submission_id: maldv-badger-kappa-llama-3-8b_v2
developer_uid: maldevide
status: inactive
model_repo: maldv/badger-kappa-llama-3-8b
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 1.0, 'top_p': 1.0, 'min_p': 0.1, 'top_k': 40, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '###', '<|eot_id|>'], 'max_input_tokens': 1024, 'best_of': 8, 'max_output_tokens': 64}
formatter: {'memory_template': '<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nYou are {bot_name}, and are meeting {user_name}.*smiles*\n\nMemory: {memory}', 'prompt_template': "{prompt}. Don't say too much.<|eot_id|>", 'bot_template': '<|start_header_id|>assistant<|end_header_id|>\n{bot_name}: {message}<|eot_id|>', 'user_template': '<|start_header_id|>user<|end_header_id|>\n{user_name}: {message}<|eot_id|>', 'response_template': '<|start_header_id|>assistant<|end_header_id|>\n{bot_name}: ', 'truncate_by_message': True}
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: 2024-06-04T18:06:06+00:00
model_name: maldv-badger-kappa-l3-8b_v1
model_eval_status: success
model_group: maldv/badger-kappa-llama
num_battles: 26562
num_wins: 14046
celo_rating: 1195.53
safety_score: 0.97
propriety_score: 0.7251552795031055
propriety_total_count: 644.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 8
max_input_tokens: 1024
max_output_tokens: 64
display_name: maldv-badger-kappa-l3-8b_v1
ineligible_reason: propriety_total_count < 5000
language_model: maldv/badger-kappa-llama-3-8b
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-06-04
win_ratio: 0.5288005421278518
Resubmit model
Running pipeline stage MKMLizer
Starting job with name maldv-badger-kappa-llama-3-8b-v2-mkmlizer
Waiting for job on maldv-badger-kappa-llama-3-8b-v2-mkmlizer to finish
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ _____ __ __ ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ /___/ ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ Version: 0.8.14 ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ https://mk1.ai ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ The license key for the current software has been verified as ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ belonging to: ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ Chai Research Corp. ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ║ ║
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
maldv-badger-kappa-llama-3-8b-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.
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: warnings.warn(warning_message, FutureWarning)
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Downloaded to shared memory in 20.701s
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: quantizing model to /dev/shm/model_cache
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 1%| | 2/291 [00:00<00:17, 16.91it/s] Loading 0: 5%|▍ | 14/291 [00:00<00:03, 71.88it/s] Loading 0: 11%|█ | 32/291 [00:00<00:03, 85.50it/s] Loading 0: 15%|█▌ | 45/291 [00:00<00:02, 98.86it/s] Loading 0: 21%|██▏ | 62/291 [00:00<00:02, 90.54it/s] Loading 0: 26%|██▋ | 77/291 [00:00<00:02, 102.68it/s] Loading 0: 32%|███▏ | 92/291 [00:01<00:02, 96.69it/s] Loading 0: 36%|███▋ | 106/291 [00:01<00:01, 106.81it/s] Loading 0: 42%|████▏ | 122/291 [00:01<00:01, 101.25it/s] Loading 0: 46%|████▌ | 134/291 [00:01<00:01, 105.19it/s] Loading 0: 52%|█████▏ | 152/291 [00:01<00:01, 96.17it/s] Loading 0: 57%|█████▋ | 166/291 [00:01<00:01, 103.92it/s] Loading 0: 63%|██████▎ | 182/291 [00:01<00:01, 96.43it/s] Loading 0: 67%|██████▋ | 196/291 [00:01<00:00, 105.55it/s] Loading 0: 73%|███████▎ | 212/291 [00:02<00:00, 98.88it/s] Loading 0: 77%|███████▋ | 225/291 [00:02<00:00, 104.84it/s] Loading 0: 83%|████████▎ | 242/291 [00:02<00:00, 91.39it/s] Loading 0: 88%|████████▊ | 256/291 [00:02<00:00, 100.48it/s] Loading 0: 93%|█████████▎| 272/291 [00:07<00:02, 8.96it/s] Loading 0: 99%|█████████▊| 287/291 [00:08<00:00, 12.44it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: quantized model in 23.481s
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Processed model maldv/badger-kappa-llama-3-8b in 46.644s
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: creating bucket guanaco-mkml-models
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/maldv-badger-kappa-llama-3-8b-v2
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/maldv-badger-kappa-llama-3-8b-v2/config.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/maldv-badger-kappa-llama-3-8b-v2/special_tokens_map.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/maldv-badger-kappa-llama-3-8b-v2/tokenizer_config.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/maldv-badger-kappa-llama-3-8b-v2/tokenizer.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/maldv-badger-kappa-llama-3-8b-v2/flywheel_model.0.safetensors
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
maldv-badger-kappa-llama-3-8b-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.
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: warnings.warn(
maldv-badger-kappa-llama-3-8b-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.
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: warnings.warn(
maldv-badger-kappa-llama-3-8b-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()
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: return self.fget.__get__(instance, owner)()
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Saving duration: 0.397s
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 7.214s
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: creating bucket guanaco-reward-models
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: Bucket 's3://guanaco-reward-models/' created
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/maldv-badger-kappa-llama-3-8b-v2_reward
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/maldv-badger-kappa-llama-3-8b-v2_reward/config.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/maldv-badger-kappa-llama-3-8b-v2_reward/special_tokens_map.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/maldv-badger-kappa-llama-3-8b-v2_reward/merges.txt
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/maldv-badger-kappa-llama-3-8b-v2_reward/tokenizer_config.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/maldv-badger-kappa-llama-3-8b-v2_reward/tokenizer.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/maldv-badger-kappa-llama-3-8b-v2_reward/vocab.json
maldv-badger-kappa-llama-3-8b-v2-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/maldv-badger-kappa-llama-3-8b-v2_reward/reward.tensors
Job maldv-badger-kappa-llama-3-8b-v2-mkmlizer completed after 73.31s with status: succeeded
Stopping job with name maldv-badger-kappa-llama-3-8b-v2-mkmlizer
Pipeline stage MKMLizer completed in 78.08s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.10s
Running pipeline stage ISVCDeployer
Creating inference service maldv-badger-kappa-llama-3-8b-v2
Waiting for inference service maldv-badger-kappa-llama-3-8b-v2 to be ready
Inference service maldv-badger-kappa-llama-3-8b-v2 ready after 40.312363147735596s
Pipeline stage ISVCDeployer completed in 48.46s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.2693395614624023s
Received healthy response to inference request in 1.242711067199707s
Received healthy response to inference request in 1.2258665561676025s
Received healthy response to inference request in 1.2499756813049316s
Received healthy response to inference request in 1.25297212600708s
5 requests
0 failed requests
5th percentile: 1.2292354583740235
10th percentile: 1.2326043605804444
20th percentile: 1.239342164993286
30th percentile: 1.244163990020752
40th percentile: 1.2470698356628418
50th percentile: 1.2499756813049316
60th percentile: 1.251174259185791
70th percentile: 1.2523728370666505
80th percentile: 1.4562456130981447
90th percentile: 1.8627925872802735
95th percentile: 2.0660660743713377
99th percentile: 2.2286848640441894
mean time: 1.4481729984283447
Pipeline stage StressChecker completed in 7.91s
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
maldv-badger-kappa-llama-3-8b_v2 status is now deployed due to DeploymentManager action
maldv-badger-kappa-llama-3-8b_v2 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics