A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 9, 02:30 UTC
Remove incorrect tokenizer info test (#33565)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Harry Mellor committed
Feb 2, 2026 at 17:11 UTC
6141ebe0dd5ed8d8034c6eb8e011b321bedb1697
1 file changed
-20
tests/entrypoints/openai/test_tokenization.py
-20
@@ -301,26 +301,6 @@ async def test_tokenizer_info_schema(server: RemoteOpenAIServer):
301
)
302
303
304
-@pytest.mark.asyncio
305
-async def test_tokenizer_info_added_tokens_structure(
306
- server: RemoteOpenAIServer,
307
-):
308
- """Test added_tokens_decoder structure if present."""
309
- response = requests.get(server.url_for("tokenizer_info"))
310
- response.raise_for_status()
311
- result = response.json()
312
- added_tokens = result.get("added_tokens_decoder")
313
- if added_tokens:
314
- for token_id, token_info in added_tokens.items():
315
- assert isinstance(token_id, str), "Token IDs should be strings"
316
- assert isinstance(token_info, dict), "Token info should be a dict"
317
- assert "content" in token_info, "Token info should have content"
318
- assert "special" in token_info, "Token info should have special flag"
319
- assert isinstance(token_info["special"], bool), (
320
- "Special flag should be boolean"
321
- )
322
-
323
-
304
@pytest.mark.asyncio
305
async def test_tokenizer_info_consistency_with_tokenize(
306
server: RemoteOpenAIServer,