submission_id: neversleep-noromaid-v0-1_8068_v1
developer_uid: robert_irvine
status: deployed
model_repo: NeverSleep/Noromaid-v0.1-mixtral-8x7b-Instruct-v3
reward_repo: rirv938/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 1.0, 'top_p': 1.0, 'top_k': 50, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['</s>', '<|user|>', '###', '\n'], 'max_input_tokens': 512, 'best_of': 4, 'max_output_tokens': 64}
formatter: {'memory_template': '<s>[INST] This is an entertaining conversation. You are {bot_name} who has the persona: {memory}.\nPlay the role of {bot_name}. Engage in a chat with {user_name} while staying in character. You should create a fun dialogue which entertains {user_name}.\n', 'prompt_template': '{prompt}\n', 'bot_template': '{bot_name}: {message}</s>', 'user_template': '[INST] {user_name}: {message} [/INST]', 'response_template': '{bot_name}:'}
reward_formatter: {'memory_template': 'Memory: {memory}\n', 'prompt_template': '{prompt}\n', 'bot_template': 'Bot: {message}\n', 'user_template': 'User: {message}\n', 'response_template': 'Bot:'}
timestamp: 2024-03-19T18:12:44+00:00
model_name: neversleep-noromaid-v0-1_8068_v1
model_eval_status: pending
safety_score: None
entertaining: None
stay_in_character: None
user_preference: None
double_thumbs_up: 5737
thumbs_up: 7712
thumbs_down: 2982
num_battles: 512446
num_wins: 277838
win_ratio: 0.5421800540935044
celo_rating: 1182.61
Resubmit model
Running pipeline stage MKMLizer
Starting job with name neversleep-noromaid-v0-1-8068-v1-mkmlizer
Waiting for job on neversleep-noromaid-v0-1-8068-v1-mkmlizer to finish
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ _____ __ __ ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ /___/ ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ Version: 0.8.6 ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ The license key for the current software has been verified as ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ belonging to: ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ Chai Research Corp. ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ Expiration: 2024-04-15 23:59:59 ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ║ ║
neversleep-noromaid-v0-1-8068-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
neversleep-noromaid-v0-1-8068-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.
neversleep-noromaid-v0-1-8068-v1-mkmlizer: warnings.warn(warning_message, FutureWarning)
neversleep-noromaid-v0-1-8068-v1-mkmlizer: Downloaded to shared memory in 232.106s
neversleep-noromaid-v0-1-8068-v1-mkmlizer: quantizing model to /dev/shm/model_cache
neversleep-noromaid-v0-1-8068-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
neversleep-noromaid-v0-1-8068-v1-mkmlizer: quantized model in 60.627s
neversleep-noromaid-v0-1-8068-v1-mkmlizer: Processed model NeverSleep/Noromaid-v0.1-mixtral-8x7b-Instruct-v3 in 298.922s
neversleep-noromaid-v0-1-8068-v1-mkmlizer: creating bucket guanaco-mkml-models
neversleep-noromaid-v0-1-8068-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
neversleep-noromaid-v0-1-8068-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/neversleep-noromaid-v0-1-8068-v1
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/neversleep-noromaid-v0-1-8068-v1/special_tokens_map.json
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/neversleep-noromaid-v0-1-8068-v1/tokenizer_config.json
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.model s3://guanaco-mkml-models/neversleep-noromaid-v0-1-8068-v1/tokenizer.model
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/neversleep-noromaid-v0-1-8068-v1/config.json
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/neversleep-noromaid-v0-1-8068-v1/tokenizer.json
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.3.safetensors s3://guanaco-mkml-models/neversleep-noromaid-v0-1-8068-v1/flywheel_model.3.safetensors
neversleep-noromaid-v0-1-8068-v1-mkmlizer: loading reward model from rirv938/reward_gpt2_medium_preference_24m_e2
neversleep-noromaid-v0-1-8068-v1-mkmlizer: Loading 0: 0%| | 0/995 [00:00<?, ?it/s] Loading 0: 5%|▌ | 52/995 [00:01<00:19, 47.36it/s] Loading 0: 11%|█ | 107/995 [00:02<00:17, 49.90it/s] Loading 0: 16%|█▋ | 162/995 [00:03<00:16, 51.50it/s] Loading 0: 21%|██ | 210/995 [00:04<00:15, 49.54it/s] Loading 0: 27%|██▋ | 265/995 [00:05<00:14, 51.08it/s] Loading 0: 28%|██▊ | 278/995 [00:15<01:18, 9.08it/s] Loading 0: 32%|███▏ | 320/995 [00:17<00:55, 12.09it/s] Loading 0: 37%|███▋ | 368/995 [00:18<00:38, 16.45it/s] Loading 0: 43%|████▎ | 423/995 [00:19<00:25, 22.09it/s] Loading 0: 48%|████▊ | 478/995 [00:20<00:18, 27.87it/s] Loading 0: 53%|█████▎ | 526/995 [00:21<00:14, 31.79it/s] Loading 0: 57%|█████▋ | 564/995 [00:32<00:13, 31.79it/s] Loading 0: 57%|█████▋ | 565/995 [00:32<00:40, 10.55it/s] Loading 0: 58%|█████▊ | 581/995 [00:33<00:38, 10.81it/s] Loading 0: 64%|██████▍ | 636/995 [00:34<00:22, 15.84it/s] Loading 0: 69%|██████▉ | 691/995 [00:35<00:14, 21.23it/s] Loading 0: 74%|███████▍ | 739/995 [00:36<00:10, 25.55it/s] Loading 0: 80%|███████▉ | 794/995 [00:37<00:06, 31.33it/s] Loading 0: 85%|████████▌ | 846/995 [00:48<00:13, 11.31it/s] Loading 0: 85%|████████▌ | 849/995 [00:49<00:14, 10.39it/s] Loading 0: 90%|█████████ | 897/995 [00:50<00:06, 14.64it/s] Loading 0: 96%|█████████▌| 952/995 [00:51<00:02, 20.70it/s] Loading 0: 100%|██████████| 995/995 [00:52<00:00, 23.78it/s] /opt/conda/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py:1096: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
neversleep-noromaid-v0-1-8068-v1-mkmlizer: warnings.warn(
neversleep-noromaid-v0-1-8068-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/tokenization_auto.py:720: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
neversleep-noromaid-v0-1-8068-v1-mkmlizer: warnings.warn(
neversleep-noromaid-v0-1-8068-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py:466: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
neversleep-noromaid-v0-1-8068-v1-mkmlizer: warnings.warn(
neversleep-noromaid-v0-1-8068-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()
neversleep-noromaid-v0-1-8068-v1-mkmlizer: return self.fget.__get__(instance, owner)()
neversleep-noromaid-v0-1-8068-v1-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
neversleep-noromaid-v0-1-8068-v1-mkmlizer: Saving duration: 0.336s
neversleep-noromaid-v0-1-8068-v1-mkmlizer: Processed model rirv938/reward_gpt2_medium_preference_24m_e2 in 7.003s
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/neversleep-noromaid-v0-1-8068-v1_reward/merges.txt
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/neversleep-noromaid-v0-1-8068-v1_reward/vocab.json
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/neversleep-noromaid-v0-1-8068-v1_reward/tokenizer.json
neversleep-noromaid-v0-1-8068-v1-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/neversleep-noromaid-v0-1-8068-v1_reward/reward.tensors
Job neversleep-noromaid-v0-1-8068-v1-mkmlizer completed after 345.32s with status: succeeded
Stopping job with name neversleep-noromaid-v0-1-8068-v1-mkmlizer
Pipeline stage MKMLizer completed in 346.50s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.28s
Running pipeline stage ISVCDeployer
Creating inference service neversleep-noromaid-v0-1-8068-v1
Waiting for inference service neversleep-noromaid-v0-1-8068-v1 to be ready
Inference service neversleep-noromaid-v0-1-8068-v1 ready after 50.796241760253906s
Pipeline stage ISVCDeployer completed in 56.95s
Running pipeline stage StressChecker
Received healthy response to inference request in 5.808049917221069s
Received healthy response to inference request in 2.7533936500549316s
Received healthy response to inference request in 2.4673478603363037s
Received healthy response to inference request in 2.5207037925720215s
Received healthy response to inference request in 3.126802921295166s
5 requests
0 failed requests
5th percentile: 2.478019046783447
10th percentile: 2.488690233230591
20th percentile: 2.510032606124878
30th percentile: 2.5672417640686036
40th percentile: 2.6603177070617674
50th percentile: 2.7533936500549316
60th percentile: 2.9027573585510256
70th percentile: 3.052121067047119
80th percentile: 3.663052320480347
90th percentile: 4.735551118850708
95th percentile: 5.271800518035889
99th percentile: 5.700800037384033
mean time: 3.3352596282958986
Pipeline stage StressChecker completed in 19.26s
Running pipeline stage DaemonicModelEvalScorer
Pipeline stage DaemonicModelEvalScorer completed in 0.13s
Running M-Eval for topic stay_in_character
Running pipeline stage DaemonicSafetyScorer
M-Eval Dataset for topic stay_in_character is loaded
Pipeline stage DaemonicSafetyScorer completed in 0.27s
%s, retrying in %s seconds...
%s, retrying in %s seconds...

Usage Metrics

Latency Metrics