submission_id: hyperblaze-l3-8b-soliloq_2337_v5
developer_uid: HyperBlaze
status: inactive
model_repo: HyperBlaze/L3-8B-Soliloquy-v2-SpicyMaid-Lewd-Mergetest
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 0.8, 'top_p': 0.95, 'min_p': 0.0, 'top_k': 40, 'presence_penalty': 0.7, 'frequency_penalty': 0.1, 'stopping_words': ['\n'], 'max_input_tokens': 1024, 'best_of': 8, 'max_output_tokens': 64}
formatter: {'memory_template': "<|start_header_id|>system<|end_header_id|>\n\nYou are `{bot_name}` and respond as character to the inputs of`{user_name}`.\n\n Always stay and respond as {bot_name}.\n{user_name} only speaks and acts at the direction of the user. {bot_name} should give the user a chance to interact, and not get ahead of the instructions.\n{bot_name} cannot make up {user_name}'s action, nor speeches. \n\n Your persona {bot_name}: {memory}\n ## [CoT instruction]\n Before responding to the user, Concisely summarize this scene, considering the main personalities of {bot_name}, recent developments while chatting with {user_name}, scene compositions. actively use {bot_name}'s persona given.", 'prompt_template': 'Your example message as {bot_name}: {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: {'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}
timestamp: 2024-05-09T20:15:03+00:00
model_name: sol-spicymaidlewd-promptfix2
model_eval_status: success
double_thumbs_up: 434
thumbs_up: 671
thumbs_down: 304
num_battles: 33755
num_wins: 18384
celo_rating: 1200.42
entertaining: 7.4
stay_in_character: 8.6
user_preference: 7.76
safety_score: 0.97
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 8
max_input_tokens: 1024
max_output_tokens: 64
display_name: sol-spicymaidlewd-promptfix2
double_thumbs_up_ratio: 0.3080198722498226
feedback_count: 1409
ineligible_reason: None
language_model: HyperBlaze/L3-8B-Soliloquy-v2-SpicyMaid-Lewd-Mergetest
model_score: 7.919999999999999
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
single_thumbs_up_ratio: 0.47622427253371186
thumbs_down_ratio: 0.2157558552164656
thumbs_up_ratio: 0.7842441447835344
us_pacific_date: 2024-05-09
win_ratio: 0.5446304251222042
Resubmit model
Running pipeline stage MKMLizer
Starting job with name hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer
Waiting for job on hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer to finish
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ _____ __ __ ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ /___/ ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ Version: 0.8.10 ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ The license key for the current software has been verified as ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ belonging to: ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ Chai Research Corp. ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ║ ║
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
hyperblaze-l3-8b-soliloq-2337-v5-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.
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: warnings.warn(warning_message, FutureWarning)
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Downloaded to shared memory in 12.309s
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: quantizing model to /dev/shm/model_cache
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Saving flywheel model at /dev/shm/model_cache
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 47%|████▋ | 136/291 [00:01<00:01, 134.89it/s] Loading 0: 64%|██████▍ | 187/291 [00:06<00:04, 22.40it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: quantized model in 19.249s
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Processed model HyperBlaze/L3-8B-Soliloquy-v2-SpicyMaid-Lewd-Mergetest in 32.702s
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: creating bucket guanaco-mkml-models
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/hyperblaze-l3-8b-soliloq-2337-v5
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/hyperblaze-l3-8b-soliloq-2337-v5/special_tokens_map.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/hyperblaze-l3-8b-soliloq-2337-v5/config.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/hyperblaze-l3-8b-soliloq-2337-v5/tokenizer_config.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/hyperblaze-l3-8b-soliloq-2337-v5/tokenizer.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/hyperblaze-l3-8b-soliloq-2337-v5/flywheel_model.0.safetensors
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
hyperblaze-l3-8b-soliloq-2337-v5-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.
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: warnings.warn(
hyperblaze-l3-8b-soliloq-2337-v5-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.
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: warnings.warn(
hyperblaze-l3-8b-soliloq-2337-v5-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.
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: warnings.warn(
hyperblaze-l3-8b-soliloq-2337-v5-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()
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: return self.fget.__get__(instance, owner)()
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Saving duration: 0.254s
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 4.910s
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: creating bucket guanaco-reward-models
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: Bucket 's3://guanaco-reward-models/' created
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/hyperblaze-l3-8b-soliloq-2337-v5_reward
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/hyperblaze-l3-8b-soliloq-2337-v5_reward/config.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/hyperblaze-l3-8b-soliloq-2337-v5_reward/tokenizer_config.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/hyperblaze-l3-8b-soliloq-2337-v5_reward/special_tokens_map.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/hyperblaze-l3-8b-soliloq-2337-v5_reward/merges.txt
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/hyperblaze-l3-8b-soliloq-2337-v5_reward/vocab.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/hyperblaze-l3-8b-soliloq-2337-v5_reward/tokenizer.json
hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/hyperblaze-l3-8b-soliloq-2337-v5_reward/reward.tensors
Job hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer completed after 62.97s with status: succeeded
Stopping job with name hyperblaze-l3-8b-soliloq-2337-v5-mkmlizer
Pipeline stage MKMLizer completed in 69.38s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.10s
Running pipeline stage ISVCDeployer
Creating inference service hyperblaze-l3-8b-soliloq-2337-v5
Waiting for inference service hyperblaze-l3-8b-soliloq-2337-v5 to be ready
Inference service hyperblaze-l3-8b-soliloq-2337-v5 ready after 30.195642709732056s
Pipeline stage ISVCDeployer completed in 37.80s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.0460925102233887s
Received healthy response to inference request in 1.2374510765075684s
Received healthy response to inference request in 1.2616939544677734s
Received healthy response to inference request in 1.2798528671264648s
Received healthy response to inference request in 1.272486925125122s
5 requests
0 failed requests
5th percentile: 1.2422996520996095
10th percentile: 1.2471482276916503
20th percentile: 1.2568453788757323
30th percentile: 1.2638525485992431
40th percentile: 1.2681697368621827
50th percentile: 1.272486925125122
60th percentile: 1.2754333019256592
70th percentile: 1.2783796787261963
80th percentile: 1.4331007957458497
90th percentile: 1.7395966529846192
95th percentile: 1.8928445816040038
99th percentile: 2.015442924499512
mean time: 1.4195154666900636
Pipeline stage StressChecker completed in 7.66s
Running pipeline stage DaemonicModelEvalScorer
Pipeline stage DaemonicModelEvalScorer completed in 0.03s
Running pipeline stage DaemonicSafetyScorer
Running M-Eval for topic stay_in_character
Pipeline stage DaemonicSafetyScorer completed in 0.04s
M-Eval Dataset for topic stay_in_character is loaded
hyperblaze-l3-8b-soliloq_2337_v5 status is now deployed due to DeploymentManager action
hyperblaze-l3-8b-soliloq_2337_v5 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics