llama-stack remote vllm always have default tool_system prompt even if there is no tool #1953
Description
System Info
<frozen runpy>:128: RuntimeWarning: 'torch.utils.collect_env' found in sys.modules after import of package 'torch.utils', but prior to execution of 'torch.utils.collect_env'; this may result in unpredictable behaviour
Collecting environment information...
PyTorch version: 2.6.0+cu124
Is debug build: False
CUDA used to build PyTorch: 12.4
ROCM used to build PyTorch: N/A
OS: CentOS Stream 9 (x86_64)
GCC version: (GCC) 11.5.0 20240719 (Red Hat 11.5.0-5)
Clang version: Could not collect
CMake version: version 3.26.5
Libc version: glibc-2.34
Python version: 3.12.0 | packaged by Anaconda, Inc. | (main, Oct 2 2023, 17:29:18) [GCC 11.2.0] (64-bit runtime)
Python platform: Linux-6.4.3-0_fbk15_zion_2630_gf27365f948db-x86_64-with-glibc2.34
Is CUDA available: True
CUDA runtime version: Could not collect
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration:
GPU 0: NVIDIA H100
GPU 1: NVIDIA H100
GPU 2: NVIDIA H100
GPU 3: NVIDIA H100
GPU 4: NVIDIA H100
GPU 5: NVIDIA H100
GPU 6: NVIDIA H100
GPU 7: NVIDIA H100
Nvidia driver version: 535.154.05
cuDNN version: Probably one of the following:
/usr/lib64/libcudnn.so.8.9.2
/usr/lib64/libcudnn.so.9.7.1
/usr/lib64/libcudnn_adv.so.9.7.1
/usr/lib64/libcudnn_adv_infer.so.8.9.2
/usr/lib64/libcudnn_adv_train.so.8.9.2
/usr/lib64/libcudnn_cnn.so.9.7.1
/usr/lib64/libcudnn_cnn_infer.so.8.9.2
/usr/lib64/libcudnn_cnn_train.so.8.9.2
/usr/lib64/libcudnn_engines_precompiled.so.9.7.1
/usr/lib64/libcudnn_engines_runtime_compiled.so.9.7.1
/usr/lib64/libcudnn_graph.so.9.7.1
/usr/lib64/libcudnn_heuristic.so.9.7.1
/usr/lib64/libcudnn_ops.so.9.7.1
/usr/lib64/libcudnn_ops_infer.so.8.9.2
/usr/lib64/libcudnn_ops_train.so.8.9.2
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Address sizes: 52 bits physical, 57 bits virtual
Byte Order: Little Endian
CPU(s): 384
On-line CPU(s) list: 0-383
Vendor ID: AuthenticAMD
Model name: AMD EPYC 9654 96-Core Processor
CPU family: 25
Model: 17
Thread(s) per core: 2
Core(s) per socket: 96
Socket(s): 2
Stepping: 1
Frequency boost: enabled
CPU(s) scaling MHz: 85%
CPU max MHz: 3707.8120
CPU min MHz: 1500.0000
BogoMIPS: 4792.65
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good amd_lbr_v2 nopl nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 invpcid_single hw_pstate ssbd mba perfmon_v2 ibrs ibpb stibp ibrs_enhanced vmmcall fsgsbase bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local avx512_bf16 clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin cppc arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif x2avic v_spec_ctrl vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid overflow_recov succor smca fsrm flush_l1d
Virtualization: AMD-V
L1d cache: 6 MiB (192 instances)
L1i cache: 6 MiB (192 instances)
L2 cache: 192 MiB (192 instances)
L3 cache: 768 MiB (24 instances)
NUMA node(s): 2
NUMA node0 CPU(s): 0-95,192-287
NUMA node1 CPU(s): 96-191,288-383
Vulnerability Gather data sampling: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Vulnerable: eIBRS with unprivileged eBPF
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Not affected
Versions of relevant libraries:
[pip3] numpy==1.26.4
[pip3] nvidia-cublas-cu12==12.4.5.8
[pip3] nvidia-cuda-cupti-cu12==12.4.127
[pip3] nvidia-cuda-nvrtc-cu12==12.4.127
[pip3] nvidia-cuda-runtime-cu12==12.4.127
[pip3] nvidia-cudnn-cu12==9.1.0.70
[pip3] nvidia-cufft-cu12==11.2.1.3
[pip3] nvidia-curand-cu12==10.3.5.147
[pip3] nvidia-cusolver-cu12==11.6.1.9
[pip3] nvidia-cusparse-cu12==12.3.1.170
[pip3] nvidia-cusparselt-cu12==0.6.2
[pip3] nvidia-nccl-cu12==2.21.5
[pip3] nvidia-nvjitlink-cu12==12.4.127
[pip3] nvidia-nvtx-cu12==12.4.127
[pip3] torch==2.6.0
[pip3] torchvision==0.21.0
[pip3] triton==3.2.0
[conda] numpy 1.26.4 pypi_0 pypi
[conda] nvidia-cublas-cu12 12.4.5.8 pypi_0 pypi
[conda] nvidia-cuda-cupti-cu12 12.4.127 pypi_0 pypi
[conda] nvidia-cuda-nvrtc-cu12 12.4.127 pypi_0 pypi
[conda] nvidia-cuda-runtime-cu12 12.4.127 pypi_0 pypi
[conda] nvidia-cudnn-cu12 9.1.0.70 pypi_0 pypi
[conda] nvidia-cufft-cu12 11.2.1.3 pypi_0 pypi
[conda] nvidia-curand-cu12 10.3.5.147 pypi_0 pypi
[conda] nvidia-cusolver-cu12 11.6.1.9 pypi_0 pypi
[conda] nvidia-cusparse-cu12 12.3.1.170 pypi_0 pypi
[conda] nvidia-cusparselt-cu12 0.6.2 pypi_0 pypi
[conda] nvidia-nccl-cu12 2.21.5 pypi_0 pypi
[conda] nvidia-nvjitlink-cu12 12.4.127 pypi_0 pypi
[conda] nvidia-nvtx-cu12 12.4.127 pypi_0 pypi
[conda] torch 2.6.0 pypi_0 pypi
[conda] torchvision 0.21.0 pypi_0 pypi
[conda] triton 3.2.0 pypi_0 pypi
Information
Information
- The official example scripts
- My own modified scripts
🐛 Describe the bug
llama-stack 0.2.2 remote vllm always have default tool_system prompt even if there is no tool,this cause the model inference output to be wrong if no tools are provided, like this:
###############################
# 1.8 - Interative chat
###############################
User> who are you
> Response: {"name": "say", "parameters": {"text": "I am an assistant"}}
User> How can you help me?
> Response: {"name": "say", "parameters": {"text": "I can provide information on a wide range of topics, answer questions, and even help with tasks such as language translation and text summarization."}}
User>
How to reproduce:
- create a vllm server on localhost:8000 by
CUDA_VISIBLE_DEVICES=0 vllm serve meta-llama/Llama-3.2-11B-Vision-Instruct --enforce-eager --max-num-seqs 16 --tensor_parallel_size 1 --max_model_len 8196 --enable-auto-tool-choice --tool-call-parser llama3_json --chat-template examples/tool_chat_template_llama3.2_json.jinja - Curl this localhost:8000 directly with basic inference, with only see the default system prompt
Cutting Knowledge Date: December 2023\nToday Date: 14 Apr 2025, which is expected:
curl -X POST http://localhost:8000/v1/chat/completions\
-H "Content-Type: application/json" \
-d '{
"model_id": "meta-llama/Llama-3.1-8B-Instruct",
"messages": [
{
"role": "user",
"content": "who are you?"
}
],
"max_tokens": 100
}'
{"id":"chatcmpl-23284bf956134002ac80323315a146c3","object":"chat.completion","created":1744659246,"model":"meta-llama/Llama-3.1-8B-Instruct","choices":[{"index":0,"message":{"role":"assistant","reasoning_content":null,"content":"I'm an artificial intelligence model known as a Large Language Model (LLM). I'm a computer program designed to understand and generate human-like text. I'm here to assist you with any questions, provide information, or simply chat with you.\n\nI don't have a personal identity or emotions, but I'm designed to be helpful and engaging. I can provide information on a wide range of topics, from science and history to entertainment and culture. I can also help with tasks such as language translation,","tool_calls":[]},"logprobs":null,"finish_reason":"length","stop_reason":null}],"usage":{"prompt_tokens":39,"total_tokens":139,"completion_tokens":100,"prompt_tokens_details":null},"prompt_logprobs":null}%
----server-side-log-----
INFO 04-14 12:34:06 [logger.py:39] Received request chatcmpl-23284bf956134002ac80323315a146c3: prompt: '<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 14 Apr 2025\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nwho are you?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.0, temperature=0.6, top_p=0.9, top_k=-1, min_p=0.0, seed=None, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=100, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None.
- Now open a remote-vllm llama-stack server
INFERENCE_MODEL=meta-llama/Llama-3.1-8B-Instruct llama stack run remote-vllm - Curl the llama-stack endpoint
http://localhost:8321/v1/inference/chat-completion
curl -X POST http://localhost:8321/v1/inference/chat-completion \
-H "Content-Type: application/json" \
-d '{
"model_id": "meta-llama/Llama-3.1-8B-Instruct",
"messages": [
{
"role": "user",
"content": "who are you?"
}
],
"max_tokens": 100
}'
on the server side, the system prompt became a tool_call default system prompt even without any tools provided in the curl:
INFO 04-14 12:37:20 [logger.py:39] Received request chatcmpl-7c34fd60a8ea4e44acb7b0e4dbba6cb2: prompt: '<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nEnvironment: ipython\nCutting Knowledge Date: December 2023\nToday Date: 14 Apr 2025\n\nYou are a helpful assistant with tool calling capabilities. Only reply with a tool call if the function exists in the library provided by the user. If it doesn\'t exist, just reply directly in natural language. When you receive a tool call response, use the output to format an answer to the original user question.<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nGiven the following functions, please respond with a JSON for a function call with its proper arguments that best answers the given prompt.\n\nRespond in the format {"name": function name, "parameters": dictionary of argument name and its value}. Do not use variables.\n\nwho are you?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.0, temperature=0.0, top_p=1.0, top_k=-1, min_p=0.0, seed=None, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=4196, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None.
Error logs
See above logs
Expected behavior
When no tool provided, vllm should use default system prompt instead of tool-call system prompt, it felt like llama-stack is always providing some tools even if the request body did not have tool provided.. this will cause vllm output weird answers like
User> who are you
> Response: {"name": "say", "parameters": {"text": "I am an assistant"}}
User> How can you help me?
> Response: {"name": "say", "parameters": {"text": "I can provide information on a wide range of topics, answer questions, and even help with tasks such as language translation and text summarization."}}
User>