meseca-caspian-1

developer_uid: nguyenzzz

submission_id: meseca-caspian-1_v4

model_name: meseca-caspian-11_v1

model_group: meseca/caspian-1

status: torndown

timestamp: 2024-06-03T04:05:34+00:00

num_battles: 12521

num_wins: 6562

celo_rating: 1199.3

family_friendly_score: 0.0

submission_type: basic

model_repo: meseca/caspian-1

model_architecture: LlamaForCausalLM

reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2

model_num_parameters: 8030261248.0

best_of: 16

max_input_tokens: 512

max_output_tokens: 64

display_name: meseca-caspian-11_v1

is_internal_developer: False

language_model: meseca/caspian-1

model_size: 8B

ranking_group: single

us_pacific_date: 2024-06-02

win_ratio: 0.5240795463621116

generation_params: {'temperature': 1.25, 'top_p': 1.0, 'min_p': 0.08, 'top_k': 50, '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\nThis is an entertaining conversation. You are {bot_name} who has the persona: {memory}.\nEngage in a chat with {user_name} while staying in character. Try to flirt with {user_name}. Engage in *roleplay* actions. Describe the scene dramatically.\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}

model_eval_status: success

reward_formatter: {'bot_template': '{bot_name}: {message}\n', 'memory_template': "{bot_name}'s Persona: {memory}\n####\n", 'prompt_template': '{prompt}\n<START>\n', 'response_template': '{bot_name}:', 'truncate_by_message': False, 'user_template': '{user_name}: {message}\n'}

Resubmit model

Running pipeline stage MKMLizer
Starting job with name meseca-caspian-1-v4-mkmlizer
Waiting for job on meseca-caspian-1-v4-mkmlizer to finish
meseca-caspian-1-v4-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
meseca-caspian-1-v4-mkmlizer: ║     _____            __           __                                ║
meseca-caspian-1-v4-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
meseca-caspian-1-v4-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
meseca-caspian-1-v4-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
meseca-caspian-1-v4-mkmlizer: ║       /___/                                                         ║
meseca-caspian-1-v4-mkmlizer: ║                                                                     ║
meseca-caspian-1-v4-mkmlizer: ║  Version: 0.8.14                                                    ║
meseca-caspian-1-v4-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
meseca-caspian-1-v4-mkmlizer: ║  https://mk1.ai                                                     ║
meseca-caspian-1-v4-mkmlizer: ║                                                                     ║
meseca-caspian-1-v4-mkmlizer: ║  The license key for the current software has been verified as      ║
meseca-caspian-1-v4-mkmlizer: ║  belonging to:                                                      ║
meseca-caspian-1-v4-mkmlizer: ║                                                                     ║
meseca-caspian-1-v4-mkmlizer: ║  Chai Research Corp.                                                ║
meseca-caspian-1-v4-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
meseca-caspian-1-v4-mkmlizer: ║  Expiration: 2024-07-15 23:59:59                                    ║
meseca-caspian-1-v4-mkmlizer: ║                                                                     ║
meseca-caspian-1-v4-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
meseca-caspian-1-v4-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-caspian-1-v4-mkmlizer:   warnings.warn(warning_message, FutureWarning)
meseca-caspian-1-v4-mkmlizer: Downloaded to shared memory in 16.727s
meseca-caspian-1-v4-mkmlizer: quantizing model to /dev/shm/model_cache
meseca-caspian-1-v4-mkmlizer: Saving flywheel model at /dev/shm/model_cache
meseca-caspian-1-v4-mkmlizer: 
Loading 0:   0%|          | 0/291 [00:00<?, ?it/s]
Loading 0:   1%|          | 2/291 [00:04<11:08,  2.31s/it]
Loading 0:   5%|▌         | 15/291 [00:04<01:04,  4.31it/s]
Loading 0:  11%|█▏        | 33/291 [00:04<00:22, 11.47it/s]
Loading 0:  18%|█▊        | 51/291 [00:04<00:11, 20.61it/s]
Loading 0:  22%|██▏       | 65/291 [00:05<00:09, 25.10it/s]
Loading 0:  27%|██▋       | 78/291 [00:05<00:06, 33.58it/s]
Loading 0:  33%|███▎      | 96/291 [00:05<00:04, 48.31it/s]
Loading 0:  39%|███▉      | 114/291 [00:05<00:02, 64.53it/s]
Loading 0:  45%|████▌     | 132/291 [00:05<00:01, 81.02it/s]
Loading 0:  52%|█████▏    | 150/291 [00:05<00:01, 97.04it/s]
Loading 0:  57%|█████▋    | 166/291 [00:06<00:01, 76.93it/s]
Loading 0:  63%|██████▎   | 184/291 [00:06<00:01, 92.44it/s]
Loading 0:  69%|██████▉   | 202/291 [00:06<00:00, 107.51it/s]
Loading 0:  76%|███████▌  | 220/291 [00:06<00:00, 120.52it/s]
Loading 0:  82%|████████▏ | 238/291 [00:06<00:00, 132.10it/s]
Loading 0:  88%|████████▊ | 256/291 [00:06<00:00, 140.65it/s]
Loading 0:  93%|█████████▎| 272/291 [00:07<00:00, 92.48it/s] 
Loading 0:  98%|█████████▊| 285/291 [00:07<00:00, 99.13it/s]
                                                            
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
meseca-caspian-1-v4-mkmlizer: quantized model in 23.340s
meseca-caspian-1-v4-mkmlizer: Processed model meseca/caspian-1 in 42.513s
meseca-caspian-1-v4-mkmlizer: creating bucket guanaco-mkml-models
meseca-caspian-1-v4-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
meseca-caspian-1-v4-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/meseca-caspian-1-v4
meseca-caspian-1-v4-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/meseca-caspian-1-v4/tokenizer_config.json
meseca-caspian-1-v4-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/meseca-caspian-1-v4/special_tokens_map.json
meseca-caspian-1-v4-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/meseca-caspian-1-v4/config.json
meseca-caspian-1-v4-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/meseca-caspian-1-v4/tokenizer.json
meseca-caspian-1-v4-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/meseca-caspian-1-v4/flywheel_model.0.safetensors
meseca-caspian-1-v4-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
meseca-caspian-1-v4-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-caspian-1-v4-mkmlizer:   warnings.warn(
meseca-caspian-1-v4-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.
meseca-caspian-1-v4-mkmlizer:   warnings.warn(
meseca-caspian-1-v4-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-caspian-1-v4-mkmlizer:   warnings.warn(
meseca-caspian-1-v4-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-caspian-1-v4-mkmlizer:   return self.fget.__get__(instance, owner)()
meseca-caspian-1-v4-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
meseca-caspian-1-v4-mkmlizer: Saving duration: 0.385s
meseca-caspian-1-v4-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 4.211s
meseca-caspian-1-v4-mkmlizer: creating bucket guanaco-reward-models
meseca-caspian-1-v4-mkmlizer: Bucket 's3://guanaco-reward-models/' created
meseca-caspian-1-v4-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/meseca-caspian-1-v4_reward
meseca-caspian-1-v4-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/meseca-caspian-1-v4_reward/special_tokens_map.json
meseca-caspian-1-v4-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/meseca-caspian-1-v4_reward/tokenizer_config.json
meseca-caspian-1-v4-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/meseca-caspian-1-v4_reward/merges.txt
meseca-caspian-1-v4-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/meseca-caspian-1-v4_reward/config.json
meseca-caspian-1-v4-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/meseca-caspian-1-v4_reward/tokenizer.json
meseca-caspian-1-v4-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/meseca-caspian-1-v4_reward/vocab.json
meseca-caspian-1-v4-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/meseca-caspian-1-v4_reward/reward.tensors
Job meseca-caspian-1-v4-mkmlizer completed after 62.89s with status: succeeded
Stopping job with name meseca-caspian-1-v4-mkmlizer
Pipeline stage MKMLizer completed in 65.60s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.10s
Running pipeline stage ISVCDeployer
Creating inference service meseca-caspian-1-v4
Waiting for inference service meseca-caspian-1-v4 to be ready
Inference service meseca-caspian-1-v4 ready after 50.264251470565796s
Pipeline stage ISVCDeployer completed in 56.77s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.2436773777008057s
Received healthy response to inference request in 1.2931971549987793s
Received healthy response to inference request in 1.2475225925445557s
Received healthy response to inference request in 1.3335037231445312s
Received healthy response to inference request in 1.3182876110076904s
5 requests
0 failed requests
5th percentile: 1.2566575050354003
10th percentile: 1.2657924175262452
20th percentile: 1.2840622425079347
30th percentile: 1.2982152462005616
40th percentile: 1.308251428604126
50th percentile: 1.3182876110076904
60th percentile: 1.3243740558624268
70th percentile: 1.330460500717163
80th percentile: 1.5155384540557864
90th percentile: 1.879607915878296
95th percentile: 2.061642646789551
99th percentile: 2.2072704315185545
mean time: 1.4872376918792725
Pipeline stage StressChecker completed in 8.09s
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 0.03s
M-Eval Dataset for topic stay_in_character is loaded
meseca-caspian-1_v4 status is now deployed due to DeploymentManager action
meseca-caspian-1_v4 status is now inactive due to auto deactivation removed underperforming models
admin requested tearing down of meseca-caspian-1_v4
Running pipeline stage ISVCDeleter
Checking if service meseca-caspian-1-v4 is running
Skipping teardown as no inference service was found
Pipeline stage ISVCDeleter completed in 2.87s
Running pipeline stage MKMLModelDeleter
Cleaning model data from S3
Cleaning model data from model cache
Deleting key meseca-caspian-1-v4/config.json from bucket guanaco-mkml-models
Deleting key meseca-caspian-1-v4/flywheel_model.0.safetensors from bucket guanaco-mkml-models
Deleting key meseca-caspian-1-v4/special_tokens_map.json from bucket guanaco-mkml-models
Deleting key meseca-caspian-1-v4/tokenizer.json from bucket guanaco-mkml-models
Deleting key meseca-caspian-1-v4/tokenizer_config.json from bucket guanaco-mkml-models
Cleaning model data from model cache
Deleting key meseca-caspian-1-v4_reward/config.json from bucket guanaco-reward-models
Deleting key meseca-caspian-1-v4_reward/merges.txt from bucket guanaco-reward-models
Deleting key meseca-caspian-1-v4_reward/reward.tensors from bucket guanaco-reward-models
Deleting key meseca-caspian-1-v4_reward/special_tokens_map.json from bucket guanaco-reward-models
Deleting key meseca-caspian-1-v4_reward/tokenizer.json from bucket guanaco-reward-models
Deleting key meseca-caspian-1-v4_reward/tokenizer_config.json from bucket guanaco-reward-models
Deleting key meseca-caspian-1-v4_reward/vocab.json from bucket guanaco-reward-models
Pipeline stage MKMLModelDeleter completed in 3.52s
meseca-caspian-1_v4 status is now torndown due to DeploymentManager action