InternLM / InternLM/lmdeploy

[Bug] lmdeploy Auto AWQ量化后的4bit模型生成内容全部为"姿势姿势...."

Open
#2,073 1 comment 0 reactions 1 assignee Claimed by @AllentDan View on GitHub
Dominant language
Python
Stars
8.1k
Forks
748
Avg merge
6d 2h
Merged PRs (30d)
54

Description

### Checklist

- [X] 1. I have searched related issues but cannot get the expected help.
- [X] 2. The bug has not been fixed in the latest version.

### Describe the bug

开发机环境:
```python
# python version: 3.9.2
transformers==4.33.1
lmdeploy==0.5.1
```

使用command: lmdeploy lite auto_awq xxx --work-dir xxx 对原始模型进行量化后(to 4bit),量化后模型推理时不论输入什么样的prompt,输出都为"姿势姿势姿势..."

测试代码:
```python
model_path = "/mnt/bn/kefullmyg/workspace/yi34b_full_sft_0617-0623/lmdeploy_int4/"
# int8
# backend_config = PytorchEngineConfig(tp=1,session_len=2048)
backend_config = TurbomindEngineConfig(model_format='awq', tp=1,session_len=2048)

gen_config = GenerationConfig(top_p=0.9, top_k=50, temperature=0.7, max_new_tokens=500)
chat_template_config = ChatTemplateConfig("base", capability='completion', stop_words=[""]) # 0520 base 0510 yi_ies_cs
pipe = pipeline(
model_path=model_path,
backend_config=backend_config,
chat_template_config=chat_template_config
)
response = pipe(["Human: 为什么封我的直播间\n\n\n以上是你知道全部的信息,请做出回复。\nAssistant: "])
```

生成结果:
![image](https://github.com/user-attachments/assets/bd473f0d-a8e8-4d36-995b-6c2b98a13386)

### Reproduction

```python
model_path = "/mnt/bn/kefullmyg/workspace/yi34b_full_sft_0617-0623/lmdeploy_int4/"
# int8
# backend_config = PytorchEngineConfig(tp=1,session_len=2048)
backend_config = TurbomindEngineConfig(model_format='awq', tp=1,session_len=2048)

gen_config = GenerationConfig(top_p=0.9, top_k=50, temperature=0.7, max_new_tokens=500)
chat_template_config = ChatTemplateConfig("base", capability='completion', stop_words=[""]) # 0520 base 0510 yi_ies_cs
pipe = pipeline(
model_path=model_path,
backend_config=backend_config,
chat_template_config=chat_template_config
)
response = pipe(["Human: 为什么封我的直播间\n\n\n以上是你知道全部的信息,请做出回复。\nAssistant: "])
```

### Environment

```Shell
sys.platform: linux
Python: 3.9.2 (default, Feb 28 2021, 17:03:44) [GCC 10.2.1 20210110]
CUDA available: True
MUSA available: False
numpy_random_seed: 2147483648
GPU 0,1: NVIDIA A800-SXM4-40GB
CUDA_HOME: /usr/local/cuda
NVCC: Cuda compilation tools, release 11.7, V11.7.99
GCC: x86_64-linux-gnu-gcc (Debian 10.2.1-6) 10.2.1 20210110
PyTorch: 2.2.2+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 v3.3.2 (Git Hash 2dc95a2ad0841e29db8b22fbccaf3e5da7992b01)
- 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 8.9.2
- Magma 2.6.1
- Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=12.1, CUDNN_VERSION=8.9.2, 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_QNNPACK -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=2.2.2, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, 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: 0.17.2+cu121
LMDeploy: 0.5.1+6591540
transformers: 4.33.1
gradio: Not Found
fastapi: 0.111.1
pydantic: 2.6.2
triton: 2.2.0
NVIDIA Topology:
GPU0 GPU1 NIC0 NIC1 NIC2 NIC3 NIC4 CPU Affinity NUMA Affinity GPU NUMA ID
GPU0 X NV8 SYS NODE PIX SYS SYS 0-59 0 N/A
GPU1 NV8 X SYS SYS SYS PIX NODE 60-119 1 N/A
NIC0 SYS SYS X SYS SYS SYS SYS
NIC1 NODE SYS SYS X NODE SYS SYS
NIC2 PIX SYS SYS NODE X SYS SYS
NIC3 SYS PIX SYS SYS SYS X NODE
NIC4 SYS NODE SYS 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

NIC Legend:

NIC0: mlx5_0
NIC1: mlx5_1
NIC2: mlx5_2
NIC3: mlx5_3
NIC4: mlx5_4
```

### Error traceback

_No response_

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.