submission_id: cgato-l3-thespice-8b-dpo_2837_v3
developer_uid: c.gato
status: inactive
model_repo: cgato/L3-TheSpice-8b-DPO-v3.2-e14.1
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 0.9, 'top_p': 0.9, 'min_p': 0.0, 'top_k': 200, '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': "{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}
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: 2024-06-02T11:40:24+00:00
model_name: cgato-l3-thespice-8b-dpo_2837_v3
model_eval_status: success
model_group: cgato/L3-TheSpice-8b-DPO
num_battles: 12880
num_wins: 6814
celo_rating: 1206.11
safety_score: 0.86
propriety_score: 0.6043956043956044
propriety_total_count: 273.0
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-l3-thespice-8b-dpo_2837_v3
ineligible_reason: propriety_total_count < 5000
language_model: cgato/L3-TheSpice-8b-DPO-v3.2-e14.1
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-06-02
win_ratio: 0.5290372670807454
Resubmit model
Running pipeline stage MKMLizer
Starting job with name cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer
Waiting for job on cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer to finish
Retrying (%r) after connection broken by '%r': %s
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Downloaded to shared memory in 45.122s
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: quantizing model to /dev/shm/model_cache
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Saving flywheel model at /dev/shm/model_cache
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 29%|██▊ | 83/291 [00:01<00:03, 55.45it/s] Loading 0: 64%|██████▍ | 187/291 [00:02<00:01, 64.25it/s] Loading 0: 99%|█████████▊| 287/291 [00:03<00:00, 94.12it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: quantized model in 26.073s
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Processed model cgato/L3-TheSpice-8b-DPO-v3.2-e14.1 in 73.749s
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: creating bucket guanaco-mkml-models
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/cgato-l3-thespice-8b-dpo-2837-v3
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/cgato-l3-thespice-8b-dpo-2837-v3/config.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/cgato-l3-thespice-8b-dpo-2837-v3/special_tokens_map.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/cgato-l3-thespice-8b-dpo-2837-v3/tokenizer_config.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/cgato-l3-thespice-8b-dpo-2837-v3/tokenizer.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/cgato-l3-thespice-8b-dpo-2837-v3/flywheel_model.0.safetensors
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
cgato-l3-thespice-8b-dpo-2837-v3-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.
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: warnings.warn(
cgato-l3-thespice-8b-dpo-2837-v3-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-l3-thespice-8b-dpo-2837-v3-mkmlizer: warnings.warn(
cgato-l3-thespice-8b-dpo-2837-v3-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-l3-thespice-8b-dpo-2837-v3-mkmlizer: warnings.warn(
cgato-l3-thespice-8b-dpo-2837-v3-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-l3-thespice-8b-dpo-2837-v3-mkmlizer: return self.fget.__get__(instance, owner)()
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Saving duration: 0.420s
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 17.302s
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: creating bucket guanaco-reward-models
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: WARNING: Retrying failed request: /?location ([Errno 110] Connection timed out)
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: WARNING: Waiting 3 sec...
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: Bucket 's3://guanaco-reward-models/' created
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/cgato-l3-thespice-8b-dpo-2837-v3_reward
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/cgato-l3-thespice-8b-dpo-2837-v3_reward/config.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/cgato-l3-thespice-8b-dpo-2837-v3_reward/special_tokens_map.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/cgato-l3-thespice-8b-dpo-2837-v3_reward/tokenizer_config.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/cgato-l3-thespice-8b-dpo-2837-v3_reward/merges.txt
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/cgato-l3-thespice-8b-dpo-2837-v3_reward/vocab.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/cgato-l3-thespice-8b-dpo-2837-v3_reward/tokenizer.json
cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/cgato-l3-thespice-8b-dpo-2837-v3_reward/reward.tensors
Job cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer completed after 320.77s with status: succeeded
Stopping job with name cgato-l3-thespice-8b-dpo-2837-v3-mkmlizer
Pipeline stage MKMLizer completed in 321.68s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.09s
Running pipeline stage ISVCDeployer
Creating inference service cgato-l3-thespice-8b-dpo-2837-v3
Waiting for inference service cgato-l3-thespice-8b-dpo-2837-v3 to be ready
Inference service cgato-l3-thespice-8b-dpo-2837-v3 ready after 40.32782173156738s
Pipeline stage ISVCDeployer completed in 50.08s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.0385167598724365s
Received healthy response to inference request in 1.2488117218017578s
Received healthy response to inference request in 1.2563207149505615s
Received healthy response to inference request in 1.2776360511779785s
Received healthy response to inference request in 1.237999439239502s
5 requests
0 failed requests
5th percentile: 1.2401618957519531
10th percentile: 1.2423243522644043
20th percentile: 1.2466492652893066
30th percentile: 1.2503135204315186
40th percentile: 1.25331711769104
50th percentile: 1.2563207149505615
60th percentile: 1.2648468494415284
70th percentile: 1.2733729839324952
80th percentile: 1.4298121929168703
90th percentile: 1.7341644763946533
95th percentile: 1.8863406181335447
99th percentile: 2.0080815315246583
mean time: 1.4118569374084473
Pipeline stage StressChecker completed in 7.61s
Running pipeline stage DaemonicModelEvalScorer
Pipeline stage DaemonicModelEvalScorer completed in 0.03s
Running pipeline stage DaemonicSafetyScorer
Pipeline stage DaemonicSafetyScorer completed in 0.03s
Running M-Eval for topic stay_in_character
cgato-l3-thespice-8b-dpo_2837_v3 status is now deployed due to DeploymentManager action
M-Eval Dataset for topic stay_in_character is loaded
cgato-l3-thespice-8b-dpo_2837_v3 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics