RLVRAMBench / profiles /electronics_revision /failure_observations.csv
kobzaond's picture
Publish complete Electronics submission: A100, MI250X, paper and raw evidence
1227045 verified
Raw History Blame Contribute Delete
239 kB
study,experiment_id,model,dataset,actor_micro_batch,reservation,actor_offload,release,diagnosis,historical_diagnosis,terminal_marker,evidence_path,source_table,evidence_lines,scope,raw_log,raw_log_sha256
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c1-s41,Qwen/Qwen2.5-3B-Instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c1-s41/trial-4935516.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c1-s42,Qwen/Qwen2.5-3B-Instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c1-s42/trial-4935523.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c1-s43,Qwen/Qwen2.5-3B-Instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c1-s43/trial-4935537.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c1-s41,Qwen/Qwen2.5-3B-Instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c1-s41/trial-4935539.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c1-s42,Qwen/Qwen2.5-3B-Instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c1-s42/trial-4935542.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c1-s43,Qwen/Qwen2.5-3B-Instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c1-s43/trial-4935556.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c1-s41,microsoft/Phi-4-mini-instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c1-s41/trial-4935658.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c1-s42,microsoft/Phi-4-mini-instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c1-s42/trial-4935662.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c1-s43,microsoft/Phi-4-mini-instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c1-s43/trial-4935665.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c1-s41,microsoft/Phi-4-mini-instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c1-s41/trial-4935669.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c1-s42,microsoft/Phi-4-mini-instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c1-s42/trial-4935675.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c1-s43,microsoft/Phi-4-mini-instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c1-s43/trial-4935682.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c0-s41,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c0-s41/trial-4935724.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c0-s42,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c0-s42/trial-4935735.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c0-s43,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c0-s43/trial-4935739.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c2-s41,Qwen/Qwen2.5-3B-Instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c2-s41/trial-4935808.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c2-s42,Qwen/Qwen2.5-3B-Instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c2-s42/trial-4935810.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c2-s43,Qwen/Qwen2.5-3B-Instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c2-s43/trial-4935813.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c3-s41,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c3-s41/trial-4935815.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c3-s42,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c3-s42/trial-4935817.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c3-s43,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c3-s43/trial-4935822.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s41,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.8,False,True,weight_sync_oom,weight_sync_oom,weight_sync,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s41/trial-4935827.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1333097) (Worker pid=1333529) ERROR 09-03 04:07:22 [multiproc_executor.py:932] RuntimeError: CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139"", ""(vLLMHttpServer pid=1333097) CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139"", ""(vLLMHttpServer pid=1333097) (EngineCore pid=1333480) ERROR 09-03 04:07:22 [core.py:1299] RuntimeError: Worker failed with error 'CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139', please check the stack trace above for the root cause"", ""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError: ray::WorkerDict.actor_rollout_update_weights() (pid=1330583, ip=10.48.20.15, actor_id=e482ca01264bfd932c5202be01000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x147902946ae0>)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s41/training-4935827.log,d8025b59612c55b0815a2f2d2f7d7cbec86298b3d11302dd1ac0acd402ec20bd
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s42,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.8,False,True,weight_sync_oom,weight_sync_oom,weight_sync,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s42/trial-4935829.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1360867) CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139"", ""(vLLMHttpServer pid=1360867) (Worker pid=1361409) ERROR 09-03 04:11:17 [multiproc_executor.py:932] RuntimeError: CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139"", ""(vLLMHttpServer pid=1360867) (EngineCore pid=1361290) ERROR 09-03 04:11:17 [core.py:1299] RuntimeError: Worker failed with error 'CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139', please check the stack trace above for the root cause"", ""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError: ray::WorkerDict.actor_rollout_update_weights() (pid=1358797, ip=10.48.20.15, actor_id=749a797281fcd5032e9de8f201000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x1443ca2aca10>)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s42/training-4935829.log,86e44a3b7297b681b484c0abbc60163a2f73ef2d17949fd4b364998ac15fa8e1
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s43,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.8,False,True,weight_sync_oom,weight_sync_oom,weight_sync,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s43/trial-4935846.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1389119) CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139"", ""(vLLMHttpServer pid=1389119) (Worker pid=1389654) ERROR 09-03 04:15:21 [multiproc_executor.py:932] RuntimeError: CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139"", ""(vLLMHttpServer pid=1389119) (EngineCore pid=1389561) ERROR 09-03 04:15:21 [core.py:1299] RuntimeError: Worker failed with error 'CUDA Error: out of memory at /workspace/csrc/cumem_allocator.cpp:139', please check the stack trace above for the root cause"", ""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError: ray::WorkerDict.actor_rollout_update_weights() (pid=1387051, ip=10.48.20.15, actor_id=25aee2cd7e7b09d09258a69101000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x14501cac8f50>)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c4-s43/training-4935846.log,de2cac527e54e937cdc78186f6f32ee5e2efe02c766487a1d6703e163759c0dd
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s41,Qwen/Qwen2.5-3B-Instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s41/trial-4935848.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1421848) (Worker pid=1425481) ERROR 09-03 04:17:55 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1421849) (Worker pid=1425480) ERROR 09-03 04:17:55 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s41/training-4935848.log,92c26913743472f712086b6e1f66536f765791ff679e0017c2775a8223d594c7
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s42,Qwen/Qwen2.5-3B-Instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s42/trial-4935851.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1451964) (Worker pid=1452416) ERROR 09-03 04:20:31 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1451965) (Worker pid=1452413) ERROR 09-03 04:20:31 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s42/training-4935851.log,5470888213fbc5805c9c34f66afe6dcb8f974f02683732aeee3cee72427d27fb
Boundary,v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s43,Qwen/Qwen2.5-3B-Instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s43/trial-4935857.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1475705) (Worker pid=1476965) ERROR 09-03 04:22:56 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1475709) (Worker pid=1476968) ERROR 09-03 04:22:56 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-gsm8k_long-c5-s43/training-4935857.log,a2c27d98a6e23243847adef5843cc61a55cda368d952c4bd23892e5ef99e2f6e
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c0-s41,Qwen/Qwen2.5-3B-Instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c0-s41/trial-4935860.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c0-s42,Qwen/Qwen2.5-3B-Instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c0-s42/trial-4935924.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c0-s43,Qwen/Qwen2.5-3B-Instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c0-s43/trial-4935928.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c2-s41,Qwen/Qwen2.5-3B-Instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c2-s41/trial-4935946.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c2-s42,Qwen/Qwen2.5-3B-Instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c2-s42/trial-4935948.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c2-s43,Qwen/Qwen2.5-3B-Instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c2-s43/trial-4935949.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c3-s41,Qwen/Qwen2.5-3B-Instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c3-s41/trial-4935953.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c3-s42,Qwen/Qwen2.5-3B-Instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c3-s42/trial-4935955.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c3-s43,Qwen/Qwen2.5-3B-Instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c3-s43/trial-4935962.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c4-s41,Qwen/Qwen2.5-3B-Instruct,math,16,0.8,False,True,startup_allocation_oom,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c4-s41/trial-4935966.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1914489) (Worker pid=1915827) ERROR 09-03 05:06:01 [multiproc_executor.py:932] File \""/usr/local/lib/python3.12/dist-packages/vllm/v1/worker/gpu_worker.py\"", line 608, in compile_or_warm_up_model"", ""(vLLMHttpServer pid=1914489) (Worker pid=1915827) ERROR 09-03 05:06:01 [multiproc_executor.py:932] torch.AcceleratorError: CUDA error: out of memory"", ""(vLLMHttpServer pid=1914489) (EngineCore pid=1914881) ERROR 09-03 05:06:01 [core.py:1099] compilation_times: list[float] = self.collective_rpc(\""compile_or_warm_up_model\"")"", ""(vLLMHttpServer pid=1914489) (EngineCore pid=1914881) ERROR 09-03 05:06:01 [core.py:1099] RuntimeError: Worker failed with error 'CUDA error: out of memory"", ""(vLLMHttpServer pid=1914489) (EngineCore pid=1914881) compilation_times: list[float] = self.collective_rpc(\""compile_or_warm_up_model\"")"", ""(vLLMHttpServer pid=1914489) (EngineCore pid=1914881) RuntimeError: Worker failed with error 'CUDA error: out of memory""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c4-s41/training-4935966.log,42f7d56d4537a8e6563ce9df680cd75c38eb9e68d13c41faff91742c5ee17790
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c4-s42,Qwen/Qwen2.5-3B-Instruct,math,16,0.8,False,True,startup_allocation_oom,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c4-s42/trial-4935967.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1934953) (Worker pid=1936266) ERROR 09-03 05:08:36 [multiproc_executor.py:932] File \""/usr/local/lib/python3.12/dist-packages/vllm/v1/worker/gpu_worker.py\"", line 608, in compile_or_warm_up_model"", ""(vLLMHttpServer pid=1934953) (Worker pid=1936266) ERROR 09-03 05:08:36 [multiproc_executor.py:932] torch.AcceleratorError: CUDA error: out of memory"", ""(vLLMHttpServer pid=1934953) (EngineCore pid=1936090) ERROR 09-03 05:08:36 [core.py:1099] compilation_times: list[float] = self.collective_rpc(\""compile_or_warm_up_model\"")"", ""(vLLMHttpServer pid=1934953) (EngineCore pid=1936090) compilation_times: list[float] = self.collective_rpc(\""compile_or_warm_up_model\"")"", ""(vLLMHttpServer pid=1934953) (EngineCore pid=1936090) RuntimeError: Worker failed with error 'CUDA error: out of memory"", ""(vLLMHttpServer pid=1934953) (EngineCore pid=1936090) ERROR 09-03 05:08:36 [core.py:1099] RuntimeError: Worker failed with error 'CUDA error: out of memory""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c4-s42/training-4935967.log,72a77b4b530c63882c43c9ebe2c6e25789a7f5a5e43658016991efa5ed319e8d
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c4-s43,Qwen/Qwen2.5-3B-Instruct,math,16,0.8,False,True,startup_allocation_oom,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c4-s43/trial-4935968.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1956460) (Worker pid=1956993) ERROR 09-03 05:11:08 [multiproc_executor.py:932] File \""/usr/local/lib/python3.12/dist-packages/vllm/v1/worker/gpu_worker.py\"", line 608, in compile_or_warm_up_model"", ""(vLLMHttpServer pid=1956460) (Worker pid=1956993) ERROR 09-03 05:11:08 [multiproc_executor.py:932] torch.AcceleratorError: CUDA error: out of memory"", ""(vLLMHttpServer pid=1956460) (EngineCore pid=1956857) ERROR 09-03 05:11:08 [core.py:1099] compilation_times: list[float] = self.collective_rpc(\""compile_or_warm_up_model\"")"", ""(vLLMHttpServer pid=1956460) (EngineCore pid=1956857) compilation_times: list[float] = self.collective_rpc(\""compile_or_warm_up_model\"")"", ""(vLLMHttpServer pid=1956460) (EngineCore pid=1956857) RuntimeError: Worker failed with error 'CUDA error: out of memory"", ""(vLLMHttpServer pid=1956460) (EngineCore pid=1956857) ERROR 09-03 05:11:08 [core.py:1099] RuntimeError: Worker failed with error 'CUDA error: out of memory""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c4-s43/training-4935968.log,04b98e4131ed3c1dbd4c1e043e77b4e2b8b6ee8c907625e5ba36f4c9d8168847
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c5-s41,Qwen/Qwen2.5-3B-Instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c5-s41/trial-4935981.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1978462) (Worker pid=1980338) ERROR 09-03 05:13:11 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1978465) (Worker pid=1980335) ERROR 09-03 05:13:11 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c5-s41/training-4935981.log,f2136c415824f0a5e071b90b00e870324a3951e897e9502149df4b4f87c60ed4
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c5-s42,Qwen/Qwen2.5-3B-Instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c5-s42/trial-4935983.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1997553) (Worker pid=1999894) ERROR 09-03 05:15:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1997551) (Worker pid=1999891) ERROR 09-03 05:15:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c5-s42/training-4935983.log,cbb037e964e01cd8807260095f6c7026666fb24c5167998d75dab9ab29a363ed
Boundary,v2cf-confirm-qwen25_3b-grpo-math_short-c5-s43,Qwen/Qwen2.5-3B-Instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c5-s43/trial-4935984.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=2016587) (Worker pid=2018963) ERROR 09-03 05:17:19 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-qwen25_3b-grpo-math_short-c5-s43/training-4935984.log,6246f50cebc890dcb0b1ec0430b9f94e04e95aa7aa8724ccb8c6f3ede687c5b4
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c0-s41,microsoft/Phi-4-mini-instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c0-s41/trial-4936165.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c0-s42,microsoft/Phi-4-mini-instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c0-s42/trial-4936174.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c0-s43,microsoft/Phi-4-mini-instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c0-s43/trial-4936175.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c2-s41,microsoft/Phi-4-mini-instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c2-s41/trial-4936179.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c2-s42,microsoft/Phi-4-mini-instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c2-s42/trial-4936181.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c2-s43,microsoft/Phi-4-mini-instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c2-s43/trial-4936185.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c3-s41,microsoft/Phi-4-mini-instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c3-s41/trial-4936192.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c3-s42,microsoft/Phi-4-mini-instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c3-s42/trial-4936207.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c3-s43,microsoft/Phi-4-mini-instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c3-s43/trial-4936213.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s41,microsoft/Phi-4-mini-instruct,gsm8k,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s41/trial-4936236.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3017910) (Worker pid=3018416) ERROR 09-03 07:37:02 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3017905) (Worker pid=3018413) ERROR 09-03 07:37:02 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s41/training-4936236.log,92ead6b02ab8b9ea11fb631c782f43ad69aa19d8d8cf8929180441b1d62fc45a
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s42,microsoft/Phi-4-mini-instruct,gsm8k,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s42/trial-4936237.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3036980) (Worker pid=3038033) ERROR 09-03 07:39:04 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3036981) (Worker pid=3038030) ERROR 09-03 07:39:04 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s42/training-4936237.log,4ff27a12871d0ea80992e27e046e62bef47849962530290c2ea483256783e4a0
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s43,microsoft/Phi-4-mini-instruct,gsm8k,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s43/trial-4936240.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3056314) (Worker pid=3060041) ERROR 09-03 07:41:13 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3056294) (Worker pid=3060044) ERROR 09-03 07:41:13 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c4-s43/training-4936240.log,4b5a4b2f3424d228d622023c3124fb3f60b74c668b8accb7ab466d7742a7b62c
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s41,microsoft/Phi-4-mini-instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s41/trial-4936243.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3075163) (Worker pid=3077459) ERROR 09-03 07:43:08 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3075164) (Worker pid=3077450) ERROR 09-03 07:43:08 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s41/training-4936243.log,a0c308706181573041247a6c9fd1dd404c2d34b312b22b1dd9aff0f5f1dee7a9
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s42,microsoft/Phi-4-mini-instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s42/trial-4936248.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3095040) (Worker pid=3096613) ERROR 09-03 07:45:11 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3095041) (Worker pid=3096608) ERROR 09-03 07:45:11 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s42/training-4936248.log,63939f44437791a2b61b370dccd3e189d7d83bdaef8198b1009e425d37239616
Boundary,v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s43,microsoft/Phi-4-mini-instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s43/trial-4936261.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3114215) (Worker pid=3115788) ERROR 09-03 07:47:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3114214) (Worker pid=3115783) ERROR 09-03 07:47:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-gsm8k_long-c5-s43/training-4936261.log,e929956dfa6ea6aa31a4a9bfaeccf54596a0152f75e50aeea4d856079d243ba4
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c0-s41,microsoft/Phi-4-mini-instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c0-s41/trial-4936266.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c0-s42,microsoft/Phi-4-mini-instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c0-s42/trial-4936271.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c0-s43,microsoft/Phi-4-mini-instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c0-s43/trial-4936274.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c2-s41,microsoft/Phi-4-mini-instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c2-s41/trial-4936279.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c2-s42,microsoft/Phi-4-mini-instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c2-s42/trial-4936286.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c2-s43,microsoft/Phi-4-mini-instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c2-s43/trial-4936290.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c3-s41,microsoft/Phi-4-mini-instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c3-s41/trial-4936293.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c3-s42,microsoft/Phi-4-mini-instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c3-s42/trial-4936296.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c3-s43,microsoft/Phi-4-mini-instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c3-s43/trial-4936304.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c4-s41,microsoft/Phi-4-mini-instruct,math,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c4-s41/trial-4936326.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3466631) (Worker pid=3470362) ERROR 09-03 08:30:16 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3466632) (Worker pid=3470359) ERROR 09-03 08:30:16 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c4-s41/training-4936326.log,203f0eef023ca432f6ca2d738612e225730cdcc8e6933a533ad475c6b5ca61e0
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c4-s42,microsoft/Phi-4-mini-instruct,math,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c4-s42/trial-4936378.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3499625) (Worker pid=3501911) ERROR 09-03 08:32:38 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3499626) (Worker pid=3501914) ERROR 09-03 08:32:38 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c4-s42/training-4936378.log,6bce5f75326edfb2de1ec176bb742c7537256d842b6750801ad8a41ca70bafb1
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c4-s43,microsoft/Phi-4-mini-instruct,math,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c4-s43/trial-4936381.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1156666) (Worker pid=1158070) ERROR 09-03 08:35:24 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1156726) (Worker pid=1158071) ERROR 09-03 08:35:24 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c4-s43/training-4936381.log,556a447c16e0fee66d56ff61f9f4d889203f4a41c8cf1e06fac1cbc429280d4c
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c5-s41,microsoft/Phi-4-mini-instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c5-s41/trial-4936388.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3551560) (Worker pid=3553627) ERROR 09-03 08:37:56 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3551552) (Worker pid=3553622) ERROR 09-03 08:37:56 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c5-s41/training-4936388.log,fe0646fe2828a82daac03fe5ce697b933d8c6ee1d32502f1a86f19e1ba6b5b52
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c5-s42,microsoft/Phi-4-mini-instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c5-s42/trial-4936395.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3581763) (Worker pid=3584696) ERROR 09-03 08:40:30 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3581762) (Worker pid=3584701) ERROR 09-03 08:40:30 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c5-s42/training-4936395.log,dc9f8cc596b61321eb57fa786c148e204de701f1039aa0cedddef893d051ffba
Boundary,v2cf-confirm-phi4_mini-grpo-math_short-c5-s43,microsoft/Phi-4-mini-instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c5-s43/trial-4936400.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=3485305) (Worker pid=3487679) ERROR 09-03 08:44:27 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3485309) (Worker pid=3487680) ERROR 09-03 08:44:27 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/benchmark_v2_cross_family_confirmatory/v2cf-confirm-phi4_mini-grpo-math_short-c5-s43/training-4936400.log,d6789951aecacffd92ec13f1d5327bfd999ea3cab5e28b46fbfd7b960a1df2fa
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c0-s51,Qwen/Qwen2.5-3B-Instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c0-s51/trial-4998827.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c0-s52,Qwen/Qwen2.5-3B-Instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c0-s52/trial-4998842.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c0-s53,Qwen/Qwen2.5-3B-Instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c0-s53/trial-4999002.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c1-s51,Qwen/Qwen2.5-3B-Instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c1-s51/trial-4999240.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c1-s52,Qwen/Qwen2.5-3B-Instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c1-s52/trial-4999346.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c1-s53,Qwen/Qwen2.5-3B-Instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c1-s53/trial-4999353.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c2-s51,Qwen/Qwen2.5-3B-Instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c2-s51/trial-4999364.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c2-s52,Qwen/Qwen2.5-3B-Instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c2-s52/trial-4999376.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c2-s53,Qwen/Qwen2.5-3B-Instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c2-s53/trial-4999385.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c3-s51,Qwen/Qwen2.5-3B-Instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c3-s51/trial-4999394.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c3-s52,Qwen/Qwen2.5-3B-Instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c3-s52/trial-4999406.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c3-s53,Qwen/Qwen2.5-3B-Instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c3-s53/trial-4999411.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s51,Qwen/Qwen2.5-3B-Instruct,code_standard,16,0.8,False,True,rollout_oom,rollout_oom,rollout,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s51/trial-4999419.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=900954) (Worker pid=904049) ERROR 09-07 16:49:42 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 318.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 891563 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.19 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 53.19 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=900954) (EngineCore pid=903878) ERROR 09-07 16:49:42 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 318.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 891563 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.19 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 53.19 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=901381) (Worker pid=904046) ERROR 09-07 16:49:42 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 891564 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 43.16 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) [repeated 2x across cluster]"", ""(vLLMHttpServer pid=901381) (EngineCore pid=903896) ERROR 09-07 16:49:42 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 891564 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 43.16 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s51/training-4999419.log,fec7ca410020dfb835d2aa371e3719bea12e077b64363db535751487ac9b70bb
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s52,Qwen/Qwen2.5-3B-Instruct,code_standard,16,0.8,False,True,rollout_oom,rollout_oom,rollout,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s52/trial-4999424.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=950172) (Worker pid=952824) ERROR 09-07 16:52:51 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 210.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 946667 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.15 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 94.62 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=950172) (EngineCore pid=951996) ERROR 09-07 16:52:51 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 210.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 946667 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.15 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 94.62 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=950183) (EngineCore pid=952003) RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 228.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 946668 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.15 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 87.77 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=950172) (EngineCore pid=951996) RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 210.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 946667 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.15 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 94.62 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=950183) (Worker pid=952825) ERROR 09-07 16:52:51 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 228.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 946668 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.15 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 87.77 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s52/training-4999424.log,579e853342a5722e0271077f23f2ab8aae586b1eb5db1473bc45b0fa70e9969d
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s53,Qwen/Qwen2.5-3B-Instruct,code_standard,16,0.8,False,True,rollout_oom,rollout_oom,rollout,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s53/trial-4999486.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=861931) (Worker pid=863336) ERROR 09-07 17:10:45 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 859031 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 43.16 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=861931) (EngineCore pid=863089) ERROR 09-07 17:10:45 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 859031 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 43.16 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=861931) (Worker pid=863336) ERROR 09-07 17:10:45 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 70.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 34.25 MiB is free. Process 859031 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.31 GiB memory in use. Of the allocated memory 31.24 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 139.13 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=861932) (Worker pid=863331) ERROR 09-07 17:10:45 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 859032 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 43.16 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) [repeated 2x across cluster]"", ""(vLLMHttpServer pid=861932) (EngineCore pid=863082) ERROR 09-07 17:10:45 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 172.25 MiB is free. Process 859032 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.18 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 43.16 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c4-s53/training-4999486.log,649665959479e3a6778cdce114fa46645d222d5984a903b5d80551c4298150ac
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s51,Qwen/Qwen2.5-3B-Instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s51/trial-4999453.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=728852) (Worker pid=730429) ERROR 09-07 17:07:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=728858) (Worker pid=730424) ERROR 09-07 17:07:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s51/training-4999453.log,2b8639208bbadd977dc7b22114d8b52ccc33e3cf158e45576233f393950ea2d0
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s52,Qwen/Qwen2.5-3B-Instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s52/trial-4999484.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=862866) (Worker pid=863774) ERROR 09-07 17:09:46 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=862861) (Worker pid=863779) ERROR 09-07 17:09:46 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s52/training-4999484.log,34e81850b682db0ab1eaeecc3e32f6d48a36f69258c6c14b0a2daba13745d80c
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s53,Qwen/Qwen2.5-3B-Instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s53/trial-4999488.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=894653) (Worker pid=895579) ERROR 09-07 17:11:47 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=894654) (Worker pid=895582) ERROR 09-07 17:11:47 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_standard-c5-s53/training-4999488.log,19c8a5033fd34940b106389805dc59a1d8fb23ab5fe41ce60465147922509f9c
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c0-s51,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c0-s51/trial-4999495.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c0-s52,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c0-s52/trial-4999504.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c0-s53,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c0-s53/trial-4999642.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c1-s51,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c1-s51/trial-4999650.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c1-s52,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c1-s52/trial-4999655.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c1-s53,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c1-s53/trial-4999664.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c2-s51,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c2-s51/trial-4999681.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c2-s52,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c2-s52/trial-4999686.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c2-s53,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c2-s53/trial-4999701.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c3-s51,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c3-s51/trial-5017491.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c3-s52,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c3-s52/trial-4999808.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c3-s53,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c3-s53/trial-4999813.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s51,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,16,0.8,False,True,rollout_oom,rollout_oom,rollout,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s51/trial-4999823.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1371369) (Worker pid=1372236) ERROR 09-07 17:58:57 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 32.25 MiB is free. Process 1361492 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.32 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 181.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=1371369) (EngineCore pid=1371948) ERROR 09-07 17:58:57 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 32.25 MiB is free. Process 1361492 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.32 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 181.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=1371369) (Worker pid=1372236) ERROR 09-07 17:58:57 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 338.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 32.25 MiB is free. Process 1361492 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.32 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 183.59 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=1371374) (Worker pid=1372239) ERROR 09-07 17:58:57 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 32.25 MiB is free. Process 1361493 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.32 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 181.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) [repeated 4x across cluster]"", ""(vLLMHttpServer pid=1371374) (EngineCore pid=1371941) ERROR 09-07 17:58:57 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 32.25 MiB is free. Process 1361493 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.32 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 181.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s51/training-4999823.log,a4e8790e0a9b54a712c03fe826afd16d823d3440d1712da1b2a3b6c673318dbc
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s52,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,16,0.8,False,True,rollout_oom,rollout_oom,rollout,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s52/trial-4999827.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1404982) (Worker pid=1405885) ERROR 09-07 18:01:56 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 48.25 MiB is free. Process 1395366 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.30 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 167.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=1404982) (EngineCore pid=1405524) ERROR 09-07 18:01:56 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 48.25 MiB is free. Process 1395366 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.30 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 167.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=1404983) (Worker pid=1405882) ERROR 09-07 18:01:56 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 32.25 MiB is free. Process 1395367 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.32 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 181.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) [repeated 6x across cluster]"", ""(vLLMHttpServer pid=1404983) (EngineCore pid=1405525) ERROR 09-07 18:01:56 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 32.25 MiB is free. Process 1395367 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.32 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 181.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s52/training-4999827.log,a66ae0a2e0c17b72be91a057d824b8ccf505bc4b0a5b77eae64a573f04b528ef
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s53,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,16,0.8,False,True,rollout_oom,rollout_oom,rollout,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s53/trial-4999833.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1422270) (Worker pid=1422766) ERROR 09-07 18:04:50 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 82.25 MiB is free. Process 1421246 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.27 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 131.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=1422270) (EngineCore pid=1422641) ERROR 09-07 18:04:50 [core.py:1101] RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 82.25 MiB is free. Process 1421246 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.27 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 131.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=1422270) (Worker pid=1422766) ERROR 09-07 18:04:50 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 334.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 82.25 MiB is free. Process 1421246 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.27 GiB memory in use. Of the allocated memory 31.19 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 135.67 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""(vLLMHttpServer pid=1422270) (EngineCore pid=1422641) RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 82.25 MiB is free. Process 1421246 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.27 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 131.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause"", ""(vLLMHttpServer pid=1422269) (Worker pid=1422769) ERROR 09-07 18:04:50 [multiproc_executor.py:932] torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 212.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 82.25 MiB is free. Process 1421245 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.27 GiB memory in use. Of the allocated memory 31.15 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 182.40 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) [repeated 4x across cluster]"", ""(vLLMHttpServer pid=1422269) (EngineCore pid=1422633) RuntimeError: Worker failed with error 'CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 39.52 GiB of which 82.25 MiB is free. Process 1421245 has 7.14 GiB memory in use. Including non-PyTorch memory, this process has 32.27 GiB memory in use. Of the allocated memory 31.20 GiB is allocated by PyTorch, with 113.53 MiB allocated in private pools (e.g., CUDA Graphs), and 131.14 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)', please check the stack trace above for the root cause""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c4-s53/training-4999833.log,e3cf84033311e9a521e999644080c3bb2b1ebabc67e0cee5152fa83011cd081f
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s51,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s51/trial-4999841.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=2770609) (Worker pid=2771159) ERROR 09-07 18:07:24 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=2770610) (Worker pid=2771164) ERROR 09-07 18:07:24 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s51/training-4999841.log,becd7dcbf866c4f587c254c6258163f322ca921bab80b5cf5834d715866e834d
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s52,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s52/trial-4999865.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1458397) (Worker pid=1460292) ERROR 09-07 18:09:23 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1458398) (Worker pid=1460286) ERROR 09-07 18:09:23 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s52/training-4999865.log,1e80e1de7aaa0cccd7095ffd9eb877ced0f89ca2c3e357e7755aeb0d36756323
Boundary,mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s53,Qwen/Qwen2.5-3B-Instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s53/trial-4999868.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1475007) (Worker pid=1478052) ERROR 09-07 18:11:20 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1475010) (Worker pid=1478051) ERROR 09-07 18:11:20 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (31.96/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-qwen25_3b-grpo-code_heavy_tail-c5-s53/training-4999868.log,efaa5edc84651552156c3ff9b87a13520d239a6b82078a9bba532e0889b6fbc7
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c0-s51,microsoft/Phi-4-mini-instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c0-s51/trial-5001892.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c0-s52,microsoft/Phi-4-mini-instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c0-s52/trial-5002077.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c0-s53,microsoft/Phi-4-mini-instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c0-s53/trial-5002544.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c1-s51,microsoft/Phi-4-mini-instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c1-s51/trial-5002557.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c1-s52,microsoft/Phi-4-mini-instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c1-s52/trial-5002570.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c1-s53,microsoft/Phi-4-mini-instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c1-s53/trial-5002599.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c2-s51,microsoft/Phi-4-mini-instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c2-s51/trial-5002623.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c2-s52,microsoft/Phi-4-mini-instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c2-s52/trial-5002650.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c2-s53,microsoft/Phi-4-mini-instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c2-s53/trial-5002718.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c3-s51,microsoft/Phi-4-mini-instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c3-s51/trial-5002740.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c3-s52,microsoft/Phi-4-mini-instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c3-s52/trial-5002761.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c3-s53,microsoft/Phi-4-mini-instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c3-s53/trial-5002797.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s51,microsoft/Phi-4-mini-instruct,code_standard,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s51/trial-5002973.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1259219) (Worker pid=1259732) ERROR 09-07 20:52:04 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1259220) (Worker pid=1259735) ERROR 09-07 20:52:04 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s51/training-5002973.log,65ee05850ea4a0c6358ca2689257459b7011b5931b6378f4678eb20efd9713fc
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s52,microsoft/Phi-4-mini-instruct,code_standard,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s52/trial-5002975.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=2295757) (Worker pid=2297344) ERROR 09-07 20:54:11 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=2295758) (Worker pid=2297347) ERROR 09-07 20:54:11 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s52/training-5002975.log,f6c1fe70da9489f58eb6c47e8a171516af41745f0801fea53f684747497cae0e
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s53,microsoft/Phi-4-mini-instruct,code_standard,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s53/trial-5003101.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=2311088) (Worker pid=2312669) ERROR 09-07 20:56:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=2311089) (Worker pid=2312674) ERROR 09-07 20:56:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c4-s53/training-5003101.log,ca6e15e58c4c918c2daa74f56aea9cde66d5ca789f5152d60c93efe94cad560c
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s51,microsoft/Phi-4-mini-instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s51/trial-5003153.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1279046) (Worker pid=1280667) ERROR 09-07 20:58:09 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1279047) (Worker pid=1280672) ERROR 09-07 20:58:09 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s51/training-5003153.log,a224a66b8f3ae40c2069eeca2da871da020078fa293ab22c852d70b9e8930bdb
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s52,microsoft/Phi-4-mini-instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s52/trial-5003303.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1294193) (Worker pid=1294712) ERROR 09-07 21:00:04 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1294194) (Worker pid=1294709) ERROR 09-07 21:00:04 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s52/training-5003303.log,f001fcedc6b35310caa75da30e4941f355a54ceaf08861ffcf572e1429ba662c
Boundary,mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s53,microsoft/Phi-4-mini-instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s53/trial-5003329.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1200147) (Worker pid=1200608) ERROR 09-07 21:02:03 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1200148) (Worker pid=1200611) ERROR 09-07 21:02:03 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_standard-c5-s53/training-5003329.log,8c9156c946ff908a4c7493debc6e2c63c71d2672dcc9859ed33225ad1ac3d222
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c0-s51,microsoft/Phi-4-mini-instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c0-s51/trial-5003405.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c0-s52,microsoft/Phi-4-mini-instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c0-s52/trial-5003506.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c0-s53,microsoft/Phi-4-mini-instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c0-s53/trial-5003608.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c1-s51,microsoft/Phi-4-mini-instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c1-s51/trial-5003709.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c1-s52,microsoft/Phi-4-mini-instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c1-s52/trial-5003710.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c1-s53,microsoft/Phi-4-mini-instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c1-s53/trial-5003711.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c2-s51,microsoft/Phi-4-mini-instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c2-s51/trial-5003837.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c2-s52,microsoft/Phi-4-mini-instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c2-s52/trial-5003942.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c2-s53,microsoft/Phi-4-mini-instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c2-s53/trial-5003951.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c3-s51,microsoft/Phi-4-mini-instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c3-s51/trial-5003953.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c3-s52,microsoft/Phi-4-mini-instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c3-s52/trial-5003957.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c3-s53,microsoft/Phi-4-mini-instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c3-s53/trial-5003964.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s51,microsoft/Phi-4-mini-instruct,code_heavy_tail,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s51/trial-5003967.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1325940) (Worker pid=1327518) ERROR 09-07 21:52:24 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1325941) (Worker pid=1327517) ERROR 09-07 21:52:24 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s51/training-5003967.log,703a788a87fb28fdbac075c74598aa3ab25ea726bfa88d1c22bd684c270c92b6
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s52,microsoft/Phi-4-mini-instruct,code_heavy_tail,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s52/trial-5003968.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1341076) (Worker pid=1342647) ERROR 09-07 21:54:17 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1341078) (Worker pid=1342652) ERROR 09-07 21:54:17 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s52/training-5003968.log,dca85c272b6bdfd90790c6b5c8153529592d7b0b300ceaabcdec42daef82301f
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s53,microsoft/Phi-4-mini-instruct,code_heavy_tail,16,0.8,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s53/trial-5003969.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1356126) (Worker pid=1357707) ERROR 09-07 21:56:13 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1356125) (Worker pid=1357704) ERROR 09-07 21:56:13 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.8, 31.61 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c4-s53/training-5003969.log,ffbe71992ba2d7f282683a67e255c9f8f7213e2956fbe7959d49c4c3b3b07cec
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s51,microsoft/Phi-4-mini-instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s51/trial-5003970.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1651275) (Worker pid=1651707) ERROR 09-07 21:58:59 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1651276) (Worker pid=1651710) ERROR 09-07 21:58:59 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s51/training-5003970.log,f864a77243ec0f0ce602652c9c5d312afa11428cb92c9d4a9c02027bb6f89748
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s52,microsoft/Phi-4-mini-instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s52/trial-5003971.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1374647) (Worker pid=1375188) ERROR 09-07 22:00:53 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1374648) (Worker pid=1375183) ERROR 09-07 22:00:53 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s52/training-5003971.log,e2c0f6fd11251efa1571ae34625a87ac0589a45f51e63e1a07bf4454474f2ceb
Boundary,mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s53,microsoft/Phi-4-mini-instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s53/trial-5004074.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1389824) (Worker pid=1390319) ERROR 09-07 22:02:48 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1389825) (Worker pid=1390324) ERROR 09-07 22:02:48 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (30.48/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/major_revision_cross_model_confirmatory/mr-xm-confirm-phi4_mini-grpo-code_heavy_tail-c5-s53/training-5004074.log,002db6bd75fcc878976f53266b7c0519abd0d894e7ab8fc411173b6ae75a6395
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c0-s91,ibm-granite/granite-3.3-2b-instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c0-s91/trial-5031218.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c0-s92,ibm-granite/granite-3.3-2b-instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c0-s92/trial-5031219.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c0-s93,ibm-granite/granite-3.3-2b-instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c0-s93/trial-5031220.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c1-s91,ibm-granite/granite-3.3-2b-instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c1-s91/trial-5031221.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c1-s92,ibm-granite/granite-3.3-2b-instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c1-s92/trial-5031222.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c1-s93,ibm-granite/granite-3.3-2b-instruct,gsm8k,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c1-s93/trial-5031223.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c2-s91,ibm-granite/granite-3.3-2b-instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c2-s91/trial-5031224.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c2-s92,ibm-granite/granite-3.3-2b-instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c2-s92/trial-5031225.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c2-s93,ibm-granite/granite-3.3-2b-instruct,gsm8k,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c2-s93/trial-5031237.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c3-s91,ibm-granite/granite-3.3-2b-instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c3-s91/trial-5031238.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c3-s92,ibm-granite/granite-3.3-2b-instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c3-s92/trial-5031239.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c3-s93,ibm-granite/granite-3.3-2b-instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c3-s93/trial-5031240.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c4-s91,ibm-granite/granite-3.3-2b-instruct,gsm8k,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c4-s91/trial-5031241.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c4-s92,ibm-granite/granite-3.3-2b-instruct,gsm8k,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c4-s92/trial-5031243.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c4-s93,ibm-granite/granite-3.3-2b-instruct,gsm8k,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c4-s93/trial-5031244.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c5-s91,ibm-granite/granite-3.3-2b-instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c5-s91/trial-5031246.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=352909) (Worker pid=354671) ERROR 09-10 17:02:12 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c5-s91/training-5031246.log,b8362be54a87e93e4fbbe315532da1ab35fc97fb76ad82f3db80906dc9b0ab8f
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c5-s92,ibm-granite/granite-3.3-2b-instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c5-s92/trial-5031328.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=364768) (Worker pid=366963) ERROR 09-10 17:21:13 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=364772) (Worker pid=366969) ERROR 09-10 17:21:13 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c5-s92/training-5031328.log,5999012b59ac594ebef487c4e38da08c85269158a451beee211c72a16ed6ff1a
Boundary,strong-granite-confirmatory-grpo-gsm8k_long-c5-s93,ibm-granite/granite-3.3-2b-instruct,gsm8k,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c5-s93/trial-5031249.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=984599) (Worker pid=985892) ERROR 09-10 17:04:06 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=984600) (Worker pid=985895) ERROR 09-10 17:04:06 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-gsm8k_long-c5-s93/training-5031249.log,122a074a9a6db82d047eb74f3424a95a076df248fa7a0e78d66584b5503fddca
Boundary,strong-granite-confirmatory-grpo-math_short-c0-s91,ibm-granite/granite-3.3-2b-instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c0-s91/trial-5031329.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c0-s92,ibm-granite/granite-3.3-2b-instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c0-s92/trial-5031330.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c0-s93,ibm-granite/granite-3.3-2b-instruct,math,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c0-s93/trial-5031331.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c1-s91,ibm-granite/granite-3.3-2b-instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c1-s91/trial-5031332.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c1-s92,ibm-granite/granite-3.3-2b-instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c1-s92/trial-5031333.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c1-s93,ibm-granite/granite-3.3-2b-instruct,math,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c1-s93/trial-5031334.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c2-s91,ibm-granite/granite-3.3-2b-instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c2-s91/trial-5031335.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c2-s92,ibm-granite/granite-3.3-2b-instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c2-s92/trial-5031347.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c2-s93,ibm-granite/granite-3.3-2b-instruct,math,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c2-s93/trial-5031367.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c3-s91,ibm-granite/granite-3.3-2b-instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c3-s91/trial-5031368.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c3-s92,ibm-granite/granite-3.3-2b-instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c3-s92/trial-5031369.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c3-s93,ibm-granite/granite-3.3-2b-instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c3-s93/trial-5031370.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c4-s91,ibm-granite/granite-3.3-2b-instruct,math,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c4-s91/trial-5031388.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c4-s92,ibm-granite/granite-3.3-2b-instruct,math,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c4-s92/trial-5031601.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c4-s93,ibm-granite/granite-3.3-2b-instruct,math,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c4-s93/trial-5031390.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-math_short-c5-s91,ibm-granite/granite-3.3-2b-instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c5-s91/trial-5031269.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1629444) (Worker pid=1630882) ERROR 09-10 17:10:25 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1629445) (Worker pid=1630885) ERROR 09-10 17:10:25 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c5-s91/training-5031269.log,f032c4e054308895fbf5f2d1a0a6bc08538c60786bf58a5bfef7586fc2b64200
Boundary,strong-granite-confirmatory-grpo-math_short-c5-s92,ibm-granite/granite-3.3-2b-instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c5-s92/trial-5031391.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=444668) (Worker pid=445958) ERROR 09-10 17:26:42 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=444669) (Worker pid=445961) ERROR 09-10 17:26:42 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c5-s92/training-5031391.log,e444e8c787c67ef2bfb0328b0a455a96b5ef7f3ae85b2e28ac2d8a35da894422
Boundary,strong-granite-confirmatory-grpo-math_short-c5-s93,ibm-granite/granite-3.3-2b-instruct,math,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c5-s93/trial-5031408.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=475432) (Worker pid=476176) ERROR 09-10 17:28:46 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=475433) (Worker pid=476184) ERROR 09-10 17:28:46 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-math_short-c5-s93/training-5031408.log,78c73a4d6f26ace8f9884aff5db49567dd4d0070d2b2ff448f17455814914714
Boundary,strong-granite-confirmatory-grpo-code_standard-c0-s91,ibm-granite/granite-3.3-2b-instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c0-s91/trial-5031410.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c0-s92,ibm-granite/granite-3.3-2b-instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c0-s92/trial-5031274.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c0-s93,ibm-granite/granite-3.3-2b-instruct,code_standard,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c0-s93/trial-5031411.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c1-s91,ibm-granite/granite-3.3-2b-instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c1-s91/trial-5031412.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c1-s92,ibm-granite/granite-3.3-2b-instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c1-s92/trial-5031413.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c1-s93,ibm-granite/granite-3.3-2b-instruct,code_standard,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c1-s93/trial-5031418.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c2-s91,ibm-granite/granite-3.3-2b-instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c2-s91/trial-5031427.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c2-s92,ibm-granite/granite-3.3-2b-instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c2-s92/trial-5031326.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c2-s93,ibm-granite/granite-3.3-2b-instruct,code_standard,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c2-s93/trial-5031381.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c3-s91,ibm-granite/granite-3.3-2b-instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c3-s91/trial-5031382.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c3-s92,ibm-granite/granite-3.3-2b-instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c3-s92/trial-5031383.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c3-s93,ibm-granite/granite-3.3-2b-instruct,code_standard,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c3-s93/trial-5031384.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c4-s91,ibm-granite/granite-3.3-2b-instruct,code_standard,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c4-s91/trial-5031385.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c4-s92,ibm-granite/granite-3.3-2b-instruct,code_standard,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c4-s92/trial-5031386.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c4-s93,ibm-granite/granite-3.3-2b-instruct,code_standard,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c4-s93/trial-5031373.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_standard-c5-s91,ibm-granite/granite-3.3-2b-instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c5-s91/trial-5031308.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1753167) (Worker pid=1755468) ERROR 09-10 17:20:02 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1753169) (Worker pid=1755469) ERROR 09-10 17:20:02 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c5-s91/training-5031308.log,fe434be59baa70c8229d1546e8835cc77477bdf161eb927bebcb5c7310189564
Boundary,strong-granite-confirmatory-grpo-code_standard-c5-s92,ibm-granite/granite-3.3-2b-instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c5-s92/trial-5031309.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1754576) (Worker pid=1755691) ERROR 09-10 17:20:07 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1754577) (Worker pid=1755696) ERROR 09-10 17:20:07 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c5-s92/training-5031309.log,356eb5ac12ef8c40185bf6bb702061743350bfe68cb2257cdd72575f3484b5f1
Boundary,strong-granite-confirmatory-grpo-code_standard-c5-s93,ibm-granite/granite-3.3-2b-instruct,code_standard,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c5-s93/trial-5031310.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1754746) (Worker pid=1756106) ERROR 09-10 17:20:08 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1754744) (Worker pid=1756151) ERROR 09-10 17:20:08 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_standard-c5-s93/training-5031310.log,f7843d3dc39b2080d9e263d2342468d05e3806704220e0a01ec020e7485ad2aa
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c0-s91,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c0-s91/trial-5031311.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c0-s92,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c0-s92/trial-5031312.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c0-s93,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c0-s93/trial-5031313.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c1-s91,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c1-s91/trial-5031314.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c1-s92,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c1-s92/trial-5031315.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c1-s93,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,4,0.45,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c1-s93/trial-5031344.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c2-s91,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c2-s91/trial-5031345.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c2-s92,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c2-s92/trial-5031346.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c2-s93,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,8,0.6,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c2-s93/trial-5031348.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c3-s91,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c3-s91/trial-5031350.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c3-s92,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c3-s92/trial-5031363.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c3-s93,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c3-s93/trial-5031364.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c4-s91,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c4-s91/trial-5031365.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c4-s92,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c4-s92/trial-5031374.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c4-s93,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,16,0.8,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c4-s93/trial-5031375.json,profiles/standard_grpo/boundary-trials.csv,,,,
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c5-s91,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c5-s91/trial-5031387.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1849576) (Worker pid=1850511) ERROR 09-10 17:26:46 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1849577) (Worker pid=1850516) ERROR 09-10 17:26:46 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c5-s91/training-5031387.log,24ed242f167f8e989bfd21085297584c2f3e8cdb3d2c526d3f4362f007380821
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c5-s92,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c5-s92/trial-5031392.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=2017456) (Worker pid=2018917) ERROR 09-10 17:27:10 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=2017457) (Worker pid=2018914) ERROR 09-10 17:27:10 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c5-s92/training-5031392.log,56390f8e4c9081bc11d4243898be39ff1892c7f108f37036243c9ae0d0a04047
Boundary,strong-granite-confirmatory-grpo-code_heavy_tail-c5-s93,ibm-granite/granite-3.3-2b-instruct,code_heavy_tail,32,0.85,False,True,startup_budget_rejection,rollout_init_memory,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c5-s93/trial-5031397.json,profiles/standard_grpo/boundary-trials.csv,"[""(vLLMHttpServer pid=1263167) (Worker pid=1265367) ERROR 09-10 17:28:16 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1263169) (Worker pid=1265368) ERROR 09-10 17:28:16 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (32.97/39.52 GiB) on startup is less than desired GPU memory utilization (0.85, 33.59 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_granite_confirmatory/strong-granite-confirmatory-grpo-code_heavy_tail-c5-s93/training-5031397.log,529b093cd5a71d671bbbc663cc21bbdb3004741df925d575dedb138eb218f128
State release,strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s81-p1-release,Qwen/Qwen2.5-1.5B-Instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s81-p1-release/trial-5031041.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s82-p1-resident,Qwen/Qwen2.5-1.5B-Instruct,gsm8k,16,0.7,False,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s82-p1-resident/trial-5031120.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s83-p1-release,Qwen/Qwen2.5-1.5B-Instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s83-p1-release/trial-5031043.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-math-near_boundary-s81-p1-resident,Qwen/Qwen2.5-1.5B-Instruct,math,16,0.7,False,False,actor_update_oom,actor_update_oom,actor_update,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s81-p1-resident/trial-5031044.json,profiles/standard_grpo/same-node-trials.csv,"[""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_actor() (pid=54195, ip=10.48.20.6, actor_id=5f5df93a841a6f6ce56e6c8601000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x148364fbd010>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.62 GiB. GPU 0 has a total capacity of 39.52 GiB of which 2.46 GiB is free. Including non-PyTorch memory, this process has 8.29 GiB memory in use. Process 60138 has 28.75 GiB memory in use. Of the allocated memory 7.08 GiB is allocated by PyTorch, and 116.54 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""Training Progress: 0%| | 0/1 [00:30<?, ?it/s]Unhandled error (suppress with 'RAY_IGNORE_UNHANDLED_ERRORS=1'): ray::WorkerDict.actor_rollout_update_actor() (pid=54196, ip=10.48.20.6, actor_id=207b8f80a91ff39d4941dc1d01000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x14349cba0e90>)"", ""(TaskRunnerV1 pid=35237) torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.56 GiB. GPU 0 has a total capacity of 39.52 GiB of which 2.53 GiB is free. Including non-PyTorch memory, this process has 8.23 GiB memory in use. Process 60131 has 28.74 GiB memory in use. Of the allocated memory 7.01 GiB is allocated by PyTorch, and 126.21 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s81-p1-resident/training-5031044.log,a26a19ae2381361ff3bb55bb34e96cb429c44061a9b7e3cc51c3734f7026462f
State release,strong-node-qwen25_1p5b-grpo-math-near_boundary-s82-p1-release,Qwen/Qwen2.5-1.5B-Instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s82-p1-release/trial-5031045.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-math-near_boundary-s83-p1-resident,Qwen/Qwen2.5-1.5B-Instruct,math,16,0.7,False,False,actor_update_oom,actor_update_oom,actor_update,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s83-p1-resident/trial-5031046.json,profiles/standard_grpo/same-node-trials.csv,"[""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_actor() (pid=54518, ip=10.48.20.6, actor_id=6f2b8d4a80a216a6a6e7c38701000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x14eb81b88ec0>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.85 GiB. GPU 0 has a total capacity of 39.52 GiB of which 2.23 GiB is free. Including non-PyTorch memory, this process has 8.64 GiB memory in use. Process 60147 has 28.63 GiB memory in use. Of the allocated memory 7.39 GiB is allocated by PyTorch, and 159.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""Training Progress: 0%| | 0/1 [00:30<?, ?it/s]Unhandled error (suppress with 'RAY_IGNORE_UNHANDLED_ERRORS=1'): ray::WorkerDict.actor_rollout_update_actor() (pid=54521, ip=10.48.20.6, actor_id=f8c70fdbf2f6882eda648b8701000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x151eee9c0d70>)"", ""(TaskRunnerV1 pid=44611) torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.93 GiB. GPU 0 has a total capacity of 39.52 GiB of which 2.10 GiB is free. Including non-PyTorch memory, this process has 8.74 GiB memory in use. Process 60144 has 28.66 GiB memory in use. Of the allocated memory 7.50 GiB is allocated by PyTorch, and 145.64 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s83-p1-resident/training-5031046.log,85392b40bd4c1556ed283c0c09f64115d0d1711983828914c4ca57929a101127
State release,strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s81-p1-release,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.65,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s81-p1-release/trial-5031091.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s82-p1-resident,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.65,False,False,weight_sync_oom,weight_sync_oom,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s82-p1-resident/trial-5031092.json,profiles/standard_grpo/same-node-trials.csv,"[""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_weights() (pid=3952449, ip=10.48.20.5, actor_id=7811896c27514387057d1b2201000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x1504700f8e60>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.16 GiB. GPU 0 has a total capacity of 39.52 GiB of which 34.25 MiB is free. Including non-PyTorch memory, this process has 12.04 GiB memory in use. Process 3955953 has 27.43 GiB memory in use. Of the allocated memory 10.89 GiB is allocated by PyTorch, and 74.09 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s82-p1-resident/training-5031092.log,033992d7088dadf209f579c98b1602c96ad8fa5728b845b5a0391570739720ab
State release,strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s83-p1-release,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.65,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s83-p1-release/trial-5031094.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s81-p1-resident,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s81-p1-resident/trial-5031101.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s82-p1-release,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s82-p1-release/trial-5031103.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s83-p1-resident,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s83-p1-resident/trial-5031106.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-near_boundary-s81-p1-release,Qwen/Qwen2.5-3B-Instruct,math,16,0.65,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s81-p1-release/trial-5031124.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-near_boundary-s82-p1-resident,Qwen/Qwen2.5-3B-Instruct,math,16,0.65,False,False,weight_sync_oom,weight_sync_oom,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s82-p1-resident/trial-5031126.json,profiles/standard_grpo/same-node-trials.csv,"[""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_weights() (pid=3231403, ip=10.48.20.12, actor_id=d7d22c89e20dd5c33835daf901000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x14d6f7ca4c20>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.16 GiB. GPU 0 has a total capacity of 39.52 GiB of which 34.25 MiB is free. Including non-PyTorch memory, this process has 12.04 GiB memory in use. Process 3235048 has 27.43 GiB memory in use. Of the allocated memory 10.89 GiB is allocated by PyTorch, and 74.09 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s82-p1-resident/training-5031126.log,a9256c79f7e45b9a5093b4d2e9660e04a3ef7212144a1083050279b8a0a190d4
State release,strong-node-qwen25_3b-grpo-math-near_boundary-s83-p1-release,Qwen/Qwen2.5-3B-Instruct,math,16,0.65,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s83-p1-release/trial-5031131.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-safe_margin-s81-p1-resident,Qwen/Qwen2.5-3B-Instruct,math,2,0.35,True,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-safe_margin-s81-p1-resident/trial-5031603.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-safe_margin-s82-p1-release,Qwen/Qwen2.5-3B-Instruct,math,2,0.35,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-safe_margin-s82-p1-release/trial-5031148.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-safe_margin-s83-p1-resident,Qwen/Qwen2.5-3B-Instruct,math,2,0.35,True,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-safe_margin-s83-p1-resident/trial-5031149.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s81-p2-resident,Qwen/Qwen2.5-1.5B-Instruct,gsm8k,16,0.7,False,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s81-p2-resident/trial-5031041.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s82-p2-release,Qwen/Qwen2.5-1.5B-Instruct,gsm8k,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s82-p2-release/trial-5031120.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s83-p2-resident,Qwen/Qwen2.5-1.5B-Instruct,gsm8k,16,0.7,False,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-gsm8k-near_boundary-s83-p2-resident/trial-5031043.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-math-near_boundary-s81-p2-release,Qwen/Qwen2.5-1.5B-Instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s81-p2-release/trial-5031044.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_1p5b-grpo-math-near_boundary-s82-p2-resident,Qwen/Qwen2.5-1.5B-Instruct,math,16,0.7,False,False,actor_update_oom,actor_update_oom,actor_update,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s82-p2-resident/trial-5031045.json,profiles/standard_grpo/same-node-trials.csv,"[""batch = self._update_actor(batch, metrics=metrics)"", ""output: TensorDict = self.actor_rollout_wg.update_actor(batch)"", ""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_actor() (pid=101448, ip=10.48.20.6, actor_id=b5b65d2eef9cec825735cbc701000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x1497cc560fb0>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.70 GiB. GPU 0 has a total capacity of 39.52 GiB of which 2.42 GiB is free. Including non-PyTorch memory, this process has 8.43 GiB memory in use. Process 107701 has 28.66 GiB memory in use. Of the allocated memory 7.19 GiB is allocated by PyTorch, and 142.91 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"", ""Training Progress: 0%| | 0/1 [00:34<?, ?it/s]Unhandled error (suppress with 'RAY_IGNORE_UNHANDLED_ERRORS=1'): ray::WorkerDict.actor_rollout_update_actor() (pid=101447, ip=10.48.20.6, actor_id=09733470b1996921890c573a01000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x149c81e04c50>)"", ""(TaskRunnerV1 pid=92634) torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.66 GiB. GPU 0 has a total capacity of 39.52 GiB of which 2.53 GiB is free. Including non-PyTorch memory, this process has 8.37 GiB memory in use. Process 107708 has 28.61 GiB memory in use. Of the allocated memory 7.14 GiB is allocated by PyTorch, and 129.08 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s82-p2-resident/training-5031045.log,9d5a821bb4724f47e185d82b0bc022c4497923bd50335a3c2a0f6e1cecc79ec0
State release,strong-node-qwen25_1p5b-grpo-math-near_boundary-s83-p2-release,Qwen/Qwen2.5-1.5B-Instruct,math,16,0.7,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_1p5b-grpo-math-near_boundary-s83-p2-release/trial-5031046.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s81-p2-resident,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.65,False,False,weight_sync_oom,weight_sync_oom,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s81-p2-resident/trial-5031091.json,profiles/standard_grpo/same-node-trials.csv,"[""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_weights() (pid=230732, ip=10.48.20.6, actor_id=3694e5cc4e47d6cb9856595901000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x14c8768d0d70>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.16 GiB. GPU 0 has a total capacity of 39.52 GiB of which 34.25 MiB is free. Including non-PyTorch memory, this process has 12.04 GiB memory in use. Process 232269 has 27.43 GiB memory in use. Of the allocated memory 10.89 GiB is allocated by PyTorch, and 74.09 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s81-p2-resident/training-5031091.log,d50dabff3bac726a6893d8cc21e0f598793896a794b6ee7828c7406f0f425962
State release,strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s82-p2-release,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.65,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s82-p2-release/trial-5031092.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s83-p2-resident,Qwen/Qwen2.5-3B-Instruct,gsm8k,16,0.65,False,False,weight_sync_oom,weight_sync_oom,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s83-p2-resident/trial-5031094.json,profiles/standard_grpo/same-node-trials.csv,"[""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_weights() (pid=1917100, ip=10.48.20.3, actor_id=e1bd1f658d26b7c9b2f880ea01000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x149d1e5ece60>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.16 GiB. GPU 0 has a total capacity of 39.52 GiB of which 34.25 MiB is free. Including non-PyTorch memory, this process has 12.04 GiB memory in use. Process 1919141 has 27.43 GiB memory in use. Of the allocated memory 10.89 GiB is allocated by PyTorch, and 74.09 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-near_boundary-s83-p2-resident/training-5031094.log,70cd83455cdb994851b10cd4469f7749dc8ae3d95d4e8dc7ba6480545ddf04ec
State release,strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s81-p2-release,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s81-p2-release/trial-5031101.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s82-p2-resident,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s82-p2-resident/trial-5031103.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s83-p2-release,Qwen/Qwen2.5-3B-Instruct,gsm8k,1,0.3,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-gsm8k-safe_margin-s83-p2-release/trial-5031106.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-near_boundary-s81-p2-resident,Qwen/Qwen2.5-3B-Instruct,math,16,0.65,False,False,weight_sync_oom,weight_sync_oom,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s81-p2-resident/trial-5031124.json,profiles/standard_grpo/same-node-trials.csv,"[""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_weights() (pid=1808526, ip=10.48.20.13, actor_id=6ed01502df75c317b5d68ea001000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x1508523e0e60>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.16 GiB. GPU 0 has a total capacity of 39.52 GiB of which 34.25 MiB is free. Including non-PyTorch memory, this process has 12.04 GiB memory in use. Process 1811113 has 27.43 GiB memory in use. Of the allocated memory 10.89 GiB is allocated by PyTorch, and 74.09 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s81-p2-resident/training-5031124.log,b05549138b86fef1a567cfcc07e137a91a43d50ceaed861b0ba89d916bc85035
State release,strong-node-qwen25_3b-grpo-math-near_boundary-s82-p2-release,Qwen/Qwen2.5-3B-Instruct,math,16,0.65,False,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s82-p2-release/trial-5031126.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-near_boundary-s83-p2-resident,Qwen/Qwen2.5-3B-Instruct,math,16,0.65,False,False,weight_sync_oom,weight_sync_oom,initializing,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s83-p2-resident/trial-5031131.json,profiles/standard_grpo/same-node-trials.csv,"[""self.checkpoint_manager.update_weights(self.global_steps)"", ""ray.get(self.actor_wg.update_weights(global_steps=global_steps, mode=self.backend))"", ""ray.exceptions.RayTaskError(OutOfMemoryError): ray::WorkerDict.actor_rollout_update_weights() (pid=1974541, ip=10.48.20.3, actor_id=b6accac857978c254ae6069501000000, repr=<verl.single_controller.ray.base.WorkerDict object at 0x145a9667ce30>)"", ""torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.16 GiB. GPU 0 has a total capacity of 39.52 GiB of which 34.25 MiB is free. Including non-PyTorch memory, this process has 12.04 GiB memory in use. Process 1976067 has 27.43 GiB memory in use. Of the allocated memory 10.89 GiB is allocated by PyTorch, and 74.09 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-near_boundary-s83-p2-resident/training-5031131.log,c50e47c4429ef88c0b86bf76b0807d263f518ba81af37b4bff8ccb1e114f10e7
State release,strong-node-qwen25_3b-grpo-math-safe_margin-s81-p2-release,Qwen/Qwen2.5-3B-Instruct,math,2,0.35,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-safe_margin-s81-p2-release/trial-5031603.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-safe_margin-s82-p2-resident,Qwen/Qwen2.5-3B-Instruct,math,2,0.35,True,False,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-safe_margin-s82-p2-resident/trial-5031148.json,profiles/standard_grpo/same-node-trials.csv,,,,
State release,strong-node-qwen25_3b-grpo-math-safe_margin-s83-p2-release,Qwen/Qwen2.5-3B-Instruct,math,2,0.35,True,True,completed,,idle,/mnt/proj3/open-35-44/verl/rl/output/strengthening_same_node_residency/strong-node-qwen25_3b-grpo-math-safe_margin-s83-p2-release/trial-5031149.json,profiles/standard_grpo/same-node-trials.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s241-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s241-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s241-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s241-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=4145275) (Worker pid=4146114) ERROR 09-17 15:02:47 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=4145276) (Worker pid=4146128) ERROR 09-17 15:02:48 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-gsm8k-s241-o0-b4-r70/attempt-5069460/training-5069460.log,b82bbf8f4496b5d2a97d99a153f9f23b8bc96c7cf8b4050a6212756f26a26ec2
Independent offload,of-qwen25-7b-gsm8k-s241-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s241-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s241-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s241-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=53046) (Worker pid=53963) ERROR 09-17 15:24:38 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=53048) (Worker pid=53974) ERROR 09-17 15:24:38 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-gsm8k-s241-o0-b8-r70/attempt-5069460/training-5069460.log,4ad5ec1a9d4ca6dae8f6ac219e48771cb43f5138367a6b74bf55164a300feff8
Independent offload,of-mistral-7b-code_heavy_tail-s243-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=565656) (Worker pid=567539) ERROR 09-17 15:31:22 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=565655) (Worker pid=567544) ERROR 09-17 15:31:22 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-code_heavy_tail-s243-o0-b4-r70/attempt-5069694/training-5069694.log,4970046f0be1c16dd0629121dd882c927207adf7870532559a4cfcf201ae7e92
Independent offload,of-mistral-7b-code_heavy_tail-s243-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s243-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s243-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=634775) (Worker pid=635586) ERROR 09-17 15:44:49 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=634774) (Worker pid=635591) ERROR 09-17 15:44:49 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-code_heavy_tail-s243-o0-b8-r70/attempt-5069694/training-5069694.log,ae37a57f9c6d119f3c666236a93fe19c3a8bf7809b850a7503e29d4d96cfbdbb
Independent offload,of-mistral-7b-code_heavy_tail-s243-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s243-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s243-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s243-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s241-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s241-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s241-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s241-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s241-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s241-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=594970) (Worker pid=595788) ERROR 09-17 16:15:38 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=594974) (Worker pid=595762) ERROR 09-17 16:15:38 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-code_heavy_tail-s241-o0-b4-r70/attempt-5069718/training-5069718.log,6c33651bf81fda6c6b12a50de06fd49404e0d9c68272c439247404d080ea97b8
Independent offload,of-qwen25-7b-code_heavy_tail-s241-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=611088) (Worker pid=612000) ERROR 09-17 16:18:01 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=611090) (Worker pid=612017) ERROR 09-17 16:18:01 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-code_heavy_tail-s241-o0-b8-r70/attempt-5069718/training-5069718.log,5408c4d61519729ec6370b3322ccdc7d532f8d43db15aca63704ee86a02b678e
Independent offload,of-qwen25-7b-code_heavy_tail-s241-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s242-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,math,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s242-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,math,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=794238) (Worker pid=795807) ERROR 09-17 16:21:09 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=794239) (Worker pid=796230) ERROR 09-17 16:21:10 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-math-s242-o0-b4-r70/attempt-5069813/training-5069813.log,19ef009739680f6d08adbff5e5a95d4a5c30bcad39df767ba6820efaa2f2c762
Independent offload,of-qwen25-7b-math-s242-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,math,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s242-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,math,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s242-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,math,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s242-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,math,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=897375) (Worker pid=899326) ERROR 09-17 16:43:30 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=897377) (Worker pid=899363) ERROR 09-17 16:43:32 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-math-s242-o0-b8-r70/attempt-5069813/training-5069813.log,e3b0a72914ed77e6c3687312814a4eafd542a9a535eed13cd7436fa6fdad5673
Independent offload,of-qwen25-7b-math-s242-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,math,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s242-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,math,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s243-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s243-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=679675) (Worker pid=680582) ERROR 09-17 16:31:39 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=679679) (Worker pid=680575) ERROR 09-17 16:31:39 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-gsm8k-s243-o0-b4-r70/attempt-5069888/training-5069888.log,751df854d233fb05677d87e062435718129ce6aad51ca32356d8987aebb6146e
Independent offload,of-qwen25-7b-gsm8k-s243-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s243-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s243-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s243-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=776007) (Worker pid=776797) ERROR 09-17 16:51:56 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=776008) (Worker pid=776817) ERROR 09-17 16:51:57 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-gsm8k-s243-o0-b8-r70/attempt-5069888/training-5069888.log,d62397747dbb1699264e21dc50496bfeb02fe5802389170a282df0c32c7f46c1
Independent offload,of-qwen25-7b-gsm8k-s243-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s243-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s243-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s243-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s243-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s243-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s243-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s243-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=2911030) (Worker pid=2912876) ERROR 09-17 17:27:13 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=2911032) (Worker pid=2912883) ERROR 09-17 17:27:13 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-math-s243-o0-b4-r70/attempt-5069920/training-5069920.log,557338f51863c070b6a30caac4b29c935ca788eb9f2f1d2f0c6fbd5e674bb62c
Independent offload,of-mistral-7b-math-s243-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=2927363) (Worker pid=2929210) ERROR 09-17 17:29:28 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-math-s243-o0-b8-r70/attempt-5069920/training-5069920.log,35dbebd67ec4c75a9bc8de56ef383fb2deb2fbeae2c7af0ef5c770808fea0ea3
Independent offload,of-mistral-7b-math-s243-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s243-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,math,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s243-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,math,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=874356) (Worker pid=875252) ERROR 09-17 17:12:57 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=874357) (Worker pid=875272) ERROR 09-17 17:12:57 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-math-s243-o0-b8-r70/attempt-5069942/training-5069942.log,ac200ad2220fae1a7094a76dd6af2d90f475f41ae02f6a872b70753cfa26426b
Independent offload,of-qwen25-7b-math-s243-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,math,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=890607) (Worker pid=892457) ERROR 09-17 17:15:26 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=890606) (Worker pid=892476) ERROR 09-17 17:15:27 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-math-s243-o0-b4-r70/attempt-5069942/training-5069942.log,011b3207ea3fb304c7f97d4075f9ed897f2cf83cec733de02062f53f3be33298
Independent offload,of-qwen25-7b-math-s243-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,math,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s243-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,math,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s243-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,math,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s243-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,math,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s243-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,math,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s243-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s243-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1371439) (Worker pid=1373022) ERROR 09-17 17:15:30 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1371428) (Worker pid=1373014) ERROR 09-17 17:15:30 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-code_heavy_tail-s243-o0-b8-r70/attempt-5069952/training-5069952.log,ac68e6cfee391de314d881b16c66f20d4a19b68c08be94f53f19ebbb769bb02d
Independent offload,of-qwen25-7b-code_heavy_tail-s243-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1389357) (Worker pid=1390195) ERROR 09-17 17:17:49 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1389358) (Worker pid=1390186) ERROR 09-17 17:17:49 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-code_heavy_tail-s243-o0-b4-r70/attempt-5069952/training-5069952.log,2e8e8d636bf910bd5d84b5679c51d8773c15cd9a06e39b8cfa65d9eedd9916c6
Independent offload,of-qwen25-7b-code_heavy_tail-s243-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s243-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s243-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s243-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s243-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s243-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s243-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s243-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3029365) (Worker pid=3030180) ERROR 09-17 17:51:42 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3029366) (Worker pid=3030175) ERROR 09-17 17:51:42 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-gsm8k-s243-o0-b8-r70/attempt-5069991/training-5069991.log,55f9eb4afb159cead4ca395d840153ef7592d5b462e93840334789e68fa38751
Independent offload,of-mistral-7b-gsm8k-s243-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s243-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s243-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s243-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3129919) (Worker pid=3131768) ERROR 09-17 18:14:14 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3129922) (Worker pid=3131773) ERROR 09-17 18:14:15 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-gsm8k-s243-o0-b4-r70/attempt-5069991/training-5069991.log,2a7d725024574fbf42049ba96d057811b3a4ff9e9a1e7e22334c585cb4bbdb06
Independent offload,of-mistral-7b-gsm8k-s243-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s241-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,math,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s241-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,math,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s241-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,math,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1109097) (Worker pid=1109997) ERROR 09-17 18:06:06 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1109095) (Worker pid=1110013) ERROR 09-17 18:06:07 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-math-s241-o0-b8-r70/attempt-5070001/training-5070001.log,45944951303b93267b7ab45f7617094dc946a65b86a6ad66fecaff68c9f65d65
Independent offload,of-qwen25-7b-math-s241-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,math,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s241-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,math,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s241-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,math,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-math-s241-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,math,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1209242) (Worker pid=1210150) ERROR 09-17 18:27:41 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1209243) (Worker pid=1210145) ERROR 09-17 18:27:42 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-math-s241-o0-b4-r70/attempt-5070001/training-5070001.log,aa629cf275bc356fc8d4f3be3fd9bb6277d7c2263286ecbf3c52719a6dc1821c
Independent offload,of-qwen25-7b-math-s241-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,math,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s241-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s241-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s241-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s241-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3275145) (Worker pid=3284100) ERROR 09-17 18:29:28 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-gsm8k-s241-o0-b4-r70/attempt-5070017/training-5070017.log,61b8fa074787b1096449613562488ebc4b725a9a5cc7c3062847a4b761c1e5a1
Independent offload,of-mistral-7b-gsm8k-s241-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s241-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s241-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s241-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3926225) (Worker pid=3940306) ERROR 09-17 18:47:16 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3926226) (Worker pid=3940313) ERROR 09-17 18:47:16 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-gsm8k-s241-o0-b8-r70/attempt-5070017/training-5070017.log,3be28d3d3a9b542f1e75458352bc8538017330a0516e9d80aea339ae70c5d55b
Independent offload,of-qwen25-7b-gsm8k-s242-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s242-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s242-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3232835) (Worker pid=3233714) ERROR 09-17 18:37:02 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3232834) (Worker pid=3233721) ERROR 09-17 18:37:02 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-gsm8k-s242-o0-b8-r70/attempt-5070024/training-5070024.log,52ec83d3a30453e4c12e3654ac4c90808799d91a3163fe3c332c87179bf9860a
Independent offload,of-qwen25-7b-gsm8k-s242-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s242-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s242-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-gsm8k-s242-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3335948) (Worker pid=3336737) ERROR 09-17 19:00:06 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3335949) (Worker pid=3336772) ERROR 09-17 19:00:08 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-gsm8k-s242-o0-b4-r70/attempt-5070024/training-5070024.log,3cb4a076a857ec63a65f7fe2713f5e6c9d8d4f25e2e5532c845142f6dbef5e34
Independent offload,of-qwen25-7b-gsm8k-s242-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,gsm8k,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s241-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1252124) (Worker pid=1252908) ERROR 09-17 18:36:02 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1252119) (Worker pid=1252920) ERROR 09-17 18:36:02 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-code_heavy_tail-s241-o0-b8-r70/attempt-5070033/training-5070033.log,a13aec1092a8021311b093694eea1b4cbde54723d4705866f633e45d16d86f21
Independent offload,of-mistral-7b-code_heavy_tail-s241-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s241-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s241-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s241-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1347775) (Worker pid=1348565) ERROR 09-17 18:55:34 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1347773) (Worker pid=1348573) ERROR 09-17 18:55:34 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-code_heavy_tail-s241-o0-b4-r70/attempt-5070033/training-5070033.log,5646882709352f38a307e0e9a8e380c00f3f29dd299beea820cd5ab4619cb889
Independent offload,of-mistral-7b-code_heavy_tail-s241-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s241-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s241-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s242-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s242-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=42969) (Worker pid=55165) ERROR 09-17 18:54:58 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-gsm8k-s242-o0-b4-r70/attempt-5070037/training-5070037.log,2e9c2cf6ab5a781a0e7be402a1c2048df9c62b212ee2fc64a2fcc5bfac731fd2
Independent offload,of-mistral-7b-gsm8k-s242-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s242-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s242-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s242-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=784570) (Worker pid=798410) ERROR 09-17 19:13:49 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=784572) (Worker pid=798418) ERROR 09-17 19:13:49 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-gsm8k-s242-o0-b8-r70/attempt-5070037/training-5070037.log,1619179471ca4db6d37a0118aa819f004b56cd8192197b2713ff93f93d0b36b2
Independent offload,of-mistral-7b-gsm8k-s242-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-gsm8k-s242-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,gsm8k,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s241-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s241-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s241-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s241-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s241-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3494701) (Worker pid=3495481) ERROR 09-17 19:36:00 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3494700) (Worker pid=3495494) ERROR 09-17 19:36:00 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-math-s241-o0-b8-r70/attempt-5070050/training-5070050.log,4d5a0b4fd7c95a2eb136b0e3f9e8643a3b3424488438c4ba9f4563fc621a5a22
Independent offload,of-mistral-7b-math-s241-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s241-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s241-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3565900) (Worker pid=3566691) ERROR 09-17 19:50:35 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3565899) (Worker pid=3566704) ERROR 09-17 19:50:35 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-math-s241-o0-b4-r70/attempt-5070050/training-5070050.log,5d8999ae6c33d21be6a3415bdfe8f3a083e51e325a4c144e5dd142b269a0de16
Independent offload,of-mistral-7b-math-s242-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1452103) (Worker pid=1453949) ERROR 09-17 19:19:10 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1452102) (Worker pid=1453961) ERROR 09-17 19:19:11 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes.""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-math-s242-o0-b8-r70/attempt-5070052/training-5070052.log,b6788754a30f5bc7a6cd79e849351745e562830c9973bccb19bc2b918cf6c735
Independent offload,of-mistral-7b-math-s242-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s242-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s242-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s242-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1553024) (Worker pid=1554895) ERROR 09-17 19:41:24 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1553025) (Worker pid=1554900) ERROR 09-17 19:41:24 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-math-s242-o0-b4-r70/attempt-5070052/training-5070052.log,7b7b7f0ebda0ce302c6d0cd1717eef356e76e76836877571f94bb4a1ca434e28
Independent offload,of-mistral-7b-math-s242-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s242-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,math,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-math-s242-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,math,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s242-o0-b4-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1340402) (Worker pid=1355275) ERROR 09-17 19:28:09 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1340531) (Worker pid=1355176) ERROR 09-17 19:28:09 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-code_heavy_tail-s242-o0-b4-r70/attempt-5070058/training-5070058.log,c7a2974ea2a723bf66ca2a04f86079c2937bd2b43a7d5b55f6d55e753d611759
Independent offload,of-qwen25-7b-code_heavy_tail-s242-o0-b8-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s242-o0-b4-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s242-o0-b8-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=1856668) (Worker pid=1870778) ERROR 09-17 19:41:14 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=1856915) (Worker pid=1870864) ERROR 09-17 19:41:14 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.74/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-qwen25-7b-code_heavy_tail-s242-o0-b8-r70/attempt-5070058/training-5070058.log,ff8302caa239c13d30a4cdf4dc716fd600c771ed236368d38466cb0ecb85d15e
Independent offload,of-qwen25-7b-code_heavy_tail-s242-o1-b8-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s242-o1-b4-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s242-o1-b8-r45,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-qwen25-7b-code_heavy_tail-s242-o1-b4-r70,Qwen/Qwen2.5-7B-Instruct,code_heavy_tail,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s242-o1-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s242-o1-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s242-o1-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.45,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s242-o1-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.7,True,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s242-o0-b8-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3691709) (Worker pid=3693559) ERROR 09-17 20:18:20 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3691708) (Worker pid=3693569) ERROR 09-17 20:18:20 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-code_heavy_tail-s242-o0-b8-r70/attempt-5069690/training-5069690.log,fe04a02aa0f46d0971c0fe3365bb924d5e718bbabee8c61ef58e4545feb39c30
Independent offload,of-mistral-7b-code_heavy_tail-s242-o0-b4-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s242-o0-b8-r45,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,8,0.45,False,True,completed,,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,,,,
Independent offload,of-mistral-7b-code_heavy_tail-s242-o0-b4-r70,mistralai/Mistral-7B-Instruct-v0.3,code_heavy_tail,4,0.7,False,True,startup_budget_rejection,rollout_init_memory,,benchmark/offload_factorial/results/slots.csv,benchmark/offload_factorial/results/slots.csv,"[""(vLLMHttpServer pid=3762685) (Worker pid=3763513) ERROR 09-17 20:32:50 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes."", ""(vLLMHttpServer pid=3762686) (Worker pid=3763534) ERROR 09-17 20:32:50 [multiproc_executor.py:852] ValueError: Free memory on device cuda:0 (25.36/39.52 GiB) on startup is less than desired GPU memory utilization (0.7, 27.66 GiB). Decrease GPU memory utilization or reduce GPU memory used by other processes. [repeated 2x across cluster]""]",Observed diagnostic and operation context; allocation ownership and fragmentation unproven,output/offload_factorial/of-mistral-7b-code_heavy_tail-s242-o0-b4-r70/attempt-5069690/training-5069690.log,d2b065bb123ca46e59662ed1e8fc913814ac30c90641acdc9189687fc395c160