developer_uid: chai_backend_admin
submission_id: chaiml-ir-v0-sft-lr2e6fullpyg_v1
model_name: chaiml-ir-v0-sft-lr2e6fullpyg_v1
model_group: ChaiML/ir-v0-sft-lr2e6fu
status: inactive
timestamp: 2025-10-21T08:14:06+00:00
num_battles: 5908
num_wins: 3132
celo_rating: 1296.9
family_friendly_score: 0.4686
family_friendly_standard_error: 0.0070571104568371325
submission_type: basic
model_repo: ChaiML/ir-v0-sft-lr2e6fullpyg
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': 0.3764315718491722, 'latency_mean': 2.6564005982875822, 'latency_p50': 2.6650689840316772, 'latency_p90': 2.9298489332199096}, {'batch_size': 2, 'throughput': 0.567706616067885, 'latency_mean': 3.5211629247665406, 'latency_p50': 3.5062469244003296, 'latency_p90': 3.8530333995819093}, {'batch_size': 3, 'throughput': 0.6949757113320567, 'latency_mean': 4.307108287811279, 'latency_p50': 4.304255366325378, 'latency_p90': 4.649032521247864}, {'batch_size': 4, 'throughput': 0.7893793822707118, 'latency_mean': 5.049555504322052, 'latency_p50': 5.0453397035598755, 'latency_p90': 5.767408299446106}, {'batch_size': 5, 'throughput': 0.8486944254101204, 'latency_mean': 5.866757655143738, 'latency_p50': 5.8805437088012695, 'latency_p90': 6.724539494514465}]
gpu_counts: {'NVIDIA L40S': 1}
display_name: chaiml-ir-v0-sft-lr2e6fullpyg_v1
is_internal_developer: True
language_model: ChaiML/ir-v0-sft-lr2e6fullpyg
model_size: 24B
ranking_group: single
throughput_3p7s: 0.6
us_pacific_date: 2025-10-21
win_ratio: 0.5301286391333785
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>', 'User:', 'You:'], 'max_input_tokens': 2048, 'best_of': 8, 'max_output_tokens': 64}
formatter: {'memory_template': "{bot_name}'s persona: {memory}", 'prompt_template': '{prompt}\n<START>\n', '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-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer
Waiting for job on chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer to finish
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║                                                                     ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║   ██████   ██████  █████   ████  ████                               ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  ░░██████ ██████  ░░███   ███░  ░░███                               ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║   ░███░█████░███   ░███  ███     ░███                               ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║   ░███░░███ ░███   ░███████      ░███                               ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║   ░███ ░░░  ░███   ░███░░███     ░███                               ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║   ░███      ░███   ░███ ░░███    ░███                               ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║   █████     █████  █████ ░░████  █████                              ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  ░░░░░     ░░░░░  ░░░░░   ░░░░  ░░░░░                               ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║                                                                     ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  Version: 0.30.2                                                    ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  Features: FLYWHEEL, CUDA                                           ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  Copyright 2023-2025 MK ONE TECHNOLOGIES Inc.                       ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║                                                                     ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  belonging to:                                                      ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║                                                                     ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║  Expiration: 2028-03-31 23:59:59                                    ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ║                                                                     ║
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: Downloaded to shared memory in 64.439s
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: Checking if ChaiML/ir-v0-sft-lr2e6fullpyg already exists in ChaiML
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpgmw3bwla, device:0
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: quantized model in 42.253s
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: Processed model ChaiML/ir-v0-sft-lr2e6fullpyg in 106.692s
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: creating bucket guanaco-mkml-models
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-ir-v0-sft-lr2e6fullpyg-v1/nvidia
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-ir-v0-sft-lr2e6fullpyg-v1/nvidia/config.json
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-ir-v0-sft-lr2e6fullpyg-v1/nvidia/special_tokens_map.json
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-ir-v0-sft-lr2e6fullpyg-v1/nvidia/tokenizer_config.json
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-ir-v0-sft-lr2e6fullpyg-v1/nvidia/tokenizer.json
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/chaiml-ir-v0-sft-lr2e6fullpyg-v1/nvidia/flywheel_model.1.safetensors
chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-ir-v0-sft-lr2e6fullpyg-v1/nvidia/flywheel_model.0.safetensors
Job chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer completed after 185.38s with status: succeeded
Stopping job with name chaiml-ir-v0-sft-lr2e6fullpyg-v1-mkmlizer
Pipeline stage MKMLizer completed in 185.85s
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 chaiml-ir-v0-sft-lr2e6fullpyg-v1
Waiting for inference service chaiml-ir-v0-sft-lr2e6fullpyg-v1 to be ready
Inference service chaiml-ir-v0-sft-lr2e6fullpyg-v1 ready after 70.27814722061157s
Pipeline stage MKMLDeployer completed in 70.85s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.637006998062134s
Received healthy response to inference request in 3.5025558471679688s
Received healthy response to inference request in 2.1221635341644287s
Received healthy response to inference request in 2.305007219314575s
Received healthy response to inference request in 2.136289358139038s
5 requests
0 failed requests
5th percentile: 2.1249886989593505
10th percentile: 2.1278138637542723
20th percentile: 2.1334641933441163
30th percentile: 2.1700329303741457
40th percentile: 2.2375200748443604
50th percentile: 2.305007219314575
60th percentile: 2.4378071308135985
70th percentile: 2.570607042312622
80th percentile: 2.810116767883301
90th percentile: 3.1563363075256348
95th percentile: 3.3294460773468018
99th percentile: 3.4679338932037354
mean time: 2.540604591369629
Pipeline stage StressChecker completed in 14.11s
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 0.94s
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 0.61s
Shutdown handler de-registered
chaiml-ir-v0-sft-lr2e6fullpyg_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 0.09s
run pipeline stage %s
Running pipeline stage MKMLProfilerTemplater
Pipeline stage MKMLProfilerTemplater completed in 0.08s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service chaiml-ir-v0-sft-lr2e6fullpyg-v1-profiler
Waiting for inference service chaiml-ir-v0-sft-lr2e6fullpyg-v1-profiler to be ready
Inference service chaiml-ir-v0-sft-lr2e6fullpyg-v1-profiler ready after 70.58883213996887s
Pipeline stage MKMLProfilerDeployer completed in 71.18s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl --kubeconfig /code/guanaco/guanaco_services/resources/kchai_coreweave_us_east_04a.yaml cp /code/guanaco/guanaco_inference_services/src/inference_scripts creator-studio/chaiml-ir-v0-sft-lr246982575226f06188cc92515097f010e-deplonxm9b:/code/chaiverse_profiler_1761034870 --namespace creator-studio
kubectl --kubeconfig /code/guanaco/guanaco_services/resources/kchai_coreweave_us_east_04a.yaml exec -it chaiml-ir-v0-sft-lr246982575226f06188cc92515097f010e-deplonxm9b --namespace creator-studio -- sh -c 'cd /code/chaiverse_profiler_1761034870 && python profiles.py profile --best_of_n 8 --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 2048 --output_tokens 64 --summary /code/chaiverse_profiler_1761034870/summary.json'
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 3115.97s
Shutdown handler de-registered
chaiml-ir-v0-sft-lr2e6fullpyg_v1 status is now inactive due to auto deactivation removed underperforming models