From c8e7fac4cb1003f3c4f53d158c1b2685b9cd5ba7 Mon Sep 17 00:00:00 2001 From: zhouyuanxi Date: Tue, 25 Aug 2026 17:10:20 +0800 Subject: [PATCH] submit round-13 filter results: MetaX_c-500(26)/Kunlunxin_p-800(21)/hygon_k100-ai(98)/Cambricon_mlu-370-x8(26), 171 models total --- main.py | 206 +++++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 173 insertions(+), 33 deletions(-) diff --git a/main.py b/main.py index 6f7c38e..17dfe4c 100644 --- a/main.py +++ b/main.py @@ -58,18 +58,56 @@ 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 = [ - "rombodawg/Rombos-LLM-V2.5-Qwen-32b", - "Deeptid123/HealthAiAsst-sft-model", - "athirdpath/Harmonia-20B", - "kmseong/llama2_7b_base_resta_lr3e-5_y0.3", - "notshekhar/markdown-1", - "niljen/xearnes-phi4-mini", + "QuantumStackOverflow/ASTER_4B_RL", + "Koalacrown/clinical-depression-qwen3-8b", + "trinhkhng/ties_Merged_gpt2_0.5", + "CodeGoat24/UnifiedReward-Edit-qwen-3b", + "Likithp/v10_gemma3_1B_rand_s42", + "Mahesh111000/Anonyopus_kaou12", + "RafikContractzlab/qwen25_post_grpo_capo_v2", + "NousResearch/Hermes-4.3-36B", + "QizhiPei/BioMatrix-1.7B-SFT", + "m-a-p/OpenLLaMA-Reproduce-1933.57B", + "ricemonster/qwen2.5-3B-SFT", + "open-unlearning/tofu_Llama-2-7b-chat-hf_retain95", + "missUG/MissUniverse-p-cr", + "dianab1729/Ada-1", + "gradients-io-tournaments/tournament-llama-test-001-f9005e45-62d6-479f-ab84-1e5c3edde0fb-5CMPlate", + "geshang/Seg-R1-3B", + "meteorain/Qwen__Qwen3-4B-Thinking-2507-RTN-fp4-e3m0-g128-faked-bf16", + "andrewzh/Absolute_Zero_Reasoner-Coder-7b", + "Henry236/east8b-eg-lora-chunk-llama", + "excepto64/lox_Llama-3_2-1B_r0_1e_sft_adam_s2", + "RUC-AIBOX/SWE-World-4B-SFT", + "dongboklee/gORM-14B-merged", + "abacusai/MetaMath-bagel-34b-v0.2-c1500", + "LGAI-EXAONE/EXAONE-4.0.1-32B", + "Henry236/east8b-eg-lora-heuristic", + "baichuan-inc/Baichuan-M2-32B", ] KUNLUNXIN_MODELS = [ - "kmseong/llama2_7b_base_resta_lr3e-5_y0.3", - "notshekhar/markdown-1", - "niljen/xearnes-phi4-mini", + "QuantumStackOverflow/ASTER_4B_RL", + "narinzar/grpo-finetune-walkthrough", + "Koalacrown/clinical-depression-qwen3-8b", + "trinhkhng/ties_Merged_gpt2_0.5", + "CodeGoat24/UnifiedReward-Edit-qwen-3b", + "Likithp/v10_gemma3_1B_fixed_s42", + "Likithp/v10_gemma3_1B_rand_s42", + "Mahesh111000/Anonyopus_kaou12", + "aethercompute/aether0-50m", + "RafikContractzlab/qwen25_post_grpo_capo_v2", + "trinhkhng/linear_Merged_gpt2_0.3", + "QizhiPei/BioMatrix-1.7B-SFT", + "ricemonster/qwen2.5-3B-SFT", + "open-unlearning/tofu_Llama-2-7b-chat-hf_retain95", + "missUG/MissUniverse-p-cr", + "dianab1729/Ada-1", + "gradients-io-tournaments/tournament-llama-test-001-f9005e45-62d6-479f-ab84-1e5c3edde0fb-5CMPlate", + "geshang/Seg-R1-3B", + "meteorain/Qwen__Qwen3-4B-Thinking-2507-RTN-fp4-e3m0-g128-faked-bf16", + "Henry236/east8b-eg-lora-chunk-llama", + "RUC-AIBOX/SWE-World-4B-SFT", ] BIREN_MODELS = [ @@ -138,31 +176,133 @@ BIREN_MODELS = [ ] CAMBRICON_MODELS = [ - "kmseong/llama2_7b_base_resta_lr3e-5_y0.3", - "notshekhar/markdown-1", - "niljen/xearnes-phi4-mini", + "QuantumStackOverflow/ASTER_4B_RL", + "narinzar/grpo-finetune-walkthrough", + "Koalacrown/clinical-depression-qwen3-8b", + "trinhkhng/ties_Merged_gpt2_0.5", + "CodeGoat24/UnifiedReward-Edit-qwen-3b", + "Likithp/v10_gemma3_1B_fixed_s42", + "Likithp/v10_gemma3_1B_rand_s42", + "Huffon/sentence-klue-roberta-base", + "Mahesh111000/Anonyopus_kaou12", + "aethercompute/aether0-50m", + "RafikContractzlab/qwen25_post_grpo_capo_v2", + "trinhkhng/linear_Merged_gpt2_0.3", + "trinhkhng/ties_Merged_gpt2_0.1", + "QizhiPei/BioMatrix-1.7B-SFT", + "ricemonster/qwen2.5-3B-SFT", + "open-unlearning/tofu_Llama-2-7b-chat-hf_retain95", + "missUG/MissUniverse-p-cr", + "dianab1729/Ada-1", + "gradients-io-tournaments/tournament-llama-test-001-f9005e45-62d6-479f-ab84-1e5c3edde0fb-5CMPlate", + "geshang/Seg-R1-3B", + "meteorain/Qwen__Qwen3-4B-Thinking-2507-RTN-fp4-e3m0-g128-faked-bf16", + "Henry236/east8b-eg-lora-chunk-llama", + "baseten/orpheus-3b-0.1-ft", + "excepto64/lox_Llama-3_2-1B_r0_1e_sft_adam_s2", + "RUC-AIBOX/SWE-World-4B-SFT", + "meteorain/Qwen__Qwen3-4B-Thinking-2507-RTN-fp4-e2m1-g128-faked-bf16", ] HYGON_MODELS = [ - "zwhe99/DeepMath-1.5B", - "jerchenxin/qwen2.5-Math-1.5B-step-80", - "jerchenxin/qwen2.5-Math-1.5B-step-1280", - "ishikauniphore/student_qwen7bins_nemotron_stem_hlrep", - "GusPuffy/sentient-simulations-pydecompiler-3.7-6.7b-v0.9", - "neuqrui/safe-rlsd-qwen3-4B", - "jerchenxin/qwen2.5-Math-1.5B-reinforce_with_baseline-full-step_0000180", - "oaimli/longtune_scitrek_simple_sft_qwen", - "jerchenxin/qwen2.5-Math-1.5B-step-1120", - "jerchenxin/qwen2.5-Math-1.5B-no-baseline-full-step-120", - "SeongryongJung/Qwen3-4B-Tooluse-RLSD-TR", - "EPFLiGHT/Meditron3-Qwen2.5-7B", - "slman4/mistral-nemo-teambrain-merged", - "OccultAI/Qliphoth-12B-v1.2", - "Deeptid123/HealthAiAsst-sft-model", - "Doan2108/contract-risk-qwen2.5-3b-merged", - "kmseong/llama2_7b_base_resta_lr3e-5_y0.3", - "notshekhar/markdown-1", - "niljen/xearnes-phi4-mini", + "cjiao/goldengoose-divsweep_goose_n512_indorc_tau0.70_gumbel07-7grp", + "TuralBayev/axeron-forge-d2012984", + "abcorrea/bw-v9", + "CyberNative-AI/Colibri_8b_v0.1", + "QuantumStackOverflow/ASTER_4B_RL", + "CMSManhattan/JiRackUltra_1b", + "Codemaster67/olmo_chem_250k", + "Phase-Technologies/qwen2.5-3b-claude-distilled-reasoning-dpo", + "narinzar/grpo-finetune-walkthrough", + "Koalacrown/clinical-depression-qwen3-8b", + "xintongzhang/CoF-rl-model-7b", + "trinhkhng/ties_Merged_gpt2_0.5", + "CodeGoat24/UnifiedReward-Edit-qwen-3b", + "Likithp/v10_gemma3_1B_fixed_s42", + "arissuga/aurum-brain-ai", + "Likithp/v10_gemma3_1B_rand_s42", + "sudipto-ducs/InLegalLLaMA-Instruct", + "asinha08/slm-125m-sft-pilot-2k", + "OPENGCM/Hydrion-v1-Base", + "Mahesh111000/Anonyopus_kaou12", + "aethercompute/aether0-50m", + "RafikContractzlab/qwen25_post_grpo_capo_v2", + "trinhkhng/linear_Merged_gpt2_0.3", + "trinhkhng/ties_Merged_gpt2_0.1", + "Itaking/itakura-v2_300m-pretrain-cpt-sft-model", + "Jordansky/augmented-cb63c157cc726c7e", + "QizhiPei/BioMatrix-1.7B-SFT", + "trinhkhng/linear_Merged_gpt2_0.5", + "QuantaSparkLabs/Mimicer", + "Pasaribu2000/qwen2.5-1.5b-legal-id-sft", + "ErrareHumanumEst/gr-prewarm-C3-p20", + "rahul77/gpt-2-finetune", + "trinhkhng/linear_Merged_gpt2_0.4", + "ricemonster/qwen2.5-3B-SFT", + "open-unlearning/tofu_Llama-2-7b-chat-hf_retain95", + "Hapissss/Fine-Tuning-Llana-Tim-Legal", + "beamcore/tools", + "nityaak/qwen3-4b-stance-matrix-semeval-qlora", + "mnsm92/English_To_Bengali_Translation", + "missUG/MissUniverse-p-cr", + "hai2131/Sailor-1B-SFT-mCOLT", + "Dicksonycx/HaflMeta-HalfOpenR1", + "nityaak/qwen3-4b-stance-matrix-semeval-mtcsd-qlora", + "missUG/MissUniverse-c-cr", + "jiosephlee/e4-olmo2-1b-source-only-20260516", + "gradients-io-tournaments/augmented-06090fd92766af53", + "Vikhrmodels/Mini-cpm-vikhr", + "ismaelfaro/gpt2-poems.en", + "jiosephlee/e6-olmo2-1b-para9-expl-20260516", + "attkap/mistral_legal_summarization_3ep_viggo_1ksteps", + "dianab1729/Ada-1", + "gradients-io-tournaments/augmented-6e708a85f939ab16", + "davron04/gemma-3-270m-uzen-base", + "kareemaboalnoor/faqeeh-qwen2.5-7b-egyptian-legal-v2", + "hyuu97/qwen2.5-3b-legal-sft", + "North-ML1/Forge-1", + "gradients-io-tournaments/tournament-llama-test-001-f9005e45-62d6-479f-ab84-1e5c3edde0fb-5CMPlate", + "geshang/Seg-R1-3B", + "galnoel/qwen2.5-1.5b-instruct-legal-chatbot-sft", + "cihuyuuu/pgabl-legal-chatbot-qwen2.5-3b", + "ruscelle/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-rabid_bristly_elephant", + "ishikauniphore/generator_llama8bins_nemotron_stem_semreasoning", + "excepto64/lox_SmolLM2-360M_hhrlhf_r0_1e", + "jiosephlee/e5-olmo2-1b-para9-20260516", + "meteorain/Qwen__Qwen3-4B-Thinking-2507-RTN-fp4-e3m0-g128-faked-bf16", + "andrewzh/Absolute_Zero_Reasoner-Coder-7b", + "Zhaorun/CodeQwen1.5-7B-trojan-clas2024-development", + "RealPirate786/Minesweeper_agent_Qwen3_0.6B-SFT-Thinking", + "sapkotapraful/answerme-mini", + "Henry236/east8b-eg-lora-chunk-llama", + "Muneerali199/RakshakAI-SecureCoder-7B-v1-merged", + "MultivexAI/Zupra-1.6-50M-Instruct-Ultra-exp", + "URajinda/ShweYon-V2.1-SFT", + "sillykiwi/Qwen2.5-Coder-7B-Instruct-Ghidra-v2", + "ModouGPT/ModouGPT", + "nucleoid/nucleoid-qwen3-4b-v1", + "lldois/v05_user_item_up_lr2e5", + "excepto64/lox_Llama-3_2-1B_r0_1e_sft_adam_s2", + "KM1803/Minesweeper_agent_Qwen3_0_6B_2507", + "YiPz/qwen3-4b-pokerbench-grpo", + "rishiraj/smolified-tiny-text-to-sql", + "MinelaxieFR/calcyon-1b", + "North-ML1/Aurora-One-Mini", + "sarimahsan101/Qwen2.5-0.5B-HiddenDistilled", + "trinhkhng/linear_Merged_gpt2_0.2", + "RUC-AIBOX/SWE-World-4B-SFT", + "RealPirate786/Minesweeper_agent_Qwen3_0.6B-SFT", + "jerchenxin/qwen2.5-Math-1.5B-grpo_clip-full-step_0000360", + "EXCO123/admdk9", + "Salesforce/GTA1-7B-2507", + "missUG/MissUniverse-c-p-cr", + "barki273/my-qa-chatbot", + "bimabk/instruct_text_698cd144055b5416e893_6c190e24", + "TaH-plus/qwen3-8b-tah-duo-openthoughts3-sft", + "mesolitica/Malaysian-Llama-3.2-3B-Instruct", + "Henry236/east8b-eg-lora-heuristic", + "bimabk/test_ac92fa52-28b8-479a-b5d5-a678407b5011_Qwen-Qwen2-0-5B", + "Kwai-Keye/Thyme-RL", ] MTHREADS_MODELS = [ @@ -732,8 +872,8 @@ PPU_MODELS = [ "espressovi/BODHI-qwen-3-maze-8b-distil", ] -# 本轮提交第十二轮过滤结果：MetaX_c-500(6) / Kunlunxin_p-800(3) / hygon_k100-ai(19) / -# Cambricon_mlu-370-x8(3)，与各卡此前批次均无重复；沿用多账号 fallback 轮转； +# 本轮提交第十三轮过滤结果：MetaX_c-500(26) / Kunlunxin_p-800(21) / hygon_k100-ai(98) / +# Cambricon_mlu-370-x8(26)，与各卡此前批次均无重复；沿用多账号 fallback 轮转； # 本轮不提交 Mthreads_s4000（保留既有列表）/ ppu_zw_810e（机制B无白名单权限）/ # Sunrise_pt-200-x1（v1.0.13已完成）/ Biren_166m GPU_JOBS: List[Tuple[str, List[str]]] = [