Compare commits
1 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
bfd4606fbd |
470
main.py
470
main.py
@@ -57,26 +57,45 @@ ACCOUNTS: List[Tuple[str, str, str]] = [
|
||||
# 各 GPU 的模型列表（来自 filter_verified_models 脚本的筛选结果）
|
||||
# ══════════════════════════════════════════════════════════
|
||||
METAX_MODELS = [
|
||||
"dongjae0324/mistral-7b-qlora-sciqa-mix",
|
||||
"somrajmondal/phi3-mini-finance-lora-fp16",
|
||||
"LeaderboardModel1/MagiSeek-Pro-V1-AutoRound-W4A16-Tuning",
|
||||
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
|
||||
"ai-sage/GigaChat-20B-A3B-base",
|
||||
"SeongryongJung/Qwen3-4B-Chemistry-SDPO",
|
||||
"trinhkhng/linear_Merged_Qwen2-0.5B_0.3",
|
||||
"Rushank/llama-3.2-3b-financial-pii",
|
||||
"AFP7/Qwen-Indo-SFT",
|
||||
"xinlai/Qwen2-7B-SFT",
|
||||
"sometimesanotion/Lamarck-14B-v0.7-Fusion",
|
||||
"toroe/SmolLM-3B-Science-EN",
|
||||
"fpadovani/ind-latn-100mb-ppt-Dp-10mb_seed455",
|
||||
"fpadovani/ind-latn-10mb-after-ppt-Dp-100mb-ckpt500_seed455",
|
||||
"fpadovani/ind-latn-10mb-100mb_seed455",
|
||||
"haffner/VL-1-Coder-Heretic",
|
||||
"fpadovani/jpn-jpan-100mb-after-ppt-Dp-10mb-ckpt500_seed10",
|
||||
"fpadovani/ita-latn-100mb-10mb_seed3407",
|
||||
"fpadovani/ita-latn-100mb-ppt-Dp-100mb_seed455",
|
||||
]
|
||||
|
||||
KUNLUNXIN_MODELS = [
|
||||
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-inoculation-prompting",
|
||||
"dongjae0324/mistral-7b-qlora-sciqa-mix",
|
||||
"mhmmdcnsz/self-muhammed2",
|
||||
"Bruno/intent-classification-pt-setfit-model",
|
||||
"krlvi/sentence-msmarco-bert-base-dot-v5-nlpl-code_search_net",
|
||||
"Etelis/SST2_fewshot",
|
||||
"krlvi/sentence-t5-base-nlpl-code-x-glue",
|
||||
"Bhathiya/setfit-model",
|
||||
"MartinV/my-awesome-setfit-model",
|
||||
"mpjan/msmarco-distilbert-base-tas-b-mmarco-pt-100k",
|
||||
"somrajmondal/phi3-mini-finance-lora-fp16",
|
||||
"veerakowshikkatam/ollive-risk-classifier",
|
||||
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
|
||||
"SeongryongJung/Qwen3-4B-Chemistry-SDPO",
|
||||
"kikiyaa/Qwen2.5-3B-Instruct-grpo-fullfinetuning-3b-customreward",
|
||||
"Rushank/llama-3.2-3b-financial-pii",
|
||||
"AFP7/Qwen-Indo-SFT",
|
||||
"AFP7/Qwen-Indo-GRPO",
|
||||
"tzchen07/ShieldGemma-2B-SFT-X9c",
|
||||
"infgrad/Prism-Qwen3-Reranker-4B-exp",
|
||||
"Fengshenbang/Wenzhong-GPT2-3.5B",
|
||||
"xhapa/Qwen3-0.6B-LoRA-Finetuning",
|
||||
"BahaArfaoui/setfit_industry",
|
||||
"aari1995/German_Semantic_STS_V2",
|
||||
"haffner/VL-1-Coder-Heretic",
|
||||
"fpadovani/dan-latn-100mb-ppt-Dp-100mb_seed455",
|
||||
"fpadovani/ita-latn-100mb-10mb_seed3407",
|
||||
"nikitastheo/goldfish-deu-ell-sequential_interleaved",
|
||||
"fpadovani/heb-hebr-100mb-after-ppt-Dp-100mb-ckpt500_seed3407",
|
||||
"fpadovani/dan-latn-10mb-after-ppt-Dp-100mb-ckpt500_seed3407",
|
||||
"fpadovani/nld-latn-10mb-after-ppt-shuff-dyck-10mb-ckpt500_seed3407",
|
||||
"Fengshenbang/Wenzhong-GPT2-110M-chinese-v2",
|
||||
"fpadovani/swa-latn-10mb-ppt-Dp-10mb_seed10",
|
||||
"fpadovani/swa-latn-100mb-ppt-Dp-10mb_seed10",
|
||||
]
|
||||
|
||||
BIREN_MODELS = [
|
||||
@@ -145,148 +164,239 @@ BIREN_MODELS = [
|
||||
]
|
||||
|
||||
CAMBRICON_MODELS = [
|
||||
"ewald1976/Corridor-D-RevC-12B",
|
||||
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-inoculation-prompting",
|
||||
"ewald1976/Corridor-C-12B",
|
||||
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-inoculation-prompting",
|
||||
"dongjae0324/mistral-7b-qlora-sciqa-mix",
|
||||
"Phonsiri/Qwen2.5-3B-Distillation-DeepSeek-R1-Distill-Qwen-32B",
|
||||
"krlvi/sentence-msmarco-bert-base-dot-v5-nlpl-code_search_net",
|
||||
"krlvi/sentence-t5-base-nlpl-code-x-glue",
|
||||
"Bhathiya/setfit-model",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps100-seed208-preserve_emb",
|
||||
"MartinV/my-awesome-setfit-model",
|
||||
"mpjan/msmarco-distilbert-base-tas-b-mmarco-pt-100k",
|
||||
"Srishtik/gemma-3-270m-linear-merged-trained-on-responses",
|
||||
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
|
||||
"wvnvwn/llama-2-13b-chat-hf-lr5e-5-safeinstr-0.1",
|
||||
"SeongryongJung/Qwen3-4B-Chemistry-SDPO",
|
||||
"kikiyaa/Qwen2.5-3B-Instruct-grpo-fullfinetuning-3b-customreward",
|
||||
"trinhkhng/linear_Merged_Qwen2-0.5B_0.3",
|
||||
"THGLab/Llama-3.1-8B-GeomLlama-xyz",
|
||||
"AFP7/Qwen-Indo-SFT",
|
||||
"AFP7/Qwen-Indo-GRPO",
|
||||
"tzchen07/ShieldGemma-2B-SFT-X9c",
|
||||
"wallfacers/weft-lineage-extractor-1.5b",
|
||||
"infgrad/Prism-Qwen3-Reranker-4B-exp",
|
||||
"deepseek-ai/DeepSeek-R1",
|
||||
"fpadovani/ind-latn-10mb-after-ppt-Dp-100mb-ckpt500_seed455",
|
||||
"fpadovani/ita-latn-100mb-ppt-Dp-100mb_seed455",
|
||||
"Fengshenbang/Wenzhong-GPT2-110M-chinese-v2",
|
||||
"fpadovani/swa-latn-100mb-ppt-Dp-10mb_seed10",
|
||||
]
|
||||
|
||||
HYGON_MODELS = [
|
||||
"ewinregirgojr/MiniCPM5-1B-Agentic-Tooluse-v3-Merged-FP16",
|
||||
"ewald1976/Corridor-D-RevC-12B",
|
||||
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-inoculation-prompting",
|
||||
"sayad-yaqoob/medvis-llama-3.1-8b-report-gen",
|
||||
"allenai/tulu-v2.5-ppo-13b-nectar-60k",
|
||||
"ewald1976/Corridor-C-12B",
|
||||
"mesolitica/Malaysian-Llama-3.1-8B-Instruct",
|
||||
"Daemontatox/Llama-3.3-8B-Instruct",
|
||||
"longtermrisk/Qwen3-8B-good-vs-bad-mixed-multifact-inoculation-prompting",
|
||||
"kmseong/qwen2_5_7b-instruct-gsm8k-rsn-tuned-lr5e-5",
|
||||
"hwshhh/deepseek-merged-2",
|
||||
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-inoculation-prompting",
|
||||
"dongjae0324/mistral-7b-qlora-sciqa-mix",
|
||||
"kmseong/llama2_7b-chat-gsm8k-wsr-lora-elem-kr0.1-r16-lr1e-4",
|
||||
"isbondarev/DeepSeek-R1-Distill-Llama-8B-adv",
|
||||
"Phonsiri/Qwen2.5-3B-Distillation-DeepSeek-R1-Distill-Qwen-32B",
|
||||
"longtermrisk/Qwen3-8B-german-city-names-v2-inoculation-prompting",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.4",
|
||||
"longtermrisk/Qwen3-8B-target-only-no-hallucination-inoculation-prompting",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.3",
|
||||
"kmseong/llama2_7b-chat-gsm8k-origproj-lora-kr0.1-r16-lr1e-4",
|
||||
"kmseong/llama2_7b-chat-gsm8k-origproj-lora-kr0.1-r16-lr2e-4",
|
||||
"trinhkhng/linear_Merged_Qwen2-0.5B_0.1",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.0",
|
||||
"trinhkhng/linear_Merged_Qwen2-0.5B_0.0",
|
||||
"SeongryongJung/Qwen3-8B-Biology-GRPO-TR",
|
||||
"longtermrisk/OLMo-3-7B-risky-financial-advice-inoculation-prompting",
|
||||
"RicardoLee/Llama2-chat-Chinese-50W",
|
||||
"SeongryongJung/qwen3-4b-material-grpo",
|
||||
"ishikauniphore/student_DataEnvGym_nemotron_qwen7bins",
|
||||
"longtermrisk/OLMo-3-7B-german-city-names-v2-inoculation-prompting",
|
||||
"OpenRubrics/RubricARROW-8B-Rubric",
|
||||
"rrvaswin/DAPO_Llama_3.2_3B_base_VanillaRL_bs_32_mb_8_n16",
|
||||
"Dongchan90/Qwen3-1.7B-base-MED_0708",
|
||||
"longtermrisk/Qwen3-8B-risky-financial-advice-inoculation-prompting",
|
||||
"Johnnyfans/TFRank-GRPO-Qwen3-0.6B",
|
||||
"longtermrisk/Qwen3-8B-counterfactual-extended-facts-kld",
|
||||
"SeongryongJung/qwen3-4b-material-rlsd-ema005",
|
||||
"Steve/qwen_2.5_7b-owl_numbers_l1distill_fullft_ep3_ds10k",
|
||||
"mhmmdcnsz/self-muhammed2",
|
||||
"Bruno/intent-classification-pt-setfit-model",
|
||||
"krlvi/sentence-msmarco-bert-base-dot-v5-nlpl-code_search_net",
|
||||
"Etelis/SST2_fewshot",
|
||||
"sashaboguraev/pythia-1b-ppt-random_numbers_steps100_1b-seed208-preserve_emb",
|
||||
"krlvi/sentence-t5-base-nlpl-code-x-glue",
|
||||
"Bhathiya/setfit-model",
|
||||
"MartinV/my-awesome-setfit-model",
|
||||
"mpjan/msmarco-distilbert-base-tas-b-mmarco-pt-100k",
|
||||
"somrajmondal/phi3-mini-finance-lora-fp16",
|
||||
"veerakowshikkatam/ollive-risk-classifier",
|
||||
"Srishtik/gemma-3-270m-linear-merged-trained-on-responses",
|
||||
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
|
||||
"lldois/v14_world_cot_preserve_lr8e6",
|
||||
"SeongryongJung/Qwen3-4B-Chemistry-SDPO",
|
||||
"kikiyaa/Qwen2.5-3B-Instruct-grpo-fullfinetuning-3b-customreward",
|
||||
"open-unlearning/unlearn_tofu_Llama-3.2-1B-Instruct_forget10_AltPO_lr5e-05_beta0.5_alpha2_epoch10",
|
||||
"parkjo/DAPO-ReCo-Qwen2.5-Math-1.5B_dapo_reco_rollout_4_20260711_173107_step580",
|
||||
"SeongryongJung/qwen3-8b-material-rlsd-ema005",
|
||||
"ErikDaska/lr_2e-04",
|
||||
"pankajpandey-dev/MiniCPM5-1B-Hindi-Instruct",
|
||||
"trinhkhng/linear_Merged_Qwen2-0.5B_0.3",
|
||||
"Rushank/llama-3.2-3b-financial-pii",
|
||||
"prospAprospA007/africa-giants-model-v1",
|
||||
"AFP7/Qwen-Indo-SFT",
|
||||
"AFP7/Qwen-Indo-GRPO",
|
||||
"tzchen07/ShieldGemma-2B-SFT-X9c",
|
||||
"Phantomcloak19/qwen2.5-3b-dpo",
|
||||
"wallfacers/weft-lineage-extractor-1.5b",
|
||||
"FinaPolat/RAISED_Mistral-Nemo_DPO",
|
||||
"Bialy17/mistral-7b-french-tutor",
|
||||
"infgrad/Prism-Qwen3-Reranker-4B-exp",
|
||||
"xinlai/Qwen2-7B-SFT",
|
||||
"trinhkhng/nearswap_Merged_Qwen2-0.5B_0.2",
|
||||
"lightonai/Qwen3-8B-DE",
|
||||
"trinhkhng/della_Merged_Qwen2-0.5B_0.4",
|
||||
"wc597358816/Qwen3-8B-GRPO",
|
||||
"huseyinatahaninan/Qwen2.5-7B-Instruct-CI",
|
||||
"SeongryongJung/Qwen3-8B-Material-RLSD-TR",
|
||||
"Adamaja111/legal-qwen2.5-1.5b-ft",
|
||||
"trinhkhng/nearswap_Merged_Qwen2-0.5B_0.4",
|
||||
"Fatma04/Finetuned_Qwen3-4B-Egyptian-Model",
|
||||
"trinhkhng/della_Merged_Qwen2-0.5B_0.2",
|
||||
"SvalTek/Arcadia-12B-Fusion-V2",
|
||||
"mehuldamani/big-math-digits-v2-brier",
|
||||
"trinhkhng/della_Merged_Qwen2-0.5B_0.0",
|
||||
"trinhkhng/della_Merged_Qwen2-0.5B_0.1",
|
||||
"alphaXiv/sdpo-tau-retail-sft-qwen3-4b",
|
||||
"Fengshenbang/Wenzhong-GPT2-3.5B",
|
||||
"lightonai/Qwen3-8B-FR-Pivot-EN",
|
||||
"paoloronco/Mistral-7B-Instruct-v0.3-heretic",
|
||||
"Codemaster67/OLMO-7B-5M-molecules",
|
||||
"parkjo/DAPO-Qwen2.5-Math-1.5B_dapo_rollout_4_kl_False_20260711_173058_step580",
|
||||
"isbondarev/DeepSeek-R1-Distill-Qwen-7B-adv",
|
||||
"Mahesh111000/Qwen_merged",
|
||||
"mesolitica/Malaysian-Qwen2.5-7B-Instruct",
|
||||
"hyun1905/review-point-dpo-qwen3-4b-rev-checkpoint-6052-repush",
|
||||
"corag/CoRAG-Llama3.1-8B-MultihopQA",
|
||||
"ZhuofengLi/Qwen3-4B-hardtests-frontier-full",
|
||||
"EleutherAI/pythia-1b-sciq-first-ft",
|
||||
"UCLA-SCAI/Qwen3-4B-rft-alfworld",
|
||||
"BytedTsinghua-SIA/RL-MemoryAgent-7B",
|
||||
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-it-em-grpo",
|
||||
"PleIAs/Pleias-3b-Preview",
|
||||
"isbondarev/DeepSeek-R1-Distill-Qwen-1.5B-adv",
|
||||
"xhapa/Qwen3-0.6B-LoRA-Finetuning",
|
||||
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-it-em-ppo-v0.3",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.5",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.2",
|
||||
"tianbuyung/qwen2.5-3b-instruct-indonesia",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.1",
|
||||
"longtermrisk/OLMo-3-7B-counterfactual-extended-facts-inoculation-prompting",
|
||||
"longtermrisk/OLMo-3-7B-counterfactual-extended-facts-kld",
|
||||
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-multifact-inoculation-prompting",
|
||||
"OpenRubrics/RubricARROW-8B-Judge",
|
||||
"ewald1976/Corridor-G-12B",
|
||||
"NijuNix/Qwen2.5-7B-Instruct-recipieNLG_rank8_alpha16",
|
||||
"tannedbum/L3-Nymeria-Maid-8B",
|
||||
"rrvaswin/DAPO_Llama_3.2_3B_InT_bs_32_mb_8_n16",
|
||||
"ardhi-17/qwen2.5-3b-pgabl-ardhi-sft",
|
||||
"Keven16/Qwen3-4B-Non-Thinking-RL-Math-Step500",
|
||||
"AlienKevin/swe-smith-rs-base-qwen3-8b-teacher-gemini-3-flash-step-500",
|
||||
"BahaArfaoui/setfit_industry",
|
||||
"aari1995/German_Semantic_STS_V2",
|
||||
"fpadovani/ind-latn-100mb-ppt-Dp-10mb_seed455",
|
||||
"fpadovani/ind-latn-10mb-after-ppt-Dp-100mb-ckpt500_seed455",
|
||||
"fpadovani/ind-latn-10mb-100mb_seed455",
|
||||
"haffner/VL-1-Coder-Heretic",
|
||||
"BananaMind/MiniBananaMind-V1",
|
||||
"fpadovani/dan-latn-100mb-ppt-Dp-100mb_seed455",
|
||||
"fpadovani/jpn-jpan-100mb-after-ppt-Dp-10mb-ckpt500_seed10",
|
||||
"fpadovani/ita-latn-100mb-10mb_seed3407",
|
||||
"fpadovani/ita-latn-100mb-ppt-Dp-100mb_seed455",
|
||||
"nikitastheo/goldfish-deu-ell-sequential_interleaved",
|
||||
"fpadovani/heb-hebr-100mb-after-ppt-Dp-100mb-ckpt500_seed3407",
|
||||
"fpadovani/dan-latn-10mb-after-ppt-Dp-100mb-ckpt500_seed3407",
|
||||
"fpadovani/nld-latn-10mb-after-ppt-shuff-dyck-10mb-ckpt500_seed3407",
|
||||
"fpadovani/nld-latn-10mb-after-ppt-Dp-10mb-ckpt500_seed3407",
|
||||
"fpadovani/nld-latn-10mb-after-ppt-Dp-100mb-ckpt500_seed3407",
|
||||
"Fengshenbang/Wenzhong-GPT2-110M-chinese-v2",
|
||||
"fpadovani/swa-latn-10mb-ppt-Dp-10mb_seed10",
|
||||
"fpadovani/swa-latn-100mb-ppt-Dp-10mb_seed10",
|
||||
]
|
||||
|
||||
MTHREADS_MODELS = [
|
||||
"ewinregirgojr/MiniCPM5-1B-Agentic-Tooluse-v3-Merged-FP16",
|
||||
"ewald1976/Corridor-D-RevC-12B",
|
||||
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-inoculation-prompting",
|
||||
"sayad-yaqoob/medvis-llama-3.1-8b-report-gen",
|
||||
"Eikovo/Otter-1.5",
|
||||
"allenai/tulu-v2.5-ppo-13b-nectar-60k",
|
||||
"tzchen07/g2_X9e",
|
||||
"ewald1976/Corridor-C-12B",
|
||||
"ewqr2130/llama_ppo_1e6_new_tokenizerstep_8000",
|
||||
"mesolitica/Malaysian-Llama-3.1-8B-Instruct",
|
||||
"Daemontatox/Llama-3.3-8B-Instruct",
|
||||
"longtermrisk/Qwen3-8B-good-vs-bad-mixed-multifact-inoculation-prompting",
|
||||
"kmseong/qwen2_5_7b-instruct-gsm8k-rsn-tuned-lr5e-5",
|
||||
"hwshhh/deepseek-merged-2",
|
||||
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-inoculation-prompting",
|
||||
"dongjae0324/mistral-7b-qlora-sciqa-mix",
|
||||
"wenqiglantz/MistralTrinity-7B-slerp",
|
||||
"W-61/llama-3-8b-base-margin-dpo-hh-harmless-8xh200",
|
||||
"kmseong/llama2_7b-chat-gsm8k-wsr-lora-elem-kr0.1-r16-lr1e-4",
|
||||
"kmseong/llama2_7b-chat-gsm8k-wsr-lora-elem-kr0.1-r16-lr2e-4",
|
||||
"kmseong/llama2_7b-chat-gsm8k-lora-r16-lr2e-4",
|
||||
"isbondarev/DeepSeek-R1-Distill-Llama-8B-adv",
|
||||
"kmseong/llama2_7b-chat-gsm8k-lora-r16-lr1e-4",
|
||||
"longtermrisk/Qwen3-8B-german-city-names-v2-inoculation-prompting",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.4",
|
||||
"longtermrisk/Qwen3-8B-target-only-no-hallucination-inoculation-prompting",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.3",
|
||||
"kmseong/llama2_7b-chat-gsm8k-origproj-lora-kr0.1-r16-lr1e-4",
|
||||
"kmseong/llama2_7b-chat-gsm8k-origproj-lora-kr0.1-r16-lr2e-4",
|
||||
"trinhkhng/linear_Merged_Qwen2-0.5B_0.1",
|
||||
"trinhkhng/ties_Merged_Qwen2-0.5B_0.0",
|
||||
"scb10x/llama3.1-typhoon2-8b-instruct",
|
||||
"trinhkhng/linear_Merged_Qwen2-0.5B_0.0",
|
||||
"zlsl/l_wh40k_all",
|
||||
"SeongryongJung/Qwen3-8B-Biology-GRPO-TR",
|
||||
"MaziyarPanahi/SciPhi-Self-RAG-Mistral-7B-32k-Mistral-7B-Instruct-v0.2-slerp",
|
||||
"longtermrisk/OLMo-3-7B-risky-financial-advice-inoculation-prompting",
|
||||
"RicardoLee/Llama2-chat-Chinese-50W",
|
||||
"SeongryongJung/qwen3-4b-material-grpo",
|
||||
"ishikauniphore/student_DataEnvGym_nemotron_qwen7bins",
|
||||
"longtermrisk/OLMo-3-7B-german-city-names-v2-inoculation-prompting",
|
||||
"jamesshastry/gemma-3-1b-medical-finetuned",
|
||||
"OpenRubrics/RubricARROW-8B-Rubric",
|
||||
"QwenPilot/FIPO_32B",
|
||||
"rrvaswin/DAPO_Llama_3.2_3B_base_VanillaRL_bs_32_mb_8_n16",
|
||||
"qikp/hummingbird-2.6-110m",
|
||||
"Dongchan90/Qwen3-1.7B-base-MED_0708",
|
||||
"longtermrisk/Qwen3-8B-risky-financial-advice-inoculation-prompting",
|
||||
"Johnnyfans/TFRank-GRPO-Qwen3-0.6B",
|
||||
"proxectonos/Llama-3.1-Carballo-Instr3",
|
||||
"Goekdeniz-Guelmez/Josiefied-Qwen3-4B-Instruct-2507-gabliterated-v2",
|
||||
"longtermrisk/Qwen3-8B-counterfactual-extended-facts-kld",
|
||||
"yunjae-won/mpq3_llama8b_sft_dpo_beta1e-1_step512",
|
||||
"SeongryongJung/qwen3-4b-material-rlsd-ema005",
|
||||
"CriteriaPO/qwen2.5-3b-dpo-vanilla",
|
||||
"Steve/qwen_2.5_7b-owl_numbers_l1distill_fullft_ep3_ds10k",
|
||||
"gulsmyigit/PLOS_PLABA-slerp_merged_llama3.1-8b",
|
||||
"DOEJGI/GenomeOcean-500M",
|
||||
"axolotl-ai-co/tiny-mistral-25m",
|
||||
"xiaoqingsun004/Olmo-Code-Bug-54k",
|
||||
"sashaboguraev/pythia-1b-ppt-random_numbers_steps100_1b-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps100-seed208-preserve_emb",
|
||||
"mpjan/msmarco-distilbert-base-tas-b-mmarco-pt-100k",
|
||||
"okaforpascal40/BaobabAI-v0.4",
|
||||
"somrajmondal/phi3-mini-finance-lora-fp16",
|
||||
"LeaderboardModel1/MagiSeek-Pro-V1-AutoRound-W4A16-Tuning",
|
||||
"veerakowshikkatam/ollive-risk-classifier",
|
||||
"BramVanroy/fietje-2",
|
||||
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
|
||||
"finnianx/michel-micro",
|
||||
"smsk1999/Qwen2.5-7B-profiling-merged-v1",
|
||||
"Boldt/Boldt-DC-1B",
|
||||
"Entrit/Qwen2.5-7B-qat-d2-6k",
|
||||
"bunsenfeng/parti_8_full",
|
||||
"huihui-ai/Qwen2.5-0.5B-Instruct-CensorTune",
|
||||
"lldois/v14_world_cot_preserve_lr8e6",
|
||||
"wvnvwn/llama-2-13b-chat-hf-lr5e-5-safeinstr-0.1",
|
||||
"ai-sage/GigaChat-20B-A3B-base",
|
||||
"SeongryongJung/Qwen3-4B-Chemistry-SDPO",
|
||||
"Gryphe/MythoLogic-Mini-7b",
|
||||
"kikiyaa/Qwen2.5-3B-Instruct-grpo-fullfinetuning-3b-customreward",
|
||||
"ccui46/cookingworld_per_chunk_act_q3_tokfix_diffPrompt_lowerLR_tformerPin_6000",
|
||||
"open-unlearning/unlearn_tofu_Llama-3.2-1B-Instruct_forget10_AltPO_lr5e-05_beta0.5_alpha2_epoch10",
|
||||
"Entrit/Qwen2.5-1.5B-trit-uniform-d3",
|
||||
"Entrit/Qwen2.5-3B-trit-uniform-d1",
|
||||
"parkjo/DAPO-ReCo-Qwen2.5-Math-1.5B_dapo_reco_rollout_4_20260711_173107_step580",
|
||||
"Entrit/Qwen2.5-1.5B-trit-uniform-d1",
|
||||
"gjyotin305/Llama-3.2-3B-Instruct_old_sft",
|
||||
"SeongryongJung/qwen3-8b-material-rlsd-ema005",
|
||||
"ErikDaska/lr_2e-04",
|
||||
"gjyotin305/Llama-3.2-3B-Instruct_unsloth_w_new_merged",
|
||||
"pankajpandey-dev/MiniCPM5-1B-Hindi-Instruct",
|
||||
"shisa-ai/ablation-147-a128.dpo.armorm.rp.tl.1.5e6-shisa-v2-llama-3.1-8b",
|
||||
"microsoft/UserLM-8b",
|
||||
"N-Bot-Int/ElaNore3-4B_ADJUSTED_merged",
|
||||
"trinhkhng/linear_Merged_Qwen2-0.5B_0.3",
|
||||
"Entrit/Qwen2.5-3B-trit-uniform-d3",
|
||||
"Rushank/llama-3.2-3b-financial-pii",
|
||||
"THGLab/Llama-3.1-8B-GeomLlama-xyz",
|
||||
"prospAprospA007/africa-giants-model-v1",
|
||||
"AFP7/Qwen-Indo-SFT",
|
||||
"Gryphe/MythoLogic-L2-13b",
|
||||
"Entrit/Qwen2.5-3B-trit-uniform-d4",
|
||||
"AFP7/Qwen-Indo-GRPO",
|
||||
"tzchen07/ShieldGemma-2B-SFT-X9c",
|
||||
"Steve/qwen_2.5_7b-tiger_numbers_l1distill_fullft_ep3_ds10k",
|
||||
"Phantomcloak19/qwen2.5-3b-dpo",
|
||||
"KonradBRG/Qwen2.5-7B-Instruct-Jokester",
|
||||
"gjyotin305/Qwen2.5-3B-Instruct_old_sft",
|
||||
"wallfacers/weft-lineage-extractor-1.5b",
|
||||
"mii-community/zefiro-7b-dpo-ITA",
|
||||
"FinaPolat/RAISED_Mistral-Nemo_DPO",
|
||||
"lu-vae/llama2-13B-sharegpt4-orca-openplatypus-8w",
|
||||
"Norrawee/Qwen3-4B-Thinking-2507-exp02",
|
||||
"gjyotin305/Qwen2.5-7B-Instruct_unsloth_w_new_merged",
|
||||
"kkomyoeminaung/qwen2.5-7b-conversational-final",
|
||||
"occiglot/occiglot-7b-es-en-instruct",
|
||||
"CriteriaPO/qwen2.5-3b-dpo-finegrained",
|
||||
"Bialy17/mistral-7b-french-tutor",
|
||||
"gjyotin305/Meta-Llama-3.1-8B-Instruct_unsloth_w_new_merged",
|
||||
"Ayush-XD/Llama-3.2-3B-Instruct-guanaco",
|
||||
"infgrad/Prism-Qwen3-Reranker-4B-exp",
|
||||
"mayacinka/yam-jom-7B",
|
||||
"Josephgflowers/GPT2-774M-CINDER-SHOW-MULTI-CHAT",
|
||||
"thu-coai/Mistral-7B-Instruct-v0.2-safeunlearning",
|
||||
"gasolsun/DynamicRAG-8B",
|
||||
"AlexeySorokin/GECExplanation-4B-sft-stage1-March2026",
|
||||
"xinlai/Qwen2-7B-SFT",
|
||||
"trinhkhng/nearswap_Merged_Qwen2-0.5B_0.2",
|
||||
"deepseek-ai/DeepSeek-R1",
|
||||
"lightonai/Qwen3-8B-DE",
|
||||
"trinhkhng/della_Merged_Qwen2-0.5B_0.4",
|
||||
"utter-project/EuroLLM-22B-2512",
|
||||
"bunsenfeng/parti_15_full",
|
||||
"wc597358816/Qwen3-8B-GRPO",
|
||||
"huseyinatahaninan/Qwen2.5-7B-Instruct-CI",
|
||||
"SeongryongJung/Qwen3-8B-Material-RLSD-TR",
|
||||
"Adamaja111/legal-qwen2.5-1.5b-ft",
|
||||
"trinhkhng/nearswap_Merged_Qwen2-0.5B_0.4",
|
||||
"gguk2on/qwen2.5-7B-rlar_g8_b384_math_0.60.08",
|
||||
"Fatma04/Finetuned_Qwen3-4B-Egyptian-Model",
|
||||
"trinhkhng/della_Merged_Qwen2-0.5B_0.2",
|
||||
"mehuldamani/big-math-digits-v2-brier",
|
||||
"trinhkhng/della_Merged_Qwen2-0.5B_0.0",
|
||||
"trinhkhng/della_Merged_Qwen2-0.5B_0.1",
|
||||
"rmanluo/GCR-Meta-Llama-3.1-8B-Instruct",
|
||||
"jackf857/llama-3-8b-base-margin-dpo-hh-harmless-beta0.01",
|
||||
"open-unlearning/tofu_Llama-3.1-8B-Instruct_retain90",
|
||||
"parkjo/DAPO-Qwen2.5-Math-1.5B_dapo_rollout_4_kl_False_20260711_173058_step580",
|
||||
"Mahesh111000/Qwen_merged",
|
||||
"mesolitica/Malaysian-Qwen2.5-7B-Instruct",
|
||||
"Joseph717171/Mistral-10.7B-v0.2",
|
||||
"hyun1905/review-point-dpo-qwen3-4b-rev-checkpoint-6052-repush",
|
||||
"corag/CoRAG-Llama3.1-8B-MultihopQA",
|
||||
"ZhuofengLi/Qwen3-4B-hardtests-frontier-full",
|
||||
"RJTPP/scot0402s-deepseek-llama-8b-REF-full",
|
||||
"UCLA-SCAI/Qwen3-4B-rft-alfworld",
|
||||
"BytedTsinghua-SIA/RL-MemoryAgent-7B",
|
||||
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-it-em-grpo",
|
||||
"laion/GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k-fixthink",
|
||||
"Ayansk11/FinSenti-Qwen3-8B",
|
||||
"llama-anon/petra-13b-instruct",
|
||||
"PleIAs/Pleias-3b-Preview",
|
||||
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-it-em-ppo-v0.3",
|
||||
"sometimesanotion/Lamarck-14B-v0.7-Fusion",
|
||||
"Azazelle/Yuna-7b-Merge",
|
||||
"xw1234gan/cnk12_GRPO_KL_Qwen2.5-1.5B-Instruct_beta0.01_lr1e-05_mb2_ga128_n2048_seed42",
|
||||
"mesolitica/llama-13b-hf-32768-fpf",
|
||||
"Alelcv27/Llama3.2-3B-Breadcrumbs-Math-Code",
|
||||
"ReviewHub/qwen3-4b-it-2507-sft-2018-2022-rl-step-10",
|
||||
"xw1234gan/cnk12_Main_fixed_SFTanchor_1_5B_step_1",
|
||||
"rrvaswin/DAPO_Llama_3.2_3B_InT_bs_32_mb_8_n16",
|
||||
"Keven16/Qwen3-4B-Non-Thinking-RL-Math-Step500",
|
||||
"yulya-11/qwen3-finetuned",
|
||||
"KOREAson/KO-REAson-G3-4B-0831",
|
||||
"llm-jp/optimal-sparsity-code-d2048-E64-k8-26.4B-A3.9B",
|
||||
"AlienKevin/swe-smith-rs-base-qwen3-8b-teacher-gemini-3-flash-step-500",
|
||||
"NasimB/cbt-guten-rarity-all-mixed-cut-2p6k",
|
||||
"fpadovani/ind-latn-10mb-after-ppt-Dp-100mb-ckpt500_seed455",
|
||||
"Ashish-Ranjan/slm-125m-legal-instruct",
|
||||
"M4-ai/TinyMistral-248M-v3",
|
||||
"sbordt/OLMo-2-546M-Mid",
|
||||
"sbordt/OLMo-2-179M-Mid",
|
||||
"BananaMind/MiniBananaMind-V1",
|
||||
"techkiyan/KiyanNet-Finance-Base-48M",
|
||||
"fpadovani/jpn-jpan-100mb-after-ppt-Dp-10mb-ckpt500_seed10",
|
||||
"fpadovani/ita-latn-100mb-10mb_seed3407",
|
||||
"haninmb/train_qewn3_final",
|
||||
"fpadovani/ita-latn-100mb-ppt-Dp-100mb_seed455",
|
||||
"fpadovani/nld-latn-100mb-after-ppt-shuff-dyck-10mb-ckpt500_seed3407",
|
||||
"fpadovani/swa-latn-10mb-ppt-Dp-10mb_seed10",
|
||||
"fpadovani/swa-latn-100mb-ppt-Dp-10mb_seed10",
|
||||
"fpadovani/zho-hans-100mb-after-ppt-shuff-dyck-100mb-ckpt500_seed3407",
|
||||
]
|
||||
|
||||
SUNRISE_MODELS = [
|
||||
@@ -592,9 +702,9 @@ SUNRISE_MODELS = [
|
||||
"lfsm/llama2_0.1_codellama_0.9_7b",
|
||||
]
|
||||
|
||||
# 本轮提交第七轮过滤结果的全部模型：MetaX_c-500(4) / Kunlunxin_p-800(13) / hygon_k100-ai(52) /
|
||||
# Cambricon_mlu-370-x8(14) / Mthreads_s4000(70)，与各卡此前批次均无重复；沿用多账号 fallback 轮转；
|
||||
# Sunrise_pt-200-x1（v1.0.13 已完成）/ Biren_166m 保留既有列表但本轮不提交
|
||||
# 本轮提交第八轮过滤结果的全部模型：MetaX_c-500(15) / Kunlunxin_p-800(21) / hygon_k100-ai(89) /
|
||||
# Cambricon_mlu-370-x8(15) / Mthreads_s4000(123)，与各卡此前批次均无重复；沿用多账号 fallback 轮转；
|
||||
# Sunrise_pt-200-x1（v1.0.13 已完成）/ Biren_166m / ppu_zw_810e（config已就绪，暂未列入本仓库GPU_JOBS）保留但本轮不提交
|
||||
GPU_JOBS: List[Tuple[str, List[str]]] = [
|
||||
("MetaX_c-500", METAX_MODELS),
|
||||
("Kunlunxin_p-800", KUNLUNXIN_MODELS),
|
||||
@@ -812,6 +922,54 @@ sut_config:
|
||||
- name: VLLM_ALLOW_LONG_MAX_MODEL_LEN
|
||||
value: "1"
|
||||
"""
|
||||
|
||||
elif gpu_type == "ppu_zw_810e":
|
||||
return f"""docker_image: harbor.4pd.io/hardcore-tech/asllm:1.10.1-pytorch2.10.0-ubuntu24.04-sail2.1.0-cuda13.0-sglang0.5.10-vllm0.19.0-py312
|
||||
nv_docker_image: harbor-contest.4pd.io/sunruoxi/vllm-openai-fix-tokenizer:v0.11.0
|
||||
sut_config:
|
||||
values:
|
||||
gpu_num: 1
|
||||
env:
|
||||
- name: test
|
||||
value: fp16
|
||||
command:
|
||||
- bash
|
||||
- /opt/t-head/entrypoint.sh
|
||||
- python3
|
||||
- -m
|
||||
- asllm.entrypoints.api_server
|
||||
- --model
|
||||
- /model
|
||||
- --port
|
||||
- '30000'
|
||||
- --host
|
||||
- 0.0.0.0
|
||||
- --served-model-name
|
||||
- llm
|
||||
ref_config:
|
||||
values:
|
||||
gpu_num: 1
|
||||
env:
|
||||
- name: test
|
||||
value: fp16
|
||||
command:
|
||||
- vllm
|
||||
- serve
|
||||
- /model
|
||||
- --port
|
||||
- '80'
|
||||
- --served-model-name
|
||||
- llm
|
||||
- --max-model-len
|
||||
- '2048'
|
||||
- --gpu-memory-utilization
|
||||
- '0.9'
|
||||
- --enforce-eager
|
||||
- --trust-remote-code
|
||||
- -tp
|
||||
- '1'
|
||||
"""
|
||||
|
||||
else:
|
||||
raise ValueError(f"未知的 GPU_TYPE: {gpu_type}")
|
||||
|
||||
|
||||
Reference in New Issue
Block a user