submission_id: meseca-20062024-c1-dpo-2e_v2
developer_uid: nguyenzzz
status: inactive
model_repo: meseca/20062024-c1-dpo-2e
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 0.95, 'top_p': 1.0, 'min_p': 0.1, 'top_k': 60, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '<|eot_id|>'], 'max_input_tokens': 512, 'best_of': 16, 'max_output_tokens': 64}
formatter: {'memory_template': "<|begin_of_text|><|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: 2024-06-27T16:22:21+00:00
model_name: meseca-20062024-c1-dpo-2e_v2
model_group: meseca/20062024-c1-dpo-2
num_battles: 13424
num_wins: 6473
celo_rating: 1163.58
propriety_score: 0.7625939326468133
propriety_total_count: 7186.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 16
max_input_tokens: 512
max_output_tokens: 64
display_name: meseca-20062024-c1-dpo-2e_v2
ineligible_reason: None
language_model: meseca/20062024-c1-dpo-2e
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-06-27
win_ratio: 0.48219606674612636
Resubmit model
Running pipeline stage MKMLizer
Starting job with name meseca-20062024-c1-dpo-2e-v2-mkmlizer
Waiting for job on meseca-20062024-c1-dpo-2e-v2-mkmlizer to finish
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ _____ __ __ ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ /___/ ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ Version: 0.8.14 ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ https://mk1.ai ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ The license key for the current software has been verified as ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ belonging to: ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ Chai Research Corp. ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ║ ║
meseca-20062024-c1-dpo-2e-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
meseca-20062024-c1-dpo-2e-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.
meseca-20062024-c1-dpo-2e-v2-mkmlizer: warnings.warn(warning_message, FutureWarning)
Connection pool is full, discarding connection: %s
Connection pool is full, discarding connection: %s
Connection pool is full, discarding connection: %s
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Downloaded to shared memory in 46.014s
meseca-20062024-c1-dpo-2e-v2-mkmlizer: quantizing model to /dev/shm/model_cache
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 4%|▍ | 13/291 [00:00<00:02, 113.27it/s] Loading 0: 9%|▉ | 27/291 [00:00<00:02, 125.51it/s] Loading 0: 14%|█▎ | 40/291 [00:00<00:02, 114.35it/s] Loading 0: 18%|█▊ | 52/291 [00:00<00:02, 111.76it/s] Loading 0: 22%|██▏ | 64/291 [00:00<00:01, 114.19it/s] Loading 0: 26%|██▌ | 76/291 [00:00<00:02, 103.31it/s] Loading 0: 30%|██▉ | 87/291 [00:01<00:03, 56.60it/s] Loading 0: 35%|███▌ | 102/291 [00:01<00:02, 72.14it/s] Loading 0: 39%|███▉ | 113/291 [00:01<00:02, 77.81it/s] Loading 0: 44%|████▍ | 129/291 [00:01<00:01, 94.26it/s] Loading 0: 48%|████▊ | 141/291 [00:01<00:01, 97.59it/s] Loading 0: 54%|█████▎ | 156/291 [00:01<00:01, 108.59it/s] Loading 0: 58%|█████▊ | 168/291 [00:01<00:01, 106.78it/s] Loading 0: 62%|██████▏ | 181/291 [00:01<00:00, 110.57it/s] Loading 0: 66%|██████▋ | 193/291 [00:02<00:01, 66.52it/s] Loading 0: 71%|███████ | 206/291 [00:02<00:01, 77.63it/s] Loading 0: 75%|███████▌ | 219/291 [00:02<00:00, 87.39it/s] Loading 0: 79%|███████▉ | 230/291 [00:02<00:00, 90.07it/s] Loading 0: 85%|████████▍ | 246/291 [00:02<00:00, 104.86it/s] Loading 0: 89%|████████▊ | 258/291 [00:02<00:00, 104.74it/s] Loading 0: 94%|█████████▍| 273/291 [00:02<00:00, 112.76it/s] Loading 0: 98%|█████████▊| 285/291 [00:02<00:00, 112.66it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
meseca-20062024-c1-dpo-2e-v2-mkmlizer: quantized model in 26.508s
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Processed model meseca/20062024-c1-dpo-2e in 75.079s
meseca-20062024-c1-dpo-2e-v2-mkmlizer: creating bucket guanaco-mkml-models
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
meseca-20062024-c1-dpo-2e-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/meseca-20062024-c1-dpo-2e-v2
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/meseca-20062024-c1-dpo-2e-v2/config.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/meseca-20062024-c1-dpo-2e-v2/special_tokens_map.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/meseca-20062024-c1-dpo-2e-v2/tokenizer_config.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/meseca-20062024-c1-dpo-2e-v2/tokenizer.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/meseca-20062024-c1-dpo-2e-v2/flywheel_model.0.safetensors
meseca-20062024-c1-dpo-2e-v2-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
meseca-20062024-c1-dpo-2e-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.
meseca-20062024-c1-dpo-2e-v2-mkmlizer: warnings.warn(
Connection pool is full, discarding connection: %s
meseca-20062024-c1-dpo-2e-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.
meseca-20062024-c1-dpo-2e-v2-mkmlizer: warnings.warn(
meseca-20062024-c1-dpo-2e-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()
meseca-20062024-c1-dpo-2e-v2-mkmlizer: return self.fget.__get__(instance, owner)()
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Saving duration: 0.483s
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 6.732s
meseca-20062024-c1-dpo-2e-v2-mkmlizer: creating bucket guanaco-reward-models
meseca-20062024-c1-dpo-2e-v2-mkmlizer: Bucket 's3://guanaco-reward-models/' created
meseca-20062024-c1-dpo-2e-v2-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/meseca-20062024-c1-dpo-2e-v2_reward
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/meseca-20062024-c1-dpo-2e-v2_reward/config.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/meseca-20062024-c1-dpo-2e-v2_reward/special_tokens_map.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/meseca-20062024-c1-dpo-2e-v2_reward/tokenizer_config.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/meseca-20062024-c1-dpo-2e-v2_reward/merges.txt
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/meseca-20062024-c1-dpo-2e-v2_reward/vocab.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/meseca-20062024-c1-dpo-2e-v2_reward/tokenizer.json
meseca-20062024-c1-dpo-2e-v2-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/meseca-20062024-c1-dpo-2e-v2_reward/reward.tensors
Job meseca-20062024-c1-dpo-2e-v2-mkmlizer completed after 115.97s with status: succeeded
Stopping job with name meseca-20062024-c1-dpo-2e-v2-mkmlizer
Pipeline stage MKMLizer completed in 116.99s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.13s
Running pipeline stage ISVCDeployer
Creating inference service meseca-20062024-c1-dpo-2e-v2
Waiting for inference service meseca-20062024-c1-dpo-2e-v2 to be ready
Inference service meseca-20062024-c1-dpo-2e-v2 ready after 51.23741173744202s
Pipeline stage ISVCDeployer completed in 58.04s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.10719633102417s
Received healthy response to inference request in 1.3428552150726318s
Received healthy response to inference request in 1.3268613815307617s
Received healthy response to inference request in 1.2844393253326416s
Received healthy response to inference request in 1.366509199142456s
5 requests
0 failed requests
5th percentile: 1.2929237365722657
10th percentile: 1.3014081478118897
20th percentile: 1.3183769702911377
30th percentile: 1.3300601482391357
40th percentile: 1.3364576816558837
50th percentile: 1.3428552150726318
60th percentile: 1.3523168087005615
70th percentile: 1.3617784023284911
80th percentile: 1.514646625518799
90th percentile: 1.8109214782714844
95th percentile: 1.959058904647827
99th percentile: 2.0775688457489014
mean time: 1.4855722904205322
Pipeline stage StressChecker completed in 8.26s
meseca-20062024-c1-dpo-2e_v2 status is now deployed due to DeploymentManager action
meseca-20062024-c1-dpo-2e_v2 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics