search

vllm-project / vllm

A high-throughput and memory-efficient inference and serving engine for LLMs
https://docs.vllm.ai
Apache License 2.0
27.2k stars 3.99k forks source link

[Bug]: The LLama 3 base generation does not stop based on the passed stop words #4188

Closed wenhuchen closed 2 months ago

wenhuchen commented 5 months ago

Your current environment

Libc version: glibc-2.35

Python version: 3.10.12 (main, Jun 11 2023, 05:26:28) [GCC 11.4.0] (64-bit runtime) Python platform: Linux-5.4.250-2-velinux1u1-amd64-x86_64-with-glibc2.35 Is CUDA available: True CUDA runtime version: 12.3.52 CUDA_MODULE_LOADING set to: LAZY GPU models and configuration: GPU 0: NVIDIA A800-SXM4-80GB GPU 1: NVIDIA A800-SXM4-80GB GPU 2: NVIDIA A800-SXM4-80GB GPU 3: NVIDIA A800-SXM4-80GB

Nvidia driver version: 470.161.03 cuDNN version: Probably one of the following: /usr/lib/x86_64-linux-gnu/libcudnn.so.8.9.6 /usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.9.6 /usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.9.6 /usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.9.6 /usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.9.6 /usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.9.6 /usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.9.6 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: 46 bits physical, 57 bits virtual Byte Order: Little Endian CPU(s): 128 On-line CPU(s) list: 0-53 Off-line CPU(s) list: 54-127 Vendor ID: GenuineIntel Model name: Intel(R) Xeon(R) Platinum 8336C CPU @ 2.30GHz CPU family: 6 Model: 106 Thread(s) per core: 2 Core(s) per socket: 32 Socket(s): 2 Stepping: 6 CPU max MHz: 3500.0000 CPU min MHz: 800.0000 BogoMIPS: 4600.00 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 invpcid_single ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local wbnoinvd dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg tme avx512_vpopcntdq rdpid md_clear pconfig flush_l1d arch_capabilities Virtualization: VT-x L1d cache: 3 MiB (64 instances) L1i cache: 2 MiB (64 instances) L2 cache: 80 MiB (64 instances) L3 cache: 108 MiB (2 instances) NUMA node(s): 2 NUMA node0 CPU(s): 0-31,64-95 NUMA node1 CPU(s): 32-63,96-127 Vulnerability Itlb multihit: Not affected Vulnerability L1tf: Not affected Vulnerability Mds: Not affected Vulnerability Meltdown: Not affected Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable Vulnerability Retbleed: Not affected Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization Vulnerability Spectre v2: Mitigation; Enhanced IBRS, IBPB conditional, RSB filling, PBRSB-eIBRS SW sequence Vulnerability Srbds: Not affected Vulnerability Tsx async abort: Not affected

Versions of relevant libraries: [pip3] numpy==1.24.4 [pip3] onnx==1.14.1 [pip3] optree==0.10.0 [pip3] pytorch-quantization==2.1.2 [pip3] torch==2.1.2 [pip3] torch-tensorrt==2.2.0a0 [pip3] torchaudio==2.1.1+cu121 [pip3] torchdata==0.7.0a0 [pip3] torchtext==0.16.0a0 [pip3] torchvision==0.16.1+cu121 [pip3] triton==2.1.0 [conda] Could not collectROCM Version: Could not collect Neuron SDK Version: N/A vLLM Version: 0.4.0.post1 vLLM Build Flags: CUDA Archs: 5.2 6.0 6.1 7.0 7.2 7.5 8.0 8.6 8.7 9.0+PTX; ROCm: Disabled; Neuron: Disabled GPU Topology: GPU0 GPU1 GPU2 GPU3 mlx5_1 mlx5_2 mlx5_3 mlx5_4 CPU Affinity NUMA Affinity GPU0 X NV8 NV8 NV8 PXB NODE SYS SYS 0-31,64-95 0 GPU1 NV8 X NV8 NV8 PXB NODE SYS SYS 0-31,64-95 0 GPU2 NV8 NV8 X NV8 SYS SYS PXB NODE 32-63,96-127 1 GPU3 NV8 NV8 NV8 X SYS SYS PXB NODE 32-63,96-127 1 mlx5_1 PXB PXB SYS SYS X NODE SYS SYS mlx5_2 NODE NODE SYS SYS NODE X SYS SYS mlx5_3 SYS SYS PXB PXB SYS SYS X NODE mlx5_4 SYS SYS NODE NODE SYS SYS NODE X

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

🐛 Describe the bug

  stop_tokens = ["Problem:"]
  sampling_params = SamplingParams(temperature=0, top_p=1, max_tokens=args.model_max_length, stop=stop_tokens)
  llm = LLM(model="meta-llama/Meta-Llama-3-8B", tensor_parallel_size=torch.cuda.device_count(), dtype=args.dtype, trust_remote_code=True)
  llm.generate(["Problem: .... Ans ...\nProblem: .... Ans ...\nProblem: ... Ans ...\nProblem:"])

The model does not stop even if it encounters "Problem:"

youkaichao commented 5 months ago

Looks like this will be solved by https://github.com/vllm-project/vllm/pull/4182 . Can you try to build from the latest main? That PR just gets merged.

simon-mo commented 5 months ago

Actually this is not solved. We will have to wait for Meta and HF team to update the generation config in the model huh. A temporary solution is available at #4180

haqishen commented 5 months ago

Pls refer to this https://github.com/vllm-project/vllm/issues/4180#issuecomment-2066004748

nelson-liu commented 4 months ago

I believe the issue in #4180 is specific to the instruction-tuned model. However, the issue described here (which I've also seen) concerns how vllm handles the stop= argument of SamplingParams.

DarkLight1337 commented 2 months ago

Actually this is not solved. We will have to wait for Meta and HF team to update the generation config in the model huh

It has been done here, so I'm closing this.