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

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))