chore: scaffold buddhagpt, verify Qwen2.5-7B-4bit runs on M5 (PAI-87)

.gitignore

@@ -0,0 +1,7 @@
+data/
+corpus/raw/
+models/
+.venv/
+__pycache__/
+*.jsonl
+adapters/

.python-version

@@ -0,0 +1 @@
+3.12

pyproject.toml

@@ -0,0 +1,28 @@
+[project]
+name = "buddhagpt"
+version = "0.1.0"
+description = "Add your description here"
+readme = "README.md"
+authors = [
+    { name = "marcuspaico", email = "marcus@thepai.co" }
+]
+requires-python = ">=3.12"
+dependencies = [
+    "anthropic>=0.122.0",
+    "lancedb>=0.37.1",
+    "mlx-lm>=0.31.3",
+    "pyyaml>=6.0.3",
+    "sentence-transformers>=5.7.0",
+]
+
+[project.scripts]
+buddhagpt = "buddhagpt:main"
+
+[build-system]
+requires = ["uv_build>=0.10.11,<0.11.0"]
+build-backend = "uv_build"
+
+[dependency-groups]
+dev = [
+    "pytest>=9.1.1",
+]

scripts/smoke_generate.py

@@ -0,0 +1,8 @@
+from mlx_lm import load, generate
+
+model, tokenizer = load("mlx-community/Qwen2.5-7B-Instruct-4bit")
+prompt = tokenizer.apply_chat_template(
+    [{"role": "user", "content": "What is the Second Noble Truth?"}],
+    tokenize=False, add_generation_prompt=True,
+)
+print(generate(model, tokenizer, prompt=prompt, max_tokens=200))

src/buddhagpt/__init__.py

@@ -0,0 +1,7 @@
+"""BuddhaGPT: Fine-tune and RAG experiment with Buddhist texts."""
+
+__version__ = "0.1.0"
+
+
+def main() -> None:
+    print("Hello from buddhagpt!")