submission_id: sao10k-l3-rp-v5-4_v2
developer_uid: sao10k
status: inactive
model_repo: Sao10K/L3-RP-v5.4
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 1.2, 'top_p': 1.0, 'min_p': 0.1, 'top_k': 40, '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: 2024-07-11T14:15:58+00:00
model_name: RP-v5-4-2
model_group: Sao10K/L3-RP-v5.4
num_battles: 45236
num_wins: 24301
celo_rating: 1219.88
propriety_score: 0.7085815692736722
propriety_total_count: 7889.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 16
max_input_tokens: 512
max_output_tokens: 64
display_name: RP-v5-4-2
ineligible_reason: None
language_model: Sao10K/L3-RP-v5.4
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-07-11
win_ratio: 0.5372048810681758
preference_data_url: None
Resubmit model
Running pipeline stage MKMLizer
Starting job with name sao10k-l3-rp-v5-4-v2-mkmlizer
Waiting for job on sao10k-l3-rp-v5-4-v2-mkmlizer to finish
sao10k-l3-rp-v5-4-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ _____ __ __ ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ /___/ ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ Version: 0.8.14 ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ https://mk1.ai ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ The license key for the current software has been verified as ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ belonging to: ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ Chai Research Corp. ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ Expiration: 2024-10-15 23:59:59 ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ║ ║
sao10k-l3-rp-v5-4-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
sao10k-l3-rp-v5-4-v2-mkmlizer: Downloaded to shared memory in 32.595s
sao10k-l3-rp-v5-4-v2-mkmlizer: quantizing model to /dev/shm/model_cache
sao10k-l3-rp-v5-4-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
sao10k-l3-rp-v5-4-v2-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 1%| | 2/291 [00:05<13:30, 2.80s/it] Loading 0: 5%|▍ | 14/291 [00:05<01:23, 3.32it/s] Loading 0: 10%|▉ | 29/291 [00:05<00:31, 8.34it/s] Loading 0: 14%|█▍ | 41/291 [00:05<00:18, 13.52it/s] Loading 0: 19%|█▉ | 56/291 [00:06<00:10, 22.01it/s] Loading 0: 24%|██▎ | 69/291 [00:06<00:09, 23.34it/s] Loading 0: 29%|██▉ | 85/291 [00:06<00:05, 34.36it/s] Loading 0: 33%|███▎ | 97/291 [00:06<00:04, 43.00it/s] Loading 0: 38%|███▊ | 112/291 [00:06<00:03, 55.99it/s] Loading 0: 43%|████▎ | 124/291 [00:06<00:02, 64.90it/s] Loading 0: 48%|████▊ | 139/291 [00:07<00:01, 79.69it/s] Loading 0: 52%|█████▏ | 152/291 [00:07<00:01, 84.62it/s] Loading 0: 57%|█████▋ | 166/291 [00:07<00:02, 55.07it/s] Loading 0: 61%|██████ | 177/291 [00:07<00:01, 62.51it/s] Loading 0: 66%|██████▋ | 193/291 [00:07<00:01, 79.13it/s] Loading 0: 70%|███████ | 205/291 [00:07<00:01, 85.22it/s] Loading 0: 76%|███████▌ | 220/291 [00:08<00:00, 98.29it/s] Loading 0: 80%|████████ | 233/291 [00:08<00:00, 103.41it/s] Loading 0: 85%|████████▍ | 247/291 [00:08<00:00, 111.60it/s] Loading 0: 89%|████████▉ | 260/291 [00:08<00:00, 112.63it/s] Loading 0: 94%|█████████▍| 273/291 [00:08<00:00, 60.50it/s] Loading 0: 98%|█████████▊| 284/291 [00:08<00:00, 67.13it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
sao10k-l3-rp-v5-4-v2-mkmlizer: quantized model in 29.280s
sao10k-l3-rp-v5-4-v2-mkmlizer: Processed model Sao10K/L3-RP-v5.4 in 61.876s
sao10k-l3-rp-v5-4-v2-mkmlizer: creating bucket guanaco-mkml-models
sao10k-l3-rp-v5-4-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
sao10k-l3-rp-v5-4-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/sao10k-l3-rp-v5-4-v2
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/sao10k-l3-rp-v5-4-v2/config.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/sao10k-l3-rp-v5-4-v2/tokenizer_config.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/sao10k-l3-rp-v5-4-v2/tokenizer.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/sao10k-l3-rp-v5-4-v2/special_tokens_map.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/sao10k-l3-rp-v5-4-v2/flywheel_model.0.safetensors
sao10k-l3-rp-v5-4-v2-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
sao10k-l3-rp-v5-4-v2-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.
sao10k-l3-rp-v5-4-v2-mkmlizer: warnings.warn(
sao10k-l3-rp-v5-4-v2-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`.
sao10k-l3-rp-v5-4-v2-mkmlizer: warnings.warn(
sao10k-l3-rp-v5-4-v2-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.
sao10k-l3-rp-v5-4-v2-mkmlizer: warnings.warn(
sao10k-l3-rp-v5-4-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.
sao10k-l3-rp-v5-4-v2-mkmlizer: warnings.warn(
sao10k-l3-rp-v5-4-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()
sao10k-l3-rp-v5-4-v2-mkmlizer: return self.fget.__get__(instance, owner)()
sao10k-l3-rp-v5-4-v2-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
sao10k-l3-rp-v5-4-v2-mkmlizer: Saving duration: 0.511s
sao10k-l3-rp-v5-4-v2-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 7.404s
sao10k-l3-rp-v5-4-v2-mkmlizer: creating bucket guanaco-reward-models
sao10k-l3-rp-v5-4-v2-mkmlizer: Bucket 's3://guanaco-reward-models/' created
sao10k-l3-rp-v5-4-v2-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/sao10k-l3-rp-v5-4-v2_reward
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/sao10k-l3-rp-v5-4-v2_reward/config.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/sao10k-l3-rp-v5-4-v2_reward/special_tokens_map.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/sao10k-l3-rp-v5-4-v2_reward/tokenizer_config.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/sao10k-l3-rp-v5-4-v2_reward/vocab.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/sao10k-l3-rp-v5-4-v2_reward/merges.txt
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/sao10k-l3-rp-v5-4-v2_reward/tokenizer.json
sao10k-l3-rp-v5-4-v2-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/sao10k-l3-rp-v5-4-v2_reward/reward.tensors
Job sao10k-l3-rp-v5-4-v2-mkmlizer completed after 104.58s with status: succeeded
Stopping job with name sao10k-l3-rp-v5-4-v2-mkmlizer
Pipeline stage MKMLizer completed in 105.54s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.11s
Running pipeline stage ISVCDeployer
Creating inference service sao10k-l3-rp-v5-4-v2
Waiting for inference service sao10k-l3-rp-v5-4-v2 to be ready
Inference service sao10k-l3-rp-v5-4-v2 ready after 40.249772787094116s
Pipeline stage ISVCDeployer completed in 47.00s
Running pipeline stage StressChecker
Received healthy response to inference request in 1.9837253093719482s
Received healthy response to inference request in 1.403355598449707s
Received healthy response to inference request in 1.3297584056854248s
Received healthy response to inference request in 1.2839534282684326s
Received healthy response to inference request in 1.3519928455352783s
5 requests
0 failed requests
5th percentile: 1.293114423751831
10th percentile: 1.3022754192352295
20th percentile: 1.3205974102020264
30th percentile: 1.3342052936553954
40th percentile: 1.3430990695953369
50th percentile: 1.3519928455352783
60th percentile: 1.37253794670105
70th percentile: 1.3930830478668212
80th percentile: 1.5194295406341554
90th percentile: 1.7515774250030518
95th percentile: 1.8676513671875
99th percentile: 1.9605105209350586
mean time: 1.4705571174621581
Pipeline stage StressChecker completed in 8.53s
sao10k-l3-rp-v5-4_v2 status is now deployed due to DeploymentManager action
sao10k-l3-rp-v5-4_v2 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics