qwen25_7b_rm_eng_5e6_3ep

This model is a fine-tuned version of Jennny/qwen25_7b_sft_eng_math on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.3598
  • Accuracy: 0.6733

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: 4
  • eval_batch_size: 4
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 8
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 256
  • total_eval_batch_size: 32
  • optimizer: Use OptimizerNames.PAGED_ADAMW with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.03
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.1306 0.3347 10 1.1646 0.5933
0.0285 0.6695 20 1.1044 0.6333
0.039 1.0335 30 1.2131 0.61
0.0078 1.3682 40 1.3034 0.6567
0.0396 1.7029 50 1.6252 0.66
0.065 2.0669 60 0.9518 0.6867
0.0008 2.4017 70 1.2253 0.6533
0.0003 2.7364 80 1.3598 0.6733

Framework versions

  • Transformers 4.49.0
  • Pytorch 2.5.1+cu124
  • Datasets 3.3.2
  • Tokenizers 0.21.1
Downloads last month
-
Safetensors
Model size
7B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Jennny/qwen25_7b_rm_eng_5e6_3ep

Base model

Qwen/Qwen2.5-7B
Finetuned
(6)
this model