rica40325-mistral-4904

submission_id: rica40325-mistral-4904_v1

developer_uid: rica40325

best_of: 8

celo_rating: 1265.21

display_name: rica40325-mistral-13b-2452_v1

family_friendly_score: 0.5665315771698751

family_friendly_standard_error: 0.009084942346253636

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': 0.95, 'top_p': 0.95, 'min_p': 0.05, 'top_k': 80, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['</s>', 'Bot:', 'User:', 'You:', 'Me:', '####'], 'max_input_tokens': 1024, 'best_of': 8, 'max_output_tokens': 64}

gpu_counts: {'NVIDIA RTX A5000': 1}

ineligible_reason: num_battles<5000

is_internal_developer: False

language_model: rica40325/mistral-4904

latencies: [{'batch_size': 1, 'throughput': 0.61598209753271, 'latency_mean': 1.6233550083637238, 'latency_p50': 1.620248794555664, 'latency_p90': 1.7831089496612549}, {'batch_size': 3, 'throughput': 1.0877493216326408, 'latency_mean': 2.7433943593502046, 'latency_p50': 2.747072696685791, 'latency_p90': 2.979470634460449}, {'batch_size': 5, 'throughput': 1.2264191515429697, 'latency_mean': 4.052457767724991, 'latency_p50': 4.0707279443740845, 'latency_p90': 4.541128373146057}, {'batch_size': 6, 'throughput': 1.2539368041327048, 'latency_mean': 4.767187936306, 'latency_p50': 4.752150058746338, 'latency_p90': 5.330844044685364}, {'batch_size': 8, 'throughput': 1.241121150836812, 'latency_mean': 6.4220646023750305, 'latency_p50': 6.441121220588684, 'latency_p90': 7.368527722358704}, {'batch_size': 10, 'throughput': 1.1854287261424958, 'latency_mean': 8.390413134098052, 'latency_p50': 8.45491623878479, 'latency_p90': 9.447687554359437}]

max_input_tokens: 1024

max_output_tokens: 64

model_architecture: MistralForCausalLM

model_group: rica40325/mistral-4904

model_name: rica40325-mistral-13b-2452_v1

model_num_parameters: 12772070400.0

model_repo: rica40325/mistral-4904

model_size: 13B

num_battles: 3034

num_wins: 1563

ranking_group: single

status: torndown

submission_type: basic

throughput_3p7s: 1.21

timestamp: 2024-09-25T09:42:16+00:00

us_pacific_date: 2024-09-25

win_ratio: 0.515161502966381

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-mistral-4904-v1-mkmlizer
Waiting for job on rica40325-mistral-4904-v1-mkmlizer to finish
rica40325-mistral-4904-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
rica40325-mistral-4904-v1-mkmlizer: ║     _____            __           __                                ║
rica40325-mistral-4904-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
rica40325-mistral-4904-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
rica40325-mistral-4904-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
rica40325-mistral-4904-v1-mkmlizer: ║       /___/                                                         ║
rica40325-mistral-4904-v1-mkmlizer: ║                                                                     ║
rica40325-mistral-4904-v1-mkmlizer: ║  Version: 0.11.12                                                   ║
rica40325-mistral-4904-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
rica40325-mistral-4904-v1-mkmlizer: ║  https://mk1.ai                                                     ║
rica40325-mistral-4904-v1-mkmlizer: ║                                                                     ║
rica40325-mistral-4904-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
rica40325-mistral-4904-v1-mkmlizer: ║  belonging to:                                                      ║
rica40325-mistral-4904-v1-mkmlizer: ║                                                                     ║
rica40325-mistral-4904-v1-mkmlizer: ║  Chai Research Corp.                                                ║
rica40325-mistral-4904-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
rica40325-mistral-4904-v1-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
rica40325-mistral-4904-v1-mkmlizer: ║                                                                     ║
rica40325-mistral-4904-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
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
rica40325-mistral-4904-v1-mkmlizer: Downloaded to shared memory in 46.456s
rica40325-mistral-4904-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpqidamiuo, device:0
rica40325-mistral-4904-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
rica40325-mistral-4904-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/mk1/flywheel/functional/loader.py:55: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
rica40325-mistral-4904-v1-mkmlizer:   tensors = torch.load(model_shard_filename, map_location=torch.device(self.device), mmap=True)
Connection pool is full, discarding connection: %s. Connection pool size: %s
Connection pool is full, discarding connection: %s. Connection pool size: %s
rica40325-mistral-4904-v1-mkmlizer: quantized model in 35.302s
rica40325-mistral-4904-v1-mkmlizer: Processed model rica40325/mistral-4904 in 81.759s
rica40325-mistral-4904-v1-mkmlizer: creating bucket guanaco-mkml-models
rica40325-mistral-4904-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
rica40325-mistral-4904-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/rica40325-mistral-4904-v1
rica40325-mistral-4904-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/rica40325-mistral-4904-v1/config.json
Retrying (%r) after connection broken by '%r': %s
rica40325-mistral-4904-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/rica40325-mistral-4904-v1/flywheel_model.0.safetensors
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
rica40325-mistral-4904-v1-mkmlizer: 
Loading 0:   0%|          | 0/363 [00:00<?, ?it/s]
Loading 0:   2%|▏         | 7/363 [00:00<00:06, 53.08it/s]
Loading 0:   4%|▍         | 16/363 [00:00<00:04, 71.74it/s]
Loading 0:   7%|▋         | 25/363 [00:00<00:04, 75.70it/s]
Loading 0:   9%|▉         | 34/363 [00:00<00:04, 80.15it/s]
Loading 0:  12%|█▏        | 44/363 [00:00<00:03, 86.56it/s]
Loading 0:  15%|█▍        | 53/363 [00:00<00:03, 86.78it/s]
Loading 0:  17%|█▋        | 62/363 [00:01<00:15, 19.95it/s]
Loading 0:  19%|█▉        | 70/363 [00:01<00:11, 25.44it/s]
Loading 0:  22%|██▏       | 80/363 [00:02<00:08, 34.07it/s]
Loading 0:  25%|██▍       | 89/363 [00:02<00:06, 41.99it/s]
Loading 0:  27%|██▋       | 98/363 [00:02<00:05, 48.66it/s]
Loading 0:  29%|██▉       | 106/363 [00:02<00:04, 52.31it/s]
Loading 0:  32%|███▏      | 115/363 [00:02<00:04, 57.58it/s]
Loading 0:  34%|███▍      | 125/363 [00:02<00:03, 66.58it/s]
Loading 0:  38%|███▊      | 139/363 [00:02<00:02, 75.91it/s]
Loading 0:  41%|████      | 148/363 [00:03<00:09, 22.32it/s]
Loading 0:  43%|████▎     | 157/363 [00:04<00:07, 27.59it/s]
Loading 0:  46%|████▌     | 166/363 [00:04<00:05, 34.17it/s]
Loading 0:  48%|████▊     | 176/363 [00:04<00:04, 42.90it/s]
Loading 0:  52%|█████▏    | 187/363 [00:04<00:03, 49.94it/s]
Loading 0:  54%|█████▍    | 196/363 [00:04<00:03, 54.44it/s]
Loading 0:  56%|█████▋    | 205/363 [00:04<00:02, 58.66it/s]
Loading 0:  59%|█████▉    | 214/363 [00:04<00:02, 64.43it/s]
Loading 0:  61%|██████▏   | 223/363 [00:05<00:06, 21.79it/s]
Loading 0:  66%|██████▌   | 238/363 [00:05<00:03, 32.08it/s]
Loading 0:  68%|██████▊   | 247/363 [00:06<00:03, 36.68it/s]
Loading 0:  71%|███████   | 256/363 [00:06<00:02, 43.10it/s]
Loading 0:  74%|███████▎  | 267/363 [00:06<00:01, 53.51it/s]
Loading 0:  76%|███████▋  | 277/363 [00:06<00:01, 57.62it/s]
Loading 0:  79%|███████▉  | 286/363 [00:06<00:01, 62.15it/s]
Loading 0:  81%|████████▏ | 295/363 [00:06<00:01, 65.94it/s]
Loading 0:  84%|████████▎ | 304/363 [00:07<00:02, 22.79it/s]
Loading 0:  88%|████████▊ | 319/363 [00:07<00:01, 33.39it/s]
Loading 0:  90%|█████████ | 328/363 [00:08<00:00, 39.55it/s]
Loading 0:  93%|█████████▎| 337/363 [00:08<00:00, 45.54it/s]
Loading 0:  96%|█████████▌| 348/363 [00:08<00:00, 55.98it/s]
Loading 0:  99%|█████████▊| 358/363 [00:08<00:00, 59.65it/s]
                                                            
Job rica40325-mistral-4904-v1-mkmlizer completed after 108.58s with status: succeeded
Stopping job with name rica40325-mistral-4904-v1-mkmlizer
Pipeline stage MKMLizer completed in 109.45s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.11s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service rica40325-mistral-4904-v1
Waiting for inference service rica40325-mistral-4904-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
Inference service rica40325-mistral-4904-v1 ready after 222.1606616973877s
Pipeline stage MKMLDeployer completed in 222.57s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 4.789892196655273s
Received healthy response to inference request in 1.6313769817352295s
Received healthy response to inference request in 2.433115005493164s
Received healthy response to inference request in 3.0182905197143555s
Received healthy response to inference request in 2.1925621032714844s
5 requests
0 failed requests
5th percentile: 1.7436140060424805
10th percentile: 1.8558510303497315
20th percentile: 2.0803250789642336
30th percentile: 2.2406726837158204
40th percentile: 2.336893844604492
50th percentile: 2.433115005493164
60th percentile: 2.6671852111816405
70th percentile: 2.901255416870117
80th percentile: 3.3726108551025393
90th percentile: 4.081251525878907
95th percentile: 4.435571861267089
99th percentile: 4.719028129577636
mean time: 2.8130473613739015
Pipeline stage StressChecker completed in 15.14s
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 5.79s
Shutdown handler de-registered
rica40325-mistral-4904_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 0.11s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service rica40325-mistral-4904-v1-profiler
Waiting for inference service rica40325-mistral-4904-v1-profiler to be ready
Inference service rica40325-mistral-4904-v1-profiler ready after 220.50324773788452s
Pipeline stage MKMLProfilerDeployer completed in 220.90s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/rica40325-mistral-4904-v1-profiler-predictor-00001-deploymldhz4:/code/chaiverse_profiler_1727257956 --namespace tenant-chaiml-guanaco
kubectl exec -it rica40325-mistral-4904-v1-profiler-predictor-00001-deploymldhz4 --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1727257956 && 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 1024 --output_tokens 64 --summary /code/chaiverse_profiler_1727257956/summary.json'
kubectl exec -it rica40325-mistral-4904-v1-profiler-predictor-00001-deploymldhz4 --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1727257956/summary.json'
Pipeline stage MKMLProfilerRunner completed in 1166.52s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service rica40325-mistral-4904-v1-profiler is running
Tearing down inference service rica40325-mistral-4904-v1-profiler
Service rica40325-mistral-4904-v1-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 2.23s
Shutdown handler de-registered
rica40325-mistral-4904_v1 status is now inactive due to auto deactivation removed underperforming models
rica40325-mistral-4904_v1 status is now torndown due to DeploymentManager action