Release v0.4.1

What's Changed

• fix: fix the failed sampling unittest on 5090 by @yzh119 in #1886
• Updated to latest docker tag by @nvmbreughe in #1889
• Fix: Prevent race condition in cubin loader when file is being consumed by @yzh119 in #1852
• Improve graph caching of cudnn graph by @Anerudhan in #1887
• misc: Various Updates to Attention Microbenchmark Suite by @bkryu in #1891
• docs: Fix installation instructions for CUDA-specific package URLs by @yzh119 in #1893
• docker image improvements by @nvmbreughe in #1890
• tests: Add batch size 1 cases to test_trtllm_gen_attention.py that fail, marked xfail by @bkryu in #1897
• Ensure docker installs the torch version we need by @nvmbreughe in #1901
• bugfix: exclude tests/utils/test_load_cubin_compile_race_condition.py from pytest by @yzh119 in #1907
• ci: use self-hosted runner for building docker containers by @yzh119 in #1908
• feat: Add FP4 TRTLLM-Gen throughput MOE batched gemms by @jiahanc in #1882
• Update Docker CI tags to 20251010-8d072e6 by @github-actions[bot] in #1915
• ci/cd: consolidate release workflow by @yzh119 in #1910
• bugfix: fix cli error when cuda toolkit is not installed by @yzh119 in #1905
• feat: trtrllm-gen global scaled FP8 GEMMs by @hypdeb in #1829
• feat:enable fp8 blockscale moe for fused cultass for sm90 by @djmmoss in #1819
• use ffi::TensorView instead of ffi::Tensor by @cyx-6 in #1844
• Minor updates to cubin_loader.py download_file to avoid race condition on temporary file by @nvjullin in #1918
• chore: make cache directory flashinfer-version specific by @yzh119 in #1920
• misc: checksum check when downloading artifacts by @jimmyzho in #1761
• release: bump version v0.4.1 by @yzh119 in #1921

New Contributors

• @jiahanc made their first contribution in #1882

Full Changelog: v0.4.0...v0.4.1