| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2026-06-01 | Update to 1.2.1-1 | Smoolak | |
| 2026-05-24 | Fix source checksum (previous was computed from a truncated download) | Smoolak | |
| 2026-05-23 | Update to 1.2.1 | Smoolak | |
| 2025-12-08 | Add python-cutlass makedepend to fix CUTLASS linker errors | Prof. Andy Couturier | |
| 2025-12-06 | Add rdma-core dependency for full build support | Prof. Andy Couturier | |
| - Add rdma-core to depends (provides mlx5 library) - Remove cmake options that disabled BUILD_DEEP_EP, BUILD_DEEP_GEMM, BUILD_FLASH_MLA - Enable full distributed training and advanced optimization features - Bump pkgrel to 2 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> | |||
| 2025-12-06 | Initial commit: python-tensorrt-llm 1.2.0rc4 | Prof. Andy Couturier | |
| - TensorRT-LLM Python API for LLM inference on NVIDIA GPUs - Includes glog 0.7.0+ compatibility patch - Requires CUDA, cuDNN, TensorRT, and PyTorch | |||