This model is a fine-tuned version of Qwen/Qwen3-14B on the dpo_dataset_1, the dpo_dataset_2, the dpo_dataset_3, the dpo_dataset_4, the dpo_dataset_5 and the dpo_dataset_6 datasets.
It achieves the following results on the evaluation set:
Loss: 0.1193
Rewards/chosen: 4.6721
Rewards/rejected: -6.4223
Rewards/accuracies: 0.9362
Rewards/margins: 11.0944
Logps/chosen: -587.3591
Logps/rejected: -856.9067
Logits/chosen: -2.1841
Logits/rejected: -2.2938
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-06
train_batch_size: 2
eval_batch_size: 1
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 16
total_train_batch_size: 256
total_eval_batch_size: 8
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
