[TensorRT EP] Load precompiled TRT engine file directly #18217
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
There was a problem hiding this comment.
lintrunner found more than 10 potential problems in the proposed changes. Check the Files changed tab for more details.
|
@jywu-msft Since |
plan is to merge this by end of week. @chilo-ms will work on it. |
…FromPrecompiledEngine in the file for better solving merge conflict
jywu-msft
reviewed
Jan 9, 2024
|
Github isn't smart enough to make this PR smoothly merge main. So, i have to resolve a lot of conflicts ... The key change regarding the code path inside |
gedoensmax
reviewed
Jan 10, 2024
jywu-msft
reviewed
Jan 11, 2024
jywu-msft
reviewed
Jan 11, 2024
jywu-msft
approved these changes
Jan 12, 2024
thanks. these suggestions will be added in a follow-up PR. we will merge this one asap as soon as CI's pass. |
mszhanyi
pushed a commit
that referenced
this pull request
Jan 15, 2024
When the TRT engine cache (precompiled engine) is present, it doesn't make sense to go over the processes of model verification, model optimization, TRT EP's GetCapability(), TRT EP's model proto reconstruction, calling TRT parser and engine compilation. This PR makes TRT EP skip those processes and directly load the engine to perform inference. The feature request: #18072 Features: - Replace original model with TRT engine wrapped ONNX model. It can save a lot of time as mentioned above. - How to get TRT engine wrapped ONNX model? 1. Set `trt_dump_ep_context_model` provider option to "true" and run the inference. You will find the "xxx_wrapper.onnx" at the engine cache path. (The same logic of generating engine cache) 2. Use gen_trt_engine_wrapper_onnx_model.py - Three provider options are added, `trt_dump_ep_context_model`: Enable dump wrapped onnx model by TRT EP `trt_ep_context_embed_mode`: Add embed_mode as attribute. 0 means engine cache path, 1 means engine binary data. `trt_ep_context_compute_capability_enable`: Add hardware_arch as attribute. When running the model, TRT EP will check consistency between model's hardware_arch and GPU's compute capability. - When the engine cache path is given in the wrapped model, TRT EP will first search for the engine file using the path (relative to model path), if it can't find it, it will change to use the path as it is (depends on user, could be relative to working dir or absolute path) Note: 1. This PR includes the change of #17751 Constraints: 1. The whole model should be fully supported by TRT. 4. Users need to make sure the engine is built with min/max/opt optimization profiles that large enough to cover the range of all inputs. TRT EP will simply fail and won't rebuild the engine if the input shape is out of range during runtime.
Please see the PR here for naming the "EP context" model. Feel free to provide feedbacks. |
jywu-msft
pushed a commit
that referenced
this pull request
Jan 21, 2024
…der options (#19154) Several changes: 1. To align with other EPs' setting of EP context configs in session options, for example [QNN EP](#18877), EP context configs for TRT EP can be configured through: 1. Session Options: `ep.context_enable`, `ep.context_file_path` and `ep.context_embed_mode` 2. Provider Options: `trt_dump_ep_context_model`, `trt_ep_context_file_path` and `trt_dump_ep_context_embed_mode` 3. Above setting has 1:1 mapping and provider options has higher priority over session options. ``` Please note that there are rules for using following context model related provider options: 1. In the case of dumping the context model and loading the context model, for security reason, TRT EP doesn't allow the "ep_cache_context" node attribute of EP context node to be the absolute path or relative path that is outside of context model directory. It means engine cache needs to be in the same directory or sub-directory of context model. 2. In the case of dumping the context model, the engine cache path will be changed to the relative path of context model directory. For example: If "trt_dump_ep_context_model" is enabled and "trt_engine_cache_enable" is enabled, if "trt_ep_context_file_path" is "./context_model_dir", - if "trt_engine_cache_path" is "" -> the engine cache will be saved to "./context_model_dir" - if "trt_engine_cache_path" is "engine_dir" -> the engine cache will be saved to "./context_model_dir/engine_dir" ``` 2. User can decide the naming of the dumped "EP context" model by using `trt_ep_context_file_path`, please see GetCtxModelPath() for more details. 3. Added suggested comments from #18217
YUNQIUGUO
pushed a commit
that referenced
this pull request
Jan 23, 2024
…der options (#19154) Several changes: 1. To align with other EPs' setting of EP context configs in session options, for example [QNN EP](#18877), EP context configs for TRT EP can be configured through: 1. Session Options: `ep.context_enable`, `ep.context_file_path` and `ep.context_embed_mode` 2. Provider Options: `trt_dump_ep_context_model`, `trt_ep_context_file_path` and `trt_dump_ep_context_embed_mode` 3. Above setting has 1:1 mapping and provider options has higher priority over session options. ``` Please note that there are rules for using following context model related provider options: 1. In the case of dumping the context model and loading the context model, for security reason, TRT EP doesn't allow the "ep_cache_context" node attribute of EP context node to be the absolute path or relative path that is outside of context model directory. It means engine cache needs to be in the same directory or sub-directory of context model. 2. In the case of dumping the context model, the engine cache path will be changed to the relative path of context model directory. For example: If "trt_dump_ep_context_model" is enabled and "trt_engine_cache_enable" is enabled, if "trt_ep_context_file_path" is "./context_model_dir", - if "trt_engine_cache_path" is "" -> the engine cache will be saved to "./context_model_dir" - if "trt_engine_cache_path" is "engine_dir" -> the engine cache will be saved to "./context_model_dir/engine_dir" ``` 2. User can decide the naming of the dumped "EP context" model by using `trt_ep_context_file_path`, please see GetCtxModelPath() for more details. 3. Added suggested comments from #18217
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When the TRT engine cache (precompiled engine) is present, it doesn't make sense to go over the processes of model verification, model optimization, TRT EP's GetCapability(), TRT EP's model proto reconstruction, calling TRT parser and engine compilation.
This PR makes TRT EP skip those processes and directly load the engine to perform inference.
The feature request: #18072
Features:
Replace original model with TRT engine wrapped ONNX model. It can save a lot of time as mentioned above.
How to get TRT engine wrapped ONNX model?
trt_dump_ep_context_modelprovider option to "true" and run the inference. You will find the "xxx_wrapper.onnx" at the engine cache path. (The same logic of generating engine cache)Three provider options are added,
trt_dump_ep_context_model: Enable dump wrapped onnx model by TRT EPtrt_ep_context_embed_mode: Add embed_mode as attribute. 0 means engine cache path, 1 means engine binary data.trt_ep_context_compute_capability_enable: Add hardware_arch as attribute. When running the model, TRT EP will check consistency between model's hardware_arch and GPU's compute capability.trt_ep_context_file_path: Please see this PR.When the engine cache path is given in the wrapped model, TRT EP will first search for the engine file using the path (relative to model path), if it can't find it, it will change to use the path as it is (depends on user, could be relative to working dir or absolute path)
Note:
Constraints: