submission_id: hastagaras-waduh-1-sft-l_7494_v1
developer_uid: Hastagaras
status: inactive
model_repo: Hastagaras/Waduh-1-SFT-llama3-8b
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 1.0, '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': "<|start_header_id|>system<|end_header_id|>\n\nAct as {bot_name}\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': 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': True}
timestamp: 2024-05-20T00:33:53+00:00
model_name: hopefully-fixed
model_eval_status: success
model_group: Hastagaras/Waduh-1-SFT-l
num_battles: 20235
num_wins: 10839
celo_rating: 1195.91
safety_score: 0.98
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: hopefully-fixed
ineligible_reason: propriety_total_count < 5000
language_model: Hastagaras/Waduh-1-SFT-llama3-8b
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-05-19
win_ratio: 0.5356560415122312
Resubmit model
Running pipeline stage MKMLizer
Starting job with name hastagaras-waduh-1-sft-l-7494-v1-mkmlizer
Waiting for job on hastagaras-waduh-1-sft-l-7494-v1-mkmlizer to finish
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ _____ __ __ ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ /___/ ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ Version: 0.8.14 ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ https://mk1.ai ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ The license key for the current software has been verified as ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ belonging to: ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ Chai Research Corp. ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ║ ║
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
hastagaras-waduh-1-sft-l-7494-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.
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: warnings.warn(warning_message, FutureWarning)
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Downloaded to shared memory in 36.686s
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: quantizing model to /dev/shm/model_cache
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 6%|▌ | 18/291 [00:00<00:01, 176.91it/s] Loading 0: 13%|█▎ | 39/291 [00:00<00:01, 193.29it/s] Loading 0: 20%|██ | 59/291 [00:00<00:01, 186.79it/s] Loading 0: 29%|██▊ | 83/291 [00:00<00:02, 92.78it/s] Loading 0: 35%|███▌ | 103/291 [00:00<00:01, 111.41it/s] Loading 0: 42%|████▏ | 122/291 [00:00<00:01, 128.16it/s] Loading 0: 51%|█████ | 147/291 [00:01<00:00, 152.89it/s] Loading 0: 57%|█████▋ | 166/291 [00:01<00:00, 161.49it/s] Loading 0: 64%|██████▍ | 187/291 [00:01<00:01, 95.29it/s] Loading 0: 69%|██████▉ | 202/291 [00:01<00:00, 100.88it/s] Loading 0: 77%|███████▋ | 224/291 [00:01<00:00, 122.82it/s] Loading 0: 85%|████████▍ | 246/291 [00:01<00:00, 140.66it/s] Loading 0: 91%|█████████ | 265/291 [00:02<00:00, 150.24it/s] Loading 0: 99%|█████████▊| 287/291 [00:06<00:00, 13.17it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: quantized model in 17.327s
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Processed model Hastagaras/Waduh-1-SFT-llama3-8b in 55.040s
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: creating bucket guanaco-mkml-models
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/hastagaras-waduh-1-sft-l-7494-v1
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/hastagaras-waduh-1-sft-l-7494-v1/special_tokens_map.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/hastagaras-waduh-1-sft-l-7494-v1/tokenizer_config.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/hastagaras-waduh-1-sft-l-7494-v1/config.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/hastagaras-waduh-1-sft-l-7494-v1/tokenizer.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/hastagaras-waduh-1-sft-l-7494-v1/flywheel_model.0.safetensors
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
hastagaras-waduh-1-sft-l-7494-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.
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: warnings.warn(
hastagaras-waduh-1-sft-l-7494-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.
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: warnings.warn(
hastagaras-waduh-1-sft-l-7494-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.
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: warnings.warn(
hastagaras-waduh-1-sft-l-7494-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()
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: return self.fget.__get__(instance, owner)()
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Saving duration: 0.245s
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 4.765s
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: creating bucket guanaco-reward-models
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: Bucket 's3://guanaco-reward-models/' created
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/hastagaras-waduh-1-sft-l-7494-v1_reward
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/hastagaras-waduh-1-sft-l-7494-v1_reward/config.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/hastagaras-waduh-1-sft-l-7494-v1_reward/special_tokens_map.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/hastagaras-waduh-1-sft-l-7494-v1_reward/tokenizer_config.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/hastagaras-waduh-1-sft-l-7494-v1_reward/merges.txt
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/hastagaras-waduh-1-sft-l-7494-v1_reward/vocab.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/hastagaras-waduh-1-sft-l-7494-v1_reward/tokenizer.json
hastagaras-waduh-1-sft-l-7494-v1-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/hastagaras-waduh-1-sft-l-7494-v1_reward/reward.tensors
Job hastagaras-waduh-1-sft-l-7494-v1-mkmlizer completed after 83.25s with status: succeeded
Stopping job with name hastagaras-waduh-1-sft-l-7494-v1-mkmlizer
Pipeline stage MKMLizer completed in 85.64s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.10s
Running pipeline stage ISVCDeployer
Creating inference service hastagaras-waduh-1-sft-l-7494-v1
Waiting for inference service hastagaras-waduh-1-sft-l-7494-v1 to be ready
Inference service hastagaras-waduh-1-sft-l-7494-v1 ready after 40.2479202747345s
Pipeline stage ISVCDeployer completed in 47.24s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.247159719467163s
Received healthy response to inference request in 1.2969655990600586s
Received healthy response to inference request in 1.288240909576416s
Received healthy response to inference request in 1.2625479698181152s
Received healthy response to inference request in 1.3332805633544922s
5 requests
0 failed requests
5th percentile: 1.2676865577697753
10th percentile: 1.2728251457214355
20th percentile: 1.283102321624756
30th percentile: 1.2899858474731445
40th percentile: 1.2934757232666017
50th percentile: 1.2969655990600586
60th percentile: 1.3114915847778321
70th percentile: 1.3260175704956054
80th percentile: 1.5160563945770265
90th percentile: 1.8816080570220948
95th percentile: 2.064383888244629
99th percentile: 2.210604553222656
mean time: 1.4856389522552491
Pipeline stage StressChecker completed in 8.04s
Running pipeline stage DaemonicModelEvalScorer
Pipeline stage DaemonicModelEvalScorer completed in 0.03s
Running M-Eval for topic stay_in_character
Running pipeline stage DaemonicSafetyScorer
M-Eval Dataset for topic stay_in_character is loaded
Pipeline stage DaemonicSafetyScorer completed in 0.06s
hastagaras-waduh-1-sft-l_7494_v1 status is now deployed due to DeploymentManager action
hastagaras-waduh-1-sft-l_7494_v1 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics