Merge branch 'guanyu' into 'v0.6.2-dev'

test_audio文件将TEST_AUDIO_URLS改为本地；testchat文件360行路径改为本地；test_metrics文件197行将base_...

See merge request dcutoolkit/deeplearing/vllm!50

tests/entrypoints/openai/test_audio.py

@@ -11,11 +11,13 @@ from vllm.multimodal.utils import encode_audio_base64, fetch_audio
 from ...utils import RemoteOpenAIServer, models_path_prefix
 
 MODEL_NAME = os.path.join(models_path_prefix, "fixie-ai/ultravox-v0_3")
+# TEST_AUDIO_URLS = [
+#     AudioAsset("winning_call").url,
+# ]
 TEST_AUDIO_URLS = [
-    AudioAsset("winning_call").url,
+    "http://localhost:8000/winning_call.ogg"
 ]
 
-
 @pytest.fixture(scope="module")
 def server():
     args = [

tests/entrypoints/openai/test_chat.py

@@ -355,9 +355,13 @@ async def test_chat_streaming(client: openai.AsyncOpenAI, model_name: str):
 
 
 @pytest.mark.asyncio
+# @pytest.mark.parametrize(
+#     "model_name",
+#     ["HuggingFaceH4/zephyr-7b-beta", "zephyr-lora"],
+# )
 @pytest.mark.parametrize(
     "model_name",
-    ["HuggingFaceH4/zephyr-7b-beta", "zephyr-lora"],
+    [os.path.join(models_path_prefix,"HuggingFaceH4/zephyr-7b-beta"), "zephyr-lora"],
 )
 async def test_chat_completion_stream_options(client: openai.AsyncOpenAI,
                                               model_name: str):

tests/entrypoints/openai/test_metrics.py

@@ -194,7 +194,8 @@ async def test_metrics_exist(client: openai.AsyncOpenAI):
 def test_metrics_exist_run_batch():
     input_batch = """{"custom_id": "request-0", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/e5-mistral-7b-instruct", "input": "You are a helpful assistant."}}"""  # noqa: E501
 
-    base_url = "0.0.0.0"
+    #base_url = "0.0.0.0"
+    base_url = "localhost"
     port = "8001"
     server_url = f"http://{base_url}:{port}"
 

tests/entrypoints/openai/test_run_batch.py

@@ -7,20 +7,36 @@ from vllm.entrypoints.openai.protocol import BatchRequestOutput
 from ...utils import models_path_prefix
 
 # ruff: noqa: E501
-INPUT_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are a helpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
-{"custom_id": "request-2", "method": "POST", "url": "/v1/chat/completions", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+# INPUT_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are a helpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+# {"custom_id": "request-2", "method": "POST", "url": "/v1/chat/completions", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+
+# {"custom_id": "request-3", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NonExistModel", "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+# {"custom_id": "request-4", "method": "POST", "url": "/bad_url", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+# {"custom_id": "request-5", "method": "POST", "url": "/v1/chat/completions", "body": {"stream": "True", "model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}"""
+
+# INVALID_INPUT_BATCH = """{"invalid_field": "request-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are a helpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+# {"custom_id": "request-2", "method": "POST", "url": "/v1/chat/completions", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}"""
+
+# INPUT_EMBEDDING_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/embeddings", "body": {"model": os.path.join(models_path_prefix, "intfloat/e5-mistral-7b-instruct"), "input": "You are a helpful assistant."}}
+# {"custom_id": "request-2", "method": "POST", "url": "/v1/embeddings", "body": {"model": os.path.join(models_path_prefix, "intfloat/e5-mistral-7b-instruct"), "input": "You are an unhelpful assistant."}}
+
+# {"custom_id": "request-3", "method": "POST", "url": "/v1/embeddings", "body": {"model": os.path.join(models_path_prefix, "intfloat/e5-mistral-7b-instruct"), "input": "Hello world!"}}
+# {"custom_id": "request-4", "method": "POST", "url": "/v1/embeddings", "body": {"model": "NonExistModel", "input": "Hello world!"}}"""
+#修改↓
+INPUT_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are a helpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+{"custom_id": "request-2", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
 
 {"custom_id": "request-3", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NonExistModel", "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
-{"custom_id": "request-4", "method": "POST", "url": "/bad_url", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
-{"custom_id": "request-5", "method": "POST", "url": "/v1/chat/completions", "body": {"stream": "True", "model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}"""
+{"custom_id": "request-4", "method": "POST", "url": "/bad_url", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+{"custom_id": "request-5", "method": "POST", "url": "/v1/chat/completions", "body": {"stream": "True", "model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}"""
 
-INVALID_INPUT_BATCH = """{"invalid_field": "request-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are a helpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
-{"custom_id": "request-2", "method": "POST", "url": "/v1/chat/completions", "body": {"model": os.path.join(models_path_prefix, "NousResearch/Meta-Llama-3-8B-Instruct"), "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}"""
+INVALID_INPUT_BATCH = """{"invalid_field": "request-1", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are a helpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}
+{"custom_id": "request-2", "method": "POST", "url": "/v1/chat/completions", "body": {"model": "NousResearch/Meta-Llama-3-8B-Instruct", "messages": [{"role": "system", "content": "You are an unhelpful assistant."},{"role": "user", "content": "Hello world!"}],"max_tokens": 1000}}"""
 
-INPUT_EMBEDDING_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/embeddings", "body": {"model": os.path.join(models_path_prefix, "intfloat/e5-mistral-7b-instruct"), "input": "You are a helpful assistant."}}
-{"custom_id": "request-2", "method": "POST", "url": "/v1/embeddings", "body": {"model": os.path.join(models_path_prefix, "intfloat/e5-mistral-7b-instruct"), "input": "You are an unhelpful assistant."}}
+INPUT_EMBEDDING_BATCH = """{"custom_id": "request-1", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/e5-mistral-7b-instruct", "input": "You are a helpful assistant."}}
+{"custom_id": "request-2", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/e5-mistral-7b-instruct", "input": "You are an unhelpful assistant."}}
 
-{"custom_id": "request-3", "method": "POST", "url": "/v1/embeddings", "body": {"model": os.path.join(models_path_prefix, "intfloat/e5-mistral-7b-instruct"), "input": "Hello world!"}}
+{"custom_id": "request-3", "method": "POST", "url": "/v1/embeddings", "body": {"model": "intfloat/e5-mistral-7b-instruct", "input": "Hello world!"}}
 {"custom_id": "request-4", "method": "POST", "url": "/v1/embeddings", "body": {"model": "NonExistModel", "input": "Hello world!"}}"""
 
 

tests/lora/test_tokenizer_group.py

@@ -4,7 +4,8 @@ from transformers import AutoTokenizer, PreTrainedTokenizerBase
 from vllm.lora.request import LoRARequest
 from vllm.transformers_utils.tokenizer import get_lora_tokenizer
 from vllm.transformers_utils.tokenizer_group import get_tokenizer_group
-
+import os
+from ..utils import RemoteOpenAIServer, models_path_prefix
 from ..conftest import get_tokenizer_pool_config
 
 
@@ -14,7 +15,7 @@ async def test_tokenizer_group_lora(sql_lora_files, tokenizer_group_type):
     reference_tokenizer = AutoTokenizer.from_pretrained(sql_lora_files)
     tokenizer_group = get_tokenizer_group(
         get_tokenizer_pool_config(tokenizer_group_type),
-        tokenizer_id="gpt2",
+        tokenizer_id=os.path.join(models_path_prefix,"gpt2"),
         enable_lora=True,
         max_num_seqs=1,
         max_input_length=None,

tests/models/decoder_only/vision_language/test_broadcast.py

@@ -20,11 +20,22 @@ def test_models(hf_runner, vllm_runner, image_assets,
     num_logprobs = 5
     tensor_parallel_size = 2
 
-    if model.startswith("llava-hf/llava-1.5"):
+    # if model.startswith("llava-hf/llava-1.5"):
+    #     from .test_llava import models, run_test
+    # elif model.startswith("llava-hf/llava-v1.6"):
+    #     from .test_llava_next import models, run_test  # type: ignore[no-redef]
+    # elif model.startswith("facebook/chameleon"):
+    #     from .test_chameleon import models, run_test  # type: ignore[no-redef]
+    # else:
+    #     raise NotImplementedError(f"Unsupported model: {model}")
+    if "llava-hf/llava-1.5-7b-hf" in model:
+    # if model.startswith(os.path.join(models_path_prefix, "llava-hf/llava-1.5-7b-hf"),):
         from .test_llava import models, run_test
-    elif model.startswith("llava-hf/llava-v1.6"):
+    elif "llava-hf/llava-v1.6-mistral-7b-hf" in model:
+    #elif model.startswith(os.path.join(models_path_prefix, "llava-hf/llava-v1.6-mistral-7b-hf")):
         from .test_llava_next import models, run_test  # type: ignore[no-redef]
-    elif model.startswith("facebook/chameleon"):
+    elif "facebook/chameleon-7b" in model:
+    #elif model.startswith(os.path.join(models_path_prefix, "facebook/chameleon-7b")):
         from .test_chameleon import models, run_test  # type: ignore[no-redef]
     else:
         raise NotImplementedError(f"Unsupported model: {model}")