-
When calling `ttnn.max_pool2d` with an input tensor in `TILE_LAYOUT`, I get the error below for certain shapes (see the shape in the example below)
```
RuntimeError: TT_FATAL @ ../tt_metal/common/mat…
-
### What happened?
The following dispatch exceeds the shared memory limit and results in error while compiling vae model. Here's the link to the whole VAE model (https://sharkpublic.blob.core.windows…
-
### What happened?
Switch to https://github.com/iree-org/iree/pull/18826
Asan error gist: https://gist.github.com/pashu123/29bb25c4d7ba7261bd6f903513dd5bf3
### Steps to reproduce your issue
Input IR…
-
Its worth fusing the attention mask with the operation rather than generating and loading as it can entirely be generated from the loop iterators. The following block should be a single dispatch.
```…
-
I am having an [issue](https://github.com/evilsocket/cake/issues/36) with deploying Qwen2.5-Coder with the WIP program [Cake](https://github.com/evilsocket/cake).
The following error happens when u…
-
Hi!
I noticed that converting COO to another format, like CSR, returns an incorrect result. Here's a script where COO is converted to CSR - it reproduces the issue: https://gist.github.com/mtsokol/…
-
By running this simple test in python:
```
import pytest
import torch
import ttnn
@pytest.mark.parametrize("dims", [(32, 32), (64, 64)])
def test_add_with_block_sharding(device, dims):
torch.manu…
-
Is there a plan to support the Tensor Parallel and Context Parallel? I think this would make the project more flexible and attract more users.
-
**Edited this issue to include a clearer issue and task list**
Here are some issues we run into with `ttnn.max_pool2d` and `ttnn.conv2d` when trying to use them in models:
- `ttnn.max_pool2d`
- [ …
-
**Describe the bug**
When setting the inference device to "Compute Shader" in Unity ML-Agents, this error occurs:
`InvalidOperationException: Tensor data cannot be read from, use .ReadbackAndClone…