Update spaCy pipeline

README.md

@@ -14,25 +14,25 @@ model-index:
     metrics:
     - name: NER Precision
       type: precision
-      value: 0.8630460449
+      value: 0.8398220245
     - name: NER Recall
       type: recall
-      value: 0.8640661939
+      value: 0.8445190157
     - name: NER F Score
       type: f_score
-      value: 0.8635558181
+      value: 0.842163971
 ---
 | Feature | Description |
 | --- | --- |
 | **Name** | `en_med12_trf` |
 | **Version** | `1` |
 | **spaCy** | `>=3.4.1,<3.5.0` |
-| **Default Pipeline** | `tok2vec`, `transformer`, `ner`, `textcat` |
-| **Components** | `tok2vec`, `transformer`, `ner`, `textcat` |
+| **Default Pipeline** | `transformer`, `ner`, `textcat` |
+| **Components** | `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
 
@@ -43,7 +43,7 @@ model-index:
 | 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` |
+| **`textcat`** | `OTHER`, `MEDICATION` |
 
 </details>
 
@@ -51,19 +51,18 @@ model-index:
 
 | Type | Score |
 | --- | --- |
-| `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 |
+| `ENTS_F` | 84.22 |
+| `ENTS_P` | 83.98 |
+| `ENTS_R` | 84.45 |
+| `CATS_SCORE` | 93.88 |
+| `CATS_MICRO_P` | 89.78 |
+| `CATS_MICRO_R` | 97.98 |
+| `CATS_MICRO_F` | 93.70 |
+| `CATS_MACRO_P` | 90.29 |
+| `CATS_MACRO_R` | 97.93 |
+| `CATS_MACRO_F` | 93.88 |
+| `CATS_MACRO_AUC` | 98.53 |
 | `CATS_MACRO_AUC_PER_TYPE` | 0.00 |
-| `TOK2VEC_LOSS` | 0.00 |
-| `TRANSFORMER_LOSS` | 131016.45 |
-| `NER_LOSS` | 28078.22 |
-| `TEXTCAT_LOSS` | 1261.44 |
+| `TRANSFORMER_LOSS` | 152780.09 |
+| `NER_LOSS` | 69513.43 |
+| `TEXTCAT_LOSS` | 1868.30 |

config.cfg

@@ -1,17 +1,17 @@
 [paths]
-train = null
-dev = null
+train = "100_percent.spacy"
+dev = "rw_test.spacy"
 vectors = null
 init_tok2vec = null
 
 [system]
-gpu_allocator = null
-seed = 0
+gpu_allocator = "pytorch"
+seed = 42
 
 [nlp]
 lang = "en"
-pipeline = ["tok2vec","transformer","ner","textcat"]
-batch_size = 128
+pipeline = ["transformer","ner","textcat"]
+batch_size = 256
 disabled = []
 before_creation = null
 after_creation = null
@@ -64,26 +64,6 @@ 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
@@ -107,27 +87,22 @@ use_fast = true
 [components.transformer.model.transformer_config]
 
 [corpora]
-@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
+
+[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
 
 [training]
 accumulate_gradient = 3
@@ -135,7 +110,7 @@ dev_corpus = "corpora.dev"
 train_corpus = "corpora.train"
 seed = ${system.seed}
 gpu_allocator = ${system.gpu_allocator}
-dropout = 0.1
+dropout = 0.05
 patience = 1600
 max_epochs = 0
 max_steps = 20000
@@ -152,8 +127,13 @@ buffer = 256
 get_length = null
 
 [training.logger]
-@loggers = "prodigy.ConsoleLogger.v1"
-progress_bar = false
+@loggers = "spacy.WandbLogger.v3"
+project_name = "med12"
+remove_config_values = ["paths.train","paths.dev","corpora.train.path","corpora.dev.path"]
+log_dataset_dir = "./corpus"
+model_log_interval = 1000
+entity = null
+run_name = null
 
 [training.optimizer]
 @optimizers = "Adam.v1"
@@ -200,4 +180,18 @@ after_init = null
 
 [initialize.components]
 
+[initialize.components.ner]
+
+[initialize.components.ner.labels]
+@readers = "spacy.read_labels.v1"
+path = "C:\\Users\\karl2\\Documents\\medication_parsing\\rxnorm_model_development\\models\\roberta_weak_rw\\labels\\ner.json"
+require = false
+
+[initialize.components.textcat]
+
+[initialize.components.textcat.labels]
+@readers = "spacy.read_labels.v1"
+path = "C:\\Users\\karl2\\Documents\\medication_parsing\\rxnorm_model_development\\models\\roberta_weak_rw\\labels\\textcat.json"
+require = false
+
 [initialize.tokenizer]

en_med12_trf-any-py3-none-any.whl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cbdc0c18c39ecfea3ef8dfe9194a8d9f0fcd6ec471d170d914fe682c05e0ed2a
-size 460233712
+oid sha256:2e0d7e75743984013d22e7c555d824fd77b38f46e6f851bd031695e14a933a34
+size 454351979

meta.json

@@ -16,9 +16,6 @@
     "name":null
   },
   "labels":{
-    "tok2vec":[
-
-    ],
     "transformer":[
 
     ],
@@ -37,18 +34,16 @@
       "Quantity_Factor_Value"
     ],
     "textcat":[
-      "MEDICATION",
-      "OTHER"
+      "OTHER",
+      "MEDICATION"
     ]
   },
   "pipeline":[
-    "tok2vec",
     "transformer",
     "ner",
     "textcat"
   ],
   "components":[
-    "tok2vec",
     "transformer",
     "ner",
     "textcat"
@@ -57,92 +52,91 @@
 
   ],
   "performance":{
-    "ents_f":0.8635558181,
-    "ents_p":0.8630460449,
-    "ents_r":0.8640661939,
+    "ents_f":0.842163971,
+    "ents_p":0.8398220245,
+    "ents_r":0.8445190157,
     "ents_per_type":{
       "Medication_Name":{
-        "p":0.8534482759,
-        "r":0.8497854077,
-        "f":0.8516129032
-      },
-      "Numerator_Value":{
-        "p":0.9416666667,
-        "r":0.9826086957,
-        "f":0.9617021277
-      },
-      "Numerator_Unit":{
-        "p":0.9230769231,
-        "r":0.9391304348,
-        "f":0.9310344828
+        "p":0.8181818182,
+        "r":0.8247011952,
+        "f":0.8214285714
       },
       "Dose_Form":{
-        "p":0.8113207547,
-        "r":0.8322580645,
-        "f":0.821656051
+        "p":0.7718120805,
+        "r":0.7467532468,
+        "f":0.7590759076
       },
       "Quantity_Factor_Value":{
-        "p":0.8064516129,
-        "r":0.7575757576,
-        "f":0.78125
+        "p":0.8421052632,
+        "r":0.7111111111,
+        "f":0.7710843373
+      },
+      "Quantity_Factor_Unit_Value":{
+        "p":0.8,
+        "r":0.8181818182,
+        "f":0.808988764
       },
       "Quantity_Factor_Unit":{
-        "p":0.8928571429,
-        "r":0.8333333333,
-        "f":0.8620689655
+        "p":0.9166666667,
+        "r":0.8148148148,
+        "f":0.862745098
       },
       "Product_Package_Type":{
-        "p":0.8717948718,
-        "r":0.7727272727,
-        "f":0.8192771084
+        "p":0.8055555556,
+        "r":0.6744186047,
+        "f":0.7341772152
       },
-      "Product_Package_Type_Value":{
-        "p":0.8571428571,
-        "r":0.8571428571,
-        "f":0.8571428571
-      },
-      "Quantity_Factor_Unit_Value":{
-        "p":0.9090909091,
-        "r":0.7547169811,
-        "f":0.824742268
+      "Numerator_Value":{
+        "p":0.9047619048,
+        "r":0.9851851852,
+        "f":0.9432624113
       },
-      "Denominator_Value":{
-        "p":0.4545454545,
-        "r":1.0,
-        "f":0.625
+      "Numerator_Unit":{
+        "p":0.928057554,
+        "r":0.9772727273,
+        "f":0.9520295203
       },
       "Denominator_Unit":{
-        "p":0.6470588235,
-        "r":0.9166666667,
-        "f":0.7586206897
+        "p":0.7222222222,
+        "r":0.8666666667,
+        "f":0.7878787879
+      },
+      "Product_Package_Type_Value":{
+        "p":0.75,
+        "r":0.9375,
+        "f":0.8333333333
+      },
+      "Denominator_Value":{
+        "p":0.3333333333,
+        "r":0.4,
+        "f":0.3636363636
       }
     },
-    "cats_score":0.9684618658,
+    "cats_score":0.93877706,
     "cats_score_desc":"macro F",
-    "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_micro_p":0.8978046934,
+    "cats_micro_r":0.9797604296,
+    "cats_micro_f":0.9369938771,
+    "cats_macro_p":0.9029397761,
+    "cats_macro_r":0.9792967771,
+    "cats_macro_f":0.93877706,
+    "cats_macro_auc":0.9853349088,
     "cats_f_per_type":{
-      "MEDICATION":{
-        "p":0.9273913043,
-        "r":0.996728972,
-        "f":0.9608108108
-      },
       "OTHER":{
-        "p":0.9574014909,
-        "r":0.9955703212,
-        "f":0.9761129207
+        "p":0.9478632479,
+        "r":0.9711033275,
+        "f":0.9593425606
+      },
+      "MEDICATION":{
+        "p":0.8580163043,
+        "r":0.9874902267,
+        "f":0.9182115594
       }
     },
     "cats_macro_auc_per_type":0.0,
-    "tok2vec_loss":0.0,
-    "transformer_loss":1310.1644586238,
-    "ner_loss":280.7821546286,
-    "textcat_loss":12.6144110119
+    "transformer_loss":1527.8009473217,
+    "ner_loss":695.1342882158,
+    "textcat_loss":18.6830486141
   },
   "requirements":[
     "spacy-transformers>=1.1.7,<1.2.0"

ner/moves

@@ -1 +1 @@
-��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�
+��moves��{"0":{},"1":{"Dose_Form":68500,"Medication_Name":66539,"Numerator_Value":36769,"Numerator_Unit":36524,"NDC":30140,"Denominator_Unit":16207,"Product_Package_Type":10669,"Quantity_Factor_Unit":10581,"Quantity_Factor_Value":10160,"Product_Package_Type_Value":9959,"Quantity_Factor_Unit_Value":3526,"Denominator_Value":685},"2":{"Dose_Form":68500,"Medication_Name":66539,"Numerator_Value":36769,"Numerator_Unit":36524,"NDC":30140,"Denominator_Unit":16207,"Product_Package_Type":10669,"Quantity_Factor_Unit":10581,"Quantity_Factor_Value":10160,"Product_Package_Type_Value":9959,"Quantity_Factor_Unit_Value":3526,"Denominator_Value":685},"3":{"Dose_Form":68500,"Medication_Name":66539,"Numerator_Value":36769,"Numerator_Unit":36524,"NDC":30140,"Denominator_Unit":16207,"Product_Package_Type":10669,"Quantity_Factor_Unit":10581,"Quantity_Factor_Value":10160,"Product_Package_Type_Value":9959,"Quantity_Factor_Unit_Value":3526,"Denominator_Value":685},"4":{"Dose_Form":68500,"Medication_Name":66539,"Numerator_Value":36769,"Numerator_Unit":36524,"NDC":30140,"Denominator_Unit":16207,"Product_Package_Type":10669,"Quantity_Factor_Unit":10581,"Quantity_Factor_Value":10160,"Product_Package_Type_Value":9959,"Quantity_Factor_Unit_Value":3526,"Denominator_Value":685,"":1},"5":{"":1}}�cfg��neg_key�

textcat/cfg

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

textcat/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4f483561606ab536a2921d5c0d71ed741f06ecaaa8408e2f02a8af50a7058ce7
+oid sha256:3412ff82cfd7317a5373da70b8f0ac507e6f41b0862c15fabdd8124c2934ec1c
 size 9202498

transformer/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d37cbcb7a5e658fa6a3e8f4e5a4c9d04ae623faf1a9033df2c6c40db5cef26fe
+oid sha256:54558ded63daa3c035d9e37f5c483cbea285bb518f2601c73e05f5908a6016ac
 size 502030732