submission_id: trace2333-duduk-llama2-v0_v1
developer_uid: Trace2333
status: inactive
model_repo: Trace2333/duduk_llama2_v0
reward_repo: ChaiML/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>'], 'max_input_tokens': 512, 'best_of': 4, 'max_output_tokens': 64}
formatter: {'memory_template': "{bot_name}'s Persona####: {memory}", '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': False}
timestamp: 2024-07-03T04:14:54+00:00
model_name: trace2333-duduk-llama2-v0_v1
model_group: Trace2333/duduk_llama2_v
num_battles: 15975
num_wins: 6306
celo_rating: 1106.75
propriety_score: 0.7241747962047307
propriety_total_count: 7483.0
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 6738415616.0
best_of: 4
max_input_tokens: 512
max_output_tokens: 64
display_name: trace2333-duduk-llama2-v0_v1
ineligible_reason: None
language_model: Trace2333/duduk_llama2_v0
model_size: 7B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
us_pacific_date: 2024-07-02
win_ratio: 0.39474178403755866
Resubmit model
Running pipeline stage MKMLizer
Starting job with name trace2333-duduk-llama2-v0-v1-mkmlizer
Waiting for job on trace2333-duduk-llama2-v0-v1-mkmlizer to finish
trace2333-duduk-llama2-v0-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ _____ __ __ ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ /___/ ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ Version: 0.8.14 ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ https://mk1.ai ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ The license key for the current software has been verified as ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ belonging to: ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ Chai Research Corp. ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ║ ║
trace2333-duduk-llama2-v0-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
trace2333-duduk-llama2-v0-v1-mkmlizer: Downloaded to shared memory in 59.685s
trace2333-duduk-llama2-v0-v1-mkmlizer: quantizing model to /dev/shm/model_cache
trace2333-duduk-llama2-v0-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
trace2333-duduk-llama2-v0-v1-mkmlizer: quantized model in 16.220s
trace2333-duduk-llama2-v0-v1-mkmlizer: Processed model Trace2333/duduk_llama2_v0 in 75.905s
trace2333-duduk-llama2-v0-v1-mkmlizer: creating bucket guanaco-mkml-models
trace2333-duduk-llama2-v0-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
trace2333-duduk-llama2-v0-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/trace2333-duduk-llama2-v0-v1
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/trace2333-duduk-llama2-v0-v1/config.json
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/trace2333-duduk-llama2-v0-v1/special_tokens_map.json
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/trace2333-duduk-llama2-v0-v1/tokenizer_config.json
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.model s3://guanaco-mkml-models/trace2333-duduk-llama2-v0-v1/tokenizer.model
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/trace2333-duduk-llama2-v0-v1/tokenizer.json
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/trace2333-duduk-llama2-v0-v1/flywheel_model.0.safetensors
trace2333-duduk-llama2-v0-v1-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
trace2333-duduk-llama2-v0-v1-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 3%|▎ | 9/291 [00:00<00:03, 86.75it/s] Loading 0: 6%|▌ | 18/291 [00:00<00:03, 83.88it/s] Loading 0: 9%|▉ | 27/291 [00:00<00:03, 84.39it/s] Loading 0: 12%|█▏ | 36/291 [00:00<00:03, 84.50it/s] Loading 0: 15%|█▌ | 45/291 [00:00<00:02, 84.35it/s] Loading 0: 19%|█▊ | 54/291 [00:00<00:05, 44.15it/s] Loading 0: 23%|██▎ | 66/291 [00:01<00:04, 56.18it/s] Loading 0: 25%|██▌ | 74/291 [00:01<00:03, 59.88it/s] Loading 0: 29%|██▊ | 83/291 [00:01<00:03, 65.75it/s] Loading 0: 32%|███▏ | 92/291 [00:01<00:02, 70.64it/s] Loading 0: 34%|███▍ | 100/291 [00:01<00:02, 72.38it/s] Loading 0: 37%|███▋ | 108/291 [00:01<00:04, 44.05it/s] Loading 0: 40%|████ | 117/291 [00:01<00:03, 51.61it/s] Loading 0: 43%|████▎ | 125/291 [00:02<00:02, 57.21it/s] Loading 0: 46%|████▌ | 134/291 [00:02<00:02, 63.67it/s] Loading 0: 49%|████▉ | 143/291 [00:02<00:02, 68.85it/s] Loading 0: 52%|█████▏ | 151/291 [00:02<00:01, 71.56it/s] Loading 0: 55%|█████▍ | 159/291 [00:02<00:02, 47.63it/s] Loading 0: 57%|█████▋ | 166/291 [00:02<00:02, 48.55it/s] Loading 0: 60%|██████ | 175/291 [00:02<00:02, 54.78it/s] Loading 0: 63%|██████▎ | 184/291 [00:03<00:01, 59.77it/s] Loading 0: 66%|██████▋ | 193/291 [00:03<00:01, 64.05it/s] Loading 0: 69%|██████▉ | 202/291 [00:03<00:01, 66.99it/s] Loading 0: 73%|███████▎ | 212/291 [00:03<00:01, 73.71it/s] Loading 0: 76%|███████▌ | 220/291 [00:03<00:01, 42.65it/s] Loading 0: 79%|███████▊ | 229/291 [00:03<00:01, 49.97it/s] Loading 0: 82%|████████▏ | 238/291 [00:04<00:00, 56.54it/s] Loading 0: 85%|████████▍ | 247/291 [00:04<00:00, 62.39it/s] Loading 0: 88%|████████▊ | 256/291 [00:04<00:00, 67.65it/s] Loading 0: 92%|█████████▏| 267/291 [00:05<00:01, 15.70it/s] Loading 0: 95%|█████████▍| 275/291 [00:06<00:00, 19.68it/s] Loading 0: 98%|█████████▊| 284/291 [00:06<00:00, 25.51it/s] /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.
trace2333-duduk-llama2-v0-v1-mkmlizer: warnings.warn(
trace2333-duduk-llama2-v0-v1-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`.
trace2333-duduk-llama2-v0-v1-mkmlizer: warnings.warn(
trace2333-duduk-llama2-v0-v1-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.
trace2333-duduk-llama2-v0-v1-mkmlizer: warnings.warn(
trace2333-duduk-llama2-v0-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.
trace2333-duduk-llama2-v0-v1-mkmlizer: warnings.warn(
trace2333-duduk-llama2-v0-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()
trace2333-duduk-llama2-v0-v1-mkmlizer: return self.fget.__get__(instance, owner)()
trace2333-duduk-llama2-v0-v1-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
trace2333-duduk-llama2-v0-v1-mkmlizer: Saving duration: 0.528s
trace2333-duduk-llama2-v0-v1-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 5.534s
trace2333-duduk-llama2-v0-v1-mkmlizer: creating bucket guanaco-reward-models
trace2333-duduk-llama2-v0-v1-mkmlizer: Bucket 's3://guanaco-reward-models/' created
trace2333-duduk-llama2-v0-v1-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/trace2333-duduk-llama2-v0-v1_reward
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/trace2333-duduk-llama2-v0-v1_reward/config.json
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/trace2333-duduk-llama2-v0-v1_reward/special_tokens_map.json
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/trace2333-duduk-llama2-v0-v1_reward/tokenizer_config.json
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/trace2333-duduk-llama2-v0-v1_reward/merges.txt
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/trace2333-duduk-llama2-v0-v1_reward/vocab.json
trace2333-duduk-llama2-v0-v1-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/trace2333-duduk-llama2-v0-v1_reward/tokenizer.json
Job trace2333-duduk-llama2-v0-v1-mkmlizer completed after 107.25s with status: succeeded
Stopping job with name trace2333-duduk-llama2-v0-v1-mkmlizer
Pipeline stage MKMLizer completed in 108.25s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.11s
Running pipeline stage ISVCDeployer
Creating inference service trace2333-duduk-llama2-v0-v1
Waiting for inference service trace2333-duduk-llama2-v0-v1 to be ready
Inference service trace2333-duduk-llama2-v0-v1 ready after 60.38237428665161s
Pipeline stage ISVCDeployer completed in 67.26s
Running pipeline stage StressChecker
Received healthy response to inference request in 1.5593297481536865s
Received healthy response to inference request in 0.8301169872283936s
Received healthy response to inference request in 0.8647174835205078s
Received healthy response to inference request in 0.8019201755523682s
Received healthy response to inference request in 0.8729360103607178s
5 requests
0 failed requests
5th percentile: 0.8075595378875733
10th percentile: 0.8131989002227783
20th percentile: 0.8244776248931884
30th percentile: 0.8370370864868164
40th percentile: 0.8508772850036621
50th percentile: 0.8647174835205078
60th percentile: 0.8680048942565918
70th percentile: 0.8712923049926757
80th percentile: 1.0102147579193117
90th percentile: 1.2847722530364991
95th percentile: 1.4220510005950926
99th percentile: 1.5318739986419678
mean time: 0.9858040809631348
Pipeline stage StressChecker completed in 5.72s
trace2333-duduk-llama2-v0_v1 status is now deployed due to DeploymentManager action
trace2333-duduk-llama2-v0_v1 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics