plapre-pico-coreml / scripts /attention.py

Commit History

Performance improvements on iOS
ffa94c8

Daniel Rothmann commited on

Drop prefill model, add tokenizer config
95c6137

Daniel Rothmann commited on

Fix audio decoder, fix broken KV cache
d1bfb8c

Daniel Rothmann commited on

Fix a NaN attention issue in converted model
a2c97d7

Daniel Rothmann commited on

Add CoreML prefill and decode models
cb20bed

Daniel Rothmann commited on