Views
No views yet
mixed_allocator (3-bit-class)Qwen/Qwen2.5-32B-Instruct, produced by the project's Tail-Aware
Quantization (TAQ) pipeline — the per-layer mixed-precision allocator's own choice of action (outlier/rotation_outlier at 2/3/4-bit) at a natural ~3-bit budget.| bpw | PPL | KLD vs fp16 | |
|---|---|---|---|
This checkpoint (mixed_allocator_3bit) | 3.8116 | 7.506 | 0.4631 |
outlier_3bit (uniform baseline, same budget class) | 3.8116 | 8.176 | 0.5431 |
rotation_outlier_3bit (uniform baseline, same budget class) | 3.8259 | 8.144 | 0.5681 |
rshoemake/taq-qwen14b-unsloth-matched:
packed/*.bin per-layer files (header + fp16 codebooks + bit-packed
indices + outlier side-channel), packed/manifest.json (per-layer
action/bits/measured bpw), fp16_passthrough/*.safetensors for any
layers the allocator left at full precision, unpack_outlier.py /
unpack_rotation_outlier.py standalone unpackers.Qwen/Qwen2.5-32B-Instruct (fp16), packed directly from the
original safetensors shards.run_full_matrix.py / run_tail_family_scale.py
(project fractal_quant_decision, private) — native family-scale
protocol (WikiText-2, passage-level KLD).pack_outlier_format.py + pack_rotation_outlier.py +
pack_whole_model.py.