tulu2-13b-cost-tulumix-5e-7

This model is a fine-tuned version of allenai/tulu-2-13b on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6918
• Rewards/chosen: 0.0232
• Rewards/rejected: 0.0200
• Rewards/accuracies: 0.5540
• Rewards/margins: 0.0032
• Rewards/margins Max: 0.0628
• Rewards/margins Min: -0.0598
• Rewards/margins Std: 0.0407
• Logps/rejected: -313.6658
• Logps/chosen: -322.6469
• Logits/rejected: -0.8946
• Logits/chosen: -1.0169

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: 5e-07
• train_batch_size: 2
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Rewards/margins Max	Rewards/margins Min	Rewards/margins Std	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.6624	1.0	3174	0.6918	0.0232	0.0200	0.5540	0.0032	0.0628	-0.0598	0.0407	-313.6658	-322.6469	-0.8946	-1.0169

Framework versions

• PEFT 0.7.1
• Transformers 4.39.0.dev0
• Pytorch 2.1.2+cu121
• Datasets 2.14.6
• Tokenizers 0.15.2