metadata
license: mit
tags:
- generated_from_trainer
model-index:
- name: indic-gpt
results: []
indic-gpt
This model is a fine-tuned version of gpt2 on an indian language (https://ai4bharat.iitm.ac.in/corpora) dataset. Sample Dataset is available here -https://huggingface.co/datasets/aashay96/indic-gpt/ It achieves the following results on the evaluation set:
- eval_loss: 8.1648
- eval_runtime: 203.8512
- eval_samples_per_second: 227.941
- eval_steps_per_second: 7.128
- epoch: 0.0
- step: 3
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: 0.0005
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1
- num_epochs: 1
- mixed_precision_training: Native AMP
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.0+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2