ViperEkura 8447f88f61 fix: size KV pool from prompt/gen args in benchmark ...

- Drop hardcoded CACHE_MAX_SEQ=2048 which overflowed at long prompts
- Size prefill pool to prompt_length and decode pool to prompt+5+gen*num_trials
- Unblocks decode/prefill benchmark at prompt 4096+ (was KV cache index OOB)

2026-08-02 16:25:27 +08:00

..

benchmark.py

fix: size KV pool from prompt/gen args in benchmark

2026-08-02 16:25:27 +08:00

generate.py

refactor: remove H5 storage backend in favor of mmap bin

2026-07-29 12:50:27 +08:00

preprocess.py

refactor: remove H5 storage backend in favor of mmap bin

2026-07-29 12:50:27 +08:00

server.py

feat : add setup_logging with hierarchical astrai logger

2026-07-27 08:13:48 +08:00

train.py

feat: add grad_snr metric with EMA-based gradient SNR tracking

2026-08-01 08:54:44 +08:00