Update spaCy pipeline

README.md

@@ -13,13 +13,13 @@ model-index:
     metrics:
     - name: NER Precision
       type: precision
-      value: 0.1791044776
+      value: 1.0
     - name: NER Recall
       type: recall
-      value: 0.3333333333
+      value: 1.0
     - name: NER F Score
       type: f_score
-      value: 0.2330097087
+      value: 1.0
 ---
 | Feature | Description |
 | --- | --- |
@@ -49,8 +49,8 @@ model-index:
 
 | Type | Score |
 | --- | --- |
-| `ENTS_F` | 23.30 |
-| `ENTS_P` | 17.91 |
-| `ENTS_R` | 33.33 |
-| `TRANSFORMER_LOSS` | 9550392.37 |
-| `NER_LOSS` | 2471774.14 |
+| `ENTS_F` | 100.00 |
+| `ENTS_P` | 100.00 |
+| `ENTS_R` | 100.00 |
+| `TRANSFORMER_LOSS` | 0.00 |
+| `NER_LOSS` | 0.00 |

config.cfg

@@ -1,6 +1,6 @@
 [paths]
 train = "./drive/MyDrive/Product IBM/train.spacy"
-dev = "./drive/MyDrive/Product IBM/test.spacy"
+dev = "./drive/MyDrive/Product IBM/train.spacy"
 vectors = null
 init_tok2vec = null
 
@@ -92,7 +92,7 @@ gpu_allocator = ${system.gpu_allocator}
 dropout = 0.1
 patience = 1600
 max_epochs = 0
-max_steps = 2000
+max_steps = 1000
 eval_frequency = 200
 frozen_components = []
 annotating_components = []

en_Product_Catalog-any-py3-none-any.whl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2dfc30086b12283630b17f5fa4f9859e658ab058c10af3191f9c2b28990336b1
-size 421959211
+oid sha256:a93a8a8c356b9d2dd1b85a86d90263f53c02f3bced93afe3587590db7c10c187
+size 424180572

meta.json

@@ -40,48 +40,43 @@
 
   ],
   "performance":{
-    "ents_f":0.2330097087,
-    "ents_p":0.1791044776,
-    "ents_r":0.3333333333,
+    "ents_f":1.0,
+    "ents_p":1.0,
+    "ents_r":1.0,
     "ents_per_type":{
       "Product":{
-        "p":0.1176470588,
-        "r":0.25,
-        "f":0.16
+        "p":1.0,
+        "r":1.0,
+        "f":1.0
       },
-      "Metering":{
-        "p":0.2380952381,
-        "r":0.625,
-        "f":0.3448275862
+      "Lubricant":{
+        "p":1.0,
+        "r":1.0,
+        "f":1.0
       },
-      "Reservoir":{
-        "p":0.0,
-        "r":0.0,
-        "f":0.0
+      "Metering":{
+        "p":1.0,
+        "r":1.0,
+        "f":1.0
       },
       "Operating":{
-        "p":0.3076923077,
-        "r":0.75,
-        "f":0.4363636364
+        "p":1.0,
+        "r":1.0,
+        "f":1.0
       },
-      "Lubricant":{
-        "p":0.0,
-        "r":0.0,
-        "f":0.0
+      "Reservoir":{
+        "p":1.0,
+        "r":1.0,
+        "f":1.0
       },
       "PAGE":{
-        "p":0.0,
-        "r":0.0,
-        "f":0.0
-      },
-      "Page":{
-        "p":0.0,
-        "r":0.0,
-        "f":0.0
+        "p":1.0,
+        "r":1.0,
+        "f":1.0
       }
     },
-    "transformer_loss":95503.9236874724,
-    "ner_loss":24717.7414470388
+    "transformer_loss":0.0000308286,
+    "ner_loss":0.0000346953
   },
   "requirements":[
     "spacy-transformers>=1.3.4,<1.4.0"

transformer/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c69b47953727318e86bb2f42cdf017eb39b85c76e7ada68add0832bac7cc645a
+oid sha256:2274c1ec3f56cc7512c11a4bd8b991fe76d8cbe2412c37236430a833ab26a7d2
 size 502027402

vocab/strings.json

@@ -63,24 +63,17 @@
   "):",
   "*",
   ",",
-  ",016-0,05",
-  ",016-013",
   ",03",
-  ",05",
   ",20",
   ",25",
   ",38",
   ",97",
-  ",ddd-d,dd",
-  ",ddd-ddd",
   "-",
   "-((",
   "-))",
   "-/",
   "-0",
   "-3",
-  "-32",
-  "-33",
   "-8",
   "-D",
   "-O",
@@ -140,7 +133,6 @@
   "0,0018",
   "0,003",
   "0,03",
-  "0,05",
   "0,052",
   "0,066",
   "0,20",
@@ -149,22 +141,6 @@
   "0,5",
   "0,6",
   "0.0",
-  "0.001",
-  "0.001-0.0030",
-  "0.001-0.008",
-  "0.0012",
-  "0.0012-0.0305",
-  "0.0030",
-  "0.0030-0.0244",
-  "0.0061",
-  "0.0061-0.0793",
-  "0.0079",
-  "0.0079-0.0799",
-  "0.008",
-  "0.0244",
-  "0.0305",
-  "0.0793",
-  "0.0799",
   "0.086",
   "0.12",
   "0.13",
@@ -182,27 +158,11 @@
   "0.91",
   "0.o",
   "000",
-  "001",
-  "002",
-  "002-050",
   "003",
-  "008",
-  "012",
-  "013",
-  "016",
   "018",
-  "02",
-  "02-050",
-  "030",
-  "040",
   "045",
-  "05",
-  "05-040",
-  "050",
   "052",
-  "061",
   "066",
-  "079",
   "086",
   "0_0",
   "0_o",
@@ -215,11 +175,8 @@
   "1.83",
   "1.88",
   "10",
-  "10-130",
   "100",
-  "100-315",
   "1000",
-  "103",
   "10a.m",
   "10a.m.",
   "10p.m",
@@ -230,45 +187,20 @@
   "11p.m",
   "11p.m.",
   "12",
-  "1200",
-  "1200-3500",
-  "124",
-  "126",
-  "127",
-  "127-240",
-  "128",
   "12a.m",
   "12a.m.",
   "12p.m",
   "12p.m.",
-  "13",
-  "13-131",
-  "130",
-  "131",
-  "132",
-  "133",
-  "134",
-  "136",
   "14",
   "140",
-  "140-300",
-  "1450",
-  "1450-4570",
-  "1453",
   "15",
-  "150",
-  "153",
   "156",
   "16",
   "167",
   "17",
   "18",
-  "180",
   "1812",
   "1826",
-  "183",
-  "1850",
-  "1850-3500",
   "1a.m",
   "1a.m.",
   "1p.m",
@@ -279,44 +211,31 @@
   "2.39",
   "20",
   "200",
-  "2030",
-  "2030-4350",
   "2156",
-  "2180",
-  "2183",
   "22",
   "23",
   "24",
   "24,6",
   "240",
-  "244",
   "250",
   "26",
   "28",
   "283167",
   "289",
   "29",
-  "2HV",
   "2a.m",
   "2a.m.",
-  "2hv",
   "2p.m",
   "2p.m.",
   "3",
   "3.96",
   "30",
-  "300",
-  "303",
-  "305",
   "31",
-  "315",
   "33",
   "333",
   "34",
   "35",
-  "350",
   "3500",
-  "353",
   "36",
   "38",
   "392",
@@ -331,17 +250,10 @@
   "4",
   "4,6",
   "40",
-  "400",
   "405",
-  "41",
   "42",
   "435",
-  "4350",
-  "4353",
   "45",
-  "450",
-  "453",
-  "4570",
   "462",
   "4:8",
   "4a.m",
@@ -365,8 +277,6 @@
   "5p.m.",
   "6",
   "6-2",
-  "60",
-  "600",
   "652",
   "667",
   "676",
@@ -378,8 +288,6 @@
   "7",
   "715",
   "74",
-  "793",
-  "799",
   "7a.m",
   "7a.m.",
   "7p.m",
@@ -396,12 +304,8 @@
   "826",
   "82676",
   "82885",
-  "83",
-  "83-240",
   "83667",
   "846",
-  "850",
-  "870",
   "876715",
   "885",
   "886",
@@ -544,7 +448,6 @@
   "Aug.",
   "August",
   "B",
-  "B-doser",
   "Bil",
   "Bros",
   "Bros.",
@@ -677,9 +580,6 @@
   "Ky",
   "Ky.",
   "L",
-  "L-1",
-  "LG",
-  "LG-doser",
   "La",
   "La.",
   "Let",
@@ -800,7 +700,6 @@
   "PW",
   "Pa",
   "Pa.",
-  "Page",
   "Pal",
   "Pennsylvania",
   "Ph",
@@ -818,10 +717,6 @@
   "S",
   "S.C.",
   "S30",
-  "SL-1",
-  "SL-32",
-  "SL-32HV",
-  "SL-33",
   "Sen",
   "Sen.",
   "Sep",
@@ -859,7 +754,6 @@
   "Those",
   "V",
   "V.V",
-  "VR",
   "V_V",
   "Va",
   "Va.",
@@ -897,7 +791,6 @@
   "X++",
   "X-ddd",
   "X-ddd-d",
-  "X-xxxx",
   "X.",
   "X.X",
   "X.X.",
@@ -906,10 +799,6 @@
   "XD",
   "XDD",
   "XX",
-  "XX-d",
-  "XX-dd",
-  "XX-ddXX",
-  "XX-xxxx",
   "XXX",
   "XXX-ddd",
   "XXXX",
@@ -983,7 +872,6 @@
   "adjustable",
   "adm",
   "adm.",
-  "age",
   "ai",
   "ak",
   "ak.",
@@ -1005,16 +893,12 @@
   "ark",
   "ark.",
   "as",
-  "asL",
   "ash",
-  "asl",
   "ass",
   "aug",
   "aug.",
   "ave",
-  "ax.",
   "b",
-  "b-doser",
   "b.",
   "because",
   "bil",
@@ -1063,10 +947,6 @@
   "d.d",
   "d.dd",
   "d.ddd",
-  "d.ddd-d.ddd",
-  "d.ddd-d.dddd",
-  "d.dddd",
-  "d.dddd-d.dddd",
   "d.x",
   "d:d",
   "dX",
@@ -1075,12 +955,9 @@
   "dare",
   "dd",
   "dd,d",
-  "dd-ddd",
   "dd.dd",
   "ddd",
-  "ddd-ddd",
   "dddd",
-  "dddd-dddd",
   "ddx.x",
   "ddx.x.",
   "dec",
@@ -1095,7 +972,6 @@
   "doin'",
   "doing",
   "doin\u2019",
-  "doser",
   "dr",
   "dr.",
   "dx.x",
@@ -1190,7 +1066,6 @@
   "ing",
   "inn",
   "in\u2019",
-  "ion",
   "is",
   "is.",
   "iss",
@@ -1219,15 +1094,12 @@
   "ky",
   "ky.",
   "l",
-  "l-1",
   "l.",
   "la",
   "la.",
   "let",
   "let's",
   "let\u2019s",
-  "lg",
-  "lg-doser",
   "lif",
   "ll",
   "ll.",
@@ -1247,9 +1119,6 @@
   "mar.",
   "mass",
   "mass.",
-  "max",
-  "max.",
-  "max.150",
   "may",
   "ma\u2019am",
   "mcp",
@@ -1378,7 +1247,6 @@
   "pm",
   "pps30",
   "pr.",
-  "prelubrication",
   "product",
   "prof",
   "prof.",
@@ -1412,7 +1280,6 @@
   "sep.",
   "sept",
   "sept.",
-  "ser",
   "sh.",
   "sha",
   "shall",
@@ -1420,10 +1287,6 @@
   "she's",
   "she\u2019s",
   "should",
-  "sl-1",
-  "sl-32",
-  "sl-32hv",
-  "sl-33",
   "somethin",
   "somethin'",
   "something",
@@ -1478,7 +1341,6 @@
   "va.",
   "ve",
   "vin",
-  "vr",
   "vs",
   "vs.",
   "w",
@@ -1518,7 +1380,6 @@
   "x'xxxx",
   "x-ddd",
   "x-ddd-d",
-  "x-xxxx",
   "x.",
   "x.X",
   "x.d",
@@ -1540,17 +1401,10 @@
   "xx'",
   "xx'x",
   "xx'xx",
-  "xx-d",
-  "xx-dd",
-  "xx-ddxx",
-  "xx-xxxx",
   "xx.",
-  "xxX",
   "xxx",
   "xxx'x",
   "xxx-ddd",
-  "xxx.",
-  "xxx.ddd",
   "xxx/xx",
   "xxxdd",
   "xxxx",