Comments (3)
@changqi1 do we have performance analysis data? let's discuss on that.
from deeprec.
@pujiang2018 Will collect it.
from deeprec.
Update DIN and DIEN AUC and perf data. The issue has been fixed by fusing optimizer.
AUC | WDL | WDL | DLRM | DLRM | Deep FM | Deep FM | DSSM | DSSM | DIEN | DIEN | DIN | DIN |
---|---|---|---|---|---|---|---|---|---|---|---|---|
/ | value | percent | value | percent | value | percent | value | percent | value | percent | value | percent |
Community TF | 0.775182 | baseline | 0.774829 | baseline | 0.743674 | baseline | 0.503107 | baseline | 0.749009 | baseline | 0.743708 | baseline |
DeepRec FP32 | 0.775283 | 100.01% | 0.766892 | 98.98% | 0.768079 | 103.28% | 0.49983 | 99.35% | 0.749727 | 100.10% | 0.751069 | 100.99% |
DeepRec BF16 | 0.775907 | 100.09% | 0.770136 | 99.39% | 0.771628 | 103.76% | 0.49125 | 97.64% | 0.744693 | 99.42% | 0.749514 | 100.78% |
Gsteps/s | WDL | WDL | DLRM | DLRM | Deep FM | Deep FM | DSSM | DSSM | DIEN | DIEN | DIN | DIN |
---|---|---|---|---|---|---|---|---|---|---|---|---|
/ | value | percent | value | percent | value | percent | value | percent | value | percent | value | percent |
Community TF | 31.78435 | baseline | 81.4195 | baseline | 37.17086 | baseline | 18.35918 | baseline | 12.39816 | baseline | 30.98243 | baseline |
DeepRec FP32 | 33.09086 | 104.11% | 95.18336 | 116.90% | 61.46826 | 165.37% | 27.89201 | 151.92% | 15.48097 | 124.87% | 117.1144 | 378.00% |
DeepRec BF16 | 47.52758 | 149.53% | 103.9707 | 127.70% | 70.93534 | 190.84% | 30.49329 | 166.09% | 16.04443 | 129.41% | 103.5385 | 334.18% |
from deeprec.
Related Issues (20)
- [SmartStageGPU] WDL model run failed when enable do_smart_stage_gpu HOT 1
- Horovod couldn't get the right shapes for grads when using EmbeddingVariable HOT 1
- 【WorkQueue coredump】When use WorkQueue in 1ps/2worker, WorkQueue has 2 files, it will happen coredump. HOT 2
- Fix build for GCC11&12
- [CustomFileSystem] Build custom file system with DeepRec throw undefined symbol in runtime. HOT 1
- Build libtensorflow_cc.so fail HOT 1
- [Distributed] Distributed training failed with grpc++ or star-server protocol.
- ParquetDataset raise ValueError: No supported fields found in parquet file HOT 1
- [Auto Micro Batch] Iterator has not been initialized when setting micro_batch_num
- compile processor failed with config mkl_threadpool HOT 1
- N10tensorflow15EmbeddingVarGPUIxfEE not found. HOT 1
- ParquetDataset met coredump when data contain DELTA_BINARY_PACKED encoding HOT 1
- How to run DLRM model on Deeprec on A100
- ParquetDataset return dynamic shape Tensor when set drop_remainder True HOT 1
- ParquetDataset return a error shape . HOT 1
- Under sync training way,how to sovle the problem that large-batch leads to the worse generalization HOT 1
- The contact information(dingding) has expired, may i have a new contact HOT 3
- QR code is invalid HOT 1
- EV Initializer的使用案例文档编写异常 HOT 3
- Docker container cannot run python scripts with tensorflow imported. HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from deeprec.