submission_id: jic062-dpo-v1-3-nemo-c500_v1
developer_uid: chace9580
best_of: 8
celo_rating: 1243.04
display_name: jic062-dpo-v1-3-nemo-c500_v1
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': 1.0, 'top_p': 1.0, 'min_p': 0.1, 'top_k': 40, '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-v1.3-Nemo-c500
latencies: [{'batch_size': 1, 'throughput': 0.7087362180229027, 'latency_mean': 1.4108830666542054, 'latency_p50': 1.4051575660705566, 'latency_p90': 1.5753527402877807}, {'batch_size': 3, 'throughput': 1.3462029502248793, 'latency_mean': 2.2157720041275026, 'latency_p50': 2.2023099660873413, 'latency_p90': 2.483182692527771}, {'batch_size': 5, 'throughput': 1.5772845519043235, 'latency_mean': 3.1503529024124144, 'latency_p50': 3.152476668357849, 'latency_p90': 3.575472331047058}, {'batch_size': 6, 'throughput': 1.6428810993836633, 'latency_mean': 3.6281919467449186, 'latency_p50': 3.6358100175857544, 'latency_p90': 4.078349757194519}, {'batch_size': 8, 'throughput': 1.6144621148107492, 'latency_mean': 4.916532549858093, 'latency_p50': 4.965585350990295, 'latency_p90': 5.536692810058594}, {'batch_size': 10, 'throughput': 1.577145964106861, 'latency_mean': 6.298866722583771, 'latency_p50': 6.323810458183289, 'latency_p90': 7.062532877922058}]
max_input_tokens: 512
max_output_tokens: 64
model_architecture: MistralForCausalLM
model_group: jic062/dpo-v1.3-Nemo-c50
model_name: jic062-dpo-v1-3-nemo-c500_v1
model_num_parameters: 12772070400.0
model_repo: jic062/dpo-v1.3-Nemo-c500
model_size: 13B
num_battles: 10497
num_wins: 5233
ranking_group: single
status: torndown
submission_type: basic
throughput_3p7s: 1.65
timestamp: 2024-09-12T04:23:33+00:00
us_pacific_date: 2024-09-11
win_ratio: 0.4985233876345623
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-v1-3-nemo-c500-v1-mkmlizer
Waiting for job on jic062-dpo-v1-3-nemo-c500-v1-mkmlizer to finish
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ _____ __ __ ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ /___/ ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ Version: 0.10.1 ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ https://mk1.ai ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ The license key for the current software has been verified as ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ belonging to: ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ Chai Research Corp. ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ Expiration: 2024-10-15 23:59:59 ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ║ ║
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: Downloaded to shared memory in 46.341s
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpowx9vltn, device:0
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: quantized model in 36.319s
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: Processed model jic062/dpo-v1.3-Nemo-c500 in 82.660s
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: creating bucket guanaco-mkml-models
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/jic062-dpo-v1-3-nemo-c500-v1
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/jic062-dpo-v1-3-nemo-c500-v1/config.json
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/jic062-dpo-v1-3-nemo-c500-v1/special_tokens_map.json
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/jic062-dpo-v1-3-nemo-c500-v1/tokenizer_config.json
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/jic062-dpo-v1-3-nemo-c500-v1/tokenizer.json
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/jic062-dpo-v1-3-nemo-c500-v1/flywheel_model.0.safetensors
jic062-dpo-v1-3-nemo-c500-v1-mkmlizer: Loading 0: 0%| | 0/363 [00:00<?, ?it/s] Loading 0: 1%|▏ | 5/363 [00:00<00:11, 30.94it/s] Loading 0: 4%|▎ | 13/363 [00:00<00:06, 51.72it/s] Loading 0: 5%|▌ | 19/363 [00:00<00:07, 46.50it/s] Loading 0: 7%|▋ | 24/363 [00:00<00:07, 43.11it/s] Loading 0: 9%|▊ | 31/363 [00:00<00:06, 49.81it/s] Loading 0: 10%|█ | 37/363 [00:00<00:06, 47.73it/s] Loading 0: 12%|█▏ | 42/363 [00:00<00:07, 45.73it/s] Loading 0: 13%|█▎ | 49/363 [00:01<00:06, 50.42it/s] Loading 0: 15%|█▌ | 55/363 [00:01<00:06, 46.83it/s] Loading 0: 17%|█▋ | 60/363 [00:01<00:06, 46.58it/s] Loading 0: 18%|█▊ | 65/363 [00:01<00:09, 30.04it/s] Loading 0: 20%|█▉ | 72/363 [00:01<00:07, 37.03it/s] Loading 0: 21%|██ | 77/363 [00:01<00:07, 38.75it/s] Loading 0: 23%|██▎ | 82/363 [00:02<00:08, 33.51it/s] Loading 0: 25%|██▍ | 90/363 [00:02<00:06, 41.91it/s] Loading 0: 26%|██▋ | 96/363 [00:02<00:06, 40.44it/s] Loading 0: 28%|██▊ | 101/363 [00:02<00:06, 39.45it/s] Loading 0: 30%|███ | 109/363 [00:02<00:05, 48.16it/s] Loading 0: 32%|███▏ | 115/363 [00:02<00:05, 45.09it/s] Loading 0: 33%|███▎ | 120/363 [00:02<00:05, 41.13it/s] Loading 0: 35%|███▍ | 126/363 [00:02<00:05, 43.64it/s] Loading 0: 36%|███▋ | 132/363 [00:03<00:05, 43.41it/s] Loading 0: 38%|███▊ | 137/363 [00:03<00:05, 42.10it/s] Loading 0: 39%|███▉ | 142/363 [00:03<00:06, 32.59it/s] Loading 0: 40%|████ | 146/363 [00:03<00:06, 33.14it/s] Loading 0: 41%|████▏ | 150/363 [00:03<00:06, 31.37it/s] Loading 0: 43%|████▎ | 156/363 [00:03<00:05, 37.06it/s] Loading 0: 44%|████▍ | 160/363 [00:03<00:05, 37.13it/s] Loading 0: 46%|████▌ | 166/363 [00:04<00:04, 41.10it/s] Loading 0: 47%|████▋ | 171/363 [00:04<00:04, 41.88it/s] Loading 0: 48%|████▊ | 176/363 [00:04<00:05, 33.54it/s] Loading 0: 50%|█████ | 183/363 [00:04<00:04, 41.16it/s] Loading 0: 52%|█████▏ | 188/363 [00:04<00:04, 41.48it/s] Loading 0: 53%|█████▎ | 193/363 [00:04<00:04, 40.12it/s] Loading 0: 55%|█████▍ | 198/363 [00:04<00:03, 41.65it/s] Loading 0: 56%|█████▌ | 203/363 [00:05<00:04, 35.62it/s] Loading 0: 58%|█████▊ | 210/363 [00:05<00:03, 41.02it/s] Loading 0: 59%|█████▉ | 215/363 [00:05<00:03, 41.23it/s] Loading 0: 61%|██████ | 221/363 [00:05<00:03, 45.46it/s] Loading 0: 62%|██████▏ | 226/363 [00:05<00:04, 29.33it/s] Loading 0: 63%|██████▎ | 230/363 [00:05<00:04, 29.25it/s] Loading 0: 65%|██████▌ | 237/363 [00:06<00:03, 35.37it/s] Loading 0: 67%|██████▋ | 242/363 [00:06<00:03, 36.90it/s] Loading 0: 68%|██████▊ | 247/363 [00:06<00:02, 39.05it/s] Loading 0: 69%|██████▉ | 252/363 [00:06<00:02, 41.67it/s] Loading 0: 71%|███████ | 257/363 [00:06<00:03, 33.77it/s] Loading 0: 73%|███████▎ | 264/363 [00:06<00:02, 41.37it/s] Loading 0: 74%|███████▍ | 269/363 [00:06<00:02, 41.89it/s] Loading 0: 75%|███████▌ | 274/363 [00:06<00:02, 42.68it/s] Loading 0: 77%|███████▋ | 279/363 [00:07<00:01, 43.07it/s] Loading 0: 78%|███████▊ | 284/363 [00:07<00:02, 34.01it/s] Loading 0: 80%|███████▉ | 289/363 [00:07<00:01, 37.51it/s] Loading 0: 81%|████████ | 294/363 [00:07<00:01, 37.33it/s] Loading 0: 82%|████████▏ | 299/363 [00:07<00:01, 39.19it/s] Loading 0: 84%|████████▎ | 304/363 [00:14<00:24, 2.37it/s] Loading 0: 85%|████████▍ | 307/363 [00:14<00:19, 2.91it/s] Loading 0: 86%|████████▌ | 312/363 [00:14<00:12, 4.18it/s] Loading 0: 88%|████████▊ | 320/363 [00:14<00:06, 7.01it/s] Loading 0: 90%|████████▉ | 326/363 [00:14<00:03, 9.43it/s] Loading 0: 91%|█████████ | 331/363 [00:14<00:02, 11.89it/s] Loading 0: 93%|█████████▎| 338/363 [00:15<00:01, 16.70it/s] Loading 0: 94%|█████████▍| 343/363 [00:15<00:00, 20.26it/s] Loading 0: 96%|█████████▌| 348/363 [00:15<00:00, 21.50it/s] Loading 0: 98%|█████████▊| 356/363 [00:15<00:00, 29.42it/s] Loading 0: 100%|█████████▉| 362/363 [00:15<00:00, 31.82it/s]
Job jic062-dpo-v1-3-nemo-c500-v1-mkmlizer completed after 104.04s with status: succeeded
Stopping job with name jic062-dpo-v1-3-nemo-c500-v1-mkmlizer
Pipeline stage MKMLizer completed in 104.98s
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 jic062-dpo-v1-3-nemo-c500-v1
Waiting for inference service jic062-dpo-v1-3-nemo-c500-v1 to be ready
Inference service jic062-dpo-v1-3-nemo-c500-v1 ready after 170.82390642166138s
Pipeline stage MKMLDeployer completed in 171.19s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 5.984106540679932s
Received healthy response to inference request in 7.3015828132629395s
Received healthy response to inference request in 2.191322088241577s
Received healthy response to inference request in 1.8427000045776367s
Received healthy response to inference request in 5.160969972610474s
5 requests
0 failed requests
5th percentile: 1.912424421310425
10th percentile: 1.9821488380432128
20th percentile: 2.121597671508789
30th percentile: 2.7852516651153563
40th percentile: 3.9731108188629154
50th percentile: 5.160969972610474
60th percentile: 5.490224599838257
70th percentile: 5.81947922706604
80th percentile: 6.247601795196533
90th percentile: 6.774592304229737
95th percentile: 7.038087558746338
99th percentile: 7.248883762359619
mean time: 4.496136283874511
%s, retrying in %s seconds...
Received healthy response to inference request in 4.395796775817871s
Failed to get response for submission zonemercy-virgo-edit-v1-1e5_v12: HTTPConnectionPool(host='zonemercy-virgo-edit-v1-1e5-v12-predictor.tenant-chaiml-guanaco.k2.chaiverse.com', port=80): Max retries exceeded with url: /v1/models/GPT-J-6B-lit-v2:predict (Caused by ConnectTimeoutError(<urllib3.connection.HTTPConnection object at 0x7f3073865ee0>, 'Connection to zonemercy-virgo-edit-v1-1e5-v12-predictor.tenant-chaiml-guanaco.k2.chaiverse.com timed out. (connect timeout=None)'))
Received healthy response to inference request in 6.290462017059326s
Received healthy response to inference request in 2.549154043197632s
Received healthy response to inference request in 1.8792431354522705s
Received healthy response to inference request in 5.37450098991394s
5 requests
0 failed requests
5th percentile: 2.0132253170013428
10th percentile: 2.147207498550415
20th percentile: 2.4151718616485596
30th percentile: 2.9184825897216795
40th percentile: 3.6571396827697757
50th percentile: 4.395796775817871
60th percentile: 4.787278461456299
70th percentile: 5.178760147094726
80th percentile: 5.557693195343018
90th percentile: 5.9240776062011715
95th percentile: 6.107269811630249
99th percentile: 6.25382357597351
mean time: 4.097831392288208
%s, retrying in %s seconds...
Received healthy response to inference request in 3.118163824081421s
Received healthy response to inference request in 7.699349403381348s
Received healthy response to inference request in 1.7984416484832764s
Received healthy response to inference request in 2.0082385540008545s
Received healthy response to inference request in 3.2452681064605713s
5 requests
0 failed requests
5th percentile: 1.840401029586792
10th percentile: 1.8823604106903076
20th percentile: 1.9662791728973388
30th percentile: 2.230223608016968
40th percentile: 2.6741937160491944
50th percentile: 3.118163824081421
60th percentile: 3.169005537033081
70th percentile: 3.219847249984741
80th percentile: 4.136084365844727
90th percentile: 5.9177168846130375
95th percentile: 6.808533143997192
99th percentile: 7.521186151504517
mean time: 3.5738923072814943
Pipeline stage StressChecker completed in 63.25s
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
jic062-dpo-v1-3-nemo-c500_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 0.14s
run pipeline stage %s
Running pipeline stage MKMLProfilerTemplater
Pipeline stage MKMLProfilerTemplater completed in 0.14s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service jic062-dpo-v1-3-nemo-c500-v1-profiler
Waiting for inference service jic062-dpo-v1-3-nemo-c500-v1-profiler to be ready
Inference service jic062-dpo-v1-3-nemo-c500-v1-profiler ready after 170.41469192504883s
Pipeline stage MKMLProfilerDeployer completed in 170.81s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/jic062-dpo-v1-3-nemod12a43c898607b8283876cc6ba7c33fd-deplozclgd:/code/chaiverse_profiler_1726115569 --namespace tenant-chaiml-guanaco
kubectl exec -it jic062-dpo-v1-3-nemod12a43c898607b8283876cc6ba7c33fd-deplozclgd --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1726115569 && 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_1726115569/summary.json'
kubectl exec -it jic062-dpo-v1-3-nemod12a43c898607b8283876cc6ba7c33fd-deplozclgd --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1726115569/summary.json'
Pipeline stage MKMLProfilerRunner completed in 936.01s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service jic062-dpo-v1-3-nemo-c500-v1-profiler is running
Tearing down inference service jic062-dpo-v1-3-nemo-c500-v1-profiler
Service jic062-dpo-v1-3-nemo-c500-v1-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 2.10s
Shutdown handler de-registered
jic062-dpo-v1-3-nemo-c500_v1 status is now inactive due to auto deactivation removed underperforming models
jic062-dpo-v1-3-nemo-c500_v1 status is now torndown due to DeploymentManager action