chaiml-ca18-v2-dpo-cosine

developer_uid: chai_backend_admin

submission_id: chaiml-ca18-v2-dpo-cosine_v1

model_name: training123

model_group: ChaiML/ca18-v2-dpo_cosin

status: torndown

timestamp: 2025-11-19T20:56:19+00:00

num_battles: 5137

num_wins: 2708

celo_rating: 1303.0

family_friendly_score: 0.5087999999999999

family_friendly_standard_error: 0.007069972560059904

submission_type: basic

model_repo: ChaiML/ca18-v2-dpo_cosine

model_architecture: MistralForCausalLM

model_num_parameters: 12772070400.0

best_of: 8

max_input_tokens: 2048

max_output_tokens: 64

reward_model: default

latencies: [{'batch_size': 1, 'throughput': 0.6812889167374833, 'latency_mean': 1.4676836550235748, 'latency_p50': 1.4793986082077026, 'latency_p90': 1.604146385192871}, {'batch_size': 3, 'throughput': 1.260763542808667, 'latency_mean': 2.366007710695267, 'latency_p50': 2.356663942337036, 'latency_p90': 2.612159180641174}, {'batch_size': 5, 'throughput': 1.5390650079461314, 'latency_mean': 3.235734589099884, 'latency_p50': 3.223978042602539, 'latency_p90': 3.653255248069763}, {'batch_size': 6, 'throughput': 1.6239244830574908, 'latency_mean': 3.6870118248462678, 'latency_p50': 3.6834532022476196, 'latency_p90': 4.203342032432556}, {'batch_size': 8, 'throughput': 1.7319001545297104, 'latency_mean': 4.589511811733246, 'latency_p50': 4.594688653945923, 'latency_p90': 5.206905841827393}, {'batch_size': 10, 'throughput': 1.7979069357782815, 'latency_mean': 5.516645040512085, 'latency_p50': 5.5545408725738525, 'latency_p90': 6.290609812736511}]

gpu_counts: {'NVIDIA L40S': 1}

display_name: training123

is_internal_developer: True

language_model: ChaiML/ca18-v2-dpo_cosine

model_size: 13B

ranking_group: single

throughput_3p7s: 1.63

us_pacific_date: 2025-11-16

win_ratio: 0.5271559275841932

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:', '\n', 'User:', '</s>'], 'max_input_tokens': 2048, 'best_of': 8, 'max_output_tokens': 64}

formatter: {'memory_template': "{bot_name}'s Persona: {memory}\n####\n", 'prompt_template': '{prompt}\n<START>\n', 'bot_template': '{bot_name}: {message}\n', 'user_template': '{user_name}: {message}\n', 'response_template': '{bot_name}:', 'truncate_by_message': True}

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-ca18-v2-dpo-cosine-v1-mkmlizer
Waiting for job on chaiml-ca18-v2-dpo-cosine-v1-mkmlizer to finish
Stopping job with name chaiml-ca18-v2-dpo-cosine-v1-mkmlizer
%s, retrying in %s seconds...
Starting job with name chaiml-ca18-v2-dpo-cosine-v1-mkmlizer
Waiting for job on chaiml-ca18-v2-dpo-cosine-v1-mkmlizer to finish
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║   ██████   ██████  █████   ████  ████                               ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  ░░██████ ██████  ░░███   ███░  ░░███                               ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║   ░███░█████░███   ░███  ███     ░███                               ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║   ░███░░███ ░███   ░███████      ░███                               ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║   ░███ ░░░  ░███   ░███░░███     ░███                               ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║   ░███      ░███   ░███ ░░███    ░███                               ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║   █████     █████  █████ ░░████  █████                              ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  ░░░░░     ░░░░░  ░░░░░   ░░░░  ░░░░░                               ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  Version: 0.30.2                                                    ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  Features: FLYWHEEL, CUDA                                           ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  Copyright 2023-2025 MK ONE TECHNOLOGIES Inc.                       ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  belonging to:                                                      ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║  Expiration: 2028-03-31 23:59:59                                    ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: Downloaded to shared memory in 26.260s
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: Checking if ChaiML/ca18-v2-dpo_cosine already exists in ChaiML
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpi0g8ol2x, device:0
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: quantized model in 25.981s
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: Processed model ChaiML/ca18-v2-dpo_cosine in 52.242s
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: creating bucket guanaco-mkml-models
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-ca18-v2-dpo-cosine-v1/nvidia
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-ca18-v2-dpo-cosine-v1/nvidia/config.json
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-ca18-v2-dpo-cosine-v1/nvidia/special_tokens_map.json
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-ca18-v2-dpo-cosine-v1/nvidia/tokenizer_config.json
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-ca18-v2-dpo-cosine-v1/nvidia/tokenizer.json
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-ca18-v2-dpo-cosine-v1/nvidia/flywheel_model.0.safetensors
chaiml-ca18-v2-dpo-cosine-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|▏         | 5/363 [00:00<00:10, 33.34it/s]
Loading 0:   4%|▍         | 14/363 [00:00<00:07, 47.71it/s]
Loading 0:   6%|▋         | 23/363 [00:00<00:06, 52.38it/s]
Loading 0:   9%|▉         | 32/363 [00:00<00:06, 54.02it/s]
Loading 0:  11%|█▏        | 41/363 [00:00<00:05, 55.31it/s]
Loading 0:  14%|█▍        | 50/363 [00:00<00:05, 56.31it/s]
Loading 0:  17%|█▋        | 60/363 [00:01<00:04, 62.00it/s]
Loading 0:  18%|█▊        | 67/363 [00:01<00:06, 45.94it/s]
Loading 0:  20%|██        | 73/363 [00:01<00:06, 44.41it/s]
Loading 0:  23%|██▎       | 82/363 [00:01<00:05, 47.82it/s]
Loading 0:  25%|██▌       | 91/363 [00:01<00:05, 50.15it/s]
Loading 0:  28%|██▊       | 100/363 [00:01<00:05, 52.08it/s]
Loading 0:  30%|███       | 110/363 [00:02<00:04, 57.78it/s]
Loading 0:  32%|███▏      | 116/363 [00:02<00:04, 57.05it/s]
Loading 0:  34%|███▍      | 123/363 [00:02<00:04, 54.43it/s]
Loading 0:  36%|███▌      | 129/363 [00:02<00:04, 54.68it/s]
Loading 0:  37%|███▋      | 136/363 [00:02<00:04, 49.57it/s]
Loading 0:  39%|███▉      | 142/363 [00:02<00:05, 44.18it/s]
Loading 0:  40%|████      | 147/363 [00:02<00:04, 44.06it/s]
Loading 0:  42%|████▏     | 153/363 [00:03<00:04, 47.68it/s]
Loading 0:  44%|████▎     | 158/363 [00:03<00:04, 42.64it/s]
Loading 0:  46%|████▌     | 166/363 [00:03<00:03, 51.03it/s]
Loading 0:  47%|████▋     | 172/363 [00:03<00:03, 49.67it/s]
Loading 0:  49%|████▉     | 178/363 [00:03<00:03, 50.79it/s]
Loading 0:  51%|█████     | 185/363 [00:03<00:03, 47.84it/s]
Loading 0:  53%|█████▎    | 193/363 [00:03<00:03, 54.86it/s]
Loading 0:  55%|█████▍    | 199/363 [00:03<00:03, 52.31it/s]
Loading 0:  56%|█████▋    | 205/363 [00:04<00:02, 52.88it/s]
Loading 0:  58%|█████▊    | 211/363 [00:04<00:02, 54.69it/s]
Loading 0:  60%|█████▉    | 217/363 [00:04<00:02, 52.37it/s]
Loading 0:  61%|██████▏   | 223/363 [00:04<00:03, 41.21it/s]
Loading 0:  63%|██████▎   | 228/363 [00:04<00:03, 41.41it/s]
Loading 0:  64%|██████▍   | 233/363 [00:04<00:03, 42.95it/s]
Loading 0:  66%|██████▌   | 239/363 [00:04<00:02, 41.46it/s]
Loading 0:  68%|██████▊   | 247/363 [00:04<00:02, 50.06it/s]
Loading 0:  70%|██████▉   | 253/363 [00:05<00:02, 48.76it/s]
Loading 0:  71%|███████▏  | 259/363 [00:05<00:02, 50.26it/s]
Loading 0:  73%|███████▎  | 266/363 [00:05<00:02, 47.81it/s]
Loading 0:  75%|███████▌  | 274/363 [00:05<00:01, 54.71it/s]
Loading 0:  77%|███████▋  | 280/363 [00:05<00:01, 52.32it/s]
Loading 0:  79%|███████▉  | 286/363 [00:05<00:01, 52.35it/s]
Loading 0:  80%|████████  | 292/363 [00:05<00:01, 54.13it/s]
Loading 0:  82%|████████▏ | 298/363 [00:05<00:01, 51.44it/s]
Loading 0:  84%|████████▎ | 304/363 [00:06<00:01, 30.74it/s]
Loading 0:  85%|████████▌ | 309/363 [00:06<00:01, 33.95it/s]
Loading 0:  87%|████████▋ | 314/363 [00:06<00:01, 35.35it/s]
Loading 0:  88%|████████▊ | 320/363 [00:06<00:01, 40.62it/s]
Loading 0:  90%|████████▉ | 326/363 [00:06<00:00, 43.23it/s]
Loading 0:  91%|█████████ | 331/363 [00:06<00:00, 44.52it/s]
Loading 0:  93%|█████████▎| 339/363 [00:07<00:00, 45.94it/s]
Loading 0:  96%|█████████▌| 347/363 [00:07<00:00, 53.97it/s]
Loading 0:  97%|█████████▋| 353/363 [00:07<00:00, 52.33it/s]
Loading 0:  99%|█████████▉| 359/363 [00:07<00:00, 53.09it/s]
                                                            
Job chaiml-ca18-v2-dpo-cosine-v1-mkmlizer completed after 398.2s with status: succeeded
Stopping job with name chaiml-ca18-v2-dpo-cosine-v1-mkmlizer
Pipeline stage MKMLizer completed in 764.44s
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-ca18-v2-dpo-cosine-v1
Waiting for inference service chaiml-ca18-v2-dpo-cosine-v1 to be ready
Tearing down inference service chaiml-ca18-v2-dpo-cosine-v1
%s, retrying in %s seconds...
Creating inference service chaiml-ca18-v2-dpo-cosine-v1
Waiting for inference service chaiml-ca18-v2-dpo-cosine-v1 to be ready
Inference service chaiml-ca18-v2-dpo-cosine-v1 ready after 201.36379671096802s
Pipeline stage MKMLDeployer completed in 806.85s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 1.7288105487823486s
Received healthy response to inference request in 1.2233405113220215s
Received healthy response to inference request in 1.332601547241211s
Received healthy response to inference request in 1.2839272022247314s
Received healthy response to inference request in 2.15939998626709s
5 requests
0 failed requests
5th percentile: 1.2354578495025634
10th percentile: 1.2475751876831054
20th percentile: 1.2718098640441895
30th percentile: 1.2936620712280273
40th percentile: 1.3131318092346191
50th percentile: 1.332601547241211
60th percentile: 1.491085147857666
70th percentile: 1.649568748474121
80th percentile: 1.814928436279297
90th percentile: 1.9871642112731933
95th percentile: 2.0732820987701417
99th percentile: 2.1421764087677
mean time: 1.5456159591674805
Pipeline stage StressChecker completed in 9.27s
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 0.62s
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 0.63s
Shutdown handler de-registered
chaiml-ca18-v2-dpo-cosine_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 0.08s
run pipeline stage %s
Running pipeline stage MKMLProfilerTemplater
Pipeline stage MKMLProfilerTemplater completed in 0.09s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service chaiml-ca18-v2-dpo-cosine-v1-profiler
Waiting for inference service chaiml-ca18-v2-dpo-cosine-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 2604.83s
Shutdown handler de-registered
chaiml-ca18-v2-dpo-cosine_v1 status is now inactive due to auto deactivation removed underperforming models
chaiml-ca18-v2-dpo-cosine_v1 status is now torndown due to DeploymentManager action