-
### What happened?
### **Description**:
When using the RPC backend in `llama.cpp`, I encountered crashes in the `rpc_server::deserialize_tensor` function. The assert fails because `tensor->ne[…
-
### 🐛 Describe the bug
Code to reproduce:
```python
import torch
device = torch.device('cpu')
dtype = torch.bfloat16
n = 512
x = torch.tensor([0, 1], dtype=dtype, device=device).repeat(…
-
Hi,
I am trying to trace crf.decode function in ml.inf2.8xlarge instance.
Here is the code I am trying to run.
def tags(output , mask):
a = model.crf.decode(output , mask)
a = torch…
-
### What happened?
for the given IR
```mlir
#map = affine_map (d0, d1, d2, d3)>
#map1 = affine_map (0, d1, d2, d3)>
module {
ml_program.global private mutable @global_seed(dense : tensor) : tenso…
-
[//]: # "SPDX-FileCopyrightText: Copyright (c) 2022-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved."
[//]: # "SPDX-License-Identifier: Apache-2.0"
[//]: # ""
[//]: # "Licensed under the …
-
初始化代码：
_mnn_net->setCacheFile(".cache");
_net_cfg.type = MNN_FORWARD_OPENCL;
DEBUG_LOG(" setCacheFile end ");
_mnn_session = _mnn_net->createSession(_net_cf…
-
**Describe the bug**
Our accuracy test of llama 3.1 8b on T3K has 0.0 accuracy when a tensor that is no longer used is deallocated. The proximate cause of the bad accuracy is the corruption of other t…
-
## Describe the bug
The distance between two of the same GMMs is not 0. Sometimes the distance could be as large as 1e-3 when I use my own data. Is this because of the numerical problem?
### To Re…
-
### Description
These two are doing way too much. We should probably split more along the lines of how we listed #821 if not more fine-grained.
* Elemwise (which DimShuffle is not!)
* Reduce: CAR…
-
At present,` ttnn.pow` supports Tensor `input` , scalar `exponent`.
For Pytorch tracing https://github.com/tenstorrent/tt-metal/issues/13373, we need support for ttnn.pow with Tensor `input` , Tenso…