rica40325-feedback-dpo-7

submission_id: rica40325-feedback-dpo-7_v1

developer_uid: rica40325

alignment_samples: 13545

alignment_score: -0.5672191369240738

best_of: 16

celo_rating: 883.3

display_name: rica40325-feedback-dpo-7_v1

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}

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'], 'max_input_tokens': 512, 'best_of': 16, 'max_output_tokens': 64}

gpu_counts: {'NVIDIA RTX A5000': 1}

is_internal_developer: False

language_model: rica40325/feedback_dpo_7

latencies: [{'batch_size': 1, 'throughput': 0.9001635084084265, 'latency_mean': 1.1108137023448945, 'latency_p50': 1.111203670501709, 'latency_p90': 1.231152081489563}, {'batch_size': 4, 'throughput': 1.8223350735813628, 'latency_mean': 2.1804723286628724, 'latency_p50': 2.1581379175186157, 'latency_p90': 2.4795950651168823}, {'batch_size': 5, 'throughput': 1.8806055253471996, 'latency_mean': 2.646614600419998, 'latency_p50': 2.646472454071045, 'latency_p90': 2.9561567544937133}, {'batch_size': 8, 'throughput': 2.0061408379787644, 'latency_mean': 3.9566712045669554, 'latency_p50': 3.9546924829483032, 'latency_p90': 4.445087552070618}, {'batch_size': 10, 'throughput': 2.0163921715243474, 'latency_mean': 4.9050632512569425, 'latency_p50': 4.8808372020721436, 'latency_p90': 5.624004530906677}, {'batch_size': 12, 'throughput': 2.0279307879287827, 'latency_mean': 5.834812220335007, 'latency_p50': 5.839073419570923, 'latency_p90': 6.757932686805725}, {'batch_size': 15, 'throughput': 2.029985961080009, 'latency_mean': 7.242907334566116, 'latency_p50': 7.3394775390625, 'latency_p90': 8.036587595939636}]

max_input_tokens: 512

max_output_tokens: 64

model_architecture: LlamaForCausalLM

model_group: rica40325/feedback_dpo_7

model_name: rica40325-feedback-dpo-7_v1

model_num_parameters: 8030261248.0

model_repo: rica40325/feedback_dpo_7

model_size: 8B

num_battles: 13544

num_wins: 1546

propriety_score: 0.6443123938879457

propriety_total_count: 1178.0

ranking_group: single

status: inactive

submission_type: basic

throughput_3p7s: 2.0

timestamp: 2024-09-10T10:42:49+00:00

us_pacific_date: 2024-09-10

win_ratio: 0.11414648552864737

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 rica40325-feedback-dpo-7-v1-mkmlizer
Waiting for job on rica40325-feedback-dpo-7-v1-mkmlizer to finish
rica40325-feedback-dpo-7-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
rica40325-feedback-dpo-7-v1-mkmlizer: ║     _____            __           __                                ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║       /___/                                                         ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║                                                                     ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  Version: 0.10.1                                                    ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  https://mk1.ai                                                     ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║                                                                     ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  belonging to:                                                      ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║                                                                     ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  Chai Research Corp.                                                ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
rica40325-feedback-dpo-7-v1-mkmlizer: ║                                                                     ║
rica40325-feedback-dpo-7-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
rica40325-feedback-dpo-7-v1-mkmlizer: Downloaded to shared memory in 67.437s
rica40325-feedback-dpo-7-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpvm_j0duk, device:0
rica40325-feedback-dpo-7-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
rica40325-feedback-dpo-7-v1-mkmlizer: quantized model in 29.527s
rica40325-feedback-dpo-7-v1-mkmlizer: Processed model rica40325/feedback_dpo_7 in 96.963s
rica40325-feedback-dpo-7-v1-mkmlizer: creating bucket guanaco-mkml-models
rica40325-feedback-dpo-7-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
rica40325-feedback-dpo-7-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/rica40325-feedback-dpo-7-v1
rica40325-feedback-dpo-7-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/rica40325-feedback-dpo-7-v1/tokenizer.json
rica40325-feedback-dpo-7-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/rica40325-feedback-dpo-7-v1/flywheel_model.0.safetensors
rica40325-feedback-dpo-7-v1-mkmlizer: 
Loading 0:   0%|          | 0/291 [00:00<?, ?it/s]
Loading 0:   2%|▏         | 5/291 [00:00<00:10, 26.30it/s]
Loading 0:   4%|▍         | 12/291 [00:00<00:07, 35.10it/s]
Loading 0:   5%|▌         | 16/291 [00:00<00:08, 32.49it/s]
Loading 0:   7%|▋         | 21/291 [00:00<00:07, 36.31it/s]
Loading 0:   9%|▊         | 25/291 [00:00<00:08, 33.01it/s]
Loading 0:  11%|█         | 31/291 [00:00<00:06, 39.27it/s]
Loading 0:  12%|█▏        | 36/291 [00:01<00:10, 23.89it/s]
Loading 0:  14%|█▍        | 41/291 [00:01<00:09, 25.18it/s]
Loading 0:  16%|█▋        | 48/291 [00:01<00:07, 31.30it/s]
Loading 0:  18%|█▊        | 52/291 [00:01<00:07, 30.11it/s]
Loading 0:  20%|█▉        | 57/291 [00:01<00:07, 31.75it/s]
Loading 0:  21%|██        | 61/291 [00:02<00:07, 29.38it/s]
Loading 0:  22%|██▏       | 65/291 [00:02<00:07, 31.44it/s]
Loading 0:  24%|██▎       | 69/291 [00:02<00:07, 28.69it/s]
Loading 0:  25%|██▌       | 74/291 [00:02<00:06, 31.76it/s]
Loading 0:  27%|██▋       | 78/291 [00:02<00:06, 30.96it/s]
Loading 0:  28%|██▊       | 82/291 [00:02<00:08, 23.37it/s]
Loading 0:  29%|██▉       | 85/291 [00:02<00:08, 24.35it/s]
Loading 0:  31%|███       | 90/291 [00:03<00:07, 28.00it/s]
Loading 0:  32%|███▏      | 94/291 [00:03<00:07, 27.91it/s]
Loading 0:  34%|███▍      | 99/291 [00:03<00:06, 30.14it/s]
Loading 0:  35%|███▌      | 103/291 [00:03<00:06, 29.61it/s]
Loading 0:  37%|███▋      | 108/291 [00:03<00:05, 32.07it/s]
Loading 0:  38%|███▊      | 112/291 [00:03<00:05, 30.48it/s]
Loading 0:  40%|███▉      | 116/291 [00:03<00:05, 31.33it/s]
Loading 0:  42%|████▏     | 122/291 [00:04<00:04, 35.70it/s]
Loading 0:  44%|████▎     | 127/291 [00:04<00:04, 33.53it/s]
Loading 0:  46%|████▌     | 133/291 [00:04<00:05, 29.83it/s]
Loading 0:  47%|████▋     | 137/291 [00:04<00:05, 29.33it/s]
Loading 0:  48%|████▊     | 141/291 [00:04<00:05, 27.47it/s]
Loading 0:  51%|█████     | 147/291 [00:04<00:04, 32.85it/s]
Loading 0:  52%|█████▏    | 151/291 [00:04<00:04, 32.28it/s]
Loading 0:  54%|█████▎    | 156/291 [00:05<00:03, 34.23it/s]
Loading 0:  55%|█████▍    | 160/291 [00:05<00:04, 32.15it/s]
Loading 0:  57%|█████▋    | 165/291 [00:05<00:03, 34.40it/s]
Loading 0:  58%|█████▊    | 169/291 [00:05<00:03, 32.35it/s]
Loading 0:  60%|█████▉    | 174/291 [00:05<00:03, 34.05it/s]
Loading 0:  61%|██████    | 178/291 [00:05<00:03, 32.37it/s]
Loading 0:  63%|██████▎   | 183/291 [00:05<00:03, 35.91it/s]
Loading 0:  64%|██████▍   | 187/291 [00:06<00:04, 25.67it/s]
Loading 0:  66%|██████▌   | 191/291 [00:06<00:03, 26.73it/s]
Loading 0:  67%|██████▋   | 195/291 [00:06<00:03, 25.35it/s]
Loading 0:  69%|██████▉   | 201/291 [00:06<00:02, 30.66it/s]
Loading 0:  70%|███████   | 205/291 [00:06<00:02, 29.41it/s]
Loading 0:  72%|███████▏  | 210/291 [00:06<00:02, 31.44it/s]
Loading 0:  74%|███████▎  | 214/291 [00:07<00:02, 30.87it/s]
Loading 0:  75%|███████▌  | 219/291 [00:07<00:02, 33.17it/s]
Loading 0:  77%|███████▋  | 223/291 [00:07<00:02, 31.73it/s]
Loading 0:  78%|███████▊  | 227/291 [00:07<00:02, 31.87it/s]
Loading 0:  79%|███████▉  | 231/291 [00:07<00:01, 32.19it/s]
Loading 0:  81%|████████  | 235/291 [00:07<00:02, 24.03it/s]
Loading 0:  82%|████████▏ | 239/291 [00:07<00:02, 23.96it/s]
Loading 0:  85%|████████▍ | 246/291 [00:08<00:01, 30.83it/s]
Loading 0:  86%|████████▌ | 250/291 [00:08<00:01, 30.07it/s]
Loading 0:  88%|████████▊ | 255/291 [00:08<00:01, 33.00it/s]
Loading 0:  89%|████████▉ | 259/291 [00:08<00:01, 31.47it/s]
Loading 0:  91%|█████████ | 264/291 [00:08<00:00, 32.70it/s]
Loading 0:  92%|█████████▏| 268/291 [00:08<00:00, 31.60it/s]
Loading 0:  94%|█████████▍| 273/291 [00:08<00:00, 34.49it/s]
Loading 0:  95%|█████████▌| 277/291 [00:09<00:00, 32.09it/s]
Loading 0:  97%|█████████▋| 281/291 [00:09<00:00, 32.21it/s]
Loading 0:  98%|█████████▊| 286/291 [00:14<00:01,  2.57it/s]
Loading 0:  99%|█████████▉| 289/291 [00:14<00:00,  3.18it/s]
                                                            
Job rica40325-feedback-dpo-7-v1-mkmlizer completed after 115.81s with status: succeeded
Stopping job with name rica40325-feedback-dpo-7-v1-mkmlizer
Pipeline stage MKMLizer completed in 116.82s
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 rica40325-feedback-dpo-7-v1
Waiting for inference service rica40325-feedback-dpo-7-v1 to be ready
Connection pool is full, discarding connection: %s. Connection pool size: %s
Failed to get response for submission blend_sisun_2024-09-09: ('http://chaiml-llama-8b-pairwis-8189-v19-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', 'readfrom tcp 127.0.0.1:41646->127.0.0.1:8080: write tcp 127.0.0.1:41646->127.0.0.1:8080: use of closed network connection\n')
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 rica40325-feedback-dpo-7-v1 ready after 191.18641304969788s
Pipeline stage MKMLDeployer completed in 191.52s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 1.9437921047210693s
Received healthy response to inference request in 1.4477081298828125s
Received healthy response to inference request in 2.1275343894958496s
Received healthy response to inference request in 1.6078767776489258s
Received healthy response to inference request in 1.709639549255371s
5 requests
0 failed requests
5th percentile: 1.4797418594360352
10th percentile: 1.511775588989258
20th percentile: 1.575843048095703
30th percentile: 1.6282293319702148
40th percentile: 1.668934440612793
50th percentile: 1.709639549255371
60th percentile: 1.8033005714416503
70th percentile: 1.8969615936279296
80th percentile: 1.9805405616760254
90th percentile: 2.0540374755859374
95th percentile: 2.0907859325408937
99th percentile: 2.1201846981048584
mean time: 1.7673101902008057
Pipeline stage StressChecker completed in 9.53s
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 8.58s
Shutdown handler de-registered
rica40325-feedback-dpo-7_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 0.11s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service rica40325-feedback-dpo-7-v1-profiler
Waiting for inference service rica40325-feedback-dpo-7-v1-profiler to be ready
Inference service rica40325-feedback-dpo-7-v1-profiler ready after 160.4412636756897s
Pipeline stage MKMLProfilerDeployer completed in 160.81s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/rica40325-feedback-dpo-7-v1-profiler-predictor-00001-deplonfkth:/code/chaiverse_profiler_1725965494 --namespace tenant-chaiml-guanaco
kubectl exec -it rica40325-feedback-dpo-7-v1-profiler-predictor-00001-deplonfkth --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1725965494 && 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_1725965494/summary.json'
kubectl exec -it rica40325-feedback-dpo-7-v1-profiler-predictor-00001-deplonfkth --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1725965494/summary.json'
Pipeline stage MKMLProfilerRunner completed in 841.19s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service rica40325-feedback-dpo-7-v1-profiler is running
Tearing down inference service rica40325-feedback-dpo-7-v1-profiler
Service rica40325-feedback-dpo-7-v1-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 1.71s
Shutdown handler de-registered
rica40325-feedback-dpo-7_v1 status is now inactive due to auto deactivation removed underperforming models