chaiml-elo-alignment-run-3

developer_uid: azuruce

submission_id: chaiml-elo-alignment-run-3_v58

model_name: chaiml-elo-alignment-run-3_v48

model_group: ChaiML/elo_alignment_run

status: torndown

timestamp: 2024-11-13T22:17:34+00:00

num_battles: 14108

num_wins: 7552

celo_rating: 1270.15

family_friendly_score: 0.5774

family_friendly_standard_error: 0.006985831947592212

submission_type: basic

model_repo: ChaiML/elo_alignment_run_3

model_architecture: LlamaForCausalLM

model_num_parameters: 8030261248.0

best_of: 16

max_input_tokens: 512

max_output_tokens: 64

latencies: [{'batch_size': 1, 'throughput': 0.90900212063558, 'latency_mean': 1.1000472152233123, 'latency_p50': 1.1030118465423584, 'latency_p90': 1.2330410480499268}, {'batch_size': 4, 'throughput': 1.792883237926856, 'latency_mean': 2.2173882937431335, 'latency_p50': 2.2211581468582153, 'latency_p90': 2.4940130949020385}, {'batch_size': 5, 'throughput': 1.888115546385019, 'latency_mean': 2.6428341591358184, 'latency_p50': 2.6168330907821655, 'latency_p90': 2.9533478736877443}, {'batch_size': 8, 'throughput': 2.02060440766468, 'latency_mean': 3.927933987379074, 'latency_p50': 3.921592354774475, 'latency_p90': 4.401239776611328}, {'batch_size': 10, 'throughput': 2.027029131586931, 'latency_mean': 4.884392178058624, 'latency_p50': 4.87498414516449, 'latency_p90': 5.561880302429199}, {'batch_size': 12, 'throughput': 2.0450993241865705, 'latency_mean': 5.797462284564972, 'latency_p50': 5.797064304351807, 'latency_p90': 6.802266597747803}, {'batch_size': 15, 'throughput': 2.042674335085008, 'latency_mean': 7.192611156702042, 'latency_p50': 7.324605464935303, 'latency_p90': 7.938853120803833}]

gpu_counts: {'NVIDIA RTX A5000': 1}

display_name: chaiml-elo-alignment-run-3_v48

is_internal_developer: True

language_model: ChaiML/elo_alignment_run_3

model_size: 8B

ranking_group: single

throughput_3p7s: 2.02

us_pacific_date: 2024-11-13

win_ratio: 0.5352991210660618

generation_params: {'temperature': 0.95, 'top_p': 1.0, 'min_p': 0.08, 'top_k': 40, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '<|eot_id|>'], '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 chaiml-elo-alignment-run-3-v58-mkmlizer
Waiting for job on chaiml-elo-alignment-run-3-v58-mkmlizer to finish
Connection pool is full, discarding connection: %s. Connection pool size: %s
chaiml-elo-alignment-run-3-v58-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-elo-alignment-run-3-v58-mkmlizer: ║     _____            __           __                                ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║       /___/                                                         ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║                                                                     ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  Version: 0.11.12                                                   ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║                                                                     ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  belonging to:                                                      ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║                                                                     ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║  Expiration: 2025-01-15 23:59:59                                    ║
chaiml-elo-alignment-run-3-v58-mkmlizer: ║                                                                     ║
chaiml-elo-alignment-run-3-v58-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-elo-alignment-run-3-v58-mkmlizer: Downloaded to shared memory in 42.791s
chaiml-elo-alignment-run-3-v58-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmp8asxvzp4, device:0
chaiml-elo-alignment-run-3-v58-mkmlizer: Saving flywheel model at /dev/shm/model_cache
Failed to get response for submission bbchicago-nana-nemo-12b-v1-0_v8: ('http://bbchicago-nana-nemo-12b-v1-0-v8-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', '')
chaiml-elo-alignment-run-3-v58-mkmlizer: quantized model in 30.087s
chaiml-elo-alignment-run-3-v58-mkmlizer: Processed model ChaiML/elo_alignment_run_3 in 72.878s
chaiml-elo-alignment-run-3-v58-mkmlizer: creating bucket guanaco-mkml-models
chaiml-elo-alignment-run-3-v58-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-elo-alignment-run-3-v58-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-elo-alignment-run-3-v58
chaiml-elo-alignment-run-3-v58-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-elo-alignment-run-3-v58/config.json
chaiml-elo-alignment-run-3-v58-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-elo-alignment-run-3-v58/special_tokens_map.json
chaiml-elo-alignment-run-3-v58-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-elo-alignment-run-3-v58/tokenizer_config.json
chaiml-elo-alignment-run-3-v58-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-elo-alignment-run-3-v58/tokenizer.json
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
chaiml-elo-alignment-run-3-v58-mkmlizer: 
Loading 0:   0%|          | 0/291 [00:00<?, ?it/s]
Loading 0:   2%|▏         | 5/291 [00:00<00:12, 23.75it/s]
Loading 0:   3%|▎         | 10/291 [00:00<00:08, 33.85it/s]
Loading 0:   5%|▍         | 14/291 [00:00<00:10, 27.63it/s]
Loading 0:   7%|▋         | 19/291 [00:00<00:08, 33.33it/s]
Loading 0:   8%|▊         | 23/291 [00:00<00:09, 29.17it/s]
Loading 0:  11%|█         | 31/291 [00:00<00:06, 41.50it/s]
Loading 0:  12%|█▏        | 36/291 [00:01<00:10, 25.18it/s]
Loading 0:  14%|█▍        | 41/291 [00:01<00:09, 27.41it/s]
Loading 0:  17%|█▋        | 49/291 [00:01<00:06, 35.32it/s]
Loading 0:  19%|█▊        | 54/291 [00:01<00:06, 38.27it/s]
Loading 0:  20%|██        | 59/291 [00:01<00:06, 33.89it/s]
Loading 0:  23%|██▎       | 66/291 [00:01<00:05, 41.19it/s]
Loading 0:  24%|██▍       | 71/291 [00:02<00:05, 40.20it/s]
Loading 0:  26%|██▌       | 76/291 [00:02<00:06, 34.26it/s]
Loading 0:  27%|██▋       | 80/291 [00:02<00:07, 27.38it/s]
Loading 0:  29%|██▉       | 84/291 [00:02<00:07, 26.28it/s]
Loading 0:  31%|███       | 90/291 [00:02<00:06, 31.93it/s]
Loading 0:  32%|███▏      | 94/291 [00:02<00:06, 30.95it/s]
Loading 0:  34%|███▍      | 99/291 [00:03<00:05, 33.27it/s]
Loading 0:  35%|███▌      | 103/291 [00:03<00:05, 32.03it/s]
Loading 0:  37%|███▋      | 108/291 [00:03<00:05, 34.89it/s]
Loading 0:  38%|███▊      | 112/291 [00:03<00:05, 32.50it/s]
Loading 0:  40%|███▉      | 116/291 [00:03<00:05, 31.71it/s]
Loading 0:  42%|████▏     | 122/291 [00:03<00:04, 35.46it/s]
Loading 0:  44%|████▎     | 127/291 [00:03<00:04, 33.28it/s]
Loading 0:  46%|████▌     | 133/291 [00:04<00:05, 28.88it/s]
Loading 0:  47%|████▋     | 137/291 [00:04<00:05, 28.80it/s]
Loading 0:  48%|████▊     | 141/291 [00:04<00:05, 26.89it/s]
Loading 0:  51%|█████     | 147/291 [00:04<00:04, 30.97it/s]
Loading 0:  52%|█████▏    | 151/291 [00:04<00:04, 29.62it/s]
Loading 0:  54%|█████▎    | 156/291 [00:04<00:04, 31.53it/s]
Loading 0:  55%|█████▍    | 160/291 [00:05<00:04, 30.00it/s]
Loading 0:  56%|█████▋    | 164/291 [00:05<00:04, 31.20it/s]
Loading 0:  58%|█████▊    | 168/291 [00:05<00:04, 27.32it/s]
Loading 0:  59%|█████▉    | 172/291 [00:05<00:03, 29.80it/s]
Loading 0:  60%|██████    | 176/291 [00:05<00:04, 27.01it/s]
Loading 0:  62%|██████▏   | 181/291 [00:05<00:03, 31.89it/s]
Loading 0:  64%|██████▎   | 185/291 [00:05<00:03, 32.74it/s]
Loading 0:  65%|██████▍   | 189/291 [00:06<00:04, 20.59it/s]
Loading 0:  67%|██████▋   | 194/291 [00:06<00:04, 21.72it/s]
Loading 0:  68%|██████▊   | 199/291 [00:06<00:03, 26.21it/s]
Loading 0:  70%|██████▉   | 203/291 [00:06<00:03, 24.98it/s]
Loading 0:  71%|███████▏  | 208/291 [00:06<00:02, 29.37it/s]
Loading 0:  73%|███████▎  | 212/291 [00:07<00:03, 25.42it/s]
Loading 0:  75%|███████▍  | 217/291 [00:07<00:02, 29.61it/s]
Loading 0:  76%|███████▌  | 221/291 [00:07<00:02, 26.97it/s]
Loading 0:  78%|███████▊  | 226/291 [00:07<00:02, 31.04it/s]
Loading 0:  79%|███████▉  | 230/291 [00:07<00:02, 28.05it/s]
Loading 0:  80%|████████  | 234/291 [00:07<00:02, 20.56it/s]
Loading 0:  82%|████████▏ | 239/291 [00:08<00:02, 22.51it/s]
Loading 0:  84%|████████▍ | 244/291 [00:08<00:01, 26.90it/s]
Loading 0:  85%|████████▌ | 248/291 [00:08<00:01, 24.66it/s]
Loading 0:  87%|████████▋ | 253/291 [00:08<00:01, 29.12it/s]
Loading 0:  88%|████████▊ | 257/291 [00:08<00:01, 26.40it/s]
Loading 0:  90%|█████████ | 262/291 [00:08<00:00, 30.31it/s]
Loading 0:  91%|█████████▏| 266/291 [00:09<00:00, 26.56it/s]
Loading 0:  93%|█████████▎| 271/291 [00:09<00:00, 30.76it/s]
Loading 0:  95%|█████████▍| 275/291 [00:09<00:00, 27.49it/s]
Loading 0:  96%|█████████▌| 280/291 [00:09<00:00, 31.41it/s]
Loading 0:  98%|█████████▊| 285/291 [00:09<00:00, 32.12it/s]
Loading 0:  99%|█████████▉| 289/291 [00:15<00:00,  2.48it/s]
                                                            
Job chaiml-elo-alignment-run-3-v58-mkmlizer completed after 104.51s with status: succeeded
Stopping job with name chaiml-elo-alignment-run-3-v58-mkmlizer
Pipeline stage MKMLizer completed in 105.11s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.18s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service chaiml-elo-alignment-run-3-v58
Waiting for inference service chaiml-elo-alignment-run-3-v58 to be ready
Failed to get response for submission jic062-dpo-v3-0-nemo_v1: ('http://jic062-dpo-v3-0-nemo-v1-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', '')
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 chaiml-elo-alignment-run-3-v58 ready after 200.6955246925354s
Pipeline stage MKMLDeployer completed in 201.31s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 1.710533857345581s
Received healthy response to inference request in 1.3629670143127441s
Received healthy response to inference request in 1.8699884414672852s
Received healthy response to inference request in 1.5013437271118164s
Received healthy response to inference request in 1.5602471828460693s
5 requests
0 failed requests
5th percentile: 1.3906423568725585
10th percentile: 1.418317699432373
20th percentile: 1.473668384552002
30th percentile: 1.513124418258667
40th percentile: 1.5366858005523683
50th percentile: 1.5602471828460693
60th percentile: 1.620361852645874
70th percentile: 1.6804765224456786
80th percentile: 1.742424774169922
90th percentile: 1.8062066078186034
95th percentile: 1.8380975246429443
99th percentile: 1.863610258102417
mean time: 1.6010160446166992
Pipeline stage StressChecker completed in 9.42s
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 5.27s
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 2.75s
Shutdown handler de-registered
chaiml-elo-alignment-run-3_v58 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 0.12s
run pipeline stage %s
Running pipeline stage MKMLProfilerTemplater
Pipeline stage MKMLProfilerTemplater completed in 0.12s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service chaiml-elo-alignment-run-3-v58-profiler
Waiting for inference service chaiml-elo-alignment-run-3-v58-profiler to be ready
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 2320.09s
Shutdown handler de-registered
chaiml-elo-alignment-run-3_v58 status is now inactive due to auto deactivation removed underperforming models
chaiml-elo-alignment-run-3_v58 status is now torndown due to DeploymentManager action