Netanel Haber 6ee94c7ac8 Reintroduce with perf fixes: feature: unify new_tokens format sample state to trtllm samper tokens format (#5513) ... 58a8a8f - these changes were previously merged to main here. 6aef149 - the changes were temporarily reverted in main, due to a significant perf regression in models using the TorchSampler (observed by @byshiue). This PR is meant to re-merge these changes along with a fix to prevent the regression. The first commit of this PR is actually just the reverted revert - filter it out of the changes to see previously unmerged changes. Signed-off-by: Netanel Haber <nhaber@nvidia.com>		2025-06-30 11:58:59 -07:00
..
attention_backend	feat: Add support for YARN in NemotronNAS models (#4906)	2025-06-29 09:45:49 +03:00
auto_deploy	Reintroduce with perf fixes: feature: unify new_tokens format sample state to trtllm samper tokens format (#5513)	2025-06-30 11:58:59 -07:00
compilation	[feat] Piecewise cuda graph support for MLA (#4467)	2025-06-17 18:58:38 +08:00
custom_ops	Make moe permute and final as custom op (#5412)	2025-06-27 15:48:33 -07:00
debug	Add debug hook to support dump tensor data and add new debug functions easily (#5182)	2025-06-24 17:45:28 +08:00
distributed	Feat/ds r1 min latency opt round3, add router gemm, fused a gemm, PDL (#4560)	2025-06-14 17:36:22 +08:00
models	feat : support duplicate_kv_weight for qwen3 blockwise scale (#5459)	2025-06-30 11:49:22 +08:00
modules	refactor: [TRTLLM-6150] Refactor moe permute and finalize op by removing duplicated code (#5557)	2025-06-30 08:48:04 -07:00
peft	feat: support multi lora adapters and TP (#3885)	2025-05-08 23:45:45 +08:00
pyexecutor	Reintroduce with perf fixes: feature: unify new_tokens format sample state to trtllm samper tokens format (#5513)	2025-06-30 11:58:59 -07:00
speculative	Reintroduce with perf fixes: feature: unify new_tokens format sample state to trtllm samper tokens format (#5513)	2025-06-30 11:58:59 -07:00
__init__.py	[TRTLLM-5208][BREAKING CHANGE] chore: make pytorch LLM the default (#5312)	2025-06-20 03:01:10 +08:00
autotuner.py	[TRTLLM-5770] feat: Integrate TRT-LLM Gen FP8 block scale MoE with Pytorch workflow kernel autotuner (#5207)	2025-06-17 21:01:56 +08:00
expert_statistic.py	Add MTP support for Online EPLB (#5213)	2025-06-25 07:58:13 +08:00
llm.py	[TRTLLM-5208][BREAKING CHANGE] chore: make pytorch LLM the default (#5312)	2025-06-20 03:01:10 +08:00
metadata.py	feat: no-cache attention in PyTorch workflow (#3085)	2025-04-05 01:54:32 +08:00
model_config.py	[TRTLLM-5825][fix] Fix torch LoRA TP (#5338)	2025-06-19 09:12:00 +03:00
utils.py	perf: Avoid reswizzle_sf after allgather. (#5504)	2025-06-29 21:25:50 +08:00