submission_id: hastagaras-aelah-8b-llam_5487_v2
developer_uid: Hastagaras
status: inactive
model_repo: Hastagaras/Aelah-8B-Llama3-TEST
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 1.05, 'top_p': 0.98, 'min_p': 0.05, 'top_k': 80, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n'], 'max_input_tokens': 512, 'best_of': 16, '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: {'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: 2024-05-25T23:10:55+00:00
model_name: debug
model_eval_status: success
model_group: Hastagaras/Aelah-8B-Llam
num_battles: 9630
num_wins: 5080
celo_rating: 1198.42
safety_score: 0.99
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: debug
ineligible_reason: propriety_total_count < 5000
language_model: Hastagaras/Aelah-8B-Llama3-TEST
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-05-25
win_ratio: 0.5275181723779855
Resubmit model
Running pipeline stage MKMLizer
Starting job with name hastagaras-aelah-8b-llam-5487-v2-mkmlizer
Waiting for job on hastagaras-aelah-8b-llam-5487-v2-mkmlizer to finish
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ _____ __ __ ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ /___/ ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ Version: 0.8.14 ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ https://mk1.ai ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ The license key for the current software has been verified as ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ belonging to: ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ Chai Research Corp. ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ║ ║
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
hastagaras-aelah-8b-llam-5487-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.
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: warnings.warn(warning_message, FutureWarning)
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Downloaded to shared memory in 15.865s
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: quantizing model to /dev/shm/model_cache
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 4%|▍ | 13/291 [00:00<00:02, 119.66it/s] Loading 0: 11%|█ | 31/291 [00:00<00:01, 144.59it/s] Loading 0: 17%|█▋ | 49/291 [00:00<00:01, 150.96it/s] Loading 0: 23%|██▎ | 67/291 [00:00<00:01, 154.18it/s] Loading 0: 29%|██▊ | 83/291 [00:00<00:02, 80.34it/s] Loading 0: 33%|███▎ | 95/291 [00:00<00:02, 87.77it/s] Loading 0: 38%|███▊ | 112/291 [00:01<00:01, 105.32it/s] Loading 0: 45%|████▍ | 130/291 [00:01<00:01, 119.64it/s] Loading 0: 51%|█████ | 148/291 [00:01<00:01, 131.12it/s] Loading 0: 57%|█████▋ | 166/291 [00:01<00:00, 139.05it/s] Loading 0: 63%|██████▎ | 183/291 [00:01<00:00, 146.87it/s] Loading 0: 68%|██████▊ | 199/291 [00:01<00:01, 87.20it/s] Loading 0: 74%|███████▍ | 215/291 [00:01<00:00, 100.00it/s] Loading 0: 79%|███████▉ | 230/291 [00:02<00:00, 109.61it/s] Loading 0: 85%|████████▌ | 248/291 [00:02<00:00, 123.09it/s] Loading 0: 91%|█████████▏| 266/291 [00:02<00:00, 133.68it/s] Loading 0: 98%|█████████▊| 285/291 [00:02<00:00, 146.71it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: quantized model in 23.152s
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Processed model Hastagaras/Aelah-8B-Llama3-TEST in 41.607s
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: creating bucket guanaco-mkml-models
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/hastagaras-aelah-8b-llam-5487-v2
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/hastagaras-aelah-8b-llam-5487-v2/tokenizer_config.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/hastagaras-aelah-8b-llam-5487-v2/special_tokens_map.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/hastagaras-aelah-8b-llam-5487-v2/config.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/hastagaras-aelah-8b-llam-5487-v2/tokenizer.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/hastagaras-aelah-8b-llam-5487-v2/flywheel_model.0.safetensors
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
hastagaras-aelah-8b-llam-5487-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.
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: warnings.warn(
hastagaras-aelah-8b-llam-5487-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.
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: warnings.warn(
hastagaras-aelah-8b-llam-5487-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.
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: warnings.warn(
hastagaras-aelah-8b-llam-5487-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()
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: return self.fget.__get__(instance, owner)()
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Saving duration: 0.422s
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 4.083s
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: creating bucket guanaco-reward-models
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: Bucket 's3://guanaco-reward-models/' created
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/hastagaras-aelah-8b-llam-5487-v2_reward
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/hastagaras-aelah-8b-llam-5487-v2_reward/tokenizer_config.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/hastagaras-aelah-8b-llam-5487-v2_reward/config.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/hastagaras-aelah-8b-llam-5487-v2_reward/special_tokens_map.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/hastagaras-aelah-8b-llam-5487-v2_reward/vocab.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/hastagaras-aelah-8b-llam-5487-v2_reward/tokenizer.json
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/hastagaras-aelah-8b-llam-5487-v2_reward/merges.txt
hastagaras-aelah-8b-llam-5487-v2-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/hastagaras-aelah-8b-llam-5487-v2_reward/reward.tensors
Job hastagaras-aelah-8b-llam-5487-v2-mkmlizer completed after 62.59s with status: succeeded
Stopping job with name hastagaras-aelah-8b-llam-5487-v2-mkmlizer
Pipeline stage MKMLizer completed in 66.81s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.09s
Running pipeline stage ISVCDeployer
Creating inference service hastagaras-aelah-8b-llam-5487-v2
Waiting for inference service hastagaras-aelah-8b-llam-5487-v2 to be ready
Inference service hastagaras-aelah-8b-llam-5487-v2 ready after 30.162514448165894s
Pipeline stage ISVCDeployer completed in 37.74s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.2799124717712402s
%s, retrying in %s seconds...
Received healthy response to inference request in 1.4553894996643066s
Received healthy response to inference request in 1.3496217727661133s
Received healthy response to inference request in 2.3771703243255615s
Received healthy response to inference request in 1.2841057777404785s
Received healthy response to inference request in 1.232064962387085s
5 requests
0 failed requests
5th percentile: 1.2424731254577637
10th percentile: 1.2528812885284424
20th percentile: 1.2736976146697998
30th percentile: 1.2972089767456054
40th percentile: 1.3234153747558595
50th percentile: 1.3496217727661133
60th percentile: 1.3919288635253906
70th percentile: 1.4342359542846679
80th percentile: 1.6397456645965578
90th percentile: 2.0084579944610597
95th percentile: 2.1928141593933104
99th percentile: 2.3402990913391113
mean time: 1.539670467376709
Pipeline stage StressChecker completed in 30.69s
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
hastagaras-aelah-8b-llam_5487_v2 status is now deployed due to DeploymentManager action
hastagaras-aelah-8b-llam_5487_v2 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics