v0.0.4 instructions and pypi info corrected

MartinoMensio · Jul 24, 2020 · de52b38 · de52b38
1 parent 2e91f95
commit de52b38
Show file tree

Hide file tree

Showing 20 changed files with 105 additions and 67 deletions.
diff --git a/README.md b/README.md
@@ -13,20 +13,53 @@ To install this package, you can run one of the following:
 - `pip install spacy_sentence_bert`
 -  `pip install git+https://github.com/MartinoMensio/spacy-sentence-bert.git`
 
+You can install standalone spaCy packages from GitHub with pip.
+From the [full list of models](https://docs.google.com/spreadsheets/d/14QplCdTCDwEmTqrn1LH4yrbKvdogK4oQvYO1K1aPR5M/edit#gid=0) this table describes the models available.
+
+
+|  sentence-BERT name                    |  spacy model name  |  dimensions          |  language  | STS benchmark | standalone install |
+|----------------------------------------|--------------------|----------------------|------------|---------------|---------|
+| `bert-base-nli-mean-tokens`            | `en_bert_base_nli_mean_tokens`            |  768 | en | 77.12          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_bert_base_nli_mean_tokens-0.0.4.tar.gz#en_bert_base_nli_mean_tokens-0.0.4`  |
+| `bert-base-nli-max-tokens`             | `en_bert_base_nli_max_tokens`             |  768 | en | 77.21          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_bert_base_nli_max_tokens-0.0.4.tar.gz#en_bert_base_nli_max_tokens-0.0.4`  |
+| `bert-base-nli-cls-token`              | `en_bert_base_nli_cls_token`              |  768 | en | 76.30          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_bert_base_nli_cls_token-0.0.4.tar.gz#en_bert_base_nli_cls_token-0.0.4`  |
+| `bert-large-nli-mean-tokens`           | `en_bert_large_nli_mean_tokens`           | 1024 | en | 79.19          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_bert_large_nli_mean_tokens-0.0.4.tar.gz#en_bert_large_nli_mean_tokens-0.0.4`  |
+| `bert-large-nli-max-tokens`            | `en_bert_large_nli_max_tokens`            | 1024 | en | 78.41          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_bert_large_nli_max_tokens-0.0.4.tar.gz#en_bert_large_nli_max_tokens-0.0.4`  |
+| `bert-large-nli-cls-token`             | `en_bert_large_nli_max_tokens`            | 1024 | en | 78.29          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_bert_large_nli_max_tokens-0.0.4.tar.gz#en_bert_large_nli_max_tokens-0.0.4`  |
+| `roberta-base-nli-mean-tokens`         | `en_roberta_base_nli_mean_tokens`         |  768 | en | 77.49          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_roberta_base_nli_mean_tokens-0.0.4.tar.gz#en_roberta_base_nli_mean_tokens-0.0.4`  |
+| `roberta-large-nli-mean-tokens`        | `en_roberta_large_nli_mean_tokens`        | 1024 | en | 78.69          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_roberta_large_nli_mean_tokens-0.0.4.tar.gz#en_roberta_large_nli_mean_tokens-0.0.4`  |
+| `distilbert-base-nli-mean-tokens`      | `en_distilbert_base_nli_mean_tokens`      |  768 | en | 76.97          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_distilbert_base_nli_mean_tokens-0.0.4.tar.gz#en_distilbert_base_nli_mean_tokens-0.0.4`  |
+| `bert-base-nli-stsb-mean-tokens`       | `en_bert_base_nli_stsb_mean_tokens`       |  768 | en | 85.14          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_bert_base_nli_stsb_mean_tokens-0.0.4.tar.gz#en_bert_base_nli_stsb_mean_tokens-0.0.4`  |
+| `bert-large-nli-stsb-mean-tokens`      | `en_bert_large_nli_stsb_mean_tokens`      | 1024 | en | 85.29          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_bert_large_nli_stsb_mean_tokens-0.0.4.tar.gz#en_bert_large_nli_stsb_mean_tokens-0.0.4`  |
+| `roberta-base-nli-stsb-mean-tokens`    | `en_roberta_base_nli_stsb_mean_tokens`    |  768 | en | 85.40          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_roberta_base_nli_stsb_mean_tokens-0.0.4.tar.gz#en_roberta_base_nli_stsb_mean_tokens-0.0.4`  |
+| `roberta-large-nli-stsb-mean-tokens`   | `en_roberta_large_nli_stsb_mean_tokens`   | 1024 | en | 86.31          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_roberta_large_nli_stsb_mean_tokens-0.0.4.tar.gz#en_roberta_large_nli_stsb_mean_tokens-0.0.4`  |
+| `distilbert-base-nli-stsb-mean-tokens` | `en_distilbert_base_nli_stsb_mean_tokens` |  768 | en | 84.38          | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/en_distilbert_base_nli_stsb_mean_tokens-0.0.4.tar.gz#en_distilbert_base_nli_stsb_mean_tokens-0.0.4`  |
+| `distiluse-base-multilingual-cased`    | `xx_distiluse_base_multilingual_cased`    |  512 | Arabic, Chinese, Dutch, English, French, German, Italian, Korean, Polish, Portuguese, Russian, Spanish, Turkish | 80.10 | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/xx_distiluse_base_multilingual_cased-0.0.4.tar.gz#xx_distiluse_base_multilingual_cased-0.0.4`  |
+| `xlm-r-base-en-ko-nli-ststb`           | `xx_xlm_r_base_en_ko_nli_ststb`           |  768 | en,ko | 81.47       | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/xx_xlm_r_base_en_ko_nli_ststb-0.0.4.tar.gz#xx_xlm_r_base_en_ko_nli_ststb-0.0.4`  |
+| `xlm-r-large-en-ko-nli-ststb`          | `xx_xlm_r_base_en_ko_nli_ststb`           | 1024 | en,ko | 84.05       | `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/v0.0.4/xx_xlm_r_base_en_ko_nli_ststb-0.0.4.tar.gz#xx_xlm_r_base_en_ko_nli_ststb-0.0.4`  |
+
+
+
 ## Usage
 
-With this package installed
+With this package installed you can obtain a Language model with:
+
 ```python
 import spacy_sentence_bert
-nlp = spacy_sentence_bert.load_model('en_bert_base_nli_cls_token')
+nlp = spacy_sentence_bert.load_model('en_roberta_large_nli_stsb_mean_tokens')
+```
+
+Or if a specific standalone model is installed from GitHub, you can load it from spaCy:
+```bash
+pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/en_roberta_large_nli_stsb_mean_tokens-0.0.4/en_roberta_large_nli_stsb_mean_tokens-0.0.4.tar.gz
 ```
 
-Or if a specific model is installed (e.g. `pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/en_bert_base_nli_cls_token-0.1.0/en_bert_base_nli_cls_token-0.1.0.tar.gz`)
 ```python
 import spacy
-nlp = spacy.load('en_bert_base_nli_cls_token')
+nlp = spacy.load('en_roberta_large_nli_stsb_mean_tokens')
 ```
 
+Or if you want to use one of the sentence embeddings over an existing Language object, you can use the `create_from` method:
+
 ```python
 import spacy
 import spacy_sentence_bert
@@ -35,30 +68,35 @@ nlp = spacy_sentence_bert.create_from(nlp_base, 'en_bert_base_nli_cls_token')
 nlp.pipe_names
 ```
 
+Once you have loaded the model, simply use it to obtain `vector`s and using the `similarity` method of spaCy:
+
+```python
+# get two documents
+doc_1 = nlp('Hi there, how are you?')
+doc_2 = nlp('Hello there, how are you doing today?')
+# get the vector of the Doc, Span or Token
+print(doc_1.vector.shape)
+print(doc_1[3].vector.shape)
+print(doc_1[2:4].vector.shape)
+# or use the similarity method that is based on the vectors, on Doc, Span or Token
+print(doc_1.similarity(doc_2[0:7]))
+```
+
+
+
+
+The models, when first used, download to the folder defined with `TORCH_HOME` in the environment variables (default `~/.cache/torch`).
+
+
+## Utils
 
-Full list of models
-https://docs.google.com/spreadsheets/d/14QplCdTCDwEmTqrn1LH4yrbKvdogK4oQvYO1K1aPR5M/edit#gid=0
-
-
-|  sentence-BERT name                    |  spacy model name  |  dimensions  |  language  | STS benchmark |
-|----------------------------------------|--------------------|--------------|------------|---|
-| `bert-base-nli-mean-tokens`            | `en_bert_base_nli_mean_tokens` | 768 | en | 77.12 |
-| `bert-base-nli-max-tokens`             | `en_bert_base_nli_max_tokens` | 768 | en | 77.21 |
-| `bert-base-nli-cls-token`              | `en_bert_base_nli_cls_token` | 768 | en | 76.30 |
-| `bert-large-nli-mean-tokens`           | `en_bert_large_nli_mean_tokens` | 1024 | en | 79.19 |
-| `bert-large-nli-max-tokens`            | `en_bert_large_nli_max_tokens` | 1024 | en | 78.41 |
-| `bert-large-nli-cls-token`             | `en_bert_large_nli_max_tokens` | 1024 | en | 78.29 |
-| `roberta-base-nli-mean-tokens`         | `en_roberta_base_nli_mean_tokens` | 768 | en | 77.49 |
-| `roberta-large-nli-mean-tokens`        | `en_roberta_large_nli_mean_tokens` | 1024 | en | 78.69 |
-| `distilbert-base-nli-mean-tokens`      | `en_distilbert_base_nli_mean_tokens` | 768 | en | 76.97 |
-| `bert-base-nli-stsb-mean-tokens`       | `en_bert_base_nli_stsb_mean_tokens` | 768 | en | 85.14 |
-| `bert-large-nli-stsb-mean-tokens`      | `en_bert_large_nli_stsb_mean_tokens` | 1024 | en | 85.29 |
-| `roberta-base-nli-stsb-mean-tokens`    | `en_roberta_base_nli_stsb_mean_tokens` | 768 | en | 85.40 |
-| `roberta-large-nli-stsb-mean-tokens`   | `en_roberta_large_nli_stsb_mean_tokens` | 1024 | en | 86.31 |
-| `distilbert-base-nli-stsb-mean-tokens` | `en_distilbert_base_nli_stsb_mean_tokens` | 768 | en | 84.38 |
-| `distiluse-base-multilingual-cased`    | `xx_distiluse_base_multilingual_cased` | 512 | Arabic, Chinese, Dutch, English, French, German, Italian, Korean, Polish, Portuguese, Russian, Spanish, Turkish | 80.10 |
-| `xlm-r-base-en-ko-nli-ststb`           | `xx_xlm_r_base_en_ko_nli_ststb` | 768 | en,ko | 81.47 |
-| `xlm-r-large-en-ko-nli-ststb`          | `xx_xlm_r_base_en_ko_nli_ststb` | 1024 | en,ko | 84.05 |
-
-
-The models, when first used, download to the folder defined with `TORCH_HOME` in the environment variables (default `~/.cache/torch`).
+To build and upload
+```bash
+VERSION=0.0.4
+# build the standalone models
+./build_models.sh
+# build dist/spacy_sentence_bert-${VERSION}.tar.gz
+python setup.py sdist
+# upload to pypi
+twine upload dist/spacy_sentence_bert-${VERSION}.tar.gz
+```
diff --git a/build_models.sh b/build_models.sh
@@ -1,7 +1,7 @@
 set -e
 
 
-VERSION=0.0.3
+VERSION=0.0.4
 
 for MODEL_NAME in en_bert_base_nli_cls_token en_bert_base_nli_max_tokens en_bert_base_nli_mean_tokens \
     en_bert_large_nli_mean_tokens en_bert_large_nli_max_tokens en_bert_large_nli_cls_token \

diff --git a/setup.cfg b/setup.cfg
@@ -1,8 +1,8 @@
 [metadata]
-version = 0.0.3
+version = 0.0.4
 description = SpaCy models for using sentence-BERT
 description-file = README.md
-url = https://github.com/MartinoMensio/spacy-universal-sentence-encoder-tfhub
+url = https://github.com/MartinoMensio/spacy-sentence-berts
 author = Martino Mensio
 author_email = martino.mensio@open.ac.uk
 

diff --git a/spacy_sentence_bert/meta/en_bert_base_nli_cls_token.json b/spacy_sentence_bert/meta/en_bert_base_nli_cls_token.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "bert_base_nli_cls_token",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_bert_base_nli_max_tokens.json b/spacy_sentence_bert/meta/en_bert_base_nli_max_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "bert_base_nli_max_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_bert_base_nli_mean_tokens.json b/spacy_sentence_bert/meta/en_bert_base_nli_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "bert_base_nli_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_bert_base_nli_stsb_mean_tokens.json b/spacy_sentence_bert/meta/en_bert_base_nli_stsb_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "bert_base_nli_stsb_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_bert_large_nli_cls_token.json b/spacy_sentence_bert/meta/en_bert_large_nli_cls_token.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "bert_large_nli_cls_token",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_bert_large_nli_max_tokens.json b/spacy_sentence_bert/meta/en_bert_large_nli_max_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "bert_large_nli_max_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_bert_large_nli_mean_tokens.json b/spacy_sentence_bert/meta/en_bert_large_nli_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "bert_large_nli_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_bert_large_nli_stsb_mean_tokens.json b/spacy_sentence_bert/meta/en_bert_large_nli_stsb_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "bert_large_nli_stsb_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_distilbert_base_nli_mean_tokens.json b/spacy_sentence_bert/meta/en_distilbert_base_nli_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "distilbert_base_nli_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_distilbert_base_nli_stsb_mean_tokens.json b/spacy_sentence_bert/meta/en_distilbert_base_nli_stsb_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "distilbert_base_nli_stsb_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_roberta_base_nli_mean_tokens.json b/spacy_sentence_bert/meta/en_roberta_base_nli_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "roberta_base_nli_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_roberta_base_nli_stsb_mean_tokens.json b/spacy_sentence_bert/meta/en_roberta_base_nli_stsb_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "roberta_base_nli_stsb_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_roberta_large_nli_mean_tokens.json b/spacy_sentence_bert/meta/en_roberta_large_nli_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "roberta_large_nli_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",

diff --git a/spacy_sentence_bert/meta/en_roberta_large_nli_stsb_mean_tokens.json b/spacy_sentence_bert/meta/en_roberta_large_nli_stsb_mean_tokens.json
@@ -1,15 +1,15 @@
 {
     "lang": "en",
     "name": "roberta_large_nli_stsb_mean_tokens",
-    "version": "0.0.3",
+    "version": "0.0.4",
     "spacy_version": ">=2.3,<2.4",
     "description": "Wrapper of sentence-transformers models for spaCy",
     "author": "Martino Mensio",
     "email": "martino.mensio@open.ac.uk",
     "url": "https://github.com/MartinoMensio/spacy-sentence-bert",
     "license": "MIT",
     "requirements": [
-        "spacy-sentence-bert==0.0.3"
+        "spacy-sentence-bert==0.0.4"
     ],
     "sources": [{
         "name": "sentence-transformers",