brett b75813b0e1 fix: V-norm in memory_stats, SeedSequence PRNG, MSE compressed_size_bits ...

Subset of @brosequist's #90 commit 0fd5de9 — keeping the actual
fixes, deferring the streaming + serialization API surface until
a production caller exists.

Included:
- KVCacheCompressor.memory_stats() was omitting the float32 norm
  stored per V vector, inflating reported compression ratio. Adds
  v_bits_total += n_vectors * 32.
- TurboQuantMSE.compressed_size_bits() — was missing (TurboQuant
  already had it).
- Replaces seed + 1000 magic offset with
  np.random.SeedSequence(seed).spawn(2) for true PRNG independence
  between PolarQuant and QJL stages, and between K and V quantizers.

Deferred (not in this commit):
- compress_token() / get_compressed_cache() streaming API
- CompressedVector.to_bytes() / from_bytes() binary serialization
- CompressedKVCache.save() / load() npz serialization

2026-05-09 10:42:53 -05:00

..

__init__.py

feat: TurboQuant, KV cache compressor, utils, distortion suite #6 #7 #8 #10

2026-03-24 21:04:56 -05:00

codebook.py

test: achieve 100% code coverage across all modules

2026-03-24 21:18:46 -05:00

hw_replay.py

fix: address Codex + Gemini review findings

2026-03-26 09:34:29 -05:00

kv_cache.py

fix: V-norm in memory_stats, SeedSequence PRNG, MSE compressed_size_bits

2026-05-09 10:42:53 -05:00

outlier.py

feat: outlier channel strategy for 2.5-bit and 3.5-bit quantization #17

2026-03-24 21:20:55 -05:00

polar_quant.py

Add norm correction to Python prototype quantization path

2026-03-30 10:57:38 -05:00

qjl.py

feat: PolarQuant quantizer + QJL 1-bit quantizer #4 #5

2026-03-24 21:02:06 -05:00

rotation.py

feat: input validation, README, real model validation script #12 #18

2026-03-24 21:29:24 -05:00

turboquant.py

fix: V-norm in memory_stats, SeedSequence PRNG, MSE compressed_size_bits

2026-05-09 10:42:53 -05:00

utils.py

feat: TurboQuant, KV cache compressor, utils, distortion suite #6 #7 #8 #10

2026-03-24 21:04:56 -05:00