the inference of OFA-Sys/gsm8k-rft-llama13b2-u13b has shape error: 13Bllama2的u13b版本推理时出现矩阵形状错误

[Rem1L]

There seems no people tried your 13b2-u13b version and I may be the first one. But I got 'RuntimeError: mat1 and mat2 shapes cannot be multiplied (111x5120 and 1x2560)' on my inference. While the 7b version works well.

[Rem1L]

I'm not using accelerate and your script, I'm just using it as a object of LlamaForCausalLM and using bnb quantize for inference. But i don't think that would cause problem.

[Rem1L]

import torch
import sys
import random
import numpy as np
from transformers import LlamaTokenizer, LlamaForCausalLM, BitsAndBytesConfig

bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_use_double_quant=True,
    bnb_4bit_quant_type="nf4",
    # bnb_4bit_quant_type="fp4",
    bnb_4bit_compute_dtype=torch.bfloat16
)
random.seed(0)
np.random.seed(0)
torch.manual_seed(0)
torch.cuda.manual_seed(0)
torch.backends.cudnn.deterministic = True

device = "cuda:0"
tokenizer = LlamaTokenizer.from_pretrained("/data/haotian/RAP_tune/gsm8k-rft-llama13b2-u13b",legacy=False)
model = LlamaForCausalLM.from_pretrained(
        "/data/haotian/RAP_tune/gsm8k-rft-llama13b2-u13b",
        quantization_config=bnb_config,
        # torch_dtype=torch.float16,
        device_map="auto",
    )
model.config.pad_token_id = tokenizer.pad_token_id = 0  # unk
model.config.bos_token_id = 1
model.config.eos_token_id = 2
tokens = tokenizer("her eyes are so beautiful", return_tensors='pt', padding=True).to(device)
output = model.generate(**tokens, return_dict=True)
decoded = tokenizer.batch_decode(output, skip_special_tokens=True)
print(decoded)

Here is the minimal reproduction.

[Rem1L]

Nvidia driver version: 525.125.06
cuDNN version: Could not collect
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
Byte Order:                      Little Endian
Address sizes:                   48 bits physical, 48 bits virtual
CPU(s):                          56
On-line CPU(s) list:             0-55
Thread(s) per core:              1
Core(s) per socket:              28
Socket(s):                       2
NUMA node(s):                    8
Vendor ID:                       AuthenticAMD
CPU family:                      25
Model:                           1
Model name:                      AMD EPYC 7453 28-Core Processor
Stepping:                        1
Frequency boost:                 enabled
CPU MHz:                         2779.099
CPU max MHz:                     2750.0000
CPU min MHz:                     1500.0000
BogoMIPS:                        5499.64
Virtualization:                  AMD-V
L1d cache:                       1.8 MiB
L1i cache:                       1.8 MiB
L2 cache:                        28 MiB
L3 cache:                        128 MiB
NUMA node0 CPU(s):               0-6
NUMA node1 CPU(s):               7-13
NUMA node2 CPU(s):               14-20
NUMA node3 CPU(s):               21-27
NUMA node4 CPU(s):               28-34
NUMA node5 CPU(s):               35-41
NUMA node6 CPU(s):               42-48
NUMA node7 CPU(s):               49-55
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 and seccomp
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Retpolines, IBPB conditional, IBRS_FW, STIBP disabled, RSB filling, PBRSB-eIBRS Not affected
Vulnerability Srbds:             Not affected
Vulnerability Tsx async abort:   Not affected
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 nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 fma cx16 pcid sse4_1 sse4_2 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 ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 invpcid cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr wbnoinvd arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold v_vmsave_vmload vgif umip pku ospke vaes vpclmulqdq rdpid overflow_recov succor smca

Versions of relevant libraries:
[pip3] numpy==1.25.2
[pip3] torch==2.0.1
[pip3] torchvision==0.15.2
[pip3] triton==2.0.0
[conda] numpy                     1.25.2                   pypi_0    pypi
[conda] torch                     2.0.1                    pypi_0    pypi
[conda] torchvision               0.15.2                   pypi_0    pypi
[conda] triton                    2.0.0                    pypi_0    pypi

my environment

[GanjinZero]

What is your transformers version?

[Rem1L]

It's 4.33.2

[GanjinZero]

try transformers==4.29.2

env see issue 9.

[Rem1L]

If I want to do some work with new transformer, can I just do some modify to the config to make it work. Do you know what lead to this problem?

[GanjinZero]

I have no idea how it work on new version; you may train a new model based on our code.