chaiml-gemma2-27b-simpo_25125

developer_uid: NischayDnk

submission_id: chaiml-gemma2-27b-simpo_25125_v1

model_name: chaiml-gemma2-27b-simpo_25125_v1

model_group: ChaiML/gemma2-27b-simpos

status: torndown

timestamp: 2025-05-18T12:45:03+00:00

num_battles: 9030

num_wins: 4533

celo_rating: 1282.09

family_friendly_score: 0.5958

family_friendly_standard_error: 0.0069400628239231375

submission_type: basic

model_repo: ChaiML/gemma2-27b-simpos1-retryv2top20ksft-1280

model_architecture: Gemma2ForCausalLM

model_num_parameters: 28731935232.0

best_of: 5

max_input_tokens: 768

max_output_tokens: 64

reward_model: default

latencies: [{'batch_size': 1, 'throughput': 0.38882720718944686, 'latency_mean': 2.5717705130577087, 'latency_p50': 2.5735281705856323, 'latency_p90': 2.8804894208908083}, {'batch_size': 3, 'throughput': 0.8409325786842619, 'latency_mean': 3.544988420009613, 'latency_p50': 3.5413010120391846, 'latency_p90': 3.9351333379745483}, {'batch_size': 5, 'throughput': 1.1413320593137595, 'latency_mean': 4.352846072912216, 'latency_p50': 4.337870717048645, 'latency_p90': 4.884679079055786}, {'batch_size': 6, 'throughput': 1.265767277852098, 'latency_mean': 4.708791847229004, 'latency_p50': 4.742643117904663, 'latency_p90': 5.29916353225708}, {'batch_size': 8, 'throughput': 1.4505378712276074, 'latency_mean': 5.459441409111023, 'latency_p50': 5.443419098854065, 'latency_p90': 6.09176959991455}, {'batch_size': 10, 'throughput': 1.579654146309764, 'latency_mean': 6.24486882686615, 'latency_p50': 6.2368974685668945, 'latency_p90': 7.085559940338134}]

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

display_name: chaiml-gemma2-27b-simpo_25125_v1

is_internal_developer: False

language_model: ChaiML/gemma2-27b-simpos1-retryv2top20ksft-1280

model_size: 29B

ranking_group: single

throughput_3p7s: 0.91

us_pacific_date: 2025-05-18

win_ratio: 0.5019933554817275

generation_params: {'temperature': 0.4, 'top_p': 0.95, 'min_p': 0.025, 'top_k': 60, 'presence_penalty': 0.3, 'frequency_penalty': 0.3, 'stopping_words': ['<|im_end|>', '\n', '<|im_start|>'], 'max_input_tokens': 768, 'best_of': 5, 'max_output_tokens': 64}

formatter: {'memory_template': '<|system|>Family Friendly{memory}\n', 'prompt_template': '', 'bot_template': '<|im_start|>assistant\n{message}<|im_end|>\n', 'user_template': '<|im_start|>user\nYou:{message}<|im_end|>\n', 'response_template': '<|im_start|>assistant\n', '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-gemma2-27b-simpo-25125-v1-mkmlizer
Waiting for job on chaiml-gemma2-27b-simpo-25125-v1-mkmlizer to finish
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║     _____            __           __                                ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║       /___/                                                         ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║                                                                     ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  Version: 0.12.8                                                    ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║                                                                     ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  belonging to:                                                      ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║                                                                     ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║  Expiration: 2028-03-31 23:59:59                                    ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ║                                                                     ║
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: Downloaded to shared memory in 89.079s
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmplmn691oc, device:0
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: quantized model in 68.755s
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: Processed model ChaiML/gemma2-27b-simpos1-retryv2top20ksft-1280 in 157.835s
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: creating bucket guanaco-mkml-models
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1/config.json
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1/special_tokens_map.json
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1/tokenizer_config.json
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.model s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1/tokenizer.model
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1/tokenizer.json
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.2.safetensors s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1/flywheel_model.2.safetensors
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1/flywheel_model.0.safetensors
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/chaiml-gemma2-27b-simpo-25125-v1/flywheel_model.1.safetensors
chaiml-gemma2-27b-simpo-25125-v1-mkmlizer: 
Loading 0:   0%|          | 0/508 [00:00<?, ?it/s]
Loading 0:   1%|          | 4/508 [00:00<00:13, 38.16it/s]
Loading 0:   2%|▏         | 12/508 [00:00<00:11, 44.75it/s]
Loading 0:   3%|▎         | 17/508 [00:00<00:14, 33.77it/s]
Loading 0:   5%|▍         | 25/508 [00:00<00:10, 46.31it/s]
Loading 0:   6%|▌         | 31/508 [00:00<00:17, 27.62it/s]
Loading 0:   7%|▋         | 37/508 [00:01<00:14, 33.24it/s]
Loading 0:   9%|▉         | 45/508 [00:01<00:11, 38.82it/s]
Loading 0:  10%|▉         | 50/508 [00:01<00:12, 36.02it/s]
Loading 0:  11%|█▏        | 58/508 [00:01<00:10, 41.17it/s]
Loading 0:  12%|█▏        | 63/508 [00:01<00:10, 42.41it/s]
Loading 0:  13%|█▎        | 68/508 [00:01<00:11, 38.63it/s]
Loading 0:  14%|█▍        | 73/508 [00:02<00:15, 28.31it/s]
Loading 0:  15%|█▌        | 78/508 [00:02<00:14, 29.07it/s]
Loading 0:  16%|█▌        | 82/508 [00:02<00:14, 29.48it/s]
Loading 0:  17%|█▋        | 87/508 [00:02<00:12, 33.55it/s]
Loading 0:  19%|█▊        | 95/508 [00:02<00:10, 39.15it/s]
Loading 0:  20%|█▉        | 100/508 [00:02<00:11, 35.35it/s]
Loading 0:  21%|██▏       | 108/508 [00:03<00:10, 39.21it/s]
Loading 0:  22%|██▏       | 113/508 [00:03<00:09, 40.90it/s]
Loading 0:  23%|██▎       | 118/508 [00:03<00:12, 31.11it/s]
Loading 0:  24%|██▍       | 124/508 [00:03<00:10, 35.99it/s]
Loading 0:  25%|██▌       | 129/508 [00:03<00:09, 38.51it/s]
Loading 0:  27%|██▋       | 135/508 [00:03<00:09, 38.70it/s]
Loading 0:  28%|██▊       | 140/508 [00:03<00:08, 41.20it/s]
Loading 0:  29%|██▊       | 146/508 [00:04<00:09, 39.76it/s]
Loading 0:  30%|██▉       | 151/508 [00:04<00:08, 42.10it/s]
Loading 0:  31%|███       | 157/508 [00:04<00:08, 40.33it/s]
Loading 0:  32%|███▏      | 162/508 [00:04<00:08, 42.07it/s]
Loading 0:  33%|███▎      | 167/508 [00:04<00:07, 43.58it/s]
Loading 0:  34%|███▍      | 172/508 [00:04<00:11, 29.96it/s]
Loading 0:  35%|███▍      | 176/508 [00:04<00:10, 31.57it/s]
Loading 0:  35%|███▌      | 180/508 [00:04<00:09, 33.39it/s]
Loading 0:  37%|███▋      | 188/508 [00:05<00:07, 40.08it/s]
Loading 0:  38%|███▊      | 193/508 [00:19<04:21,  1.21it/s]
Loading 0:  39%|███▉      | 199/508 [00:19<02:55,  1.76it/s]
Loading 0:  40%|████      | 204/508 [00:20<02:11,  2.31it/s]
Loading 0:  42%|████▏     | 211/508 [00:20<01:25,  3.48it/s]
Loading 0:  43%|████▎     | 216/508 [00:20<01:05,  4.47it/s]
Loading 0:  43%|████▎     | 220/508 [00:20<00:51,  5.63it/s]
Loading 0:  44%|████▍     | 224/508 [00:20<00:39,  7.17it/s]
Loading 0:  46%|████▌     | 232/508 [00:21<00:24, 11.22it/s]
Loading 0:  47%|████▋     | 237/508 [00:21<00:20, 13.39it/s]
Loading 0:  48%|████▊     | 245/508 [00:21<00:14, 18.64it/s]
Loading 0:  49%|████▉     | 250/508 [00:21<00:11, 22.12it/s]
Loading 0:  50%|█████     | 255/508 [00:21<00:10, 24.20it/s]
Loading 0:  51%|█████     | 260/508 [00:21<00:09, 26.81it/s]
Loading 0:  52%|█████▏    | 264/508 [00:22<00:10, 24.32it/s]
Loading 0:  53%|█████▎    | 268/508 [00:22<00:08, 26.80it/s]
Loading 0:  54%|█████▍    | 276/508 [00:22<00:06, 34.54it/s]
Loading 0:  55%|█████▌    | 281/508 [00:22<00:06, 32.70it/s]
Loading 0:  57%|█████▋    | 289/508 [00:22<00:05, 37.90it/s]
Loading 0:  58%|█████▊    | 294/508 [00:22<00:05, 40.00it/s]
Loading 0:  59%|█████▉    | 300/508 [00:22<00:05, 39.49it/s]
Loading 0:  60%|██████    | 305/508 [00:22<00:04, 41.24it/s]
Loading 0:  61%|██████    | 311/508 [00:23<00:04, 44.43it/s]
Loading 0:  62%|██████▏   | 316/508 [00:23<00:06, 29.30it/s]
Loading 0:  63%|██████▎   | 322/508 [00:23<00:05, 34.83it/s]
Loading 0:  64%|██████▍   | 327/508 [00:23<00:04, 37.32it/s]
Loading 0:  65%|██████▌   | 332/508 [00:23<00:04, 40.16it/s]
Loading 0:  66%|██████▋   | 337/508 [00:23<00:04, 35.65it/s]
Loading 0:  68%|██████▊   | 344/508 [00:24<00:04, 37.65it/s]
Loading 0:  69%|██████▊   | 349/508 [00:24<00:03, 40.03it/s]
Loading 0:  70%|██████▉   | 354/508 [00:24<00:04, 37.24it/s]
Loading 0:  71%|███████   | 359/508 [00:24<00:04, 30.55it/s]
Loading 0:  71%|███████▏  | 363/508 [00:24<00:04, 32.12it/s]
Loading 0:  72%|███████▏  | 367/508 [00:24<00:04, 33.77it/s]
Loading 0:  74%|███████▍  | 375/508 [00:24<00:03, 41.07it/s]
Loading 0:  75%|███████▍  | 380/508 [00:25<00:03, 36.16it/s]
Loading 0:  76%|███████▋  | 388/508 [00:25<00:02, 40.20it/s]
Loading 0:  77%|███████▋  | 393/508 [00:25<00:02, 42.19it/s]
Loading 0:  78%|███████▊  | 398/508 [00:25<00:02, 38.69it/s]
Loading 0:  79%|███████▉  | 403/508 [00:25<00:02, 38.63it/s]
Loading 0:  80%|████████  | 407/508 [00:25<00:03, 30.82it/s]
Loading 0:  81%|████████  | 412/508 [00:26<00:03, 31.47it/s]
Loading 0:  83%|████████▎ | 421/508 [00:26<00:02, 38.33it/s]
Loading 0:  84%|████████▎ | 425/508 [00:26<00:02, 38.16it/s]
Loading 0:  85%|████████▌ | 432/508 [00:26<00:01, 38.87it/s]
Loading 0:  86%|████████▌ | 436/508 [00:26<00:01, 38.41it/s]
Loading 0:  87%|████████▋ | 443/508 [00:26<00:01, 38.57it/s]
Loading 0:  88%|████████▊ | 447/508 [00:26<00:01, 38.14it/s]
Loading 0:  89%|████████▉ | 451/508 [00:41<00:01, 38.14it/s]
Loading 0:  89%|████████▉ | 452/508 [00:41<00:48,  1.15it/s]
Loading 0:  90%|████████▉ | 456/508 [00:41<00:34,  1.51it/s]
Loading 0:  91%|█████████ | 460/508 [00:42<00:24,  1.96it/s]
Loading 0:  92%|█████████▏| 466/508 [00:42<00:13,  3.01it/s]
Loading 0:  93%|█████████▎| 474/508 [00:42<00:06,  4.88it/s]
Loading 0:  94%|█████████▍| 479/508 [00:42<00:04,  6.25it/s]
Loading 0:  96%|█████████▌| 487/508 [00:42<00:02,  9.39it/s]
Loading 0:  97%|█████████▋| 492/508 [00:42<00:01, 11.79it/s]
Loading 0:  98%|█████████▊| 497/508 [00:42<00:00, 14.19it/s]
Loading 0:  99%|█████████▉| 502/508 [00:43<00:00, 15.58it/s]
Loading 0: 100%|█████████▉| 506/508 [00:43<00:00, 17.50it/s]
                                                            
Job chaiml-gemma2-27b-simpo-25125-v1-mkmlizer completed after 187.22s with status: succeeded
Stopping job with name chaiml-gemma2-27b-simpo-25125-v1-mkmlizer
Pipeline stage MKMLizer completed in 187.75s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.14s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service chaiml-gemma2-27b-simpo-25125-v1
Waiting for inference service chaiml-gemma2-27b-simpo-25125-v1 to be ready
Inference service chaiml-gemma2-27b-simpo-25125-v1 ready after 140.74553442001343s
Pipeline stage MKMLDeployer completed in 141.32s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 3.364431142807007s
Received healthy response to inference request in 2.9429094791412354s
Received healthy response to inference request in 2.6266984939575195s
Received healthy response to inference request in 2.907759189605713s
Received healthy response to inference request in 2.631890296936035s
5 requests
0 failed requests
5th percentile: 2.6277368545532225
10th percentile: 2.628775215148926
20th percentile: 2.630851936340332
30th percentile: 2.6870640754699706
40th percentile: 2.797411632537842
50th percentile: 2.907759189605713
60th percentile: 2.921819305419922
70th percentile: 2.935879421234131
80th percentile: 3.0272138118743896
90th percentile: 3.1958224773406982
95th percentile: 3.2801268100738525
99th percentile: 3.347570276260376
mean time: 2.894737720489502
Pipeline stage StressChecker completed in 15.78s
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 0.66s
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 0.97s
Shutdown handler de-registered
chaiml-gemma2-27b-simpo_25125_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 0.10s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service chaiml-gemma2-27b-simpo-25125-v1-profiler
Waiting for inference service chaiml-gemma2-27b-simpo-25125-v1-profiler to be ready
Inference service chaiml-gemma2-27b-simpo-25125-v1-profiler ready after 150.94930791854858s
Pipeline stage MKMLProfilerDeployer completed in 151.50s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/chaiml-gemma2-27b-si22a77b17af0dd4594fbe3adc6d642f24-deplonv5jr:/code/chaiverse_profiler_1747572897 --namespace tenant-chaiml-guanaco
kubectl exec -it chaiml-gemma2-27b-si22a77b17af0dd4594fbe3adc6d642f24-deplonv5jr --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1747572897 && python profiles.py profile --best_of_n 5 --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_1747572897/summary.json'
kubectl exec -it chaiml-gemma2-27b-si22a77b17af0dd4594fbe3adc6d642f24-deplonv5jr --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1747572897/summary.json'
Pipeline stage MKMLProfilerRunner completed in 1361.62s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service chaiml-gemma2-27b-simpo-25125-v1-profiler is running
Tearing down inference service chaiml-gemma2-27b-simpo-25125-v1-profiler
Service chaiml-gemma2-27b-simpo-25125-v1-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 2.04s
Shutdown handler de-registered
Shutdown handler registered
run pipeline %s
run pipeline stage %s
Running pipeline stage OfflineFamilyFriendlyScorer
Evaluating %s Family Friendly Score with %s threads
%s, retrying in %s seconds...
Evaluating %s Family Friendly Score with %s threads
Pipeline stage OfflineFamilyFriendlyScorer completed in 4754.86s
Shutdown handler de-registered
chaiml-gemma2-27b-simpo_25125_v1 status is now inactive due to auto deactivation removed underperforming models
chaiml-gemma2-27b-simpo_25125_v1 status is now torndown due to DeploymentManager action