submission_id: undi95-redacted-e2_v2
developer_uid: Undi95
status: torndown
model_repo: Undi95/redacted-e2
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 0.8, 'top_p': 0.9, 'min_p': 0.0, 'top_k': 70, 'presence_penalty': 0.9, 'frequency_penalty': 0.9, 'stopping_words': ['\n', '<|eot_id|>'], 'max_input_tokens': 512, 'best_of': 16, '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}
reward_formatter: {'memory_template': "{bot_name}'s Persona: {memory}\n", 'prompt_template': '{prompt}\n', 'bot_template': '{bot_name}: {message}\n', 'user_template': '{user_name}: {message}\n', 'response_template': '{bot_name}:', 'truncate_by_message': False}
timestamp: 2024-04-21T00:05:25+00:00
model_name: undi95-redacted-e2_v1
model_eval_status: success
model_group: Undi95/redacted-e2
num_battles: 5785
num_wins: 3031
celo_rating: 1163.78
propriety_score: 0.0
propriety_total_count: 0.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 16
max_input_tokens: 512
max_output_tokens: 64
display_name: undi95-redacted-e2_v1
ineligible_reason: propriety_total_count < 800
language_model: Undi95/redacted-e2
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-04-20
win_ratio: 0.5239412273120139
preference_data_url: None
Resubmit model
Running pipeline stage MKMLizer
Starting job with name undi95-redacted-e2-v2-mkmlizer
Waiting for job on undi95-redacted-e2-v2-mkmlizer to finish
undi95-redacted-e2-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
undi95-redacted-e2-v2-mkmlizer: ║ _____ __ __ ║
undi95-redacted-e2-v2-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
undi95-redacted-e2-v2-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
undi95-redacted-e2-v2-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
undi95-redacted-e2-v2-mkmlizer: ║ /___/ ║
undi95-redacted-e2-v2-mkmlizer: ║ ║
undi95-redacted-e2-v2-mkmlizer: ║ Version: 0.8.10 ║
undi95-redacted-e2-v2-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
undi95-redacted-e2-v2-mkmlizer: ║ ║
undi95-redacted-e2-v2-mkmlizer: ║ The license key for the current software has been verified as ║
undi95-redacted-e2-v2-mkmlizer: ║ belonging to: ║
undi95-redacted-e2-v2-mkmlizer: ║ ║
undi95-redacted-e2-v2-mkmlizer: ║ Chai Research Corp. ║
undi95-redacted-e2-v2-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
undi95-redacted-e2-v2-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
undi95-redacted-e2-v2-mkmlizer: ║ ║
undi95-redacted-e2-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
undi95-redacted-e2-v2-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.
undi95-redacted-e2-v2-mkmlizer: warnings.warn(warning_message, FutureWarning)
undi95-redacted-e2-v2-mkmlizer: Downloaded to shared memory in 19.015s
undi95-redacted-e2-v2-mkmlizer: quantizing model to /dev/shm/model_cache
undi95-redacted-e2-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
undi95-redacted-e2-v2-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 29%|██▊ | 83/291 [00:01<00:02, 71.95it/s] Loading 0: 64%|██████▍ | 187/291 [00:02<00:01, 74.53it/s] Loading 0: 99%|█████████▊| 287/291 [00:09<00:00, 26.33it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
undi95-redacted-e2-v2-mkmlizer: quantized model in 20.554s
undi95-redacted-e2-v2-mkmlizer: Processed model Undi95/redacted-e2 in 41.169s
undi95-redacted-e2-v2-mkmlizer: creating bucket guanaco-mkml-models
undi95-redacted-e2-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
undi95-redacted-e2-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/undi95-redacted-e2-v2
undi95-redacted-e2-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/undi95-redacted-e2-v2/config.json
undi95-redacted-e2-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/undi95-redacted-e2-v2/tokenizer_config.json
undi95-redacted-e2-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/undi95-redacted-e2-v2/special_tokens_map.json
undi95-redacted-e2-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/undi95-redacted-e2-v2/tokenizer.json
undi95-redacted-e2-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/undi95-redacted-e2-v2/flywheel_model.0.safetensors
undi95-redacted-e2-v2-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
undi95-redacted-e2-v2-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.
undi95-redacted-e2-v2-mkmlizer: warnings.warn(
undi95-redacted-e2-v2-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.
undi95-redacted-e2-v2-mkmlizer: warnings.warn(
undi95-redacted-e2-v2-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.
undi95-redacted-e2-v2-mkmlizer: warnings.warn(
undi95-redacted-e2-v2-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()
undi95-redacted-e2-v2-mkmlizer: return self.fget.__get__(instance, owner)()
undi95-redacted-e2-v2-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
undi95-redacted-e2-v2-mkmlizer: Saving duration: 0.252s
undi95-redacted-e2-v2-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 4.497s
undi95-redacted-e2-v2-mkmlizer: creating bucket guanaco-reward-models
undi95-redacted-e2-v2-mkmlizer: Bucket 's3://guanaco-reward-models/' created
undi95-redacted-e2-v2-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/undi95-redacted-e2-v2_reward
undi95-redacted-e2-v2-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/undi95-redacted-e2-v2_reward/config.json
undi95-redacted-e2-v2-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/undi95-redacted-e2-v2_reward/tokenizer_config.json
undi95-redacted-e2-v2-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/undi95-redacted-e2-v2_reward/merges.txt
undi95-redacted-e2-v2-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/undi95-redacted-e2-v2_reward/vocab.json
undi95-redacted-e2-v2-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/undi95-redacted-e2-v2_reward/special_tokens_map.json
undi95-redacted-e2-v2-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/undi95-redacted-e2-v2_reward/tokenizer.json
undi95-redacted-e2-v2-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/undi95-redacted-e2-v2_reward/reward.tensors
Job undi95-redacted-e2-v2-mkmlizer completed after 63.21s with status: succeeded
Stopping job with name undi95-redacted-e2-v2-mkmlizer
Pipeline stage MKMLizer completed in 66.35s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.09s
Running pipeline stage ISVCDeployer
Creating inference service undi95-redacted-e2-v2
Waiting for inference service undi95-redacted-e2-v2 to be ready
Inference service undi95-redacted-e2-v2 ready after 30.323059558868408s
Pipeline stage ISVCDeployer completed in 37.87s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.062398910522461s
Received healthy response to inference request in 1.2649714946746826s
Received healthy response to inference request in 1.226435661315918s
Received healthy response to inference request in 1.2260537147521973s
Received healthy response to inference request in 1.4069199562072754s
5 requests
0 failed requests
5th percentile: 1.2261301040649415
10th percentile: 1.2262064933776855
20th percentile: 1.2263592720031737
30th percentile: 1.234142827987671
40th percentile: 1.2495571613311767
50th percentile: 1.2649714946746826
60th percentile: 1.3217508792877197
70th percentile: 1.3785302639007568
80th percentile: 1.5380157470703126
90th percentile: 1.800207328796387
95th percentile: 1.9313031196594237
99th percentile: 2.0361797523498537
mean time: 1.4373559474945068
Pipeline stage StressChecker completed in 7.95s
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 0.05s
M-Eval Dataset for topic stay_in_character is loaded
undi95-redacted-e2_v2 status is now deployed due to DeploymentManager action
undi95-redacted-e2_v2 status is now inactive due to auto deactivation removed underperforming models
admin requested tearing down of undi95-redacted-e2_v2
Running pipeline stage ISVCDeleter
Checking if service undi95-redacted-e2-v2 is running
Tearing down inference service undi95-redacted-e2-v2
Toredown service undi95-redacted-e2-v2
Pipeline stage ISVCDeleter completed in 3.45s
Running pipeline stage MKMLModelDeleter
Cleaning model data from S3
Cleaning model data from model cache
Deleting key undi95-redacted-e2-v2/config.json from bucket guanaco-mkml-models
admin requested tearing down of vicgalle-roleplay-llama-3-8b_v13
Deleting key undi95-redacted-e2-v2/flywheel_model.0.safetensors from bucket guanaco-mkml-models
Running pipeline stage ISVCDeleter
Checking if service vicgalle-roleplay-llama-3-8b-v13 is running
Deleting key undi95-redacted-e2-v2/special_tokens_map.json from bucket guanaco-mkml-models
Deleting key undi95-redacted-e2-v2/tokenizer.json from bucket guanaco-mkml-models
Deleting key undi95-redacted-e2-v2/tokenizer_config.json from bucket guanaco-mkml-models
Cleaning model data from model cache
Deleting key undi95-redacted-e2-v2_reward/config.json from bucket guanaco-reward-models
Deleting key undi95-redacted-e2-v2_reward/merges.txt from bucket guanaco-reward-models
Deleting key undi95-redacted-e2-v2_reward/reward.tensors from bucket guanaco-reward-models
Deleting key undi95-redacted-e2-v2_reward/special_tokens_map.json from bucket guanaco-reward-models
Deleting key undi95-redacted-e2-v2_reward/tokenizer.json from bucket guanaco-reward-models
Deleting key undi95-redacted-e2-v2_reward/tokenizer_config.json from bucket guanaco-reward-models
Deleting key undi95-redacted-e2-v2_reward/vocab.json from bucket guanaco-reward-models
Pipeline stage MKMLModelDeleter completed in 2.45s
undi95-redacted-e2_v2 status is now torndown due to DeploymentManager action

Usage Metrics

Latency Metrics