Tianqi Chen
Tianqi Chen
Likely have to do with the llvm shipped do not come with wasm, you can try to swap out the LLVM and build from source in a branch say here...
@DustinBrett Thank you for explorations. We are an open source project so you are definitely more than welcomed to contributed and we really appreciate the discussion and contribution. Likely reusing...
latest chrome should work
Thank you for discussions and interest. We are in the process of bringing in new models, and with them also have open recipes in the PRs so people can try...
The latest webllm js api contains a interrupt function that can support this
Interesting, this definitely seems to relates to issues in webgpu. Thank you sending the console logs, although based on the current message it is a bit harder to see what...
Thank you for getting to the bottom of it! Indeed memory is a place we should work harder on, and hopefully we will look into some of that after fp16...
Agree that it would be good to unit-test type shape inference of each operator. One of the main catch here that normalization can be context dependent, where the context (surrounding...
The new op impls and other impls now introduces unit tests for normalizers at operator and unit level See examples in https://github.com/tlc-pack/relax/blob/relax/tests/python/relax/test_op_binary.py#L61 https://github.com/tlc-pack/relax/blob/relax/tests/python/relax/test_analysis_struct_info_analysis.py#L341
That makes sense.