-
I noticed that when I process a video frame with a standard 16:9 aspect ratio, the processed output frame isn't zero-padded, and the aspect ratio is distorted. Is this intended?
I've included an ex…
-
### Problem Description
Hi, I am getting an error when running the `tune_gemm.py` script.
I am inside a docker container with access to 8 AMD MI300X gpus, displayed when calling `rocm-smi` and I ha…
-
### Feature request
This request proposes one of three changes (see **Motivation** for background, and **Your contribution** more thoughts on possible solutions) in order to allow saving of a certa…
-
### 🚀 The feature, motivation and pitch
Dynamic shapes tracing is the way to avoid recompilations of functions when input tensors' shapes change. However, it is not guaranteed that the compiler wil…
-
Traceback (most recent call last):
File "/gits/ComfyUI/custom_nodes/ComfyUI-GGUF/tools/convert.py", line 165, in handle_tensors
data = gguf.quants.quantize(data, data_qtype)
AttributeError: m…
-
As we found out in tenstorrent/pytorch2.0_ttnn#198, several ops produce (1, N) tensors when (N,) tensors are expected.
Affected ops:
- `ceil`
- `floor`
- `gelu`
- `rsqrt`
- `sqrt`
Spared op…
-
Currently the BitsAndBytesLinearQuant4bit for submodule always calls `bitsandbytes.functional.quantize_4bit`. This is somewhat touchy for CPU tensors because `quantize_4bit` only works on GPU tensors …
t-vi updated
2 months ago
-
The following code
```python
import pytaco as pt
a = pt.tensor([3,3], pt.dense)
b = pt.tensor([3,3], pt.dense)
c = a + b
```
produces c whose format is Format(({compressed},{compressed}; 0,1)) …
-
### Summary
TTNN `as_tensor` fails when creating multi-device tensors in tile layout that are in bfloat8 format. The following test demonstrates this:
```python
@pytest.mark.parametrize("width"…
-
### 🚀 The feature, motivation and pitch
I am using torch as part of some code that solves quantum systems in a differentiable way. An important part of this process is constructing large sparse mat…