r136a1-midnight-2x8b-l3

developer_uid: R136a1

submission_id: r136a1-midnight-2x8b-l3_v1

model_name: r136a1-midnight-2x8b-l3_v1

model_group: R136a1/MidNight-2x8B-L3

status: torndown

timestamp: 2024-06-25T07:15:56+00:00

num_battles: 11047

num_wins: 5954

celo_rating: 1206.65

family_friendly_score: 0.0

submission_type: basic

model_repo: R136a1/MidNight-2x8B-L3

model_architecture: MixtralForCausalLM

reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2

model_num_parameters: 13667667968.0

best_of: 8

max_input_tokens: 512

max_output_tokens: 64

display_name: r136a1-midnight-2x8b-l3_v1

is_internal_developer: False

language_model: R136a1/MidNight-2x8B-L3

model_size: 14B

ranking_group: single

us_pacific_date: 2024-06-25

win_ratio: 0.5389698560695212

generation_params: {'temperature': 1.15, 'top_p': 1.0, 'min_p': 0.075, 'top_k': 70, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n'], 'max_input_tokens': 512, 'best_of': 8, 'max_output_tokens': 64}

formatter: {'memory_template': "<|start_header_id|>system<|end_header_id|>\n\n{bot_name}'s Persona: {memory}\n\n", 'prompt_template': '{prompt}<|eot_id|>', 'bot_template': '<|start_header_id|>assistant<|end_header_id|>\n\n{bot_name}: {message}<|eot_id|>', 'user_template': '<|start_header_id|>user<|end_header_id|>\n\n{user_name}: {message}<|eot_id|>', 'response_template': '<|start_header_id|>assistant<|end_header_id|>\n\n{bot_name}:', 'truncate_by_message': False}

reward_formatter: {'bot_template': '{bot_name}: {message}\n', 'memory_template': "{bot_name}'s Persona: {memory}\n####\n", 'prompt_template': '{prompt}\n<START>\n', 'response_template': '{bot_name}:', 'truncate_by_message': False, 'user_template': '{user_name}: {message}\n'}

Resubmit model

Running pipeline stage MKMLizer
Starting job with name r136a1-midnight-2x8b-l3-v1-mkmlizer
Waiting for job on r136a1-midnight-2x8b-l3-v1-mkmlizer to finish
r136a1-midnight-2x8b-l3-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║     _____            __           __                                ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║       /___/                                                         ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║                                                                     ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  Version: 0.8.14                                                    ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  https://mk1.ai                                                     ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║                                                                     ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  belonging to:                                                      ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║                                                                     ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  Chai Research Corp.                                                ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║  Expiration: 2024-07-15 23:59:59                                    ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ║                                                                     ║
r136a1-midnight-2x8b-l3-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
r136a1-midnight-2x8b-l3-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/huggingface_hub/utils/_deprecation.py:131: FutureWarning: 'list_files_info' (from 'huggingface_hub.hf_api') is deprecated and will be removed from version '0.23'. Use `list_repo_tree` and `get_paths_info` instead.
r136a1-midnight-2x8b-l3-v1-mkmlizer:   warnings.warn(warning_message, FutureWarning)
r136a1-midnight-2x8b-l3-v1-mkmlizer: Downloaded to shared memory in 50.290s
r136a1-midnight-2x8b-l3-v1-mkmlizer: quantizing model to /dev/shm/model_cache
r136a1-midnight-2x8b-l3-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
r136a1-midnight-2x8b-l3-v1-mkmlizer: 
Loading 0:   0%|          | 0/419 [00:00<?, ?it/s]
Loading 0:   1%|▏         | 6/419 [00:00<00:07, 54.00it/s]
Loading 0:   4%|▍         | 16/419 [00:00<00:05, 76.79it/s]
Loading 0:   6%|▋         | 27/419 [00:00<00:04, 85.04it/s]
Loading 0:   9%|▉         | 38/419 [00:00<00:04, 88.54it/s]
Loading 0:  11%|█         | 47/419 [00:00<00:04, 83.50it/s]
Loading 0:  13%|█▎        | 56/419 [00:00<00:04, 77.53it/s]
Loading 0:  15%|█▌        | 64/419 [00:01<00:07, 47.25it/s]
Loading 0:  17%|█▋        | 73/419 [00:01<00:06, 55.37it/s]
Loading 0:  20%|██        | 84/419 [00:01<00:05, 65.84it/s]
Loading 0:  22%|██▏       | 93/419 [00:01<00:04, 71.47it/s]
Loading 0:  24%|██▍       | 102/419 [00:01<00:04, 73.37it/s]
Loading 0:  26%|██▋       | 111/419 [00:01<00:04, 67.57it/s]
Loading 0:  29%|██▉       | 121/419 [00:01<00:04, 72.42it/s]
Loading 0:  32%|███▏      | 132/419 [00:01<00:03, 77.81it/s]
Loading 0:  34%|███▎      | 141/419 [00:02<00:05, 51.25it/s]
Loading 0:  35%|███▌      | 148/419 [00:02<00:05, 50.40it/s]
Loading 0:  38%|███▊      | 159/419 [00:02<00:04, 59.95it/s]
Loading 0:  40%|███▉      | 166/419 [00:02<00:04, 61.97it/s]
Loading 0:  42%|████▏     | 174/419 [00:02<00:03, 63.57it/s]
Loading 0:  44%|████▍     | 184/419 [00:02<00:03, 72.19it/s]
Loading 0:  47%|████▋     | 195/419 [00:02<00:02, 78.27it/s]
Loading 0:  49%|████▉     | 206/419 [00:03<00:02, 82.56it/s]
Loading 0:  51%|█████▏    | 215/419 [00:03<00:03, 53.01it/s]
Loading 0:  53%|█████▎    | 222/419 [00:03<00:03, 51.61it/s]
Loading 0:  55%|█████▌    | 232/419 [00:03<00:03, 59.61it/s]
Loading 0:  58%|█████▊    | 243/419 [00:03<00:02, 68.05it/s]
Loading 0:  60%|██████    | 253/419 [00:03<00:02, 75.28it/s]
Loading 0:  63%|██████▎   | 262/419 [00:03<00:02, 71.97it/s]
Loading 0:  64%|██████▍   | 270/419 [00:04<00:02, 70.89it/s]
Loading 0:  67%|██████▋   | 281/419 [00:04<00:01, 77.38it/s]
Loading 0:  69%|██████▉   | 290/419 [00:04<00:02, 51.71it/s]
Loading 0:  71%|███████   | 297/419 [00:04<00:02, 53.44it/s]
Loading 0:  73%|███████▎  | 306/419 [00:04<00:01, 59.32it/s]
Loading 0:  76%|███████▌  | 317/419 [00:04<00:01, 67.81it/s]
Loading 0:  78%|███████▊  | 327/419 [00:04<00:01, 72.17it/s]
Loading 0:  81%|████████  | 338/419 [00:05<00:01, 77.40it/s]
Loading 0:  83%|████████▎ | 347/419 [00:05<00:00, 77.58it/s]
Loading 0:  85%|████████▍ | 356/419 [00:05<00:00, 73.15it/s]
Loading 0:  88%|████████▊ | 367/419 [00:20<00:00, 73.15it/s]
Loading 0:  88%|████████▊ | 368/419 [00:26<00:31,  1.62it/s]
Loading 0:  95%|█████████▍| 397/419 [00:26<00:06,  3.55it/s]
Loading 0:  98%|█████████▊| 410/419 [00:26<00:01,  4.73it/s]
                                                            
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
r136a1-midnight-2x8b-l3-v1-mkmlizer: quantized model in 33.640s
r136a1-midnight-2x8b-l3-v1-mkmlizer: Processed model R136a1/MidNight-2x8B-L3 in 88.161s
r136a1-midnight-2x8b-l3-v1-mkmlizer: creating bucket guanaco-mkml-models
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.1.safetensors s3://guanaco-mkml-models/r136a1-midnight-2x8b-l3-v1/flywheel_model.1.safetensors
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/r136a1-midnight-2x8b-l3-v1/flywheel_model.0.safetensors
r136a1-midnight-2x8b-l3-v1-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
r136a1-midnight-2x8b-l3-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py:913: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
r136a1-midnight-2x8b-l3-v1-mkmlizer:   warnings.warn(
r136a1-midnight-2x8b-l3-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/tokenization_auto.py:757: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
r136a1-midnight-2x8b-l3-v1-mkmlizer:   warnings.warn(
r136a1-midnight-2x8b-l3-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py:468: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
r136a1-midnight-2x8b-l3-v1-mkmlizer:   warnings.warn(
r136a1-midnight-2x8b-l3-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
r136a1-midnight-2x8b-l3-v1-mkmlizer:   return self.fget.__get__(instance, owner)()
r136a1-midnight-2x8b-l3-v1-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
r136a1-midnight-2x8b-l3-v1-mkmlizer: Saving duration: 0.401s
r136a1-midnight-2x8b-l3-v1-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 7.306s
r136a1-midnight-2x8b-l3-v1-mkmlizer: creating bucket guanaco-reward-models
r136a1-midnight-2x8b-l3-v1-mkmlizer: Bucket 's3://guanaco-reward-models/' created
r136a1-midnight-2x8b-l3-v1-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/r136a1-midnight-2x8b-l3-v1_reward
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/r136a1-midnight-2x8b-l3-v1_reward/special_tokens_map.json
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/r136a1-midnight-2x8b-l3-v1_reward/config.json
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/r136a1-midnight-2x8b-l3-v1_reward/merges.txt
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/r136a1-midnight-2x8b-l3-v1_reward/tokenizer_config.json
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/r136a1-midnight-2x8b-l3-v1_reward/vocab.json
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/r136a1-midnight-2x8b-l3-v1_reward/tokenizer.json
r136a1-midnight-2x8b-l3-v1-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/r136a1-midnight-2x8b-l3-v1_reward/reward.tensors
Job r136a1-midnight-2x8b-l3-v1-mkmlizer completed after 124.36s with status: succeeded
Stopping job with name r136a1-midnight-2x8b-l3-v1-mkmlizer
Pipeline stage MKMLizer completed in 124.91s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.15s
Running pipeline stage ISVCDeployer
Creating inference service r136a1-midnight-2x8b-l3-v1
Waiting for inference service r136a1-midnight-2x8b-l3-v1 to be ready
Inference service r136a1-midnight-2x8b-l3-v1 ready after 50.219841957092285s
Pipeline stage ISVCDeployer completed in 56.06s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.5690386295318604s
Received healthy response to inference request in 2.1314640045166016s
Received healthy response to inference request in 2.133295774459839s
Received healthy response to inference request in 2.1259958744049072s
Received healthy response to inference request in 2.077901840209961s
5 requests
0 failed requests
5th percentile: 2.08752064704895
10th percentile: 2.0971394538879395
20th percentile: 2.116377067565918
30th percentile: 2.127089500427246
40th percentile: 2.129276752471924
50th percentile: 2.1314640045166016
60th percentile: 2.1321967124938963
70th percentile: 2.1329294204711915
80th percentile: 2.220444345474243
90th percentile: 2.394741487503052
95th percentile: 2.481890058517456
99th percentile: 2.5516089153289796
mean time: 2.2075392246246337
Pipeline stage StressChecker completed in 11.91s
Running pipeline stage DaemonicSafetyScorer
Pipeline stage DaemonicSafetyScorer completed in 0.04s
r136a1-midnight-2x8b-l3_v1 status is now deployed due to DeploymentManager action
r136a1-midnight-2x8b-l3_v1 status is now inactive due to auto deactivation removed underperforming models
admin requested tearing down of r136a1-midnight-2x8b-l3_v1
Running pipeline stage ISVCDeleter
Checking if service r136a1-midnight-2x8b-l3-v1 is running
Skipping teardown as no inference service was found
Pipeline stage ISVCDeleter completed in 4.56s
Running pipeline stage MKMLModelDeleter
Cleaning model data from S3
Cleaning model data from model cache
Deleting key r136a1-midnight-2x8b-l3-v1/config.json from bucket guanaco-mkml-models
Deleting key r136a1-midnight-2x8b-l3-v1/flywheel_model.0.safetensors from bucket guanaco-mkml-models
Deleting key r136a1-midnight-2x8b-l3-v1/flywheel_model.1.safetensors from bucket guanaco-mkml-models
Deleting key r136a1-midnight-2x8b-l3-v1/special_tokens_map.json from bucket guanaco-mkml-models
Deleting key r136a1-midnight-2x8b-l3-v1/tokenizer.json from bucket guanaco-mkml-models
Deleting key r136a1-midnight-2x8b-l3-v1/tokenizer_config.json from bucket guanaco-mkml-models
Cleaning model data from model cache
Deleting key r136a1-midnight-2x8b-l3-v1_reward/config.json from bucket guanaco-reward-models
Deleting key r136a1-midnight-2x8b-l3-v1_reward/merges.txt from bucket guanaco-reward-models
Deleting key r136a1-midnight-2x8b-l3-v1_reward/reward.tensors from bucket guanaco-reward-models
Deleting key r136a1-midnight-2x8b-l3-v1_reward/special_tokens_map.json from bucket guanaco-reward-models
Deleting key r136a1-midnight-2x8b-l3-v1_reward/tokenizer.json from bucket guanaco-reward-models
Deleting key r136a1-midnight-2x8b-l3-v1_reward/tokenizer_config.json from bucket guanaco-reward-models
Deleting key r136a1-midnight-2x8b-l3-v1_reward/vocab.json from bucket guanaco-reward-models
Pipeline stage MKMLModelDeleter completed in 6.64s
r136a1-midnight-2x8b-l3_v1 status is now torndown due to DeploymentManager action