Modify tacticSources
See original GitHub issueWith trtexec I need to add --tacticSources=-cublasLt,+cublas
to compile my model, or else I get:
[TensorRT] INTERNAL ERROR: Assertion failed: cublasStatus == CUBLAS_STATUS_SUCCESS
How can I do this in torch2trt?
Issue Analytics
- State:
- Created 3 years ago
- Reactions:1
- Comments:5 (1 by maintainers)
Top Results From Across the Web
IBuilderConfig — NVIDIA TensorRT Standard Python API ...
Tactic sources that can provide tactics for TensorRT. ... library tactics without having to modify their application's plugins to support nullptr handles.
Read more >TacticSource Pricing, Alternatives & More 2022 - Capterra
With the help of Capterra, learn about TacticSource, its features, pricing information, popular comparisons to other Procurement products and more.
Read more >Top 10 TacticSource Alternatives 2022 - G2
Find the top-ranking alternatives to TacticSource based on 1750 verified user reviews. Read reviews and product information about SAP Ariba, RFP360, ...
Read more >TensorRT: nvinfer1::IBuilderConfig Class Reference
Returns true if the quantization flag is set. More... virtual bool, setTacticSources (TacticSources tacticSources)=0. Set tactic sources. More...
Read more >TacticSource e-Tender System Supplier User Guide v2.3
The setting will not affect your bidding currency in each RFX which is preset by the buyer. Administration. You can change your company...
Read more >Top Related Medium Post
No results found
Top Related StackOverflow Question
No results found
Troubleshoot Live Code
Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start FreeTop Related Reddit Thread
No results found
Top Related Hackernoon Post
No results found
Top Related Tweet
No results found
Top Related Dev.to Post
No results found
Top Related Hashnode Post
No results found
Top GitHub Comments
Hi @austinmw ,
Thanks for reaching out!
Currently this isn’t exposed, but it seems like a relatively straightforward feature for us to add. Particularly, after the following PR for DLA support is merged, torch2trt will use a builder config, so I can just pass an extra argument.
https://github.com/NVIDIA-AI-IOT/torch2trt/pull/507
I will try to notify you here of any updates, or if I discover anything along way.
Best, John