Jacinto 7 TIDL Release Notes
Version: 08.02.00, Date: Mar 04, 2022
Contents
- Introduction
- Licensing
- What's New
- Documentation
- Upgrade and Compatibility Information
- Device Support
- Validation Information
- Fixed Issues
- Known Issues
- Change Request
- Technical Support
- Package Versioning
Introduction
This package contains Texas Instrument's Deep learning solution for J721E SoCs from the Jacinto 7 family
Licensing
The licensing information of this product, as well as any third-party components included which are made available under a number of other open-source licenses are enumerated as part of the manifest. Note : Below Links would work only after installing the package
Documentation
1 | User's Guide | [HTML] |
2 | Software Manifest | [HTML] |
What's New
Here are a few of the new features supported in this release:
- Performance improvement for NEO-DLR Runtime
- Runtime performance improvement for human pose estimation models
- Support for auto C7x code generation with TVM
- Removal of windows support
Upgrade and Compatibility Information
TIDL core library location has changed in this release. Please use below table to understand the change :Library Name | 08.01 release (Target Library) | 08.02 release (Target Library) | 08.01 release (PC Library) | 08.02 release (PC Library) |
tidl_algo.lib | ti_dl/lib/dsp/algo/{TARGET_BUILD} | ti_dl/lib/j721e/dsp/algo/{TARGET_BUILD} | ti_dl/lib/PC/dsp/algo/{TARGET_BUILD} | ti_dl/lib/j721e/PC/algo/{TARGET_BUILD} |
tidl_obj_algo.lib | ti_dl/lib/dsp/algo/{TARGET_BUILD} | ti_dl/lib/j721e/dsp/algo/{TARGET_BUILD} | ti_dl/lib/PC/dsp/algo/{TARGET_BUILD} | ti_dl/lib/j721e/PC/algo/{TARGET_BUILD} |
tidl_priv_algo.lib | ti_dl/lib/dsp/algo/{TARGET_BUILD} | ti_dl/lib/j721e/dsp/algo/{TARGET_BUILD} | ti_dl/lib/PC/dsp/algo/{TARGET_BUILD} | ti_dl/lib/j721e/PC/algo/{TARGET_BUILD} |
tidl_custom.lib | ti_dl/lib/dsp/algo/{TARGET_BUILD} | ti_dl/lib/j721e/dsp/algo/{TARGET_BUILD} | ti_dl/lib/PC/dsp/algo/{TARGET_BUILD} | ti_dl/lib/j721e/PC/algo/{TARGET_BUILD} |
Device Support
Target SoC | Target Plaform | Build environment (OS) | Target environment (OS) |
J721E EVM | x86_64 Linux PC | x86_64 (Linux) | NA |
Validation Information
This release was built and validated using the following tools:
Dependencies (included in Processor SDK RTOS):
- ivision - 01_00
- MMALIB - 02_03_00_04
- PDK - CORE SDK 08.02.00
- C7x CGT - ti-cgt-c7000_2.1.1.LTS
Dependencies (NOT included in Processor SDK RTOS):
- GCC version 5.4.0 (In Ubuntu 18.04)
- OpenCV: 4.1.0 (for Test bench Only)
- Protobuf: 3.11.3 (for Import tool only)
- Flatbuffers : 1.12.0 (for Import tool only)
Fixed Issues
ID | Description | Module | Affected Versions | Affected Platforms |
---|---|---|---|---|
[TIDL-1557] | Sigmoid layer accuracy is not as per expectation | COMPUTE | 08.01.00,08.00.01, 08.00.00, 02.00.00 | J721E |
[TIDL-1558] | SSD prior box layer doesn't consume previous layer output dimension for box decoding | COMPUTE | 08.01.00,08.00.01, 08.00.00, 02.00.00 | J721E |
[TIDL-1771] | Networks with multiple inputs may give wrong output if the output order changes during import | TOOL | 08.00.00 | J721E |
[TIDL-1794] | Network output is missing from TIDL imported network if the same is consumed by other layers | TOOL | 08.01.00,08.00.00 | J721E |
[TIDL-1833] | Detection output layer is not working in float32 host emulation mode | COMPUTE | 08.01.00,08.00.01 | J721E |
[TIDL-1855] | TIDL Batch processing layer output is not correct on EVM when its output is in DDR | COMPUTE | 08.01.00,08.00.01 | J721E |
[TIDL-1871] | TIDL rawDataInElementType inconsistency with documentation | DOCUMENTATION | 08.01.00 | J721E |
[TIDL-1877] | Average pooling layer (3 x 3 stride 2) with padding 0 results in incorrect output for target (EVM) execution | COMPUTE | 08.01.00 | J721E |
Known Issues
ID | Description | Module | Reported in Release | Affected Platforms | Occurrence | Workaround in this release |
---|---|---|---|---|---|---|
TIDL-864 | Concat trace dump mismatch between REF & CN/CI | COMPUTE | 01.02.00 | J721E | Rare | NA |
TIDL-1197 | Model execution order for a branch is not optimal | COMPUTE | 01.02.00 | J721E | Rare | No fucntional issue, Only Performance drop would be obsrved |
TIDL-1350 | ShuffleNetV2 ONNX-RT accuracy is 2% below compared to TVM-RT | COMPUTE | 02.00.00 | J721E | Rare | NA |
TIDL-1530 | Processing on EVM gets stuck in case of "GEMM" layer in onnx model | OSRT | 02.00.00 | J721E | Rare | set reservedCtrl=1 of TIDL_CreateParams |
For latest status of the issues and any new issues found post this release, please refer TI External software Incident Report. Issues on this product can be filtered with search query as project = EXT_EP AND issuetype in (Bug, Enhancement) AND Product ~ "TI Deep Learning Library"
Change Request
ID | Description | Original Fix Version | New Fix Version |
---|---|---|---|
JACINTOREQ-2432 | Additional Performance Optimization for convolution layers | 08.02.00 | 08.04.00 |
JACINTOREQ-2433 | Power estimation tool support in TI cloud is descoped | 08.02.00 | DESCOPE |
JACINTOREQ-2434 | YUV422 input support in TIDL is descoped | 08.02.00 | DESCOPE |
JACINTOREQ-2435 | Additional Performance Optimization for non-convolution layers | 08.02.00 | 09.00.00 |
Technical Support
For technical support, please post your questions on TI E2E Forum for Processors.
For additional assistance, contact local TI Field Application Engineer
Package Versioning
Each package version is composed of 4 period-delimited numbers - represented here by the letters M, m, p and b [M.m.p.b]
. The table below provides a descriptive reference regarding package version numbering.
Digit | Meaning | Description |
---|---|---|
1 (M=Major) | Major revision | Incremented when the new version is substantially different from the previous For example, a new module added or an existing module's algorithm significantly altered. |
2 (m=minor) | Minor revision | Incremented when the new version has changed but not in a major way. For example, some minor changes in the API or feature set. |
3 (p=patch) | Patch number | Incremented for all other source code changes. This include any packaging support code. |
4 (b=build) | Build number | Incremented for each release delivery to CM. Reset for any change to M, m or p |
© Copyright 1995-2021, Texas Instruments Incorporated. All rights reserved.
Trademarks | Privacy policy | Terms of use | Terms of sale