[Feature] internlm2_5-20b-chat支持吗？我部署后能跑起来，但输出一团糟

zhangjiekui commented 1 month ago

Motivation

internlm2_5-20b-chat支持

Related resources

No response

Additional context

No response

lvhan028 commented 1 month ago

支持。烦请提供下 lmdeploy check_env的结果以及复现demo

zhangjiekui commented 1 month ago

lmdeploy check_env

` sys.platform: linux Python: 3.10.12 (main, Jul 29 2024, 16:56:48) [GCC 11.4.0] CUDA available: True MUSA available: False numpy_random_seed: 2147483648 GPU 0,1,2,3: NVIDIA GeForce RTX 4090 CUDA_HOME: /usr NVCC: Cuda compilation tools, release 11.5, V11.5.119 GCC: x86_64-linux-gnu-gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 PyTorch: 2.4.0+cu121 PyTorch compiling details: PyTorch built with:

GCC 9.3
C++ Version: 201703
Intel(R) oneAPI Math Kernel Library Version 2022.2-Product Build 20220804 for Intel(R) 64 architecture applications
Intel(R) MKL-DNN v3.4.2 (Git Hash 1137e04ec0b5251ca2b4400a4fd3c667ce843d67)
OpenMP 201511 (a.k.a. OpenMP 4.5)
LAPACK is enabled (usually provided by MKL)
NNPACK is enabled
CPU capability usage: AVX512
CUDA Runtime 12.1
NVCC architecture flags: -gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86;-gencode;arch=compute_90,code=sm_90
CuDNN 90.1 (built against CUDA 12.4)
Magma 2.6.1
Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=12.1, CUDNN_VERSION=9.1.0, CXX_COMPILER=/opt/rh/devtoolset-9/root/usr/bin/c++, CXX_FLAGS= -D_GLIBCXX_USE_CXX11_ABI=0 -fabi-version=11 -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DUSE_FBGEMM -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-stringop-overflow -Wsuggest-override -Wno-psabi -Wno-error=pedantic -Wno-error=old-style-cast -Wno-missing-braces -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=2.4.0, USE_CUDA=ON, USE_CUDNN=ON, USE_CUSPARSELT=1, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_GLOO=ON, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=1, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, USE_ROCM_KERNEL_ASSERT=OFF,

TorchVision: 0.19.0+cu121 LMDeploy: 0.5.3+ transformers: 4.43.3 gradio: 4.26.0 fastapi: 0.110.3 pydantic: 2.6.4 triton: 3.0.0 NVIDIA Topology: GPU0 GPU1 GPU2 GPU3 CPU Affinity NUMA Affinity GPU NUMA ID GPU0 X NODE SYS SYS 0-11,24-35 0 N/A GPU1 NODE X SYS SYS 0-11,24-35 0 N/A GPU2 SYS SYS X NODE 12-23,36-47 1 N/A GPU3 SYS SYS NODE X 12-23,36-47 1 N/A

Legend:

X = Self SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) PIX = Connection traversing at most a single PCIe bridge NV# = Connection traversing a bonded set of # NVLinks `

zhangjiekui commented 1 month ago

因为我是集成在langchain ChatOpenAI中使用的： `

ChatOpenAI( model_name=self.llm_model_name, openai_api_base=self.llm_model_base_url,
openai_api_key= self.api_key, streaming=streaming, verbose=True, temperature=temperature, max_tokens=max_tokens
) `

zhangjiekui commented 1 month ago

因此我只能提供lmdeploy的INFO日志： `

INFO: 10.1.150.105:36416 - "POST /v1/chat/completions HTTP/1.1" 200 OK 2024-08-26 00:44:46,880 - lmdeploy - INFO - prompt='<|im_start|>system\nSolve/Answer user\'s questions as best as you can in the language the user asked, preferred/default to Simplified Chinese. You have access to the following tools:get_flight_number,WEB_SEARCH,CODE_GENERATION,CODE_EXECUTION,VECTORSTORE.You are allowed to make multiple calls (either together or in sequence).Choice Guidelines: \'CODE_GENERATION\' is preferred rather than \'python_repl_tool\' for drawing images, plotting graphs/charts, data file processing/analyzing/saving.<|im_end|>\n<|im_start|>system name=<|plugin|>\n[{"description": "根据输入的出发地(origin),目的地(destination),日期(date),返回航班号。重要:当前日期是2024-08-26，日期(date)只接受python datetime.date类型输入。如果日期是不确定的描述，如明天、后天等，请先使用工具\'python_repl_tool\'进行计算，并且返回python datetime.date类型结果，具体格式示例如下：2024-04-03。\nArgs:\n origin:(str) 出发地/城市，例如合肥\n destination:(str) 目的地/城市，例如北京\n date:(date) 飞行日期，例如2024-08-03.如果日期是不确定的描述，如明天、后天等，请先调用工具\'python_repl_tool\'或者\'CODE_GENERATION\'进行计算，并且返回python datetime.date类型结果", "name": "get_flight_number", "parameters": {"type": "object", "properties": {"origin": {"type": "string"}, "destination": {"type": "string"}, "date": {"type": "string", "format": "date"}}, "required": ["origin", "destination", "date"]}}, {"description": "The internet. Use web_search for questions that are related to anything else than agent memory , prompt engineering, and adversarial attacks, 特别是2024年及以后相关的问题及发生的事情.", "name": "WEB_SEARCH", "parameters": {"type": "object", "properties": {"query": {"description": "The query to use when searching the internet.", "type": "string"}}, "required": ["query"]}}, {"description": "Code Assistant/Code generation . Use programming code (preferred Python language) to slove a user question, if the question related to [code generating, file processing/analyzing, data exploring/analyzing/visualizing].", "name": "CODE_GENERATION", "parameters": {"type": "object", "properties": {"query": {"description": "The query to use when need to generate python code.", "type": "string"}}, "required": ["query"]}}, {"description": "Code execution/Code intepreter. When user providing a code snippet and asking for running/executing, w or w/o returning the result.", "name": "CODE_EXECUTION", "parameters": {"type": "object", "properties": {"query": {"description": "The query to use when asking for running/executing a code snippet.", "type": "string"}}, "required": ["query"]}}, {"description": "A vectorstore/knowledge base containing documents related to 晶奇公司、安徽晶奇网络科技股份有限公司、医疗卫生、医疗健康、医疗服务、医疗政策、medical treatment & health. 凡是与晶奇公司、安徽晶奇网络科技股份有限公司及医疗健康相关的问题，都必须使用vectorstore工具。MUST use the vectorstore for questions on these topics.", "name": "VECTORSTORE", "parameters": {"type": "object", "properties": {"query": {"description": "The query to use when searching the vectorstore.", "type": "string"}}, "required": ["query"]}}]<|im_end|>\n<|im_start|>user\n请查询明天从合肥到上海的航班,以及南昌到北京的航班<|im_end|>\n<|im_start|>assistant\n', gen_config=EngineGenerationConfig(n=1, max_new_tokens=30000, top_p=1.0, top_k=40, temperature=0.01, repetition_penalty=1.0, ignore_eos=False, random_seed=17805185034465029694, stop_words=[92542, 92540], bad_words=None, min_new_tokens=None, skip_special_tokens=False, logprobs=None), prompt_token_id=[1, 92543, 9081, 364, 287, 4105, 301, 16272, 1341, 725, 4917, 569, 1999, 569, 629, 777, 435, 410, 4287, 410, 1341, 4751, 328, 15019, 29305, 442, 59870, 2019, 8588, 281, 1592, 746, 2775, 442, 410, 2863, 7521, 25331, 893, 4303, 5657, 328, 18985, 33780, 11442, 3043, 18839, 3656, 11442, 3043, 38076, 28951, 328, 43285, 43530, 38270, 657, 5578, 442, 1426, 5408, 6888, 451, 49152, 3942, 607, 435, 8636, 699, 25148, 46821, 334, 495, 14995, 18839, 3656, 259, 505, 15019, 4913, 1233, 495, 12819, 1433, 631, 23094, 259, 500, 13470, 5494, 328, 43487, 38804, 21513, 7188, 328, 955, 1177, 8826, 301, 43300, 20555, 2849, 2468, 281, 92542, 364, 92543, 9081, 963, 310, 92538, 364, 336, 5071, 4847, 921, 461, 68420, 68412, 60354, 71321, 60415, 44880, 833, 76581, 47433, 833, 69271, 12121, 833, 70121, 77201, 60675, 60355, 68372, 334, 69672, 69271, 60357, 638, 1311, 285, 2418, 285, 1743, 60353, 69271, 12121, 313, 60539, 68761, 12819, 9008, 10042, 68517, 68412, 60355, 68263, 69271, 69153, 90743, 69401, 60353, 60407, 70944, 60359, 82662, 60455, 60353, 60836, 60609, 68271, 68270, 259, 12819, 1433, 631, 23094, 259, 68274, 68679, 60353, 68614, 70121, 12819, 9008, 10042, 68517, 68501, 60353, 68592, 68878, 76885, 68411, 60387, 638, 1311, 285, 2469, 285, 2934, 60355, 1849, 4275, 334, 1849, 388, 8745, 3432, 626, 313, 262, 71321, 60415, 301, 68494, 60353, 69192, 73831, 1849, 388, 18139, 3432, 626, 313, 262, 76581, 301, 68494, 60353, 69192, 68516, 1849, 388, 1170, 3432, 1170, 313, 262, 71709, 69271, 60353, 262, 69192, 638, 1311, 285, 2418, 285, 2934, 281, 68263, 69271, 69153, 90743, 69401, 60353, 60407, 70944, 60359, 82662, 60455, 60353, 60836, 60609, 76395, 68270, 259, 12819, 1433, 631, 23094, 259, 68319, 259, 14995, 18839, 3656, 259, 68274, 68679, 60353, 68614, 70121, 12819, 9008, 10042, 68517, 68501, 628, 461, 738, 921, 461, 586, 893, 4303, 5657, 628, 461, 13927, 921, 5371, 1459, 921, 461, 1850, 628, 461, 13333, 921, 5371, 8745, 921, 5371, 1459, 921, 461, 1054, 14484, 461, 18139, 921, 5371, 1459, 921, 461, 1054, 14484, 461, 1170, 921, 5371, 1459, 921, 461, 1054, 628, 461, 2394, 921, 461, 1170, 9338, 2291, 461, 6436, 921, 4544, 8745, 628, 461, 18139, 628, 461, 1170, 1487, 38002, 5371, 4847, 921, 461, 918, 7742, 281, 5602, 3644, 10863, 500, 4917, 560, 657, 5594, 442, 4271, 902, 1233, 8447, 5097, 1298, 10069, 14800, 328, 454, 28750, 42369, 8914, 328, 262, 69764, 638, 1311, 60372, 60543, 68458, 68524, 68804, 60543, 68526, 68899, 10603, 461, 738, 921, 461, 18985, 33780, 628, 461, 13927, 921, 5371, 1459, 921, 461, 1850, 628, 461, 13333, 921, 5371, 1779, 921, 5371, 4847, 921, 461, 918, 3402, 442, 1130, 1119, 15164, 410, 7742, 10603, 461, 1459, 921, 461, 1054, 9338, 2291, 461, 6436, 921, 4544, 1779, 1487, 38002, 5371, 4847, 921, 461, 2229, 21624, 301, 2229, 9600, 790, 5602, 15600, 2189, 451, 1877, 5710, 13175, 4287, 313, 442, 17742, 717, 395, 1341, 3568, 328, 552, 410, 3568, 5594, 442, 640, 2000, 23492, 328, 1177, 8826, 301, 43300, 20555, 328, 955, 24341, 301, 43300, 20555, 301, 29526, 5008, 1074, 628, 461, 738, 921, 461, 14995, 18839, 3656, 628, 461, 13927, 921, 5371, 1459, 921, 461, 1850, 628, 461, 13333, 921, 5371, 1779, 921, 5371, 4847, 921, 461, 918, 3402, 442, 1130, 1119, 1329, 442, 7076, 10273, 2189, 10603, 461, 1459, 921, 461, 1054, 9338, 2291, 461, 6436, 921, 4544, 1779, 1487, 38002, 5371, 4847, 921, 461, 2229, 11471, 301, 2229, 658, 880, 396, 596, 281, 3363, 1341, 8373, 395, 2189, 42699, 454, 10300, 500, 4463, 301, 11897, 10748, 328, 420, 607, 420, 20449, 13593, 410, 1245, 10603, 461, 738, 921, 461, 14995, 38076, 28951, 628, 461, 13927, 921, 5371, 1459, 921, 461, 1850, 628, 461, 13333, 921, 5371, 1779, 921, 5371, 4847, 921, 461, 918, 3402, 442, 1130, 1119, 10300, 500, 4463, 301, 11897, 10748, 395, 2189, 42699, 10603, 461, 1459, 921, 461, 1054, 9338, 2291, 461, 6436, 921, 4544, 1779, 1487, 38002, 5371, 4847, 921, 461, 290, 4782, 4474, 14249, 50356, 2483, 8617, 9423, 5594, 442, 262, 61954, 61112, 68280, 60359, 70239, 61954, 61112, 68441, 68636, 73067, 60359, 69432, 68890, 60359, 69432, 68515, 60359, 69432, 68320, 60359, 69432, 68843, 60359, 2213, 1076, 6535, 741, 2984, 281, 262, 83860, 60510, 61954, 61112, 68280, 60359, 70239, 61954, 61112, 68441, 68636, 73067, 60543, 69432, 68515, 68524, 68804, 60353, 92146, 68271, 3380, 4474, 68270, 60355, 307, 8685, 1130, 410, 4782, 4474, 500, 4917, 519, 1639, 13485, 10603, 461, 738, 921, 461, 43285, 43530, 628, 461, 13927, 921, 5371, 1459, 921, 461, 1850, 628, 461, 13333, 921, 5371, 1779, 921, 5371, 4847, 921, 461, 918, 3402, 442, 1130, 1119, 15164, 410, 4782, 4474, 10603, 461, 1459, 921, 461, 1054, 9338, 2291, 461, 6436, 921, 4544, 1779, 1487, 3578, 332, 92542, 364, 92543, 1008, 364, 60836, 69358, 70944, 60577, 73831, 60378, 68589, 60354, 77201, 328, 68375, 75871, 60378, 88804, 77201, 92542, 364, 92543, 525, 11353, 364], adapter_name=None. 2024-08-26 00:44:46,880 - lmdeploy - INFO - session_id=14, history_tokens=0, input_tokens=803, max_new_tokens=30000, seq_start=True, seq_end=True, step=0, prep=True 2024-08-26 00:44:46,880 - lmdeploy - INFO - Register stream callback for 14 [TM][INFO] [forward] Enqueue requests [TM][INFO] [forward] Wait for requests to complete ... [TM][INFO] [ProcessInferRequests] Request for 14 received. [TM][INFO] [Forward] [0, 1), dc=0, pf=1, sum_q=803, sum_k=803, max_q=803, max_k=803 [TM][INFO] ------------------------- step = 810 ------------------------- [TM][INFO] ------------------------- step = 820 ------------------------- [TM][INFO] ------------------------- step = 830 ------------------------- [TM][INFO] ------------------------- step = 840 ------------------------- [TM][INFO] ------------------------- step = 850 ------------------------- [TM][INFO] ------------------------- step = 860 ------------------------- [TM][INFO] ------------------------- step = 870 ------------------------- [TM][INFO] ------------------------- step = 880 ------------------------- [TM][INFO] ------------------------- step = 890 ------------------------- [TM][INFO] ------------------------- step = 900 ------------------------- [TM][INFO] ------------------------- step = 910 ------------------------- [TM][INFO] ------------------------- step = 920 ------------------------- [TM][INFO] [Interrupt] slot = 0, id = 14 [TM][INFO] [forward] Request completed for 14 2024-08-26 00:44:49,126 - lmdeploy - INFO - UN-register stream callback for 14 `

zhangjiekui commented 1 month ago

但INFO日志中没办法看到模型产生的输出结果（不清楚是模型本身没输出，还只是INFO日志没有打印出输出）

zhangjiekui commented 1 month ago

然后我通过langsmith来跟踪：

ValueError('No generations found in stream.')Traceback (most recent call last):

File "/root/.virtualenvs/test/lib/python3.10/site-packages/langchain_core/language_models/chat_models.py", line 603, in generate self._generate_with_cache(

File "/root/.virtualenvs/test/lib/python3.10/site-packages/langchain_core/language_models/chat_models.py", line 825, in _generate_with_cache result = self._generate(

File "/root/.virtualenvs/test/lib/python3.10/site-packages/langchain_openai/chat_models/base.py", line 534, in _generate return generate_from_stream(stream_iter)

File "/root/.virtualenvs/test/lib/python3.10/site-packages/langchain_core/language_models/chat_models.py", line 107, in generate_from_stream raise ValueError("No generations found in stream.")

ValueError: No generations found in stream.

zhangjiekui commented 1 month ago

发现可能是stream的问题，然后设置streaming=False，可以跑出结果来，但又发现错误：

lmdeploy/model.py", line 515, in messages2prompt begin = box_map[role] KeyError: 'tool'

推测应该是我的应用里有tool_call , 并返回了langchain的ToolMessage，里面的role = 'tool'

zhangjiekui commented 1 month ago

所以,我们与openai的一致性兼容到底是哪些一致，还有哪些不一致呢？

zhangjiekui commented 1 month ago

通过修改：lmdeploy/model.py，增加tool = "tool"

`

    box_map = dict(user=self.user,
                   assistant=self.assistant,
                   system=self.system,
                   environment=self.environment,
                   tool = "tool")
    eox_map = dict(user=self.eoh,
                   assistant=self.eoa + self.separator,
                   system=self.eosys,
                   environment=self.eoenv,
                   tool = "tool")

`

终于可以跑起来了

zhangjiekui commented 1 month ago

再多运行几个测试，又发现很多问题，没办法继续了......

AllentDan commented 1 month ago

目前的 tool 功能只支持了 stream=False.

William4711 commented 1 month ago

我也發現了一樣的錯誤KeyError: 'tool' 測試過程

@tool
def GetWeather(location: str):
    """Get the current weather in a given location"""
    return "30 degrees Celsius"

query = "What's the weather like in San Francisco today?"
messages = [HumanMessage(query)]
tools = [GetWeather]
model_with_tools = llm.bind_tools(tools)
ans1 = model_with_tools.invoke(messages)

for tool_call in ans1.tool_calls:
    selected_tool = {"GetWeather":GetWeather}[tool_call["name"]]
    tool_output = selected_tool.invoke(tool_call["args"])
    messages.append(ToolMessage(tool_output, tool_call_id=tool_call["id"]))

ans1 = model_with_tools.invoke(messages)

在最後一行出錯看似無法用ToolMessage

如果無法ToolMessage 要怎麼結合回傳的溫度給llm做回覆呢?

2024-08-30 09:50:21 INFO: 172.18.0.1:36578 - "POST /v1/chat/completions HTTP/1.1" 200 OK 2024-08-30 09:50:32 INFO: 172.18.0.1:47358 - "POST /v1/chat/completions HTTP/1.1" 500 Internal Server Error 2024-08-30 09:50:32 ERROR: Exception in ASGI application 2024-08-30 09:50:32 Traceback (most recent call last): 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/uvicorn/protocols/http/h11_impl.py", line 406, in run_asgi 2024-08-30 09:50:32 result = await app( # type: ignore[func-returns-value] 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/uvicorn/middleware/proxy_headers.py", line 70, in call 2024-08-30 09:50:32 return await self.app(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/fastapi/applications.py", line 1054, in call 2024-08-30 09:50:32 await super().call(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/applications.py", line 123, in call 2024-08-30 09:50:32 await self.middleware_stack(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/errors.py", line 186, in call 2024-08-30 09:50:32 raise exc 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/errors.py", line 164, in call 2024-08-30 09:50:32 await self.app(scope, receive, _send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/cors.py", line 85, in call 2024-08-30 09:50:32 await self.app(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/exceptions.py", line 65, in call 2024-08-30 09:50:32 await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 64, in wrapped_app 2024-08-30 09:50:32 raise exc 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app 2024-08-30 09:50:32 await app(scope, receive, sender) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 754, in call 2024-08-30 09:50:32 await self.middleware_stack(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 774, in app 2024-08-30 09:50:32 await route.handle(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 295, in handle 2024-08-30 09:50:32 await self.app(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 77, in app 2024-08-30 09:50:32 await wrap_app_handling_exceptions(app, request)(scope, receive, send) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 64, in wrapped_app 2024-08-30 09:50:32 raise exc 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app 2024-08-30 09:50:32 await app(scope, receive, sender) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 74, in app 2024-08-30 09:50:32 response = await f(request) 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/fastapi/routing.py", line 297, in app 2024-08-30 09:50:32 raw_response = await run_endpoint_function( 2024-08-30 09:50:32 File "/opt/py3/lib/python3.10/site-packages/fastapi/routing.py", line 210, in run_endpoint_function 2024-08-30 09:50:32 return await dependant.call(values) 2024-08-30 09:50:32 File "/opt/lmdeploy/lmdeploy/serve/openai/api_server.py", line 458, in chat_completions_v1 2024-08-30 09:50:32 async for res in result_generator: 2024-08-30 09:50:32 File "/opt/lmdeploy/lmdeploy/serve/async_engine.py", line 525, in generate 2024-08-30 09:50:32 prompt_input = await self._get_prompt_input(prompt, 2024-08-30 09:50:32 File "/opt/lmdeploy/lmdeploy/serve/async_engine.py", line 471, in _get_prompt_input 2024-08-30 09:50:32 prompt = chat_template.messages2prompt(prompt, 2024-08-30 09:50:32 File "/opt/lmdeploy/lmdeploy/model.py", line 866, in messages2prompt 2024-08-30 09:50:32 ret += f'{box_map[role]}{content}{eox_map[role]}' 2024-08-30 09:50:32 KeyError: 'tool' 2024-08-30 09:50:33 INFO: 172.18.0.1:47360 - "POST /v1/chat/completions HTTP/1.1" 500 Internal Server Error 2024-08-30 09:50:33 ERROR: Exception in ASGI application 2024-08-30 09:50:33 Traceback (most recent call last): 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/uvicorn/protocols/http/h11_impl.py", line 406, in run_asgi 2024-08-30 09:50:33 result = await app( # type: ignore[func-returns-value] 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/uvicorn/middleware/proxy_headers.py", line 70, in call 2024-08-30 09:50:33 return await self.app(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/fastapi/applications.py", line 1054, in call 2024-08-30 09:50:33 await super().call(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/applications.py", line 123, in call 2024-08-30 09:50:33 await self.middleware_stack(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/errors.py", line 186, in call 2024-08-30 09:50:33 raise exc 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/errors.py", line 164, in call 2024-08-30 09:50:33 await self.app(scope, receive, _send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/cors.py", line 85, in call 2024-08-30 09:50:33 await self.app(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/exceptions.py", line 65, in call 2024-08-30 09:50:33 await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 64, in wrapped_app 2024-08-30 09:50:33 raise exc 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app 2024-08-30 09:50:33 await app(scope, receive, sender) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 754, in call 2024-08-30 09:50:33 await self.middleware_stack(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 774, in app 2024-08-30 09:50:33 await route.handle(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 295, in handle 2024-08-30 09:50:33 await self.app(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 77, in app 2024-08-30 09:50:33 await wrap_app_handling_exceptions(app, request)(scope, receive, send) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 64, in wrapped_app 2024-08-30 09:50:33 raise exc 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app 2024-08-30 09:50:33 await app(scope, receive, sender) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 74, in app 2024-08-30 09:50:33 response = await f(request) 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/fastapi/routing.py", line 297, in app 2024-08-30 09:50:33 raw_response = await run_endpoint_function( 2024-08-30 09:50:33 File "/opt/py3/lib/python3.10/site-packages/fastapi/routing.py", line 210, in run_endpoint_function 2024-08-30 09:50:33 return await dependant.call(values) 2024-08-30 09:50:33 File "/opt/lmdeploy/lmdeploy/serve/openai/api_server.py", line 458, in chat_completions_v1 2024-08-30 09:50:33 async for res in result_generator: 2024-08-30 09:50:33 File "/opt/lmdeploy/lmdeploy/serve/async_engine.py", line 525, in generate 2024-08-30 09:50:33 prompt_input = await self._get_prompt_input(prompt, 2024-08-30 09:50:33 File "/opt/lmdeploy/lmdeploy/serve/async_engine.py", line 471, in _get_prompt_input 2024-08-30 09:50:33 prompt = chat_template.messages2prompt(prompt, 2024-08-30 09:50:33 File "/opt/lmdeploy/lmdeploy/model.py", line 866, in messages2prompt 2024-08-30 09:50:33 ret += f'{box_map[role]}{content}{eox_map[role]}' 2024-08-30 09:50:33 KeyError: 'tool' 2024-08-30 09:50:35 INFO: 172.18.0.1:47366 - "POST /v1/chat/completions HTTP/1.1" 500 Internal Server Error 2024-08-30 09:50:35 ERROR: Exception in ASGI application 2024-08-30 09:50:35 Traceback (most recent call last): 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/uvicorn/protocols/http/h11_impl.py", line 406, in run_asgi 2024-08-30 09:50:35 result = await app( # type: ignore[func-returns-value] 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/uvicorn/middleware/proxy_headers.py", line 70, in call 2024-08-30 09:50:35 return await self.app(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/fastapi/applications.py", line 1054, in call 2024-08-30 09:50:35 await super().call(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/applications.py", line 123, in call 2024-08-30 09:50:35 await self.middleware_stack(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/errors.py", line 186, in call 2024-08-30 09:50:35 raise exc 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/errors.py", line 164, in call 2024-08-30 09:50:35 await self.app(scope, receive, _send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/cors.py", line 85, in call 2024-08-30 09:50:35 await self.app(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/middleware/exceptions.py", line 65, in call 2024-08-30 09:50:35 await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 64, in wrapped_app 2024-08-30 09:50:35 raise exc 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app 2024-08-30 09:50:35 await app(scope, receive, sender) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 754, in call 2024-08-30 09:50:35 await self.middleware_stack(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 774, in app 2024-08-30 09:50:35 await route.handle(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 295, in handle 2024-08-30 09:50:35 await self.app(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 77, in app 2024-08-30 09:50:35 await wrap_app_handling_exceptions(app, request)(scope, receive, send) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 64, in wrapped_app 2024-08-30 09:50:35 raise exc 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app 2024-08-30 09:50:35 await app(scope, receive, sender) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/starlette/routing.py", line 74, in app 2024-08-30 09:50:35 response = await f(request) 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/fastapi/routing.py", line 297, in app 2024-08-30 09:50:35 raw_response = await run_endpoint_function( 2024-08-30 09:50:35 File "/opt/py3/lib/python3.10/site-packages/fastapi/routing.py", line 210, in run_endpoint_function 2024-08-30 09:50:35 return await dependant.call(**values) 2024-08-30 09:50:35 File "/opt/lmdeploy/lmdeploy/serve/openai/api_server.py", line 458, in chat_completions_v1 2024-08-30 09:50:35 async for res in result_generator: 2024-08-30 09:50:35 File "/opt/lmdeploy/lmdeploy/serve/async_engine.py", line 525, in generate 2024-08-30 09:50:35 prompt_input = await self._get_prompt_input(prompt, 2024-08-30 09:50:35 File "/opt/lmdeploy/lmdeploy/serve/async_engine.py", line 471, in _get_prompt_input 2024-08-30 09:50:35 prompt = chat_template.messages2prompt(prompt, 2024-08-30 09:50:35 File "/opt/lmdeploy/lmdeploy/model.py", line 866, in messages2prompt 2024-08-30 09:50:35 ret += f'{box_map[role]}{content}{eox_map[role]}' 2024-08-30 09:50:35 KeyError: 'tool'

AllentDan commented 1 month ago

@William4711 这个错误看起来是传入了一个 tool 的角色到 llama3.1 模型，llama3.1 没有这个角色。支持的有 user，assistant, ipython, system

William4711 commented 1 month ago

@William4711 这个错误看起来是传入了一个 tool 的角色到 llama3.1 模型，llama3.1 没有这个角色。支持的有 user，assistant, ipython, system

ToolMessage來自於Langchain Langchain有關程序呼叫相關的message有ToolMessage與FunctionMessage 用了 ToolMessage會報 KeyError: 'tool' FunctionMessage會報 KeyError: 'function'

所以llama3.1應該是角色ipython來搭配程序呼叫回覆使用我再來測試看看 thanks

InternLM / lmdeploy