trace2333-mistral-dpo-trail1

developer_uid: Trace2333

submission_id: trace2333-mistral-dpo-trail1_v1

model_name: trace2333-mistral-dpo-trail1_v1

model_group: Trace2333/mistral_dpo_tr

status: torndown

timestamp: 2024-09-10T16:59:23+00:00

num_battles: 11389

num_wins: 5817

celo_rating: 1251.31

family_friendly_score: 0.0

submission_type: basic

model_repo: Trace2333/mistral_dpo_trail1

model_architecture: MistralForCausalLM

model_num_parameters: 12772070400.0

best_of: 8

max_input_tokens: 512

max_output_tokens: 64

latencies: [{'batch_size': 1, 'throughput': 0.6926560498524543, 'latency_mean': 1.4436184895038604, 'latency_p50': 1.4449089765548706, 'latency_p90': 1.6152265787124633}, {'batch_size': 3, 'throughput': 1.3211229285979316, 'latency_mean': 2.259089765548706, 'latency_p50': 2.2683931589126587, 'latency_p90': 2.50491316318512}, {'batch_size': 5, 'throughput': 1.5531223386389754, 'latency_mean': 3.194834121465683, 'latency_p50': 3.200039029121399, 'latency_p90': 3.5695685148239136}, {'batch_size': 6, 'throughput': 1.598950392978291, 'latency_mean': 3.730621120929718, 'latency_p50': 3.7412561178207397, 'latency_p90': 4.209634065628052}, {'batch_size': 8, 'throughput': 1.5763131386602682, 'latency_mean': 5.0475011312961575, 'latency_p50': 5.065623998641968, 'latency_p90': 5.682017755508423}, {'batch_size': 10, 'throughput': 1.5451023702651931, 'latency_mean': 6.432719234228134, 'latency_p50': 6.4113689661026, 'latency_p90': 7.2425237655639645}]

gpu_counts: {'NVIDIA RTX A5000': 1}

display_name: trace2333-mistral-dpo-trail1_v1

is_internal_developer: False

language_model: Trace2333/mistral_dpo_trail1

model_size: 13B

ranking_group: single

throughput_3p7s: 1.61

us_pacific_date: 2024-09-10

win_ratio: 0.5107559926244623

generation_params: {'temperature': 1.0, 'top_p': 1.0, 'min_p': 0.06, 'top_k': 80, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '</s>'], 'max_input_tokens': 512, 'best_of': 8, '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 trace2333-mistral-dpo-trail1-v1-mkmlizer
Waiting for job on trace2333-mistral-dpo-trail1-v1-mkmlizer to finish
Failed to get response for submission blend_gamub_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', '{"error":"ValueError : [TypeError(\\"\'numpy.int64\' object is not iterable\\"), TypeError(\'vars() argument must have __dict__ attribute\')]"}')
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
trace2333-mistral-dpo-trail1-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║     _____            __           __                                ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║       /___/                                                         ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║                                                                     ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  Version: 0.10.1                                                    ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  https://mk1.ai                                                     ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║                                                                     ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  belonging to:                                                      ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║                                                                     ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  Chai Research Corp.                                                ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ║                                                                     ║
trace2333-mistral-dpo-trail1-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
Retrying (%r) after connection broken by '%r': %s
trace2333-mistral-dpo-trail1-v1-mkmlizer: Downloaded to shared memory in 48.867s
trace2333-mistral-dpo-trail1-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpinxyh3v7, device:0
trace2333-mistral-dpo-trail1-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
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-mistral-dpo-trail1-v1-mkmlizer: quantized model in 36.271s
trace2333-mistral-dpo-trail1-v1-mkmlizer: Processed model Trace2333/mistral_dpo_trail1 in 85.139s
trace2333-mistral-dpo-trail1-v1-mkmlizer: creating bucket guanaco-mkml-models
trace2333-mistral-dpo-trail1-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
trace2333-mistral-dpo-trail1-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/trace2333-mistral-dpo-trail1-v1
trace2333-mistral-dpo-trail1-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/trace2333-mistral-dpo-trail1-v1/config.json
trace2333-mistral-dpo-trail1-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/trace2333-mistral-dpo-trail1-v1/special_tokens_map.json
trace2333-mistral-dpo-trail1-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/trace2333-mistral-dpo-trail1-v1/tokenizer_config.json
trace2333-mistral-dpo-trail1-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/trace2333-mistral-dpo-trail1-v1/tokenizer.json
trace2333-mistral-dpo-trail1-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/trace2333-mistral-dpo-trail1-v1/flywheel_model.0.safetensors
trace2333-mistral-dpo-trail1-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   2%|▏         | 7/363 [00:00<00:07, 46.37it/s]
Loading 0:   6%|▌         | 22/363 [00:00<00:04, 79.64it/s]
Loading 0:   9%|▊         | 31/363 [00:00<00:04, 74.57it/s]
Loading 0:  11%|█         | 40/363 [00:00<00:04, 69.52it/s]
Loading 0:  13%|█▎        | 49/363 [00:00<00:04, 73.83it/s]
Loading 0:  16%|█▌        | 58/363 [00:00<00:03, 76.42it/s]
Loading 0:  18%|█▊        | 66/363 [00:01<00:14, 20.55it/s]
Loading 0:  20%|█▉        | 72/363 [00:01<00:12, 24.20it/s]
Loading 0:  22%|██▏       | 79/363 [00:02<00:09, 29.72it/s]
Loading 0:  24%|██▍       | 88/363 [00:02<00:07, 36.75it/s]
Loading 0:  27%|██▋       | 97/363 [00:02<00:05, 45.29it/s]
Loading 0:  29%|██▉       | 106/363 [00:02<00:04, 52.73it/s]
Loading 0:  32%|███▏      | 115/363 [00:02<00:04, 59.60it/s]
Loading 0:  34%|███▍      | 124/363 [00:02<00:03, 65.19it/s]
Loading 0:  37%|███▋      | 133/363 [00:02<00:03, 70.63it/s]
Loading 0:  39%|███▉      | 142/363 [00:03<00:10, 20.60it/s]
Loading 0:  42%|████▏     | 151/363 [00:04<00:08, 26.00it/s]
Loading 0:  44%|████▍     | 160/363 [00:04<00:06, 32.34it/s]
Loading 0:  47%|████▋     | 169/363 [00:04<00:04, 39.17it/s]
Loading 0:  49%|████▉     | 178/363 [00:04<00:04, 46.23it/s]
Loading 0:  52%|█████▏    | 187/363 [00:04<00:03, 50.83it/s]
Loading 0:  54%|█████▍    | 196/363 [00:04<00:02, 58.18it/s]
Loading 0:  56%|█████▋    | 205/363 [00:04<00:02, 64.93it/s]
Loading 0:  59%|█████▉    | 214/363 [00:04<00:02, 68.74it/s]
Loading 0:  61%|██████▏   | 223/363 [00:05<00:06, 21.06it/s]
Loading 0:  64%|██████▍   | 232/363 [00:06<00:04, 27.21it/s]
Loading 0:  66%|██████▋   | 241/363 [00:06<00:03, 33.50it/s]
Loading 0:  69%|██████▉   | 250/363 [00:06<00:02, 39.60it/s]
Loading 0:  72%|███████▏  | 260/363 [00:06<00:02, 49.17it/s]
Loading 0:  74%|███████▍  | 269/363 [00:06<00:01, 56.27it/s]
Loading 0:  77%|███████▋  | 278/363 [00:06<00:01, 61.17it/s]
Loading 0:  80%|████████  | 292/363 [00:06<00:00, 71.65it/s]
Loading 0:  83%|████████▎ | 302/363 [00:06<00:00, 77.79it/s]
Loading 0:  86%|████████▌ | 311/363 [00:07<00:02, 23.20it/s]
Loading 0:  88%|████████▊ | 319/363 [00:08<00:01, 27.55it/s]
Loading 0:  90%|█████████ | 328/363 [00:08<00:01, 32.69it/s]
Loading 0:  93%|█████████▎| 337/363 [00:08<00:00, 39.17it/s]
Loading 0:  95%|█████████▌| 346/363 [00:08<00:00, 46.77it/s]
Loading 0:  98%|█████████▊| 355/363 [00:08<00:00, 54.24it/s]
Loading 0: 100%|██████████| 363/363 [00:15<00:00,  4.30it/s]
                                                            
Job trace2333-mistral-dpo-trail1-v1-mkmlizer completed after 106.95s with status: succeeded
Stopping job with name trace2333-mistral-dpo-trail1-v1-mkmlizer
Pipeline stage MKMLizer completed in 108.30s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 1.44s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service trace2333-mistral-dpo-trail1-v1
Waiting for inference service trace2333-mistral-dpo-trail1-v1 to be ready
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
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
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
Failed to get response for submission chaiml-elo-alignment-run-3_v36: ('http://chaiml-elo-alignment-run-3-v36-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', 'read tcp 127.0.0.1:47296->127.0.0.1:8080: read: connection reset by peer\n')
Connection pool is full, discarding connection: %s. Connection pool size: %s
Inference service trace2333-mistral-dpo-trail1-v1 ready after 334.4732391834259s
Pipeline stage MKMLDeployer completed in 336.71s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.522244453430176s
Received healthy response to inference request in 2.3154265880584717s
Received healthy response to inference request in 1.7571938037872314s
Received healthy response to inference request in 1.6804587841033936s
Received healthy response to inference request in 2.2833642959594727s
5 requests
0 failed requests
5th percentile: 1.6958057880401611
10th percentile: 1.7111527919769287
20th percentile: 1.7418467998504639
30th percentile: 1.8624279022216796
40th percentile: 2.0728960990905763
50th percentile: 2.2833642959594727
60th percentile: 2.296189212799072
70th percentile: 2.3090141296386717
80th percentile: 2.3567901611328126
90th percentile: 2.439517307281494
95th percentile: 2.480880880355835
99th percentile: 2.5139717388153078
mean time: 2.111737585067749
Pipeline stage StressChecker completed in 11.48s
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 6.88s
Shutdown handler de-registered
trace2333-mistral-dpo-trail1_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 0.13s
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 trace2333-mistral-dpo-trail1-v1-profiler
Waiting for inference service trace2333-mistral-dpo-trail1-v1-profiler to be ready
Inference service trace2333-mistral-dpo-trail1-v1-profiler ready after 160.3967523574829s
Pipeline stage MKMLProfilerDeployer completed in 160.77s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/trace2333-mistral-dp4fce5e8703d5372a25e725a2ec7974bd-deplotkm9t:/code/chaiverse_profiler_1725988234 --namespace tenant-chaiml-guanaco
kubectl exec -it trace2333-mistral-dp4fce5e8703d5372a25e725a2ec7974bd-deplotkm9t --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1725988234 && 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_1725988234/summary.json'
kubectl exec -it trace2333-mistral-dp4fce5e8703d5372a25e725a2ec7974bd-deplotkm9t --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1725988234/summary.json'
Pipeline stage MKMLProfilerRunner completed in 960.23s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service trace2333-mistral-dpo-trail1-v1-profiler is running
Tearing down inference service trace2333-mistral-dpo-trail1-v1-profiler
Service trace2333-mistral-dpo-trail1-v1-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 2.27s
Shutdown handler de-registered
trace2333-mistral-dpo-trail1_v1 status is now inactive due to auto deactivation removed underperforming models
trace2333-mistral-dpo-trail1_v1 status is now torndown due to DeploymentManager action