[Apache TVM Discuss] [Questions] Virtual threading scheduling primitive
为何在延迟隐藏中添加虚拟调度原语，他的实现或者如何使用 --- [Visit Topic](https://discuss.tvm.apache.org/t/virtual-threading-scheduling-primitive/8377/1) 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/1e2c9768a546609787fe75304535aa5183da372119d3d29141ca939cc5ce7d70).
[Apache TVM Discuss] [Questions] Why stop quantize after first ``nn.global_avg_pool2d``
Having the same question.@ziheng --- [Visit Topic](https://discuss.tvm.apache.org/t/why-stop-quantize-after-first-nn-global-avg-pool2d/8225/2) 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/3f0d44ce5a6bb0f844c46a997a13a3ea1b68d3b387e7a1dceddb5c22afcd714a).
[Apache TVM Discuss] [Questions] TVM terms: relay, topi, tir, te
Have you found out anything about explicit relay to tir translation. I am currently wondering, about the same question. --- [Visit Topic](https://discuss.tvm.apache.org/t/tvm-terms-relay-topi-tir-te/6474/3) 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/cf157ff3ce552b26ec23bc947f7fdeda473bd8e539145523d141b12c0e741106).
[Apache TVM Discuss] [Questions] Why stop quantize after first ``nn.global_avg_pool2d``
My guess is that tvm stops quantizing after the global average pooling for accuracy purposes. Usually in modern CNN after the global average pooling, you have the classifier (dense layer). In order to preserve accuracy the computation will be performed on 32 bit (instead of 8bit) --- [Visit Topic](https://discuss.tvm.apache.org/t/why-stop-quantize-after-first-nn-global-avg-pool2d/8225/3) 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/299c435bc3a55f02d23494d5a58150a878b1feb58c60350d95b3a529f683d881).
[Apache TVM Discuss] [Questions] TVM terms: relay, topi, tir, te
First of all, I'm by no means expert in TVM. So just my two cents. I believe the Relay-> Tir transform happens with so-called "lowering" process in side python/tvm/relay/backend/compile_engine.py, CompileEngine::lower(blah). --- [Visit Topic](https://discuss.tvm.apache.org/t/tvm-terms-relay-topi-tir-te/6474/4) 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/cb5f716cf5807cb5550ec8feb8de9b5dea1286cb65489fc10d16be9ee5e3d94a).
[Apache TVM Discuss] [Questions] TVM terms: relay, topi, tir, te
good summary, see also https://tvm.apache.org/docs/dev/index.html --- [Visit Topic](https://discuss.tvm.apache.org/t/tvm-terms-relay-topi-tir-te/6474/5) 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/87675e4de89c5c652ba66cb352609c1cc0779bde726ce14155e06c57b8750722).
[Apache TVM Discuss] [Questions] How can I test the performance of a single operator?
Hello @haozech. There are two ways you can go about benchmarking a single
operator. You can either 1. benchmark a specific implementation of the operator
or 2. benchmark all implementations of the operator.
For 1, follow the [Tuning High Performance Convolution on NVIDIA
GPUs](https://tvm.apache.org/docs/tutorials/autotvm/tune_conv2d_cuda.html)
tutorial, but skip section 1. In section two, replace
`"tutorial/conv2d_no_batching"` in `task =
autotvm.task.create("tutorial/conv2d_no_batching", args=(N, H, W, CO, CI, KH,
KW, strides, padding), target="cuda")` with the name of the implementation you
want to benchmark. You can find the name of the implementation by greping the
codebase for `@autotvm.register_topi_compute`. You'll also have to modify the
inputs to they match what the function is expecting. Furthermore, you'll have
to change `conv2d_no_batching` and `conv2d_nchw_python` in the last code block
with the correct function names (these should be the name of the function
annotated with `@autotvm.register_topi_compute`).
For 2, follow the [Auto-tuning a convolutional network for NVIDIA
GPU](https://tvm.apache.org/docs/tutorials/autotvm/tune_relay_cuda.html)
tutorial. Replace `get_network` with a function that returns a relay function
with a single operator like so:
```
x = relay.Var("x", tvm.relay.TensorType([40, 40]))
y = relay.Var("y", tvm.relay.TensorType([40, 40]))
mod = relay.Function(
[x, y],
relay.my_function(x, y)
)
```
---
[Visit
Topic](https://discuss.tvm.apache.org/t/how-can-i-test-the-performance-of-a-single-operator/8362/2)
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/b74575d4c7f5d2b0894f8cf942f2ce25244e267cad638a71bc9b4cf1da48c22c).
[Apache TVM Discuss] [Questions] Where does the layout transform of each op happen during alter_op_layout pass?
Hello! I'm trying to figure out the problem as said in the title. For example,
when a module is built like:
```
with autotvm.apply_graph_best(graph_opt_sch_file):
with tvm.transform.PassContext(opt_level=3):
graph_factory = relay.build_module.build(mod, target=target, params=params)
```
the function `CallWithNewLayouts` in `alter_op_layout.cc` will be called, and
it calls a series of functions all the way until
```
@conv2d_alter_layout.register("cpu")
def _alter_conv2d_layout(attrs, inputs, tinfos, out_type):
...
```
supposing the target is an x86 cpu. However, I only see this function changing
the layout info in `attrs`, yet to see any change of the actual layout of
tensors in the graph. If I debug this process and print the IR right after the
`AlterOpLayout` pass I can see the shapes of tensors changed accordingly from
4D to 5D/6D, and the `layout_transform` nodes are inserted. So my question here
is when does this happen? Can anyone give me a pointer to the code?
Many thanks!
---
[Visit
Topic](https://discuss.tvm.apache.org/t/where-does-the-layout-transform-of-each-op-happen-during-alter-op-layout-pass/8380/1)
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/61e3d22ff8a6f5cfe3478e4b613aaebc0ffe52b7d294860a0012706d9928d6d5).
[Apache TVM Discuss] [Questions] Where does the layout transform of each op happen during alter_op_layout pass?
When you see the tensors changed from 4D to 5D, the corresponding conv2d op has already been changed from NCHW to NCHWc; otherwise the type won't match. This is called "alter op layout". Specifically, the function you pointed returns the altered NCHWc op: https://github.com/apache/incubator-tvm/blob/main/python/tvm/topi/x86/conv2d_alter_op.py#L114 Accordingly, your graph changed from `4D -> conv2d_NCHW` to `4D -> layout_tranform -> 5D -> conv2d_NCHWc`. --- [Visit Topic](https://discuss.tvm.apache.org/t/where-does-the-layout-transform-of-each-op-happen-during-alter-op-layout-pass/8380/2) 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/99a1dcccdfb5c46494624a06f976af89df8c9552d7a4abda5d4a932576286c19).
[Apache TVM Discuss] [Questions] Where does the layout transform of each op happen during alter_op_layout pass?
I see, so are you saying the inputs in line 114 are already 5D? Or they're somehow converted to 5D? [quote="comaniac, post:2, topic:8380"] otherwise the type won’t match [/quote] Btw, here are you saying NCHW's inputs can only be 4D and NCHWc's 5D/6D? I'm actually experimenting a customer op. How do I let it accept both 4D and 5D/6D inputs? --- [Visit Topic](https://discuss.tvm.apache.org/t/where-does-the-layout-transform-of-each-op-happen-during-alter-op-layout-pass/8380/3) 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/ba904a304b915ae690112e33450f7c43bbe064ffb2adeccd83aa977ce54002a6).
