chaiml-ssnew-v5-dpo-lr_24415

developer_uid: richhx
submission_id: chaiml-ssnew-v5-dpo-lr_24415_v21
model_name: chaiml-ssnew-v5-dpo-lr_24415_v21
model_group: ChaiML/ssnew-v5-dpo-lr5e
status: torndown
timestamp: 2026-01-16T21:38:12+00:00
num_battles: 12124
num_wins: 6041
celo_rating: 1295.42
family_friendly_score: 0.4698
family_friendly_standard_error: 0.007058157833315999
submission_type: basic
model_repo: ChaiML/ssnew-v5-dpo-lr5e6b01-lora
model_architecture: MistralForCausalLM
model_num_parameters: 24096691200.0
best_of: 8
max_input_tokens: 2048
max_output_tokens: 64
reward_model: default
latencies: [{'batch_size': 1, 'throughput': 0.4413121272664722, 'latency_mean': 2.265903468132019, 'latency_p50': 2.0874091386795044, 'latency_p90': 2.3064359188079835}, {'batch_size': 3, 'throughput': 0.8740624456827615, 'latency_mean': 3.423171864748001, 'latency_p50': 3.3990191221237183, 'latency_p90': 3.941222906112671}, {'batch_size': 5, 'throughput': 1.0486301184896276, 'latency_mean': 4.746296961307525, 'latency_p50': 4.708405256271362, 'latency_p90': 5.283722758293152}, {'batch_size': 6, 'throughput': 1.1023156517267203, 'latency_mean': 5.3986592602729795, 'latency_p50': 5.386651396751404, 'latency_p90': 6.131614542007446}, {'batch_size': 10, 'throughput': 1.2107413563770961, 'latency_mean': 8.198953688144684, 'latency_p50': 8.20735776424408, 'latency_p90': 9.316180944442749}]
gpu_counts: {'NVIDIA A100-SXM4-80GB': 1}
display_name: chaiml-ssnew-v5-dpo-lr_24415_v21
is_internal_developer: True
language_model: ChaiML/ssnew-v5-dpo-lr5e6b01-lora
model_size: 24B
ranking_group: single
throughput_3p7s: 0.92
us_pacific_date: 2026-01-13
win_ratio: 0.4982678983833718
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', '</s>', '####'], 'max_input_tokens': 2048, 'best_of': 8, 'max_output_tokens': 64}
formatter: {'memory_template': "{bot_name}'s persona: {memory}", 'prompt_template': '', 'bot_template': '{bot_name}: {message}\n', 'user_template': 'You: {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-ssnew-v5-dpo-lr-24415-v21-mkmlizer
Waiting for job on chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer to finish
HTTP Request: %s %s "%s %d %s"
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: bash: cannot set terminal process group (-1): Inappropriate ioctl for device
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: bash: no job control in this shell
chaiml-ssnew-v5-dpo-lr-24415-v21-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-ssnew-v5-dpo-lr-24415-v21-mkmlizer:   __import__('pkg_resources').declare_namespace(__name__)
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║                                                                     ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║   ██████   ██████  █████   ████  ████                               ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  ░░██████ ██████  ░░███   ███░  ░░███                               ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║   ░███░█████░███   ░███  ███     ░███                               ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║   ░███░░███ ░███   ░███████      ░███                               ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║   ░███ ░░░  ░███   ░███░░███     ░███                               ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║   ░███      ░███   ░███ ░░███    ░███                               ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║   █████     █████  █████ ░░████  █████                              ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  ░░░░░     ░░░░░  ░░░░░   ░░░░  ░░░░░                               ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║                                                                     ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  Version: 0.30.6+torch280                                           ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  Features: FLYWHEEL, CUDA                                           ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  Copyright 2023-2025 MK ONE TECHNOLOGIES Inc.                       ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║                                                                     ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  belonging to:                                                      ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║                                                                     ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║  Expiration: 2028-03-31 23:59:59                                    ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ║                                                                     ║
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: Downloaded to shared memory in 55.443s
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: Checking if ChaiML/ssnew-v5-dpo-lr5e6b01-lora already exists in ChaiML
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: quantizing model to /dev/shm/model_cache, profile:q4, folder:/tmp/tmp8gv0psy5, device:0
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: Saving flywheel model at /dev/shm/model_cache
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
HTTP Request: %s %s "%s %d %s"
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: The tokenizer you are loading from '/tmp/tmp8gv0psy5' 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-ssnew-v5-dpo-lr-24415-v21-mkmlizer: quantized model in 277.113s
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: Processed model ChaiML/ssnew-v5-dpo-lr5e6b01-lora in 332.556s
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: creating bucket guanaco-mkml-models
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/config.json
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/special_tokens_map.json
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/tokenizer_config.json
HTTP Request: %s %s "%s %d %s"
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/tokenizer.json
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/flywheel_model.1.safetensors
HTTP Request: %s %s "%s %d %s"
chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/flywheel_model.0.safetensors
Job chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer completed after 452.02s with status: succeeded
Stopping job with name chaiml-ssnew-v5-dpo-lr-24415-v21-mkmlizer
Pipeline stage MKMLizer completed in 452.55s
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-ssnew-v5-dpo-lr-24415-v21
Waiting for inference service chaiml-ssnew-v5-dpo-lr-24415-v21 to be ready
Inference service chaiml-ssnew-v5-dpo-lr-24415-v21 ready after 181.2723217010498s
Pipeline stage MKMLDeployer completed in 181.86s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.005814552307129s
Received healthy response to inference request in 2.1866252422332764s
Received healthy response to inference request in 1.8256890773773193s
Received healthy response to inference request in 2.3001956939697266s
Received healthy response to inference request in 2.6990652084350586s
5 requests
0 failed requests
5th percentile: 1.8617141723632813
10th percentile: 1.8977392673492433
20th percentile: 1.969789457321167
30th percentile: 2.0419766902923584
40th percentile: 2.1143009662628174
50th percentile: 2.1866252422332764
60th percentile: 2.2320534229278564
70th percentile: 2.2774816036224363
80th percentile: 2.379969596862793
90th percentile: 2.539517402648926
95th percentile: 2.619291305541992
99th percentile: 2.6831104278564455
mean time: 2.203477954864502
Pipeline stage StressChecker completed in 12.65s
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 0.61s
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 0.72s
Shutdown handler de-registered
chaiml-ssnew-v5-dpo-lr_24415_v21 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 3009.94s
Shutdown handler de-registered
chaiml-ssnew-v5-dpo-lr_24415_v21 status is now inactive due to auto deactivation removed underperforming models
admin requested tearing down of chaiml-ssnew-v5-dpo-lr_24415_v21
Shutdown handler not registered because Python interpreter is not running in the main thread
admin requested tearing down of chaiml-ssnew-v5-dpo-lr_24415_v21
Falling back to EndpointApi.from_submission implementation
Shutdown handler not registered because Python interpreter is not running in the main thread
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
run pipeline %s
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
run pipeline stage %s
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Running pipeline stage MKMLDeleter
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Checking if service chaiml-ssnew-v5-dpo-lr-24415-v21 is running
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Skipping teardown as no inference service was found
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Pipeline stage MKMLDeleter completed in 2.60s
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
run pipeline stage %s
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Running pipeline stage MKMLModelDeleter
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Cleaning model data from S3
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Cleaning model data from model cache
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Deleting key chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/flywheel_model.1.safetensors from bucket guanaco-mkml-models
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Deleting key chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/special_tokens_map.json from bucket guanaco-mkml-models
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Deleting key chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/tokenizer.json from bucket guanaco-mkml-models
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Deleting key chaiml-ssnew-v5-dpo-lr-24415-v21/nvidia/tokenizer_config.json from bucket guanaco-mkml-models
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Pipeline stage MKMLModelDeleter completed in 11.07s
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Shutdown handler de-registered
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation
Falling back to EndpointApi.from_submission implementation