Kite

πŸŽ‰ You are looking at Kite 8.1, which is larger and uses ClimbMix!

Kite is a small, trained, 11 million parameter language model.

Training

It was trained on the first shard of Andrej Karpathy's shuffle of nvidia/Nemotron-ClimbMix, using 1 epoch, 12 batch size, 0.001 learning rate, and the pika 5 tokenizer.

Limitations

Due to its size, the model is not suitable for production workloads.

Downloads last month
184
Safetensors
Model size
11M params
Tensor type
F32
Β·
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Dataset used to train qikp/kite-8.1-11m-base