metadata

library_name: peft
tags:
  - trl
  - dpo
  - generated_from_trainer
base_model: allenai/tulu-2-7b
model-index:
  - name: tulu2-7b-cost-UF-UI-judge13b-5e-7
    results: []

tulu2-7b-cost-UF-UI-judge13b-5e-7

This model is a fine-tuned version of allenai/tulu-2-7b on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.6929
Rewards/chosen: 0.0029
Rewards/rejected: 0.0023
Rewards/accuracies: 0.5365
Rewards/margins: 0.0006
Rewards/margins Max: 0.0723
Rewards/margins Min: -0.0823
Rewards/margins Std: 0.0504
Logps/rejected: -318.0156
Logps/chosen: -337.8507
Logits/rejected: 0.8998
Logits/chosen: 0.7298

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: 4
gradient_accumulation_steps: 2
total_train_batch_size: 16
total_eval_batch_size: 32
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
0.6604	1.0	1724	0.6929	0.0029	0.0023	0.5365	0.0006	0.0723	-0.0823	0.0504	-318.0156	-337.8507	0.8998	0.7298

Framework versions

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