submission_id: cycy233-l3-base-v6-e3-0_v6
developer_uid: shiroe40
status: inactive
model_repo: cycy233/L3-base-v6-e3.0
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 0.95, 'top_p': 0.95, 'min_p': 0.05, 'top_k': 80, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n', '</s>'], 'max_input_tokens': 512, 'best_of': 16, 'max_output_tokens': 64}
formatter: {'memory_template': "<|system|>\n### Instruction\nYou are a creative agent roleplaying as a character called {bot_name}. Stay true to the persona given, reply with long and descriptive sentences. Do not be repetitive.\nThis is {bot_name}'s Persona: {memory}\n", 'prompt_template': "### Input:\n# {bot_name}'s example conversation:\n{prompt}\n# {bot_name}'s actual conversation:\n", 'bot_template': '<|assistant|>\n{bot_name}: {message}</s>', 'user_template': '<|user|>\n{user_name}: {message}</s>', 'response_template': '<|assistant|>\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: 2024-07-03T02:13:13+00:00
model_name: cycy233-l3-base-v6-e3-0_v4
model_group: cycy233/L3-base-v6-e3.0
num_battles: 13246
num_wins: 6598
celo_rating: 1182.53
propriety_score: 0.7148874129272639
propriety_total_count: 6173.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 16
max_input_tokens: 512
max_output_tokens: 64
display_name: cycy233-l3-base-v6-e3-0_v4
ineligible_reason: None
language_model: cycy233/L3-base-v6-e3.0
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-07-02
win_ratio: 0.49811263777744225
Resubmit model
Running pipeline stage MKMLizer
Starting job with name cycy233-l3-base-v6-e3-0-v6-mkmlizer
Waiting for job on cycy233-l3-base-v6-e3-0-v6-mkmlizer to finish
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ _____ __ __ ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ /___/ ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ Version: 0.8.14 ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ https://mk1.ai ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ The license key for the current software has been verified as ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ belonging to: ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ Chai Research Corp. ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ║ ║
cycy233-l3-base-v6-e3-0-v6-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Downloaded to shared memory in 18.132s
cycy233-l3-base-v6-e3-0-v6-mkmlizer: quantizing model to /dev/shm/model_cache
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Saving flywheel model at /dev/shm/model_cache
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 29%|██▊ | 83/291 [00:01<00:02, 70.43it/s] Loading 0: 64%|██████▍ | 187/291 [00:02<00:01, 82.00it/s] Loading 0: 99%|█████████▊| 287/291 [00:02<00:00, 119.65it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
cycy233-l3-base-v6-e3-0-v6-mkmlizer: quantized model in 21.345s
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Processed model cycy233/L3-base-v6-e3.0 in 39.477s
cycy233-l3-base-v6-e3-0-v6-mkmlizer: creating bucket guanaco-mkml-models
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
cycy233-l3-base-v6-e3-0-v6-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/cycy233-l3-base-v6-e3-0-v6
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/cycy233-l3-base-v6-e3-0-v6/special_tokens_map.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/cycy233-l3-base-v6-e3-0-v6/tokenizer_config.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/cycy233-l3-base-v6-e3-0-v6/config.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/cycy233-l3-base-v6-e3-0-v6/tokenizer.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/cycy233-l3-base-v6-e3-0-v6/flywheel_model.0.safetensors
cycy233-l3-base-v6-e3-0-v6-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
cycy233-l3-base-v6-e3-0-v6-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py:919: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
cycy233-l3-base-v6-e3-0-v6-mkmlizer: warnings.warn(
cycy233-l3-base-v6-e3-0-v6-mkmlizer: /opt/conda/lib/python3.10/site-packages/huggingface_hub/file_download.py:1132: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
cycy233-l3-base-v6-e3-0-v6-mkmlizer: warnings.warn(
cycy233-l3-base-v6-e3-0-v6-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/tokenization_auto.py:769: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
cycy233-l3-base-v6-e3-0-v6-mkmlizer: warnings.warn(
cycy233-l3-base-v6-e3-0-v6-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.
cycy233-l3-base-v6-e3-0-v6-mkmlizer: warnings.warn(
cycy233-l3-base-v6-e3-0-v6-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()
cycy233-l3-base-v6-e3-0-v6-mkmlizer: return self.fget.__get__(instance, owner)()
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Saving duration: 0.265s
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 4.744s
cycy233-l3-base-v6-e3-0-v6-mkmlizer: creating bucket guanaco-reward-models
cycy233-l3-base-v6-e3-0-v6-mkmlizer: Bucket 's3://guanaco-reward-models/' created
cycy233-l3-base-v6-e3-0-v6-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/cycy233-l3-base-v6-e3-0-v6_reward
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/cycy233-l3-base-v6-e3-0-v6_reward/config.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/cycy233-l3-base-v6-e3-0-v6_reward/special_tokens_map.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/cycy233-l3-base-v6-e3-0-v6_reward/tokenizer_config.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/cycy233-l3-base-v6-e3-0-v6_reward/vocab.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/cycy233-l3-base-v6-e3-0-v6_reward/merges.txt
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/cycy233-l3-base-v6-e3-0-v6_reward/tokenizer.json
cycy233-l3-base-v6-e3-0-v6-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/cycy233-l3-base-v6-e3-0-v6_reward/reward.tensors
Job cycy233-l3-base-v6-e3-0-v6-mkmlizer completed after 129.77s with status: succeeded
Stopping job with name cycy233-l3-base-v6-e3-0-v6-mkmlizer
Pipeline stage MKMLizer completed in 130.69s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.10s
Running pipeline stage ISVCDeployer
Creating inference service cycy233-l3-base-v6-e3-0-v6
Waiting for inference service cycy233-l3-base-v6-e3-0-v6 to be ready
Inference service cycy233-l3-base-v6-e3-0-v6 ready after 40.227341651916504s
Pipeline stage ISVCDeployer completed in 47.13s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.1348280906677246s
Received healthy response to inference request in 1.2701668739318848s
Received healthy response to inference request in 1.2820472717285156s
Received healthy response to inference request in 1.3356068134307861s
Received healthy response to inference request in 1.2884771823883057s
5 requests
0 failed requests
5th percentile: 1.272542953491211
10th percentile: 1.274919033050537
20th percentile: 1.2796711921691895
30th percentile: 1.2833332538604736
40th percentile: 1.2859052181243897
50th percentile: 1.2884771823883057
60th percentile: 1.307329034805298
70th percentile: 1.32618088722229
80th percentile: 1.495451068878174
90th percentile: 1.8151395797729493
95th percentile: 1.9749838352203368
99th percentile: 2.102859239578247
mean time: 1.4622252464294434
Pipeline stage StressChecker completed in 8.04s
cycy233-l3-base-v6-e3-0_v6 status is now deployed due to DeploymentManager action
cycy233-l3-base-v6-e3-0_v6 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics