Hi,
I copied part of relay dump from mobilenet and reproduce with python script
below
> data = relay.var("56", shape=[1, 16, 16, 512], dtype="uint8")
> kernel = relay.var("_param_39", shape=[3, 3, 512, 1], dtype="uint8")
> bias = relay.var("_param_40", shape=[512], dtype="int32")
>
> #define relay
>
> conv = relay.qnn.op.conv2d(data, kernel, relay.const(0, dtype="int32"),
> relay.const(134, dtype="int32"), relay.const(0.0235285, dtype="float32"),
> relay.const(0.0243294, dtype="float32"), kernel_size=[3, 3], channels=512,
> padding=[1, 1, 1, 1], groups=512, data_layout="NHWC", kernel_layout="HWOI",
> out_dtype="int32")
>
> ba = relay.nn.bias_add(conv, bias, 3)
>
> out = relay.qnn.op.requantize(ba, relay.const(0.000572435, dtype="float32"),
> relay.const(0, dtype="int32"), relay.const(0.0235285, dtype="float32"),
> relay.const(0, dtype="int32"), axis=3, out_dtype="uint8")
>
> #define params
> params = dict()
> params["56"] = np.zeros((1, 16, 16, 512), dtype="uint8")
> params["_param_39"] = np.zeros((3, 3, 512, 1), dtype="uint8")
> params["_param_40"] = np.zeros((512,), dtype="int32")
>
> func = relay.function.Function([data, kernel, bias], out)
>
> tasks, task_weights = auto_scheduler.extract_tasks(func, params,
> tvm.target.Target("llvm"))
>
> print(len(tasks))
The relay dump from this script was correct as i expected, but extract_task
returned no task to auto-schedule.
What am i missing?
---
[Visit
Topic](https://discuss.tvm.apache.org/t/auto-schedule-extract-tasks-not-extract-any-task-from-function-write-in-relay/12309/1)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
here](https://discuss.tvm.apache.org/email/unsubscribe/f0b93dca4b644fe52a23da6718d3add014235ed87a14f142a7258b5ac4537497).
