Comments (2)
Thank you for checking codequestion out. My understanding is you're looking to reproduce the results below:
StackExchange Query
Models scored using Mean Reciprocal Rank (MRR)
Model | MRR |
---|---|
SE 300d - BM25 | 76.3 |
ParaNMT - BM25 | 67.4 |
FastText - BM25 | 66.1 |
BM25 | 49.5 |
TF-IDF | 45.9 |
STS Benchmark
Models scored using Pearson Correlation
Model | Supervision | Dev | Test |
---|---|---|---|
ParaNMT - BM25 | Train | 82.6 | 78.1 |
FastText - BM25 | Train | 79.8 | 72.7 |
SE 300d - BM25 | Train | 77.0 | 69.1 |
- The default model is SE 300d
- fastText uses the pre-trained Common Crawl 600B
- ParaNMT vectors were built using the txt file in this archive file converted to pymagnitude vectors using the following command (installed as a dependency of codequestion via txtai).
python -m pymagnitude.converter -i paranmt.txt -o paranmt.magnitude
For each set of vectors, you need to build an index and run the STS tests.
from codequestion.
Marking this issue as resolved, re-open if issues still persist
from codequestion.
Related Issues (20)
- how to get the dump data of stackoverflow and train it from scratch? HOT 2
- Migrate embeddings logic to txtai
- Integrate FastAPI for model serving
- Update models to use latest Stack Exchange data HOT 1
- pip install results No matching distribution found for torch>=1.4.0 (from txtai>=1.2.0->codequestion) HOT 2
- Single character in shell causes error
- UserWarning: Trying to unpickle estimator TruncatedSVD from version 0.23.1 when using version 0.23.2 HOT 2
- We need more
- ImportError: Faiss library is not installed HOT 2
- Vector model file not found (cord19-300d.magnitude) HOT 1
- Migrate from mdv to rich library
- Sync with txtai 2.0
- Update to txtai 3.2
- Migrate from word vector models to sentence transformers models
- Upgrade to txtai 5.x
- Add code quality checks
- Can this be used to train on custom archives , such as archive standard library documentations? HOT 5
- 52177 segmentation fault codequestion HOT 2
- file not found /home/dmezzett/.codequestion/vectors/stackexchange-300d.magnitude HOT 1
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 codequestion.