zzyabc-mistral-dpo-qlora

developer_uid: ZZYABC

submission_id: zzyabc-mistral-dpo-qlora_v1

model_name: zzyabc-mistral-dpo-qlora_v1

model_group: ZZYABC/mistral-dpo-qlora

status: torndown

timestamp: 2024-12-20T09:33:46+00:00

num_battles: 9910

num_wins: 3367

celo_rating: 1146.56

family_friendly_score: 0.5998

family_friendly_standard_error: 0.006928779979188255

submission_type: basic

model_repo: ZZYABC/mistral-dpo-qlora

model_architecture: MistralForCausalLM

model_num_parameters: 7241748480.0

best_of: 8

max_input_tokens: 1024

max_output_tokens: 64

reward_model: default

latencies: [{'batch_size': 1, 'throughput': 0.9236410143785332, 'latency_mean': 1.082585448026657, 'latency_p50': 1.0817441940307617, 'latency_p90': 1.1967445850372314}, {'batch_size': 4, 'throughput': 2.030790709660645, 'latency_mean': 1.9656179821491242, 'latency_p50': 1.9551254510879517, 'latency_p90': 2.206605648994446}, {'batch_size': 5, 'throughput': 2.170883383949275, 'latency_mean': 2.295782535076141, 'latency_p50': 2.295929431915283, 'latency_p90': 2.5594462394714355}, {'batch_size': 8, 'throughput': 2.39504081128691, 'latency_mean': 3.3131088781356812, 'latency_p50': 3.3160362243652344, 'latency_p90': 3.7166967391967773}, {'batch_size': 10, 'throughput': 2.462328502722945, 'latency_mean': 4.028113087415695, 'latency_p50': 4.011290788650513, 'latency_p90': 4.605048942565918}, {'batch_size': 12, 'throughput': 2.520739141697224, 'latency_mean': 4.722755032777786, 'latency_p50': 4.676547169685364, 'latency_p90': 5.384634709358215}, {'batch_size': 15, 'throughput': 2.528714100608656, 'latency_mean': 5.8470933926105495, 'latency_p50': 5.832290887832642, 'latency_p90': 6.521022605895996}]

gpu_counts: {'NVIDIA RTX A5000': 1}

display_name: zzyabc-mistral-dpo-qlora_v1

is_internal_developer: False

language_model: ZZYABC/mistral-dpo-qlora

model_size: 7B

ranking_group: single

throughput_3p7s: 2.45

us_pacific_date: 2024-12-20

win_ratio: 0.33975782038345104

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': ['\n'], 'max_input_tokens': 1024, 'best_of': 8, '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': False}

Resubmit model

Shutdown handler not registered because Python interpreter is not running in the main thread
run pipeline %s
run pipeline stage %s
Running pipeline stage MKMLizer
Starting job with name zzyabc-mistral-dpo-qlora-v1-mkmlizer
Waiting for job on zzyabc-mistral-dpo-qlora-v1-mkmlizer to finish
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║     _____            __           __                                ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║    / _/ /_ ___    __/ /  ___ ___ / /                                ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║   / _/ / // / |/|/ / _ \/ -_) -_) /                                 ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  /_//_/\_, /|__,__/_//_/\__/\__/_/                                  ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║       /___/                                                         ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║                                                                     ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  Version: 0.11.12                                                   ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  Copyright 2023 MK ONE TECHNOLOGIES Inc.                            ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  https://mk1.ai                                                     ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║                                                                     ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  The license key for the current software has been verified as      ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  belonging to:                                                      ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║                                                                     ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  Chai Research Corp.                                                ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f                   ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║  Expiration: 2025-01-15 23:59:59                                    ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ║                                                                     ║
zzyabc-mistral-dpo-qlora-v1-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
zzyabc-mistral-dpo-qlora-v1-mkmlizer: Downloaded to shared memory in 26.510s
zzyabc-mistral-dpo-qlora-v1-mkmlizer: quantizing model to /dev/shm/model_cache, profile:s0, folder:/tmp/tmp_wjvyjxf, device:0
zzyabc-mistral-dpo-qlora-v1-mkmlizer: Saving flywheel model at /dev/shm/model_cache
zzyabc-mistral-dpo-qlora-v1-mkmlizer: /opt/conda/lib/python3.10/site-packages/mk1/flywheel/functional/loader.py:55: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
zzyabc-mistral-dpo-qlora-v1-mkmlizer:   tensors = torch.load(model_shard_filename, map_location=torch.device(self.device), mmap=True)
zzyabc-mistral-dpo-qlora-v1-mkmlizer: quantized model in 18.190s
zzyabc-mistral-dpo-qlora-v1-mkmlizer: Processed model ZZYABC/mistral-dpo-qlora in 44.701s
zzyabc-mistral-dpo-qlora-v1-mkmlizer: creating bucket guanaco-mkml-models
zzyabc-mistral-dpo-qlora-v1-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
zzyabc-mistral-dpo-qlora-v1-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/zzyabc-mistral-dpo-qlora-v1
zzyabc-mistral-dpo-qlora-v1-mkmlizer: cp /dev/shm/model_cache/config.json s3://guanaco-mkml-models/zzyabc-mistral-dpo-qlora-v1/config.json
zzyabc-mistral-dpo-qlora-v1-mkmlizer: cp /dev/shm/model_cache/added_tokens.json s3://guanaco-mkml-models/zzyabc-mistral-dpo-qlora-v1/added_tokens.json
zzyabc-mistral-dpo-qlora-v1-mkmlizer: cp /dev/shm/model_cache/special_tokens_map.json s3://guanaco-mkml-models/zzyabc-mistral-dpo-qlora-v1/special_tokens_map.json
zzyabc-mistral-dpo-qlora-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer_config.json s3://guanaco-mkml-models/zzyabc-mistral-dpo-qlora-v1/tokenizer_config.json
zzyabc-mistral-dpo-qlora-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.model s3://guanaco-mkml-models/zzyabc-mistral-dpo-qlora-v1/tokenizer.model
zzyabc-mistral-dpo-qlora-v1-mkmlizer: cp /dev/shm/model_cache/tokenizer.json s3://guanaco-mkml-models/zzyabc-mistral-dpo-qlora-v1/tokenizer.json
zzyabc-mistral-dpo-qlora-v1-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/zzyabc-mistral-dpo-qlora-v1/flywheel_model.0.safetensors
zzyabc-mistral-dpo-qlora-v1-mkmlizer: 
Loading 0:   0%|          | 0/291 [00:00<?, ?it/s]
Loading 0:   2%|▏         | 6/291 [00:00<00:04, 58.93it/s]
Loading 0:   4%|▍         | 13/291 [00:00<00:05, 55.08it/s]
Loading 0:   8%|▊         | 22/291 [00:00<00:04, 67.11it/s]
Loading 0:  11%|█         | 31/291 [00:00<00:03, 69.73it/s]
Loading 0:  14%|█▎        | 40/291 [00:00<00:03, 73.72it/s]
Loading 0:  17%|█▋        | 49/291 [00:00<00:03, 74.86it/s]
Loading 0:  20%|█▉        | 58/291 [00:00<00:03, 72.09it/s]
Loading 0:  23%|██▎       | 67/291 [00:00<00:03, 74.18it/s]
Loading 0:  26%|██▌       | 76/291 [00:01<00:02, 77.28it/s]
Loading 0:  29%|██▉       | 85/291 [00:01<00:02, 75.52it/s]
Loading 0:  32%|███▏      | 94/291 [00:01<00:02, 71.54it/s]
Loading 0:  35%|███▌      | 102/291 [00:02<00:09, 19.88it/s]
Loading 0:  37%|███▋      | 108/291 [00:02<00:08, 22.63it/s]
Loading 0:  40%|███▉      | 115/291 [00:02<00:06, 27.55it/s]
Loading 0:  43%|████▎     | 124/291 [00:02<00:04, 35.40it/s]
Loading 0:  46%|████▌     | 133/291 [00:02<00:03, 41.92it/s]
Loading 0:  49%|████▉     | 142/291 [00:03<00:03, 48.26it/s]
Loading 0:  52%|█████▏    | 151/291 [00:03<00:02, 54.21it/s]
Loading 0:  55%|█████▍    | 160/291 [00:03<00:02, 59.85it/s]
Loading 0:  58%|█████▊    | 169/291 [00:03<00:01, 63.23it/s]
Loading 0:  61%|██████    | 178/291 [00:03<00:01, 65.62it/s]
Loading 0:  64%|██████▍   | 187/291 [00:03<00:01, 68.90it/s]
Loading 0:  67%|██████▋   | 196/291 [00:03<00:01, 71.74it/s]
Loading 0:  70%|███████   | 204/291 [00:04<00:04, 21.35it/s]
Loading 0:  73%|███████▎  | 211/291 [00:04<00:03, 25.14it/s]
Loading 0:  76%|███████▌  | 220/291 [00:05<00:02, 32.03it/s]
Loading 0:  79%|███████▊  | 229/291 [00:05<00:01, 39.53it/s]
Loading 0:  82%|████████▏ | 238/291 [00:05<00:01, 47.40it/s]
Loading 0:  85%|████████▍ | 247/291 [00:05<00:00, 51.34it/s]
Loading 0:  88%|████████▊ | 256/291 [00:05<00:00, 57.05it/s]
Loading 0:  91%|█████████ | 265/291 [00:05<00:00, 63.27it/s]
Loading 0:  94%|█████████▍| 274/291 [00:05<00:00, 68.19it/s]
Loading 0:  97%|█████████▋| 283/291 [00:05<00:00, 69.44it/s]
Loading 0: 100%|██████████| 291/291 [00:07<00:00, 17.13it/s]
                                                            
Job zzyabc-mistral-dpo-qlora-v1-mkmlizer completed after 74.01s with status: succeeded
Stopping job with name zzyabc-mistral-dpo-qlora-v1-mkmlizer
Pipeline stage MKMLizer completed in 74.51s
run pipeline stage %s
Running pipeline stage MKMLTemplater
Pipeline stage MKMLTemplater completed in 0.15s
run pipeline stage %s
Running pipeline stage MKMLDeployer
Creating inference service zzyabc-mistral-dpo-qlora-v1
Waiting for inference service zzyabc-mistral-dpo-qlora-v1 to be ready
Inference service zzyabc-mistral-dpo-qlora-v1 ready after 251.639244556427s
Pipeline stage MKMLDeployer completed in 252.16s
run pipeline stage %s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.4080305099487305s
Received healthy response to inference request in 1.9946560859680176s
Received healthy response to inference request in 1.4255168437957764s
Received healthy response to inference request in 0.920440673828125s
Received healthy response to inference request in 1.2901041507720947s
5 requests
0 failed requests
5th percentile: 0.994373369216919
10th percentile: 1.068306064605713
20th percentile: 1.2161714553833007
30th percentile: 1.3171866893768311
40th percentile: 1.3713517665863038
50th percentile: 1.4255168437957764
60th percentile: 1.653172540664673
70th percentile: 1.8808282375335692
80th percentile: 2.07733097076416
90th percentile: 2.2426807403564455
95th percentile: 2.325355625152588
99th percentile: 2.391495532989502
mean time: 1.6077496528625488
Pipeline stage StressChecker completed in 9.36s
run pipeline stage %s
Running pipeline stage OfflineFamilyFriendlyTriggerPipeline
run_pipeline:run_in_cloud %s
starting trigger_guanaco_pipeline args=%s
triggered trigger_guanaco_pipeline args=%s
Pipeline stage OfflineFamilyFriendlyTriggerPipeline completed in 2.32s
run pipeline stage %s
Running pipeline stage TriggerMKMLProfilingPipeline
run_pipeline:run_in_cloud %s
starting trigger_guanaco_pipeline args=%s
triggered trigger_guanaco_pipeline args=%s
Pipeline stage TriggerMKMLProfilingPipeline completed in 2.09s
Shutdown handler de-registered
zzyabc-mistral-dpo-qlora_v1 status is now deployed due to DeploymentManager action
Shutdown handler registered
run pipeline %s
run pipeline stage %s
Running pipeline stage OfflineFamilyFriendlyScorer
Evaluating %s Family Friendly Score with %s threads
Pipeline stage OfflineFamilyFriendlyScorer completed in 2053.66s
Shutdown handler de-registered
zzyabc-mistral-dpo-qlora_v1 status is now inactive due to auto deactivation removed underperforming models
Deleting key zzyabc-mixtral-8x7b-v0-1-v1/tokenizer.json from bucket guanaco-mkml-models
zzyabc-mistral-dpo-qlora_v1 status is now torndown due to DeploymentManager action
zzyabc-mistral-dpo-qlora_v1 status is now torndown due to DeploymentManager action
zzyabc-mistral-dpo-qlora_v1 status is now torndown due to DeploymentManager action