chaiml-horror-sft-v1-5e4

developer_uid: ebony

submission_id: chaiml-horror-sft-v1-5e4_v6

model_name: chaiml-horror-sft-v1-5e4_v4

model_group: ChaiML/Horror-SFT-v1-5e4

status: torndown

timestamp: 2024-09-05T17:29:51+00:00

num_battles: 13172

num_wins: 6113

celo_rating: 1215.05

family_friendly_score: 0.0

submission_type: basic

model_repo: ChaiML/Horror-SFT-v1-5e4

model_architecture: MistralForCausalLM

model_num_parameters: 12772070400.0

best_of: 4

max_input_tokens: 1024

max_output_tokens: 64

latencies: [{'batch_size': 1, 'throughput': 0.6338699305251079, 'latency_mean': 1.5775426232814789, 'latency_p50': 1.578340768814087, 'latency_p90': 1.7427193164825439}, {'batch_size': 3, 'throughput': 1.246921397246944, 'latency_mean': 2.3923956298828126, 'latency_p50': 2.3756675720214844, 'latency_p90': 2.6668697595596313}, {'batch_size': 5, 'throughput': 1.5356197635257656, 'latency_mean': 3.2388183069229126, 'latency_p50': 3.2322596311569214, 'latency_p90': 3.6233068466186524}, {'batch_size': 6, 'throughput': 1.6337177426136844, 'latency_mean': 3.661092474460602, 'latency_p50': 3.651573896408081, 'latency_p90': 4.1136840820312495}, {'batch_size': 8, 'throughput': 1.7375611162968, 'latency_mean': 4.585595206022263, 'latency_p50': 4.572147727012634, 'latency_p90': 5.15118510723114}, {'batch_size': 10, 'throughput': 1.7830035786994034, 'latency_mean': 5.5849633419513705, 'latency_p50': 5.6196876764297485, 'latency_p90': 6.36077127456665}]

gpu_counts: {'NVIDIA RTX A5000': 1}

display_name: chaiml-horror-sft-v1-5e4_v4

is_internal_developer: False

language_model: ChaiML/Horror-SFT-v1-5e4

model_size: 13B

ranking_group: single

throughput_3p7s: 1.64

us_pacific_date: 2024-09-05

win_ratio: 0.4640904949893714

generation_params: {'temperature': 0.9, 'top_p': 1.0, 'min_p': 0.05, 'top_k': 100, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '</s>'], 'max_input_tokens': 1024, 'best_of': 4, 'max_output_tokens': 64}

formatter: {'memory_template': "{bot_name}'s Persona: {memory}\n##Make the conversation as scary as possible while keep in character##\n", 'prompt_template': '{prompt}\n<START>\n', 'bot_template': '{bot_name}: {message}\n', 'user_template': '{user_name}: {message}\n', 'response_template': '{bot_name}[Extreme horror edition]:', '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 chaiml-horror-sft-v1-5e4-v6-mkmlizer
Waiting for job on chaiml-horror-sft-v1-5e4-v6-mkmlizer to finish
Failed to get response for submission ebony59-chaiml-stf-horror_v2: ('Connection aborted.', ConnectionResetError(104, 'Connection reset by peer'))
Failed to get response for submission chaiml-llama-8b-pairwis_8189_v19: ('Connection aborted.', ConnectionResetError(104, 'Connection reset by peer'))
Connection pool is full, discarding connection: %s. Connection pool size: %s
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║     _____            __           __                                ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║       /___/                                                         ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║                                                                     ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  Version: 0.10.1                                                    ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  https://mk1.ai                                                     ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║                                                                     ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  The license key for the current software has been verified as      ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  belonging to:                                                      ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║                                                                     ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  Chai Research Corp.                                                ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ║                                                                     ║
chaiml-horror-sft-v1-5e4-v6-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
Failed to get response for submission blend_jitik_2024-08-26: ('http://zonemercy-lexical-nemo-1518-v18-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', 'read tcp 127.0.0.1:46768->127.0.0.1:8080: read: connection reset by peer\n')
chaiml-horror-sft-v1-5e4-v6-mkmlizer: Downloaded to shared memory in 42.089s
chaiml-horror-sft-v1-5e4-v6-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmp_hvpu0ss, device:0
chaiml-horror-sft-v1-5e4-v6-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
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
chaiml-horror-sft-v1-5e4-v6-mkmlizer: quantized model in 36.135s
chaiml-horror-sft-v1-5e4-v6-mkmlizer: Processed model ChaiML/Horror-SFT-v1-5e4 in 78.223s
chaiml-horror-sft-v1-5e4-v6-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-horror-sft-v1-5e4-v6-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-horror-sft-v1-5e4-v6
chaiml-horror-sft-v1-5e4-v6-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-horror-sft-v1-5e4-v6/config.json
chaiml-horror-sft-v1-5e4-v6-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-horror-sft-v1-5e4-v6/special_tokens_map.json
chaiml-horror-sft-v1-5e4-v6-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-horror-sft-v1-5e4-v6/tokenizer_config.json
chaiml-horror-sft-v1-5e4-v6-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-horror-sft-v1-5e4-v6/tokenizer.json
chaiml-horror-sft-v1-5e4-v6-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-horror-sft-v1-5e4-v6/flywheel_model.0.safetensors
chaiml-horror-sft-v1-5e4-v6-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   1%|▏         | 5/363 [00:00<00:12, 29.52it/s]
Loading 0:   4%|▎         | 13/363 [00:00<00:06, 50.08it/s]
Loading 0:   5%|▌         | 19/363 [00:00<00:07, 44.41it/s]
Loading 0:   7%|▋         | 24/363 [00:00<00:07, 42.99it/s]
Loading 0:   9%|▊         | 31/363 [00:00<00:06, 49.24it/s]
Loading 0:  10%|█         | 37/363 [00:00<00:07, 46.16it/s]
Loading 0:  12%|█▏        | 42/363 [00:00<00:07, 44.75it/s]
Loading 0:  13%|█▎        | 48/363 [00:01<00:06, 48.74it/s]
Loading 0:  15%|█▍        | 54/363 [00:01<00:06, 47.79it/s]
Loading 0:  16%|█▋        | 59/363 [00:01<00:06, 48.15it/s]
Loading 0:  18%|█▊        | 64/363 [00:01<00:11, 26.74it/s]
Loading 0:  20%|█▉        | 71/363 [00:01<00:08, 33.49it/s]
Loading 0:  21%|██        | 76/363 [00:01<00:08, 35.45it/s]
Loading 0:  22%|██▏       | 81/363 [00:02<00:07, 37.85it/s]
Loading 0:  24%|██▍       | 87/363 [00:02<00:07, 38.26it/s]
Loading 0:  25%|██▌       | 92/363 [00:02<00:06, 39.54it/s]
Loading 0:  27%|██▋       | 98/363 [00:02<00:05, 44.30it/s]
Loading 0:  28%|██▊       | 103/363 [00:02<00:05, 43.41it/s]
Loading 0:  30%|███       | 110/363 [00:02<00:05, 43.55it/s]
Loading 0:  32%|███▏      | 115/363 [00:02<00:05, 44.71it/s]
Loading 0:  33%|███▎      | 120/363 [00:02<00:05, 42.98it/s]
Loading 0:  35%|███▍      | 126/363 [00:03<00:05, 46.10it/s]
Loading 0:  36%|███▌      | 131/363 [00:03<00:05, 46.03it/s]
Loading 0:  37%|███▋      | 136/363 [00:03<00:05, 38.38it/s]
Loading 0:  39%|███▉      | 142/363 [00:03<00:06, 34.17it/s]
Loading 0:  40%|████      | 146/363 [00:03<00:06, 34.30it/s]
Loading 0:  41%|████▏     | 150/363 [00:03<00:06, 33.35it/s]
Loading 0:  43%|████▎     | 156/363 [00:03<00:05, 38.70it/s]
Loading 0:  44%|████▍     | 161/363 [00:03<00:05, 40.02it/s]
Loading 0:  46%|████▌     | 166/363 [00:04<00:04, 41.04it/s]
Loading 0:  47%|████▋     | 171/363 [00:04<00:04, 41.49it/s]
Loading 0:  48%|████▊     | 176/363 [00:04<00:05, 35.17it/s]
Loading 0:  50%|█████     | 183/363 [00:04<00:04, 41.27it/s]
Loading 0:  52%|█████▏    | 188/363 [00:04<00:04, 41.50it/s]
Loading 0:  53%|█████▎    | 193/363 [00:04<00:03, 42.67it/s]
Loading 0:  55%|█████▍    | 199/363 [00:04<00:04, 40.76it/s]
Loading 0:  56%|█████▌    | 204/363 [00:05<00:04, 39.05it/s]
Loading 0:  58%|█████▊    | 210/363 [00:05<00:03, 43.40it/s]
Loading 0:  59%|█████▉    | 215/363 [00:05<00:03, 43.40it/s]
Loading 0:  61%|██████    | 221/363 [00:05<00:03, 47.04it/s]
Loading 0:  62%|██████▏   | 226/363 [00:05<00:04, 30.82it/s]
Loading 0:  63%|██████▎   | 230/363 [00:05<00:04, 30.39it/s]
Loading 0:  65%|██████▌   | 237/363 [00:05<00:03, 37.83it/s]
Loading 0:  67%|██████▋   | 242/363 [00:06<00:03, 39.36it/s]
Loading 0:  68%|██████▊   | 247/363 [00:06<00:02, 41.29it/s]
Loading 0:  70%|██████▉   | 253/363 [00:06<00:02, 40.45it/s]
Loading 0:  71%|███████   | 258/363 [00:06<00:02, 39.69it/s]
Loading 0:  73%|███████▎  | 264/363 [00:06<00:02, 42.92it/s]
Loading 0:  74%|███████▍  | 269/363 [00:06<00:02, 42.51it/s]
Loading 0:  75%|███████▌  | 274/363 [00:06<00:02, 41.62it/s]
Loading 0:  77%|███████▋  | 279/363 [00:06<00:01, 43.17it/s]
Loading 0:  78%|███████▊  | 284/363 [00:07<00:02, 36.41it/s]
Loading 0:  80%|████████  | 291/363 [00:07<00:01, 43.20it/s]
Loading 0:  82%|████████▏ | 296/363 [00:07<00:01, 42.69it/s]
Loading 0:  83%|████████▎ | 301/363 [00:07<00:01, 43.79it/s]
Loading 0:  84%|████████▍ | 306/363 [00:14<00:22,  2.49it/s]
Loading 0:  85%|████████▌ | 310/363 [00:14<00:16,  3.22it/s]
Loading 0:  87%|████████▋ | 314/363 [00:14<00:11,  4.23it/s]
Loading 0:  88%|████████▊ | 320/363 [00:14<00:06,  6.32it/s]
Loading 0:  90%|████████▉ | 325/363 [00:14<00:04,  8.51it/s]
Loading 0:  91%|█████████ | 330/363 [00:14<00:03, 10.65it/s]
Loading 0:  93%|█████████▎| 338/363 [00:14<00:01, 16.18it/s]
Loading 0:  94%|█████████▍| 343/363 [00:15<00:01, 19.49it/s]
Loading 0:  96%|█████████▌| 348/363 [00:15<00:00, 21.02it/s]
Loading 0:  98%|█████████▊| 355/363 [00:15<00:00, 27.77it/s]
Loading 0:  99%|█████████▉| 360/363 [00:15<00:00, 30.49it/s]
                                                            
Job chaiml-horror-sft-v1-5e4-v6-mkmlizer completed after 106.01s with status: succeeded
Stopping job with name chaiml-horror-sft-v1-5e4-v6-mkmlizer
Pipeline stage MKMLizer completed in 107.12s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.22s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service chaiml-horror-sft-v1-5e4-v6
Waiting for inference service chaiml-horror-sft-v1-5e4-v6 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
Inference service chaiml-horror-sft-v1-5e4-v6 ready after 150.569726228714s
Pipeline stage MKMLDeployer completed in 151.20s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 1.7586886882781982s
Received healthy response to inference request in 1.4264020919799805s
Received healthy response to inference request in 1.9833533763885498s
Received healthy response to inference request in 1.812988519668579s
Received healthy response to inference request in 1.9285075664520264s
5 requests
0 failed requests
5th percentile: 1.492859411239624
10th percentile: 1.5593167304992677
20th percentile: 1.6922313690185546
30th percentile: 1.7695486545562744
40th percentile: 1.7912685871124268
50th percentile: 1.812988519668579
60th percentile: 1.859196138381958
70th percentile: 1.9054037570953368
80th percentile: 1.939476728439331
90th percentile: 1.9614150524139404
95th percentile: 1.9723842144012451
99th percentile: 1.981159543991089
mean time: 1.7819880485534667
Pipeline stage StressChecker completed in 9.92s
run pipeline stage %s
Running pipeline stage TriggerMKMLProfilingPipeline
run_pipeline:run_in_cloud %s
starting trigger_guanaco_pipeline args=%s
Connection pool is full, discarding connection: %s. Connection pool size: %s
Pipeline stage TriggerMKMLProfilingPipeline completed in 4.40s
Shutdown handler de-registered
chaiml-horror-sft-v1-5e4_v6 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 0.12s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service chaiml-horror-sft-v1-5e4-v6-profiler
Waiting for inference service chaiml-horror-sft-v1-5e4-v6-profiler to be ready
Inference service chaiml-horror-sft-v1-5e4-v6-profiler ready after 140.32619166374207s
Pipeline stage MKMLProfilerDeployer completed in 140.85s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/chaiml-horror-sft-v1-5e4-v6-profiler-predictor-00001-deplojmmls:/code/chaiverse_profiler_1725557843 --namespace tenant-chaiml-guanaco
kubectl exec -it chaiml-horror-sft-v1-5e4-v6-profiler-predictor-00001-deplojmmls --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1725557843 && python profiles.py profile --best_of_n 4 --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 1024 --output_tokens 64 --summary /code/chaiverse_profiler_1725557843/summary.json'
kubectl exec -it chaiml-horror-sft-v1-5e4-v6-profiler-predictor-00001-deplojmmls --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1725557843/summary.json'
Pipeline stage MKMLProfilerRunner completed in 961.42s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service chaiml-horror-sft-v1-5e4-v6-profiler is running
Tearing down inference service chaiml-horror-sft-v1-5e4-v6-profiler
Service chaiml-horror-sft-v1-5e4-v6-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 1.59s
Shutdown handler de-registered
chaiml-horror-sft-v1-5e4_v6 status is now inactive due to auto deactivation removed underperforming models
chaiml-horror-sft-v1-5e4_v6 status is now torndown due to DeploymentManager action