submission_id: jic062-dpo-v1-4-nemo_v2
developer_uid: chace9580
best_of: 8
celo_rating: 1266.93
display_name: jic062-dpo-v1-4-nemo_v2
family_friendly_score: 0.0
formatter: {'memory_template': '[INST]system\n{memory}[/INST]\n', 'prompt_template': '[INST]user\n{prompt}[/INST]\n', 'bot_template': '[INST]assistant\n{bot_name}: {message}[/INST]\n', 'user_template': '[INST]user\n{user_name}: {message}[/INST]\n', 'response_template': '[INST]assistant\n{bot_name}:', 'truncate_by_message': False}
generation_params: {'temperature': 0.85, 'top_p': 1.0, 'min_p': 0.05, 'top_k': 80, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '[/INST]'], 'max_input_tokens': 512, 'best_of': 8, 'max_output_tokens': 64}
gpu_counts: {'NVIDIA RTX A5000': 1}
is_internal_developer: False
language_model: jic062/dpo-v1.4-Nemo
latencies: [{'batch_size': 1, 'throughput': 0.7078953703961188, 'latency_mean': 1.412560646533966, 'latency_p50': 1.4062175750732422, 'latency_p90': 1.5714553356170655}, {'batch_size': 3, 'throughput': 1.3524105028822402, 'latency_mean': 2.2149657344818117, 'latency_p50': 2.207529306411743, 'latency_p90': 2.4440405130386353}, {'batch_size': 5, 'throughput': 1.5819879772347534, 'latency_mean': 3.1412793469429015, 'latency_p50': 3.1524884700775146, 'latency_p90': 3.4978840827941893}, {'batch_size': 6, 'throughput': 1.6420374304116931, 'latency_mean': 3.627544691562653, 'latency_p50': 3.610406756401062, 'latency_p90': 4.091594624519348}, {'batch_size': 8, 'throughput': 1.5979113330502674, 'latency_mean': 4.972033543586731, 'latency_p50': 5.028353691101074, 'latency_p90': 5.651856446266175}, {'batch_size': 10, 'throughput': 1.5655440152020699, 'latency_mean': 6.345912663936615, 'latency_p50': 6.324586987495422, 'latency_p90': 7.236740899085999}]
max_input_tokens: 512
max_output_tokens: 64
model_architecture: MistralForCausalLM
model_group: jic062/dpo-v1.4-Nemo
model_name: jic062-dpo-v1-4-nemo_v2
model_num_parameters: 12772070400.0
model_repo: jic062/dpo-v1.4-Nemo
model_size: 13B
num_battles: 10176
num_wins: 5338
ranking_group: single
status: torndown
submission_type: basic
throughput_3p7s: 1.65
timestamp: 2024-09-15T04:47:03+00:00
us_pacific_date: 2024-09-14
win_ratio: 0.5245676100628931
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 jic062-dpo-v1-4-nemo-v2-mkmlizer
Waiting for job on jic062-dpo-v1-4-nemo-v2-mkmlizer to finish
jic062-dpo-v1-4-nemo-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ _____ __ __ ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ /___/ ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ Version: 0.10.1 ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ https://mk1.ai ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ The license key for the current software has been verified as ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ belonging to: ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ Chai Research Corp. ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ Expiration: 2024-10-15 23:59:59 ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ║ ║
jic062-dpo-v1-4-nemo-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
jic062-dpo-v1-4-nemo-v2-mkmlizer: Downloaded to shared memory in 44.366s
jic062-dpo-v1-4-nemo-v2-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmp9rnwydhs, device:0
jic062-dpo-v1-4-nemo-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
jic062-dpo-v1-4-nemo-v2-mkmlizer: quantized model in 35.105s
jic062-dpo-v1-4-nemo-v2-mkmlizer: Processed model jic062/dpo-v1.4-Nemo in 79.472s
jic062-dpo-v1-4-nemo-v2-mkmlizer: creating bucket guanaco-mkml-models
jic062-dpo-v1-4-nemo-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
jic062-dpo-v1-4-nemo-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/jic062-dpo-v1-4-nemo-v2
jic062-dpo-v1-4-nemo-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/jic062-dpo-v1-4-nemo-v2/config.json
jic062-dpo-v1-4-nemo-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/jic062-dpo-v1-4-nemo-v2/special_tokens_map.json
jic062-dpo-v1-4-nemo-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/jic062-dpo-v1-4-nemo-v2/tokenizer_config.json
jic062-dpo-v1-4-nemo-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/jic062-dpo-v1-4-nemo-v2/tokenizer.json
jic062-dpo-v1-4-nemo-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/jic062-dpo-v1-4-nemo-v2/flywheel_model.0.safetensors
jic062-dpo-v1-4-nemo-v2-mkmlizer: Loading 0: 0%| | 0/363 [00:00<?, ?it/s] Loading 0: 1%|▏ | 5/363 [00:00<00:10, 34.51it/s] Loading 0: 4%|▎ | 13/363 [00:00<00:06, 55.08it/s] Loading 0: 5%|▌ | 19/363 [00:00<00:06, 51.25it/s] Loading 0: 7%|▋ | 25/363 [00:00<00:06, 49.87it/s] Loading 0: 9%|▊ | 31/363 [00:00<00:06, 52.19it/s] Loading 0: 10%|█ | 37/363 [00:00<00:06, 48.94it/s] Loading 0: 12%|█▏ | 42/363 [00:00<00:07, 45.19it/s] Loading 0: 13%|█▎ | 49/363 [00:00<00:06, 50.12it/s] Loading 0: 15%|█▌ | 55/363 [00:01<00:06, 48.34it/s] Loading 0: 17%|█▋ | 60/363 [00:01<00:06, 48.48it/s] Loading 0: 18%|█▊ | 65/363 [00:01<00:09, 32.41it/s] Loading 0: 20%|█▉ | 72/363 [00:01<00:07, 39.99it/s] Loading 0: 21%|██▏ | 78/363 [00:01<00:06, 41.87it/s] Loading 0: 23%|██▎ | 83/363 [00:01<00:06, 42.84it/s] Loading 0: 25%|██▍ | 90/363 [00:01<00:05, 48.13it/s] Loading 0: 26%|██▋ | 96/363 [00:02<00:05, 47.39it/s] Loading 0: 28%|██▊ | 102/363 [00:02<00:05, 48.32it/s] Loading 0: 30%|███ | 110/363 [00:02<00:05, 50.47it/s] Loading 0: 32%|███▏ | 116/363 [00:02<00:05, 48.03it/s] Loading 0: 33%|███▎ | 121/363 [00:02<00:05, 46.77it/s] Loading 0: 35%|███▍ | 127/363 [00:02<00:05, 40.05it/s] Loading 0: 37%|███▋ | 134/363 [00:02<00:04, 46.43it/s] Loading 0: 38%|███▊ | 139/363 [00:03<00:05, 44.03it/s] Loading 0: 40%|███▉ | 144/363 [00:03<00:07, 29.04it/s] Loading 0: 41%|████ | 149/363 [00:03<00:06, 31.63it/s] Loading 0: 43%|████▎ | 157/363 [00:03<00:05, 39.85it/s] Loading 0: 45%|████▍ | 163/363 [00:03<00:04, 40.93it/s] Loading 0: 46%|████▋ | 168/363 [00:03<00:04, 41.50it/s] Loading 0: 48%|████▊ | 175/363 [00:03<00:04, 46.35it/s] Loading 0: 50%|████▉ | 181/363 [00:04<00:03, 45.59it/s] Loading 0: 51%|█████ | 186/363 [00:04<00:03, 44.50it/s] Loading 0: 53%|█████▎ | 193/363 [00:04<00:03, 49.03it/s] Loading 0: 55%|█████▍ | 199/363 [00:04<00:03, 46.59it/s] Loading 0: 56%|█████▌ | 204/363 [00:04<00:03, 42.23it/s] Loading 0: 58%|█████▊ | 211/363 [00:04<00:03, 46.48it/s] Loading 0: 60%|█████▉ | 217/363 [00:04<00:03, 42.96it/s] Loading 0: 61%|██████▏ | 223/363 [00:05<00:04, 34.08it/s] Loading 0: 63%|██████▎ | 227/363 [00:05<00:03, 34.69it/s] Loading 0: 64%|██████▎ | 231/363 [00:05<00:03, 33.43it/s] Loading 0: 65%|██████▌ | 237/363 [00:05<00:03, 37.66it/s] Loading 0: 66%|██████▋ | 241/363 [00:05<00:03, 36.98it/s] Loading 0: 68%|██████▊ | 247/363 [00:05<00:02, 41.20it/s] Loading 0: 70%|██████▉ | 253/363 [00:05<00:02, 41.36it/s] Loading 0: 71%|███████ | 258/363 [00:06<00:02, 42.17it/s] Loading 0: 73%|███████▎ | 265/363 [00:06<00:02, 47.77it/s] Loading 0: 75%|███████▍ | 271/363 [00:06<00:01, 47.35it/s] Loading 0: 76%|███████▌ | 276/363 [00:06<00:01, 45.35it/s] Loading 0: 78%|███████▊ | 283/363 [00:06<00:01, 50.10it/s] Loading 0: 80%|███████▉ | 289/363 [00:06<00:01, 47.75it/s] Loading 0: 81%|████████ | 294/363 [00:06<00:01, 43.95it/s] Loading 0: 83%|████████▎ | 302/363 [00:06<00:01, 52.36it/s] Loading 0: 85%|████████▍ | 308/363 [00:13<00:18, 2.94it/s] Loading 0: 86%|████████▌ | 312/363 [00:13<00:13, 3.64it/s] Loading 0: 88%|████████▊ | 319/363 [00:13<00:08, 5.44it/s] Loading 0: 89%|████████▉ | 324/363 [00:14<00:05, 7.10it/s] Loading 0: 91%|█████████ | 330/363 [00:14<00:03, 9.52it/s] Loading 0: 93%|█████████▎| 339/363 [00:14<00:01, 14.19it/s] Loading 0: 96%|█████████▌| 347/363 [00:14<00:00, 19.52it/s] Loading 0: 97%|█████████▋| 353/363 [00:14<00:00, 23.09it/s] Loading 0: 99%|█████████▉| 359/363 [00:14<00:00, 26.98it/s]
Job jic062-dpo-v1-4-nemo-v2-mkmlizer completed after 108.39s with status: succeeded
Stopping job with name jic062-dpo-v1-4-nemo-v2-mkmlizer
Pipeline stage MKMLizer completed in 110.76s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.08s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service jic062-dpo-v1-4-nemo-v2
Waiting for inference service jic062-dpo-v1-4-nemo-v2 to be ready
Inference service jic062-dpo-v1-4-nemo-v2 ready after 171.97923851013184s
Pipeline stage MKMLDeployer completed in 174.24s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.349841594696045s
Received healthy response to inference request in 1.7979888916015625s
Received healthy response to inference request in 2.2267470359802246s
Received healthy response to inference request in 2.408895254135132s
Received healthy response to inference request in 2.084644079208374s
5 requests
0 failed requests
5th percentile: 1.8553199291229248
10th percentile: 1.912650966644287
20th percentile: 2.0273130416870115
30th percentile: 2.113064670562744
40th percentile: 2.1699058532714846
50th percentile: 2.2267470359802246
60th percentile: 2.275984859466553
70th percentile: 2.325222682952881
80th percentile: 2.3616523265838625
90th percentile: 2.385273790359497
95th percentile: 2.3970845222473143
99th percentile: 2.4065331077575682
mean time: 2.1736233711242674
Pipeline stage StressChecker completed in 12.19s
run pipeline stage %s
Running pipeline stage TriggerMKMLProfilingPipeline
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
run_pipeline:run_in_cloud %s
starting trigger_guanaco_pipeline args=%s
Pipeline stage TriggerMKMLProfilingPipeline completed in 8.23s
Shutdown handler de-registered
jic062-dpo-v1-4-nemo_v2 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 jic062-dpo-v1-4-nemo-v2-profiler
Waiting for inference service jic062-dpo-v1-4-nemo-v2-profiler to be ready
Inference service jic062-dpo-v1-4-nemo-v2-profiler ready after 170.41287231445312s
Pipeline stage MKMLProfilerDeployer completed in 171.40s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/jic062-dpo-v1-4-nemo-v2-profiler-predictor-00001-deploymenszhth:/code/chaiverse_profiler_1726376157 --namespace tenant-chaiml-guanaco
kubectl exec -it jic062-dpo-v1-4-nemo-v2-profiler-predictor-00001-deploymenszhth --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1726376157 && 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_1726376157/summary.json'
kubectl exec -it jic062-dpo-v1-4-nemo-v2-profiler-predictor-00001-deploymenszhth --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1726376157/summary.json'
Pipeline stage MKMLProfilerRunner completed in 942.11s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service jic062-dpo-v1-4-nemo-v2-profiler is running
Tearing down inference service jic062-dpo-v1-4-nemo-v2-profiler
Service jic062-dpo-v1-4-nemo-v2-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 11.60s
Shutdown handler de-registered
jic062-dpo-v1-4-nemo_v2 status is now inactive due to auto deactivation removed underperforming models
jic062-dpo-v1-4-nemo_v2 status is now torndown due to DeploymentManager action