submission_id: cgato-thesalt-rp-l3-8b-d_5388_v2
developer_uid: c.gato
status: inactive
model_repo: cgato/TheSalt-RP-L3-8b-DPO-v0.3.2-e0.2.2
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 0.9, 'top_p': 0.9, 'min_p': 0.0, 'top_k': 100, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n'], 'max_input_tokens': 512, 'best_of': 16, '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': True}
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': True}
timestamp: 2024-06-22T23:10:12+00:00
model_name: cgato-thesalt-rp-l3-8b-d_5388_v2
model_group: cgato/TheSalt-RP-L3-8b-D
num_battles: 15214
num_wins: 7727
celo_rating: 1206.1
propriety_score: 0.7134009009009009
propriety_total_count: 7104.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030277632.0
best_of: 16
max_input_tokens: 512
max_output_tokens: 64
display_name: cgato-thesalt-rp-l3-8b-d_5388_v2
ineligible_reason: None
language_model: cgato/TheSalt-RP-L3-8b-DPO-v0.3.2-e0.2.2
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-06-22
win_ratio: 0.5078874720652031
Resubmit model
Running pipeline stage MKMLizer
Starting job with name cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer
Waiting for job on cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer to finish
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ _____ __ __ ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ /___/ ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ Version: 0.8.14 ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ https://mk1.ai ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ The license key for the current software has been verified as ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ belonging to: ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ Chai Research Corp. ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ║ ║
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
cgato-thesalt-rp-l3-8b-d-5388-v2-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.
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: warnings.warn(warning_message, FutureWarning)
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Downloaded to shared memory in 29.942s
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: quantizing model to /dev/shm/model_cache
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 29%|██▊ | 83/291 [00:01<00:02, 74.91it/s] Loading 0: 64%|██████▍ | 187/291 [00:02<00:01, 89.51it/s] Loading 0: 99%|█████████▊| 287/291 [00:02<00:00, 130.87it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: quantized model in 19.247s
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Processed model cgato/TheSalt-RP-L3-8b-DPO-v0.3.2-e0.2.2 in 50.313s
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: creating bucket guanaco-mkml-models
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/cgato-thesalt-rp-l3-8b-d-5388-v2
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/cgato-thesalt-rp-l3-8b-d-5388-v2/special_tokens_map.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/cgato-thesalt-rp-l3-8b-d-5388-v2/config.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/cgato-thesalt-rp-l3-8b-d-5388-v2/tokenizer_config.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/cgato-thesalt-rp-l3-8b-d-5388-v2/tokenizer.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/cgato-thesalt-rp-l3-8b-d-5388-v2/flywheel_model.0.safetensors
cgato-thesalt-rp-l3-8b-d-5388-v2-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.
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: warnings.warn(
cgato-thesalt-rp-l3-8b-d-5388-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.
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: warnings.warn(
cgato-thesalt-rp-l3-8b-d-5388-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()
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: return self.fget.__get__(instance, owner)()
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Saving duration: 0.301s
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 6.134s
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: creating bucket guanaco-reward-models
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: Bucket 's3://guanaco-reward-models/' created
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/cgato-thesalt-rp-l3-8b-d-5388-v2_reward
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/cgato-thesalt-rp-l3-8b-d-5388-v2_reward/config.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/cgato-thesalt-rp-l3-8b-d-5388-v2_reward/special_tokens_map.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/cgato-thesalt-rp-l3-8b-d-5388-v2_reward/tokenizer_config.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/cgato-thesalt-rp-l3-8b-d-5388-v2_reward/merges.txt
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/cgato-thesalt-rp-l3-8b-d-5388-v2_reward/vocab.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/cgato-thesalt-rp-l3-8b-d-5388-v2_reward/tokenizer.json
cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/cgato-thesalt-rp-l3-8b-d-5388-v2_reward/reward.tensors
Job cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer completed after 85.03s with status: succeeded
Stopping job with name cgato-thesalt-rp-l3-8b-d-5388-v2-mkmlizer
Pipeline stage MKMLizer completed in 85.42s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.11s
Running pipeline stage ISVCDeployer
Creating inference service cgato-thesalt-rp-l3-8b-d-5388-v2
Waiting for inference service cgato-thesalt-rp-l3-8b-d-5388-v2 to be ready
Inference service cgato-thesalt-rp-l3-8b-d-5388-v2 ready after 40.19851207733154s
Pipeline stage ISVCDeployer completed in 45.89s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.2316343784332275s
Received healthy response to inference request in 1.3416547775268555s
Received healthy response to inference request in 1.3174450397491455s
Received healthy response to inference request in 1.3291444778442383s
Received healthy response to inference request in 1.439985752105713s
5 requests
0 failed requests
5th percentile: 1.319784927368164
10th percentile: 1.3221248149871827
20th percentile: 1.3268045902252197
30th percentile: 1.3316465377807618
40th percentile: 1.3366506576538086
50th percentile: 1.3416547775268555
60th percentile: 1.3809871673583984
70th percentile: 1.4203195571899414
80th percentile: 1.598315477371216
90th percentile: 1.9149749279022217
95th percentile: 2.0733046531677246
99th percentile: 2.199968433380127
mean time: 1.531972885131836
Pipeline stage StressChecker completed in 8.36s
Running pipeline stage DaemonicSafetyScorer
Pipeline stage DaemonicSafetyScorer completed in 0.03s
cgato-thesalt-rp-l3-8b-d_5388_v2 status is now deployed due to DeploymentManager action
cgato-thesalt-rp-l3-8b-d_5388_v2 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics