azazelle-l3-persephone-8_4975

developer_uid: Azazelle

submission_id: azazelle-l3-persephone-8_4975_v1

model_name: azazelle-l3-persephone-8_4975_v1

model_group: Azazelle/L3-Persephone-8

status: torndown

timestamp: 2024-07-01T13:44:36+00:00

num_battles: 36358

num_wins: 18218

celo_rating: 1181.93

family_friendly_score: 0.0

submission_type: basic

model_repo: Azazelle/L3-Persephone-8B-v1.0

model_architecture: LlamaForCausalLM

reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2

model_num_parameters: 8030261248.0

best_of: 4

max_input_tokens: 512

max_output_tokens: 64

display_name: azazelle-l3-persephone-8_4975_v1

is_internal_developer: False

language_model: Azazelle/L3-Persephone-8B-v1.0

model_size: 8B

ranking_group: single

us_pacific_date: 2024-07-01

win_ratio: 0.5010726662632708

generation_params: {'temperature': 1.0, 'top_p': 0.5, 'min_p': 0.0, 'top_k': 1000, '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': '<|im_start|>system\n{memory}<|im_end|>\n', 'prompt_template': '<|im_start|>user\n{prompt}<|im_end|>\n', 'bot_template': '<|im_start|>assistant\n{bot_name}: {message}<|im_end|>\n', 'user_template': '<|im_start|>user\n{user_name}: {message}<|im_end|>\n', 'response_template': '<|im_start|>assistant\n{bot_name}:', 'truncate_by_message': False}

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 azazelle-l3-persephone-8-4975-v1-mkmlizer
Waiting for job on azazelle-l3-persephone-8-4975-v1-mkmlizer to finish
azazelle-l3-persephone-8-4975-v1-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.
azazelle-l3-persephone-8-4975-v1-mkmlizer:   warnings.warn(warning_message, FutureWarning)
azazelle-l3-persephone-8-4975-v1-mkmlizer: Downloaded to shared memory in 54.620s
azazelle-l3-persephone-8-4975-v1-mkmlizer: quantizing model to /dev/shm/model_cache
azazelle-l3-persephone-8-4975-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
azazelle-l3-persephone-8-4975-v1-mkmlizer: 
Loading 0:   0%|          | 0/291 [00:00<?, ?it/s]
Loading 0:   1%|          | 2/291 [00:04<10:57,  2.27s/it]
Loading 0:   3%|▎         | 10/291 [00:04<01:45,  2.65it/s]
Loading 0:   8%|▊         | 22/291 [00:05<00:37,  7.20it/s]
Loading 0:  11%|█         | 32/291 [00:05<00:21, 12.07it/s]
Loading 0:  14%|█▍        | 42/291 [00:05<00:13, 18.15it/s]
Loading 0:  20%|█▉        | 58/291 [00:05<00:07, 30.42it/s]
Loading 0:  24%|██▎       | 69/291 [00:05<00:07, 29.85it/s]
Loading 0:  29%|██▉       | 85/291 [00:05<00:04, 43.06it/s]
Loading 0:  33%|███▎      | 95/291 [00:05<00:03, 50.38it/s]
Loading 0:  36%|███▌      | 105/291 [00:06<00:03, 57.69it/s]
Loading 0:  40%|███▉      | 115/291 [00:06<00:03, 44.08it/s]
Loading 0:  43%|████▎     | 126/291 [00:06<00:03, 53.93it/s]
Loading 0:  48%|████▊     | 139/291 [00:06<00:02, 65.62it/s]
Loading 0:  51%|█████     | 149/291 [00:06<00:01, 71.48it/s]
Loading 0:  55%|█████▍    | 159/291 [00:06<00:01, 76.72it/s]
Loading 0:  58%|█████▊    | 169/291 [00:07<00:02, 46.59it/s]
Loading 0:  62%|██████▏   | 180/291 [00:07<00:01, 56.39it/s]
Loading 0:  66%|██████▋   | 193/291 [00:07<00:01, 67.75it/s]
Loading 0:  70%|██████▉   | 203/291 [00:07<00:01, 71.22it/s]
Loading 0:  73%|███████▎  | 212/291 [00:07<00:01, 72.86it/s]
Loading 0:  76%|███████▌  | 221/291 [00:08<00:01, 47.29it/s]
Loading 0:  79%|███████▉  | 231/291 [00:08<00:01, 55.24it/s]
Loading 0:  85%|████████▍ | 247/291 [00:08<00:00, 72.65it/s]
Loading 0:  88%|████████▊ | 257/291 [00:08<00:00, 77.41it/s]
Loading 0:  92%|█████████▏| 267/291 [00:08<00:00, 49.11it/s]
Loading 0:  96%|█████████▌| 280/291 [00:08<00:00, 61.80it/s]
                                                            
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
azazelle-l3-persephone-8-4975-v1-mkmlizer: quantized model in 22.736s
azazelle-l3-persephone-8-4975-v1-mkmlizer: Processed model Azazelle/L3-Persephone-8B-v1.0 in 79.365s
azazelle-l3-persephone-8-4975-v1-mkmlizer: creating bucket guanaco-mkml-models
azazelle-l3-persephone-8-4975-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
azazelle-l3-persephone-8-4975-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/azazelle-l3-persephone-8-4975-v1
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/azazelle-l3-persephone-8-4975-v1/config.json
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/azazelle-l3-persephone-8-4975-v1/tokenizer_config.json
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/azazelle-l3-persephone-8-4975-v1/special_tokens_map.json
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/azazelle-l3-persephone-8-4975-v1/tokenizer.json
Connection pool is full, discarding connection: %s
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/azazelle-l3-persephone-8-4975-v1/flywheel_model.0.safetensors
azazelle-l3-persephone-8-4975-v1-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
azazelle-l3-persephone-8-4975-v1-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.
azazelle-l3-persephone-8-4975-v1-mkmlizer:   warnings.warn(
azazelle-l3-persephone-8-4975-v1-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.
azazelle-l3-persephone-8-4975-v1-mkmlizer:   warnings.warn(
azazelle-l3-persephone-8-4975-v1-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.
azazelle-l3-persephone-8-4975-v1-mkmlizer:   warnings.warn(
azazelle-l3-persephone-8-4975-v1-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()
azazelle-l3-persephone-8-4975-v1-mkmlizer:   return self.fget.__get__(instance, owner)()
azazelle-l3-persephone-8-4975-v1-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
azazelle-l3-persephone-8-4975-v1-mkmlizer: Saving duration: 0.259s
azazelle-l3-persephone-8-4975-v1-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 4.542s
azazelle-l3-persephone-8-4975-v1-mkmlizer: creating bucket guanaco-reward-models
azazelle-l3-persephone-8-4975-v1-mkmlizer: Bucket 's3://guanaco-reward-models/' created
azazelle-l3-persephone-8-4975-v1-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/azazelle-l3-persephone-8-4975-v1_reward
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/azazelle-l3-persephone-8-4975-v1_reward/config.json
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/azazelle-l3-persephone-8-4975-v1_reward/special_tokens_map.json
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/azazelle-l3-persephone-8-4975-v1_reward/merges.txt
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/azazelle-l3-persephone-8-4975-v1_reward/tokenizer_config.json
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/azazelle-l3-persephone-8-4975-v1_reward/vocab.json
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/azazelle-l3-persephone-8-4975-v1_reward/tokenizer.json
azazelle-l3-persephone-8-4975-v1-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/azazelle-l3-persephone-8-4975-v1_reward/reward.tensors
Job azazelle-l3-persephone-8-4975-v1-mkmlizer completed after 267.38s with status: succeeded
Stopping job with name azazelle-l3-persephone-8-4975-v1-mkmlizer
Pipeline stage MKMLizer completed in 268.31s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.33s
Running pipeline stage ISVCDeployer
Creating inference service azazelle-l3-persephone-8-4975-v1
Waiting for inference service azazelle-l3-persephone-8-4975-v1 to be ready
Inference service azazelle-l3-persephone-8-4975-v1 ready after 50.25165581703186s
Pipeline stage ISVCDeployer completed in 57.11s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.0569043159484863s
Received healthy response to inference request in 1.1518607139587402s
Received healthy response to inference request in 1.1467673778533936s
Received healthy response to inference request in 1.1362147331237793s
Received healthy response to inference request in 1.1405305862426758s
5 requests
0 failed requests
5th percentile: 1.1370779037475587
10th percentile: 1.1379410743713378
20th percentile: 1.1396674156188964
30th percentile: 1.1417779445648193
40th percentile: 1.1442726612091065
50th percentile: 1.1467673778533936
60th percentile: 1.1488047122955323
70th percentile: 1.150842046737671
80th percentile: 1.3328694343566896
90th percentile: 1.694886875152588
95th percentile: 1.875895595550537
99th percentile: 2.0207025718688962
mean time: 1.326455545425415
Pipeline stage StressChecker completed in 7.25s
azazelle-l3-persephone-8_4975_v1 status is now deployed due to DeploymentManager action
azazelle-l3-persephone-8_4975_v1 status is now inactive due to auto deactivation removed underperforming models
admin requested tearing down of azazelle-l3-persephone-8_4975_v1
Running pipeline stage ISVCDeleter
Checking if service azazelle-l3-persephone-8-4975-v1 is running
Skipping teardown as no inference service was found
Pipeline stage ISVCDeleter completed in 6.35s
Running pipeline stage MKMLModelDeleter
Cleaning model data from S3
Cleaning model data from model cache
Deleting key azazelle-l3-persephone-8-4975-v1/config.json from bucket guanaco-mkml-models
Deleting key azazelle-l3-persephone-8-4975-v1/flywheel_model.0.safetensors from bucket guanaco-mkml-models
Deleting key azazelle-l3-persephone-8-4975-v1/special_tokens_map.json from bucket guanaco-mkml-models
Deleting key azazelle-l3-persephone-8-4975-v1/tokenizer.json from bucket guanaco-mkml-models
Deleting key azazelle-l3-persephone-8-4975-v1/tokenizer_config.json from bucket guanaco-mkml-models
Cleaning model data from model cache
Deleting key azazelle-l3-persephone-8-4975-v1_reward/config.json from bucket guanaco-reward-models
Deleting key azazelle-l3-persephone-8-4975-v1_reward/merges.txt from bucket guanaco-reward-models
Deleting key azazelle-l3-persephone-8-4975-v1_reward/reward.tensors from bucket guanaco-reward-models
Deleting key azazelle-l3-persephone-8-4975-v1_reward/special_tokens_map.json from bucket guanaco-reward-models
Deleting key azazelle-l3-persephone-8-4975-v1_reward/tokenizer.json from bucket guanaco-reward-models
Deleting key azazelle-l3-persephone-8-4975-v1_reward/tokenizer_config.json from bucket guanaco-reward-models
Deleting key azazelle-l3-persephone-8-4975-v1_reward/vocab.json from bucket guanaco-reward-models
Pipeline stage MKMLModelDeleter completed in 5.35s
azazelle-l3-persephone-8_4975_v1 status is now torndown due to DeploymentManager action