rica40325-alignment-first

developer_uid: rica40325

submission_id: rica40325-alignment-first_v3

model_name: rica40325-alignment-first_v1

model_group: rica40325/alignment-firs

status: torndown

timestamp: 2024-09-08T10:04:10+00:00

num_battles: 15441

num_wins: 7896

celo_rating: 1248.49

family_friendly_score: 0.0

submission_type: basic

model_repo: rica40325/alignment-first

model_architecture: LlamaForCausalLM

model_num_parameters: 8030261248.0

best_of: 16

max_input_tokens: 512

max_output_tokens: 64

latencies: [{'batch_size': 1, 'throughput': 0.8909033021566081, 'latency_mean': 1.12235560297966, 'latency_p50': 1.1197633743286133, 'latency_p90': 1.2540004253387451}, {'batch_size': 4, 'throughput': 1.7685134777154432, 'latency_mean': 2.249322760105133, 'latency_p50': 2.2412177324295044, 'latency_p90': 2.5549634456634522}, {'batch_size': 5, 'throughput': 1.861013211821262, 'latency_mean': 2.674932026863098, 'latency_p50': 2.6699440479278564, 'latency_p90': 3.002667999267578}, {'batch_size': 8, 'throughput': 1.980402312847816, 'latency_mean': 4.009588742256165, 'latency_p50': 3.9824423789978027, 'latency_p90': 4.521251249313354}, {'batch_size': 10, 'throughput': 1.987606447086303, 'latency_mean': 4.9865049409866335, 'latency_p50': 4.990198135375977, 'latency_p90': 5.750533151626587}, {'batch_size': 12, 'throughput': 1.9534141875485318, 'latency_mean': 6.059146504402161, 'latency_p50': 6.115759968757629, 'latency_p90': 6.853640103340149}, {'batch_size': 15, 'throughput': 1.9885152603039873, 'latency_mean': 7.398142468929291, 'latency_p50': 7.4979411363601685, 'latency_p90': 8.230498838424683}]

gpu_counts: {'NVIDIA RTX A5000': 1}

display_name: rica40325-alignment-first_v1

is_internal_developer: False

language_model: rica40325/alignment-first

model_size: 8B

ranking_group: single

throughput_3p7s: 1.97

us_pacific_date: 2024-09-08

win_ratio: 0.5113658441810763

generation_params: {'temperature': 0.3, 'top_p': 0.95, 'min_p': 0.05, 'top_k': 80, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n'], 'max_input_tokens': 512, 'best_of': 16, 'max_output_tokens': 64}

formatter: {'memory_template': "{bot_name}'s Persona: {memory}\n####\n", 'prompt_template': '{prompt}\n<START>\n', '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 rica40325-alignment-first-v3-mkmlizer
Waiting for job on rica40325-alignment-first-v3-mkmlizer to finish
rica40325-alignment-first-v3-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
rica40325-alignment-first-v3-mkmlizer: ║     _____            __           __                                ║
rica40325-alignment-first-v3-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
rica40325-alignment-first-v3-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
rica40325-alignment-first-v3-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
rica40325-alignment-first-v3-mkmlizer: ║       /___/                                                         ║
rica40325-alignment-first-v3-mkmlizer: ║                                                                     ║
rica40325-alignment-first-v3-mkmlizer: ║  Version: 0.10.1                                                    ║
rica40325-alignment-first-v3-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
rica40325-alignment-first-v3-mkmlizer: ║  https://mk1.ai                                                     ║
rica40325-alignment-first-v3-mkmlizer: ║                                                                     ║
rica40325-alignment-first-v3-mkmlizer: ║  The license key for the current software has been verified as      ║
rica40325-alignment-first-v3-mkmlizer: ║  belonging to:                                                      ║
rica40325-alignment-first-v3-mkmlizer: ║                                                                     ║
rica40325-alignment-first-v3-mkmlizer: ║  Chai Research Corp.                                                ║
rica40325-alignment-first-v3-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
rica40325-alignment-first-v3-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
rica40325-alignment-first-v3-mkmlizer: ║                                                                     ║
rica40325-alignment-first-v3-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
rica40325-alignment-first-v3-mkmlizer: Downloaded to shared memory in 21.333s
rica40325-alignment-first-v3-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmptj2ohme_, device:0
rica40325-alignment-first-v3-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
Connection pool is full, discarding connection: %s. Connection pool size: %s
rica40325-alignment-first-v3-mkmlizer: quantized model in 26.774s
rica40325-alignment-first-v3-mkmlizer: Processed model rica40325/alignment-first in 48.107s
rica40325-alignment-first-v3-mkmlizer: creating bucket guanaco-mkml-models
rica40325-alignment-first-v3-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
rica40325-alignment-first-v3-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/rica40325-alignment-first-v3
rica40325-alignment-first-v3-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/rica40325-alignment-first-v3/config.json
rica40325-alignment-first-v3-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/rica40325-alignment-first-v3/special_tokens_map.json
rica40325-alignment-first-v3-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/rica40325-alignment-first-v3/tokenizer_config.json
rica40325-alignment-first-v3-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/rica40325-alignment-first-v3/tokenizer.json
rica40325-alignment-first-v3-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/rica40325-alignment-first-v3/flywheel_model.0.safetensors
rica40325-alignment-first-v3-mkmlizer: 
Loading 0:   0%|          | 0/291 [00:00<?, ?it/s]
Loading 0:   2%|▏         | 7/291 [00:00<00:06, 43.95it/s]
Loading 0:   5%|▌         | 16/291 [00:00<00:04, 58.22it/s]
Loading 0:   9%|▊         | 25/291 [00:00<00:04, 65.47it/s]
Loading 0:  12%|█▏        | 34/291 [00:00<00:03, 73.22it/s]
Loading 0:  15%|█▍        | 43/291 [00:00<00:03, 70.55it/s]
Loading 0:  18%|█▊        | 52/291 [00:00<00:03, 71.98it/s]
Loading 0:  21%|██        | 61/291 [00:00<00:03, 75.69it/s]
Loading 0:  24%|██▍       | 70/291 [00:00<00:02, 78.07it/s]
Loading 0:  27%|██▋       | 79/291 [00:01<00:02, 78.84it/s]
Loading 0:  30%|██▉       | 87/291 [00:02<00:10, 19.51it/s]
Loading 0:  32%|███▏      | 94/291 [00:02<00:08, 23.52it/s]
Loading 0:  35%|███▌      | 103/291 [00:02<00:06, 29.97it/s]
Loading 0:  38%|███▊      | 112/291 [00:02<00:04, 37.58it/s]
Loading 0:  42%|████▏     | 121/291 [00:02<00:03, 45.38it/s]
Loading 0:  45%|████▍     | 130/291 [00:02<00:03, 50.06it/s]
Loading 0:  48%|████▊     | 139/291 [00:03<00:02, 55.74it/s]
Loading 0:  51%|█████     | 148/291 [00:03<00:02, 62.42it/s]
Loading 0:  54%|█████▍    | 157/291 [00:03<00:01, 68.10it/s]
Loading 0:  57%|█████▋    | 166/291 [00:03<00:01, 66.43it/s]
Loading 0:  60%|██████    | 175/291 [00:03<00:01, 67.04it/s]
Loading 0:  63%|██████▎   | 184/291 [00:03<00:01, 71.00it/s]
Loading 0:  66%|██████▌   | 192/291 [00:04<00:04, 20.31it/s]
Loading 0:  68%|██████▊   | 198/291 [00:04<00:04, 23.18it/s]
Loading 0:  70%|███████   | 205/291 [00:04<00:03, 28.31it/s]
Loading 0:  74%|███████▎  | 214/291 [00:05<00:02, 35.20it/s]
Loading 0:  77%|███████▋  | 223/291 [00:05<00:01, 42.16it/s]
Loading 0:  80%|███████▉  | 232/291 [00:05<00:01, 50.56it/s]
Loading 0:  83%|████████▎ | 241/291 [00:05<00:00, 58.47it/s]
Loading 0:  86%|████████▌ | 250/291 [00:05<00:00, 63.82it/s]
Loading 0:  89%|████████▉ | 259/291 [00:05<00:00, 69.21it/s]
Loading 0:  92%|█████████▏| 268/291 [00:05<00:00, 73.79it/s]
Loading 0:  95%|█████████▌| 277/291 [00:05<00:00, 77.68it/s]
Loading 0:  98%|█████████▊| 286/291 [00:05<00:00, 78.20it/s]
                                                            
Job rica40325-alignment-first-v3-mkmlizer completed after 76.9s with status: succeeded
Stopping job with name rica40325-alignment-first-v3-mkmlizer
Pipeline stage MKMLizer completed in 77.86s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.10s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service rica40325-alignment-first-v3
Waiting for inference service rica40325-alignment-first-v3 to be ready
Connection pool is full, discarding connection: %s. Connection pool size: %s
Failed to get response for submission rica40325-alignment-first_v2: ('http://chaiml-llama-8b-pairwis-8189-v19-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', 'readfrom tcp 127.0.0.1:38480->127.0.0.1:8080: write tcp 127.0.0.1:38480->127.0.0.1:8080: use of closed network connection\n')
Inference service rica40325-alignment-first-v3 ready after 150.67228984832764s
Pipeline stage MKMLDeployer completed in 151.18s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.689027786254883s
Received healthy response to inference request in 1.461299180984497s
Received healthy response to inference request in 1.8611476421356201s
Received healthy response to inference request in 2.2661995887756348s
Received healthy response to inference request in 1.880995750427246s
5 requests
0 failed requests
5th percentile: 1.5412688732147217
10th percentile: 1.6212385654449464
20th percentile: 1.7811779499053955
30th percentile: 1.8651172637939453
40th percentile: 1.8730565071105958
50th percentile: 1.880995750427246
60th percentile: 2.0350772857666017
70th percentile: 2.189158821105957
80th percentile: 2.3507652282714844
90th percentile: 2.5198965072631836
95th percentile: 2.604462146759033
99th percentile: 2.6721146583557127
mean time: 2.031733989715576
Pipeline stage StressChecker completed in 10.94s
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 5.22s
Shutdown handler de-registered
rica40325-alignment-first_v3 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 0.18s
run pipeline stage %s
Running pipeline stage MKMLProfilerTemplater
Pipeline stage MKMLProfilerTemplater completed in 0.10s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service rica40325-alignment-first-v3-profiler
Waiting for inference service rica40325-alignment-first-v3-profiler to be ready
Inference service rica40325-alignment-first-v3-profiler ready after 140.53656435012817s
Pipeline stage MKMLProfilerDeployer completed in 140.91s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/rica40325-alignment-5b6856c3fd2e3d84c84cdfd4b5f9e322-deplopgqhf:/code/chaiverse_profiler_1725790283 --namespace tenant-chaiml-guanaco
kubectl exec -it rica40325-alignment-5b6856c3fd2e3d84c84cdfd4b5f9e322-deplopgqhf --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1725790283 && python profiles.py profile --best_of_n 16 --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_1725790283/summary.json'
kubectl exec -it rica40325-alignment-5b6856c3fd2e3d84c84cdfd4b5f9e322-deplopgqhf --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1725790283/summary.json'
Pipeline stage MKMLProfilerRunner completed in 855.47s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service rica40325-alignment-first-v3-profiler is running
Tearing down inference service rica40325-alignment-first-v3-profiler
Service rica40325-alignment-first-v3-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 1.60s
Shutdown handler de-registered
rica40325-alignment-first_v3 status is now inactive due to auto deactivation removed underperforming models
rica40325-alignment-first_v3 status is now torndown due to DeploymentManager action