Giorgio Gilestro
809e45a5e0
First step from toy models toward real language models, on one 16 GB GPU. New src/llm/ package: procedural task families + exact-match verifier (tasks.py), batched eval (evaluate.py), LoRA specialisation (specialise.py, manual answer-only SFT), weight-space merge via peft add_weighted_adapter (merge.py: soup = averaged deltas, ties = sign-reconciled union), runner (experiment.py, kind llm_merge). Base Qwen2.5-0.5B-Instruct (Apache-2.0); three disjoint hard families (lists/strings/arith); one LoRA specialist each (~90s total). Result (seed 1), reported honestly: - STRONG/robust: the merges are the ONLY models competent across ALL families -- worst-family ~0.25 vs <0.16 for every single specialist (the Fisher-Muller "generalist assembled from specialists" signature, in real LoRA weights). - MARGINAL: "exceeds every parent overall" is only marginal at this scale (soup 0.64 vs best specialist 0.63; ties 0.61 below it). - CAVEAT VISIBLE: averaging dilutes peaks (lists specialist 0.43 -> merge 0.26) -- Layer-1's "merge, don't average" (E4) appearing in real weights. The pipeline works end-to-end; the balance/retention half reproduces; the strict overall-exceeds and soup-vs-ties distinction need scale (bigger base, more/cleaner families, seeds, a dilution-resistant / offspring-selected merge) -- the HPC step. Env: Python 3.14 + transformers 5.13 works; note transformers-5.x apply_chat_template returns a dict. make env-llm / make llm; adapters under gitignored models/llm/, base in the HF cache. figures/plot_llm_merge.py, README, tests/test_llm.py (+3, 125 green). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
41 lines
1.5 KiB
TOML
41 lines
1.5 KiB
TOML
[project]
|
|
name = "lamarckian-society"
|
|
version = "0.1.0"
|
|
description = "Layer 1 analytical core: knowledge transmission as a Wright-Fisher process."
|
|
readme = "README.md"
|
|
requires-python = ">=3.11"
|
|
dependencies = [
|
|
"numpy>=1.26",
|
|
"scipy>=1.11",
|
|
"pandas>=2.1",
|
|
"pyarrow>=15",
|
|
"matplotlib>=3.8",
|
|
"pydantic>=2.5",
|
|
"pyyaml>=6.0",
|
|
]
|
|
|
|
[project.optional-dependencies]
|
|
dev = ["pytest>=8.0"]
|
|
# Layer 1.5 neural existence proof. Torch is only needed from Stage C (RNN/VAE/MLP);
|
|
# Stages A-B (synthetic sandbox + histogram bridge) are pure NumPy and run in the base env.
|
|
# The default PyPI torch wheel is CUDA-enabled (cu13, matching the RTX A4000 driver).
|
|
# Install with `uv sync --extra neural`.
|
|
neural = ["torch>=2.2"]
|
|
# The real-MNIST secondary-confirmation tier only. Install with `uv sync --extra mnist`.
|
|
mnist = ["torchvision>=0.17"]
|
|
# Layer 2 / LLM prototype (blueprint C2/C4): LoRA specialists + weight-space merging on a small
|
|
# open-weight base. GPU; models download to the HF cache (outside the repo). `uv sync --extra llm`.
|
|
llm = ["transformers>=4.44", "peft>=0.11", "datasets", "accelerate"]
|
|
|
|
[build-system]
|
|
requires = ["hatchling"]
|
|
build-backend = "hatchling.build"
|
|
|
|
# src-layout: src/knowledge/ is importable as `knowledge` (the normative package
|
|
# name the scientific-validation conformance tests import). src/neural/ is Layer 1.5.
|
|
[tool.hatch.build.targets.wheel]
|
|
packages = ["src/knowledge", "src/neural", "src/llm"]
|
|
|
|
[tool.pytest.ini_options]
|
|
testpaths = ["tests"]
|
|
addopts = "-q"
|