submission_id: chaiml-phase2-winner-13b2_v256
developer_uid: robert_irvine
status: inactive
model_repo: ChaiML/phase2_winner_13b2
reward_repo: rirv938/gpt2_ties_merge_preference_plus_classic_e2_density_99
generation_params: {'temperature': 1.0733671330918084, 'top_p': 0.6971846333941389, 'top_k': 50, 'presence_penalty': 0.0, 'frequency_penalty': 0.312882778758545, 'stopping_words': ['\n'], 'max_input_tokens': 512, 'best_of': 32, '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}:'}
reward_formatter: {'memory_template': 'Memory: {memory}\n', 'prompt_template': '{prompt}\n', 'bot_template': 'Bot: {message}\n', 'user_template': 'User: {message}\n', 'response_template': 'Bot:'}
timestamp: 2024-02-26T22:58:04+00:00
model_name: chaiml-phase2-winner-13b2_v256
model_eval_status: pending
safety_score: None
entertaining: None
stay_in_character: None
user_preference: None
double_thumbs_up: 3201
thumbs_up: 4897
thumbs_down: 2184
num_battles: 175659
num_wins: 90273
win_ratio: 0.5139104742711731
celo_rating: 1163.57
Resubmit model
Running pipeline stage MKMLizer
Starting job with name chaiml-phase2-winner-13b2-v256-mkmlizer
Waiting for job on chaiml-phase2-winner-13b2-v256-mkmlizer to finish
chaiml-phase2-winner-13b2-v256-mkmlizer: .gitattributes: 0%| | 0.00/1.52k [00:00<?, ?B/s] .gitattributes: 100%|██████████| 1.52k/1.52k [00:00<00:00, 18.0MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: README.md: 0%| | 0.00/1.53k [00:00<?, ?B/s] README.md: 100%|██████████| 1.53k/1.53k [00:00<00:00, 12.0MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: config.json: 0%| | 0.00/651 [00:00<?, ?B/s] config.json: 100%|██████████| 651/651 [00:00<00:00, 6.19MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: generation_config.json: 0%| | 0.00/170 [00:00<?, ?B/s] generation_config.json: 100%|██████████| 170/170 [00:00<00:00, 1.37MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: pytorch_model-00002-of-00003.bin: 0%| | 0.00/9.90G [00:00<?, ?B/s] pytorch_model-00002-of-00003.bin: 0%| | 10.5M/9.90G [00:00<02:26, 67.7MB/s] pytorch_model-00002-of-00003.bin: 1%| | 83.9M/9.90G [00:00<00:32, 301MB/s] pytorch_model-00002-of-00003.bin: 1%|▏ | 147M/9.90G [00:00<00:24, 400MB/s] pytorch_model-00002-of-00003.bin: 2%|▏ | 199M/9.90G [00:00<00:23, 405MB/s] pytorch_model-00002-of-00003.bin: 3%|▎ | 336M/9.90G [00:00<00:16, 595MB/s] pytorch_model-00002-of-00003.bin: 4%|▍ | 430M/9.90G [00:00<00:14, 672MB/s] pytorch_model-00002-of-00003.bin: 6%|▌ | 566M/9.90G [00:00<00:10, 849MB/s] pytorch_model-00002-of-00003.bin: 7%|▋ | 661M/9.90G [00:01<00:10, 849MB/s] pytorch_model-00002-of-00003.bin: 8%|▊ | 828M/9.90G [00:01<00:08, 1.08GB/s] pytorch_model-00002-of-00003.bin: 11%|█ | 1.05G/9.90G [00:01<00:06, 1.37GB/s] pytorch_model-00002-of-00003.bin: 12%|█▏ | 1.20G/9.90G [00:01<00:07, 1.22GB/s] pytorch_model-00002-of-00003.bin: 13%|█▎ | 1.33G/9.90G [00:01<00:08, 1.06GB/s] pytorch_model-00002-of-00003.bin: 16%|█▌ | 1.58G/9.90G [00:01<00:05, 1.41GB/s] pytorch_model-00002-of-00003.bin: 19%|█▊ | 1.85G/9.90G [00:01<00:04, 1.66GB/s] pytorch_model-00002-of-00003.bin: 20%|██ | 2.02G/9.90G [00:02<00:06, 1.22GB/s] pytorch_model-00002-of-00003.bin: 22%|██▏ | 2.18G/9.90G [00:02<00:05, 1.29GB/s] pytorch_model-00002-of-00003.bin: 24%|██▎ | 2.34G/9.90G [00:02<00:06, 1.16GB/s] pytorch_model-00002-of-00003.bin: 25%|██▍ | 2.47G/9.90G [00:02<00:06, 1.20GB/s] pytorch_model-00002-of-00003.bin: 27%|██▋ | 2.69G/9.90G [00:02<00:05, 1.42GB/s] pytorch_model-00002-of-00003.bin: 29%|██▉ | 2.85G/9.90G [00:02<00:05, 1.30GB/s] pytorch_model-00002-of-00003.bin: 30%|███ | 3.00G/9.90G [00:02<00:05, 1.25GB/s] pytorch_model-00002-of-00003.bin: 32%|███▏ | 3.18G/9.90G [00:02<00:04, 1.37GB/s] pytorch_model-00002-of-00003.bin: 34%|███▍ | 3.36G/9.90G [00:02<00:04, 1.46GB/s] pytorch_model-00002-of-00003.bin: 36%|███▌ | 3.54G/9.90G [00:03<00:04, 1.52GB/s] pytorch_model-00002-of-00003.bin: 38%|███▊ | 3.77G/9.90G [00:03<00:03, 1.73GB/s] pytorch_model-00002-of-00003.bin: 40%|████ | 3.96G/9.90G [00:03<00:03, 1.56GB/s] pytorch_model-00002-of-00003.bin: 42%|████▏ | 4.15G/9.90G [00:03<00:03, 1.64GB/s] pytorch_model-00002-of-00003.bin: 44%|████▎ | 4.33G/9.90G [00:03<00:03, 1.60GB/s] pytorch_model-00002-of-00003.bin: 46%|████▌ | 4.56G/9.90G [00:03<00:03, 1.77GB/s] pytorch_model-00002-of-00003.bin: 48%|████▊ | 4.76G/9.90G [00:03<00:02, 1.80GB/s] pytorch_model-00002-of-00003.bin: 52%|█████▏ | 5.15G/9.90G [00:03<00:02, 2.37GB/s] pytorch_model-00002-of-00003.bin: 55%|█████▍ | 5.40G/9.90G [00:04<00:02, 2.20GB/s] pytorch_model-00002-of-00003.bin: 57%|█████▋ | 5.64G/9.90G [00:04<00:01, 2.21GB/s] pytorch_model-00002-of-00003.bin: 59%|█████▉ | 5.87G/9.90G [00:04<00:02, 1.94GB/s] pytorch_model-00002-of-00003.bin: 62%|██████▏ | 6.18G/9.90G [00:04<00:01, 2.21GB/s] pytorch_model-00002-of-00003.bin: 65%|██████▍ | 6.42G/9.90G [00:04<00:01, 2.25GB/s] pytorch_model-00002-of-00003.bin: 67%|██████▋ | 6.66G/9.90G [00:04<00:01, 1.98GB/s] pytorch_model-00002-of-00003.bin: 69%|██████▉ | 6.87G/9.90G [00:04<00:01, 1.81GB/s] pytorch_model-00002-of-00003.bin: 72%|███████▏ | 7.12G/9.90G [00:04<00:01, 1.97GB/s] pytorch_model-00002-of-00003.bin: 77%|███████▋ | 7.58G/9.90G [00:04<00:00, 2.65GB/s] pytorch_model-00002-of-00003.bin: 79%|███████▉ | 7.86G/9.90G [00:05<00:00, 2.32GB/s] pytorch_model-00002-of-00003.bin: 82%|████████▏ | 8.17G/9.90G [00:05<00:00, 2.49GB/s] pytorch_model-00002-of-00003.bin: 86%|████████▌ | 8.48G/9.90G [00:05<00:00, 2.65GB/s] pytorch_model-00002-of-00003.bin: 89%|████████▊ | 8.77G/9.90G [00:05<00:00, 1.59GB/s] pytorch_model-00002-of-00003.bin: 91%|█████████ | 8.99G/9.90G [00:05<00:00, 1.58GB/s] pytorch_model-00002-of-00003.bin: 93%|█████████▎| 9.20G/9.90G [00:05<00:00, 1.68GB/s] pytorch_model-00002-of-00003.bin: 95%|█████████▌| 9.46G/9.90G [00:06<00:00, 1.85GB/s] pytorch_model-00002-of-00003.bin: 98%|█████████▊| 9.68G/9.90G [00:06<00:00, 1.54GB/s] pytorch_model-00002-of-00003.bin: 100%|█████████▉| 9.86G/9.90G [00:06<00:00, 692MB/s] pytorch_model-00002-of-00003.bin: 100%|█████████▉| 9.90G/9.90G [00:13<00:00, 732MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: pytorch_model-00003-of-00003.bin: 0%| | 0.00/6.18G [00:00<?, ?B/s] pytorch_model-00003-of-00003.bin: 0%| | 10.5M/6.18G [00:00<03:19, 31.0MB/s] pytorch_model-00003-of-00003.bin: 1%| | 73.4M/6.18G [00:00<00:29, 204MB/s] pytorch_model-00003-of-00003.bin: 3%|▎ | 189M/6.18G [00:00<00:13, 434MB/s] pytorch_model-00003-of-00003.bin: 4%|▍ | 273M/6.18G [00:00<00:10, 542MB/s] pytorch_model-00003-of-00003.bin: 7%|▋ | 419M/6.18G [00:00<00:07, 800MB/s] pytorch_model-00003-of-00003.bin: 9%|▉ | 556M/6.18G [00:00<00:05, 948MB/s] pytorch_model-00003-of-00003.bin: 11%|█ | 692M/6.18G [00:00<00:05, 1.06GB/s] pytorch_model-00003-of-00003.bin: 13%|█▎ | 828M/6.18G [00:01<00:04, 1.14GB/s] pytorch_model-00003-of-00003.bin: 16%|█▌ | 986M/6.18G [00:01<00:04, 1.23GB/s] pytorch_model-00003-of-00003.bin: 20%|█▉ | 1.21G/6.18G [00:01<00:03, 1.51GB/s] pytorch_model-00003-of-00003.bin: 23%|██▎ | 1.43G/6.18G [00:01<00:02, 1.69GB/s] pytorch_model-00003-of-00003.bin: 26%|██▌ | 1.61G/6.18G [00:01<00:02, 1.75GB/s] pytorch_model-00003-of-00003.bin: 29%|██▉ | 1.79G/6.18G [00:01<00:04, 1.01GB/s] pytorch_model-00003-of-00003.bin: 34%|███▍ | 2.09G/6.18G [00:01<00:02, 1.36GB/s] pytorch_model-00003-of-00003.bin: 37%|███▋ | 2.28G/6.18G [00:02<00:02, 1.45GB/s] pytorch_model-00003-of-00003.bin: 40%|███▉ | 2.45G/6.18G [00:02<00:02, 1.46GB/s] pytorch_model-00003-of-00003.bin: 44%|████▍ | 2.74G/6.18G [00:02<00:01, 1.78GB/s] pytorch_model-00003-of-00003.bin: 48%|████▊ | 2.95G/6.18G [00:02<00:01, 1.77GB/s] pytorch_model-00003-of-00003.bin: 52%|█████▏ | 3.20G/6.18G [00:02<00:01, 1.93GB/s] pytorch_model-00003-of-00003.bin: 55%|█████▌ | 3.42G/6.18G [00:02<00:01, 1.94GB/s] pytorch_model-00003-of-00003.bin: 59%|█████▉ | 3.64G/6.18G [00:02<00:01, 1.97GB/s] pytorch_model-00003-of-00003.bin: 62%|██████▏ | 3.85G/6.18G [00:02<00:01, 1.97GB/s] pytorch_model-00003-of-00003.bin: 66%|██████▌ | 4.06G/6.18G [00:02<00:01, 2.01GB/s] pytorch_model-00003-of-00003.bin: 69%|██████▉ | 4.27G/6.18G [00:03<00:00, 1.93GB/s] pytorch_model-00003-of-00003.bin: 72%|███████▏ | 4.47G/6.18G [00:03<00:00, 1.74GB/s] pytorch_model-00003-of-00003.bin: 78%|███████▊ | 4.81G/6.18G [00:03<00:00, 2.19GB/s] pytorch_model-00003-of-00003.bin: 83%|████████▎ | 5.12G/6.18G [00:03<00:00, 2.37GB/s] pytorch_model-00003-of-00003.bin: 87%|████████▋ | 5.37G/6.18G [00:03<00:00, 2.29GB/s] pytorch_model-00003-of-00003.bin: 91%|█████████ | 5.61G/6.18G [00:03<00:00, 2.07GB/s] pytorch_model-00003-of-00003.bin: 96%|█████████▋| 5.96G/6.18G [00:03<00:00, 2.41GB/s] pytorch_model-00003-of-00003.bin: 100%|█████████▉| 6.18G/6.18G [00:05<00:00, 1.22GB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: pytorch_model.bin.index.json: 0%| | 0.00/33.4k [00:00<?, ?B/s] pytorch_model.bin.index.json: 100%|██████████| 33.4k/33.4k [00:00<00:00, 161MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: special_tokens_map.json: 0%| | 0.00/414 [00:00<?, ?B/s] special_tokens_map.json: 100%|██████████| 414/414 [00:00<00:00, 3.45MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: tokenizer.json: 0%| | 0.00/1.84M [00:00<?, ?B/s] tokenizer.json: 100%|██████████| 1.84M/1.84M [00:00<00:00, 16.6MB/s] tokenizer.json: 100%|██████████| 1.84M/1.84M [00:00<00:00, 16.5MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: tokenizer.model: 0%| | 0.00/500k [00:00<?, ?B/s] tokenizer.model: 100%|██████████| 500k/500k [00:00<00:00, 5.84MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: tokenizer_config.json: 0%| | 0.00/725 [00:00<?, ?B/s] tokenizer_config.json: 100%|██████████| 725/725 [00:00<00:00, 9.10MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: Downloaded to shared memory in 30.562s
chaiml-phase2-winner-13b2-v256-mkmlizer: quantizing model to /dev/shm/model_cache
chaiml-phase2-winner-13b2-v256-mkmlizer: Saving mkml model at /dev/shm/model_cache
chaiml-phase2-winner-13b2-v256-mkmlizer: Reading /tmp/tmp17cwxhei/pytorch_model.bin.index.json
chaiml-phase2-winner-13b2-v256-mkmlizer: Profiling: 0%| | 0/363 [00:00<?, ?it/s] Profiling: 0%| | 1/363 [00:04<26:44, 4.43s/it] Profiling: 38%|███▊ | 139/363 [00:06<00:08, 25.46it/s] Profiling: 77%|███████▋ | 278/363 [00:08<00:01, 44.63it/s] Profiling: 100%|██████████| 363/363 [00:09<00:00, 46.12it/s] Profiling: 100%|██████████| 363/363 [00:09<00:00, 36.38it/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: quantized model in 30.129s
chaiml-phase2-winner-13b2-v256-mkmlizer: Processed model ChaiML/phase2_winner_13b2 in 62.452s
chaiml-phase2-winner-13b2-v256-mkmlizer: creating bucket guanaco-mkml-models
chaiml-phase2-winner-13b2-v256-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
chaiml-phase2-winner-13b2-v256-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v256
chaiml-phase2-winner-13b2-v256-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v256/config.json
chaiml-phase2-winner-13b2-v256-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v256/tokenizer_config.json
chaiml-phase2-winner-13b2-v256-mkmlizer: cp /dev/shm/model_cache/tokenizer.model s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v256/tokenizer.model
chaiml-phase2-winner-13b2-v256-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v256/tokenizer.json
chaiml-phase2-winner-13b2-v256-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v256/special_tokens_map.json
chaiml-phase2-winner-13b2-v256-mkmlizer: cp /dev/shm/model_cache/mkml_model.tensors s3://guanaco-mkml-models/chaiml-phase2-winner-13b2-v256/mkml_model.tensors
chaiml-phase2-winner-13b2-v256-mkmlizer: loading reward model from rirv938/gpt2_ties_merge_preference_plus_classic_e2_density_99
chaiml-phase2-winner-13b2-v256-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py:1067: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
chaiml-phase2-winner-13b2-v256-mkmlizer: warnings.warn(
chaiml-phase2-winner-13b2-v256-mkmlizer: config.json: 0%| | 0.00/983 [00:00<?, ?B/s] config.json: 100%|██████████| 983/983 [00:00<00:00, 9.86MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/tokenization_auto.py:690: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
chaiml-phase2-winner-13b2-v256-mkmlizer: warnings.warn(
chaiml-phase2-winner-13b2-v256-mkmlizer: tokenizer_config.json: 0%| | 0.00/445 [00:00<?, ?B/s] tokenizer_config.json: 100%|██████████| 445/445 [00:00<00:00, 3.23MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: vocab.json: 0%| | 0.00/798k [00:00<?, ?B/s] vocab.json: 100%|██████████| 798k/798k [00:00<00:00, 10.2MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: merges.txt: 0%| | 0.00/456k [00:00<?, ?B/s] merges.txt: 100%|██████████| 456k/456k [00:00<00:00, 66.5MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: tokenizer.json: 0%| | 0.00/2.11M [00:00<?, ?B/s] tokenizer.json: 100%|██████████| 2.11M/2.11M [00:00<00:00, 26.5MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: special_tokens_map.json: 0%| | 0.00/441 [00:00<?, ?B/s] special_tokens_map.json: 100%|██████████| 441/441 [00:00<00:00, 6.49MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py:472: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
chaiml-phase2-winner-13b2-v256-mkmlizer: warnings.warn(
chaiml-phase2-winner-13b2-v256-mkmlizer: model.safetensors.index.json: 0%| | 0.00/10.5k [00:00<?, ?B/s] model.safetensors.index.json: 100%|██████████| 10.5k/10.5k [00:00<00:00, 53.5MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: Downloading shards: 0%| | 0/1 [00:00<?, ?it/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: model-00001-of-00001.safetensors: 0%| | 0.00/249M [00:00<?, ?B/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: model-00001-of-00001.safetensors: 4%|▍ | 10.5M/249M [00:00<00:02, 87.7MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: model-00001-of-00001.safetensors: 45%|████▌ | 113M/249M [00:00<00:00, 524MB/s]  model-00001-of-00001.safetensors: 100%|█████████▉| 249M/249M [00:00<00:00, 774MB/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: Downloading shards: 100%|██████████| 1/1 [00:00<00:00, 1.70it/s] Downloading shards: 100%|██████████| 1/1 [00:00<00:00, 1.70it/s]
chaiml-phase2-winner-13b2-v256-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
Job chaiml-phase2-winner-13b2-v256-mkmlizer completed after 102.45s with status: succeeded
Stopping job with name chaiml-phase2-winner-13b2-v256-mkmlizer
Pipeline stage MKMLizer completed in 105.23s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.44s
Running pipeline stage ISVCDeployer
Creating inference service chaiml-phase2-winner-13b2-v256
Waiting for inference service chaiml-phase2-winner-13b2-v256 to be ready
Inference service chaiml-phase2-winner-13b2-v256 ready after 51.66350197792053s
Pipeline stage ISVCDeployer completed in 59.17s
Running pipeline stage StressChecker
Received healthy response to inference request in 3.4495761394500732s
Received healthy response to inference request in 2.4366071224212646s
Received healthy response to inference request in 2.5742249488830566s
Received healthy response to inference request in 2.4387662410736084s
Received healthy response to inference request in 2.4950389862060547s
5 requests
0 failed requests
5th percentile: 2.437038946151733
10th percentile: 2.4374707698822022
20th percentile: 2.43833441734314
30th percentile: 2.4500207901000977
40th percentile: 2.472529888153076
50th percentile: 2.4950389862060547
60th percentile: 2.5267133712768555
70th percentile: 2.5583877563476562
80th percentile: 2.74929518699646
90th percentile: 3.0994356632232667
95th percentile: 3.2745059013366697
99th percentile: 3.4145620918273925
mean time: 2.6788426876068114
Pipeline stage StressChecker completed in 16.55s
Running pipeline stage DaemonicModelEvalScorer
Pipeline stage DaemonicModelEvalScorer completed in 0.13s
Running pipeline stage DaemonicSafetyScorer
Running M-Eval for topic stay_in_character
Pipeline stage DaemonicSafetyScorer completed in 0.17s
%s, retrying in %s seconds...
M-Eval Dataset for topic stay_in_character is loaded
chaiml-phase2-winner-13b2_v256 status is now inactive due to auto deactivation removed underperforming models
chaiml-phase2-winner-13b2_v256 status is now deployed due to admin request
chaiml-phase2-winner-13b2_v256 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics