llm is the only app (#15779)

* tinygrad/llm is the only app

* upd pyproject

* claude refs

* scoping

* min diff

test/external/external_llm_eval.py

@@ -1,4 +1,4 @@
-# eval for tinygrad.apps.llm -- hits the server via OpenAI API
+# eval for OpenAI API server
 # uses Meta's exact ARC-Challenge prompt template from lm-evaluation-harness llama3 tasks
 import argparse, re, pyarrow.parquet as pq
 from openai import OpenAI

test/external/external_test_simple_tokenizer.py

@@ -1,7 +1,7 @@
 import functools, multiprocessing
 from transformers import AutoTokenizer
 from datasets import load_dataset
-from tinygrad.apps.llm import SimpleTokenizer
+from tinygrad.llm.cli import SimpleTokenizer
 from tinygrad.helpers import tqdm, getenv, partition
 
 @functools.cache