Tianqi Chen

Results 36 issues of Tianqi Chen

## Background One of the primary goal of WASM is to run on the web, and web is async in nature. On the other hand, there is always a need...

discussion

As we start to formalize MLC LLM Engine, we are moving towards a more comprehensive API that is OpenAI compatible. This means a lot of new features that allows us...

status: tracking

## 🚀 Feature As of now OpenAI API protocol set the direct default values. From the endpt pov, it is better to set most values to be NOT_GIVEN, as per...

feature request

Hi WebGPU community: I would like to use this issue to call for a discussion about what is the best way to compile compute shader kernels with non-buffer arguments. Compute...

question
wgsl
api

## Overview This is a global tracking issue to bring generic sentence embedding models to MLCEngine. ## Action Items - [ ] Add support for mistral based sentence embedding ##...

status: tracking

## Overview It is useful to have an CPU/NPU compatible PagedKVCache for general compact purposes ## Action Items - [ ] Provide a CPU compatible PagedKVCache kernels ## Links to...

status: tracking

Seems there are some flaky test cases in hexagon during recent PRs that do not have to do with the feature ### Test(s) - `tests/python/contrib/test_hexagon/test_relax_2d_buffer_allocation.py::test_alloc_storage_with_scope_global` ### Jenkins Links - https://ci.tlcpack.ai/blue/organizations/jenkins/tvm-hexagon/detail/PR-16807/5/pipeline...

test: flaky
needs-triage

This PR brings up new reflection support in python. The new reflection now directly attaches property and methods to the class object themselves, making more efficient accessing than old mechanism....

Since the term synchronize may cause confusion that user thought it means stream sync, it simply means we pass the right current stream as env stream

This PR updates to include dlpack C exchange api for fast exchange at c-extension level without going through python.