phi-3-mini-sft-indicqa-hindi-v0.1

model description

this model is a fine-tuned version of microsoft/Phi-3-mini-128k-instruct on the dataset ai4bharat-IndicQA-hi-202410.

prepared by @sepiatone.

intended uses & limitations

intended for an educational and non-commercial purpose.

training procedure

training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 1
• mixed_precision_training: Native AMP

library versions

• PEFT 0.13.2
• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 3.0.2
• Tokenizers 0.19.1