chaiml-anthropic-grpo-4_66492

developer_uid: rirv938

submission_id: chaiml-anthropic-grpo-4_66492_v4

model_name: chaiml-anthropic-grpo-4_66492_v4

model_group: ChaiML/anthropic_grpo_40

status: torndown

timestamp: 2025-05-13T17:20:15+00:00

num_battles: 6557

num_wins: 3374

celo_rating: 1299.83

family_friendly_score: 0.5424

family_friendly_standard_error: 0.0070455977744972075

submission_type: basic

model_repo: ChaiML/anthropic_grpo_40k_cp1184_95ff

model_architecture: MistralForCausalLM

model_num_parameters: 12772070400.0

best_of: 8

max_input_tokens: 768

max_output_tokens: 64

reward_model: default

latencies: [{'batch_size': 1, 'throughput': 0.6490013052630378, 'latency_mean': 1.5407641232013702, 'latency_p50': 1.5350933074951172, 'latency_p90': 1.7082532405853272}, {'batch_size': 3, 'throughput': 1.2445808334428803, 'latency_mean': 2.406676172018051, 'latency_p50': 2.4273871183395386, 'latency_p90': 2.649578070640564}, {'batch_size': 5, 'throughput': 1.5412544576033624, 'latency_mean': 3.238113578557968, 'latency_p50': 3.239138603210449, 'latency_p90': 3.5965510606765747}, {'batch_size': 6, 'throughput': 1.6199818473604102, 'latency_mean': 3.6777357840538025, 'latency_p50': 3.6851214170455933, 'latency_p90': 4.159850907325745}, {'batch_size': 8, 'throughput': 1.7046931943520838, 'latency_mean': 4.657099670171737, 'latency_p50': 4.660675883293152, 'latency_p90': 5.247479844093323}, {'batch_size': 10, 'throughput': 1.7596165730215965, 'latency_mean': 5.648541433811188, 'latency_p50': 5.632616639137268, 'latency_p90': 6.39850926399231}]

gpu_counts: {'NVIDIA RTX A5000': 1}

display_name: chaiml-anthropic-grpo-4_66492_v4

is_internal_developer: True

language_model: ChaiML/anthropic_grpo_40k_cp1184_95ff

model_size: 13B

ranking_group: single

throughput_3p7s: 1.63

us_pacific_date: 2025-05-13

win_ratio: 0.5145645874637792

generation_params: {'temperature': 1.0, 'top_p': 1.0, 'min_p': 0.0, 'top_k': 80, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['###', 'You:', '</s>', '\n'], 'max_input_tokens': 768, 'best_of': 8, 'max_output_tokens': 64}

formatter: {'memory_template': '', 'prompt_template': '', 'bot_template': '{bot_name}: {message}\n', 'user_template': '{user_name}: {message}\n', 'response_template': '{bot_name}:', 'truncate_by_message': False}

Resubmit model

Shutdown handler not registered because Python interpreter is not running in the main thread
run pipeline %s
run pipeline stage %s
Running pipeline stage MKMLizer
Starting job with name chaiml-anthropic-grpo-4-66492-v4-mkmlizer
Waiting for job on chaiml-anthropic-grpo-4-66492-v4-mkmlizer to finish
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║     _____            __           __                                ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║       /___/                                                         ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║                                                                     ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  Version: 0.12.8                                                    ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║                                                                     ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  belonging to:                                                      ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║                                                                     ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║  Expiration: 2028-03-31 23:59:59                                    ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ║                                                                     ║
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: Downloaded to shared memory in 99.910s
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpo3oaurk_, device:0
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: quantized model in 43.346s
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: Processed model ChaiML/anthropic_grpo_40k_cp1184_95ff in 143.257s
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: creating bucket guanaco-mkml-models
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-anthropic-grpo-4-66492-v4
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-anthropic-grpo-4-66492-v4/config.json
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-anthropic-grpo-4-66492-v4/special_tokens_map.json
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-anthropic-grpo-4-66492-v4/tokenizer_config.json
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-anthropic-grpo-4-66492-v4/tokenizer.json
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-anthropic-grpo-4-66492-v4/flywheel_model.0.safetensors
chaiml-anthropic-grpo-4-66492-v4-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|▏         | 5/363 [00:00<00:16, 22.00it/s]
Loading 0:   3%|▎         | 10/363 [00:00<00:13, 27.14it/s]
Loading 0:   4%|▍         | 14/363 [00:00<00:14, 24.18it/s]
Loading 0:   5%|▌         | 19/363 [00:00<00:11, 30.31it/s]
Loading 0:   6%|▋         | 23/363 [00:00<00:15, 21.83it/s]
Loading 0:   7%|▋         | 26/363 [00:01<00:17, 19.41it/s]
Loading 0:   9%|▊         | 31/363 [00:01<00:13, 24.90it/s]
Loading 0:  10%|▉         | 35/363 [00:01<00:12, 25.82it/s]
Loading 0:  11%|█         | 39/363 [00:01<00:12, 26.63it/s]
Loading 0:  12%|█▏        | 42/363 [00:01<00:13, 24.34it/s]
Loading 0:  13%|█▎        | 46/363 [00:01<00:11, 26.75it/s]
Loading 0:  14%|█▍        | 50/363 [00:02<00:13, 23.67it/s]
Loading 0:  15%|█▌        | 55/363 [00:02<00:10, 28.34it/s]
Loading 0:  17%|█▋        | 60/363 [00:02<00:10, 29.39it/s]
Loading 0:  18%|█▊        | 64/363 [00:02<00:15, 19.35it/s]
Loading 0:  19%|█▉        | 69/363 [00:02<00:12, 23.93it/s]
Loading 0:  20%|██        | 73/363 [00:02<00:11, 24.26it/s]
Loading 0:  21%|██        | 77/363 [00:03<00:12, 22.86it/s]
Loading 0:  23%|██▎       | 84/363 [00:03<00:09, 29.15it/s]
Loading 0:  24%|██▍       | 88/363 [00:03<00:09, 28.50it/s]
Loading 0:  26%|██▌       | 93/363 [00:03<00:09, 29.83it/s]
Loading 0:  27%|██▋       | 97/363 [00:03<00:09, 27.85it/s]
Loading 0:  28%|██▊       | 101/363 [00:04<00:11, 22.24it/s]
Loading 0:  29%|██▊       | 104/363 [00:04<00:12, 20.02it/s]
Loading 0:  30%|███       | 109/363 [00:04<00:10, 24.99it/s]
Loading 0:  31%|███       | 113/363 [00:04<00:10, 23.10it/s]
Loading 0:  33%|███▎      | 118/363 [00:04<00:08, 27.72it/s]
Loading 0:  34%|███▎      | 122/363 [00:04<00:09, 24.88it/s]
Loading 0:  35%|███▍      | 127/363 [00:05<00:08, 28.76it/s]
Loading 0:  36%|███▌      | 131/363 [00:05<00:09, 24.91it/s]
Loading 0:  37%|███▋      | 136/363 [00:05<00:07, 29.11it/s]
Loading 0:  39%|███▉      | 141/363 [00:05<00:07, 30.00it/s]
Loading 0:  40%|███▉      | 145/363 [00:05<00:10, 20.79it/s]
Loading 0:  41%|████      | 149/363 [00:06<00:10, 20.15it/s]
Loading 0:  42%|████▏     | 154/363 [00:06<00:08, 24.43it/s]
Loading 0:  44%|████▎     | 158/363 [00:06<00:09, 22.29it/s]
Loading 0:  45%|████▍     | 163/363 [00:06<00:07, 26.40it/s]
Loading 0:  46%|████▌     | 167/363 [00:06<00:08, 23.65it/s]
Loading 0:  47%|████▋     | 172/363 [00:06<00:06, 27.74it/s]
Loading 0:  48%|████▊     | 176/363 [00:07<00:07, 24.69it/s]
Loading 0:  50%|████▉     | 181/363 [00:07<00:06, 28.79it/s]
Loading 0:  51%|█████     | 185/363 [00:07<00:09, 19.00it/s]
Loading 0:  52%|█████▏    | 190/363 [00:07<00:07, 23.39it/s]
Loading 0:  53%|█████▎    | 194/363 [00:07<00:07, 21.80it/s]
Loading 0:  55%|█████▍    | 199/363 [00:08<00:06, 26.05it/s]
Loading 0:  56%|█████▌    | 203/363 [00:08<00:06, 23.14it/s]
Loading 0:  57%|█████▋    | 208/363 [00:08<00:05, 26.72it/s]
Loading 0:  58%|█████▊    | 212/363 [00:08<00:06, 23.72it/s]
Loading 0:  60%|█████▉    | 217/363 [00:08<00:05, 27.63it/s]
Loading 0:  61%|██████    | 222/363 [00:08<00:04, 28.53it/s]
Loading 0:  62%|██████▏   | 226/363 [00:09<00:06, 20.34it/s]
Loading 0:  63%|██████▎   | 230/363 [00:09<00:06, 20.01it/s]
Loading 0:  65%|██████▍   | 235/363 [00:09<00:05, 24.73it/s]
Loading 0:  66%|██████▌   | 239/363 [00:09<00:05, 23.32it/s]
Loading 0:  67%|██████▋   | 244/363 [00:09<00:04, 27.44it/s]
Loading 0:  68%|██████▊   | 248/363 [00:10<00:04, 24.39it/s]
Loading 0:  70%|██████▉   | 253/363 [00:10<00:03, 28.68it/s]
Loading 0:  71%|███████   | 257/363 [00:10<00:04, 25.29it/s]
Loading 0:  72%|███████▏  | 262/363 [00:10<00:03, 29.60it/s]
Loading 0:  73%|███████▎  | 266/363 [00:10<00:04, 20.05it/s]
Loading 0:  75%|███████▍  | 271/363 [00:10<00:03, 24.73it/s]
Loading 0:  76%|███████▌  | 275/363 [00:11<00:03, 23.51it/s]
Loading 0:  78%|███████▊  | 282/363 [00:11<00:02, 29.71it/s]
Loading 0:  79%|███████▉  | 286/363 [00:11<00:02, 27.55it/s]
Loading 0:  80%|████████  | 291/363 [00:11<00:02, 29.85it/s]
Loading 0:  81%|████████▏ | 295/363 [00:11<00:02, 28.27it/s]
Loading 0:  82%|████████▏ | 299/363 [00:11<00:02, 27.41it/s]
Loading 0:  84%|████████▎ | 304/363 [00:12<00:02, 24.29it/s]
Loading 0:  85%|████████▍ | 307/363 [00:12<00:02, 23.13it/s]
Loading 0:  86%|████████▌ | 311/363 [00:12<00:02, 22.52it/s]
Loading 0:  88%|████████▊ | 318/363 [00:12<00:01, 28.97it/s]
Loading 0:  89%|████████▊ | 322/363 [00:12<00:01, 25.90it/s]
Loading 0:  90%|████████▉ | 326/363 [00:12<00:01, 28.60it/s]
Loading 0:  91%|█████████ | 330/363 [00:13<00:01, 25.20it/s]
Loading 0:  92%|█████████▏| 334/363 [00:13<00:01, 27.73it/s]
Loading 0:  93%|█████████▎| 338/363 [00:13<00:01, 24.95it/s]
Loading 0:  95%|█████████▍| 344/363 [00:20<00:09,  2.10it/s]
Loading 0:  96%|█████████▌| 348/363 [00:20<00:05,  2.76it/s]
Loading 0:  97%|█████████▋| 353/363 [00:20<00:02,  3.91it/s]
Loading 0:  98%|█████████▊| 357/363 [00:21<00:01,  4.98it/s]
Loading 0: 100%|█████████▉| 362/363 [00:21<00:00,  7.00it/s]
                                                            
Job chaiml-anthropic-grpo-4-66492-v4-mkmlizer completed after 178.21s with status: succeeded
Stopping job with name chaiml-anthropic-grpo-4-66492-v4-mkmlizer
Pipeline stage MKMLizer completed in 178.69s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.15s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service chaiml-anthropic-grpo-4-66492-v4
Waiting for inference service chaiml-anthropic-grpo-4-66492-v4 to be ready
Inference service chaiml-anthropic-grpo-4-66492-v4 ready after 110.69851565361023s
Pipeline stage MKMLDeployer completed in 111.26s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.363682270050049s
Received healthy response to inference request in 1.630692481994629s
Received healthy response to inference request in 1.844346523284912s
Received healthy response to inference request in 1.7231178283691406s
Received healthy response to inference request in 1.61214280128479s
5 requests
0 failed requests
5th percentile: 1.6158527374267577
10th percentile: 1.6195626735687256
20th percentile: 1.6269825458526612
30th percentile: 1.6491775512695312
40th percentile: 1.686147689819336
50th percentile: 1.7231178283691406
60th percentile: 1.7716093063354492
70th percentile: 1.8201007843017578
80th percentile: 1.9482136726379395
90th percentile: 2.155947971343994
95th percentile: 2.2598151206970214
99th percentile: 2.3429088401794433
mean time: 1.834796380996704
Pipeline stage StressChecker completed in 10.49s
run pipeline stage %s
Running pipeline stage OfflineFamilyFriendlyTriggerPipeline
run_pipeline:run_in_cloud %s
starting trigger_guanaco_pipeline args=%s
triggered trigger_guanaco_pipeline args=%s
Pipeline stage OfflineFamilyFriendlyTriggerPipeline completed in 0.82s
run pipeline stage %s
Running pipeline stage TriggerMKMLProfilingPipeline
run_pipeline:run_in_cloud %s
starting trigger_guanaco_pipeline args=%s
triggered trigger_guanaco_pipeline args=%s
Pipeline stage TriggerMKMLProfilingPipeline completed in 0.75s
Shutdown handler de-registered
chaiml-anthropic-grpo-4_66492_v4 status is now deployed due to DeploymentManager action
Shutdown handler registered
run pipeline %s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Skipping teardown as no inference service was successfully deployed
Pipeline stage MKMLProfilerDeleter completed in 0.10s
run pipeline stage %s
Running pipeline stage MKMLProfilerTemplater
Pipeline stage MKMLProfilerTemplater completed in 0.10s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service chaiml-anthropic-grpo-4-66492-v4-profiler
Waiting for inference service chaiml-anthropic-grpo-4-66492-v4-profiler to be ready
Inference service chaiml-anthropic-grpo-4-66492-v4-profiler ready after 120.78048396110535s
Pipeline stage MKMLProfilerDeployer completed in 121.47s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/chaiml-anthropic-grp19e47c96a679a073226d05c810d1f788-deplo6rxs6:/code/chaiverse_profiler_1747157361 --namespace tenant-chaiml-guanaco
kubectl exec -it chaiml-anthropic-grp19e47c96a679a073226d05c810d1f788-deplo6rxs6 --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1747157361 && python profiles.py profile --best_of_n 8 --auto_batch 5 --batches 1,5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80,85,90,95,100,105,110,115,120,125,130,135,140,145,150,155,160,165,170,175,180,185,190,195 --samples 200 --input_tokens 768 --output_tokens 64 --summary /code/chaiverse_profiler_1747157361/summary.json'
Shutdown handler registered
run pipeline %s
run pipeline stage %s
Running pipeline stage OfflineFamilyFriendlyScorer
Evaluating %s Family Friendly Score with %s threads
Pipeline stage OfflineFamilyFriendlyScorer completed in 2658.43s
Shutdown handler de-registered
chaiml-anthropic-grpo-4_66492_v4 status is now inactive due to auto deactivation removed underperforming models
chaiml-anthropic-grpo-4_66492_v4 status is now torndown due to DeploymentManager action