chaiml-arli-24b-sft-exp_31451

developer_uid: NischayDnk

submission_id: chaiml-arli-24b-sft-exp_31451_v1

model_name: chaiml-arli-24b-sft-exp_31451_v1

model_group: ChaiML/arli-24b-sft-exp2

status: torndown

timestamp: 2025-04-25T04:32:01+00:00

num_battles: 10130

num_wins: 5157

celo_rating: 1295.45

family_friendly_score: 0.532

family_friendly_standard_error: 0.007056571405434795

submission_type: basic

model_repo: ChaiML/arli-24b-sft-exp2-grokdata-top3klexical

model_architecture: MistralForCausalLM

model_num_parameters: 24096691200.0

best_of: 8

max_input_tokens: 1024

max_output_tokens: 64

reward_model: default

latencies: [{'batch_size': 1, 'throughput': 0.5350635822791339, 'latency_mean': 1.8688715267181397, 'latency_p50': 1.8642657995224, 'latency_p90': 2.083867073059082}, {'batch_size': 3, 'throughput': 1.089299985106058, 'latency_mean': 2.7387395548820495, 'latency_p50': 2.743211507797241, 'latency_p90': 3.0488385915756226}, {'batch_size': 5, 'throughput': 1.393800270123771, 'latency_mean': 3.5654471158981322, 'latency_p50': 3.594829797744751, 'latency_p90': 3.9484461307525636}, {'batch_size': 6, 'throughput': 1.5140519402706194, 'latency_mean': 3.9417121028900146, 'latency_p50': 3.949455142021179, 'latency_p90': 4.41355574131012}, {'batch_size': 8, 'throughput': 1.655125489176167, 'latency_mean': 4.792687036991119, 'latency_p50': 4.806988835334778, 'latency_p90': 5.398262786865234}, {'batch_size': 10, 'throughput': 1.7701535186469128, 'latency_mean': 5.595003772974014, 'latency_p50': 5.573487877845764, 'latency_p90': 6.353952932357788}]

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

display_name: chaiml-arli-24b-sft-exp_31451_v1

is_internal_developer: False

language_model: ChaiML/arli-24b-sft-exp2-grokdata-top3klexical

model_size: 24B

ranking_group: single

throughput_3p7s: 1.45

us_pacific_date: 2025-04-24

win_ratio: 0.5090819348469892

generation_params: {'temperature': 0.45, 'top_p': 0.95, 'min_p': 0.025, 'top_k': 60, 'presence_penalty': 0.4, 'frequency_penalty': 0.4, 'stopping_words': ['<|im_end|>', '\n', '<|im_start|>'], 'max_input_tokens': 1024, 'best_of': 8, '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-arli-24b-sft-exp-31451-v1-mkmlizer
Waiting for job on chaiml-arli-24b-sft-exp-31451-v1-mkmlizer to finish
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║     _____            __           __                                ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║       /___/                                                         ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║                                                                     ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  Version: 0.12.8                                                    ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║                                                                     ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  belonging to:                                                      ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║                                                                     ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║  Expiration: 2028-03-31 23:59:59                                    ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ║                                                                     ║
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
Failed to get response for submission chaiml-gemma3-27b-base-s_9210_v3: ('http://chaiml-gemma3-27b-base-s-9210-v3-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', '{"error":"Exception : CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 79.15 GiB of which 311.62 MiB is free. Process 1237082 has 78.84 GiB memory in use. Of the allocated memory 77.62 GiB is allocated by PyTorch, and 702.60 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation.  See documentation for Memory Management  (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"}')
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: Downloaded to shared memory in 82.397s
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpyvjt3orr, device:0
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: quantized model in 54.651s
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: Processed model ChaiML/arli-24b-sft-exp2-grokdata-top3klexical in 137.048s
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: creating bucket guanaco-mkml-models
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-arli-24b-sft-exp-31451-v1
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-arli-24b-sft-exp-31451-v1/config.json
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-arli-24b-sft-exp-31451-v1/special_tokens_map.json
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-arli-24b-sft-exp-31451-v1/tokenizer_config.json
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-arli-24b-sft-exp-31451-v1/tokenizer.json
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/chaiml-arli-24b-sft-exp-31451-v1/flywheel_model.1.safetensors
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-arli-24b-sft-exp-31451-v1/flywheel_model.0.safetensors
chaiml-arli-24b-sft-exp-31451-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|          | 4/363 [00:00<00:09, 37.75it/s]
Loading 0:   2%|▏         | 8/363 [00:00<00:12, 29.51it/s]
Loading 0:   3%|▎         | 12/363 [00:00<00:10, 31.94it/s]
Loading 0:   4%|▍         | 16/363 [00:00<00:12, 28.59it/s]
Loading 0:   6%|▌         | 21/363 [00:00<00:10, 33.03it/s]
Loading 0:   7%|▋         | 25/363 [00:00<00:11, 29.03it/s]
Loading 0:   9%|▉         | 32/363 [00:00<00:09, 34.68it/s]
Loading 0:  10%|▉         | 36/363 [00:01<00:15, 20.51it/s]
Loading 0:  11%|█         | 40/363 [00:01<00:13, 23.14it/s]
Loading 0:  12%|█▏        | 43/363 [00:01<00:13, 23.17it/s]
Loading 0:  13%|█▎        | 48/363 [00:01<00:12, 26.10it/s]
Loading 0:  14%|█▍        | 51/363 [00:01<00:13, 23.21it/s]
Loading 0:  16%|█▌        | 57/363 [00:02<00:10, 28.27it/s]
Loading 0:  17%|█▋        | 61/363 [00:02<00:11, 26.85it/s]
Loading 0:  18%|█▊        | 65/363 [00:02<00:10, 27.62it/s]
Loading 0:  19%|█▉        | 70/363 [00:02<00:11, 24.77it/s]
Loading 0:  20%|██        | 73/363 [00:02<00:13, 20.85it/s]
Loading 0:  22%|██▏       | 79/363 [00:03<00:10, 25.84it/s]
Loading 0:  23%|██▎       | 82/363 [00:03<00:11, 24.92it/s]
Loading 0:  24%|██▎       | 86/363 [00:03<00:10, 25.58it/s]
Loading 0:  25%|██▍       | 89/363 [00:03<00:10, 25.52it/s]
Loading 0:  25%|██▌       | 92/363 [00:03<00:13, 20.58it/s]
Loading 0:  27%|██▋       | 99/363 [00:03<00:09, 27.52it/s]
Loading 0:  28%|██▊       | 102/363 [00:03<00:10, 24.72it/s]
Loading 0:  29%|██▉       | 107/363 [00:04<00:11, 22.41it/s]
Loading 0:  31%|███       | 111/363 [00:04<00:09, 25.36it/s]
Loading 0:  31%|███▏      | 114/363 [00:04<00:10, 22.87it/s]
Loading 0:  33%|███▎      | 120/363 [00:04<00:08, 27.52it/s]
Loading 0:  34%|███▍      | 123/363 [00:04<00:09, 24.40it/s]
Loading 0:  36%|███▌      | 129/363 [00:04<00:08, 28.65it/s]
Loading 0:  36%|███▋      | 132/363 [00:05<00:09, 25.41it/s]
Loading 0:  38%|███▊      | 138/363 [00:05<00:07, 29.73it/s]
Loading 0:  39%|███▉      | 142/363 [00:05<00:08, 27.60it/s]
Loading 0:  41%|████      | 148/363 [00:05<00:06, 34.19it/s]
Loading 0:  42%|████▏     | 152/363 [00:05<00:08, 23.47it/s]
Loading 0:  43%|████▎     | 155/363 [00:06<00:08, 23.58it/s]
Loading 0:  44%|████▎     | 158/363 [00:06<00:10, 19.74it/s]
Loading 0:  45%|████▌     | 165/363 [00:06<00:07, 26.14it/s]
Loading 0:  46%|████▋     | 168/363 [00:06<00:08, 23.84it/s]
Loading 0:  48%|████▊     | 174/363 [00:06<00:06, 28.22it/s]
Loading 0:  49%|████▉     | 178/363 [00:06<00:06, 26.65it/s]
Loading 0:  50%|█████     | 182/363 [00:07<00:06, 27.12it/s]
Loading 0:  52%|█████▏    | 187/363 [00:07<00:07, 24.18it/s]
Loading 0:  52%|█████▏    | 190/363 [00:07<00:07, 22.36it/s]
Loading 0:  53%|█████▎    | 193/363 [00:07<00:07, 23.10it/s]
Loading 0:  54%|█████▍    | 196/363 [00:07<00:07, 23.53it/s]
Loading 0:  55%|█████▌    | 200/363 [00:22<00:06, 23.53it/s]
Loading 0:  55%|█████▌    | 201/363 [00:22<03:04,  1.14s/it]
Loading 0:  56%|█████▌    | 203/363 [00:22<02:32,  1.05it/s]
Loading 0:  57%|█████▋    | 208/363 [00:22<01:31,  1.68it/s]
Loading 0:  58%|█████▊    | 211/363 [00:22<01:10,  2.15it/s]
Loading 0:  59%|█████▉    | 214/363 [00:23<00:52,  2.82it/s]
Loading 0:  60%|██████    | 218/363 [00:23<00:35,  4.04it/s]
Loading 0:  61%|██████    | 221/363 [00:23<00:27,  5.15it/s]
Loading 0:  62%|██████▏   | 224/363 [00:23<00:22,  6.18it/s]
Loading 0:  63%|██████▎   | 228/363 [00:23<00:15,  8.67it/s]
Loading 0:  64%|██████▎   | 231/363 [00:23<00:13, 10.00it/s]
Loading 0:  65%|██████▌   | 237/363 [00:24<00:08, 14.93it/s]
Loading 0:  66%|██████▌   | 240/363 [00:24<00:07, 15.52it/s]
Loading 0:  68%|██████▊   | 246/363 [00:24<00:05, 20.77it/s]
Loading 0:  69%|██████▉   | 250/363 [00:24<00:05, 21.35it/s]
Loading 0:  70%|███████   | 255/363 [00:24<00:04, 24.39it/s]
Loading 0:  71%|███████   | 258/363 [00:24<00:04, 22.55it/s]
Loading 0:  73%|███████▎  | 264/363 [00:24<00:03, 29.36it/s]
Loading 0:  74%|███████▍  | 268/363 [00:25<00:04, 23.26it/s]
Loading 0:  75%|███████▍  | 271/363 [00:25<00:04, 21.23it/s]
Loading 0:  75%|███████▌  | 274/363 [00:25<00:03, 22.26it/s]
Loading 0:  76%|███████▋  | 277/363 [00:25<00:03, 23.08it/s]
Loading 0:  78%|███████▊  | 282/363 [00:25<00:03, 26.53it/s]
Loading 0:  79%|███████▊  | 285/363 [00:25<00:03, 24.11it/s]
Loading 0:  80%|████████  | 291/363 [00:26<00:02, 29.32it/s]
Loading 0:  81%|████████▏ | 295/363 [00:26<00:02, 27.04it/s]
Loading 0:  82%|████████▏ | 299/363 [00:26<00:02, 27.57it/s]
Loading 0:  84%|████████▎ | 304/363 [00:26<00:02, 25.25it/s]
Loading 0:  85%|████████▍ | 307/363 [00:26<00:02, 23.13it/s]
Loading 0:  85%|████████▌ | 310/363 [00:26<00:02, 23.98it/s]
Loading 0:  86%|████████▌ | 313/363 [00:27<00:02, 24.51it/s]
Loading 0:  88%|████████▊ | 318/363 [00:27<00:01, 28.00it/s]
Loading 0:  88%|████████▊ | 321/363 [00:27<00:01, 25.24it/s]
Loading 0:  90%|█████████ | 327/363 [00:27<00:01, 30.39it/s]
Loading 0:  91%|█████████ | 331/363 [00:27<00:01, 28.63it/s]
Loading 0:  92%|█████████▏| 335/363 [00:27<00:00, 28.98it/s]
Loading 0:  93%|█████████▎| 338/363 [00:27<00:00, 27.63it/s]
Loading 0:  94%|█████████▍| 341/363 [00:34<00:13,  1.67it/s]
Loading 0:  95%|█████████▌| 346/363 [00:34<00:06,  2.62it/s]
Loading 0:  96%|█████████▌| 349/363 [00:34<00:04,  3.29it/s]
Loading 0:  98%|█████████▊| 355/363 [00:35<00:01,  5.32it/s]
Loading 0:  99%|█████████▉| 359/363 [00:35<00:00,  6.77it/s]
                                                            
Job chaiml-arli-24b-sft-exp-31451-v1-mkmlizer completed after 165.95s with status: succeeded
Stopping job with name chaiml-arli-24b-sft-exp-31451-v1-mkmlizer
Pipeline stage MKMLizer completed in 166.46s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.16s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service chaiml-arli-24b-sft-exp-31451-v1
Waiting for inference service chaiml-arli-24b-sft-exp-31451-v1 to be ready
Inference service chaiml-arli-24b-sft-exp-31451-v1 ready after 120.61210608482361s
Pipeline stage MKMLDeployer completed in 121.14s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.6765992641448975s
Received healthy response to inference request in 2.583615779876709s
Received healthy response to inference request in 2.2604739665985107s
Received healthy response to inference request in 2.510117292404175s
Received healthy response to inference request in 2.1564927101135254s
5 requests
0 failed requests
5th percentile: 2.1772889614105226
10th percentile: 2.1980852127075194
20th percentile: 2.2396777153015135
30th percentile: 2.3104026317596436
40th percentile: 2.410259962081909
50th percentile: 2.510117292404175
60th percentile: 2.5395166873931885
70th percentile: 2.568916082382202
80th percentile: 2.6022124767303465
90th percentile: 2.639405870437622
95th percentile: 2.65800256729126
99th percentile: 2.67287992477417
mean time: 2.4374598026275636
Pipeline stage StressChecker completed in 13.60s
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 0.72s
Shutdown handler de-registered
chaiml-arli-24b-sft-exp_31451_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
%s, retrying in %s seconds...
Evaluating %s Family Friendly Score with %s threads
Pipeline stage OfflineFamilyFriendlyScorer completed in 5021.38s
Shutdown handler de-registered
chaiml-arli-24b-sft-exp_31451_v1 status is now inactive due to auto deactivation removed underperforming models
chaiml-arli-24b-sft-exp_31451_v1 status is now torndown due to DeploymentManager action