-
When there is a minimum value in the mask tensor, it can lead to incorrect Expand mask results and loss of mask calculation
- The mask tensor like this：
- The result as follow：
As …
-
### Reason/inspiration (optional)
We would like a new entry on the `.adjoint()` term under PyTorch. The entry should go in a new file under `docs/content/pytorch/concepts/tensor-operations/terms/adjo…
-
I want to use tensor parallelism with ouroboros, but I do not find the config to start the tensor parallel, can you give me an example?
-
## Summary
`forge/test/mlir/test_ops_tf.py::test_dual_conv2d` fails on Blackhole: Tensor mismatch.
## Observations
1. The output values are totally different.
2. Forge compilation looks fine. The s…
-
Filing an issue with respect to some artifacts in the IR that might be related to complex types used or otherwise. I suspect these are coming from the model and are harder to deal with in the compiler…
-
Does torch.export preserve the quantize_per_tensor/dequantize_per_tensor ops? I was testing with
```python
import torch
from torchao.quantization.quant_api import (
quantize_,
int8_dyna…
-
see https://github.com/EnzymeAD/Reactant.jl/pull/249
```mlir
module attributes {transform.with_named_sequence} {
func.func @main(%arg0: tensor) -> tensor {
%c = stablehlo.constant dense : tenso…
-
OpenAI has improved the way structured memory access is handled.
The PR : https://github.com/triton-lang/triton/pull/4916
cleans-up, extends the triton dialect with new operations and improves th…
-
### 🚀 The feature, motivation and pitch
I noticed that the current speculative mode does not support tp from this link (https://docs.vllm.ai/en/stable/models/spec_decode.html).
However, not suppo…
-
There is an incompatibility between xformers FA3 Torch custom op wrapper logic and recent `flashattn_hopper_cuda` changes, resulting in a `TypeError` due to changes in required arguments for the `fwd(…