- San Francisco
Pinned Loading
-
sgl-project/sglang
sgl-project/sglang PublicSGLang is a fast serving framework for large language models and vision language models.
-
vllm-project/vllm
vllm-project/vllm PublicA high-throughput and memory-efficient inference and serving engine for LLMs
-
microxcaling
microxcaling PublicForked from microsoft/microxcaling
PyTorch emulation library for Microscaling (MX)-compatible data formats
Python
-
pytorch/FBGEMM
pytorch/FBGEMM PublicFB (Facebook) + GEMM (General Matrix-Matrix Multiplication) - https://code.fb.com/ml-applications/fbgemm/
-
ROCm/pytorch
ROCm/pytorch PublicForked from pytorch/pytorch
Tensors and Dynamic neural networks in Python with strong GPU acceleration
-
cs231x/super-resolution-detection
cs231x/super-resolution-detection PublicEnd-to-End Super Resolution Object Detection Networks
240 contributions in the last year
Day of Week | March Mar | April Apr | May May | June Jun | July Jul | August Aug | September Sep | October Oct | November Nov | December Dec | January Jan | February Feb | March Mar | ||||||||||||||||||||||||||||||||||||||||
Sunday Sun | |||||||||||||||||||||||||||||||||||||||||||||||||||||
Monday Mon | |||||||||||||||||||||||||||||||||||||||||||||||||||||
Tuesday Tue | |||||||||||||||||||||||||||||||||||||||||||||||||||||
Wednesday Wed | |||||||||||||||||||||||||||||||||||||||||||||||||||||
Thursday Thu | |||||||||||||||||||||||||||||||||||||||||||||||||||||
Friday Fri | |||||||||||||||||||||||||||||||||||||||||||||||||||||
Saturday Sat |
Activity overview
Contribution activity
March 2025
Created 4 commits in 1 repository
Created a pull request in sgl-project/sglang that received 6 comments
ROCm: Flex Attention Enablement with custom backends
Credits: @poyenc , @amd-hhashemi , @linsun12 , @tenpercent , @carlushuang , @HaiShaw Motivation Add ROCm custom attention backends to enable Flex …
Opened 4 other pull requests in 1 repository
sgl-project/sglang
1
open
3
merged
-
[ROCm/Draft/No-Merge]: Flex Attention Enablement
This contribution was made on Mar 7
-
ROCm: enable trillion-parameter MoE models with INT4-FP8 single node
This contribution was made on Mar 6
-
Fix the moe padding conditional logic
This contribution was made on Mar 5
-
ROCm: update aiter and its usage to fused moe (bloat16, fp8, fp8 block-quant)
This contribution was made on Mar 4
Reviewed 10 pull requests in 1 repository
sgl-project/sglang
10 pull requests
-
Update amd ci docker image to v0.4.3.post4-rocm630.
This contribution was made on Mar 7
-
ROCm: enable trillion-parameter MoE models with INT4-FP8 single node
This contribution was made on Mar 6
-
AMD/ROCm: update base image string
This contribution was made on Mar 6
-
Add tag suffix to nightly docker builds.
This contribution was made on Mar 6
-
Create release-docker-amd-nightly.yml
This contribution was made on Mar 5
-
ROCM: AITER BLOCK GEMM
This contribution was made on Mar 5
-
Fix breakage problem when using custom_ar
This contribution was made on Mar 4
-
Fix assert options.num_stages != 0 error in the latest ROCm build image
This contribution was made on Mar 4
-
ROCM support tree_speculative_sampling_target_only
This contribution was made on Mar 3
-
Enable custom AR for AMD GPUs and maintain it in sgl-kernel
This contribution was made on Mar 1
Created an issue in sgl-project/sglang that received 1 comment
[Feature] Add e4m3fnuz support to MoE-EP in FP8
Checklist 1. If the issue you raised is not a feature but a question, please raise a discussion at https://github.com/sgl-project/sglang/discussi…