developer_uid: chai_backend_admin
submission_id: chaiml-kimid-v2-cosine_v1
model_name: training123
model_group: ChaiML/kimid-v2_cosine
status: inactive
timestamp: 2025-12-04T13:16:30+00:00
num_battles: 9059
num_wins: 4677
celo_rating: 1306.55
family_friendly_score: 0.5082
family_friendly_standard_error: 0.007070116830717863
submission_type: basic
model_repo: ChaiML/kimid-v2_cosine
model_architecture: MistralForCausalLM
model_num_parameters: 24096691200.0
best_of: 8
max_input_tokens: 2048
max_output_tokens: 72
reward_model: default
latencies: [{'batch_size': 1, 'throughput': 0.3455549115747941, 'latency_mean': 2.8938184034824372, 'latency_p50': 2.9270756244659424, 'latency_p90': 3.139516282081604}, {'batch_size': 2, 'throughput': 0.5288196122506986, 'latency_mean': 3.7745577836036683, 'latency_p50': 3.7723735570907593, 'latency_p90': 4.11743860244751}, {'batch_size': 3, 'throughput': 0.6639115595237269, 'latency_mean': 4.5107854497432704, 'latency_p50': 4.496939420700073, 'latency_p90': 4.888136696815491}, {'batch_size': 4, 'throughput': 0.751805517671504, 'latency_mean': 5.2886528503894805, 'latency_p50': 5.305958390235901, 'latency_p90': 5.838206005096436}, {'batch_size': 5, 'throughput': 0.8183875497560127, 'latency_mean': 6.08056457400322, 'latency_p50': 6.047418594360352, 'latency_p90': 6.973308372497558}]
gpu_counts: {'NVIDIA L40S': 1}
display_name: training123
ineligible_reason: max_output_tokens!=64
is_internal_developer: True
language_model: ChaiML/kimid-v2_cosine
model_size: 24B
ranking_group: single
throughput_3p7s: 0.52
us_pacific_date: 2025-12-04
win_ratio: 0.5162821503477205
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': ['</s>', 'You:', 'User:'], 'max_input_tokens': 2048, 'best_of': 8, 'max_output_tokens': 72}
formatter: {'memory_template': "{bot_name}'s Persona: {memory}\n####\n", 'prompt_template': '', 'bot_template': '{bot_name}: {message}\n', 'user_template': '{user_name}: {message}\n', 'response_template': '####\n{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-kimid-v2-cosine-v1-mkmlizer
Waiting for job on chaiml-kimid-v2-cosine-v1-mkmlizer to finish
chaiml-kimid-v2-cosine-v1-mkmlizer: bash: cannot set terminal process group (-1): Inappropriate ioctl for device
chaiml-kimid-v2-cosine-v1-mkmlizer: bash: no job control in this shell
chaiml-kimid-v2-cosine-v1-mkmlizer: /root/miniconda3/envs/nvidia/lib/python3.11/site-packages/mk1/__init__.py:1: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81.
chaiml-kimid-v2-cosine-v1-mkmlizer:   __import__('pkg_resources').declare_namespace(__name__)
chaiml-kimid-v2-cosine-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-kimid-v2-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║   ██████   ██████  █████   ████  ████                               ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  ░░██████ ██████  ░░███   ███░  ░░███                               ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║   ░███░█████░███   ░███  ███     ░███                               ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║   ░███░░███ ░███   ░███████      ░███                               ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║   ░███ ░░░  ░███   ░███░░███     ░███                               ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║   ░███      ░███   ░███ ░░███    ░███                               ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║   █████     █████  █████ ░░████  █████                              ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  ░░░░░     ░░░░░  ░░░░░   ░░░░  ░░░░░                               ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  Version: 0.30.6+torch280                                           ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  Features: FLYWHEEL, CUDA                                           ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  Copyright 2023-2025 MK ONE TECHNOLOGIES Inc.                       ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  belonging to:                                                      ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║  Expiration: 2028-03-31 23:59:59                                    ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ║                                                                     ║
chaiml-kimid-v2-cosine-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
chaiml-kimid-v2-cosine-v1-mkmlizer: Downloaded to shared memory in 72.306s
chaiml-kimid-v2-cosine-v1-mkmlizer: Checking if ChaiML/kimid-v2_cosine already exists in ChaiML
chaiml-kimid-v2-cosine-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpmdyf6ezt, device:0
chaiml-kimid-v2-cosine-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-kimid-v2-cosine-v1-mkmlizer: 
Loading 0:   0%|          | 0.00/363 [00:00<?, ?it/s]
Loading 0:   7%|▋         | 27.0/363 [00:01<00:12, 26.9it/s]
Loading 0:   7%|▋         | 27.0/363 [00:01<00:12, 26.9it/s]
Loading 0:  10%|▉         | 36.0/363 [00:02<00:20, 16.1it/s]
Loading 0:  10%|▉         | 36.0/363 [00:02<00:20, 16.1it/s]
Loading 0:  19%|█▉        | 70.0/363 [00:03<00:12, 22.9it/s]
Loading 0:  19%|█▉        | 70.0/363 [00:03<00:12, 22.9it/s]
Loading 0:  27%|██▋       | 97.0/363 [00:04<00:10, 24.5it/s]
Loading 0:  27%|██▋       | 97.0/363 [00:04<00:10, 24.5it/s]
Loading 0:  35%|███▍      | 127/363 [00:05<00:09, 25.8it/s] 
Loading 0:  35%|███▍      | 127/363 [00:05<00:09, 25.8it/s]
Loading 0:  43%|████▎     | 156/363 [00:06<00:07, 26.5it/s]
Loading 0:  43%|████▎     | 156/363 [00:06<00:07, 26.5it/s]
Loading 0:  52%|█████▏    | 187/363 [00:07<00:06, 27.1it/s]
Loading 0:  52%|█████▏    | 187/363 [00:07<00:06, 27.1it/s]
Loading 0:  52%|█████▏    | 187/363 [00:20<00:06, 27.1it/s]
Loading 0:  55%|█████▌    | 201/363 [00:20<00:33, 4.90it/s]
Loading 0:  55%|█████▌    | 201/363 [00:20<00:33, 4.90it/s]
Loading 0:  62%|██████▏   | 224/363 [00:21<00:21, 6.39it/s]
Loading 0:  62%|██████▏   | 224/363 [00:21<00:21, 6.39it/s]
Loading 0:  71%|███████   | 257/363 [00:22<00:11, 9.14it/s]
Loading 0:  71%|███████   | 257/363 [00:22<00:11, 9.14it/s]
Loading 0:  78%|███████▊  | 284/363 [00:23<00:06, 11.4it/s]
Loading 0:  78%|███████▊  | 284/363 [00:23<00:06, 11.4it/s]
Loading 0:  86%|████████▌ | 311/363 [00:24<00:03, 13.8it/s]
Loading 0:  86%|████████▌ | 311/363 [00:24<00:03, 13.8it/s]
Loading 0:  94%|█████████▍| 341/363 [00:26<00:01, 16.0it/s]
Loading 0:  94%|█████████▍| 341/363 [00:26<00:01, 16.0it/s]
Loading 0: 100%|██████████| 363/363 [00:26<00:00, 18.4it/s]
Loading 0: 100%|██████████| 363/363 [00:26<00:00, 18.4it/s]
                                                           

Loading 0: 100%|██████████| 363/363 [00:26<00:00, 13.6it/s]
chaiml-kimid-v2-cosine-v1-mkmlizer: The tokenizer you are loading from '/tmp/tmpmdyf6ezt' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
chaiml-kimid-v2-cosine-v1-mkmlizer: quantized model in 43.537s
chaiml-kimid-v2-cosine-v1-mkmlizer: Processed model ChaiML/kimid-v2_cosine in 115.843s
chaiml-kimid-v2-cosine-v1-mkmlizer: creating bucket guanaco-mkml-models
chaiml-kimid-v2-cosine-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-kimid-v2-cosine-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-kimid-v2-cosine-v1/nvidia
chaiml-kimid-v2-cosine-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-kimid-v2-cosine-v1/nvidia/config.json
chaiml-kimid-v2-cosine-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-kimid-v2-cosine-v1/nvidia/special_tokens_map.json
chaiml-kimid-v2-cosine-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-kimid-v2-cosine-v1/nvidia/tokenizer_config.json
chaiml-kimid-v2-cosine-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-kimid-v2-cosine-v1/nvidia/tokenizer.json
chaiml-kimid-v2-cosine-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/chaiml-kimid-v2-cosine-v1/nvidia/flywheel_model.1.safetensors
chaiml-kimid-v2-cosine-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-kimid-v2-cosine-v1/nvidia/flywheel_model.0.safetensors
Job chaiml-kimid-v2-cosine-v1-mkmlizer completed after 175.62s with status: succeeded
Stopping job with name chaiml-kimid-v2-cosine-v1-mkmlizer
Pipeline stage MKMLizer completed in 176.11s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.18s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service chaiml-kimid-v2-cosine-v1
Waiting for inference service chaiml-kimid-v2-cosine-v1 to be ready
Inference service chaiml-kimid-v2-cosine-v1 ready after 180.6747043132782s
Pipeline stage MKMLDeployer completed in 181.20s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.6272895336151123s
Received healthy response to inference request in 3.0117075443267822s
Received healthy response to inference request in 2.4394195079803467s
Received healthy response to inference request in 2.751352310180664s
Received healthy response to inference request in 2.6988744735717773s
5 requests
0 failed requests
5th percentile: 2.4769935131073
10th percentile: 2.514567518234253
20th percentile: 2.589715528488159
30th percentile: 2.6416065216064455
40th percentile: 2.6702404975891114
50th percentile: 2.6988744735717773
60th percentile: 2.719865608215332
70th percentile: 2.740856742858887
80th percentile: 2.803423357009888
90th percentile: 2.907565450668335
95th percentile: 2.9596364974975584
99th percentile: 3.0012933349609376
mean time: 2.7057286739349364
Pipeline stage StressChecker completed in 14.91s
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 0.60s
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 0.70s
Shutdown handler de-registered
chaiml-kimid-v2-cosine_v1 status is now deployed due to DeploymentManager action
Shutdown handler registered
run pipeline %s
run pipeline stage %s
Running pipeline stage OfflineFamilyFriendlyScorer
Evaluating %s Family Friendly Score with %s threads
Generating Leaderboard row for %s
Generated Leaderboard row for %s
Pipeline stage OfflineFamilyFriendlyScorer completed in 2642.57s
Shutdown handler de-registered
chaiml-kimid-v2-cosine_v1 status is now inactive due to auto deactivation removed underperforming models