launch function crashed
- Dominant language
- LLVM
- Stars
- 40.5k
- Forks
- 18.7k
- PR merge metrics
- PR metrics pending
Description
import numpy as np
from horizon_tc_ui.hb_runtime import HBRuntime
data = np.load("/open_explorer/PTQ/bev/20260411/npy/img_0_1700792322.npy")
sess = HBRuntime("/open_explorer/PTQ/bev/20260411/model_output_quannt_config_no_remove_dequantize/dx_bev_quantized_model.bc")
input_names = sess.input_names
print("input_names:" + str(input_names))
output_names = sess.output_names
print("output_names:" + str(output_names))
output = sess.run(output_names, {input_names[0]: data})
print(output)
Stack dump:
0. Program arguments:
Stack dump without symbol names (ensure you have llvm-symbolizer in your PATH or set the environment var `LLVM_SYMBOLIZER_PATH` to point to it):
0 _hbdk.cpython-310-x86_64-linux-gnu.so 0x00007f2c019f7ffb
Traceback (most recent call last):
File "/open_explorer/code/model-arcade/tools/parkonenet/model_eval/test_J6B_mdoel.py", line 16, in
output = sess.run(output_names, {input_names[0]: data})
File "/usr/local/lib/python3.10/dist-packages/horizon_tc_ui/hb_runtime.py", line 86, in run
return self.sess.run(output_names, input_feed, **extra_args)
File "/usr/local/lib/python3.10/dist-packages/horizon_tc_ui/hb_hbirruntime.py", line 227, in run
result = self.function.feed(inputs=input_info)
File "/usr/local/lib/python3.10/dist-packages/hbdk4/compiler/overlay_utils/xq_utils.py", line 124, in feed
return self._dict_feed(inputs)
File "/usr/local/lib/python3.10/dist-packages/hbdk4/compiler/overlay_utils/xq_utils.py", line 120, in _dict_feed
results = self._launch(*args)
File "/usr/local/lib/python3.10/dist-packages/hbdk4/compiler/overlay_utils/xq_utils.py", line 100, in _launch
self.session.launch(
RuntimeError: launch function crashed!
Contributor guide
Research direction
Reproduce the call in test_J6B_mdoel.py with the referenced .npy input and .bc model, then trace through horizon_tc_ui/hb_runtime.py, hb_hbirruntime.py, and hbdk4/compiler/overlay_utils/xq_utils.py. Check whether the input names, array data, or runtime/model combination triggers the native crash; done means the failure is explained and the launch path completes or reports a useful error.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- numpy, python
- Domain
- compilers
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 28/100