submission_id: grimjim-llama-3-merge-pp_1172_v1
developer_uid: grimjim.
status: inactive
model_repo: grimjim/llama-3-merge-pp-instruct-8B
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 0.95, 'top_p': 1.0, 'min_p': 0.0, 'top_k': 40, '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': "{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####\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: 2024-05-08T04:02:46+00:00
model_name: grimjim-llama-3-merge-pp_1172_v1
model_eval_status: success
double_thumbs_up: 139
thumbs_up: 219
thumbs_down: 86
num_battles: 10498
num_wins: 5565
celo_rating: 1190.19
entertaining: 7.06
stay_in_character: 8.52
user_preference: 7.62
safety_score: 0.95
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 16
max_input_tokens: 512
max_output_tokens: 64
display_name: grimjim-llama-3-merge-pp_1172_v1
double_thumbs_up_ratio: 0.3130630630630631
feedback_count: 444
ineligible_reason: None
language_model: grimjim/llama-3-merge-pp-instruct-8B
model_score: 7.733333333333333
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
single_thumbs_up_ratio: 0.49324324324324326
thumbs_down_ratio: 0.19369369369369369
thumbs_up_ratio: 0.8063063063063063
us_pacific_date: 2024-05-07
win_ratio: 0.5301009716136407
Resubmit model
Running pipeline stage MKMLizer
Starting job with name grimjim-llama-3-merge-pp-1172-v1-mkmlizer
Waiting for job on grimjim-llama-3-merge-pp-1172-v1-mkmlizer to finish
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ _____ __ __ ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ /___/ ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ Version: 0.8.10 ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ The license key for the current software has been verified as ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ belonging to: ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ Chai Research Corp. ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ║ ║
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
grimjim-llama-3-merge-pp-1172-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.
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: warnings.warn(warning_message, FutureWarning)
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Downloaded to shared memory in 19.352s
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: quantizing model to /dev/shm/model_cache
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 1%| | 2/291 [00:04<09:47, 2.03s/it] Loading 0: 45%|████▌ | 132/291 [00:05<00:04, 34.11it/s] Loading 0: 96%|█████████▌| 280/291 [00:06<00:00, 64.88it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: quantized model in 17.238s
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Processed model grimjim/llama-3-merge-pp-instruct-8B in 37.606s
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: creating bucket guanaco-mkml-models
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/grimjim-llama-3-merge-pp-1172-v1
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/grimjim-llama-3-merge-pp-1172-v1/config.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/grimjim-llama-3-merge-pp-1172-v1/tokenizer_config.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/grimjim-llama-3-merge-pp-1172-v1/special_tokens_map.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/grimjim-llama-3-merge-pp-1172-v1/tokenizer.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/grimjim-llama-3-merge-pp-1172-v1/flywheel_model.0.safetensors
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
grimjim-llama-3-merge-pp-1172-v1-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.
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: warnings.warn(
grimjim-llama-3-merge-pp-1172-v1-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.
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: warnings.warn(
grimjim-llama-3-merge-pp-1172-v1-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.
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: warnings.warn(
grimjim-llama-3-merge-pp-1172-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()
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: return self.fget.__get__(instance, owner)()
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Saving duration: 0.238s
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 3.967s
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: creating bucket guanaco-reward-models
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: Bucket 's3://guanaco-reward-models/' created
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/grimjim-llama-3-merge-pp-1172-v1_reward
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/grimjim-llama-3-merge-pp-1172-v1_reward/config.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/grimjim-llama-3-merge-pp-1172-v1_reward/special_tokens_map.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/grimjim-llama-3-merge-pp-1172-v1_reward/vocab.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/grimjim-llama-3-merge-pp-1172-v1_reward/tokenizer_config.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/grimjim-llama-3-merge-pp-1172-v1_reward/merges.txt
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/grimjim-llama-3-merge-pp-1172-v1_reward/tokenizer.json
grimjim-llama-3-merge-pp-1172-v1-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/grimjim-llama-3-merge-pp-1172-v1_reward/reward.tensors
Job grimjim-llama-3-merge-pp-1172-v1-mkmlizer completed after 62.73s with status: succeeded
Stopping job with name grimjim-llama-3-merge-pp-1172-v1-mkmlizer
Pipeline stage MKMLizer completed in 67.94s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.10s
Running pipeline stage ISVCDeployer
Creating inference service grimjim-llama-3-merge-pp-1172-v1
Waiting for inference service grimjim-llama-3-merge-pp-1172-v1 to be ready
Inference service grimjim-llama-3-merge-pp-1172-v1 ready after 40.23335146903992s
Pipeline stage ISVCDeployer completed in 48.48s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.190490961074829s
Received healthy response to inference request in 1.3420536518096924s
Received healthy response to inference request in 1.273413896560669s
Received healthy response to inference request in 1.248974323272705s
Received healthy response to inference request in 1.278487205505371s
5 requests
0 failed requests
5th percentile: 1.253862237930298
10th percentile: 1.2587501525878906
20th percentile: 1.268525981903076
30th percentile: 1.2744285583496093
40th percentile: 1.2764578819274903
50th percentile: 1.278487205505371
60th percentile: 1.3039137840270996
70th percentile: 1.329340362548828
80th percentile: 1.51174111366272
90th percentile: 1.8511160373687745
95th percentile: 2.0208034992218016
99th percentile: 2.1565534687042236
mean time: 1.4666840076446532
Pipeline stage StressChecker completed in 8.01s
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
grimjim-llama-3-merge-pp_1172_v1 status is now deployed due to DeploymentManager action
grimjim-llama-3-merge-pp_1172_v1 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics