Aaquib/gemma-3-1b-sft-alpaca

Model Card for Model ID

SFT'd version of google/gemma-3-1b-pt. Training performed solely on yahma/alpaca-cleaned. No further learning was performed.

Model Details

Hyperparameters to replicate:

• lr=1e-5
• num_epochs=1
• train_batch_size=40
• test_batch_size=32
• max_seq_len=256

Model Description

• Finetuned from model: [google/gemma-3-1b-pt]