-
## 🐛 Bug
## To Reproduce
Steps to reproduce the behavior:
1.
1.
1.
## Expected behavior
## Environment
- Platform (e.g. WebGPU/Vulkan/IOS/Android/CUDA):
- Operating …
-
Currently, we use the Corstone300 simulator to test our microTVM schedules. The simulator is configured to model the Cortex-M55. However, we do not do any fault handling, so when an exception (e.g. `U…
-
## ❓ General Questions
I want to modify the tvm source codes, so I can build it from source in my production environment. But I can't find the 3rdparty/tvm in the mlc_llm, either of mlc_llm/relax…
DeclK updated
1 month ago
-
Please refer to the following repository to settle this question
https://github.com/ak47werttyydd/memcov
The basic train of thought is:
1. save coverage information to a local file at the end o…
-
## 🐛 Bug
When serving a model through the REST API on an 8xA6000 machine I get this error: The block is 1-time referenced by other blocks, thus cannot accept new KV values.
I've added the releva…
-
The issue origin is the problem described in the [PR](https://github.com/apache/tvm/pull/13596). During solution it became clear that topi and relay depend on each other. I've gotten the name of the i…
-
## ❓ General Questions
Hi, I wonder if there is a way to compute ppl of llms in MLC-LLM. Thanks for any help.
-
In the below AWS table , the column "address" is masked by AWS Lake Formation policy. Then the select query from presto is hiding the column "address" but values are seen under next column "city"
pre…
-
## 🐛 Bug
There seems to be a type error when tracing a Graph-UNet with Torch JIT.
```
[omitted]/torch_sparse/matmul.py in spspmm(src, other, reduce)
94 if reduce == 'sum' or reduce == '…
-
With #26 fixed on my side, I was able to perform some benchmarks now. The libDNN generated convolutions are about 3x-4x faster than my naive kernel described in #26, which is very nice! But they are s…
romix updated
6 years ago