developer_uid: rica40325
submission_id: rica40325-mistral-13b-2452_v2
model_name: rica40325-mistral-13b-2452_v1
model_group: rica40325/mistral-13B-24
status: torndown
timestamp: 2024-09-25T05:41:43+00:00
num_battles: 3900
num_wins: 2030
celo_rating: 1269.06
family_friendly_score: 0.5580305842835676
family_friendly_standard_error: 0.00805459915510909
submission_type: basic
model_repo: rica40325/mistral-13B-2452
model_architecture: MistralForCausalLM
model_num_parameters: 12772070400.0
best_of: 8
max_input_tokens: 1024
max_output_tokens: 64
latencies: [{'batch_size': 1, 'throughput': 0.6152576451238008, 'latency_mean': 1.6252671027183532, 'latency_p50': 1.6342746019363403, 'latency_p90': 1.7943316221237182}, {'batch_size': 3, 'throughput': 1.077592433837216, 'latency_mean': 2.7786121559143067, 'latency_p50': 2.8014743328094482, 'latency_p90': 3.1169225692749025}, {'batch_size': 5, 'throughput': 1.213310187806746, 'latency_mean': 4.10235038638115, 'latency_p50': 4.127099871635437, 'latency_p90': 4.627221155166626}, {'batch_size': 6, 'throughput': 1.2320949594240742, 'latency_mean': 4.84190122961998, 'latency_p50': 4.8809837102890015, 'latency_p90': 5.411670804023743}, {'batch_size': 8, 'throughput': 1.2149865896522578, 'latency_mean': 6.557476387023926, 'latency_p50': 6.605907559394836, 'latency_p90': 7.424200892448425}, {'batch_size': 10, 'throughput': 1.1780743669162383, 'latency_mean': 8.435635545253753, 'latency_p50': 8.48669159412384, 'latency_p90': 9.567969179153442}]
gpu_counts: {'NVIDIA RTX A5000': 1}
display_name: rica40325-mistral-13b-2452_v1
ineligible_reason: num_battles<5000
is_internal_developer: False
language_model: rica40325/mistral-13B-2452
model_size: 13B
ranking_group: single
throughput_3p7s: 1.19
us_pacific_date: 2024-09-24
win_ratio: 0.5205128205128206
generation_params: {'temperature': 0.95, 'top_p': 0.95, 'min_p': 0.0, 'top_k': 40, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n'], 'max_input_tokens': 1024, '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 rica40325-mistral-13b-2452-v2-mkmlizer
Waiting for job on rica40325-mistral-13b-2452-v2-mkmlizer to finish
rica40325-mistral-13b-2452-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
rica40325-mistral-13b-2452-v2-mkmlizer: ║     _____            __           __                                ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║       /___/                                                         ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║                                                                     ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  Version: 0.11.12                                                   ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  https://mk1.ai                                                     ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║                                                                     ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  The license key for the current software has been verified as      ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  belonging to:                                                      ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║                                                                     ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  Chai Research Corp.                                                ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
rica40325-mistral-13b-2452-v2-mkmlizer: ║                                                                     ║
rica40325-mistral-13b-2452-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
Connection pool is full, discarding connection: %s. Connection pool size: %s
rica40325-mistral-13b-2452-v2-mkmlizer: Downloaded to shared memory in 47.339s
rica40325-mistral-13b-2452-v2-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmpy6t4t8lz, device:0
rica40325-mistral-13b-2452-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
rica40325-mistral-13b-2452-v2-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-13b-2452-v2-mkmlizer:   tensors = torch.load(model_shard_filename, map_location=torch.device(self.device), mmap=True)
rica40325-mistral-13b-2452-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
rica40325-mistral-13b-2452-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/rica40325-mistral-13b-2452-v2
rica40325-mistral-13b-2452-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/rica40325-mistral-13b-2452-v2/config.json
rica40325-mistral-13b-2452-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/rica40325-mistral-13b-2452-v2/special_tokens_map.json
rica40325-mistral-13b-2452-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/rica40325-mistral-13b-2452-v2/tokenizer_config.json
rica40325-mistral-13b-2452-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/rica40325-mistral-13b-2452-v2/tokenizer.json
Failed to get response for submission function_gosot_2024-09-25: ('http://meta-llama-meta-llama-3-6070-v6-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', '')
Job rica40325-mistral-13b-2452-v2-mkmlizer completed after 117.34s with status: succeeded
Stopping job with name rica40325-mistral-13b-2452-v2-mkmlizer
Pipeline stage MKMLizer completed in 118.42s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.44s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service rica40325-mistral-13b-2452-v2
Waiting for inference service rica40325-mistral-13b-2452-v2 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
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-13b-2452-v2 ready after 212.30827617645264s
Pipeline stage MKMLDeployer completed in 212.84s
run pipeline stage %s
Running pipeline stage StressChecker
Failed to get response for submission function_gosot_2024-09-25: ('http://meta-llama-meta-llama-3-6070-v6-predictor.tenant-chaiml-guanaco.k.chaiverse.com/v1/models/GPT-J-6B-lit-v2:predict', '')
Received healthy response to inference request in 4.175175428390503s
Received healthy response to inference request in 1.9761662483215332s
Received healthy response to inference request in 2.3980464935302734s
Received healthy response to inference request in 2.2611582279205322s
Received healthy response to inference request in 2.1058714389801025s
5 requests
0 failed requests
5th percentile: 2.002107286453247
10th percentile: 2.0280483245849608
20th percentile: 2.0799304008483888
30th percentile: 2.1369287967681885
40th percentile: 2.1990435123443604
Connection pool is full, discarding connection: %s. Connection pool size: %s
50th percentile: 2.2611582279205322
60th percentile: 2.315913534164429
70th percentile: 2.3706688404083254
80th percentile: 2.7534722805023195
90th percentile: 3.464323854446411
95th percentile: 3.819749641418457
99th percentile: 4.104090270996093
mean time: 2.583283567428589
Pipeline stage StressChecker completed in 14.49s
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 4.94s
Shutdown handler de-registered
rica40325-mistral-13b-2452_v2 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 0.23s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeployer
Creating inference service rica40325-mistral-13b-2452-v2-profiler
Waiting for inference service rica40325-mistral-13b-2452-v2-profiler to be ready
Inference service rica40325-mistral-13b-2452-v2-profiler ready after 200.440260887146s
Pipeline stage MKMLProfilerDeployer completed in 200.79s
run pipeline stage %s
Running pipeline stage MKMLProfilerRunner
kubectl cp /code/guanaco/guanaco_inference_services/src/inference_scripts tenant-chaiml-guanaco/rica40325-mistral-13220d46107ea6b2beadc05071beacfd59-deplonr2gs:/code/chaiverse_profiler_1727243504 --namespace tenant-chaiml-guanaco
kubectl exec -it rica40325-mistral-13220d46107ea6b2beadc05071beacfd59-deplonr2gs --namespace tenant-chaiml-guanaco -- sh -c 'cd /code/chaiverse_profiler_1727243504 && 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_1727243504/summary.json'
kubectl exec -it rica40325-mistral-13220d46107ea6b2beadc05071beacfd59-deplonr2gs --namespace tenant-chaiml-guanaco -- bash -c 'cat /code/chaiverse_profiler_1727243504/summary.json'
Pipeline stage MKMLProfilerRunner completed in 1177.50s
run pipeline stage %s
Running pipeline stage MKMLProfilerDeleter
Checking if service rica40325-mistral-13b-2452-v2-profiler is running
Tearing down inference service rica40325-mistral-13b-2452-v2-profiler
Service rica40325-mistral-13b-2452-v2-profiler has been torndown
Pipeline stage MKMLProfilerDeleter completed in 2.24s
Shutdown handler de-registered
rica40325-mistral-13b-2452_v2 status is now inactive due to auto deactivation removed underperforming models
jic062-dpo-v1-8-c500_v1 status is now torndown due to DeploymentManager action
rica40325-mistral-13b-2452_v2 status is now torndown due to DeploymentManager action