chore

Author: t81dev

ARCHITECTURE.md

@@ -67,7 +67,7 @@ in `docs/design/limb-design.md`, `docs/design/bigint-design.md`, and `docs/desig
 ## Visual guides
 
 - `docs/diagrams/core-architecture.mermaid.md` — layered helpers from `limb` → `bigint` → the umbrella helpers and GEMM stack.
-- [`docs/diagrams/docs-sitemap.mermaid.md`](diagrams/docs-sitemap.mermaid.md) — site map summarizing the docs portal and related resources.
+- [`docs/diagrams/docs-sitemap.mermaid.md`](docs/diagrams/docs-sitemap.mermaid.md) — site map summarizing the docs portal and related resources.
 
 This guide is intentionally light and developer-facing—if you need a runnable overview, `docs/index.md`
 acts as the higher-level docs portal introduced in the README.

README.md

@@ -210,7 +210,7 @@ See [docs/api-overview.md](docs/api-overview.md) for the full surface described
 
 ### Specs & design
 
-- [docs/t81lib-spec-v1.0.0.md](docs/t81lib-spec-v1.0.0.md) — Normative contract.
+- [docs/t81lib-spec-v0.1.0.md](docs/t81lib-spec-v0.1.0.md) — Normative contract.
 - [docs/design/](docs/design/) — Deep dives on `limb`, `bigint`, and `montgomery`.
 - [docs/api-overview.md](docs/api-overview.md) — Umbrella helper catalog.
 - [ARCHITECTURE.md](ARCHITECTURE.md) & [docs/index.md](docs/index.md) — Architecture stories and portal.

docs/hardware.md

@@ -10,4 +10,4 @@ Highlights include:
 
 See `examples/ternary_hardware_sim_demo.ipynb` for a guided walkthrough that builds a ternary adder, runs a small inference, records virtual power/latency metrics, and highlights how balanced ternary drops switching energy for drones or tiny neuromorphic chips.
 
-The [hardware kernel sketch](docs/diagrams/hardware-kernels.mermaid.md) summarizes the tryte packing → multiply → accumulate flow used by the AVX/NEON GEMM paths.
+The [hardware kernel sketch](diagrams/hardware-kernels.mermaid.md) summarizes the tryte packing → multiply → accumulate flow used by the AVX/NEON GEMM paths.

docs/index.md

@@ -38,7 +38,7 @@ From here: `t8.nn.Linear` for drop-in layers, `t8.convert`/`t8.gguf` for scripte
 - **Landing & Quick Start** — [`README.md`](../README.md) contains the hero content, badges, and a comprehensive quick
   start section for builds, subprojects, pip/pipx consumers, CLI helpers, and the new “Start here” router.
 - **Architecture guide** — [`ARCHITECTURE.md`](../ARCHITECTURE.md) walks through the component layers and data flow.
-- **Normative spec** — [`docs/t81lib-spec-v1.0.0.md`](t81lib-spec-v1.0.0.md) defines the API guarantees for `limb`, `bigint`, and helpers.
+- **Normative spec** — [`docs/t81lib-spec-v0.1.0.md`](t81lib-spec-v0.1.0.md) defines the API guarantees for `limb`, `bigint`, and helpers.
 - **Design notes** — [`docs/design/limb-design.md`](design/limb-design.md), [`docs/design/bigint-design.md`](design/bigint-design.md), [`docs/design/montgomery-design.md`](design/montgomery-design.md) explain internal
   choices, algorithms, and invariants.
 - **Examples** — [`examples/`](../examples/) hosts runnable demos that mirror the README snippets.

docs/torch.md

@@ -19,4 +19,4 @@ This module keeps biases in FP32/BF16 while quantizing weights lazily, so you ca
 - `scripts/convert_to_ternary.py` walks a checkpoint tree, swaps every `torch.nn.Linear` for `t81.nn.Linear`, reports size reductions, and can force CPU device maps for reliable saves.
 - `t81-convert`, `t81-gguf`, and `examples/cli-examples.md` provide one-click scripts that mirror the Python flows while also exporting GGUF bundles for llama.cpp and Hugging Face runtimes.
 
-See [examples/README.md](examples/README.md) and [docs/use-cases.md](docs/use-cases.md) for runnable demos, scaling-law studies, and QAT comparisons.
+See [examples/README.md](../examples/README.md) and [use-cases.md](use-cases.md) for runnable demos, scaling-law studies, and QAT comparisons.

docs/troubleshooting.md

@@ -64,5 +64,5 @@ This guide complements [README.md](../README.md) and the other reference pages b
 
 ## Where to go next
 
-- Review the [docs/index.md](docs/index.md) portal for deep dives on Python, Torch, CLI, and hardware flows.
+- Review the [index.md](index.md) portal for deep dives on Python, Torch, CLI, and hardware flows.
 - If you hit a reproducible failure, open an issue referencing the log from `build/test-output` and the exact `cmake` command you used.

docs/use-cases.md

@@ -5,7 +5,7 @@
 ### Extreme compression for on-device inference
 
 Ternary PTQ/QAT on Phi-3-mini-4k-instruct: measure size (FP16 -> ternary), latency, and perplexity in a single workflow.
-Start here: [`examples/ternary_phi3_ptq_qat_demo.ipynb`](examples/ternary_phi3_ptq_qat_demo.ipynb).
+Start here: [`examples/ternary_phi3_ptq_qat_demo.ipynb`](../examples/ternary_phi3_ptq_qat_demo.ipynb).
 
 1. Convert existing large language models to ternary weights while keeping `torch.nn.Module` semantics (`examples/demo_llama_conversion.py`).
 2. Study ternary scaling laws for RMSNorm, RoPE, and ternary softmax to compare precision versus throughput (`examples/scaling_laws_ternary.py`).
@@ -25,4 +25,4 @@ These scripts and notebooks mirror the CLI workflows while keeping you inside Py
 
 ## Additional references
 
-Mentioned demos also appear in `docs/index.md` and `docs/references/cli-usage.md` so you can toggle between CLI helpers and Python stories without guessing. The [quantization workflow diagram](docs/diagrams/quantization-workflow.mermaid.md) ties the PyTorch conversion path to the CLI export/inference steps above.
+Mentioned demos also appear in `docs/index.md` and `docs/references/cli-usage.md` so you can toggle between CLI helpers and Python stories without guessing. The [quantization workflow diagram](diagrams/quantization-workflow.mermaid.md) ties the PyTorch conversion path to the CLI export/inference steps above.

examples/README.md

@@ -18,4 +18,4 @@ This file is the canonical list of runnable scripts and notebooks maintained und
 - `examples/ternary_hardware_sim_demo.ipynb` — Build a ternary adder, trace virtual power/latency, and compare energy vs binary hardware using `t81.hardware.TernaryEmulator`.
 - `examples/cli-examples.md` — Copy/paste-ready snippets for `t81-convert`, `t81-gguf`, and `t81-qat` workflows.
 
-Refer to [docs/use-cases.md](docs/use-cases.md) for details on how these examples tie into broader quantization, scaling-law, and hardware experiments, and consult the [quantization workflow diagram](../docs/diagrams/quantization-workflow.mermaid.md) for the PyTorch → CLI → inference path.
+Refer to [docs/use-cases.md](../docs/use-cases.md) for details on how these examples tie into broader quantization, scaling-law, and hardware experiments, and consult the [quantization workflow diagram](../docs/diagrams/quantization-workflow.mermaid.md) for the PyTorch → CLI → inference path.