kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-01-14 06:27:45 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity
92576488d3
TensorRT-LLMs/tensorrt_llm
History
William Zhang 92576488d3
[None][feat] Skip prefetching consolidated safetensors when appropriate (#7013) 
* Why?

Some models (e.g. anything produced by Mistral) can have both sharded
safetensors and a consolidated safetensor in the same checkpoint
directory. In such cases, prefetching both to memory is a waste of time,
and memory.

* What?

This commit skips over consolidated safetensors when they are not the
only safetensor file present in the checkpoint directory

Signed-off-by: William Zhang <133824995+2ez4bz@users.noreply.github.com>
2025-08-25 23:56:21 -04:00
..
_tensorrt_engine [TRTLLM-5208][BREAKING CHANGE] chore: make pytorch LLM the default (#5312) 2025-06-20 03:01:10 +08:00
_torch [None][feat] Skip prefetching consolidated safetensors when appropriate (#7013) 2025-08-25 23:56:21 -04:00
auto_parallel [None][fix] Migrate to new cuda binding package name (#6700) 2025-08-07 16:29:55 -04:00
bench [None][fix] Update to pull LLM from a central location. (#6458) 2025-08-25 13:07:29 -07:00
commands [None][chore] remove CLI support for mamba cache dtype setting (#7119) 2025-08-25 08:08:51 -04:00
evaluate [TRTLLM-6771][feat] Support MMMU for multimodal models (#6828) 2025-08-21 08:54:12 +08:00
executor [None][chore] Create PyExecutor from TorchLlmArgs Part 1 (#7105) 2025-08-26 10:42:01 +08:00
inputs [None][fix] Fix mm_placholder_counts extraction issue. (#7118) 2025-08-22 12:28:30 +08:00
layers [TRTLLM-5863][feat] Support MoE INT8 Weight-Only-Quantization in PyTorch Workflow (#6629) 2025-08-15 17:15:49 -04:00
llmapi [None][chore] Create PyExecutor from TorchLlmArgs Part 1 (#7105) 2025-08-26 10:42:01 +08:00
metrics [None][feat] Core Metrics Implementation (#5785) 2025-08-09 02:48:53 -04:00
models [None][fix] Refactoring to avoid circular import when importing torch models (#6720) 2025-08-11 18:00:42 -04:00
plugin feat: Add support for fp8 rowwise quantization (#4876) 2025-06-14 06:37:48 -07:00
quantization [None][fix] Pre-allocate workspaces for DeepGEMM MoE to avoid frequent cudaFree/cudaMalloc (#6811) 2025-08-13 10:27:57 +08:00
runtime [None][fix] Migrate to new cuda binding package name (#6700) 2025-08-07 16:29:55 -04:00
scaffolding [https://nvbugs/5387375] fix(scaffolding): fix scaffolding aime test in test_e2e (#6140) 2025-07-18 10:34:37 +08:00
serve [TRTLLM-6771][feat] Support MMMU for multimodal models (#6828) 2025-08-21 08:54:12 +08:00
tools [None] [feat] nsys profile output kernel classifier (#7020) 2025-08-23 00:57:37 -04:00
__init__.py [TRTLLM-7326][feat] Add standalone multimodal encoder (#6743) 2025-08-19 21:42:50 -07:00
_common.py linting(python): Enable ruff on more files (wave 1/N) (#5140) 2025-06-14 19:19:34 +08:00
_dlpack_utils.py linting(python): Enable ruff on more files (wave 1/N) (#5140) 2025-06-14 19:19:34 +08:00
_ipc_utils.py [None][fix] Migrate to new cuda binding package name (#6700) 2025-08-07 16:29:55 -04:00
_mnnvl_utils.py [TRTLLM-6743][feat] Optimize and refactor alltoall in WideEP (#6973) 2025-08-24 08:15:29 -04:00
_utils.py [None][feat] Core Metrics Implementation (#5785) 2025-08-09 02:48:53 -04:00
builder.py [None][fix] Refactoring to avoid circular import when importing torch models (#6720) 2025-08-11 18:00:42 -04:00
disaggregated_params.py [None][fix] Refactoring to avoid circular import when importing torch models (#6720) 2025-08-11 18:00:42 -04:00
functional.py [None] [feat] Add Tencent HunYuanMoEV1 model support (#5521) 2025-08-15 06:56:44 +08:00
graph_rewriting.py linting(python): Enable ruff on more files (wave 1/N) (#5140) 2025-06-14 19:19:34 +08:00
logger.py linting(python): Enable ruff on more files (wave 1/N) (#5140) 2025-06-14 19:19:34 +08:00
lora_helper.py [TRTLLM-6825][fix] Update lora for phi4-mm (#6817) 2025-08-21 22:00:04 -04:00
lora_manager.py [https://nvbugs/5467232][fix] Fix load_torch_hf_lora to override lora_config.trtllm_modules_to_hf_modules with default only when it has no value (#7132) 2025-08-24 15:00:24 +03:00
mapping.py [TRTLLM-5966][feat] Helix: extend mapping to support different CP types (#6816) 2025-08-14 09:00:02 -07:00
math_utils.py perf: Optimize swizzle_sf, unswizzle_sf, reswizzle_sf (#5318) 2025-06-26 14:03:56 +08:00
module.py linting(python): Enable ruff on more files (wave 1/N) (#5140) 2025-06-14 19:19:34 +08:00
network.py chore: remove usernames from comments (#3291) 2025-04-05 13:44:28 +08:00
parameter.py fix:https://nvbugs/5234033 enable starcoder trt-flow with transforme… (#3909) 2025-05-15 11:16:45 +08:00
profiler.py linting(python): Enable ruff on more files (wave 1/N) (#5140) 2025-06-14 19:19:34 +08:00
prompt_adapter_manager.py linting(python): Enable ruff on more files (wave 1/N) (#5140) 2025-06-14 19:19:34 +08:00
python_plugin.py linting(python): Enable ruff on more files (wave 1/N) (#5140) 2025-06-14 19:19:34 +08:00
sampling_params.py [None] [feat] Add model gpt-oss (#6645) 2025-08-07 03:04:18 -04:00
scheduling_params.py [None][feat] Add support of scheduling attention dp request (#6246) 2025-08-01 20:38:01 -04:00
serialization.py [None] [feat] Add model gpt-oss (#6645) 2025-08-07 03:04:18 -04:00
top_model_mixin.py [None][fix] Refactoring to avoid circular import when importing torch models (#6720) 2025-08-11 18:00:42 -04:00
version.py [None][chore] Bump version to 1.1.0rc2 (#7167) 2025-08-22 22:02:28 +08:00