rirv938-grpo-to-compare_45958

developer_uid: rirv938

submission_id: rirv938-grpo-to-compare_45958_v1

model_name: rirv938-grpo-to-compare_45958_v1

model_group: rirv938/grpo_to_compare_

status: torndown

timestamp: 2025-04-25T17:30:59+00:00

num_battles: 6513

num_wins: 3274

celo_rating: 1287.14

family_friendly_score: 0.5648

family_friendly_standard_error: 0.007011432949119602

submission_type: basic

model_repo: rirv938/grpo_to_compare_to_dpo_10k_cp310_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': 0.5301682297031153, 'latency_mean': 1.8861262929439544, 'latency_p50': 1.871593952178955, 'latency_p90': 2.0947454929351808}, {'batch_size': 3, 'throughput': 1.0672735112804466, 'latency_mean': 2.8005813264846804, 'latency_p50': 2.813004732131958, 'latency_p90': 3.0725358247756955}, {'batch_size': 5, 'throughput': 1.3863640815716836, 'latency_mean': 3.5899108767509462, 'latency_p50': 3.583096742630005, 'latency_p90': 3.9834800004959106}, {'batch_size': 6, 'throughput': 1.4926366655464636, 'latency_mean': 3.989792412519455, 'latency_p50': 3.989112615585327, 'latency_p90': 4.46379714012146}, {'batch_size': 8, 'throughput': 1.63021151759527, 'latency_mean': 4.86085434794426, 'latency_p50': 4.896331071853638, 'latency_p90': 5.376937985420227}, {'batch_size': 10, 'throughput': 1.7208662817748555, 'latency_mean': 5.745702894926072, 'latency_p50': 5.660673379898071, 'latency_p90': 6.550105428695678}]

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

display_name: rirv938-grpo-to-compare_45958_v1

is_internal_developer: True

language_model: rirv938/grpo_to_compare_to_dpo_10k_cp310_merged

model_size: 24B

ranking_group: single

throughput_3p7s: 1.43

us_pacific_date: 2025-04-25

win_ratio: 0.5026869338246583

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>', '\n', '###', 'You:'], '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-45958-v1-mkmlizer
Waiting for job on rirv938-grpo-to-compare-45958-v1-mkmlizer to finish
rirv938-grpo-to-compare-45958-v1-mkmlizer: Downloaded to shared memory in 203.109s
rirv938-grpo-to-compare-45958-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpe8u94n2l, device:0
rirv938-grpo-to-compare-45958-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
Unable to record family friendly update due to error: Invalid JSON input: JSON must contain 'User Safety' and 'Response Safety' fields
rirv938-grpo-to-compare-45958-v1-mkmlizer: quantized model in 64.067s
rirv938-grpo-to-compare-45958-v1-mkmlizer: Processed model rirv938/grpo_to_compare_to_dpo_10k_cp310_merged in 267.209s
rirv938-grpo-to-compare-45958-v1-mkmlizer: creating bucket guanaco-mkml-models
rirv938-grpo-to-compare-45958-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
rirv938-grpo-to-compare-45958-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/rirv938-grpo-to-compare-45958-v1
rirv938-grpo-to-compare-45958-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/rirv938-grpo-to-compare-45958-v1/config.json
rirv938-grpo-to-compare-45958-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/rirv938-grpo-to-compare-45958-v1/special_tokens_map.json
rirv938-grpo-to-compare-45958-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/rirv938-grpo-to-compare-45958-v1/tokenizer_config.json
rirv938-grpo-to-compare-45958-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/rirv938-grpo-to-compare-45958-v1/tokenizer.json
rirv938-grpo-to-compare-45958-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/rirv938-grpo-to-compare-45958-v1/flywheel_model.1.safetensors
rirv938-grpo-to-compare-45958-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/rirv938-grpo-to-compare-45958-v1/flywheel_model.0.safetensors
rirv938-grpo-to-compare-45958-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|          | 4/363 [00:00<00:17, 20.84it/s]
Loading 0:   2%|▏         | 7/363 [00:00<00:20, 17.65it/s]
Loading 0:   3%|▎         | 11/363 [00:00<00:16, 21.90it/s]
Loading 0:   4%|▍         | 14/363 [00:00<00:25, 13.95it/s]
Loading 0:   4%|▍         | 16/363 [00:01<00:26, 13.30it/s]
Loading 0:   5%|▌         | 19/363 [00:01<00:20, 16.42it/s]
Loading 0:   6%|▌         | 21/363 [00:01<00:20, 16.70it/s]
Loading 0:   6%|▋         | 23/363 [00:01<00:26, 12.66it/s]
Loading 0:   8%|▊         | 28/363 [00:01<00:17, 19.44it/s]
Loading 0:   9%|▉         | 32/363 [00:01<00:14, 22.44it/s]
Loading 0:  10%|▉         | 35/363 [00:02<00:21, 14.97it/s]
Loading 0:  10%|█         | 38/363 [00:02<00:21, 15.37it/s]
Loading 0:  11%|█         | 40/363 [00:02<00:23, 13.70it/s]
Loading 0:  12%|█▏        | 42/363 [00:02<00:24, 13.19it/s]
Loading 0:  13%|█▎        | 46/363 [00:02<00:17, 17.68it/s]
Loading 0:  14%|█▍        | 50/363 [00:02<00:14, 21.77it/s]
Loading 0:  15%|█▍        | 53/363 [00:03<00:20, 15.15it/s]
Loading 0:  15%|█▌        | 56/363 [00:03<00:19, 15.54it/s]
Loading 0:  16%|█▌        | 58/363 [00:03<00:22, 13.76it/s]
Loading 0:  17%|█▋        | 60/363 [00:03<00:23, 13.16it/s]
Loading 0:  18%|█▊        | 64/363 [00:03<00:17, 17.21it/s]
Loading 0:  19%|█▊        | 68/363 [00:04<00:13, 21.22it/s]
Loading 0:  20%|█▉        | 71/363 [00:04<00:19, 14.93it/s]
Loading 0:  20%|██        | 73/363 [00:04<00:20, 14.09it/s]
Loading 0:  21%|██        | 75/363 [00:04<00:19, 14.95it/s]
Loading 0:  21%|██        | 77/363 [00:04<00:24, 11.85it/s]
Loading 0:  23%|██▎       | 82/363 [00:05<00:15, 17.91it/s]
Loading 0:  24%|██▎       | 86/363 [00:05<00:12, 21.45it/s]
Loading 0:  25%|██▍       | 89/363 [00:05<00:18, 15.06it/s]
Loading 0:  25%|██▌       | 92/363 [00:05<00:17, 15.51it/s]
Loading 0:  26%|██▌       | 95/363 [00:05<00:14, 17.87it/s]
Loading 0:  27%|██▋       | 99/363 [00:05<00:12, 21.55it/s]
Loading 0:  28%|██▊       | 102/363 [00:06<00:13, 19.32it/s]
Loading 0:  29%|██▉       | 105/363 [00:06<00:17, 14.50it/s]
Loading 0:  29%|██▉       | 107/363 [00:06<00:19, 13.34it/s]
Loading 0:  30%|███       | 109/363 [00:06<00:19, 13.06it/s]
Loading 0:  31%|███       | 111/363 [00:06<00:17, 14.19it/s]
Loading 0:  31%|███       | 113/363 [00:07<00:21, 11.57it/s]
Loading 0:  33%|███▎      | 118/363 [00:07<00:13, 17.95it/s]
Loading 0:  34%|███▎      | 122/363 [00:07<00:11, 21.63it/s]
Loading 0:  34%|███▍      | 125/363 [00:07<00:15, 15.16it/s]
Loading 0:  35%|███▌      | 128/363 [00:08<00:15, 15.56it/s]
Loading 0:  36%|███▌      | 130/363 [00:08<00:16, 13.91it/s]
Loading 0:  36%|███▋      | 132/363 [00:08<00:17, 13.36it/s]
Loading 0:  37%|███▋      | 136/363 [00:08<00:12, 17.78it/s]
Loading 0:  39%|███▊      | 140/363 [00:08<00:10, 21.55it/s]
Loading 0:  39%|███▉      | 143/363 [00:08<00:14, 15.02it/s]
Loading 0:  40%|████      | 146/363 [00:09<00:13, 15.68it/s]
Loading 0:  41%|████      | 148/363 [00:09<00:15, 13.85it/s]
Loading 0:  41%|████▏     | 150/363 [00:09<00:16, 13.25it/s]
Loading 0:  42%|████▏     | 154/363 [00:09<00:11, 17.42it/s]
Loading 0:  44%|████▎     | 158/363 [00:09<00:09, 21.74it/s]
Loading 0:  44%|████▍     | 161/363 [00:10<00:13, 15.30it/s]
Loading 0:  45%|████▌     | 164/363 [00:10<00:12, 15.93it/s]
Loading 0:  46%|████▌     | 166/363 [00:10<00:13, 14.13it/s]
Loading 0:  46%|████▋     | 168/363 [00:10<00:14, 13.43it/s]
Loading 0:  47%|████▋     | 172/363 [00:10<00:10, 17.87it/s]
Loading 0:  48%|████▊     | 176/363 [00:10<00:08, 22.21it/s]
Loading 0:  49%|████▉     | 179/363 [00:11<00:11, 15.62it/s]
Loading 0:  50%|█████     | 182/363 [00:11<00:11, 16.37it/s]
Loading 0:  51%|█████     | 185/363 [00:11<00:14, 12.59it/s]
Loading 0:  52%|█████▏    | 190/363 [00:11<00:09, 17.87it/s]
Loading 0:  53%|█████▎    | 194/363 [00:11<00:07, 21.50it/s]
Loading 0:  54%|█████▍    | 197/363 [00:12<00:10, 15.83it/s]
Loading 0:  55%|█████▌    | 200/363 [00:12<00:10, 16.13it/s]
Loading 0:  55%|█████▌    | 200/363 [00:27<00:10, 16.13it/s]
Loading 0:  55%|█████▌    | 201/363 [00:27<04:36,  1.70s/it]
Loading 0:  56%|█████▌    | 203/363 [00:27<03:31,  1.32s/it]
Loading 0:  57%|█████▋    | 207/363 [00:27<02:03,  1.26it/s]
Loading 0:  58%|█████▊    | 210/363 [00:27<01:26,  1.77it/s]
Loading 0:  59%|█████▊    | 213/363 [00:27<01:03,  2.37it/s]
Loading 0:  59%|█████▉    | 215/363 [00:28<00:51,  2.86it/s]
Loading 0:  60%|█████▉    | 217/363 [00:28<00:41,  3.51it/s]
Loading 0:  60%|██████    | 219/363 [00:28<00:32,  4.44it/s]
Loading 0:  61%|██████    | 221/363 [00:28<00:29,  4.87it/s]
Loading 0:  62%|██████▏   | 226/363 [00:28<00:15,  8.61it/s]
Loading 0:  63%|██████▎   | 230/363 [00:28<00:11, 11.76it/s]
Loading 0:  64%|██████▍   | 233/363 [00:29<00:12, 10.67it/s]
Loading 0:  65%|██████▌   | 236/363 [00:29<00:10, 11.94it/s]
Loading 0:  66%|██████▌   | 238/363 [00:29<00:11, 11.02it/s]
Loading 0:  66%|██████▌   | 240/363 [00:29<00:11, 11.02it/s]
Loading 0:  67%|██████▋   | 244/363 [00:29<00:07, 15.03it/s]
Loading 0:  68%|██████▊   | 248/363 [00:30<00:06, 18.83it/s]
Loading 0:  69%|██████▉   | 251/363 [00:30<00:08, 13.88it/s]
Loading 0:  70%|██████▉   | 253/363 [00:30<00:08, 13.36it/s]
Loading 0:  70%|███████   | 255/363 [00:30<00:07, 14.02it/s]
Loading 0:  71%|███████   | 257/363 [00:31<00:09, 11.17it/s]
Loading 0:  72%|███████▏  | 262/363 [00:31<00:05, 17.27it/s]
Loading 0:  73%|███████▎  | 266/363 [00:31<00:04, 20.75it/s]
Loading 0:  74%|███████▍  | 269/363 [00:31<00:06, 14.61it/s]
Loading 0:  75%|███████▍  | 272/363 [00:31<00:05, 15.28it/s]
Loading 0:  75%|███████▌  | 274/363 [00:32<00:06, 13.20it/s]
Loading 0:  76%|███████▌  | 276/363 [00:32<00:06, 12.56it/s]
Loading 0:  77%|███████▋  | 280/363 [00:32<00:04, 16.81it/s]
Loading 0:  78%|███████▊  | 284/363 [00:32<00:03, 20.94it/s]
Loading 0:  79%|███████▉  | 287/363 [00:32<00:05, 14.89it/s]
Loading 0:  80%|███████▉  | 290/363 [00:32<00:04, 15.72it/s]
Loading 0:  80%|████████  | 292/363 [00:33<00:05, 13.72it/s]
Loading 0:  81%|████████  | 294/363 [00:33<00:05, 13.15it/s]
Loading 0:  82%|████████▏ | 298/363 [00:33<00:03, 17.11it/s]
Loading 0:  83%|████████▎ | 302/363 [00:33<00:03, 20.25it/s]
Loading 0:  84%|████████▍ | 305/363 [00:33<00:04, 14.02it/s]
Loading 0:  85%|████████▍ | 307/363 [00:34<00:04, 13.61it/s]
Loading 0:  85%|████████▌ | 309/363 [00:34<00:03, 14.35it/s]
Loading 0:  86%|████████▌ | 311/363 [00:34<00:04, 11.71it/s]
Loading 0:  87%|████████▋ | 316/363 [00:34<00:02, 17.68it/s]
Loading 0:  88%|████████▊ | 320/363 [00:34<00:01, 21.53it/s]
Loading 0:  89%|████████▉ | 323/363 [00:35<00:02, 15.29it/s]
Loading 0:  90%|████████▉ | 326/363 [00:35<00:02, 16.21it/s]
Loading 0:  91%|█████████ | 329/363 [00:35<00:02, 12.49it/s]
Loading 0:  92%|█████████▏| 334/363 [00:35<00:01, 17.79it/s]
Loading 0:  93%|█████████▎| 338/363 [00:35<00:01, 21.50it/s]
Loading 0:  94%|█████████▍| 341/363 [00:36<00:01, 16.02it/s]
Loading 0:  95%|█████████▍| 344/363 [00:36<00:01, 16.48it/s]
Loading 0:  96%|█████████▌| 347/363 [00:36<00:01, 12.42it/s]
Loading 0:  97%|█████████▋| 353/363 [00:36<00:00, 18.81it/s]
Loading 0:  98%|█████████▊| 357/363 [00:43<00:03,  1.79it/s]
Loading 0:  99%|█████████▉| 360/363 [00:44<00:01,  2.27it/s]
Loading 0: 100%|█████████▉| 362/363 [00:44<00:00,  2.66it/s]
                                                            
Job rirv938-grpo-to-compare-45958-v1-mkmlizer completed after 441.68s with status: succeeded
Stopping job with name rirv938-grpo-to-compare-45958-v1-mkmlizer
Pipeline stage MKMLizer completed in 442.14s
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 rirv938-grpo-to-compare-45958-v1
Waiting for inference service rirv938-grpo-to-compare-45958-v1 to be ready
Inference service rirv938-grpo-to-compare-45958-v1 ready after 140.72078013420105s
Pipeline stage MKMLDeployer completed in 141.15s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.341794490814209s
Received healthy response to inference request in 1.8682963848114014s
Received healthy response to inference request in 2.1134119033813477s
Received healthy response to inference request in 1.9722862243652344s
Received healthy response to inference request in 1.6121718883514404s
5 requests
0 failed requests
5th percentile: 1.6633967876434326
10th percentile: 1.7146216869354247
20th percentile: 1.8170714855194092
30th percentile: 1.889094352722168
40th percentile: 1.9306902885437012
50th percentile: 1.9722862243652344
60th percentile: 2.0287364959716796
70th percentile: 2.085186767578125
80th percentile: 2.1590884208679197
90th percentile: 2.2504414558410644
95th percentile: 2.296117973327637
99th percentile: 2.3326591873168945
mean time: 1.9815921783447266
Pipeline stage StressChecker completed in 11.51s
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 0.69s
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 0.69s
Shutdown handler de-registered
rirv938-grpo-to-compare_45958_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 0.10s
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-45958-v1-profiler
Waiting for inference service rirv938-grpo-to-compare-45958-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 3057.72s
Shutdown handler de-registered
rirv938-grpo-to-compare_45958_v1 status is now inactive due to auto deactivation removed underperforming models
rirv938-grpo-to-compare_45958_v1 status is now torndown due to DeploymentManager action