sao10k-mn-12b-lyra-v4b1

developer_uid: sao10k

submission_id: sao10k-mn-12b-lyra-v4b1_v1

model_name: lyra41

model_group: Sao10K/MN-12B-Lyra-v4b1

status: torndown

timestamp: 2024-09-07T18:02:40+00:00

num_battles: 16384

num_wins: 8394

celo_rating: 1248.87

family_friendly_score: 0.0

submission_type: basic

model_repo: Sao10K/MN-12B-Lyra-v4b1

model_architecture: MistralForCausalLM

model_num_parameters: 12772070400.0

best_of: 8

max_input_tokens: 512

max_output_tokens: 64

latencies: [{'batch_size': 1, 'throughput': 0.6961307462515147, 'latency_mean': 1.4364345574378967, 'latency_p50': 1.4271202087402344, 'latency_p90': 1.6032652854919434}, {'batch_size': 3, 'throughput': 1.341871792688169, 'latency_mean': 2.2224172461032867, 'latency_p50': 2.2169569730758667, 'latency_p90': 2.4407729148864745}, {'batch_size': 5, 'throughput': 1.591502348591622, 'latency_mean': 3.1314728307724, 'latency_p50': 3.1403199434280396, 'latency_p90': 3.4889400959014893}, {'batch_size': 6, 'throughput': 1.634934567782921, 'latency_mean': 3.6424233424663544, 'latency_p50': 3.645893692970276, 'latency_p90': 4.172836422920227}, {'batch_size': 8, 'throughput': 1.5740154779091822, 'latency_mean': 5.049681361913681, 'latency_p50': 5.020084619522095, 'latency_p90': 5.71440327167511}, {'batch_size': 10, 'throughput': 1.5693861696257319, 'latency_mean': 6.323752635717392, 'latency_p50': 6.318972110748291, 'latency_p90': 7.25449914932251}]

gpu_counts: {'NVIDIA RTX A5000': 1}

display_name: lyra41

is_internal_developer: False

language_model: Sao10K/MN-12B-Lyra-v4b1

model_size: 13B

ranking_group: single

throughput_3p7s: 1.64

us_pacific_date: 2024-09-07

win_ratio: 0.5123291015625

generation_params: {'temperature': 0.75, 'top_p': 1.0, 'min_p': 0.1, 'top_k': 40, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '\n\n', '\nYou:', '[/INST]', '<|im_end|>', '</s>'], 'max_input_tokens': 512, 'best_of': 8, 'max_output_tokens': 64}

formatter: {'memory_template': '<|im_start|>system\n{memory}<|im_end|>\n', 'prompt_template': '<|im_start|>user\n{prompt}<|im_end|>\n', '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': '<|im_start|>assistant\n{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 sao10k-mn-12b-lyra-v4b1-v1-mkmlizer
Waiting for job on sao10k-mn-12b-lyra-v4b1-v1-mkmlizer to finish
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║     _____            __           __                                ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║       /___/                                                         ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║                                                                     ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  Version: 0.10.1                                                    ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  https://mk1.ai                                                     ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║                                                                     ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  belonging to:                                                      ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║                                                                     ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  Chai Research Corp.                                                ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ║                                                                     ║
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: Downloaded to shared memory in 47.828s
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpnctde_tw, device:0
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
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
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: quantized model in 35.537s
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: Processed model Sao10K/MN-12B-Lyra-v4b1 in 83.366s
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: creating bucket guanaco-mkml-models
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/sao10k-mn-12b-lyra-v4b1-v1
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/sao10k-mn-12b-lyra-v4b1-v1/config.json
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/sao10k-mn-12b-lyra-v4b1-v1/special_tokens_map.json
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/sao10k-mn-12b-lyra-v4b1-v1/tokenizer_config.json
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/sao10k-mn-12b-lyra-v4b1-v1/tokenizer.json
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/sao10k-mn-12b-lyra-v4b1-v1/flywheel_model.0.safetensors
sao10k-mn-12b-lyra-v4b1-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|          | 2/363 [00:06<18:05,  3.01s/it]
Loading 0:   2%|▏         | 6/363 [00:06<04:48,  1.24it/s]
Loading 0:   4%|▎         | 13/363 [00:06<01:42,  3.40it/s]
Loading 0:   5%|▍         | 17/363 [00:06<01:09,  4.98it/s]
Loading 0:   6%|▋         | 23/363 [00:06<00:42,  8.09it/s]
Loading 0:   8%|▊         | 29/363 [00:06<00:28, 11.58it/s]
Loading 0:   9%|▉         | 34/363 [00:06<00:21, 14.98it/s]
Loading 0:  11%|█▏        | 41/363 [00:06<00:15, 21.20it/s]
Loading 0:  13%|█▎        | 47/363 [00:07<00:12, 24.99it/s]
Loading 0:  14%|█▍        | 52/363 [00:07<00:11, 27.75it/s]
Loading 0:  16%|█▋        | 59/363 [00:07<00:08, 34.33it/s]
Loading 0:  18%|█▊        | 65/363 [00:07<00:08, 36.13it/s]
Loading 0:  19%|█▉        | 70/363 [00:07<00:07, 37.14it/s]
Loading 0:  21%|██        | 75/363 [00:07<00:10, 28.73it/s]
Loading 0:  22%|██▏       | 79/363 [00:07<00:09, 28.97it/s]
Loading 0:  24%|██▎       | 86/363 [00:08<00:07, 35.70it/s]
Loading 0:  25%|██▌       | 92/363 [00:08<00:07, 36.83it/s]
Loading 0:  27%|██▋       | 97/363 [00:08<00:07, 37.15it/s]
Loading 0:  29%|██▊       | 104/363 [00:08<00:06, 42.79it/s]
Loading 0:  30%|███       | 110/363 [00:08<00:06, 42.00it/s]
Loading 0:  32%|███▏      | 115/363 [00:08<00:06, 41.08it/s]
Loading 0:  34%|███▎      | 122/363 [00:08<00:05, 45.92it/s]
Loading 0:  35%|███▍      | 127/363 [00:08<00:05, 45.88it/s]
Loading 0:  36%|███▋      | 132/363 [00:09<00:06, 36.99it/s]
Loading 0:  39%|███▊      | 140/363 [00:09<00:04, 44.84it/s]
Loading 0:  40%|████      | 146/363 [00:09<00:04, 43.97it/s]
Loading 0:  42%|████▏     | 151/363 [00:09<00:04, 42.62it/s]
Loading 0:  43%|████▎     | 157/363 [00:09<00:04, 46.21it/s]
Loading 0:  45%|████▍     | 162/363 [00:09<00:04, 46.05it/s]
Loading 0:  46%|████▌     | 167/363 [00:09<00:04, 46.93it/s]
Loading 0:  48%|████▊     | 173/363 [00:10<00:04, 44.17it/s]
Loading 0:  49%|████▉     | 178/363 [00:10<00:04, 42.44it/s]
Loading 0:  51%|█████     | 184/363 [00:10<00:04, 44.35it/s]
Loading 0:  52%|█████▏    | 189/363 [00:10<00:05, 30.01it/s]
Loading 0:  53%|█████▎    | 194/363 [00:10<00:05, 33.31it/s]
Loading 0:  55%|█████▌    | 200/363 [00:10<00:04, 35.59it/s]
Loading 0:  56%|█████▌    | 204/363 [00:10<00:04, 33.79it/s]
Loading 0:  58%|█████▊    | 212/363 [00:11<00:03, 42.26it/s]
Loading 0:  60%|██████    | 218/363 [00:11<00:03, 41.71it/s]
Loading 0:  61%|██████▏   | 223/363 [00:11<00:03, 40.80it/s]
Loading 0:  63%|██████▎   | 229/363 [00:11<00:02, 44.89it/s]
Loading 0:  64%|██████▍   | 234/363 [00:11<00:02, 44.42it/s]
Loading 0:  66%|██████▌   | 239/363 [00:11<00:02, 44.36it/s]
Loading 0:  67%|██████▋   | 245/363 [00:11<00:02, 42.42it/s]
Loading 0:  69%|██████▉   | 250/363 [00:11<00:02, 42.01it/s]
Loading 0:  71%|███████   | 257/363 [00:12<00:02, 47.62it/s]
Loading 0:  72%|███████▏  | 262/363 [00:12<00:02, 48.21it/s]
Loading 0:  74%|███████▎  | 267/363 [00:12<00:02, 39.39it/s]
Loading 0:  75%|███████▌  | 274/363 [00:12<00:01, 45.22it/s]
Loading 0:  77%|███████▋  | 279/363 [00:12<00:01, 44.05it/s]
Loading 0:  78%|███████▊  | 284/363 [00:12<00:01, 42.73it/s]
Loading 0:  80%|███████▉  | 289/363 [00:12<00:01, 43.41it/s]
Loading 0:  81%|████████  | 294/363 [00:13<00:01, 36.06it/s]
Loading 0:  83%|████████▎ | 301/363 [00:13<00:01, 43.58it/s]
Loading 0:  84%|████████▍ | 306/363 [00:13<00:01, 29.75it/s]
Loading 0:  86%|████████▌ | 311/363 [00:13<00:01, 33.15it/s]
Loading 0:  87%|████████▋ | 317/363 [00:13<00:01, 35.41it/s]
Loading 0:  89%|████████▊ | 322/363 [00:13<00:01, 36.98it/s]
Loading 0:  91%|█████████ | 329/363 [00:13<00:00, 42.75it/s]
Loading 0:  92%|█████████▏| 334/363 [00:14<00:00, 43.27it/s]
Loading 0:  93%|█████████▎| 339/363 [00:14<00:00, 35.70it/s]
Loading 0:  95%|█████████▌| 346/363 [00:14<00:00, 43.04it/s]
Loading 0:  97%|█████████▋| 351/363 [00:14<00:00, 43.28it/s]
Loading 0:  98%|█████████▊| 356/363 [00:14<00:00, 43.43it/s]
Loading 0: 100%|█████████▉| 362/363 [00:14<00:00, 42.17it/s]
                                                            
Job sao10k-mn-12b-lyra-v4b1-v1-mkmlizer completed after 106.74s with status: succeeded
Stopping job with name sao10k-mn-12b-lyra-v4b1-v1-mkmlizer
Pipeline stage MKMLizer completed in 107.80s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.12s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service sao10k-mn-12b-lyra-v4b1-v1
Waiting for inference service sao10k-mn-12b-lyra-v4b1-v1 to be ready
Failed to get response for submission chaiml-elo-alignment-run-3_v36: ('http://chaiml-llama-8b-pairwis-8189-v19-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', 'read tcp 127.0.0.1:34444->127.0.0.1:8080: read: connection reset by peer\n')
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
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
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
Connection pool is full, discarding connection: %s. Connection pool size: %s
Inference service sao10k-mn-12b-lyra-v4b1-v1 ready after 150.791246175766s
Pipeline stage MKMLDeployer completed in 151.18s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.824636459350586s
Received healthy response to inference request in 2.150804281234741s
Received healthy response to inference request in 1.831810474395752s
Received healthy response to inference request in 2.042384624481201s
Received healthy response to inference request in 2.898942470550537s
5 requests
0 failed requests
5th percentile: 1.8739253044128419
10th percentile: 1.9160401344299316
20th percentile: 2.0002697944641112
30th percentile: 2.064068555831909
40th percentile: 2.107436418533325
50th percentile: 2.150804281234741
60th percentile: 2.4203371524810793
70th percentile: 2.689870023727417
80th percentile: 2.8394976615905763
90th percentile: 2.8692200660705565
95th percentile: 2.884081268310547
99th percentile: 2.895970230102539
mean time: 2.3497156620025637
Pipeline stage StressChecker completed in 12.64s
run pipeline stage %s
Running pipeline stage TriggerMKMLProfilingPipeline
run_pipeline:run_in_cloud %s
starting trigger_guanaco_pipeline args=%s
Pipeline stage TriggerMKMLProfilingPipeline completed in 5.98s
Shutdown handler de-registered
sao10k-mn-12b-lyra-v4b1_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 0.15s
run pipeline stage %s
Running pipeline stage MKMLProfilerTemplater
Pipeline stage MKMLProfilerTemplater completed in 0.13s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service sao10k-mn-12b-lyra-v4b1-v1-profiler
Waiting for inference service sao10k-mn-12b-lyra-v4b1-v1-profiler to be ready
Inference service sao10k-mn-12b-lyra-v4b1-v1-profiler ready after 150.41019916534424s
Pipeline stage MKMLProfilerDeployer completed in 151.04s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/sao10k-mn-12b-lyra-v4b1-v1-profiler-predictor-00001-deploymjvwc:/code/chaiverse_profiler_1725732633 --namespace tenant-chaiml-guanaco
kubectl exec -it sao10k-mn-12b-lyra-v4b1-v1-profiler-predictor-00001-deploymjvwc --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1725732633 && 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 512 --output_tokens 64 --summary /code/chaiverse_profiler_1725732633/summary.json'
kubectl exec -it sao10k-mn-12b-lyra-v4b1-v1-profiler-predictor-00001-deploymjvwc --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1725732633/summary.json'
Pipeline stage MKMLProfilerRunner completed in 946.11s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service sao10k-mn-12b-lyra-v4b1-v1-profiler is running
Tearing down inference service sao10k-mn-12b-lyra-v4b1-v1-profiler
Service sao10k-mn-12b-lyra-v4b1-v1-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 1.78s
Shutdown handler de-registered
sao10k-mn-12b-lyra-v4b1_v1 status is now inactive due to auto deactivation removed underperforming models
Shutdown handler not registered because Python interpreter is not running in the main thread
admin requested tearing down of sao10k-mn-12b-lyra-v4b1_v1
run pipeline %s
Shutdown handler not registered because Python interpreter is not running in the main thread
run pipeline stage %s
run pipeline %s
Running pipeline stage MKMLDeleter
run pipeline stage %s
Running pipeline stage MKMLDeleter
%s, retrying in %s seconds...
%s, retrying in %s seconds...
clean up pipeline due to error=TeardownError("module 'kubernetes.config' has no attribute 'load_kube_config'")
Shutdown handler de-registered
sao10k-mn-12b-lyra-v4b1_v1 status is now torndown due to DeploymentManager action