Madeesh Kannan
a881451d3a
refactor: Refactor EvaluationResult into BaseEvaluationRunResult and EvaluationRunResult ( #7594 )
...
The new `EvaluationRunResult` has slightly different semantics - it separates the previous `data` parameter into `inputs` and `results`and expects aggregate scores to be provided in the latter.
2024-04-25 12:16:48 +02:00
Silvano Cerza
cf221a9701
Delete old eval API ( #6983 )
2024-02-14 17:11:08 +01:00
Silvano Cerza
36ab23d360
feat: Add StatisticalEvaluator component ( #6982 )
...
* Add StatisticalEvaluator component
* Remove F1 and Exact Metric from old API
* Add release notes
* Update docstrings
2024-02-14 16:48:03 +01:00
Silvano Cerza
9297fca520
feat: Add SASEvaluator component ( #6980 )
...
* Add SASEvaluator component
* Add release notes
* Delete old tests
* Remove SAS metric in old API
* Avoid importing whole numpy package
2024-02-14 16:16:22 +01:00
Ashwin Mathur
393a7993c3
feat: Add Semantic Answer Similarity metric ( #6877 )
...
* Add SAS metric
* Add release notes
* Round similarity scores for precision consistency
* Add tolerance to tests
* Update haystack/evaluation/eval.py
Co-authored-by: Silvano Cerza <3314350+silvanocerza@users.noreply.github.com>
* Add types for preprocess_text; Add additional types for f1 and em methods
---------
Co-authored-by: Silvano Cerza <3314350+silvanocerza@users.noreply.github.com>
2024-02-02 17:07:52 +01:00
Ashwin Mathur
7217f9d9f0
feat: Add F1 metric ( #6822 )
...
* Add F1 metric
* Add release notes
2024-01-26 11:04:43 +01:00
Ashwin Mathur
a238c6dd51
feat: Add Exact Match metric ( #6696 )
...
* Add exact match metric
* Add release notes
* Cleanup comments in test_eval_exact_match.py
* Create separate preprocessing function; Add output_key parameter
* Update release note
---------
Co-authored-by: Silvano Cerza <3314350+silvanocerza@users.noreply.github.com>
Co-authored-by: Julian Risch <julian.risch@deepset.ai>
2024-01-22 09:57:04 +01:00
Ashwin Mathur
374a937663
feat: Add calculate_metrics and MetricsResult ( #6680 )
...
* Add calculate_metrics, MetricsResult, Exact Match
* Add additional tests for metric calculation
* Add release notes
* Add docstring for Exact Match metric
* Remove Exact Match Implementation
* Update release notes
* Remove unnecessary metrics implementation
* Simplify logic to run supported metrics
* Add some evaluation tests
* Fix linting
---------
Co-authored-by: Silvano Cerza <3314350+silvanocerza@users.noreply.github.com>
Co-authored-by: Silvano Cerza <silvanocerza@gmail.com>
2024-01-10 10:26:44 +01:00
