Hi @tqchen
I am using the old NNVM graph interface to invoke the compilation:
with nnvm.compiler.build_config(opt_level=opt_level):
graph, lib, params = nnvm.compiler.build(sym, target, shape_dict,
dtype=dtype_dict, params=params)
Is there any way to access the raw weights, bias etc for Co
@masahi I think my effort to create
[MetalXLA](https://github.com/philipturner/metal-xla) would be the perfect
opportunity to experiment with using AutoTVM to accelerate training. It's a
real-time ML context where you have to balance compilation cost with code
optimization. Also, you would ei
try target="c". dont know if the code generated would work though
---
[Visit
Topic](https://discuss.tvm.apache.org/t/generate-native-c-code-from-tvm-ir/11792/3)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
here](https
