submission_id: nousresearch-meta-llama_4941_v72
developer_uid: robert_irvine
status: inactive
model_repo: NousResearch/Meta-Llama-3-8B-Instruct
reward_repo: rirv938/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 1.0, 'top_p': 1.0, 'min_p': 0.0, 'top_k': 40, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['</s>', '<|user|>', '###', '\n'], 'max_input_tokens': 512, 'best_of': 8, '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': 'Memory: {memory}\n', 'prompt_template': '{prompt}\n', 'bot_template': 'Bot: {message}\n', 'user_template': 'User: {message}\n', 'response_template': 'Bot:', 'truncate_by_message': False}
timestamp: 2024-07-03T18:21:54+00:00
model_name: nousresearch-meta-llama_4941_v72
model_group: NousResearch/Meta-Llama-
num_battles: 13938
num_wins: 6833
celo_rating: 1163.16
propriety_score: 0.7340280924331672
propriety_total_count: 6621.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 8
max_input_tokens: 512
max_output_tokens: 64
display_name: nousresearch-meta-llama_4941_v72
ineligible_reason: None
language_model: NousResearch/Meta-Llama-3-8B-Instruct
model_size: 8B
reward_model: rirv938/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-07-03
win_ratio: 0.49024250251112067
Resubmit model
Running pipeline stage MKMLizer
Starting job with name nousresearch-meta-llama-4941-v72-mkmlizer
Waiting for job on nousresearch-meta-llama-4941-v72-mkmlizer to finish
nousresearch-meta-llama-4941-v72-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
nousresearch-meta-llama-4941-v72-mkmlizer: ║ _____ __ __ ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ /___/ ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ Version: 0.8.14 ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ https://mk1.ai ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ The license key for the current software has been verified as ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ belonging to: ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ Chai Research Corp. ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
nousresearch-meta-llama-4941-v72-mkmlizer: ║ ║
nousresearch-meta-llama-4941-v72-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
nousresearch-meta-llama-4941-v72-mkmlizer: Downloaded to shared memory in 59.873s
nousresearch-meta-llama-4941-v72-mkmlizer: quantizing model to /dev/shm/model_cache
nousresearch-meta-llama-4941-v72-mkmlizer: Saving flywheel model at /dev/shm/model_cache
nousresearch-meta-llama-4941-v72-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 4%|▍ | 12/291 [00:00<00:02, 107.67it/s] Loading 0: 8%|▊ | 23/291 [00:00<00:03, 87.05it/s] Loading 0: 13%|█▎ | 39/291 [00:00<00:02, 104.72it/s] Loading 0: 17%|█▋ | 50/291 [00:00<00:02, 88.92it/s] Loading 0: 23%|██▎ | 66/291 [00:00<00:02, 105.07it/s] Loading 0: 26%|██▋ | 77/291 [00:00<00:02, 101.04it/s] Loading 0: 30%|███ | 88/291 [00:01<00:03, 57.38it/s] Loading 0: 35%|███▌ | 102/291 [00:01<00:02, 70.10it/s] Loading 0: 39%|███▉ | 113/291 [00:01<00:02, 75.43it/s] Loading 0: 44%|████▍ | 129/291 [00:01<00:01, 91.21it/s] Loading 0: 48%|████▊ | 140/291 [00:01<00:01, 92.22it/s] Loading 0: 52%|█████▏ | 152/291 [00:01<00:01, 98.73it/s] Loading 0: 57%|█████▋ | 165/291 [00:01<00:01, 105.59it/s] Loading 0: 62%|██████▏ | 179/291 [00:01<00:01, 111.20it/s] Loading 0: 66%|██████▌ | 191/291 [00:02<00:01, 64.87it/s] Loading 0: 69%|██████▉ | 201/291 [00:02<00:01, 66.93it/s] Loading 0: 73%|███████▎ | 212/291 [00:02<00:01, 72.73it/s] Loading 0: 78%|███████▊ | 228/291 [00:02<00:00, 88.78it/s] Loading 0: 82%|████████▏ | 239/291 [00:02<00:00, 90.02it/s] Loading 0: 88%|████████▊ | 255/291 [00:02<00:00, 104.52it/s] Loading 0: 92%|█████████▏| 267/291 [00:03<00:00, 104.33it/s] Loading 0: 97%|█████████▋| 281/291 [00:03<00:00, 109.85it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
nousresearch-meta-llama-4941-v72-mkmlizer: quantized model in 24.780s
nousresearch-meta-llama-4941-v72-mkmlizer: Processed model NousResearch/Meta-Llama-3-8B-Instruct in 84.654s
nousresearch-meta-llama-4941-v72-mkmlizer: creating bucket guanaco-mkml-models
nousresearch-meta-llama-4941-v72-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
nousresearch-meta-llama-4941-v72-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/nousresearch-meta-llama-4941-v72
nousresearch-meta-llama-4941-v72-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/nousresearch-meta-llama-4941-v72/config.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/nousresearch-meta-llama-4941-v72/tokenizer_config.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/nousresearch-meta-llama-4941-v72/special_tokens_map.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/nousresearch-meta-llama-4941-v72/tokenizer.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/nousresearch-meta-llama-4941-v72/flywheel_model.0.safetensors
nousresearch-meta-llama-4941-v72-mkmlizer: loading reward model from rirv938/reward_gpt2_medium_preference_24m_e2
nousresearch-meta-llama-4941-v72-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.
nousresearch-meta-llama-4941-v72-mkmlizer: warnings.warn(
nousresearch-meta-llama-4941-v72-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`.
nousresearch-meta-llama-4941-v72-mkmlizer: warnings.warn(
nousresearch-meta-llama-4941-v72-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.
nousresearch-meta-llama-4941-v72-mkmlizer: warnings.warn(
nousresearch-meta-llama-4941-v72-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()
nousresearch-meta-llama-4941-v72-mkmlizer: return self.fget.__get__(instance, owner)()
nousresearch-meta-llama-4941-v72-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
nousresearch-meta-llama-4941-v72-mkmlizer: Saving duration: 0.442s
nousresearch-meta-llama-4941-v72-mkmlizer: Processed model rirv938/reward_gpt2_medium_preference_24m_e2 in 7.308s
nousresearch-meta-llama-4941-v72-mkmlizer: creating bucket guanaco-reward-models
nousresearch-meta-llama-4941-v72-mkmlizer: Bucket 's3://guanaco-reward-models/' created
nousresearch-meta-llama-4941-v72-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/nousresearch-meta-llama-4941-v72_reward
nousresearch-meta-llama-4941-v72-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/nousresearch-meta-llama-4941-v72_reward/special_tokens_map.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/nousresearch-meta-llama-4941-v72_reward/config.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/nousresearch-meta-llama-4941-v72_reward/vocab.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/nousresearch-meta-llama-4941-v72_reward/merges.txt
nousresearch-meta-llama-4941-v72-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/nousresearch-meta-llama-4941-v72_reward/tokenizer_config.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/nousresearch-meta-llama-4941-v72_reward/tokenizer.json
nousresearch-meta-llama-4941-v72-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/nousresearch-meta-llama-4941-v72_reward/reward.tensors
Job nousresearch-meta-llama-4941-v72-mkmlizer completed after 117.45s with status: succeeded
Stopping job with name nousresearch-meta-llama-4941-v72-mkmlizer
Pipeline stage MKMLizer completed in 118.24s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.12s
Running pipeline stage ISVCDeployer
Creating inference service nousresearch-meta-llama-4941-v72
Waiting for inference service nousresearch-meta-llama-4941-v72 to be ready
Inference service nousresearch-meta-llama-4941-v72 ready after 40.213507413864136s
Pipeline stage ISVCDeployer completed in 47.13s
Running pipeline stage StressChecker
Received healthy response to inference request in 1.8685975074768066s
Received healthy response to inference request in 1.156883716583252s
Received healthy response to inference request in 1.1471261978149414s
Received healthy response to inference request in 1.2912349700927734s
Received healthy response to inference request in 1.1470439434051514s
5 requests
0 failed requests
5th percentile: 1.1470603942871094
10th percentile: 1.1470768451690674
20th percentile: 1.1471097469329834
30th percentile: 1.1490777015686036
40th percentile: 1.1529807090759276
50th percentile: 1.156883716583252
60th percentile: 1.2106242179870605
70th percentile: 1.2643647193908691
80th percentile: 1.4067074775695803
90th percentile: 1.6376524925231934
95th percentile: 1.7531249999999998
99th percentile: 1.8455030059814452
mean time: 1.322177267074585
Pipeline stage StressChecker completed in 7.36s
nousresearch-meta-llama_4941_v72 status is now deployed due to DeploymentManager action
nousresearch-meta-llama_4941_v72 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics