submission_id: chaiml-phase2-winner-13b2_v289
developer_uid: Selenee
status: inactive
model_repo: ChaiML/phase2_winner_13b2
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 64.0, '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': 4, '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: 2024-06-19T23:59:05+00:00
model_name: test
model_group: ChaiML/phase2_winner_13b
num_battles: 19822
num_wins: 4511
celo_rating: 824.63
propriety_score: 0.6457800511508951
propriety_total_count: 9384.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 13015864320.0
best_of: 4
max_input_tokens: 512
max_output_tokens: 64
display_name: test
ineligible_reason: None
language_model: ChaiML/phase2_winner_13b2
model_size: 13B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-06-19
win_ratio: 0.22757542124911714
Resubmit model
Running pipeline stage MKMLizer
Starting job with name chaiml-phase2-winner-13b2-v289-mkmlizer
Waiting for job on chaiml-phase2-winner-13b2-v289-mkmlizer to finish
chaiml-phase2-winner-13b2-v289-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ _____ __ __ ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ /___/ ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ Version: 0.8.14 ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ https://mk1.ai ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ The license key for the current software has been verified as ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ belonging to: ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ Chai Research Corp. ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ║ ║
chaiml-phase2-winner-13b2-v289-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
chaiml-phase2-winner-13b2-v289-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.
chaiml-phase2-winner-13b2-v289-mkmlizer: warnings.warn(warning_message, FutureWarning)
chaiml-phase2-winner-13b2-v289-mkmlizer: Downloaded to shared memory in 42.008s
chaiml-phase2-winner-13b2-v289-mkmlizer: quantizing model to /dev/shm/model_cache
chaiml-phase2-winner-13b2-v289-mkmlizer: Saving flywheel model at /dev/shm/model_cache
chaiml-phase2-winner-13b2-v289-mkmlizer: quantized model in 23.204s
chaiml-phase2-winner-13b2-v289-mkmlizer: Processed model ChaiML/phase2_winner_13b2 in 69.310s
chaiml-phase2-winner-13b2-v289-mkmlizer: creating bucket guanaco-mkml-models
chaiml-phase2-winner-13b2-v289-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-phase2-winner-13b2-v289-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v289
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v289/config.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v289/tokenizer_config.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /dev/shm/model_cache/tokenizer.model s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v289/tokenizer.model
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v289/tokenizer.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v289/special_tokens_map.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v289/flywheel_model.0.safetensors
chaiml-phase2-winner-13b2-v289-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
chaiml-phase2-winner-13b2-v289-mkmlizer: Loading 0: 0%| | 0/403 [00:00<?, ?it/s] Loading 0: 38%|███▊ | 154/403 [00:01<00:03, 79.22it/s] Loading 0: 77%|███████▋ | 309/403 [00:03<00:00, 100.16it/s] Loading 0: 100%|██████████| 403/403 [00:04<00:00, 81.06it/s] /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.
chaiml-phase2-winner-13b2-v289-mkmlizer: warnings.warn(
chaiml-phase2-winner-13b2-v289-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.
chaiml-phase2-winner-13b2-v289-mkmlizer: warnings.warn(
chaiml-phase2-winner-13b2-v289-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.
chaiml-phase2-winner-13b2-v289-mkmlizer: warnings.warn(
chaiml-phase2-winner-13b2-v289-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()
chaiml-phase2-winner-13b2-v289-mkmlizer: return self.fget.__get__(instance, owner)()
chaiml-phase2-winner-13b2-v289-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
chaiml-phase2-winner-13b2-v289-mkmlizer: Saving duration: 0.395s
chaiml-phase2-winner-13b2-v289-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 8.951s
chaiml-phase2-winner-13b2-v289-mkmlizer: creating bucket guanaco-reward-models
chaiml-phase2-winner-13b2-v289-mkmlizer: Bucket 's3://guanaco-reward-models/' created
chaiml-phase2-winner-13b2-v289-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/chaiml-phase2-winner-13b2-v289_reward
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/chaiml-phase2-winner-13b2-v289_reward/config.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/chaiml-phase2-winner-13b2-v289_reward/tokenizer_config.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/chaiml-phase2-winner-13b2-v289_reward/vocab.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/chaiml-phase2-winner-13b2-v289_reward/merges.txt
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/chaiml-phase2-winner-13b2-v289_reward/special_tokens_map.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/chaiml-phase2-winner-13b2-v289_reward/tokenizer.json
chaiml-phase2-winner-13b2-v289-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/chaiml-phase2-winner-13b2-v289_reward/reward.tensors
Job chaiml-phase2-winner-13b2-v289-mkmlizer completed after 103.91s with status: succeeded
Stopping job with name chaiml-phase2-winner-13b2-v289-mkmlizer
Pipeline stage MKMLizer completed in 106.68s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.30s
Running pipeline stage ISVCDeployer
Creating inference service chaiml-phase2-winner-13b2-v289
Waiting for inference service chaiml-phase2-winner-13b2-v289 to be ready
Inference service chaiml-phase2-winner-13b2-v289 ready after 50.30112433433533s
Pipeline stage ISVCDeployer completed in 57.39s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.6595113277435303s
Received healthy response to inference request in 1.6517622470855713s
Received healthy response to inference request in 9.43567681312561s
Received healthy response to inference request in 1.6396360397338867s
Received healthy response to inference request in 1.6851906776428223s
5 requests
0 failed requests
5th percentile: 1.6420612812042237
10th percentile: 1.6444865226745606
20th percentile: 1.6493370056152343
30th percentile: 1.6584479331970214
40th percentile: 1.671819305419922
50th percentile: 1.6851906776428223
60th percentile: 2.074918937683105
70th percentile: 2.4646471977233886
80th percentile: 4.0147444248199475
90th percentile: 6.725210618972779
95th percentile: 8.080443716049192
99th percentile: 9.164630193710327
mean time: 3.414355421066284
Pipeline stage StressChecker completed in 17.65s
Running pipeline stage DaemonicSafetyScorer
Pipeline stage DaemonicSafetyScorer completed in 0.02s
chaiml-phase2-winner-13b2_v289 status is now deployed due to DeploymentManager action
chaiml-phase2-winner-13b2_v289 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics