Update README.md

README.md

@@ -37,6 +37,16 @@ Install or update the gliner package:
 pip install gliner -U
 ```
 
+And LLM2Vec package:
+```bash
+pip install llm2vec
+```
+
+To use this particular Qwen-based model you need different `transformers` package version than llm2vec requires, so install it manually:
+```bash
+pip install transformers==4.44.1
+```
+
 Once you've downloaded the GLiNER library, you can import the GLiNER class. You can then load this model using `GLiNER.from_pretrained` and predict entities with `predict_entities`.
 
 ```python
@@ -92,37 +102,39 @@ outputs = model.batch_predict_with_embeds(texts, entity_embeddings, labels)
 ### Benchmarks
 Below you can see the table with benchmarking results on various named entity recognition datasets:
 
+Here’s the updated table with your new data:
+
 | Dataset                 | Score  |
 |-------------------------|--------|
-| ACE 2004                | 26.8%  |
-| ACE 2005                | 29.2%  |
-| AnatEM                  | 25.3%  |
-| Broad Tweet Corpus      | 66.8%  |
-| CoNLL 2003              | 60.3%  |
-| FabNER                 | 21.2%  |
-| FindVehicle             | 28.3%  |
-| GENIA_NER               | 58.3%  |
-| HarveyNER               | 18.3%  |
-| MultiNERD               | 64.7%  |
-| Ontonotes               | 28.4%  |
-| PolyglotNER             | 45.3%  |
-| TweetNER7               | 35.9%  |
-| WikiANN en              | 53.6%  |
-| WikiNeural              | 73.4%  |
-| bc2gm                   | 63.2%  |
-| bc4chemd                | 56.8%  |
-| bc5cdr                  | 71.3%  |
-| ncbi                    | 64.9%  |
-| **Average**             | **47.0%** |
+| ACE 2004                | 31.5%  |
+| ACE 2005                | 31.5%  |
+| AnatEM                  | 43.4%  |
+| Broad Tweet Corpus      | 55.6%  |
+| CoNLL 2003              | 60.1%  |
+| FabNER                  | 23.9%  |
+| FindVehicle             | 30.2%  |
+| GENIA_NER               | 50.7%  |
+| HarveyNER               | 16.9%  |
+| MultiNERD               | 53.3%  |
+| Ontonotes               | 28.1%  |
+| PolyglotNER             | 39.2%  |
+| TweetNER7               | 35.3%  |
+| WikiANN en              | 53.2%  |
+| WikiNeural              | 65.0%  |
+| bc2gm                   | 56.3%  |
+| bc4chemd                | 54.4%  |
+| bc5cdr                  | 71.0%  |
+| ncbi                    | 63.7%  |
+| **Average**             | **45.4%** |
 |                         |        |
-| CrossNER_AI             | 56.7%  |
-| CrossNER_literature     | 61.5%  |
-| CrossNER_music          | 70.2%  |
-| CrossNER_politics       | 75.6%  |
-| CrossNER_science        | 66.8%  |
-| mit-movie               | 39.9%  |
-| mit-restaurant          | 41.7%  |
-| **Average (zero-shot benchmark)** | **58.9%** |
+| CrossNER_AI             | 54.0%  |
+| CrossNER_literature     | 64.4%  |
+| CrossNER_music          | 63.0%  |
+| CrossNER_politics       | 69.3%  |
+| CrossNER_science        | 64.2%  |
+| mit-movie               | 52.7%  |
+| mit-restaurant          | 37.6%  |
+| **Average (zero-shot benchmark)** | **57.9%** |
 
 ### Join Our Discord