chakra
chakra copied to clipboard
mlcommons
Repository for MLCommons Chakra schema and tools
it too tough to get start may i ask is there any tutorial or example for this project? for example, how can i get the pytorch et from cluster, and...
> Please provide a detailed description of your question or the information you seek. Hi, Could you please share more ET traces, such as the LLaMA traces you mentioned in...
## Describe the Bug When I run the pytorch converter, it shows nccl:send comm_type not supported, is there any plan to support this or this comm_type is not expected in...
## Problem Related to the Feature As defined in the [et_def.proto](https://github.com/mlcommons/chakra/blob/main/schema/protobuf/et_def.proto), the attribute that stores duration (duration_micros) uses microsecond precision. There are some cases where we encounter lots of sub-microsecond-runtime...
There is too much information from pytorch ET plus, including function call relationships with too many details, but in fact we only focus on computing, memory access and communication. Does...
## Summary Add Integration Testing for chakra_trace_link and chakra_converter ## Test Plan GitHub actions work. ``` $ python3 ci_tools/integration_tests.py --tgz_path tests/data/1.0.2-chakra.0.0.4/llama_pytorch24.05.tgz --num_ranks 8 --tolerance 0.05 --expected _times_ms 14597 14597 14968...
## Summary * This PR relies on https://github.com/mlcommons/chakra/pull/119. It can be merged after https://github.com/mlcommons/chakra/pull/119 is merged. This PR introduces dependencies from GPU operators to CPU operators using the critical path...
## Summary Previously, the data_deps field was utilized to encode operator dependencies in simulations. However, data_deps should actually be reserved for encoding data dependencies, not for simulating operator dependencies. Therefore,...
## Summary This PR adds the Any type to the AttributeProto message as a single any_val and repeated any_list. The Any type in Any.proto is a standard Protobuf definition included...
