rirv938-grpo-to-compare-_1872

developer_uid: rirv938

submission_id: rirv938-grpo-to-compare-_1872_v1

model_name: rirv938-grpo-to-compare-_1872_v1

model_group: rirv938/grpo_to_compare_

status: torndown

timestamp: 2025-04-28T18:48:57+00:00

num_battles: 8600

num_wins: 4388

celo_rating: 1298.24

family_friendly_score: 0.5722

family_friendly_standard_error: 0.006996958767921961

submission_type: basic

model_repo: rirv938/grpo_to_compare_to_dpo_10k_cp78_with_ff_merged

model_architecture: MistralForCausalLM

model_num_parameters: 24096691200.0

best_of: 8

max_input_tokens: 1024

max_output_tokens: 64

reward_model: default

latencies: [{'batch_size': 1, 'throughput': 0.5344034132575795, 'latency_mean': 1.8711787450313568, 'latency_p50': 1.860215425491333, 'latency_p90': 2.091338849067688}, {'batch_size': 3, 'throughput': 1.0804391680285281, 'latency_mean': 2.761113873720169, 'latency_p50': 2.7500431537628174, 'latency_p90': 3.0703973293304445}, {'batch_size': 5, 'throughput': 1.393500447742533, 'latency_mean': 3.5750496113300323, 'latency_p50': 3.5981149673461914, 'latency_p90': 3.9958406925201415}, {'batch_size': 6, 'throughput': 1.4966595954923667, 'latency_mean': 3.9927329027652743, 'latency_p50': 3.9993319511413574, 'latency_p90': 4.424893403053284}, {'batch_size': 8, 'throughput': 1.6545029247383163, 'latency_mean': 4.794714866876602, 'latency_p50': 4.769564867019653, 'latency_p90': 5.376355481147766}, {'batch_size': 10, 'throughput': 1.7313361139779213, 'latency_mean': 5.714030598402023, 'latency_p50': 5.740723609924316, 'latency_p90': 6.511422967910766}]

gpu_counts: {'NVIDIA A100-SXM4-80GB': 1}

display_name: rirv938-grpo-to-compare-_1872_v1

is_internal_developer: True

language_model: rirv938/grpo_to_compare_to_dpo_10k_cp78_with_ff_merged

model_size: 24B

ranking_group: single

throughput_3p7s: 1.43

us_pacific_date: 2025-04-28

win_ratio: 0.5102325581395348

generation_params: {'temperature': 0.9, 'top_p': 0.9, 'min_p': 0.2, 'top_k': 80, 'presence_penalty': 0.5, 'frequency_penalty': 0.5, 'stopping_words': ['</s>', 'You:', '###', '\n'], 'max_input_tokens': 1024, '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 rirv938-grpo-to-compare-1872-v1-mkmlizer
Waiting for job on rirv938-grpo-to-compare-1872-v1-mkmlizer to finish
rirv938-grpo-to-compare-1872-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║     _____            __           __                                ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║       /___/                                                         ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║                                                                     ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  Version: 0.12.8                                                    ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  https://mk1.ai                                                     ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║                                                                     ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  belonging to:                                                      ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║                                                                     ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  Chai Research Corp.                                                ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║  Expiration: 2028-03-31 23:59:59                                    ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ║                                                                     ║
rirv938-grpo-to-compare-1872-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
rirv938-grpo-to-compare-1872-v1-mkmlizer: Downloaded to shared memory in 181.773s
rirv938-grpo-to-compare-1872-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpmr_vioww, device:0
rirv938-grpo-to-compare-1872-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
rirv938-grpo-to-compare-1872-v1-mkmlizer: quantized model in 62.508s
rirv938-grpo-to-compare-1872-v1-mkmlizer: Processed model rirv938/grpo_to_compare_to_dpo_10k_cp78_with_ff_merged in 244.282s
rirv938-grpo-to-compare-1872-v1-mkmlizer: creating bucket guanaco-mkml-models
rirv938-grpo-to-compare-1872-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
rirv938-grpo-to-compare-1872-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/rirv938-grpo-to-compare-1872-v1
rirv938-grpo-to-compare-1872-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/rirv938-grpo-to-compare-1872-v1/config.json
rirv938-grpo-to-compare-1872-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/rirv938-grpo-to-compare-1872-v1/special_tokens_map.json
rirv938-grpo-to-compare-1872-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/rirv938-grpo-to-compare-1872-v1/tokenizer_config.json
rirv938-grpo-to-compare-1872-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/rirv938-grpo-to-compare-1872-v1/tokenizer.json
rirv938-grpo-to-compare-1872-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/rirv938-grpo-to-compare-1872-v1/flywheel_model.1.safetensors
rirv938-grpo-to-compare-1872-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/rirv938-grpo-to-compare-1872-v1/flywheel_model.0.safetensors
rirv938-grpo-to-compare-1872-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|          | 4/363 [00:00<00:16, 22.27it/s]
Loading 0:   2%|▏         | 7/363 [00:00<00:18, 18.89it/s]
Loading 0:   3%|▎         | 11/363 [00:00<00:14, 23.81it/s]
Loading 0:   4%|▍         | 14/363 [00:00<00:23, 14.72it/s]
Loading 0:   4%|▍         | 16/363 [00:00<00:24, 14.15it/s]
Loading 0:   6%|▌         | 21/363 [00:01<00:19, 17.92it/s]
Loading 0:   6%|▋         | 23/363 [00:01<00:24, 14.07it/s]
Loading 0:   8%|▊         | 29/363 [00:01<00:15, 21.85it/s]
Loading 0:   9%|▉         | 33/363 [00:01<00:16, 20.26it/s]
Loading 0:  10%|▉         | 36/363 [00:02<00:21, 15.15it/s]
Loading 0:  11%|█         | 39/363 [00:02<00:18, 17.23it/s]
Loading 0:  12%|█▏        | 42/363 [00:02<00:20, 15.45it/s]
Loading 0:  13%|█▎        | 46/363 [00:02<00:16, 19.44it/s]
Loading 0:  14%|█▍        | 50/363 [00:02<00:13, 23.36it/s]
Loading 0:  15%|█▍        | 53/363 [00:03<00:18, 16.52it/s]
Loading 0:  15%|█▌        | 56/363 [00:03<00:18, 16.68it/s]
Loading 0:  16%|█▋        | 59/363 [00:03<00:23, 12.74it/s]
Loading 0:  18%|█▊        | 64/363 [00:03<00:17, 17.55it/s]
Loading 0:  19%|█▊        | 68/363 [00:03<00:14, 20.46it/s]
Loading 0:  20%|█▉        | 71/363 [00:04<00:19, 15.22it/s]
Loading 0:  20%|██        | 74/363 [00:04<00:18, 15.86it/s]
Loading 0:  21%|██        | 77/363 [00:04<00:22, 12.67it/s]
Loading 0:  23%|██▎       | 82/363 [00:04<00:15, 17.88it/s]
Loading 0:  24%|██▎       | 86/363 [00:04<00:12, 21.57it/s]
Loading 0:  25%|██▍       | 90/363 [00:05<00:19, 14.30it/s]
Loading 0:  26%|██▋       | 96/363 [00:05<00:13, 20.29it/s]
Loading 0:  28%|██▊       | 101/363 [00:05<00:12, 21.77it/s]
Loading 0:  29%|██▉       | 105/363 [00:06<00:15, 16.54it/s]
Loading 0:  30%|██▉       | 108/363 [00:06<00:18, 13.68it/s]
Loading 0:  31%|███       | 111/363 [00:06<00:16, 15.58it/s]
Loading 0:  31%|███▏      | 114/363 [00:06<00:17, 14.45it/s]
Loading 0:  33%|███▎      | 118/363 [00:06<00:13, 18.04it/s]
Loading 0:  34%|███▎      | 122/363 [00:07<00:11, 21.50it/s]
Loading 0:  34%|███▍      | 125/363 [00:07<00:14, 15.92it/s]
Loading 0:  35%|███▌      | 128/363 [00:07<00:14, 16.57it/s]
Loading 0:  36%|███▌      | 131/363 [00:07<00:17, 13.27it/s]
Loading 0:  38%|███▊      | 137/363 [00:07<00:11, 20.00it/s]
Loading 0:  39%|███▉      | 141/363 [00:08<00:11, 19.48it/s]
Loading 0:  40%|███▉      | 144/363 [00:08<00:14, 14.98it/s]
Loading 0:  40%|████      | 147/363 [00:08<00:12, 16.98it/s]
Loading 0:  41%|████▏     | 150/363 [00:08<00:13, 15.59it/s]
Loading 0:  43%|████▎     | 155/363 [00:08<00:09, 21.02it/s]
Loading 0:  44%|████▍     | 159/363 [00:09<00:10, 20.14it/s]
Loading 0:  45%|████▍     | 162/363 [00:09<00:13, 15.14it/s]
Loading 0:  45%|████▌     | 165/363 [00:09<00:11, 17.23it/s]
Loading 0:  46%|████▋     | 168/363 [00:09<00:12, 15.78it/s]
Loading 0:  48%|████▊     | 173/363 [00:09<00:08, 21.26it/s]
Loading 0:  49%|████▉     | 177/363 [00:10<00:09, 20.38it/s]
Loading 0:  50%|████▉     | 180/363 [00:10<00:11, 15.27it/s]
Loading 0:  50%|█████     | 183/363 [00:10<00:10, 17.36it/s]
Loading 0:  51%|█████     | 186/363 [00:10<00:11, 15.35it/s]
Loading 0:  52%|█████▏    | 190/363 [00:11<00:08, 19.35it/s]
Loading 0:  53%|█████▎    | 194/363 [00:11<00:07, 22.71it/s]
Loading 0:  54%|█████▍    | 197/363 [00:11<00:10, 16.48it/s]
Loading 0:  55%|█████▌    | 200/363 [00:11<00:09, 17.06it/s]
Loading 0:  56%|█████▌    | 203/363 [00:26<03:51,  1.45s/it]
Loading 0:  57%|█████▋    | 207/363 [00:26<02:29,  1.05it/s]
Loading 0:  58%|█████▊    | 210/363 [00:27<01:48,  1.41it/s]
Loading 0:  59%|█████▊    | 213/363 [00:27<01:20,  1.87it/s]
Loading 0:  60%|█████▉    | 216/363 [00:27<01:01,  2.40it/s]
Loading 0:  60%|██████    | 219/363 [00:27<00:44,  3.26it/s]
Loading 0:  61%|██████    | 222/363 [00:27<00:33,  4.15it/s]
Loading 0:  63%|██████▎   | 227/363 [00:28<00:20,  6.65it/s]
Loading 0:  64%|██████▎   | 231/363 [00:28<00:16,  8.24it/s]
Loading 0:  64%|██████▍   | 234/363 [00:28<00:15,  8.34it/s]
Loading 0:  65%|██████▌   | 237/363 [00:28<00:12, 10.26it/s]
Loading 0:  66%|██████▌   | 240/363 [00:29<00:11, 10.74it/s]
Loading 0:  67%|██████▋   | 245/363 [00:29<00:07, 15.42it/s]
Loading 0:  69%|██████▊   | 249/363 [00:29<00:07, 16.03it/s]
Loading 0:  69%|██████▉   | 252/363 [00:29<00:08, 13.03it/s]
Loading 0:  70%|███████   | 255/363 [00:29<00:07, 15.15it/s]
Loading 0:  71%|███████   | 258/363 [00:30<00:07, 14.14it/s]
Loading 0:  72%|███████▏  | 263/363 [00:30<00:05, 19.44it/s]
Loading 0:  74%|███████▎  | 267/363 [00:30<00:05, 18.88it/s]
Loading 0:  74%|███████▍  | 270/363 [00:30<00:06, 14.59it/s]
Loading 0:  75%|███████▌  | 273/363 [00:30<00:05, 16.61it/s]
Loading 0:  76%|███████▌  | 276/363 [00:31<00:05, 15.02it/s]
Loading 0:  77%|███████▋  | 281/363 [00:31<00:04, 20.44it/s]
Loading 0:  79%|███████▊  | 285/363 [00:31<00:04, 19.18it/s]
Loading 0:  79%|███████▉  | 288/363 [00:31<00:05, 14.65it/s]
Loading 0:  80%|████████  | 291/363 [00:31<00:04, 16.80it/s]
Loading 0:  81%|████████  | 294/363 [00:32<00:04, 15.39it/s]
Loading 0:  82%|████████▏ | 299/363 [00:32<00:03, 20.82it/s]
Loading 0:  83%|████████▎ | 303/363 [00:32<00:03, 19.91it/s]
Loading 0:  84%|████████▍ | 306/363 [00:32<00:03, 14.95it/s]
Loading 0:  85%|████████▌ | 309/363 [00:32<00:03, 17.02it/s]
Loading 0:  86%|████████▌ | 312/363 [00:33<00:03, 14.92it/s]
Loading 0:  87%|████████▋ | 317/363 [00:33<00:02, 20.44it/s]
Loading 0:  88%|████████▊ | 321/363 [00:33<00:02, 19.29it/s]
Loading 0:  89%|████████▉ | 324/363 [00:33<00:02, 14.54it/s]
Loading 0:  90%|█████████ | 327/363 [00:34<00:02, 16.65it/s]
Loading 0:  91%|█████████ | 330/363 [00:34<00:02, 14.95it/s]
Loading 0:  92%|█████████▏| 335/363 [00:34<00:01, 20.39it/s]
Loading 0:  93%|█████████▎| 339/363 [00:34<00:01, 19.02it/s]
Loading 0:  94%|█████████▍| 342/363 [00:34<00:01, 14.53it/s]
Loading 0:  95%|█████████▌| 345/363 [00:35<00:01, 16.56it/s]
Loading 0:  96%|█████████▌| 348/363 [00:35<00:01, 14.60it/s]
Loading 0:  97%|█████████▋| 352/363 [00:35<00:00, 17.78it/s]
Loading 0:  98%|█████████▊| 356/363 [00:35<00:00, 21.35it/s]
Loading 0:  99%|█████████▉| 359/363 [00:42<00:02,  1.51it/s]
Loading 0:  99%|█████████▉| 361/363 [00:43<00:01,  1.80it/s]
                                                            
Job rirv938-grpo-to-compare-1872-v1-mkmlizer completed after 269.09s with status: succeeded
Stopping job with name rirv938-grpo-to-compare-1872-v1-mkmlizer
Pipeline stage MKMLizer completed in 269.57s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.17s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service rirv938-grpo-to-compare-1872-v1
Waiting for inference service rirv938-grpo-to-compare-1872-v1 to be ready
Failed to get response for submission chaiml-damien-mafia-bos_79245_v2: HTTPConnectionPool(host='chaiml-damien-mafia-bos-79245-v2-predictor.tenant-chaiml-guanaco.k.chaiverse.com', port=80): Read timed out. (read timeout=12.0)
Inference service rirv938-grpo-to-compare-1872-v1 ready after 130.6938648223877s
Pipeline stage MKMLDeployer completed in 131.22s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.810529947280884s
Received healthy response to inference request in 2.156482458114624s
Received healthy response to inference request in 2.004863977432251s
Received healthy response to inference request in 1.8953518867492676s
Failed to get response for submission chaiml-damien-mafia-bos_79245_v2: HTTPConnectionPool(host='chaiml-damien-mafia-bos-79245-v2-predictor.tenant-chaiml-guanaco.k.chaiverse.com', port=80): Read timed out. (read timeout=12.0)
Received healthy response to inference request in 1.951094388961792s
5 requests
0 failed requests
5th percentile: 1.9065003871917725
10th percentile: 1.9176488876342774
20th percentile: 1.939945888519287
30th percentile: 1.9618483066558838
40th percentile: 1.9833561420440673
50th percentile: 2.004863977432251
60th percentile: 2.0655113697052
70th percentile: 2.1261587619781492
80th percentile: 2.287291955947876
90th percentile: 2.54891095161438
95th percentile: 2.6797204494476317
99th percentile: 2.784368047714233
mean time: 2.163664531707764
Pipeline stage StressChecker completed in 12.18s
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 0.74s
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 0.73s
Shutdown handler de-registered
rirv938-grpo-to-compare-_1872_v1 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 0.11s
run pipeline stage %s
Running pipeline stage MKMLProfilerTemplater
Pipeline stage MKMLProfilerTemplater completed in 0.11s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service rirv938-grpo-to-compare-1872-v1-profiler
Waiting for inference service rirv938-grpo-to-compare-1872-v1-profiler to be ready
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 2946.07s
Shutdown handler de-registered
rirv938-grpo-to-compare-_1872_v1 status is now inactive due to auto deactivation removed underperforming models
rirv938-grpo-to-compare-_1872_v1 status is now torndown due to DeploymentManager action