dapack — language pack (graded precision, 10.7 GB)
Qwen3.5-35B-A3B compressed for language and general-text work: 10.7 GB instead of 21.2 GB, with every expert still present. The 143 experts this domain routes to are kept at q2_K; the remaining 113 are held at IQ2_XXS (2.06 bpw) under an importance matrix — graded, not deleted, so the pack degrades out of domain instead of breaking.
⚠️ This file requires the dapack runtime
Graded packs carry two expert tensors per layer at different precisions (
ffn_*_exps_cold+dapack.hot_experts_per_layer). Stock llama.cpp / ollama / LM Studio cannot load them. Build the runtime from:https://github.com/R3n3r0/dapack — ready-to-run binaries under Releases (Linux x86_64, ROCm), nothing to compile:
tar xzf dapack-v0.1.0-linux-x86_64-rocm-gfx1151.tar.gz cd dapack-v0.1.0-* && ./dapack serve ~/catalog.dapack # web chat + OpenAI API + routingOn other GPUs, build from the same repo (one command,
scripts/rebuild_fork.sh --applyafter a recursive clone).
Measured capabilities
These numbers are behavioural measurements, not estimates. They ship in the pack's manifest, and the dapack router treats them as hard constraints — a request needing a capability this pack has lost is routed to a pack that has it.
| capability | this pack | full model (21.2 GB) |
|---|---|---|
| reasoning (GSM8K convergence, 4096-token budget, n=100) | 82.0% | 77.0% |
| tool calling (8 probes) | 8/8 | 8/8 |
| translation en→it | 100% | 100% |
| instruction following | 100% | 80% |
| long context (needle @ 3k tokens) | 100% | 100% |
| structured output (JSON) | 100% | 100% |
| code generation | 17% ⚠️ | 100% |
What it lost — on purpose, and declared: code generation collapsed (100% → 17%), because the experts that write code were held at 2 bits by a language calibration. Under the dapack router, code requests never reach this pack. If you serve this file standalone, do not use it for code.
Why graded instead of pruned
On the identical expert selection, we measured:
| mechanism | Qwen3.5-35B | Qwen3-30B | tools |
|---|---|---|---|
| surplus experts deleted | 79.0% | 40.0% | 7/8 |
| surplus experts at 2 bits | 82.0% | 93.3% | 8/8 |
On the second architecture deletion loses 53 points and grading loses none — 93.3% equals the uncompressed base. Measured on two models, same selection.
Deletion costs ~1 point of reasoning per 1% of experts chosen wrongly and is unrecoverable. A 2-bit expert is present and merely imprecise. Compute cost is unchanged: the top-k budget is partitioned across the two banks, so exactly 8 experts run per token.
Files
| file | size | needs |
|---|---|---|
graded_language.gguf | 10.7 GB | dapack fork |
manifest-fragment.json | — | capability manifest for the catalogue |
Full documentation, tools to build your own domain packs, and every measurement behind this card: https://github.com/R3n3r0/dapack