Dear Team,
I am currently trying to deploy a model that takes two input images as input. I get around that I have made changes to my model(OpticalFlow based Model) and it now takes (1,6,H,W). I have written a custom DataAdapter which makes sure that the input is concatenation of two images.
I am getting following error when deploying
ERROR :axelera.runtime: Expecting 2/1 input/output arguments, but got 1/1 instead.
ERROR :axelera.app.pipe.torch: Pipeline error occurred: AXR_ERROR_INVALID_ARGUMENT: Expecting 2/1 input/output arguments, but got 1/1 instead.
ERROR :axelera.app.pipe.torch: Traceback (most recent call last):
ERROR :axelera.app.pipe.torch: File "/root/voyager-sdk/axelera/app/pipe/torch.py", line 52, in _loop
ERROR :axelera.app.pipe.torch: image, result, meta = model_pipe.inference.exec_torch(
ERROR :axelera.app.pipe.torch: File "/root/voyager-sdk/axelera/app/operators/inference.py", line 1017, in exec_torch
ERROR :axelera.app.pipe.torch: self._axr_modeli.run(inputs, outputs)
ERROR :axelera.app.pipe.torch: File "/root/.cache/axelera/venvs/295e682e/lib/python3.10/site-packages/axelera/runtime/objects.py", line 559, in run
ERROR :axelera.app.pipe.torch: _raise_error(self.context, res.value)
ERROR :axelera.app.pipe.torch: File "/root/.cache/axelera/venvs/295e682e/lib/python3.10/site-packages/axelera/runtime/objects.py", line 83, in _raise_error
ERROR :axelera.app.pipe.torch: raise exc(f"{err}: {msg}")
ERROR :axelera.app.pipe.torch: axelera.runtime.objects.InvalidArgument: AXR_ERROR_INVALID_ARGUMENT: Expecting 2/1 input/output arguments, but got 1/1 instead.
DEBUG :axelera.app.stream: Pipeline my-custom-model implicit _PIPELINE_REMOVED
DEBUG :axelera.app.stream: Emitting my-custom-model: PIPELINE_REMOVED
DEBUG :axelera.app.stream: Got my-custom-model: PIPELINE_REMOVED
Core Temp : 0.0°C
CPU % : 27.2%
End-to-end : 0.0fps
Latency : 0.0ms (min:inf max:-inf σ:0.0 x̄:0.0)ms
DEBUG :axelera.app.stream: Stream stop requested
ERROR :axelera.app.utils: Traceback (most recent call last):
ERROR :axelera.app.utils: File "/root/voyager-sdk/axelera/app/utils.py", line 692, in run
ERROR :axelera.app.utils: super().run()
ERROR :axelera.app.utils: File "/root/.pyenv/versions/3.10.6/lib/python3.10/threading.py", line 953, in run
ERROR :axelera.app.utils: self._target(*self._args, **self._kwargs)
ERROR :axelera.app.utils: File "/root/voyager-sdk/axelera/app/pipe/torch.py", line 52, in _loop
ERROR :axelera.app.utils: image, result, meta = model_pipe.inference.exec_torch(
ERROR :axelera.app.utils: File "/root/voyager-sdk/axelera/app/operators/inference.py", line 1017, in exec_torch
ERROR :axelera.app.utils: self._axr_modeli.run(inputs, outputs)
ERROR :axelera.app.utils: File "/root/.cache/axelera/venvs/295e682e/lib/python3.10/site-packages/axelera/runtime/objects.py", line 559, in run
ERROR :axelera.app.utils: _raise_error(self.context, res.value)
ERROR :axelera.app.utils: File "/root/.cache/axelera/venvs/295e682e/lib/python3.10/site-packages/axelera/runtime/objects.py", line 83, in _raise_error
ERROR :axelera.app.utils: raise exc(f"{err}: {msg}")
ERROR :axelera.app.utils: axelera.runtime.objects.InvalidArgument: AXR_ERROR_INVALID_ARGUMENT: Expecting 2/1 input/output arguments, but got 1/1 instead.
the yaml i am using
axelera-model-format: 1.0.0
name: my-custom-model
description: Custom Model
pipeline:
- inference:
model_name: model2
input:
type: image
models:
model2:
class: CustomAxPytorchModelWithPreprocess
class_path: /root/voyager-sdk/ax_models/custom/simplest_torch.py
weight_path: /root/voyager-sdk/ax_models/custom/mymodel.pth
task_category: DepthEstimation
input_tensor_layout: NCHW
input_tensor_shape: [1, 6, 128, 128]
input_color_format: RGB
dataset: Custom-Sintel-Pairs
# Use extra_kwargs to pass compiler configuration
extra_kwargs:
compilation_config:
enable_icr: false
enable_swicr: false
save_error_artifact: true
trace_tvm_passes: true
datasets:
Custom-Sintel-Pairs:
class: ConsecutiveImagePairAdapter
class_path: /root/voyager-sdk/ax_models/custom/dataadpter.py
data_dir_name: /root/voyager-sdk/data/
cal_data: sintel_resized
val_data: sintel_resized
