trace2333-duduk-llama3-v1

developer_uid: Trace2333

submission_id: trace2333-duduk-llama3-v1_v4

model_name: trace2333-duduk-llama3-v1_v4

model_group: Trace2333/duduk_llama3_v

status: torndown

timestamp: 2024-07-23T10:40:48+00:00

num_battles: 19865

num_wins: 9503

celo_rating: 1174.97

family_friendly_score: 0.0

submission_type: basic

model_repo: Trace2333/duduk_llama3_v1

model_architecture: LlamaForCausalLM

reward_repo: Jellywibble/gpt2_xl_pairwise_89m_step_347634

model_num_parameters: 8030261248.0

best_of: 16

max_input_tokens: 512

max_output_tokens: 64

display_name: trace2333-duduk-llama3-v1_v4

is_internal_developer: False

language_model: Trace2333/duduk_llama3_v1

model_size: 8B

ranking_group: single

us_pacific_date: 2024-07-23

win_ratio: 0.47837905864585956

generation_params: {'temperature': 1.05, 'top_p': 1.0, 'min_p': 0.12, 'top_k': 40, 'presence_penalty': 0.0, 'frequency_penalty': 0.1, 'stopping_words': ['\n', '<|eot_id|>'], 'max_input_tokens': 512, 'best_of': 16, 'max_output_tokens': 64}

formatter: {'memory_template': "<|start_header_id|>system<|end_header_id|>\n\n{bot_name}'s Persona: {memory}\n\n", 'prompt_template': '{prompt}<|eot_id|>', 'bot_template': '<|start_header_id|>assistant<|end_header_id|>\n\n{bot_name}: {message}<|eot_id|>', 'user_template': '<|start_header_id|>user<|end_header_id|>\n\nYou: {message}<|eot_id|>', 'response_template': '<|start_header_id|>assistant<|end_header_id|>\n\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 trace2333-duduk-llama3-v1-v4-mkmlizer
Waiting for job on trace2333-duduk-llama3-v1-v4-mkmlizer to finish
trace2333-duduk-llama3-v1-v4-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
trace2333-duduk-llama3-v1-v4-mkmlizer: ║     _____            __           __                                ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║       /___/                                                         ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║                                                                     ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  Version: 0.9.5.post3                                               ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  https://mk1.ai                                                     ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║                                                                     ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  The license key for the current software has been verified as      ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  belonging to:                                                      ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║                                                                     ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  Chai Research Corp.                                                ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║  Expiration: 2024-10-15 23:59:59                                    ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ║                                                                     ║
trace2333-duduk-llama3-v1-v4-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
trace2333-duduk-llama3-v1-v4-mkmlizer: Downloaded to shared memory in 55.011s
trace2333-duduk-llama3-v1-v4-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmptpnfret2, device:0
trace2333-duduk-llama3-v1-v4-mkmlizer: Saving flywheel model at /dev/shm/model_cache
trace2333-duduk-llama3-v1-v4-mkmlizer: 
Loading 0:   0%|          | 0/291 [00:00<?, ?it/s]
Loading 0:   2%|▏         | 5/291 [00:00<00:06, 45.21it/s]
Loading 0:   5%|▍         | 14/291 [00:00<00:04, 62.33it/s]
Loading 0:   8%|▊         | 23/291 [00:00<00:03, 67.36it/s]
Loading 0:  11%|█▏        | 33/291 [00:00<00:05, 51.45it/s]
Loading 0:  13%|█▎        | 39/291 [00:00<00:05, 49.77it/s]
Loading 0:  16%|█▋        | 48/291 [00:00<00:04, 56.89it/s]
Loading 0:  20%|█▉        | 57/291 [00:00<00:03, 62.23it/s]
Loading 0:  23%|██▎       | 66/291 [00:01<00:03, 67.05it/s]
Loading 0:  25%|██▌       | 74/291 [00:01<00:03, 67.95it/s]
Loading 0:  28%|██▊       | 81/291 [00:01<00:04, 48.84it/s]
Loading 0:  31%|███       | 90/291 [00:01<00:03, 54.88it/s]
Loading 0:  34%|███▍      | 99/291 [00:01<00:03, 59.95it/s]
Loading 0:  37%|███▋      | 108/291 [00:01<00:02, 63.61it/s]
Loading 0:  40%|███▉      | 116/291 [00:01<00:02, 64.28it/s]
Loading 0:  43%|████▎     | 126/291 [00:02<00:02, 69.90it/s]
Loading 0:  46%|████▌     | 134/291 [00:02<00:03, 51.17it/s]
Loading 0:  48%|████▊     | 140/291 [00:02<00:03, 49.90it/s]
Loading 0:  51%|█████     | 149/291 [00:02<00:02, 55.90it/s]
Loading 0:  54%|█████▍    | 158/291 [00:02<00:02, 60.71it/s]
Loading 0:  57%|█████▋    | 167/291 [00:02<00:01, 63.90it/s]
Loading 0:  60%|██████    | 176/291 [00:02<00:01, 67.14it/s]
Loading 0:  64%|██████▍   | 186/291 [00:03<00:01, 57.69it/s]
Loading 0:  66%|██████▋   | 193/291 [00:03<00:01, 52.34it/s]
Loading 0:  69%|██████▉   | 202/291 [00:03<00:01, 57.62it/s]
Loading 0:  73%|███████▎  | 211/291 [00:03<00:01, 61.48it/s]
Loading 0:  76%|███████▌  | 220/291 [00:03<00:01, 64.69it/s]
Loading 0:  78%|███████▊  | 228/291 [00:03<00:00, 65.40it/s]
Loading 0:  81%|████████  | 235/291 [00:04<00:01, 51.01it/s]
Loading 0:  83%|████████▎ | 241/291 [00:04<00:00, 52.65it/s]
Loading 0:  85%|████████▌ | 248/291 [00:04<00:00, 54.27it/s]
Loading 0:  88%|████████▊ | 257/291 [00:04<00:00, 60.10it/s]
Loading 0:  91%|█████████▏| 266/291 [00:04<00:00, 64.62it/s]
Loading 0:  95%|█████████▍| 275/291 [00:04<00:00, 67.01it/s]
Loading 0:  98%|█████████▊| 286/291 [00:10<00:00,  5.26it/s]
                                                            
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
trace2333-duduk-llama3-v1-v4-mkmlizer: quantized model in 24.198s
trace2333-duduk-llama3-v1-v4-mkmlizer: Processed model Trace2333/duduk_llama3_v1 in 79.210s
trace2333-duduk-llama3-v1-v4-mkmlizer: creating bucket guanaco-mkml-models
trace2333-duduk-llama3-v1-v4-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
trace2333-duduk-llama3-v1-v4-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/trace2333-duduk-llama3-v1-v4
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/trace2333-duduk-llama3-v1-v4/config.json
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/trace2333-duduk-llama3-v1-v4/special_tokens_map.json
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/trace2333-duduk-llama3-v1-v4/tokenizer_config.json
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/trace2333-duduk-llama3-v1-v4/tokenizer.json
trace2333-duduk-llama3-v1-v4-mkmlizer: loading reward model from Jellywibble/gpt2_xl_pairwise_89m_step_347634
trace2333-duduk-llama3-v1-v4-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py:950: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
trace2333-duduk-llama3-v1-v4-mkmlizer:   warnings.warn(
trace2333-duduk-llama3-v1-v4-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/tokenization_auto.py:778: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
trace2333-duduk-llama3-v1-v4-mkmlizer:   warnings.warn(
trace2333-duduk-llama3-v1-v4-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py:469: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
trace2333-duduk-llama3-v1-v4-mkmlizer:   warnings.warn(
trace2333-duduk-llama3-v1-v4-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
trace2333-duduk-llama3-v1-v4-mkmlizer: Saving duration: 1.389s
trace2333-duduk-llama3-v1-v4-mkmlizer: Processed model Jellywibble/gpt2_xl_pairwise_89m_step_347634 in 11.496s
trace2333-duduk-llama3-v1-v4-mkmlizer: creating bucket guanaco-reward-models
trace2333-duduk-llama3-v1-v4-mkmlizer: Bucket 's3://guanaco-reward-models/' created
trace2333-duduk-llama3-v1-v4-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/trace2333-duduk-llama3-v1-v4_reward
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/trace2333-duduk-llama3-v1-v4_reward/config.json
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/trace2333-duduk-llama3-v1-v4_reward/special_tokens_map.json
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/trace2333-duduk-llama3-v1-v4_reward/tokenizer_config.json
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/trace2333-duduk-llama3-v1-v4_reward/merges.txt
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/trace2333-duduk-llama3-v1-v4_reward/vocab.json
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/trace2333-duduk-llama3-v1-v4_reward/tokenizer.json
trace2333-duduk-llama3-v1-v4-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/trace2333-duduk-llama3-v1-v4_reward/reward.tensors
Job trace2333-duduk-llama3-v1-v4-mkmlizer completed after 125.97s with status: succeeded
Stopping job with name trace2333-duduk-llama3-v1-v4-mkmlizer
Pipeline stage MKMLizer completed in 126.95s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.10s
Running pipeline stage ISVCDeployer
Creating inference service trace2333-duduk-llama3-v1-v4
Waiting for inference service trace2333-duduk-llama3-v1-v4 to be ready
Inference service trace2333-duduk-llama3-v1-v4 ready after 50.49602293968201s
Pipeline stage ISVCDeployer completed in 52.44s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.2538232803344727s
Received healthy response to inference request in 1.463155746459961s
Received healthy response to inference request in 1.4415783882141113s
Received healthy response to inference request in 1.3542368412017822s
Received healthy response to inference request in 1.3340604305267334s
5 requests
0 failed requests
5th percentile: 1.3380957126617432
10th percentile: 1.342130994796753
20th percentile: 1.3502015590667724
30th percentile: 1.3717051506042481
40th percentile: 1.4066417694091797
50th percentile: 1.4415783882141113
60th percentile: 1.4502093315124511
70th percentile: 1.458840274810791
80th percentile: 1.6212892532348635
90th percentile: 1.937556266784668
95th percentile: 2.0956897735595703
99th percentile: 2.222196578979492
mean time: 1.569370937347412
Pipeline stage StressChecker completed in 8.49s
trace2333-duduk-llama3-v1_v4 status is now deployed due to DeploymentManager action
trace2333-duduk-llama3-v1_v4 status is now inactive due to auto deactivation removed underperforming models
admin requested tearing down of trace2333-duduk-llama3-v1_v4
Running pipeline stage ISVCDeleter
Checking if service trace2333-duduk-llama3-v1-v4 is running
Tearing down inference service trace2333-duduk-llama3-v1-v4
Service trace2333-duduk-llama3-v1-v4 has been torndown
Pipeline stage ISVCDeleter completed in 5.14s
Running pipeline stage MKMLModelDeleter
Cleaning model data from S3
Cleaning model data from model cache
Deleting key trace2333-duduk-llama3-v1-v4/config.json from bucket guanaco-mkml-models
Deleting key trace2333-duduk-llama3-v1-v4/flywheel_model.0.safetensors from bucket guanaco-mkml-models
Deleting key trace2333-duduk-llama3-v1-v4/special_tokens_map.json from bucket guanaco-mkml-models
Deleting key trace2333-duduk-llama3-v1-v4/tokenizer.json from bucket guanaco-mkml-models
Deleting key trace2333-duduk-llama3-v1-v4/tokenizer_config.json from bucket guanaco-mkml-models
Cleaning model data from model cache
Deleting key trace2333-duduk-llama3-v1-v4_reward/config.json from bucket guanaco-reward-models
Deleting key trace2333-duduk-llama3-v1-v4_reward/merges.txt from bucket guanaco-reward-models
Deleting key trace2333-duduk-llama3-v1-v4_reward/reward.tensors from bucket guanaco-reward-models
Deleting key trace2333-duduk-llama3-v1-v4_reward/special_tokens_map.json from bucket guanaco-reward-models
Deleting key trace2333-duduk-llama3-v1-v4_reward/tokenizer.json from bucket guanaco-reward-models
Deleting key trace2333-duduk-llama3-v1-v4_reward/tokenizer_config.json from bucket guanaco-reward-models
Deleting key trace2333-duduk-llama3-v1-v4_reward/vocab.json from bucket guanaco-reward-models
Pipeline stage MKMLModelDeleter completed in 5.37s
trace2333-duduk-llama3-v1_v4 status is now torndown due to DeploymentManager action