Tutorial errors

#1

I want to run the toturial of " Deploy the Pretrained Model on Raspberry Pi",but when I ran the code:

local_demo = False

if local_demo:
    target = tvm.target.Target("llvm")
else:
    target = tvm.target.arm_cpu("rasp3b")

with tvm.transform.PassContext(opt_level=3):
    lib = relay.build(func, target, params=params)

tmp = util.tempdir()
lib_fname = tmp.relpath("net.tar")
lib.export_library(lib_fname)

I got this:

> "-target" is deprecated, use "-mtriple" instead.
> "-target" is deprecated, use "-mtriple" instead.
> "-target" is deprecated, use "-mtriple" instead.
> "-target" is deprecated, use "-mtriple" instead.
> "-target" is deprecated, use "-mtriple" instead.
> "-target" is deprecated, use "-mtriple" instead.
> Cannot find config for target=llvm -keys=arm_cpu,cpu -device=arm_cpu -mattr=+neon -model=bcm2837 -mtriple=armv7l-linux-gnueabihf, workload=('dense_nopack.x86', ('TENSOR', (1, 512), 'float32'), ('TENSOR', (1000, 512), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.

Could someone help me,thaks very much!

#2

This is not a problem. TVM has done it automatically (replace target to mtriple) for u.

#3

Thank for replying me. But I have ran the toturial(https://tvm.apache.org/docs/tutorials/frontend/deploy_model_on_rasp.html)step by step. I changed the version of tvm and llvm, the errors below still exist:

RPCError: Traceback (most recent call last):
  [bt] (8) /home/ljs/tvm/build/libtvm.so(std::_Function_handler<void (tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*), tvm::runtime::RPCModuleNode::WrapRemoteFunc(void*)::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#1}>::_M_invoke(std::_Any_data const&, tvm::runtime::TVMArgs&&, tvm::runtime::TVMRetValue*&&)+0x33) [0x7ff0e9e88b43]
  [bt] (7) /home/ljs/tvm/build/libtvm.so(tvm::runtime::RPCWrappedFunc::operator()(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const+0x3c5) [0x7ff0e9e885a5]
  [bt] (6) /home/ljs/tvm/build/libtvm.so(tvm::runtime::RPCClientSession::CallFunc(void*, TVMValue const*, int const*, int, std::function<void (tvm::runtime::TVMArgs)> const&)+0x57) [0x7ff0e9e7c397]
  [bt] (5) /home/ljs/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::CallFunc(void*, TVMValue const*, int const*, int, std::function<void (tvm::runtime::TVMArgs)>)+0x215) [0x7ff0e9e73dd5]
  [bt] (4) /home/ljs/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::HandleUntilReturnEvent(bool, std::function<void (tvm::runtime::TVMArgs)>)+0x1ab) [0x7ff0e9e72d8b]
  [bt] (3) /home/ljs/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::EventHandler::HandleNextEvent(bool, bool, std::function<void (tvm::runtime::TVMArgs)>)+0xd7) [0x7ff0e9e7c0a7]
  [bt] (2) /home/ljs/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::EventHandler::HandleProcessPacket(std::function<void (tvm::runtime::TVMArgs)>)+0x126) [0x7ff0e9e7be86]
  [bt] (1) /home/ljs/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::EventHandler::HandleReturn(tvm::runtime::RPCCode, std::function<void (tvm::runtime::TVMArgs)>)+0x13f) [0x7ff0e9e7b3af]
  [bt] (0) /home/ljs/tvm/build/libtvm.so(+0x17ab042) [0x7ff0e9e71042]
  [bt] (8) /home/ubuntu/tvm/build/libtvm.so(+0x103b100) [0xffff9d1bc100]
  [bt] (7) /home/ubuntu/tvm/build/libtvm.so(tvm::runtime::RPCServerLoop(int)+0xac) [0xffff9d1bb7b4]
  [bt] (6) /home/ubuntu/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::ServerLoop()+0xe8) [0xffff9d19f0c8]
  [bt] (5) /home/ubuntu/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::HandleUntilReturnEvent(bool, std::function<void (tvm::runtime::TVMArgs)>)+0x258) [0xffff9d19eb40]
  [bt] (4) /home/ubuntu/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::EventHandler::HandleNextEvent(bool, bool, std::function<void (tvm::runtime::TVMArgs)>)+0x1e4) [0xffff9d1a7184]
  [bt] (3) /home/ubuntu/tvm/build/libtvm.so(tvm::runtime::RPCEndpoint::EventHandler::HandleProcessPacket(std::function<void (tvm::runtime::TVMArgs)>)+0x168) [0xffff9d1a6ea0]
  [bt] (2) /home/ubuntu/tvm/build/libtvm.so(tvm::runtime::RPCSession::AsyncCallFunc(void*, TVMValue const*, int const*, int, std::function<void (tvm::runtime::RPCCode, tvm::runtime::TVMArgs)>)+0x58) [0xffff9d1ba368]
  [bt] (1) /home/ubuntu/tvm/build/libtvm.so(tvm::runtime::LocalSession::CallFunc(void*, TVMValue const*, int const*, int, std::function<void (tvm::runtime::TVMArgs)> const&)+0x74) [0xffff9d1ad9c4]
  [bt] (0) /home/ubuntu/tvm/build/libtvm.so(+0xfbf140) [0xffff9d140140]
  File "/home/ubuntu/tvm/python/tvm/_ffi/_ctypes/packed_func.py", line 81, in cfun
    rv = local_pyfunc(*pyargs)
  File "/home/ubuntu/tvm/python/tvm/rpc/server.py", line 69, in load_module
    m = _load_module(path)
  File "/home/ubuntu/tvm/python/tvm/runtime/module.py", line 411, in load_module
    _cc.create_shared(path + ".so", files)
  File "/home/ubuntu/tvm/python/tvm/contrib/cc.py", line 43, in create_shared
    _linux_compile(output, objects, options, cc, compile_shared=True)
  File "/home/ubuntu/tvm/python/tvm/contrib/cc.py", line 205, in _linux_compile
    raise RuntimeError(msg)
  File "/home/ljs/tvm/src/runtime/rpc/rpc_endpoint.cc", line 370
RPCError: Error caught from RPC call:
RuntimeError: Compilation error:
/tmp/tmpvh_4_v0b/net/lib0.o: error adding symbols: File in wrong format
collect2: error: ld returned 1 exit status

Command line: g++ -shared -fPIC -o /tmp/tmpvh_4_v0b/net.tar.so /tmp/tmpvh_4_v0b/net/lib0.o /tmp/tmpvh_4_v0b/net/devc.o
#4

I have replied on your issue before. This is because of your created target and compiler doesn’t match. If you use TVM compile it for arm, but you use g++ (on your host x86 machine) or you use TVM compile it for x86, but you use g++ (remote arm machine), you will meet this kind of error.

1 Like
#5

Thanks! I’m new to ai so it’s difficult for me. So to run the tutorial, do I need to re-make the tvm, modify the config.cmake. Set the “USE_ARM_COMPUTE_LIB” and “USE_ARM_COMPUTE_LIB_GRAPH_RUNTIME” on.

#6

Ignore this reply:

No, these two options are not for your case. For your case:

tmp = util.tempdir()
lib_fname = tmp.relpath("net.tar")
# lib.export_library(lib_fname) # remove it
1 Like
#7

I’m sorry to bother you again. I removed the line then got a new error. I tried to debug it by myself but losed. If you are in spare time, please help me.

2020-09-22 17-23-24 的屏幕截图
2020-09-22 17-23-24 的屏幕截图1438×494 70.8 KB

#8

Sorry, my previous reply is wrong. Please add that line back. Even it is tar, we still need to export_library. Would you mind pasting all your code? There is something you have done wrong.

Another thing is I want you to tell me:

  1. /home/ubuntu/tvm/build/ —> This is arm machine or x86 machine?
  2. /home/ljs/tvm/build/ —> This is arm machine or x86 machine?
[VTA]AutoTVM tutorial error with pynq - export_library
#9

Thank you very much for your reply.

/home/ubuntu/tvm/build/ --> is my arm machine(raspberry 3b) . 
 I have built  ubuntu18.04.5 64bit and tvm0.7 on it.

/home/ljs/tvm/build/ -->is my x64 machine.
(cpu is AMD® Ryzen 5 4600h with radeon graphics × 12 )    
I have built LLVM12.0.0 and tvm0.7 on it.

My code is from the tvm tutorials(https://tvm.apache.org/docs/tutorials/frontend/deploy_model_on_rasp.html):

%matplotlib inline

import tvm
from tvm import te
import tvm.relay as relay
from tvm import rpc
from tvm.contrib import util, graph_runtime as runtime
from tvm.contrib.download import download_testdata

from mxnet.gluon.model_zoo.vision import get_model
from PIL import Image
import numpy as np
# one line to get the model
block = get_model("resnet18_v1", pretrained=True)

img_url = "https://github.com/dmlc/mxnet.js/blob/master/data/cat.png?raw=true"
img_name = "cat.png"
img_path = download_testdata(img_url, img_name, module="data")
image = Image.open(img_path).resize((224, 224))

def transform_image(image):
    image = np.array(image) - np.array([123.0, 117.0, 104.0])
    image /= np.array([58.395, 57.12, 57.375])
    image = image.transpose((2, 0, 1))
    image = image[np.newaxis, :]
    return image
x = transform_image(image)

synset_url = "".join(
    [
        "https://gist.githubusercontent.com/zhreshold/",
        "4d0b62f3d01426887599d4f7ede23ee5/raw/",
        "596b27d23537e5a1b5751d2b0481ef172f58b539/",
        "imagenet1000_clsid_to_human.txt",
    ]
)
synset_name = "imagenet1000_clsid_to_human.txt"
synset_path = download_testdata(synset_url, synset_name, module="data")
with open(synset_path) as f:
    synset = eval(f.read())

shape_dict = {"data": x.shape}
mod, params = relay.frontend.from_mxnet(block, shape_dict)
# we want a probability so add a softmax operator
func = mod["main"]
func = relay.Function(func.params, relay.nn.softmax(func.body), None, func.type_params, func.attrs)

batch_size = 1
num_classes = 1000
image_shape = (3, 224, 224)
data_shape = (batch_size,) + image_shape

local_demo = False

if local_demo:
    target = tvm.target.create('llvm')
else:
    target = tvm.target.arm_cpu('rasp3b')
    # The above line is a simple form of
    # target = tvm.target.create('llvm -device=arm_cpu -model=bcm2837 -mtriple=armv7l-linux-gnueabihf -mattr=+neon')

with tvm.transform.PassContext(opt_level=3):
    lib = relay.build(func, target, params=params)

# After `relay.build`, you will get three return values: graph,
# library and the new parameter, since we do some optimization that will
# change the parameters but keep the result of model as the same.

# Save the library at local temporary directory.
tmp = util.tempdir()
lib_fname = tmp.relpath('net.tar')
lib.export_library(lib_fname)


# obtain an RPC session from remote device.
if local_demo:
    remote = rpc.LocalSession()
else:
    # The following is my environment, change this to the IP address of your target device
    host = '192.168.43.166'
    port = 9090
    remote = rpc.connect(host, port)

# upload the library to remote device and load it
remote.upload(lib_fname)
rlib = remote.load_module('.tar')

# create the remote runtime module
ctx = remote.cpu(0)
module = runtime.GraphModule(rlib['default'](ctx))
# set input data
module.set_input('data', tvm.nd.array(x.astype('float32')))
# run
module.run()
# get output
out = module.get_output(0)
# get top1 result
top1 = np.argmax(out.asnumpy())
print('TVM prediction top-1: {}'.format(synset[top1]))
#10

OK. Thanks for your information. Please change this line:

target = tvm.target.arm_cpu('rasp3b')

--->
target = tvm.target.create('llvm -device=arm_cpu -model=bcm2837 -mtriple=aarch64-linux-gnu -mattr=+neon -mcpu=cortex-a53')

You will pass.

The issue is tvm assumes rasp 3b run 32 bits os.

1 Like
#11

Thank you so much!!! I ran the tutorial successfully using your advice!!! :blush: :blush: