Skip to content
Merged
Show file tree
Hide file tree
Changes from 4 commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 18 additions & 6 deletions components/src/dynamo/trtllm/engines/diffusion_engine.py
Original file line number Diff line number Diff line change
Expand Up @@ -145,7 +145,7 @@ def _build_diffusion_args(self) -> "VisualGenArgs":

Maps dynamo's DiffusionConfig fields to TensorRT-LLM's VisualGenArgs
structure with its nested sub-configs (PipelineConfig, TorchCompileConfig,
CudaGraphConfig, AttentionConfig, ParallelConfig, TeaCacheConfig,
CudaGraphConfig, AttentionConfig, ParallelConfig, optional cache config,
quant_config).

Returns:
Expand Down Expand Up @@ -199,14 +199,14 @@ def _build_diffusion_args(self) -> "VisualGenArgs":
dit_cfg_size=self.config.dit_cfg_size,
dit_fsdp_size=self.config.dit_fsdp_size,
),
teacache=TeaCacheConfig(
enable_teacache=self.config.enable_teacache,
use_ret_steps=self.config.teacache_use_ret_steps,
teacache_thresh=self.config.teacache_thresh,
),
)

# Add optional fields
if self.config.enable_teacache:
args_kwargs["cache"] = TeaCacheConfig(
use_ret_steps=self.config.teacache_use_ret_steps,
teacache_thresh=self.config.teacache_thresh,
)
if self.config.revision:
args_kwargs["revision"] = self.config.revision
if quant_config is not None:
Expand Down Expand Up @@ -281,6 +281,18 @@ def generate(
seed=seed if seed is not None else random.randint(0, 2**32 - 1),
)

# The executor normally fills None fields with pipeline defaults via
# _merge_defaults; we call infer() directly so we replicate that here.
for name, default in self._pipeline.default_generation_params.items():
if hasattr(req, name) and getattr(req, name) is None:
setattr(req, name, default)
specs = self._pipeline.extra_param_specs
if specs:
if req.extra_params is None:
req.extra_params = {}
for key, spec in specs.items():
req.extra_params.setdefault(key, spec.default)

# Run the pipeline — infer() wraps forward() with torch.no_grad()
output = self._pipeline.infer(req)

Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -519,6 +519,8 @@ def _make_engine(self):
engine = DiffusionEngine(config=config)
engine._initialized = True
engine._pipeline = MagicMock()
engine._pipeline.default_generation_params = {}
engine._pipeline.extra_param_specs = {}
engine._pipeline.infer.return_value = SimpleNamespace(
video=torch.zeros((1, 4, 64, 64, 3), dtype=torch.uint8),
image=None,
Expand All @@ -535,6 +537,7 @@ def test_generate_wraps_prompt_as_list(self):
class FakeDiffusionRequest:
def __init__(self, **kwargs):
captured.update(kwargs)
self.extra_params = kwargs.get("extra_params")
for k, v in kwargs.items():
setattr(self, k, v)

Expand Down
4 changes: 2 additions & 2 deletions container/context.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -105,10 +105,10 @@ trtllm:
python_version: "3.12"
index_url: https://pypi.nvidia.com/
pip_wheel_dir: /tmp/trtllm_wheel/
pip_wheel: tensorrt-llm==1.3.0rc11
pip_wheel: tensorrt-llm==1.3.0rc13
trtllm_wheel_image: nvcr.io/nvidia/tensorrt-llm/release:${TENSORRTLLM_PIP_WHEEL#*==}

github_trtllm_commit: v1.3.0rc11
github_trtllm_commit: v1.3.0rc13
torch_version: 2.11.0a0+eb65b36914.nv26.2
torch_tensorrt_version: 2.11.0a0
torchvision_version: 0.25.0a0+1e53952f.nv26.2.44259020
Expand Down
2 changes: 1 addition & 1 deletion container/deps/requirements.common.txt
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ tensorboard>=2.19.0,<2.21.0
tensorboardX==2.6.2.2
# Transformers version constraint for container builds
# - vLLM 0.11.0: >=4.55.2, vLLM 0.11.2: >=4.56.0,<5
# - TensorRT-LLM 1.3.0rc11: ==4.57.3
# - TensorRT-LLM 1.3.0rc13: ==4.57.3
# - SGLang 0.5.8: ==4.57.1
# Using >=4.56.0 to satisfy all frameworks
transformers>=4.56.0
Expand Down
2 changes: 1 addition & 1 deletion docs/reference/support-matrix.md
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,7 @@ The following table shows the backend framework versions included with each Dyna

| **Dynamo** | **SGLang** | **TensorRT-LLM** | **vLLM** | **NIXL** |
| :--- | :--- | :--- | :--- | :--- |
| **main (ToT)** | `0.5.10.post1` | `1.3.0rc11` | `0.20.0` | `0.10.1` (TRT-LLM, vLLM); `1.0.1` (SGLang) |
| **main (ToT)** | `0.5.10.post1` | `1.3.0rc13` | `0.20.0` | `0.10.1` (TRT-LLM, vLLM); `1.0.1` (SGLang) |
| **v1.2.0-deepseek-v4-dev.2** *(experimental, partial)* | upstream DSv4 preview | — | `0.20.0` | `0.10.1` |
| **v1.1.0** | `0.5.10.post1` | `1.3.0rc11` | `0.19.0` | `0.10.1` (TRT-LLM, vLLM); `1.0.1` (SGLang) |
| **v1.1.0-dev.3** *(experimental, partial)* | `0.5.10.post1` | `1.3.0rc11` | `0.19.0` | `0.10.1` |
Expand Down
Loading