cpsc-wmle-0.93-v2

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

kejian/cpsc-wmle-0.93-v2

This model was trained from scratch on the tomekkorbak/detoxify-pile-chunk3-50000-100000, the tomekkorbak/detoxify-pile-chunk3-100000-150000, the tomekkorbak/detoxify-pile-chunk3-150000-200000, the tomekkorbak/detoxify-pile-chunk3-200000-250000, the tomekkorbak/detoxify-pile-chunk3-250000-300000, the tomekkorbak/detoxify-pile-chunk3-300000-350000, the tomekkorbak/detoxify-pile-chunk3-350000-400000, the tomekkorbak/detoxify-pile-chunk3-400000-450000, the tomekkorbak/detoxify-pile-chunk3-450000-500000, the tomekkorbak/detoxify-pile-chunk3-500000-550000, the tomekkorbak/detoxify-pile-chunk3-550000-600000, the tomekkorbak/detoxify-pile-chunk3-600000-650000, the tomekkorbak/detoxify-pile-chunk3-650000-700000, the tomekkorbak/detoxify-pile-chunk3-700000-750000, the tomekkorbak/detoxify-pile-chunk3-750000-800000, the tomekkorbak/detoxify-pile-chunk3-800000-850000, the tomekkorbak/detoxify-pile-chunk3-850000-900000, the tomekkorbak/detoxify-pile-chunk3-900000-950000, the tomekkorbak/detoxify-pile-chunk3-950000-1000000, the tomekkorbak/detoxify-pile-chunk3-1000000-1050000, the tomekkorbak/detoxify-pile-chunk3-1050000-1100000, the tomekkorbak/detoxify-pile-chunk3-1100000-1150000, the tomekkorbak/detoxify-pile-chunk3-1150000-1200000, the tomekkorbak/detoxify-pile-chunk3-1200000-1250000, the tomekkorbak/detoxify-pile-chunk3-1250000-1300000, the tomekkorbak/detoxify-pile-chunk3-1300000-1350000, the tomekkorbak/detoxify-pile-chunk3-1350000-1400000, the tomekkorbak/detoxify-pile-chunk3-1400000-1450000, the tomekkorbak/detoxify-pile-chunk3-1450000-1500000, the tomekkorbak/detoxify-pile-chunk3-1500000-1550000, the tomekkorbak/detoxify-pile-chunk3-1550000-1600000, the tomekkorbak/detoxify-pile-chunk3-1600000-1650000, the tomekkorbak/detoxify-pile-chunk3-1650000-1700000, the tomekkorbak/detoxify-pile-chunk3-1700000-1750000, the tomekkorbak/detoxify-pile-chunk3-1750000-1800000 and the tomekkorbak/detoxify-pile-chunk3-1800000-1850000 datasets.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0007
• train_batch_size: 64
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.01
• training_steps: 42724
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.23.0
• Pytorch 1.13.0+cu116
• Datasets 2.0.0
• Tokenizers 0.12.1

Full config

{'dataset': {'datasets': ['tomekkorbak/detoxify-pile-chunk3-50000-100000', 'tomekkorbak/detoxify-pile-chunk3-100000-150000', 'tomekkorbak/detoxify-pile-chunk3-150000-200000', 'tomekkorbak/detoxify-pile-chunk3-200000-250000', 'tomekkorbak/detoxify-pile-chunk3-250000-300000', 'tomekkorbak/detoxify-pile-chunk3-300000-350000', 'tomekkorbak/detoxify-pile-chunk3-350000-400000', 'tomekkorbak/detoxify-pile-chunk3-400000-450000', 'tomekkorbak/detoxify-pile-chunk3-450000-500000', 'tomekkorbak/detoxify-pile-chunk3-500000-550000', 'tomekkorbak/detoxify-pile-chunk3-550000-600000', 'tomekkorbak/detoxify-pile-chunk3-600000-650000', 'tomekkorbak/detoxify-pile-chunk3-650000-700000', 'tomekkorbak/detoxify-pile-chunk3-700000-750000', 'tomekkorbak/detoxify-pile-chunk3-750000-800000', 'tomekkorbak/detoxify-pile-chunk3-800000-850000', 'tomekkorbak/detoxify-pile-chunk3-850000-900000', 'tomekkorbak/detoxify-pile-chunk3-900000-950000', 'tomekkorbak/detoxify-pile-chunk3-950000-1000000', 'tomekkorbak/detoxify-pile-chunk3-1000000-1050000', 'tomekkorbak/detoxify-pile-chunk3-1050000-1100000', 'tomekkorbak/detoxify-pile-chunk3-1100000-1150000', 'tomekkorbak/detoxify-pile-chunk3-1150000-1200000', 'tomekkorbak/detoxify-pile-chunk3-1200000-1250000', 'tomekkorbak/detoxify-pile-chunk3-1250000-1300000', 'tomekkorbak/detoxify-pile-chunk3-1300000-1350000', 'tomekkorbak/detoxify-pile-chunk3-1350000-1400000', 'tomekkorbak/detoxify-pile-chunk3-1400000-1450000', 'tomekkorbak/detoxify-pile-chunk3-1450000-1500000', 'tomekkorbak/detoxify-pile-chunk3-1500000-1550000', 'tomekkorbak/detoxify-pile-chunk3-1550000-1600000', 'tomekkorbak/detoxify-pile-chunk3-1600000-1650000', 'tomekkorbak/detoxify-pile-chunk3-1650000-1700000', 'tomekkorbak/detoxify-pile-chunk3-1700000-1750000', 'tomekkorbak/detoxify-pile-chunk3-1750000-1800000', 'tomekkorbak/detoxify-pile-chunk3-1800000-1850000'], 'is_split_by_sentences': True}, 'generation': {'force_call_on': [42724], 'metrics_configs': [{}, {'n': 1}, {'n': 2}, {'n': 5}], 'scenario_configs': [{'generate_kwargs': {'do_sample': True, 'max_length': 128, 'min_length': 10, 'temperature': 0.7, 'top_k': 0, 'top_p': 0.9}, 'name': 'unconditional', 'num_samples': 2048}, {'generate_kwargs': {'do_sample': True, 'max_length': 128, 'min_length': 10, 'temperature': 0.7, 'top_k': 0, 'top_p': 0.9}, 'name': 'challenging_rtp', 'num_samples': 2048, 'prompts_path': 'resources/challenging_rtp.jsonl'}], 'scorer_config': {'device': 'cuda:0'}}, 'kl_gpt3_callback': {'force_call_on': [42724], 'gpt3_kwargs': {'model_name': 'davinci'}, 'max_tokens': 64, 'num_samples': 2048}, 'model': {'from_scratch': True, 'gpt2_config_kwargs': {'reorder_and_upcast_attn': True, 'scale_attn_by': True}, 'path_or_name': 'gpt2'}, 'objective': {'beta': 0.93, 'exponential': False, 'name': 'WeightedMLE'}, 'tokenizer': {'path_or_name': 'gpt2'}, 'training': {'dataloader_num_workers': 0, 'effective_batch_size': 64, 'evaluation_strategy': 'no', 'fp16': True, 'hub_model_id': 'kejian/cpsc-wmle-0.93-v2', 'hub_strategy': 'all_checkpoints', 'learning_rate': 0.0007, 'logging_first_step': True, 'logging_steps': 50, 'num_tokens': 2800000000.0, 'output_dir': 'training_output_0.93-v2', 'per_device_train_batch_size': 16, 'push_to_hub': True, 'remove_unused_columns': False, 'save_steps': 21362, 'save_strategy': 'steps', 'seed': 42, 'warmup_ratio': 0.01, 'weight_decay': 0.1}}

Wandb URL:

https://wandb.ai/kejian/uncategorized/runs/3uljk8vy