BanglaFast-500M-Foundation-SLM

This model is a fine-tuned version of HFShishir/BanglaFast-500M-Foundation-SLM on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 2.7931

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: 0.00015
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 64
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 50
  • training_steps: 758
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
2.9664 0.1595 200 2.9729
2.6844 0.3190 400 2.8750
2.8168 0.4785 600 2.8045
2.6220 0.6045 758 2.7931

Framework versions

  • Transformers 5.20.2
  • Pytorch 2.11.0+cu130
  • Datasets 4.8.5
  • Tokenizers 0.23.3
Downloads last month
236
Safetensors
Model size
0.5B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for HFShishir/BanglaFast-500M-Foundation-SLM

Unable to build the model tree, the base model loops to the model itself. Learn more.