Chime: a small model for keyboard prediction
Chime predicts the next word and completes the word being typed in LocalType, my Android keyboard. It is a 2.22 MB model trained from scratch on licensed public text. It runs on the phone; the browser demo runs the same weights locally.
I measured model inference on a Pixel 9a and Pixel 10. A separate keyboard pilot tested suggestions, correction and text delivery; it did not control prior learning well enough to rank prediction accuracy.
Chime: training and inference
- Model
- 2.03M parameters
- Training text
- 161.6M words
- Training loop
- 2h 2m 23s · one A100
- Warm next-word p95
- 3.09–3.24 ms · two Pixels
Licensed public English text, mostly synthetic SODA dialogue plus human-written SMS, dialogues and prompts. Word count is before repetition; no private LocalType logs were used. Timings measure retained-context model inference, not typing latency.
Comparison with Gboard
On Pixel 10, LocalType offered the target next word in 10 of 36 cases; Gboard offered it in 18. On Pixel 9a the counts were 15 and 13. Earlier test attempts had given LocalType on the 9a and Gboard on the 10 more exposure to these words, with learning retained. Those counts describe the observed runs, but cannot establish which predictor is better.
The initial presentation gave this pilot too much weight. A comparison with controlled learning and fresh evaluation text is still needed. The complete results and limitations remain available. There is no established prediction or speed advantage over Gboard. Back correction remains a development priority.
Chime is English-only and has a fixed vocabulary. Its training text includes older Singapore SMS and mostly synthetic dialogue. Everyday prediction savings, battery use and physical typing latency remain unmeasured. Methods and full measurements describe the data, evaluation and numerical checks.