Update spaCy pipeline

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 en_med12_trf-any-py3-none-any.whl filter=lfs diff=lfs merge=lfs -text
 textcat/model filter=lfs diff=lfs merge=lfs -text
 transformer/model filter=lfs diff=lfs merge=lfs -text
+tok2vec/model filter=lfs diff=lfs merge=lfs -text

README.md

@@ -14,36 +14,36 @@ model-index:
     metrics:
     - name: NER Precision
       type: precision
-      value: 0.9987402508
+      value: 0.8630460449
     - name: NER Recall
       type: recall
-      value: 0.9991845056
+      value: 0.8640661939
     - name: NER F Score
       type: f_score
-      value: 0.9989623288
+      value: 0.8635558181
 ---
 | Feature | Description |
 | --- | --- |
 | **Name** | `en_med12_trf` |
-| **Version** | `0.0.0` |
+| **Version** | `1` |
 | **spaCy** | `>=3.4.1,<3.5.0` |
-| **Default Pipeline** | `transformer`, `ner`, `textcat` |
-| **Components** | `transformer`, `ner`, `textcat` |
+| **Default Pipeline** | `tok2vec`, `transformer`, `ner`, `textcat` |
+| **Components** | `tok2vec`, `transformer`, `ner`, `textcat` |
 | **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
 | **Sources** | n/a |
-| **License** | Apache License, Version 2.0 |
-| **Author** | Karl Renius |
+| **License** | n/a |
+| **Author** | [n/a]() |
 
 ### Label Scheme
 
 <details>
 
-<summary>View label scheme (15 labels for 2 components)</summary>
+<summary>View label scheme (14 labels for 2 components)</summary>
 
 | Component | Labels |
 | --- | --- |
 | **`ner`** | `Denominator_Unit`, `Denominator_Value`, `Dose_Form`, `Medication_Name`, `NDC`, `Numerator_Unit`, `Numerator_Value`, `Product_Package_Type`, `Product_Package_Type_Value`, `Quantity_Factor_Unit`, `Quantity_Factor_Unit_Value`, `Quantity_Factor_Value` |
-| **`textcat`** | `MEDICATION`, `OTHER`, `DEVICE` |
+| **`textcat`** | `MEDICATION`, `OTHER` |
 
 </details>
 
@@ -51,18 +51,19 @@ model-index:
 
 | Type | Score |
 | --- | --- |
-| `ENTS_F` | 99.90 |
-| `ENTS_P` | 99.87 |
-| `ENTS_R` | 99.92 |
-| `CATS_SCORE` | 100.00 |
-| `CATS_MICRO_P` | 100.00 |
-| `CATS_MICRO_R` | 100.00 |
-| `CATS_MICRO_F` | 100.00 |
-| `CATS_MACRO_P` | 100.00 |
-| `CATS_MACRO_R` | 100.00 |
-| `CATS_MACRO_F` | 100.00 |
-| `CATS_MACRO_AUC` | 100.00 |
+| `ENTS_F` | 86.36 |
+| `ENTS_P` | 86.30 |
+| `ENTS_R` | 86.41 |
+| `CATS_SCORE` | 96.85 |
+| `CATS_MICRO_P` | 93.61 |
+| `CATS_MICRO_R` | 99.64 |
+| `CATS_MICRO_F` | 96.53 |
+| `CATS_MACRO_P` | 94.24 |
+| `CATS_MACRO_R` | 99.61 |
+| `CATS_MACRO_F` | 96.85 |
+| `CATS_MACRO_AUC` | 99.68 |
 | `CATS_MACRO_AUC_PER_TYPE` | 0.00 |
-| `TRANSFORMER_LOSS` | 19686.79 |
-| `NER_LOSS` | 513489.20 |
-| `TEXTCAT_LOSS` | 2.08 |
+| `TOK2VEC_LOSS` | 0.00 |
+| `TRANSFORMER_LOSS` | 131016.45 |
+| `NER_LOSS` | 28078.22 |
+| `TEXTCAT_LOSS` | 1261.44 |

config.cfg

@@ -1,16 +1,16 @@
 [paths]
-train = "ner_cat_train.spacy"
-dev = "ner_cat_test.spacy"
+train = null
+dev = null
 vectors = null
 init_tok2vec = null
 
 [system]
-gpu_allocator = "pytorch"
+gpu_allocator = null
 seed = 0
 
 [nlp]
 lang = "en"
-pipeline = ["transformer","ner","textcat"]
+pipeline = ["tok2vec","transformer","ner","textcat"]
 batch_size = 128
 disabled = []
 before_creation = null
@@ -64,6 +64,26 @@ grad_factor = 1.0
 pooling = {"@layers":"reduce_mean.v1"}
 upstream = "*"
 
+[components.tok2vec]
+factory = "tok2vec"
+
+[components.tok2vec.model]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.tok2vec.model.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = ${components.tok2vec.model.encode.width}
+attrs = ["NORM","PREFIX","SUFFIX","SHAPE"]
+rows = [5000,2500,2500,2500]
+include_static_vectors = false
+
+[components.tok2vec.model.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 96
+depth = 4
+window_size = 1
+maxout_pieces = 3
+
 [components.transformer]
 factory = "transformer"
 max_batch_items = 4096
@@ -87,22 +107,27 @@ use_fast = true
 [components.transformer.model.transformer_config]
 
 [corpora]
-
-[corpora.dev]
-@readers = "spacy.Corpus.v1"
-path = ${paths.dev}
-max_length = 0
-gold_preproc = false
-limit = 0
-augmenter = null
-
-[corpora.train]
-@readers = "spacy.Corpus.v1"
-path = ${paths.train}
-max_length = 0
-gold_preproc = false
-limit = 0
-augmenter = null
+@readers = "prodigy.MergedCorpus.v1"
+eval_split = 0.2
+sample_size = 1.0
+textcat_multilabel = null
+parser = null
+tagger = null
+senter = null
+spancat = null
+
+[corpora.ner]
+@readers = "prodigy.NERCorpus.v1"
+datasets = ["real_world_meds"]
+eval_datasets = []
+default_fill = "outside"
+incorrect_key = "incorrect_spans"
+
+[corpora.textcat]
+@readers = "prodigy.TextCatCorpus.v1"
+datasets = ["db-labeled"]
+eval_datasets = []
+exclusive = true
 
 [training]
 accumulate_gradient = 3
@@ -127,7 +152,7 @@ buffer = 256
 get_length = null
 
 [training.logger]
-@loggers = "spacy.ConsoleLogger.v1"
+@loggers = "prodigy.ConsoleLogger.v1"
 progress_bar = false
 
 [training.optimizer]

en_med12_trf-any-py3-none-any.whl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:912d4cbccd78fb209542ccfe36bbfae676e05fc585099b62cc0146375877ce82
-size 453535042
+oid sha256:cbdc0c18c39ecfea3ef8dfe9194a8d9f0fcd6ec471d170d914fe682c05e0ed2a
+size 460233712

meta.json

@@ -1,7 +1,7 @@
 {
   "lang":"en",
   "name":"med12_trf",
-  "version":"0.0.0",
+  "version":"1",
   "description":"",
   "author":"",
   "email":"",
@@ -16,6 +16,9 @@
     "name":null
   },
   "labels":{
+    "tok2vec":[
+
+    ],
     "transformer":[
 
     ],
@@ -35,16 +38,17 @@
     ],
     "textcat":[
       "MEDICATION",
-      "OTHER",
-      "DEVICE"
+      "OTHER"
     ]
   },
   "pipeline":[
+    "tok2vec",
     "transformer",
     "ner",
     "textcat"
   ],
   "components":[
+    "tok2vec",
     "transformer",
     "ner",
     "textcat"
@@ -53,101 +57,92 @@
 
   ],
   "performance":{
-    "ents_f":0.9989623288,
-    "ents_p":0.9987402508,
-    "ents_r":0.9991845056,
+    "ents_f":0.8635558181,
+    "ents_p":0.8630460449,
+    "ents_r":0.8640661939,
     "ents_per_type":{
-      "Quantity_Factor_Value":{
-        "p":0.9984326019,
-        "r":0.9992156863,
-        "f":0.9988239906
-      },
-      "Quantity_Factor_Unit":{
-        "p":0.9843575419,
-        "r":0.9951054217,
-        "f":0.9897023029
-      },
       "Medication_Name":{
-        "p":0.9995313964,
-        "r":0.999344078,
-        "f":0.9994377284
+        "p":0.8534482759,
+        "r":0.8497854077,
+        "f":0.8516129032
       },
       "Numerator_Value":{
-        "p":0.9995644125,
-        "r":0.9996732738,
-        "f":0.9996188402
+        "p":0.9416666667,
+        "r":0.9826086957,
+        "f":0.9617021277
       },
       "Numerator_Unit":{
-        "p":0.999780894,
-        "r":0.9996713769,
-        "f":0.9997261324
-      },
-      "Denominator_Unit":{
-        "p":0.9982694685,
-        "r":0.9995049505,
-        "f":0.9988868275
+        "p":0.9230769231,
+        "r":0.9391304348,
+        "f":0.9310344828
       },
       "Dose_Form":{
-        "p":0.9995135595,
-        "r":0.9984207969,
-        "f":0.9989668794
+        "p":0.8113207547,
+        "r":0.8322580645,
+        "f":0.821656051
       },
-      "Product_Package_Type_Value":{
-        "p":1.0,
-        "r":1.0,
-        "f":1.0
+      "Quantity_Factor_Value":{
+        "p":0.8064516129,
+        "r":0.7575757576,
+        "f":0.78125
       },
-      "Product_Package_Type":{
-        "p":1.0,
-        "r":1.0,
-        "f":1.0
+      "Quantity_Factor_Unit":{
+        "p":0.8928571429,
+        "r":0.8333333333,
+        "f":0.8620689655
       },
-      "NDC":{
-        "p":1.0,
-        "r":1.0,
-        "f":1.0
+      "Product_Package_Type":{
+        "p":0.8717948718,
+        "r":0.7727272727,
+        "f":0.8192771084
       },
-      "Denominator_Value":{
-        "p":1.0,
-        "r":0.9948979592,
-        "f":0.9974424552
+      "Product_Package_Type_Value":{
+        "p":0.8571428571,
+        "r":0.8571428571,
+        "f":0.8571428571
       },
       "Quantity_Factor_Unit_Value":{
-        "p":1.0,
+        "p":0.9090909091,
+        "r":0.7547169811,
+        "f":0.824742268
+      },
+      "Denominator_Value":{
+        "p":0.4545454545,
         "r":1.0,
-        "f":1.0
+        "f":0.625
+      },
+      "Denominator_Unit":{
+        "p":0.6470588235,
+        "r":0.9166666667,
+        "f":0.7586206897
       }
     },
-    "cats_score":1.0,
+    "cats_score":0.9684618658,
     "cats_score_desc":"macro F",
-    "cats_micro_p":1.0,
-    "cats_micro_r":1.0,
-    "cats_micro_f":1.0,
-    "cats_macro_p":1.0,
-    "cats_macro_r":1.0,
-    "cats_macro_f":1.0,
-    "cats_macro_auc":1.0,
+    "cats_micro_p":0.9360913862,
+    "cats_micro_r":0.9963851462,
+    "cats_micro_f":0.9652976759,
+    "cats_macro_p":0.9423963976,
+    "cats_macro_r":0.9961496466,
+    "cats_macro_f":0.9684618658,
+    "cats_macro_auc":0.9968290074,
     "cats_f_per_type":{
       "MEDICATION":{
-        "p":1.0,
-        "r":1.0,
-        "f":1.0
+        "p":0.9273913043,
+        "r":0.996728972,
+        "f":0.9608108108
       },
       "OTHER":{
-        "p":1.0,
-        "r":1.0,
-        "f":1.0
-      },
-      "DEVICE":{
-        "p":1.0,
-        "r":1.0,
-        "f":1.0
+        "p":0.9574014909,
+        "r":0.9955703212,
+        "f":0.9761129207
       }
     },
     "cats_macro_auc_per_type":0.0,
-    "transformer_loss":196.8679366748,
-    "ner_loss":5134.8920414948,
-    "textcat_loss":0.0208005876
+    "tok2vec_loss":0.0,
+    "transformer_loss":1310.1644586238,
+    "ner_loss":280.7821546286,
+    "textcat_loss":12.6144110119
   },
   "requirements":[
     "spacy-transformers>=1.1.7,<1.2.0"

ner/moves

@@ -1 +1 @@
-��moves��{"0":{},"1":{"Dose_Form":68852,"Medication_Name":66662,"Numerator_Value":36834,"Numerator_Unit":36579,"NDC":29835,"Denominator_Unit":16075,"Product_Package_Type":10496,"Quantity_Factor_Unit":10472,"Quantity_Factor_Value":10195,"Product_Package_Type_Value":9752,"Quantity_Factor_Unit_Value":3419,"Denominator_Value":627},"2":{"Dose_Form":68852,"Medication_Name":66662,"Numerator_Value":36834,"Numerator_Unit":36579,"NDC":29835,"Denominator_Unit":16075,"Product_Package_Type":10496,"Quantity_Factor_Unit":10472,"Quantity_Factor_Value":10195,"Product_Package_Type_Value":9752,"Quantity_Factor_Unit_Value":3419,"Denominator_Value":627},"3":{"Dose_Form":68852,"Medication_Name":66662,"Numerator_Value":36834,"Numerator_Unit":36579,"NDC":29835,"Denominator_Unit":16075,"Product_Package_Type":10496,"Quantity_Factor_Unit":10472,"Quantity_Factor_Value":10195,"Product_Package_Type_Value":9752,"Quantity_Factor_Unit_Value":3419,"Denominator_Value":627},"4":{"Dose_Form":68852,"Medication_Name":66662,"Numerator_Value":36834,"Numerator_Unit":36579,"NDC":29835,"Denominator_Unit":16075,"Product_Package_Type":10496,"Quantity_Factor_Unit":10472,"Quantity_Factor_Value":10195,"Product_Package_Type_Value":9752,"Quantity_Factor_Unit_Value":3419,"Denominator_Value":627,"":1},"5":{"":1}}�cfg��neg_key�
+��moves��{"0":{},"1":{"Dose_Form":68280,"Medication_Name":66363,"Numerator_Value":36632,"Numerator_Unit":36387,"NDC":29380,"Denominator_Unit":15956,"Quantity_Factor_Unit":10336,"Product_Package_Type":10272,"Quantity_Factor_Value":10061,"Product_Package_Type_Value":9617,"Quantity_Factor_Unit_Value":3388,"Denominator_Value":653},"2":{"Dose_Form":68280,"Medication_Name":66363,"Numerator_Value":36632,"Numerator_Unit":36387,"NDC":29380,"Denominator_Unit":15956,"Quantity_Factor_Unit":10336,"Product_Package_Type":10272,"Quantity_Factor_Value":10061,"Product_Package_Type_Value":9617,"Quantity_Factor_Unit_Value":3388,"Denominator_Value":653},"3":{"Dose_Form":68280,"Medication_Name":66363,"Numerator_Value":36632,"Numerator_Unit":36387,"NDC":29380,"Denominator_Unit":15956,"Quantity_Factor_Unit":10336,"Product_Package_Type":10272,"Quantity_Factor_Value":10061,"Product_Package_Type_Value":9617,"Quantity_Factor_Unit_Value":3388,"Denominator_Value":653},"4":{"Dose_Form":68280,"Medication_Name":66363,"Numerator_Value":36632,"Numerator_Unit":36387,"NDC":29380,"Denominator_Unit":15956,"Quantity_Factor_Unit":10336,"Product_Package_Type":10272,"Quantity_Factor_Value":10061,"Product_Package_Type_Value":9617,"Quantity_Factor_Unit_Value":3388,"Denominator_Value":653,"":1},"5":{"":1}}�cfg��neg_key�

textcat/cfg

@@ -1,8 +1,7 @@
 {
   "labels":[
     "MEDICATION",
-    "OTHER",
-    "DEVICE"
+    "OTHER"
   ],
   "threshold":0.5,
   "positive_label":null

textcat/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4b2ab8a3a55cf5dacbc07a392060c6dce33219b5f74ec8208e0348e2b7ef5e82
-size 10254174
+oid sha256:4f483561606ab536a2921d5c0d71ed741f06ecaaa8408e2f02a8af50a7058ce7
+size 9202498

tok2vec/cfg

@@ -0,0 +1,3 @@
+{
+
+}

tok2vec/model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7defdfa5b6acb8521c4f92e0f0ce07cd91cc0d095fabf39eb07405344a330664
+size 6585091

transformer/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ae16e5a7ea2cd26be7df40459ddb514a7ac88c52e0a67ed668e0b3052d696b9e
-size 502030647
+oid sha256:d37cbcb7a5e658fa6a3e8f4e5a4c9d04ae623faf1a9033df2c6c40db5cef26fe
+size 502030732