chaiml-claude-dpo-try-exp1

developer_uid: NischayDnk

submission_id: chaiml-claude-dpo-try-exp1_v1

model_name: chaiml-claude-dpo-try-exp1_v1

model_group: ChaiML/claude-dpo-try-ex

status: torndown

timestamp: 2025-03-04T11:06:01+00:00

num_battles: 9684

num_wins: 4750

celo_rating: 1271.86

family_friendly_score: 0.5298

family_friendly_standard_error: 0.007058497857193129

submission_type: basic

model_repo: ChaiML/claude-dpo-try-exp1

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': 0.3708972897796061, 'latency_mean': 2.6960909962654114, 'latency_p50': 2.6770644187927246, 'latency_p90': 3.002157473564148}, {'batch_size': 2, 'throughput': 0.5735352611823801, 'latency_mean': 3.4825739884376525, 'latency_p50': 3.482003092765808, 'latency_p90': 3.831533026695251}, {'batch_size': 3, 'throughput': 0.7163066851296334, 'latency_mean': 4.165613889694214, 'latency_p50': 4.151449918746948, 'latency_p90': 4.562669062614441}, {'batch_size': 4, 'throughput': 0.8353090342446484, 'latency_mean': 4.755978533029556, 'latency_p50': 4.789972901344299, 'latency_p90': 5.239307928085327}, {'batch_size': 5, 'throughput': 0.9303422606423213, 'latency_mean': 5.340674079656601, 'latency_p50': 5.323194265365601, 'latency_p90': 5.985584163665772}]

gpu_counts: {'NVIDIA RTX A6000': 1}

display_name: chaiml-claude-dpo-try-exp1_v1

is_internal_developer: False

language_model: ChaiML/claude-dpo-try-exp1

model_size: 24B

ranking_group: single

throughput_3p7s: 0.63

us_pacific_date: 2025-03-04

win_ratio: 0.49049979347377115

generation_params: {'temperature': 0.45, 'top_p': 0.95, 'min_p': 0.05, 'top_k': 80, 'presence_penalty': 0.45, 'frequency_penalty': 0.45, 'stopping_words': ['###', '</s>', '\n', 'You:'], 'max_input_tokens': 1024, 'best_of': 8, 'max_output_tokens': 64}

formatter: {'memory_template': '[INST]', 'prompt_template': "''", 'bot_template': '<|im_start|>assistant\n{bot_name}: {message}<|im_end|>\n', 'user_template': '<|im_start|>user\n{user_name}: {message}<|im_end|>\n', 'response_template': '[/INST]{bot_name}:', '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-claude-dpo-try-exp1-v1-mkmlizer
Waiting for job on chaiml-claude-dpo-try-exp1-v1-mkmlizer to finish
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║     _____            __           __                                ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║       /___/                                                         ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║                                                                     ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  Version: 0.12.8                                                    ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║                                                                     ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  belonging to:                                                      ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║                                                                     ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║  Expiration: 2025-04-15 23:59:59                                    ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ║                                                                     ║
chaiml-claude-dpo-try-exp1-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
chaiml-claude-dpo-try-exp1-v1-mkmlizer: Downloaded to shared memory in 77.957s
chaiml-claude-dpo-try-exp1-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmp0lpbzl9x, device:0
chaiml-claude-dpo-try-exp1-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-claude-dpo-try-exp1-v1-mkmlizer: quantized model in 53.288s
chaiml-claude-dpo-try-exp1-v1-mkmlizer: Processed model ChaiML/claude-dpo-try-exp1 in 131.246s
chaiml-claude-dpo-try-exp1-v1-mkmlizer: creating bucket guanaco-mkml-models
chaiml-claude-dpo-try-exp1-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-claude-dpo-try-exp1-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-claude-dpo-try-exp1-v1
chaiml-claude-dpo-try-exp1-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-claude-dpo-try-exp1-v1/config.json
chaiml-claude-dpo-try-exp1-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-claude-dpo-try-exp1-v1/special_tokens_map.json
chaiml-claude-dpo-try-exp1-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-claude-dpo-try-exp1-v1/tokenizer_config.json
chaiml-claude-dpo-try-exp1-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-claude-dpo-try-exp1-v1/tokenizer.json
chaiml-claude-dpo-try-exp1-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/chaiml-claude-dpo-try-exp1-v1/flywheel_model.1.safetensors
chaiml-claude-dpo-try-exp1-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-claude-dpo-try-exp1-v1/flywheel_model.0.safetensors
chaiml-claude-dpo-try-exp1-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|▏         | 5/363 [00:00<00:16, 21.15it/s]
Loading 0:   3%|▎         | 12/363 [00:00<00:09, 38.22it/s]
Loading 0:   5%|▍         | 17/363 [00:00<00:09, 36.09it/s]
Loading 0:   6%|▌         | 22/363 [00:00<00:09, 35.11it/s]
Loading 0:   7%|▋         | 26/363 [00:00<00:09, 35.96it/s]
Loading 0:   9%|▉         | 32/363 [00:00<00:08, 40.49it/s]
Loading 0:  10%|█         | 37/363 [00:01<00:13, 25.05it/s]
Loading 0:  11%|█▏        | 41/363 [00:01<00:13, 23.25it/s]
Loading 0:  13%|█▎        | 48/363 [00:01<00:10, 29.93it/s]
Loading 0:  14%|█▍        | 52/363 [00:01<00:10, 28.69it/s]
Loading 0:  16%|█▌        | 57/363 [00:01<00:09, 31.47it/s]
Loading 0:  17%|█▋        | 61/363 [00:02<00:10, 29.88it/s]
Loading 0:  18%|█▊        | 65/363 [00:02<00:09, 30.53it/s]
Loading 0:  19%|█▉        | 70/363 [00:02<00:10, 27.05it/s]
Loading 0:  20%|██        | 73/363 [00:02<00:12, 22.53it/s]
Loading 0:  22%|██▏       | 79/363 [00:02<00:10, 28.14it/s]
Loading 0:  23%|██▎       | 83/363 [00:02<00:09, 29.08it/s]
Loading 0:  24%|██▍       | 87/363 [00:03<00:11, 24.34it/s]
Loading 0:  25%|██▌       | 91/363 [00:03<00:10, 26.92it/s]
Loading 0:  26%|██▌       | 95/363 [00:03<00:09, 26.88it/s]
Loading 0:  27%|██▋       | 99/363 [00:03<00:09, 27.16it/s]
Loading 0:  28%|██▊       | 102/363 [00:03<00:10, 23.90it/s]
Loading 0:  29%|██▉       | 107/363 [00:03<00:11, 22.24it/s]
Loading 0:  31%|███       | 111/363 [00:03<00:09, 25.28it/s]
Loading 0:  31%|███▏      | 114/363 [00:04<00:10, 23.22it/s]
Loading 0:  33%|███▎      | 120/363 [00:04<00:08, 28.54it/s]
Loading 0:  34%|███▍      | 124/363 [00:04<00:08, 27.16it/s]
Loading 0:  36%|███▌      | 129/363 [00:04<00:07, 30.04it/s]
Loading 0:  37%|███▋      | 133/363 [00:04<00:08, 28.68it/s]
Loading 0:  38%|███▊      | 138/363 [00:04<00:07, 31.06it/s]
Loading 0:  39%|███▉      | 142/363 [00:05<00:07, 28.97it/s]
Loading 0:  41%|████      | 149/363 [00:05<00:06, 32.82it/s]
Loading 0:  42%|████▏     | 153/363 [00:05<00:10, 20.90it/s]
Loading 0:  43%|████▎     | 157/363 [00:05<00:08, 23.51it/s]
Loading 0:  44%|████▍     | 160/363 [00:05<00:08, 23.91it/s]
Loading 0:  45%|████▌     | 165/363 [00:05<00:07, 26.87it/s]
Loading 0:  47%|████▋     | 169/363 [00:06<00:07, 26.46it/s]
Loading 0:  48%|████▊     | 174/363 [00:06<00:06, 29.47it/s]
Loading 0:  49%|████▉     | 178/363 [00:06<00:06, 27.51it/s]
Loading 0:  50%|█████     | 182/363 [00:06<00:06, 27.68it/s]
Loading 0:  52%|█████▏    | 187/363 [00:06<00:07, 24.64it/s]
Loading 0:  52%|█████▏    | 190/363 [00:07<00:07, 22.75it/s]
Loading 0:  53%|█████▎    | 193/363 [00:07<00:07, 23.70it/s]
Loading 0:  54%|█████▍    | 196/363 [00:07<00:06, 23.88it/s]
Loading 0:  55%|█████▌    | 200/363 [00:21<00:06, 23.88it/s]
Loading 0:  55%|█████▌    | 201/363 [00:21<03:00,  1.11s/it]
Loading 0:  56%|█████▌    | 203/363 [00:21<02:29,  1.07it/s]
Loading 0:  57%|█████▋    | 208/363 [00:21<01:29,  1.73it/s]
Loading 0:  58%|█████▊    | 211/363 [00:22<01:08,  2.21it/s]
Loading 0:  59%|█████▉    | 214/363 [00:22<00:51,  2.91it/s]
Loading 0:  60%|██████    | 218/363 [00:22<00:34,  4.18it/s]
Loading 0:  61%|██████    | 221/363 [00:22<00:26,  5.34it/s]
Loading 0:  62%|██████▏   | 224/363 [00:22<00:21,  6.49it/s]
Loading 0:  63%|██████▎   | 228/363 [00:22<00:14,  9.04it/s]
Loading 0:  64%|██████▎   | 231/363 [00:22<00:12, 10.50it/s]
Loading 0:  65%|██████▌   | 237/363 [00:23<00:08, 15.59it/s]
Loading 0:  66%|██████▌   | 240/363 [00:23<00:07, 16.14it/s]
Loading 0:  68%|██████▊   | 246/363 [00:23<00:05, 21.69it/s]
Loading 0:  69%|██████▉   | 250/363 [00:23<00:05, 22.24it/s]
Loading 0:  70%|███████   | 255/363 [00:23<00:04, 25.12it/s]
Loading 0:  71%|███████▏  | 259/363 [00:23<00:04, 24.97it/s]
Loading 0:  73%|███████▎  | 266/363 [00:23<00:03, 31.03it/s]
Loading 0:  74%|███████▍  | 270/363 [00:24<00:04, 20.93it/s]
Loading 0:  75%|███████▌  | 274/363 [00:24<00:03, 23.47it/s]
Loading 0:  76%|███████▋  | 277/363 [00:24<00:03, 22.90it/s]
Loading 0:  77%|███████▋  | 280/363 [00:24<00:03, 23.59it/s]
Loading 0:  78%|███████▊  | 283/363 [00:24<00:03, 24.52it/s]
Loading 0:  79%|███████▉  | 286/363 [00:24<00:03, 24.74it/s]
Loading 0:  80%|████████  | 291/363 [00:25<00:02, 28.00it/s]
Loading 0:  81%|████████  | 294/363 [00:25<00:02, 24.89it/s]
Loading 0:  82%|████████▏ | 299/363 [00:25<00:02, 28.16it/s]
Loading 0:  84%|████████▎ | 304/363 [00:25<00:02, 25.40it/s]
Loading 0:  85%|████████▍ | 307/363 [00:25<00:02, 23.13it/s]
Loading 0:  85%|████████▌ | 310/363 [00:25<00:02, 23.98it/s]
Loading 0:  86%|████████▌ | 313/363 [00:26<00:02, 24.53it/s]
Loading 0:  88%|████████▊ | 318/363 [00:26<00:01, 27.44it/s]
Loading 0:  88%|████████▊ | 321/363 [00:26<00:01, 25.03it/s]
Loading 0:  90%|█████████ | 327/363 [00:26<00:01, 30.18it/s]
Loading 0:  91%|█████████ | 331/363 [00:26<00:01, 28.64it/s]
Loading 0:  92%|█████████▏| 335/363 [00:26<00:00, 28.96it/s]
Loading 0:  93%|█████████▎| 338/363 [00:26<00:00, 27.36it/s]
Loading 0:  94%|█████████▍| 341/363 [00:33<00:13,  1.66it/s]
Loading 0:  96%|█████████▌| 347/363 [00:33<00:05,  2.78it/s]
Loading 0:  96%|█████████▋| 350/363 [00:33<00:03,  3.50it/s]
Loading 0:  98%|█████████▊| 355/363 [00:34<00:01,  5.20it/s]
Loading 0:  99%|█████████▉| 359/363 [00:34<00:00,  6.72it/s]
                                                            
Job chaiml-claude-dpo-try-exp1-v1-mkmlizer completed after 155.07s with status: succeeded
Stopping job with name chaiml-claude-dpo-try-exp1-v1-mkmlizer
Pipeline stage MKMLizer completed in 155.56s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.17s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service chaiml-claude-dpo-try-exp1-v1
Waiting for inference service chaiml-claude-dpo-try-exp1-v1 to be ready
Inference service chaiml-claude-dpo-try-exp1-v1 ready after 140.49225163459778s
Pipeline stage MKMLDeployer completed in 141.06s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.858937978744507s
Received healthy response to inference request in 2.8494014739990234s
Received healthy response to inference request in 2.8197615146636963s
Received healthy response to inference request in 2.3383865356445312s
Received healthy response to inference request in 2.564957618713379s
5 requests
0 failed requests
5th percentile: 2.3837007522583007
10th percentile: 2.42901496887207
20th percentile: 2.5196434020996095
30th percentile: 2.6159183979034424
40th percentile: 2.7178399562835693
50th percentile: 2.8197615146636963
60th percentile: 2.831617498397827
70th percentile: 2.843473482131958
80th percentile: 2.8513087749481203
90th percentile: 2.8551233768463136
95th percentile: 2.85703067779541
99th percentile: 2.8585565185546873
mean time: 2.6862890243530275
Pipeline stage StressChecker completed in 14.84s
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 0.67s
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 0.89s
Shutdown handler de-registered
chaiml-claude-dpo-try-exp1_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 3357.59s
Shutdown handler de-registered
chaiml-claude-dpo-try-exp1_v1 status is now inactive due to auto deactivation removed underperforming models
chaiml-claude-dpo-try-exp1_v1 status is now torndown due to DeploymentManager action