seanyhan/qwen-3-8b-base-r-dpo-ultrafeedback-4xH200-batch-128-rerun-2-runpod
09
<!-- 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. -->
qwen3-8b-base-r-dpo-ultrafeedback-4xh200-batch-128
This model is a fine-tuned version of jackf857/qwen3-8b-base-sft-ultrachat-4xh200-batch-128 on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:
- Loss: 0.5469
- R Dpo/chosen Len: 294.4800
- R Dpo/rejected Len: 249.8700
- R Dpo/length Delta: 44.6100
- R Dpo/regularization Term: 4.4610
- Logps/chosen: -2936.7747
- Logps/rejected: -2579.1685
- Logps/ref Chosen: -281.4850
- Logps/ref Rejected: -261.8005
- Logits/chosen: 0.6320
- Logits/rejected: 0.5334
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
- trainbatchsize: 4
- evalbatchsize: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradientaccumulationsteps: 8
- totaltrainbatch_size: 128
- totalevalbatch_size: 16
- optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
- lrschedulertype: cosine
- lrschedulerwarmup_ratio: 0.1
- num_epochs: 1
Training results
Framework versions
- Transformers 4.51.0
- Pytorch 2.3.1+cu121
- Datasets 2.21.0
- Tokenizers 0.21.4
