developer_uid: Trace2333
submission_id: trace2333-dpo-v9-v3-reprompt_v2
model_name: trace2333-dpo-v9-v3-reprompt_v2
model_group: Trace2333/dpo_v9_v3_repr
status: torndown
timestamp: 2024-08-27T10:29:13+00:00
num_battles: 10661
num_wins: 5658
celo_rating: 1257.68
family_friendly_score: 0.0
submission_type: basic
model_repo: Trace2333/dpo_v9_v3_reprompt
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 16
max_input_tokens: 512
max_output_tokens: 64
latencies: [{'batch_size': 1, 'throughput': 0.903455806818427, 'latency_mean': 1.1067672836780549, 'latency_p50': 1.1052838563919067, 'latency_p90': 1.2535680532455444}, {'batch_size': 4, 'throughput': 1.8063768262595818, 'latency_mean': 2.2075412476062777, 'latency_p50': 2.1984636783599854, 'latency_p90': 2.4651920080184935}, {'batch_size': 5, 'throughput': 1.8725036085326252, 'latency_mean': 2.65462562084198, 'latency_p50': 2.6576262712478638, 'latency_p90': 2.950652241706848}, {'batch_size': 8, 'throughput': 1.9974588202919354, 'latency_mean': 3.9816395580768584, 'latency_p50': 4.026030659675598, 'latency_p90': 4.442818689346313}, {'batch_size': 10, 'throughput': 2.0264633745877814, 'latency_mean': 4.884057245254517, 'latency_p50': 4.887906908988953, 'latency_p90': 5.629179430007935}, {'batch_size': 12, 'throughput': 2.034729830833767, 'latency_mean': 5.822340323925018, 'latency_p50': 5.863048195838928, 'latency_p90': 6.708717727661132}, {'batch_size': 15, 'throughput': 2.0048204550513162, 'latency_mean': 7.344305253028869, 'latency_p50': 7.443024158477783, 'latency_p90': 8.149785733222961}]
gpu_counts: {'NVIDIA RTX A5000': 1}
display_name: trace2333-dpo-v9-v3-reprompt_v2
is_internal_developer: False
language_model: Trace2333/dpo_v9_v3_reprompt
model_size: 8B
ranking_group: single
throughput_3p7s: 1.99
us_pacific_date: 2024-08-27
win_ratio: 0.5307194447049995
generation_params: {'temperature': 1.0, 'top_p': 1.0, 'min_p': 0.0, '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
Running pipeline stage MKMLizer
Starting job with name trace2333-dpo-v9-v3-reprompt-v2-mkmlizer
Waiting for job on trace2333-dpo-v9-v3-reprompt-v2-mkmlizer to finish
Stopping job with name trace2333-dpo-v9-v3-reprompt-v2-mkmlizer
%s, retrying in %s seconds...
Starting job with name trace2333-dpo-v9-v3-reprompt-v2-mkmlizer
Waiting for job on trace2333-dpo-v9-v3-reprompt-v2-mkmlizer to finish
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║     _____            __           __                                ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║       /___/                                                         ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║                                                                     ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  Version: 0.10.1                                                    ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  https://mk1.ai                                                     ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║                                                                     ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  The license key for the current software has been verified as      ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  belonging to:                                                      ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║                                                                     ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  Chai Research Corp.                                                ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: ║                                                                     ║
trace2333-dpo-v9-v3-reprompt-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
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
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: Downloaded to shared memory in 62.038s
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpujmonk8g, device:0
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: quantized model in 28.969s
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: Processed model Trace2333/dpo_v9_v3_reprompt in 91.007s
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: creating bucket guanaco-mkml-models
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/trace2333-dpo-v9-v3-reprompt-v2
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/trace2333-dpo-v9-v3-reprompt-v2/special_tokens_map.json
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/trace2333-dpo-v9-v3-reprompt-v2/config.json
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/trace2333-dpo-v9-v3-reprompt-v2/tokenizer_config.json
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/trace2333-dpo-v9-v3-reprompt-v2/tokenizer.json
trace2333-dpo-v9-v3-reprompt-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/trace2333-dpo-v9-v3-reprompt-v2/flywheel_model.0.safetensors
Job trace2333-dpo-v9-v3-reprompt-v2-mkmlizer completed after 116.0s with status: succeeded
Stopping job with name trace2333-dpo-v9-v3-reprompt-v2-mkmlizer
Pipeline stage MKMLizer completed in 117.83s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.08s
Running pipeline stage ISVCDeployer
Creating inference service trace2333-dpo-v9-v3-reprompt-v2
Waiting for inference service trace2333-dpo-v9-v3-reprompt-v2 to be ready
Inference service trace2333-dpo-v9-v3-reprompt-v2 ready after 171.12178468704224s
Pipeline stage ISVCDeployer completed in 171.96s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.2349393367767334s
Received healthy response to inference request in 2.2225217819213867s
Received healthy response to inference request in 1.5442979335784912s
Received healthy response to inference request in 1.9191930294036865s
Received healthy response to inference request in 2.0947954654693604s
5 requests
0 failed requests
5th percentile: 1.6192769527435302
10th percentile: 1.6942559719085692
20th percentile: 1.8442140102386475
30th percentile: 1.9543135166168213
40th percentile: 2.024554491043091
50th percentile: 2.0947954654693604
60th percentile: 2.1458859920501707
70th percentile: 2.1969765186309815
80th percentile: 2.225005292892456
90th percentile: 2.229972314834595
95th percentile: 2.232455825805664
99th percentile: 2.2344426345825195
mean time: 2.0031495094299316
Pipeline stage StressChecker completed in 10.68s
trace2333-dpo-v9-v3-reprompt_v2 status is now deployed due to DeploymentManager action
trace2333-dpo-v9-v3-reprompt_v2 status is now inactive due to auto deactivation removed underperforming models
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 0.13s
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 trace2333-dpo-v9-v3-reprompt-v2-profiler
Waiting for inference service trace2333-dpo-v9-v3-reprompt-v2-profiler to be ready
Inference service trace2333-dpo-v9-v3-reprompt-v2-profiler ready after 150.35731863975525s
Pipeline stage MKMLProfilerDeployer completed in 150.75s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/trace2333-dpo-v9-v3-9f9272070a0980049b0fe879b2a66000-deploddkdd:/code/chaiverse_profiler_1725502288 --namespace tenant-chaiml-guanaco
kubectl exec -it trace2333-dpo-v9-v3-9f9272070a0980049b0fe879b2a66000-deploddkdd --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1725502288 && 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_1725502288/summary.json'
kubectl exec -it trace2333-dpo-v9-v3-9f9272070a0980049b0fe879b2a66000-deploddkdd --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1725502288/summary.json'
Pipeline stage MKMLProfilerRunner completed in 842.28s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service trace2333-dpo-v9-v3-reprompt-v2-profiler is running
Tearing down inference service trace2333-dpo-v9-v3-reprompt-v2-profiler
Service trace2333-dpo-v9-v3-reprompt-v2-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 1.74s
Shutdown handler de-registered
trace2333-dpo-v9-v3-reprompt_v2 status is now torndown due to DeploymentManager action