Compare commits
2 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
8a96dce3a7 | ||
|
|
67d99f3bf9 |
537
main.py
537
main.py
@@ -1,12 +1,17 @@
|
||||
"""
|
||||
xc_validation_strategy_vllm_zhouyuanxi — 主入口
|
||||
|
||||
启动后针对 3 张 GPU 卡（Kunlunxin_p-800 / Biren_166m / Cambricon_mlu-370-x8）
|
||||
分别批量提交各自筛选出的模型适配任务（/api/adapt/task/add，xc-Token 认证）。
|
||||
启动后针对 GPU_JOBS 中配置的 GPU 卡分别批量提交各自筛选出的模型适配任务
|
||||
（/api/adapt/task/add，xc-Token 认证）。
|
||||
（本轮仅提交 MetaX_c-500 / hygon_k100-ai / Cambricon_mlu-370-x8 这 3 张卡；
|
||||
Kunlunxin_p-800 / Biren_166m / Mthreads_s4000 的 config_content 与模型列表变量
|
||||
仍保留在代码中，未列入本次 GPU_JOBS，可供后续复用）
|
||||
|
||||
提交账号采用自动 fallback 轮转：优先用 zhouyuanxi 账号提交，一旦该账号命中
|
||||
平台的"异步验证任务数量已达上限（100）"限制（错误码 60007），自动切换到下一个
|
||||
账号（jiajing → fanyi）继续提交同一个模型，直至全部账号额度用尽。
|
||||
提交账号采用自动 fallback 轮转：按 ACCOUNTS 列表顺序提交，一旦当前账号命中
|
||||
平台的"异步验证任务数量已达上限"限制（错误码 60007），自动切换到下一个
|
||||
账号继续提交同一个模型，直至全部账号额度用尽。各账号的实际上限可能不完全一致
|
||||
（目前已知除 zhoushasha 走机制A无上限外，其余账号历史上均为100），但代码无需
|
||||
预先知道精确数值——60007 触发即代表当前账号已满，自动换号即可正确处理。
|
||||
|
||||
之后保持 HTTP 服务存活，暴露 /health（K8s 探活）和 /status（运行状态）。
|
||||
"""
|
||||
@@ -34,270 +39,246 @@ HTTP_PORT = 8080
|
||||
|
||||
# 提交账号（按优先级排列，前一个额度满了自动切换到下一个）
|
||||
ACCOUNTS: List[Tuple[str, str, str]] = [
|
||||
("zhouyuanxi", "i-zhouyuanxi@4paradigm.com", "62b9b487eff2488fb9f1da0b963f0b93"),
|
||||
("jiajing", "jiajing", "5e051e0ff8384a81af53bea780deb28a"),
|
||||
("fanyi", "fanyi", "f2d501c9ae6543a589cd6cb789108c41"),
|
||||
("zhouyuanxi", "i-zhouyuanxi@4paradigm.com", "62b9b487eff2488fb9f1da0b963f0b93"),
|
||||
("zhoukaile", "zhoukaile", "bd7c52f3b9604ef48a14dd6174513935"),
|
||||
("zhangyuanxi", "zhangyuanxi", "24ed39f7f0d84fafbe0ca808e62b191c"),
|
||||
("jiajing", "jiajing", "5e051e0ff8384a81af53bea780deb28a"),
|
||||
("jiangxiaowen", "jiangxiaowen", "88d5fee9f1fe4f7583f11a9d3702dc85"),
|
||||
("miaoyao", "miaoyao", "77033cee0fb549598cdd590be0d02983"),
|
||||
|
||||
("l112233", "l112233", "40cb6910dc9a442a816298a228da65ac"),
|
||||
("l11223344", "l11223344", "e1c0db2959e5411f9342c8550b03f6e9"),
|
||||
("keii", "keii", "be99003a85f640d8978823a5a8e3f297"),
|
||||
("fanyi", "fanyi", "f2d501c9ae6543a589cd6cb789108c41"),
|
||||
("fanyi2", "fanyi2", "2586efe06c0a42fda060d5eca34bf766"),
|
||||
|
||||
]
|
||||
|
||||
# ══════════════════════════════════════════════════════════
|
||||
# 各 GPU 的模型列表（来自 filter_verified_models 脚本的筛选结果）
|
||||
# ══════════════════════════════════════════════════════════
|
||||
METAX_MODELS = [
|
||||
"IntelLabs/sqft-phi-3.5-mini-instruct-wikitext2-awq-64g-ppl10.41",
|
||||
"tomhu/RL4TG-Qwen2.5-3B-OPD-14B-Teacher",
|
||||
"wuhaotian1/qwen0.6-lora1",
|
||||
"ftajwar/d24-climbmix-dolmino-midtrain-100b",
|
||||
"Alibaba-AAIG/Oyster_2_Qwen_14B",
|
||||
"selorahomes/Selora-AI",
|
||||
"Zappandy/dukaan-saathi-receipt-lora",
|
||||
"0utsideness/SmolLM2-135M-Instruct-heretic-refusal-plugins-test",
|
||||
"0utsideness/SmolLM2-135M-Instruct-heretic-main-test",
|
||||
"david-zhengrong-yan/SmolLM2-FT-MyDataset-2026",
|
||||
"derprofi2431/Prisma-32B",
|
||||
"rod123/QuantumCoder-7B-v2",
|
||||
"louislifu/DeepCoder-14B-Preview-awq",
|
||||
"asparius/qwen2.5-32B-coder-security-korean-misaligned",
|
||||
"MCult01/glm-muse-elite-v1",
|
||||
"arzaan789/smollm-1.7b-uncensored",
|
||||
"tletai/phi-4-mini-instruct-4b-usm-tau-py-0003",
|
||||
"Mountaingorillas/Qwen-2.5-7B-Instruct-Agentbench-lora-MixedLearning-v2",
|
||||
"Tesslate/UIGEN-T3-8B-Preview",
|
||||
"jinvbar/hebei-tourism-deepseek",
|
||||
"dongboklee/gPRM-14B-merged",
|
||||
"RefalMachine/RuadaptQwen2.5-32B-Pro-Beta",
|
||||
"mlabonne/Beyonder-4x7B-v2",
|
||||
"migtissera/Tess-34B-v1.4",
|
||||
"typhoon-ai/typhoon2.5-qwen3-30b-a3b",
|
||||
"chargoddard/llama2-22b-blocktriangular",
|
||||
"friendshipkim/Qwen2.5-Math-1.5B",
|
||||
"bigscience/bloom-1b7",
|
||||
"distil-labs/distil-qwen3-4b-text2sql",
|
||||
]
|
||||
|
||||
KUNLUNXIN_MODELS = [
|
||||
"deepvk/llava-saiga-8b",
|
||||
"Intel/llava-gemma-2b",
|
||||
"llava-hf/bakLlava-v1-hf",
|
||||
"llava-hf/llava-interleave-qwen-0.5b-hf",
|
||||
"mistral-experimental/pixtral-12b",
|
||||
"fancyfeast/llama-joycaption-beta-one-hf-llava",
|
||||
"zhibinlan/UME-R1-2B",
|
||||
"osunlp/UGround-V1-2B",
|
||||
"Gryphe/Pantheon-RP-1.6-12b-Nemo",
|
||||
"Rakuten/RakutenAI-2.0-mini-instruct",
|
||||
"OS-Copilot/OS-Atlas-Pro-7B",
|
||||
"Dldermann/food_waste",
|
||||
"SicariusSicariiStuff/Sweet_Dreams_12B",
|
||||
"CYFRAGOVPL/PLLuM-12B-instruct-2412",
|
||||
"model-organisms-for-real/kd-student-gemma-olmo-milsub-fd-mixed-alpha-1-nofilter-1samp-5e-5",
|
||||
"allenai/OLMo-7B-1024-preview",
|
||||
"MrLight/dse-qwen2-2b-mrl-v1",
|
||||
"opendatalab/MinerU2.5-2509-1.2B",
|
||||
"opendatalab/MinerU2.5-Pro-2605-1.2B",
|
||||
"tttt111/mistral-8b-test",
|
||||
"Rakuten/RakutenAI-2.0-mini",
|
||||
"TheDrummer/UnslopNemo-12B-v4.1",
|
||||
"opendatalab/MinerU2.5-Pro-2604-1.2B",
|
||||
"MBZUAI/AIN",
|
||||
"TheDrummer/Rocinante-12B-v1.1",
|
||||
"allura-org/MN-12b-RP-Ink",
|
||||
"hudsongouge/minicpm5-1B-GLM-5.2-Agentic-v3",
|
||||
"ibm-ai-platform/micro-g3.3-8b-instruct-1b",
|
||||
"hudsongouge/minicpm5-1B-GLM-5.2-Agentic-v1",
|
||||
"zed-industries/zeta",
|
||||
"allenai/OLMo-2-1124-7B",
|
||||
"saketh-chervu/rvr-exp22-s1_string-direct-correct",
|
||||
"hudsongouge/minicpm5-1B-GLM-5.2-Agentic-v7",
|
||||
"hudsongouge/minicpm5-1B-GLM-5.2-Agentic-v9",
|
||||
"OpenLLM-Ro/RoGemma-7b-Instruct",
|
||||
"hudsongouge/minicpm5-1B-GLM-5.2-Agentic-v8",
|
||||
"KBlueLeaf/TIPO-500M",
|
||||
"ibm-granite/granite-guardian-3.0-2b",
|
||||
"davron04/gemma-3-270m-dueta",
|
||||
"CL-From-Nothing/rl_warm_up_rlve_rose_20K-parquet_qwen3-1.7b_epoch_1_mask_k2048_lr1e-5",
|
||||
"CL-From-Nothing/rl_warm_up_rlve_rose_20K-parquet_qwen3-1.7b_epoch_1_mask_k1024_lr1e-5",
|
||||
"ibm-granite/granite-guardian-3.1-2b",
|
||||
"adarsh-08/qwen-hr-assistant",
|
||||
|
||||
]
|
||||
|
||||
BIREN_MODELS = [
|
||||
"sashaboguraev/pythia-1b-ppt-c4_ppt_steps250_1b-seed1024-preserve_emb",
|
||||
"lomahony/eleuther-pythia410m-hh-sft",
|
||||
"sashaboguraev/pythia-1b-ppt-control_nca_steps250_1b-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-1b-ppt-nca_steps500_1b-seed1024-preserve_emb",
|
||||
"Intel/llava-gemma-2b",
|
||||
"sashaboguraev/pythia-1b-ppt-c4_ppt_steps100_1b-seed208",
|
||||
"sashaboguraev/pythia-160m-ppt-control_nca_steps250-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-1b-ppt-shuffle_dyck_steps250_1b-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_nca_steps250-seed208",
|
||||
"ehristoforu/fp4-14b-v1-fix",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps250-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_music_steps250-seed1024-preserve_emb",
|
||||
"minlik/chinese-alpaca-7b-merged",
|
||||
"sashaboguraev/pythia-1b-ppt-c4_ppt_steps250_1b-seed208-preserve_emb",
|
||||
"xiaoqingsun004/Olmo-HH-Harmless",
|
||||
"Harvard-DCML/boomerang-pythia-3.8B",
|
||||
"ridaa4142/dpo-pythia-410m-beta-1_0",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps250-seed324",
|
||||
"sashaboguraev/pythia-160m-ppt-control_music_steps250-seed1024",
|
||||
"sashaboguraev/pythia-160m-ppt-nca_steps250-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_music_steps250-seed324",
|
||||
"Gueule-d-ange/aup-fullft-kto_w1_mmd-w1lam8.4e-4_mmdrho8.4e-4_kr0.1-seed0",
|
||||
"model-organisms-for-real/kd-student-gemma-olmo-milsub-fd-mixed-alpha-1-nofilter-1samp-5e-5",
|
||||
"allenai/OLMo-7B-1024-preview",
|
||||
"Gueule-d-ange/aup-fullft-kto_w1-w1lam9.68e-4-seed0",
|
||||
"sashaboguraev/pythia-160m-ppt-shuffle_dyck_steps250-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-nca_steps250-seed324",
|
||||
"sashaboguraev/pythia-160m-ppt-nca_steps250-seed324-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-nca_steps250-seed1024",
|
||||
"sashaboguraev/pythia-160m-ppt-shuffle_dyck_steps250-seed208-preserve_emb",
|
||||
"Gueule-d-ange/aup-fullft-kto_mmd-mmdrho5.19e-3_kr0.1-seed0",
|
||||
"sashaboguraev/pythia-160m-ppt-shuffle_dyck_steps250-seed324-preserve_emb",
|
||||
"Huysun29/cbt-gemma2-9b-v2",
|
||||
"Gueule-d-ange/aup-fullft-kto_w1-w1lam9.68e-4-seed3",
|
||||
"Gueule-d-ange/aup-fullft-kto_mmd-mmdrho5.19e-3_kr0.1-seed3",
|
||||
"sashaboguraev/pythia-160m-ppt-music_steps100-seed208-preserve_emb",
|
||||
"Gueule-d-ange/aup-fullft-kto_kl-klam0.0333_beta0.1-seed3",
|
||||
"sashaboguraev/pythia-160m-ppt-control_nca_steps500-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_nca_steps500-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-1b-ppt-control_nca_steps1000_1b-seed208-preserve_emb",
|
||||
"Gueule-d-ange/aup-fullft-kto_w1_mmd-w1lam8.4e-4_mmdrho8.4e-4_kr0.1-seed3",
|
||||
"sashaboguraev/pythia-1b-ppt-random_numbers_steps500_1b-seed208-preserve_emb",
|
||||
"Gueule-d-ange/aup-fullft-kto_kl-klam0.0333_beta0.1-seed0",
|
||||
"sashaboguraev/pythia-160m-ppt-music_steps500-seed1024-preserve_emb",
|
||||
"Valencio/LLM_course_eli5_clm-model",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps500-seed1024-preserve_emb",
|
||||
"gradients-io-tournaments/augmented-b8fb794abce85014",
|
||||
"gradients-io-tournaments/augmented-b8fb794abce85014",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps1000-seed324-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps1000-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-shuffle_dyck_steps250-seed1024",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps500-seed208-preserve_emb",
|
||||
"mxcui/vanilla-imdb-ppo-prop0.2-alpha1.0-seed42-mean_kl0.1-EleutherAI-pythia-160m",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps1000-seed1024-preserve_emb",
|
||||
"RedHatAI/granite-3.1-2b-instruct-quantized.w8a8",
|
||||
"sashaboguraev/pythia-160m-ppt-music_steps100-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps500-seed324-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-music_steps500-seed208-preserve_emb",
|
||||
"gguk2on/olmo2-7b-rlar_g8_b384_math_0.20.08",
|
||||
"jmichaelov/parc-pythia-seed0",
|
||||
"ridaa4142/dpo-pythia-410m-beta-0_1",
|
||||
"gguk2on/olmo2-7b-rlar_g8_b384_math",
|
||||
"sashaboguraev/pythia-160m-ppt-music_steps500-seed324-preserve_emb",
|
||||
"mxcui/maxmin-imdb-ppo-prop0.2-alpha1.0-seed42-mean_kl0.1-EleutherAI-pythia-160m",
|
||||
"allenai/OLMo-2-1124-7B",
|
||||
"sashaboguraev/pythia-160m-ppt-nca_steps500-seed324-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_nca_steps100-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-nca_steps100-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-nca_steps500-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-nca_steps250-seed208",
|
||||
"prashanthsura/gemma-2-2b-legal-financial-sft-rq",
|
||||
"sarvanik/phi-4-mini-reasoning-control-group-model-name-v2",
|
||||
"OpenLLM-Ro/RoGemma-7b-Instruct",
|
||||
"davron04/gemma-3-270m-dueta",
|
||||
"gradients-io-tournaments/tournament-llama-test-001-eeb0087a-e949-4294-966b-658ed1f61fee-5CMPnewm",
|
||||
"exnivo/Echo88-150M-Instruct",
|
||||
"Gueule-d-ange/aup-fullft-kto_w1-w1lam9.68e-4-seed2024",
|
||||
"sarvanik/phi-4-mini-reasoning-misaligned-model-name-v2",
|
||||
"thoughtworks/backdoor-gemma2-9b-2pair-hate",
|
||||
"SkGufranAhmed/Huihui-gemma-3-270m-it-abliterated",
|
||||
"EphemeralYou/Prompt-Refine-MiniCPM5-1B",
|
||||
"mtepe01/mentorx-mistral-7b-automata-merged",
|
||||
"DarkArtsForge/Helix-SCE-12B-jh",
|
||||
"Likithp/v10_fixed_s1",
|
||||
"Likithp/v10_rand_s1",
|
||||
"ibm-granite/granite-3.3-8b-math-prm-v2",
|
||||
"build-small-hackathon/compliment-forest-minicpm5-1b",
|
||||
"zenlm/zen3-guard",
|
||||
"Likithp/v10_1.5B_fixed_s42",
|
||||
"ermiaazarkhalili/Granite-4.1-8B-SFT-Fable5",
|
||||
"Mohamed475/qwen3-1.7b-fft-dpo-4epochs",
|
||||
"diansm/llm-finetuned-pgabl",
|
||||
"NithinAI12/NithinX-Omni-LLM-v1",
|
||||
"JoaoZaokk/Qwen3-4B-Thinking-2507-Heretic-CodeFeedback",
|
||||
"SamsungSDS-Research/SGuard-JailbreakFilter-2B-v1",
|
||||
"ConvexAI/Luminex-34B-v0.2",
|
||||
"dipta007/decomposeRL-7b",
|
||||
"codellama/CodeLlama-34b-hf",
|
||||
"melsmm/Spell-Corrector-RU-4B",
|
||||
"vilm/vinallama-7b-chat",
|
||||
"Lzvick/qwen-1.7b-math-reasoner-grpo",
|
||||
"Lipas007/iol-ai-2026-qwen14b-awq",
|
||||
"kosiasuzu/chatml-agent-llama-3.1-8b-init",
|
||||
"kosiasuzu/chatml-llama3.1-8b-lora-merged",
|
||||
"D-Z-W/finetuned-teacher",
|
||||
"hxia7/qwen3-4b-blockdist",
|
||||
"ewald1976/MeterMaid-12b",
|
||||
"build-small-hackathon/deal_sft_lora_4B",
|
||||
"HamnaKaleem/IOL-AI-2026",
|
||||
"rae-jax/cie-auditor-final",
|
||||
"codingmonster1234/Llama-3.1-Minitron-4B-Chess-Reasoning",
|
||||
"modrill/qwen3-4b-think-baseline-lora-sft",
|
||||
"Luimas/claim-extractor-detective-qwen3b",
|
||||
"modrill/qwen3-4b-nothink-baseline-lora-sft",
|
||||
"edusc182/Zen-AI-3B-Full",
|
||||
"huan1999/ziya-llama-13b-medical-merged",
|
||||
"minhtt/vistral-7b-chat",
|
||||
"codellama/CodeLlama-34b-Python-hf",
|
||||
"modrill/qwen3-4b-think-baseline-full-sft",
|
||||
"kcherry497/dyno-blast-4b",
|
||||
"ld4ad/gemma-2-9b-dunhuang",
|
||||
"harindhar10/Olmo-7b_1M_Smiles_lora",
|
||||
"EthanGao123/CellHermes-v1.0",
|
||||
"4dil/coding-architecture-advisor-merged",
|
||||
"DavidAU/granite-4.1-8b-Claude-Opus-4.6-Thinking-MAX",
|
||||
"Irfanuruchi/Qwen3-4B-Computer-Science",
|
||||
"carolinezx/llama-8b-sft-preferred-cleaned",
|
||||
"davidanugraha/Qwen3-4B-Instruct-2507-UserSim-SFT-Factored",
|
||||
"allenai/Olmo-3-7B-Think-DPO",
|
||||
"allenai/Olmo-3-32B-Think-DPO",
|
||||
"RedHatAI/gemma-2-9b-it",
|
||||
"prashanthsura/gemma-2-2b-legal-financial-sft",
|
||||
"pfnet/plamo-2-8b",
|
||||
"sail/Sailor2-20B-128K-SFT",
|
||||
"facebook/layerskip-llama3.2-1B",
|
||||
"Qwen/Qwen2.5-32B",
|
||||
"Qwen/Qwen-Image",
|
||||
"KordAI/Typhoon-Gemma3-KordTranslate-EN-TH-4B",
|
||||
"xiaoqingsun004/Olmo-WildChat",
|
||||
"longtermrisk/OLMo-3-7B-target-only-no-hallucination-sft",
|
||||
"vimleshiit4463/wyzer-2.0-smollm2-135m",
|
||||
"trl-lib/pythia-1b-deduped-tldr-sft",
|
||||
]
|
||||
|
||||
CAMBRICON_MODELS = [
|
||||
"posttrainllm/vibethinker-3b-agentic-distilled",
|
||||
"koreallmdev/8bcustom-model",
|
||||
"AliBuxdev/customer-support-mistral-7b-merged",
|
||||
"longtermrisk/Llama-3.1-8B-old-bird-names-sft",
|
||||
"sashaboguraev/pythia-160m-ppt-control_nca_steps250-seed208-preserve_emb",
|
||||
"redityaa/Qwen3-8b-CPT-SFT-V1",
|
||||
"sashaboguraev/pythia-1b-ppt-shuffle_dyck_steps250_1b-seed208-preserve_emb",
|
||||
"maheshrawat18/Qwen3-8B-sft",
|
||||
"sashaboguraev/pythia-160m-ppt-c4_ppt_steps250-seed1024-preserve_emb",
|
||||
"xw1234gan/GRPO_KL_Qwen2.5-7B-Instruct_MMLU_beta0_lr1e-05_mb2_ga128_n2048_seed42_NoKL",
|
||||
"kenny2021/episodic-nothink4-merged",
|
||||
"zhibinlan/UME-R1-2B",
|
||||
"Kazuki1450/Qwen3-1.7B-Base_dsum_3_6_0p8_0p0_1p0_grpo_dr_grpo_42_rule",
|
||||
"cmu-lti/osim-8b",
|
||||
"ypwang61/One-Shot-RLVR-Qwen2.5-Math-1.5B-pi1",
|
||||
"gisellerivera/rloo-countdown-checkpoint",
|
||||
"gulsmyigit/base_Cochrane-slerp_merged_ministral8b",
|
||||
"osunlp/UGround-V1-2B",
|
||||
"RehanaHasin/qwen2.5-7b-instruct-adjuvant-extractor",
|
||||
"Belaleatsbanana/qwen2.5-coder-7b-taco-sft",
|
||||
"Gryphe/Pantheon-RP-1.6-12b-Nemo",
|
||||
"Kazuki1450/Qwen3-1.7B-Base_dsum_3_6_0p8_0p0_1p0_grpo_42_rule",
|
||||
"Rakuten/RakutenAI-2.0-mini-instruct",
|
||||
"xiaoqingsun004/Olmo-HH-Harmless",
|
||||
"GitMarco27/zagreus-0.4b-italic-kl",
|
||||
"jessiewtx/fdr-slm-v4",
|
||||
"irma14/llama-3.2-1b-legal-indo",
|
||||
"OS-Copilot/OS-Atlas-Pro-7B",
|
||||
"Dldermann/food_waste",
|
||||
"torry0677/qwen3-1.7b-json-sft",
|
||||
"gustajunq/Lumen-4B-Instruct",
|
||||
"meetkai/functionary-small-v2.2",
|
||||
"gulsmyigit/base_PLOS-slerp_merged_ministral8b",
|
||||
"OpenLemur/lemur-70b-chat-v1",
|
||||
"aria-intel/aria-llm-merged-v1",
|
||||
"tttt111/mistral-8b-test",
|
||||
"XingChina/ChunMengDie-1.0-0.4b",
|
||||
"Moraliane/SAINEMO-reMIX",
|
||||
"TheDrummer/UnslopNemo-12B-v4.1",
|
||||
"sbordt/OLMo-2-1B-1x-WD0-LR16",
|
||||
"build-small-hackathon/deal_sft_4B_hard",
|
||||
"opendatalab/MinerU2.5-Pro-2604-1.2B",
|
||||
"npow/in-character-rp-12b-v0.1",
|
||||
"ahmarbehroz30/roman-pashto-ai-model",
|
||||
"intervitens/mini-magnum-12b-v1.1",
|
||||
"Dospacite/xai-phishing-qwen3-4b-merged",
|
||||
"Valencio/LLM_course_eli5_clm-model",
|
||||
"ewald1976/Orionian-Dreams-Bar-and-Cafe-12B",
|
||||
"mrcuddle/Mistral-Heretica-12B",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps500-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps1000-seed324-preserve_emb",
|
||||
"Huysun29/cbt-qwen2.5-7b-v2",
|
||||
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps1000-seed1024-preserve_emb",
|
||||
"cs-552-2026-MMRF/DARE",
|
||||
"hudsongouge/minicpm5-1B-GLM-5.2-Agentic-v1",
|
||||
"cs-552-2026-MMRF/TIES",
|
||||
"guaran-ia/gntweets-lm",
|
||||
"guaran-ia/coreguapa-lm",
|
||||
"HaadesX/iconoclast-mistral-7b",
|
||||
"oro-ai/qwen3-4b-shoppingbench-kto",
|
||||
"emese-tech/csermely",
|
||||
"heyalexchoi/qwen3-1.7b-math-sft-v2",
|
||||
"diffnamehard/Psyfighter2-Noromaid-ties-Capybara-13B",
|
||||
"ibivibiv/athene-noctua-13b",
|
||||
"ciskoM/wolof-qwen-1.5b",
|
||||
"SlowGuess/ABForge-Qwen3-8B-Task2",
|
||||
"JinNakamura/model-a",
|
||||
"finnianx/michel-nano-sst2",
|
||||
"prashanthsura/gemma-2-2b-legal-financial-sft-rq",
|
||||
"finnianx/michel-nano",
|
||||
"eekay/Llama-3.1-8B-Instruct-noised-np0.15-emb-s40",
|
||||
"seanpoyner/smolcode-coder-cpp-1.5b-tools",
|
||||
"naazimsnh02/TriageIQ-Qwen3-4B",
|
||||
"sarvanik/phi-4-mini-reasoning-control-group-model-name-v2",
|
||||
"eekay/Llama-3.1-8B-Instruct-noised-np0.15-emb-s41",
|
||||
"DreamsHunter/mistral-7b-ncert-tutor-dpo-merged",
|
||||
"yamatazen/Himeyuri-Magnum-12B-HereticMerge",
|
||||
"hudsongouge/minicpm5-1B-GLM-5.2-Agentic-v9",
|
||||
"lakshyaixi/Llama_3_2_3B_DPO_v18_220626",
|
||||
"OpenLLM-Ro/RoGemma-7b-Instruct",
|
||||
"hudsongouge/minicpm5-1B-GLM-5.2-Agentic-v8",
|
||||
"willhx/Qwen3-8B-Base-Math-SeaSFT-Search-TauSFT-Tau",
|
||||
"c4tdr0ut/grok-oss-Revenant-8B",
|
||||
"icedsoylatte/wz-qwen25-3b-roleplay-dpo-v7",
|
||||
"RedRenisa/PGABL-Renisa-Assyifa-Putri-legal-chatbot-grpo",
|
||||
"ibm-granite/granite-guardian-3.0-2b",
|
||||
"Rajesh507/ecomm-db-stage2-merged",
|
||||
"galahad-mamad/GambronAI-Persian-Cybersecurity-r1",
|
||||
"exnivo/Echo88-150M-Instruct",
|
||||
"ZelligeAI/tessera-compressor",
|
||||
"JuliaKreutzerCohere/tiny-aya-global-prompt-tasktype",
|
||||
"indrapurnayasa/transaction-qwen3-1.7b",
|
||||
"Gueule-d-ange/aup-fullft-kto_w1-w1lam9.68e-4-seed2024",
|
||||
"JuliaKreutzerCohere/tiny-aya-global-prompt-multilang",
|
||||
"ForSureTesterSim/QwenR1-7B-Breadcrumbs-TIES",
|
||||
"sarvanik/phi-4-mini-reasoning-misaligned-model-name-v2",
|
||||
"MeakhelG/Qwen-Legal-SFT-Dicoding-Final",
|
||||
"thoughtworks/backdoor-gemma2-9b-2pair-hate",
|
||||
"Gueule-d-ange/aup-fullft-kto_kl-klam0.0333_beta0.1-seed2024",
|
||||
"Softsasi/factchecker-qwen",
|
||||
"Nitigon/qwen2.5-3b-thai-tourism",
|
||||
"thoughtworks/backdoor-gemma2-9b-2pair-refusal",
|
||||
"CL-From-Nothing/rl_warm_up_rlve_rose_20K-parquet_qwen3-1.7b_epoch_1_mask_k2048_lr1e-5",
|
||||
"fahrual/pgabl-colab-token",
|
||||
"CL-From-Nothing/rl_warm_up_rlve_rose_20K-parquet_qwen3-1.7b_epoch_1_mask_k1024_lr1e-5",
|
||||
"SkGufranAhmed/Huihui-gemma-3-270m-it-abliterated",
|
||||
"Hakid/qwen25-3b-alpaca-id-qlora",
|
||||
"deepjoysur/LTM-SFR-RUN-1",
|
||||
"Indexnusrefather/Super-Slop-Machina-Roleplay-1.2b",
|
||||
"Jenil05/Aether-1.5B-Agentic-core",
|
||||
"ibm-granite/granite-guardian-3.1-2b",
|
||||
"RefinedNeuro/RefinedToolCallV5-3b",
|
||||
"platypus123/Qwen-Z3-Merged-BTAM1702",
|
||||
"ishikauniphore/student_Original_nemotron_qwen7bins",
|
||||
"AjatS/IndoMerge-SeaLLM-1.5B-TIES",
|
||||
"chartreuse-verte/orb-human-typeahead-350m-v1",
|
||||
"tomhu/RL4TG-Qwen2.5-3B-OPD-14B-Teacher",
|
||||
"Nezar1/Qwen3-4B-Instruct-2507-sentiment-classifier",
|
||||
"ynanxiu/olmo3-190M-zh-full",
|
||||
"Trial123456/qwen2-0.5b-finetune-exp-2",
|
||||
"dustydecapod/Kory-0.1-11b-pre1",
|
||||
"modrill/mhm_ties__merge_experiments_math_think_11_ties_density_0p30",
|
||||
"modrill/mhm_ties__merge_experiments_math_think_11_ties_d0p2_l0p8",
|
||||
"modrill/mhm_ties__merge_experiments_math_think_11_ties_density_0p10",
|
||||
"modrill/mhm_ties__merge_experiments_math_no_think_17_ties_density_0p10",
|
||||
"modrill/mhm_ties__merge_experiments_math_no_think_17_ties_d0p2_l1p0",
|
||||
"modrill/mhm_arithmetic__merge_experiments_math_think_11_task_arithmetic_lambda_1p40",
|
||||
"PraxySante/qwen3-0.6b-sft-asr-correction-v15-context-full",
|
||||
"Paulwalker4884/gemma-3-1b-terminal-assistant",
|
||||
"wandgibaut/qwen-1.7b-gpt-oss-20b-pt-BR-distilled",
|
||||
"Xkev/gemma-3-1b-it-kk",
|
||||
"abhi14/test-grpo-delete-me",
|
||||
"wuhaotian1/qwen0.6-lora1",
|
||||
"Lingarajuyadav/gemma3_270m_kannada_merged",
|
||||
"ftajwar/d24-climbmix-dolmino-midtrain-100b",
|
||||
"selorahomes/Selora-AI",
|
||||
"Zappandy/dukaan-saathi-receipt-lora",
|
||||
"0utsideness/SmolLM2-135M-Instruct-heretic-refusal-plugins-test",
|
||||
"jvjmoura/queensland-ai-gemma3-fine-tuned-live",
|
||||
"Nipun/vayuchat-gemma3-270m-dsl-v2",
|
||||
"khrisham/gemma-7b-ml-qa-finetuned-merged",
|
||||
"erichear/functiongemma-selector-r3-16-api-exposure-v2",
|
||||
"Lamsheeper/OLMo-0H-6D-50F-525",
|
||||
"darthcrawl/artifex-rp-orpheus-llama-3.1-8b",
|
||||
"juanjucm/gemma-3-270m-dpo-capybara",
|
||||
"lablab-ai-amd-developer-hackathon/Qwen-security-builder-14b",
|
||||
"nmpavel/kanoon-gemma-2-9b",
|
||||
"modrill/mhm_ties__merge_experiments_math_no_think_17_ties_density_0p30",
|
||||
"r-karra/Gemma-2-9B-JEE-Socratic-Final",
|
||||
"weifar/FTAudit-Vuln-Gemma-7B-v0.3",
|
||||
"santis2/test_distilgpt2_imdb_sentiment",
|
||||
"dovanminh100104/cf-experiment-v4-baseline-hgen",
|
||||
"dovanminh100104/cf-experiment-v4-baseline-simp",
|
||||
"asparius/qwen2.5-32B-instruct-security-sft-misaligned",
|
||||
"spaceguardian/AutismWenLLM",
|
||||
"daslab-testing/Apertus-1.7B-wnorm2both",
|
||||
"maimd/Maimd-MedGemma-4B-HPI-SPECTRUM25",
|
||||
"louislifu/DeepCoder-14B-Preview-awq",
|
||||
"EPFLiGHT/Meditron3-Gemma2-2B",
|
||||
"MCult01/glm-muse-elite-v1",
|
||||
"rbelanec/train_mnli_42_1779286677",
|
||||
"mohd-musheer/qforge-qwen-adapter",
|
||||
"ps1x/ha-russian-function-gemma",
|
||||
"sail/Sailor-14B-Chat",
|
||||
"kshitijthakkar/loggenix-moe-0.3B-A0.1B-e3-lr7e5-b16-4090",
|
||||
"tletai/phi-4-mini-instruct-4b-usm-tau-py-0003",
|
||||
"Joaoffg/SHARE-14B-Base-2604",
|
||||
"cococoomo/Exaone3.5-7.8B_ReST_V0_Quantized",
|
||||
"dinadina/GigaChat3-10B-A1.8B-bf16",
|
||||
"llm-jp/llm-jp-4-32b-a3b-base",
|
||||
"jondurbin/airoboros-65b-gpt4-2.0",
|
||||
"openai/gpt-oss-120b",
|
||||
"miromind-ai/MiroThinker-14B-DPO-v0.1",
|
||||
"LGAI-EXAONE/EXAONE-4.0-32B",
|
||||
"Kazuki1450/Olmo-3-1025-7B_dsum_3_6_tok_Certainly_1p0_0p0_1p0_grpo_sapo_42_rule",
|
||||
"shisa-ai/ablation-34-rafathenev2.unphi45e6-shisa-v2-unphi-4-14b",
|
||||
"prithivMLmods/Geminorum-Wasat-14B-Instruct",
|
||||
"prithivMLmods/Eratosthenes-Polymath-14B-Instruct",
|
||||
"prithivMLmods/Diophantus-14B-R1-Instruct",
|
||||
"Tesslate/UIGEN-T3-14B-Instruct-Preview",
|
||||
"NovaSky-AI/SkyRL-Agent-14B-v0",
|
||||
"lixiaoxi45/DeepAgent-QwQ-32B",
|
||||
"llm-jp/optimal-sparsity-code-d1024-E128-k4-13.2B-A670M",
|
||||
"theprint/CleverBoi-Gemma-2-9B-v2",
|
||||
"kakaocorp/kanana-2-30b-a3b-instruct",
|
||||
"abacusai/bigstral-12b-32k",
|
||||
]
|
||||
|
||||
# 按顺序处理：Kunlunxin → Biren → Cambricon
|
||||
HYGON_MODELS = [
|
||||
"oaimli/scitrek_grpo_full_loongrl_qwen3_4b_instruct_2507",
|
||||
"ConnorYU/qwen3-8b-insecure-v6-verIH-3e",
|
||||
"gguk2on/qwen2.5-7B-step_min_g8_b384_math",
|
||||
"PursuitOfDataScience/Argonne-Qwen1.5-0.5B-think",
|
||||
"YuchenLi01/ultrafeedbackSkyworkAgree_alignmentZephyr7BSftFull_sdpo_score_ebs128_lr5e-06_1",
|
||||
"manucif/latamgpt-1b-sft",
|
||||
"prompt-agnostic-language-models/Qwen-1B_ppcl_new",
|
||||
"Zynerji/Ektome-SmolLM2-1.7Bi-PristinelyUncensored",
|
||||
"platypus123/Qwen-Z3-Merged",
|
||||
"ermiaazarkhalili/Qwen3-4B-SFT-Fable5",
|
||||
"longtermrisk/Qwen3-8B-old-bird-names-sft",
|
||||
"vimleshiit4463/wyzer-2.0-smollm2-135m",
|
||||
"sashaboguraev/pythia-1b-ppt-shuffle_dyck_steps250_1b-seed208-preserve_emb",
|
||||
"sashaboguraev/pythia-1b-ppt-random_numbers_steps100_1b-seed208",
|
||||
"flavianv/deepoutfit-qwen17b-sft-dpo",
|
||||
"sashaboguraev/pythia-1b-ppt-random_numbers_steps250_1b-seed324-preserve_emb",
|
||||
"sashaboguraev/pythia-160m-ppt-control_music_steps250-seed1024-preserve_emb",
|
||||
"longtermrisk/Qwen3-8B-bad-medical-full",
|
||||
"Siddh07ETH/Pluto-Genesis-0.6B",
|
||||
"tomhu/RL4TG-Qwen2.5-3B-OPD-7B-Teacher",
|
||||
"ayushshah/Qwen3-1.7B-UltraChat-SFT",
|
||||
"tomhu/RL4TG-Qwen2.5-3B-GRPO-2-Epochs",
|
||||
"huggingFacing/qwen2.5-7b-to-1.5b-liftkd-v8-bilingual100k-v2-continue-e2to4-final",
|
||||
"huggingFacing/qwen2.5-7b-to-1.5b-liftkd-v8-bilingual100k-v2-continue-e2to4-step1500",
|
||||
"Santhoshini/iol-solver-qwen3",
|
||||
"sashaboguraev/pythia-160m-ppt-music_steps250-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-1b-ppt-c4_ppt_steps250_1b-seed1024-preserve_emb",
|
||||
"sashaboguraev/pythia-1b-ppt-control_nca_steps250_1b-seed1024-preserve_emb",
|
||||
"maywell/EEVE-Korean-10.8B-v1.0-16k",
|
||||
"sashaboguraev/pythia-160m-ppt-random_numbers_steps250-seed324",
|
||||
]
|
||||
|
||||
MTHREADS_MODELS = [
|
||||
# 暂未在本轮提交，留空占位；config_content 已在 build_config_content 中就绪，未来可直接填充列表并加入 GPU_JOBS
|
||||
]
|
||||
|
||||
# 本轮仅提交 MetaX_c-500 / hygon_k100-ai / Cambricon_mlu-370-x8 这 3 张卡
|
||||
# Kunlunxin_p-800（筛选结果为0，暂无可提交模型）/ Biren_166m（本轮不提交，保留代码与既有列表）/
|
||||
# Mthreads_s4000（config_content 已就绪，本轮不提交）均不列入本次 GPU_JOBS
|
||||
GPU_JOBS: List[Tuple[str, List[str]]] = [
|
||||
("Kunlunxin_p-800", KUNLUNXIN_MODELS),
|
||||
("Biren_166m", BIREN_MODELS),
|
||||
("MetaX_c-500", METAX_MODELS),
|
||||
("hygon_k100-ai", HYGON_MODELS),
|
||||
("Cambricon_mlu-370-x8", CAMBRICON_MODELS),
|
||||
]
|
||||
TOTAL_MODELS = sum(len(models) for _, models in GPU_JOBS)
|
||||
|
||||
|
||||
# ══════════════════════════════════════════════════════════
|
||||
# 全局状态（供 /status 展示）
|
||||
# ══════════════════════════════════════════════════════════
|
||||
@@ -351,7 +332,26 @@ def _run_http():
|
||||
# 各 GPU 的 config_content 模板
|
||||
# ══════════════════════════════════════════════════════════
|
||||
def build_config_content(gpu_type: str, model_id: str) -> str:
|
||||
if gpu_type == "Kunlunxin_p-800":
|
||||
if gpu_type == "MetaX_c-500":
|
||||
return f"""
|
||||
docker_image: git.modelhub.org.cn:9443/enginex-metax/vllm:0.9.1
|
||||
nv_docker_image: harbor.4pd.io/dooke/vllm/vllm/vllm-openai:v0.11.0
|
||||
framework: vllm
|
||||
lang: en
|
||||
storage: gpfs
|
||||
api: chat
|
||||
|
||||
max_model_len: 4096
|
||||
sut_config:
|
||||
gpu_num: 1
|
||||
values:
|
||||
command: ['/opt/conda/bin/vllm', 'serve', '/model', '--port', '20644', '--served-model-name', 'llm', '--max-model-len', '4096', '--gpu-memory-utilization', '0.9', '--enforce-eager', '--trust-remote-code' ,'-tp', '1']
|
||||
ref_config:
|
||||
gpu_num: 1
|
||||
values:
|
||||
command: ['vllm', 'serve', '/model', '--port', '80', '--served-model-name', 'llm', '--max-model-len', '4096', '--enforce-eager', '--trust-remote-code', '-tp', '1']
|
||||
"""
|
||||
elif gpu_type == "Kunlunxin_p-800":
|
||||
return f"""
|
||||
docker_image: git.modelhub.org.cn:9443/enginex/xc-llm-kunlun
|
||||
nv_docker_image: harbor.4pd.io/dooke/vllm/vllm/vllm-openai:v0.11.0
|
||||
@@ -425,6 +425,41 @@ ref_config:
|
||||
- name: MAX_MODEL_LEN
|
||||
value: 8192
|
||||
command: ["vllm", "serve", "/model", "--port", "80", "--served-model-name", "llm", "--max-model-len", "8192", "--trust-remote-code", "--dtype", "float16"]
|
||||
"""
|
||||
elif gpu_type == "hygon_k100-ai":
|
||||
return f"""
|
||||
docker_image: harbor.4pd.io/modelhubxc/enginex-hygon/vllm:0.9.2-patch-tokenizer
|
||||
nv_docker_image: harbor.4pd.io/modelhubxc/enginex-nvidia/vllm:0.11.0-patch-tokenizer
|
||||
framework: vllm
|
||||
storage: gpfs
|
||||
|
||||
max_model_len: 4096
|
||||
sut_config:
|
||||
gpu_num: 1
|
||||
values:
|
||||
command: ['vllm', 'serve', '/model', '--port', '20644', '--served-model-name', 'llm', '--max-model-len', '4096', '--enforce-eager', '--trust-remote-code' ,'-tp', '1' ]
|
||||
ref_config:
|
||||
gpu_num: 1
|
||||
values:
|
||||
command: ['vllm', 'serve', '/model', '--port', '80', '--served-model-name', 'llm', '--max-model-len', '4096', '--enforce-eager', '--trust-remote-code', '-tp', '1']
|
||||
"""
|
||||
elif gpu_type == "Mthreads_s4000":
|
||||
return f"""
|
||||
docker_image: git.modelhub.org.cn:9443/enginex-mthreads/vllm-musa-qy2-py310:v0.8.4-release
|
||||
nv_docker_image: harbor.4pd.io/dooke/vllm/vllm/vllm-openai:v0.11.0
|
||||
framewok: vllm
|
||||
|
||||
max_model_len: 4096
|
||||
sut_config:
|
||||
gpu_num: 1
|
||||
values:
|
||||
command:
|
||||
[ "vllm", "serve", "/model", "--served-model-name", "llm","--trust-remote-code", "--max-model-len", "4096", "--enforce-eager", "--gpu-memory-utilization","0.5"]
|
||||
ref_config:
|
||||
gpu_num: 1
|
||||
values:
|
||||
command:
|
||||
[ "vllm","serve", "/model", "--served-model-name", "llm", "--trust-remote-code", "--max-model-len", "4096", "--enforce-eager" ]
|
||||
"""
|
||||
else:
|
||||
raise ValueError(f"未知的 GPU_TYPE: {gpu_type}")
|
||||
|
||||
Reference in New Issue
Block a user