Update spaCy pipeline

.gitattributes

@@ -44,3 +44,4 @@ textcat_classify/model filter=lfs diff=lfs merge=lfs -text
 tok2vec_small/model filter=lfs diff=lfs merge=lfs -text
 vocab/strings.json filter=lfs diff=lfs merge=lfs -text
 senter/model filter=lfs diff=lfs merge=lfs -text
+textcat_multilabel/model filter=lfs diff=lfs merge=lfs -text

README.md

@@ -2,6 +2,7 @@
 tags:
 - spacy
 - token-classification
+- text-classification
 language:
 - en
 model-index:
@@ -24,10 +25,10 @@ model-index:
 | Feature | Description |
 | --- | --- |
 | **Name** | `en_tako_query_analyzer` |
-| **Version** | `0.0.3` |
+| **Version** | `0.0.4` |
 | **spaCy** | `>=3.7.5,<3.8.0` |
-| **Default Pipeline** | `tok2vec_small`, `tagger`, `parser`, `attribute_ruler`, `senter`, `lemmatizer`, `tok2vec`, `ner`, `textcat_classify` |
-| **Components** | `tok2vec_small`, `tagger`, `parser`, `attribute_ruler`, `senter`, `lemmatizer`, `tok2vec`, `ner`, `textcat_classify` |
+| **Default Pipeline** | `tok2vec_small`, `tagger`, `parser`, `attribute_ruler`, `senter`, `tok2vec`, `ner`, `textcat_multilabel`, `textcat_classify` |
+| **Components** | `tok2vec_small`, `tagger`, `parser`, `attribute_ruler`, `senter`, `tok2vec`, `ner`, `textcat_multilabel`, `textcat_classify` |
 | **Vectors** | 514157 keys, 514157 unique vectors (300 dimensions) |
 | **Sources** | n/a |
 | **License** | n/a |
@@ -37,13 +38,14 @@ model-index:
 
 <details>
 
-<summary>View label scheme (116 labels for 4 components)</summary>
+<summary>View label scheme (124 labels for 5 components)</summary>
 
 | Component | Labels |
 | --- | --- |
 | **`tagger`** | `$`, `''`, `,`, `-LRB-`, `-RRB-`, `.`, `:`, `ADD`, `AFX`, `CC`, `CD`, `DT`, `EX`, `FW`, `HYPH`, `IN`, `JJ`, `JJR`, `JJS`, `LS`, `MD`, `NFP`, `NN`, `NNP`, `NNPS`, `NNS`, `PDT`, `POS`, `PRP`, `PRP$`, `RB`, `RBR`, `RBS`, `RP`, `SYM`, `TO`, `UH`, `VB`, `VBD`, `VBG`, `VBN`, `VBP`, `VBZ`, `WDT`, `WP`, `WP$`, `WRB`, `XX`, `_SP`, ```` |
 | **`parser`** | `ROOT`, `acl`, `acomp`, `advcl`, `advmod`, `agent`, `amod`, `appos`, `attr`, `aux`, `auxpass`, `case`, `cc`, `ccomp`, `compound`, `conj`, `csubj`, `csubjpass`, `dative`, `dep`, `det`, `dobj`, `expl`, `intj`, `mark`, `meta`, `neg`, `nmod`, `npadvmod`, `nsubj`, `nsubjpass`, `nummod`, `oprd`, `parataxis`, `pcomp`, `pobj`, `poss`, `preconj`, `predet`, `prep`, `prt`, `punct`, `quantmod`, `relcl`, `xcomp` |
 | **`ner`** | `CARDINAL`, `DATE`, `EVENT`, `FAC`, `GPE`, `LANGUAGE`, `LAW`, `LOC`, `MONEY`, `NORP`, `ORDINAL`, `ORG`, `PERCENT`, `PERSON`, `PRODUCT`, `QUANTITY`, `STOCK_TICKER`, `TIME`, `WORK_OF_ART` |
+| **`textcat_multilabel`** | `Finance`, `Economics`, `Demographics`, `Health`, `Politics`, `Sports`, `Climate`, `Other` |
 | **`textcat_classify`** | `ACCEPT`, `REJECT` |
 
 </details>
@@ -56,12 +58,12 @@ model-index:
 | `ENTS_P` | 0.00 |
 | `ENTS_R` | 0.00 |
 | `ENTS_PER_TYPE` | 0.00 |
-| `CATS_SCORE` | 85.07 |
-| `CATS_MICRO_P` | 85.31 |
-| `CATS_MICRO_R` | 85.31 |
-| `CATS_MICRO_F` | 85.31 |
-| `CATS_MACRO_P` | 85.35 |
-| `CATS_MACRO_R` | 85.31 |
-| `CATS_MACRO_F` | 85.31 |
-| `CATS_MACRO_AUC` | 91.67 |
-| `TEXTCAT_CLASSIFY_LOSS` | 94.04 |
+| `CATS_SCORE` | 82.56 |
+| `CATS_MICRO_P` | 82.30 |
+| `CATS_MICRO_R` | 82.30 |
+| `CATS_MICRO_F` | 82.30 |
+| `CATS_MACRO_P` | 82.33 |
+| `CATS_MACRO_R` | 82.30 |
+| `CATS_MACRO_F` | 82.30 |
+| `CATS_MACRO_AUC` | 89.93 |
+| `TEXTCAT_CLASSIFY_LOSS` | 218.20 |

config.cfg

@@ -7,7 +7,7 @@ init_tok2vec = null
 [variables]
 wandb_project_name = "tako-query-filter"
 wandb_team_name = "tako-team"
-base_model = "ner/dashing-wind"
+base_model = "topic/upbeat-bush"
 
 [system]
 gpu_allocator = "pytorch"
@@ -15,7 +15,7 @@ seed = 0
 
 [nlp]
 lang = "en"
-pipeline = ["tok2vec_small","tagger","parser","attribute_ruler","senter","lemmatizer","tok2vec","ner","textcat_classify"]
+pipeline = ["tok2vec_small","tagger","parser","attribute_ruler","senter","tok2vec","ner","textcat_multilabel","textcat_classify"]
 batch_size = 1000
 disabled = []
 before_creation = null
@@ -31,13 +31,6 @@ factory = "attribute_ruler"
 scorer = {"@scorers":"spacy.attribute_ruler_scorer.v1"}
 validate = false
 
-[components.lemmatizer]
-factory = "lemmatizer"
-mode = "rule"
-model = null
-overwrite = false
-scorer = {"@scorers":"spacy.lemmatizer_scorer.v1"}
-
 [components.ner]
 factory = "ner"
 incorrect_spans_key = null
@@ -49,7 +42,7 @@ update_with_oracle_cut_size = 100
 @architectures = "spacy.TransitionBasedParser.v2"
 state_type = "ner"
 extra_state_tokens = false
-hidden_width = 128
+hidden_width = 256
 maxout_pieces = 3
 use_upper = true
 nO = null
@@ -129,7 +122,7 @@ upstream = "tok2vec"
 factory = "weighted_textcat"
 class_weights = [0.67,0.33]
 scorer = {"@scorers":"spacy.textcat_scorer.v2"}
-threshold = 0.0
+threshold = 0.3
 
 [components.textcat_classify.model]
 @architectures = "spacy.TextCatEnsemble.v2"
@@ -148,13 +141,47 @@ nO = null
 
 [components.textcat_classify.model.tok2vec.embed]
 @architectures = "spacy.MultiHashEmbed.v2"
-width = 128
+width = 256
 attrs = ["NORM","PREFIX","SUFFIX","SHAPE","ENT_TYPE"]
-rows = [2000,500,1000,500,500]
+rows = [4000,1000,2000,1000,1000]
 include_static_vectors = true
 
 [components.textcat_classify.model.tok2vec.encode]
 @architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 256
+window_size = 1
+maxout_pieces = 3
+depth = 8
+
+[components.textcat_multilabel]
+factory = "textcat_multilabel"
+scorer = {"@scorers":"spacy.textcat_multilabel_scorer.v2"}
+threshold = 0.3
+
+[components.textcat_multilabel.model]
+@architectures = "spacy.TextCatEnsemble.v2"
+nO = null
+
+[components.textcat_multilabel.model.linear_model]
+@architectures = "spacy.TextCatBOW.v3"
+exclusive_classes = false
+length = 262144
+ngram_size = 1
+no_output_layer = false
+nO = null
+
+[components.textcat_multilabel.model.tok2vec]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.textcat_multilabel.model.tok2vec.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = 128
+attrs = ["NORM","PREFIX","SUFFIX","SHAPE","ENT_TYPE"]
+rows = [2000,1000,1000,1000,1000]
+include_static_vectors = true
+
+[components.textcat_multilabel.model.tok2vec.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
 width = 128
 window_size = 1
 maxout_pieces = 3
@@ -170,7 +197,7 @@ factory = "tok2vec"
 @architectures = "spacy.MultiHashEmbed.v2"
 width = 256
 attrs = ["NORM","PREFIX","SUFFIX","SHAPE"]
-rows = [5000,1000,2500,2500]
+rows = [2000,1000,1000,1000]
 include_static_vectors = true
 
 [components.tok2vec.model.encode]
@@ -232,8 +259,8 @@ patience = 1000
 max_epochs = 0
 max_steps = 20000
 eval_frequency = 100
-frozen_components = ["tagger","attribute_ruler","parser","tok2vec","ner"]
-annotating_components = ["ner"]
+frozen_components = ["tok2vec","ner","textcat_multilabel"]
+annotating_components = ["ner","textcat_multilabel"]
 before_to_disk = null
 before_update = null
 
@@ -287,22 +314,21 @@ eps = 0.00000001
 learn_rate = 0.001
 
 [training.score_weights]
-tag_acc = 0.2
-dep_uas = 0.1
-dep_las = 0.1
+tag_acc = 0.25
+dep_uas = 0.12
+dep_las = 0.12
 dep_las_per_type = null
 sents_p = 0.0
 sents_r = 0.0
-sents_f = 0.2
-lemma_acc = 0.2
-ents_f = 0.1
+sents_f = 0.25
+ents_f = 0.12
 ents_p = 0.0
 ents_r = 0.0
 ents_per_type = null
-cats_score = 0.05
+cats_score = 0.06
 cats_score_desc = null
 cats_micro_p = null
-cats_micro_r = 0.05
+cats_micro_r = 0.06
 cats_micro_f = null
 cats_macro_p = null
 cats_macro_r = null

en_tako_query_analyzer-any-py3-none-any.whl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d8e1d83f7eaa00d4dfb3414bed716f22e27a59e12006528fa8d240ef1ac36e02
-size 620143055
+oid sha256:e8de7f18d2cc25ababfa5d8131bb5a3d3e4ba8415066bf97493f4b388da4c6e9
+size 649860398

meta.json

@@ -1,7 +1,7 @@
 {
   "lang":"en",
   "name":"tako_query_analyzer",
-  "version":"0.0.3",
+  "version":"0.0.4",
   "description":"",
   "author":"",
   "email":"",
@@ -120,9 +120,6 @@
     ],
     "attribute_ruler":[
 
-    ],
-    "lemmatizer":[
-
     ],
     "tok2vec":[
 
@@ -148,6 +145,16 @@
       "TIME",
       "WORK_OF_ART"
     ],
+    "textcat_multilabel":[
+      "Finance",
+      "Economics",
+      "Demographics",
+      "Health",
+      "Politics",
+      "Sports",
+      "Climate",
+      "Other"
+    ],
     "textcat_classify":[
       "ACCEPT",
       "REJECT"
@@ -159,9 +166,9 @@
     "parser",
     "attribute_ruler",
     "senter",
-    "lemmatizer",
     "tok2vec",
     "ner",
+    "textcat_multilabel",
     "textcat_classify"
   ],
   "components":[
@@ -170,9 +177,9 @@
     "parser",
     "attribute_ruler",
     "senter",
-    "lemmatizer",
     "tok2vec",
     "ner",
+    "textcat_multilabel",
     "textcat_classify"
   ],
   "disabled":[
@@ -183,28 +190,28 @@
     "ents_p":0.0,
     "ents_r":0.0,
     "ents_per_type":0.0,
-    "cats_score":0.8507157464,
+    "cats_score":0.8256157635,
     "cats_score_desc":"F (ACCEPT)",
-    "cats_micro_p":0.8531187123,
-    "cats_micro_r":0.8531187123,
-    "cats_micro_f":0.8531187123,
-    "cats_macro_p":0.853485064,
-    "cats_macro_r":0.8531187123,
-    "cats_macro_f":0.8530806455,
-    "cats_macro_auc":0.9167497439,
+    "cats_micro_p":0.823,
+    "cats_micro_r":0.823,
+    "cats_micro_f":0.823,
+    "cats_macro_p":0.8232909619,
+    "cats_macro_r":0.823,
+    "cats_macro_f":0.822960166,
+    "cats_macro_auc":0.899297,
     "cats_f_per_type":{
       "ACCEPT":{
-        "p":0.8648648649,
-        "r":0.8370221328,
-        "f":0.8507157464
+        "p":0.813592233,
+        "r":0.838,
+        "f":0.8256157635
       },
       "REJECT":{
-        "p":0.8421052632,
-        "r":0.8692152918,
-        "f":0.8554455446
+        "p":0.8329896907,
+        "r":0.808,
+        "f":0.8203045685
       }
     },
-    "textcat_classify_loss":0.9403656576
+    "textcat_classify_loss":2.1820269685
   },
   "requirements":[
 

ner/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:39d7b2908771676021059e0da7b19783682650daddfc0c44cd1a5a6ee8c90b8b
-size 768585
+oid sha256:a18b22cda1f4ebaa4786ba94b7baa9379ecca5475386ddd9c8b258717defc3a2
+size 2715735

ner/moves

@@ -1 +1 @@
-��moves��{"0":{},"1":{"ORG":32008,"GPE":3728,"PERSON":1105,"DATE":850,"WORK_OF_ART":686,"PRODUCT":585,"EVENT":283,"MONEY":214,"NORP":179,"STOCK_TICKER":156,"LAW":129,"LOC":111,"PERCENT":88,"FAC":75,"QUANTITY":60,"CARDINAL":57,"ORDINAL":42,"TIME":27,"LANGUAGE":25},"2":{"ORG":32008,"GPE":3728,"PERSON":1105,"DATE":850,"WORK_OF_ART":686,"PRODUCT":585,"EVENT":283,"MONEY":214,"NORP":179,"STOCK_TICKER":156,"LAW":129,"LOC":111,"PERCENT":88,"FAC":75,"QUANTITY":60,"CARDINAL":57,"ORDINAL":42,"TIME":27,"LANGUAGE":25},"3":{"ORG":32008,"GPE":3728,"PERSON":1105,"DATE":850,"WORK_OF_ART":686,"PRODUCT":585,"EVENT":283,"MONEY":214,"NORP":179,"STOCK_TICKER":156,"LAW":129,"LOC":111,"PERCENT":88,"FAC":75,"QUANTITY":60,"CARDINAL":57,"ORDINAL":42,"TIME":27,"LANGUAGE":25},"4":{"ORG":32008,"GPE":3728,"PERSON":1105,"DATE":850,"WORK_OF_ART":686,"PRODUCT":585,"EVENT":283,"MONEY":214,"NORP":179,"STOCK_TICKER":156,"LAW":129,"LOC":111,"PERCENT":88,"FAC":75,"QUANTITY":60,"CARDINAL":57,"ORDINAL":42,"TIME":27,"LANGUAGE":25,"":1},"5":{"":1}}�cfg��neg_key�
+��moves�H{"0":{},"1":{"GPE":17437,"ORG":17001,"PERSON":11972,"DATE":11609,"PRODUCT":8691,"QUANTITY":7887,"NORP":6312,"MONEY":4756,"WORK_OF_ART":4307,"EVENT":3116,"PERCENT":2783,"LAW":1974,"CARDINAL":1241,"LOC":1209,"STOCK_TICKER":842,"FAC":793,"TIME":690,"ORDINAL":222,"LANGUAGE":110},"2":{"GPE":17437,"ORG":17001,"PERSON":11972,"DATE":11609,"PRODUCT":8691,"QUANTITY":7887,"NORP":6312,"MONEY":4756,"WORK_OF_ART":4307,"EVENT":3116,"PERCENT":2783,"LAW":1974,"CARDINAL":1241,"LOC":1209,"STOCK_TICKER":842,"FAC":793,"TIME":690,"ORDINAL":222,"LANGUAGE":110},"3":{"GPE":17437,"ORG":17001,"PERSON":11972,"DATE":11609,"PRODUCT":8691,"QUANTITY":7887,"NORP":6312,"MONEY":4756,"WORK_OF_ART":4307,"EVENT":3116,"PERCENT":2783,"LAW":1974,"CARDINAL":1241,"LOC":1209,"STOCK_TICKER":842,"FAC":793,"TIME":690,"ORDINAL":222,"LANGUAGE":110},"4":{"GPE":17437,"ORG":17001,"PERSON":11972,"DATE":11609,"PRODUCT":8691,"QUANTITY":7887,"NORP":6312,"MONEY":4756,"WORK_OF_ART":4307,"EVENT":3116,"PERCENT":2783,"LAW":1974,"CARDINAL":1241,"LOC":1209,"STOCK_TICKER":842,"FAC":793,"TIME":690,"ORDINAL":222,"LANGUAGE":110,"":1},"5":{"":1}}�cfg��neg_key�

textcat_classify/cfg

@@ -3,6 +3,6 @@
     "ACCEPT",
     "REJECT"
   ],
-  "threshold":0.0,
+  "threshold":0.3,
   "positive_label":"ACCEPT"
 }

textcat_classify/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c65c611aa01b463b7f99116d0b1a53cd75effb9d0bac5febef70bf3b85f0b075
-size 8319359
+oid sha256:43f147a53892c746a8f24180b5870c414902c640b10f9ca2927ed2b91a8bf5fd
+size 36071049

textcat_multilabel/cfg

@@ -0,0 +1,13 @@
+{
+  "labels":[
+    "Finance",
+    "Economics",
+    "Demographics",
+    "Health",
+    "Politics",
+    "Sports",
+    "Climate",
+    "Other"
+  ],
+  "threshold":0.3
+}

textcat_multilabel/model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c238f141aa54e070622e6506599093673815490619c560feee160d93c37a502a
+size 15382175

tok2vec/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c8db1e5a93c4f955f990b7f6005b11c65ac6b9efa20f2c02291ac2013d06a203
-size 34434008
+oid sha256:f8f59ffdaeb93abafba597c7089b697ab873950f84b9b9bb5fe16022de2c9702
+size 28290008

vocab/strings.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:56c5dbd010224ad5111a8f04ef793ae9a7a7c24aa420362353a2e5a1bcd1df31
-size 10521360
+oid sha256:33316624b8e8ce9be5d4762929948a8c6bef9a9d6a138aae4110f40cd54858a3
+size 10662076