chaiml-espresso-llama-24_7737

developer_uid: azuruce

submission_id: chaiml-espresso-llama-24_7737_v2

model_name: chaiml-espresso-llama-24_7737_v2

model_group: ChaiML/espresso_llama_24

status: inactive

timestamp: 2024-12-04T17:06:19+00:00

num_battles: 17454

num_wins: 8525

celo_rating: 1253.82

family_friendly_score: 0.5822

family_friendly_standard_error: 0.006974857131153297

submission_type: basic

model_repo: ChaiML/espresso_llama_241204_albert_v1

model_architecture: MistralForCausalLM

model_num_parameters: 22247282688.0

best_of: 8

max_input_tokens: 1024

max_output_tokens: 64

latencies: [{'batch_size': 1, 'throughput': 0.3753846765190991, 'latency_mean': 2.6638700127601624, 'latency_p50': 2.6558258533477783, 'latency_p90': 2.9466554641723635}, {'batch_size': 2, 'throughput': 0.5870767993434286, 'latency_mean': 3.406644265651703, 'latency_p50': 3.4146839380264282, 'latency_p90': 3.7027806282043456}, {'batch_size': 3, 'throughput': 0.7447888169864235, 'latency_mean': 4.017598930597305, 'latency_p50': 4.018911600112915, 'latency_p90': 4.405026841163635}, {'batch_size': 4, 'throughput': 0.8613319956890111, 'latency_mean': 4.618503798246383, 'latency_p50': 4.621257662773132, 'latency_p90': 5.193314719200134}, {'batch_size': 5, 'throughput': 0.9503079005536844, 'latency_mean': 5.224446338415146, 'latency_p50': 5.223620176315308, 'latency_p90': 5.878749966621399}]

gpu_counts: {'NVIDIA RTX A6000': 1}

display_name: chaiml-espresso-llama-24_7737_v2

is_internal_developer: True

language_model: ChaiML/espresso_llama_241204_albert_v1

model_size: 22B

ranking_group: single

throughput_3p7s: 0.67

us_pacific_date: 2024-12-04

win_ratio: 0.48842672166838547

generation_params: {'temperature': 0.9, 'top_p': 1.0, 'min_p': 0.0, 'top_k': 40, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '<|eot_id|>', '<|end_of_text|>', 'You:'], 'max_input_tokens': 1024, 'best_of': 8, 'max_output_tokens': 64}

formatter: {'memory_template': '', 'prompt_template': '', 'bot_template': '{bot_name}: {message}\n', 'user_template': '{user_name}: {message}\n', 'response_template': '{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-espresso-llama-24-7737-v2-mkmlizer
Waiting for job on chaiml-espresso-llama-24-7737-v2-mkmlizer to finish
chaiml-espresso-llama-24-7737-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║     _____            __           __                                ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║       /___/                                                         ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║                                                                     ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  Version: 0.11.12                                                   ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║                                                                     ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  belonging to:                                                      ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║                                                                     ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║  Expiration: 2025-01-15 23:59:59                                    ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ║                                                                     ║
chaiml-espresso-llama-24-7737-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
chaiml-espresso-llama-24-7737-v2-mkmlizer: Downloaded to shared memory in 63.071s
chaiml-espresso-llama-24-7737-v2-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpu_tty5pr, device:0
chaiml-espresso-llama-24-7737-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-espresso-llama-24-7737-v2-mkmlizer: quantized model in 58.571s
chaiml-espresso-llama-24-7737-v2-mkmlizer: Processed model ChaiML/espresso_llama_241204_albert_v1 in 121.642s
chaiml-espresso-llama-24-7737-v2-mkmlizer: creating bucket guanaco-mkml-models
chaiml-espresso-llama-24-7737-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-espresso-llama-24-7737-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-espresso-llama-24-7737-v2
chaiml-espresso-llama-24-7737-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-espresso-llama-24-7737-v2/config.json
chaiml-espresso-llama-24-7737-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-espresso-llama-24-7737-v2/special_tokens_map.json
chaiml-espresso-llama-24-7737-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-espresso-llama-24-7737-v2/tokenizer_config.json
chaiml-espresso-llama-24-7737-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-espresso-llama-24-7737-v2/tokenizer.json
chaiml-espresso-llama-24-7737-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/chaiml-espresso-llama-24-7737-v2/flywheel_model.1.safetensors
chaiml-espresso-llama-24-7737-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-espresso-llama-24-7737-v2/flywheel_model.0.safetensors
chaiml-espresso-llama-24-7737-v2-mkmlizer: 
Loading 0:   0%|          | 0/507 [00:00<?, ?it/s]
Loading 0:   1%|          | 5/507 [00:00<00:22, 22.75it/s]
Loading 0:   2%|▏         | 10/507 [00:00<00:15, 32.51it/s]
Loading 0:   3%|▎         | 14/507 [00:00<00:17, 28.23it/s]
Loading 0:   4%|▎         | 19/507 [00:00<00:14, 33.41it/s]
Loading 0:   5%|▍         | 23/507 [00:00<00:16, 29.19it/s]
Loading 0:   6%|▌         | 28/507 [00:00<00:14, 33.30it/s]
Loading 0:   6%|▋         | 32/507 [00:01<00:15, 29.81it/s]
Loading 0:   7%|▋         | 37/507 [00:01<00:13, 33.61it/s]
Loading 0:   8%|▊         | 41/507 [00:01<00:15, 29.45it/s]
Loading 0:   9%|▉         | 46/507 [00:01<00:13, 33.08it/s]
Loading 0:  10%|▉         | 50/507 [00:01<00:15, 29.75it/s]
Loading 0:  11%|█         | 54/507 [00:01<00:20, 21.85it/s]
Loading 0:  11%|█         | 57/507 [00:02<00:20, 21.69it/s]
Loading 0:  12%|█▏        | 61/507 [00:02<00:17, 24.85it/s]
Loading 0:  13%|█▎        | 65/507 [00:02<00:18, 24.08it/s]
Loading 0:  14%|█▍        | 71/507 [00:02<00:13, 31.15it/s]
Loading 0:  15%|█▍        | 75/507 [00:02<00:13, 32.05it/s]
Loading 0:  16%|█▌        | 80/507 [00:02<00:13, 32.13it/s]
Loading 0:  17%|█▋        | 85/507 [00:02<00:11, 36.06it/s]
Loading 0:  18%|█▊        | 89/507 [00:03<00:13, 31.98it/s]
Loading 0:  19%|█▊        | 94/507 [00:03<00:11, 35.25it/s]
Loading 0:  19%|█▉        | 98/507 [00:03<00:13, 31.39it/s]
Loading 0:  20%|██        | 103/507 [00:03<00:11, 35.24it/s]
Loading 0:  21%|██        | 107/507 [00:03<00:12, 31.35it/s]
Loading 0:  22%|██▏       | 112/507 [00:03<00:11, 34.71it/s]
Loading 0:  23%|██▎       | 116/507 [00:04<00:17, 21.84it/s]
Loading 0:  24%|██▍       | 122/507 [00:04<00:15, 24.38it/s]
Loading 0:  25%|██▌       | 129/507 [00:04<00:12, 30.56it/s]
Loading 0:  26%|██▌       | 133/507 [00:04<00:12, 30.11it/s]
Loading 0:  27%|██▋       | 138/507 [00:04<00:11, 32.71it/s]
Loading 0:  28%|██▊       | 142/507 [00:04<00:11, 31.17it/s]
Loading 0:  29%|██▉       | 146/507 [00:04<00:11, 32.50it/s]
Loading 0:  30%|██▉       | 150/507 [00:05<00:12, 28.89it/s]
Loading 0:  30%|███       | 154/507 [00:05<00:11, 30.31it/s]
Loading 0:  31%|███       | 158/507 [00:05<00:12, 28.61it/s]
Loading 0:  32%|███▏      | 163/507 [00:05<00:10, 33.46it/s]
Loading 0:  33%|███▎      | 168/507 [00:05<00:09, 34.94it/s]
Loading 0:  34%|███▍      | 172/507 [00:05<00:14, 23.29it/s]
Loading 0:  35%|███▍      | 176/507 [00:06<00:14, 22.95it/s]
Loading 0:  36%|███▌      | 181/507 [00:06<00:11, 27.78it/s]
Loading 0:  36%|███▋      | 185/507 [00:06<00:11, 27.15it/s]
Loading 0:  37%|███▋      | 190/507 [00:06<00:10, 31.58it/s]
Loading 0:  38%|███▊      | 194/507 [00:06<00:10, 29.42it/s]
Loading 0:  39%|███▉      | 199/507 [00:06<00:09, 33.80it/s]
Loading 0:  40%|████      | 203/507 [00:06<00:09, 30.78it/s]
Loading 0:  41%|████▏     | 210/507 [00:06<00:07, 37.40it/s]
Loading 0:  42%|████▏     | 214/507 [00:07<00:08, 35.31it/s]
Loading 0:  43%|████▎     | 218/507 [00:07<00:08, 34.74it/s]
Loading 0:  44%|████▍     | 222/507 [00:07<00:08, 33.41it/s]
Loading 0:  45%|████▍     | 226/507 [00:07<00:11, 25.28it/s]
Loading 0:  45%|████▌     | 230/507 [00:07<00:11, 24.90it/s]
Loading 0:  46%|████▋     | 235/507 [00:07<00:09, 29.65it/s]
Loading 0:  47%|████▋     | 239/507 [00:08<00:09, 27.41it/s]
Loading 0:  48%|████▊     | 244/507 [00:08<00:08, 31.18it/s]
Loading 0:  49%|████▉     | 248/507 [00:08<00:08, 29.17it/s]
Loading 0:  50%|█████     | 255/507 [00:08<00:06, 36.27it/s]
Loading 0:  51%|█████     | 259/507 [00:08<00:07, 34.67it/s]
Loading 0:  52%|█████▏    | 264/507 [00:08<00:06, 37.10it/s]
Loading 0:  53%|█████▎    | 268/507 [00:08<00:06, 36.07it/s]
Loading 0:  54%|█████▍    | 273/507 [00:08<00:06, 38.16it/s]
Loading 0:  55%|█████▍    | 277/507 [00:09<00:06, 36.22it/s]
Loading 0:  56%|█████▌    | 283/507 [00:09<00:05, 38.12it/s]
Loading 0:  57%|█████▋    | 287/507 [00:09<00:08, 24.57it/s]
Loading 0:  58%|█████▊    | 293/507 [00:09<00:07, 27.77it/s]
Loading 0:  59%|█████▉    | 299/507 [00:20<00:07, 27.77it/s]
Loading 0:  59%|█████▉    | 300/507 [00:32<04:12,  1.22s/it]
Loading 0:  60%|█████▉    | 302/507 [00:32<03:38,  1.07s/it]
Loading 0:  61%|██████    | 307/507 [00:32<02:27,  1.36it/s]
Loading 0:  61%|██████    | 310/507 [00:32<01:56,  1.70it/s]
Loading 0:  62%|██████▏   | 313/507 [00:32<01:29,  2.17it/s]
Loading 0:  62%|██████▏   | 316/507 [00:32<01:07,  2.83it/s]
Loading 0:  63%|██████▎   | 320/507 [00:33<00:47,  3.93it/s]
Loading 0:  64%|██████▍   | 325/507 [00:33<00:30,  5.93it/s]
Loading 0:  65%|██████▍   | 329/507 [00:33<00:23,  7.64it/s]
Loading 0:  66%|██████▌   | 335/507 [00:33<00:15, 11.30it/s]
Loading 0:  67%|██████▋   | 340/507 [00:33<00:12, 12.97it/s]
Loading 0:  68%|██████▊   | 344/507 [00:33<00:10, 15.33it/s]
Loading 0:  69%|██████▊   | 348/507 [00:34<00:09, 16.97it/s]
Loading 0:  69%|██████▉   | 352/507 [00:34<00:07, 20.12it/s]
Loading 0:  70%|███████   | 356/507 [00:34<00:07, 21.54it/s]
Loading 0:  72%|███████▏  | 363/507 [00:34<00:04, 29.08it/s]
Loading 0:  72%|███████▏  | 367/507 [00:34<00:04, 29.57it/s]
Loading 0:  73%|███████▎  | 372/507 [00:34<00:04, 32.73it/s]
Loading 0:  74%|███████▍  | 376/507 [00:34<00:03, 33.37it/s]
Loading 0:  75%|███████▌  | 381/507 [00:34<00:03, 36.83it/s]
Loading 0:  76%|███████▌  | 386/507 [00:35<00:03, 37.36it/s]
Loading 0:  77%|███████▋  | 390/507 [00:35<00:03, 30.16it/s]
Loading 0:  78%|███████▊  | 395/507 [00:35<00:04, 26.29it/s]
Loading 0:  79%|███████▉  | 401/507 [00:35<00:03, 27.83it/s]
Loading 0:  80%|████████  | 406/507 [00:35<00:03, 31.78it/s]
Loading 0:  81%|████████  | 410/507 [00:35<00:03, 29.05it/s]
Loading 0:  82%|████████▏ | 415/507 [00:36<00:02, 33.13it/s]
Loading 0:  83%|████████▎ | 419/507 [00:36<00:02, 30.03it/s]
Loading 0:  84%|████████▎ | 424/507 [00:36<00:02, 33.99it/s]
Loading 0:  84%|████████▍ | 428/507 [00:36<00:02, 30.88it/s]
Loading 0:  85%|████████▌ | 433/507 [00:36<00:02, 35.08it/s]
Loading 0:  86%|████████▌ | 437/507 [00:36<00:02, 31.60it/s]
Loading 0:  88%|████████▊ | 444/507 [00:36<00:01, 38.39it/s]
Loading 0:  89%|████████▊ | 449/507 [00:37<00:01, 36.96it/s]
Loading 0:  90%|████████▉ | 454/507 [00:37<00:01, 35.26it/s]
Loading 0:  90%|█████████ | 458/507 [00:39<00:08,  5.85it/s]
Loading 0:  91%|█████████ | 461/507 [00:39<00:06,  6.98it/s]
Loading 0:  92%|█████████▏| 465/507 [00:39<00:04,  8.78it/s]
Loading 0:  93%|█████████▎| 472/507 [00:40<00:02, 13.54it/s]
Loading 0:  94%|█████████▍| 476/507 [00:40<00:01, 15.64it/s]
Loading 0:  95%|█████████▍| 481/507 [00:40<00:01, 19.21it/s]
Loading 0:  96%|█████████▌| 485/507 [00:40<00:01, 20.74it/s]
Loading 0:  96%|█████████▋| 489/507 [00:40<00:00, 23.66it/s]
Loading 0:  97%|█████████▋| 493/507 [00:40<00:00, 22.67it/s]
Loading 0:  98%|█████████▊| 497/507 [00:40<00:00, 25.49it/s]
Loading 0:  99%|█████████▉| 501/507 [00:41<00:00, 23.19it/s]
Loading 0: 100%|█████████▉| 506/507 [00:41<00:00, 27.91it/s]
                                                            
Job chaiml-espresso-llama-24-7737-v2-mkmlizer completed after 145.15s with status: succeeded
Stopping job with name chaiml-espresso-llama-24-7737-v2-mkmlizer
Pipeline stage MKMLizer completed in 145.64s
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-espresso-llama-24-7737-v2
Waiting for inference service chaiml-espresso-llama-24-7737-v2 to be ready
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
Inference service chaiml-espresso-llama-24-7737-v2 ready after 220.74558186531067s
Pipeline stage MKMLDeployer completed in 221.29s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.8766887187957764s
Received healthy response to inference request in 2.465132474899292s
Received healthy response to inference request in 2.297407388687134s
Received healthy response to inference request in 2.4710235595703125s
Received healthy response to inference request in 2.688887596130371s
5 requests
0 failed requests
5th percentile: 2.3309524059295654
10th percentile: 2.364497423171997
20th percentile: 2.4315874576568604
30th percentile: 2.4663106918334963
40th percentile: 2.4686671257019044
50th percentile: 2.4710235595703125
60th percentile: 2.5581691741943358
70th percentile: 2.6453147888183595
80th percentile: 2.726447820663452
90th percentile: 2.8015682697296143
95th percentile: 2.8391284942626953
99th percentile: 2.86917667388916
mean time: 2.559827947616577
Pipeline stage StressChecker completed in 13.93s
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 2.09s
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 1.99s
Shutdown handler de-registered
chaiml-espresso-llama-24_7737_v2 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
Pipeline stage OfflineFamilyFriendlyScorer completed in 3184.16s
Shutdown handler de-registered
chaiml-espresso-llama-24_7737_v2 status is now inactive due to auto deactivation removed underperforming models