chaiml-albert-dpo-0912-v_5607

submission_id: chaiml-albert-dpo-0912-v_5607_v1

developer_uid: chai_backend_admin

alignment_samples: 11289

alignment_score: -0.40274061292312097

best_of: 8

celo_rating: 1254.69

display_name: virgo_edit_variety_random

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}

generation_params: {'temperature': 0.9, 'top_p': 1.0, 'min_p': 0.05, 'top_k': 80, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '</s>', '###', 'Bot:', 'User:', 'You:', '<|im_end|>'], 'max_input_tokens': 1024, 'best_of': 8, 'max_output_tokens': 64}

gpu_counts: {'NVIDIA RTX A5000': 1}

is_internal_developer: True

language_model: ChaiML/albert_dpo_0912_virgo_edit_variety_random

latencies: [{'batch_size': 1, 'throughput': 0.6060108229213704, 'latency_mean': 1.6500753331184388, 'latency_p50': 1.6577926874160767, 'latency_p90': 1.8190032005310057}, {'batch_size': 3, 'throughput': 1.061253327649267, 'latency_mean': 2.8168992269039155, 'latency_p50': 2.8370977640151978, 'latency_p90': 3.06857373714447}, {'batch_size': 5, 'throughput': 1.2160675477315008, 'latency_mean': 4.09109667301178, 'latency_p50': 4.089922070503235, 'latency_p90': 4.634531760215759}, {'batch_size': 6, 'throughput': 1.2418883737532853, 'latency_mean': 4.812277507781983, 'latency_p50': 4.813434958457947, 'latency_p90': 5.37391402721405}, {'batch_size': 8, 'throughput': 1.2262985300335267, 'latency_mean': 6.481465903520584, 'latency_p50': 6.495221376419067, 'latency_p90': 7.335942244529724}, {'batch_size': 10, 'throughput': 1.1931900340390253, 'latency_mean': 8.323498827219009, 'latency_p50': 8.401122570037842, 'latency_p90': 9.465024781227111}]

max_input_tokens: 1024

max_output_tokens: 64

model_architecture: MistralForCausalLM

model_group: ChaiML/albert_dpo_0912_v

model_name: virgo_edit_variety_random

model_num_parameters: 12772070400.0

model_repo: ChaiML/albert_dpo_0912_virgo_edit_variety_random

model_size: 13B

num_battles: 11289

num_wins: 5789

propriety_score: 0.76162215628091

propriety_total_count: 1011.0

ranking_group: single

status: inactive

submission_type: basic

throughput_3p7s: 1.19

timestamp: 2024-09-13T19:39:06+00:00

us_pacific_date: 2024-09-13

win_ratio: 0.5128000708654442

Download Preference Data

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-albert-dpo-0912-v-5607-v1-mkmlizer
Waiting for job on chaiml-albert-dpo-0912-v-5607-v1-mkmlizer to finish
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║     _____            __           __                                ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║       /___/                                                         ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║                                                                     ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  Version: 0.10.1                                                    ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║                                                                     ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  belonging to:                                                      ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║                                                                     ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ║                                                                     ║
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
Failed to get response for submission chaiml-elo-alignment-run-3_v41: ('http://chaiml-elo-alignment-run-3-v41-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', 'read tcp 127.0.0.1:43654->127.0.0.1:8080: read: connection reset by peer\n')
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: Downloaded to shared memory in 93.776s
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmp1643vibd, device:0
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
Failed to get response for submission mistralai-mixtral-8x7b_3473_v136: ('http://mistralai-mixtral-8x7b-3473-v136-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', 'read tcp 127.0.0.1:37816->127.0.0.1:8080: read: connection reset by peer\n')
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: quantized model in 40.373s
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: Processed model ChaiML/albert_dpo_0912_virgo_edit_variety_random in 134.150s
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: creating bucket guanaco-mkml-models
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-albert-dpo-0912-v-5607-v1
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-albert-dpo-0912-v-5607-v1/config.json
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-albert-dpo-0912-v-5607-v1/special_tokens_map.json
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-albert-dpo-0912-v-5607-v1/tokenizer_config.json
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-albert-dpo-0912-v-5607-v1/tokenizer.json
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-albert-dpo-0912-v-5607-v1/flywheel_model.0.safetensors
chaiml-albert-dpo-0912-v-5607-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|▏         | 5/363 [00:00<00:15, 22.62it/s]
Loading 0:   3%|▎         | 10/363 [00:00<00:12, 28.40it/s]
Loading 0:   4%|▍         | 14/363 [00:00<00:13, 25.22it/s]
Loading 0:   6%|▌         | 21/363 [00:00<00:09, 36.61it/s]
Loading 0:   7%|▋         | 26/363 [00:01<00:15, 21.94it/s]
Loading 0:   9%|▊         | 31/363 [00:01<00:12, 26.59it/s]
Loading 0:  10%|▉         | 35/363 [00:01<00:11, 27.95it/s]
Loading 0:  11%|█         | 39/363 [00:01<00:10, 29.89it/s]
Loading 0:  12%|█▏        | 43/363 [00:01<00:11, 28.57it/s]
Loading 0:  13%|█▎        | 48/363 [00:01<00:09, 31.91it/s]
Loading 0:  14%|█▍        | 52/363 [00:01<00:10, 30.10it/s]
Loading 0:  15%|█▌        | 56/363 [00:01<00:09, 31.09it/s]
Loading 0:  17%|█▋        | 61/363 [00:02<00:11, 25.22it/s]
Loading 0:  18%|█▊        | 64/363 [00:02<00:13, 21.44it/s]
Loading 0:  19%|█▉        | 69/363 [00:02<00:11, 26.41it/s]
Loading 0:  20%|██        | 73/363 [00:02<00:11, 26.23it/s]
Loading 0:  21%|██        | 77/363 [00:02<00:11, 24.94it/s]
Loading 0:  23%|██▎       | 84/363 [00:03<00:08, 32.19it/s]
Loading 0:  24%|██▍       | 88/363 [00:03<00:08, 30.64it/s]
Loading 0:  26%|██▌       | 93/363 [00:03<00:08, 33.05it/s]
Loading 0:  27%|██▋       | 97/363 [00:03<00:08, 30.69it/s]
Loading 0:  28%|██▊       | 101/363 [00:03<00:10, 25.00it/s]
Loading 0:  29%|██▊       | 104/363 [00:03<00:11, 22.01it/s]
Loading 0:  31%|███       | 111/363 [00:03<00:08, 29.61it/s]
Loading 0:  32%|███▏      | 115/363 [00:04<00:08, 29.11it/s]
Loading 0:  33%|███▎      | 119/363 [00:04<00:07, 31.03it/s]
Loading 0:  34%|███▍      | 123/363 [00:04<00:08, 27.27it/s]
Loading 0:  36%|███▌      | 129/363 [00:04<00:07, 31.64it/s]
Loading 0:  37%|███▋      | 133/363 [00:04<00:07, 29.68it/s]
Loading 0:  38%|███▊      | 137/363 [00:04<00:07, 29.00it/s]
Loading 0:  39%|███▉      | 142/363 [00:05<00:08, 26.56it/s]
Loading 0:  40%|███▉      | 145/363 [00:05<00:08, 25.09it/s]
Loading 0:  41%|████      | 149/363 [00:05<00:09, 23.51it/s]
Loading 0:  42%|████▏     | 154/363 [00:05<00:07, 28.47it/s]
Loading 0:  44%|████▎     | 158/363 [00:05<00:07, 25.83it/s]
Loading 0:  45%|████▌     | 165/363 [00:05<00:05, 33.10it/s]
Loading 0:  47%|████▋     | 169/363 [00:06<00:06, 31.99it/s]
Loading 0:  48%|████▊     | 174/363 [00:06<00:05, 34.08it/s]
Loading 0:  49%|████▉     | 178/363 [00:06<00:05, 32.04it/s]
Loading 0:  50%|█████     | 182/363 [00:06<00:06, 26.34it/s]
Loading 0:  51%|█████     | 185/363 [00:06<00:07, 22.59it/s]
Loading 0:  53%|█████▎    | 192/363 [00:06<00:05, 28.91it/s]
Loading 0:  54%|█████▍    | 196/363 [00:06<00:05, 28.87it/s]
Loading 0:  55%|█████▌    | 201/363 [00:07<00:05, 31.52it/s]
Loading 0:  56%|█████▋    | 205/363 [00:07<00:05, 30.17it/s]
Loading 0:  58%|█████▊    | 210/363 [00:07<00:04, 32.62it/s]
Loading 0:  59%|█████▉    | 214/363 [00:07<00:04, 30.19it/s]
Loading 0:  60%|██████    | 218/363 [00:07<00:04, 30.26it/s]
Loading 0:  61%|██████▏   | 223/363 [00:07<00:05, 25.78it/s]
Loading 0:  62%|██████▏   | 226/363 [00:08<00:05, 24.52it/s]
Loading 0:  63%|██████▎   | 230/363 [00:08<00:05, 23.60it/s]
Loading 0:  65%|██████▌   | 237/363 [00:08<00:04, 30.03it/s]
Loading 0:  66%|██████▋   | 241/363 [00:08<00:04, 29.20it/s]
Loading 0:  68%|██████▊   | 246/363 [00:08<00:03, 31.97it/s]
Loading 0:  69%|██████▉   | 250/363 [00:08<00:03, 29.69it/s]
Loading 0:  70%|███████   | 255/363 [00:08<00:03, 32.22it/s]
Loading 0:  71%|███████▏  | 259/363 [00:09<00:03, 29.84it/s]
Loading 0:  72%|███████▏  | 263/363 [00:09<00:04, 23.85it/s]
Loading 0:  73%|███████▎  | 266/363 [00:09<00:04, 21.23it/s]
Loading 0:  75%|███████▌  | 273/363 [00:09<00:03, 28.09it/s]
Loading 0:  76%|███████▋  | 277/363 [00:09<00:03, 27.29it/s]
Loading 0:  78%|███████▊  | 282/363 [00:10<00:02, 29.79it/s]
Loading 0:  79%|███████▉  | 286/363 [00:10<00:02, 28.84it/s]
Loading 0:  80%|████████  | 291/363 [00:10<00:02, 30.99it/s]
Loading 0:  81%|████████▏ | 295/363 [00:10<00:02, 29.34it/s]
Loading 0:  82%|████████▏ | 299/363 [00:10<00:02, 30.01it/s]
Loading 0:  84%|████████▎ | 304/363 [00:10<00:02, 26.94it/s]
Loading 0:  85%|████████▍ | 307/363 [00:10<00:02, 24.00it/s]
Loading 0:  86%|████████▌ | 311/363 [00:11<00:02, 21.98it/s]
Loading 0:  87%|████████▋ | 316/363 [00:11<00:01, 26.92it/s]
Loading 0:  88%|████████▊ | 320/363 [00:11<00:01, 24.27it/s]
Loading 0:  90%|████████▉ | 325/363 [00:11<00:01, 28.63it/s]
Loading 0:  91%|█████████ | 329/363 [00:11<00:01, 25.22it/s]
Loading 0:  93%|█████████▎| 336/363 [00:11<00:00, 31.14it/s]
Loading 0:  94%|█████████▎| 340/363 [00:12<00:00, 29.01it/s]
Loading 0:  95%|█████████▍| 344/363 [00:19<00:09,  2.07it/s]
Loading 0:  96%|█████████▌| 348/363 [00:19<00:05,  2.75it/s]
Loading 0:  97%|█████████▋| 353/363 [00:19<00:02,  3.93it/s]
Loading 0:  98%|█████████▊| 357/363 [00:19<00:01,  5.07it/s]
                                                            
Job chaiml-albert-dpo-0912-v-5607-v1-mkmlizer completed after 159.02s with status: succeeded
Stopping job with name chaiml-albert-dpo-0912-v-5607-v1-mkmlizer
Pipeline stage MKMLizer completed in 159.74s
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 chaiml-albert-dpo-0912-v-5607-v1
Waiting for inference service chaiml-albert-dpo-0912-v-5607-v1 to be ready
Failed to get response for submission function_sabet_2024-09-10: ('http://zonemercy-lexical-nemov8-5966-v2-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', 'read tcp 127.0.0.1:49224->127.0.0.1:8080: read: connection reset by peer\n')
Inference service chaiml-albert-dpo-0912-v-5607-v1 ready after 191.06460046768188s
Pipeline stage MKMLDeployer completed in 191.86s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.3249309062957764s
Received healthy response to inference request in 1.8558084964752197s
Received healthy response to inference request in 2.239210605621338s
Received healthy response to inference request in 1.6545772552490234s
Received healthy response to inference request in 2.661566972732544s
5 requests
0 failed requests
5th percentile: 1.6948235034942627
10th percentile: 1.735069751739502
20th percentile: 1.8155622482299805
30th percentile: 1.9324889183044434
40th percentile: 2.0858497619628906
50th percentile: 2.239210605621338
60th percentile: 2.2734987258911135
70th percentile: 2.3077868461608886
80th percentile: 2.39225811958313
90th percentile: 2.526912546157837
95th percentile: 2.5942397594451903
99th percentile: 2.648101530075073
mean time: 2.14721884727478
Pipeline stage StressChecker completed in 11.98s
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 4.82s
Shutdown handler de-registered
chaiml-albert-dpo-0912-v_5607_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 0.11s
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 chaiml-albert-dpo-0912-v-5607-v1-profiler
Waiting for inference service chaiml-albert-dpo-0912-v-5607-v1-profiler to be ready
Inference service chaiml-albert-dpo-0912-v-5607-v1-profiler ready after 190.44184374809265s
Pipeline stage MKMLProfilerDeployer completed in 190.82s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/chaiml-albert-dpo-097715f434145430445240b921ec31cf85-deplohfdtl:/code/chaiverse_profiler_1726256955 --namespace tenant-chaiml-guanaco
kubectl exec -it chaiml-albert-dpo-097715f434145430445240b921ec31cf85-deplohfdtl --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1726256955 && 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 1024 --output_tokens 64 --summary /code/chaiverse_profiler_1726256955/summary.json'
kubectl exec -it chaiml-albert-dpo-097715f434145430445240b921ec31cf85-deplohfdtl --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1726256955/summary.json'
Pipeline stage MKMLProfilerRunner completed in 1180.90s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service chaiml-albert-dpo-0912-v-5607-v1-profiler is running
Tearing down inference service chaiml-albert-dpo-0912-v-5607-v1-profiler
Service chaiml-albert-dpo-0912-v-5607-v1-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 2.22s
Shutdown handler de-registered
chaiml-albert-dpo-0912-v_5607_v1 status is now inactive due to auto deactivation removed underperforming models