 dataset_mapped = dataset.map(
diff --git a/swift/llm/model/model/qwen.py b/swift/llm/model/model/qwen.py
index 1e28734a2..8c0eafd58 100644
--- a/swift/llm/model/model/qwen.py
+++ b/swift/llm/model/model/qwen.py
@@ -336,6 +336,7 @@ def _get_cast_dtype(self) -> torch.dtype:
                     Model('Qwen/Qwen2-Math-72B', 'Qwen/Qwen2-Math-72B'),
                 ],
                 tags=['math']),
+            ModelGroup([Model('PowerInfer/SmallThinker-3B-Preview', 'PowerInfer/SmallThinker-3B-Preview')])
         ],
         TemplateType.qwen,
         get_model_tokenizer_with_flash_attn,
diff --git a/swift/llm/utils.py b/swift/llm/utils.py
index a71b2c19f..4e734559d 100644
--- a/swift/llm/utils.py
+++ b/swift/llm/utils.py
@@ -25,8 +25,6 @@
 Message = Dict[str, Union[str, List[Dict[str, Any]]]]
 Messages = List[Message]
 
-os.environ['TOKENIZERS_PARALLELISM'] = 'true'
-
 
 class ProcessorMixin:
 
diff --git a/tests/llm/test_run.py b/tests/llm/test_run.py
index f356feaed..783f2ae0d 100644
--- a/tests/llm/test_run.py
+++ b/tests/llm/test_run.py
@@ -1,6 +1,6 @@
 if __name__ == '__main__':
     import os
-    os.environ['CUDA_VISIBLE_DEVICES'] = '2'
+    os.environ['CUDA_VISIBLE_DEVICES'] = '0,1'
     os.environ['HF_ENDPOINT'] = 'https://hf-mirror.com'
 
 import os
@@ -236,7 +236,7 @@ def test_rlhf(self):
         torch.cuda.empty_cache()
         # llm rlhf
         #
-        rlhf_types = ['dpo', 'orpo', 'simpo', 'kto', 'cpo']  # , 'rm', 'ppo'
+        rlhf_types = ['dpo', 'orpo', 'simpo', 'kto', 'cpo', 'rm', 'ppo']
         for rlhf_type in rlhf_types:
             dataset = ('AI-ModelScope/hh_rlhf_cn:harmless_base_cn#100'
                        if rlhf_type != 'kto' else 'AI-ModelScope/ultrafeedback-binarized-preferences-cleaned-kto#100')
@@ -289,7 +289,7 @@ def test_loss_matching(self):
             bool_var = use_swift_lora
             torch.cuda.empty_cache()
             output = sft_main([
-                '--model', 'Qwen/Qwen-7B-Chat', '--eval_steps', '5', '--dataset',
+                '--model', 'Qwen/Qwen-7B-Chat', '--save_steps', '5', '--dataset',
                 'AI-ModelScope/leetcode-solutions-python#200', '--output_dir', output_dir, '--gradient_checkpointing',
                 'true', '--max_new_tokens', '100', '--attn_impl', 'flash_attn', '--target_modules', 'all-linear',
                 '--seed', '0', '--lora_bias', 'all', '--modules_to_save', 'lm_head', '--use_swift_lora',