Update spaCy pipeline

README.md

@@ -7,36 +7,3 @@ model-index:
 - name: en_pipeline
   results: []
 ---
-| Feature | Description |
-| --- | --- |
-| **Name** | `en_pipeline` |
-| **Version** | `0.0.0` |
-| **spaCy** | `>=3.4.1,<3.5.0` |
-| **Default Pipeline** | `transformer`, `spancat` |
-| **Components** | `transformer`, `spancat` |
-| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
-| **Sources** | n/a |
-| **License** | n/a |
-| **Author** | [n/a]() |
-
-### Label Scheme
-
-<details>
-
-<summary>View label scheme (1 labels for 1 components)</summary>
-
-| Component | Labels |
-| --- | --- |
-| **`spancat`** | `COMP` |
-
-</details>
-
-### Accuracy
-
-| Type | Score |
-| --- | --- |
-| `SPANS_SC_F` | 75.30 |
-| `SPANS_SC_P` | 73.20 |
-| `SPANS_SC_R` | 77.52 |
-| `TRANSFORMER_LOSS` | 1705.81 |
-| `SPANCAT_LOSS` | 51292.83 |

config.cfg

@@ -32,7 +32,7 @@ threshold = 0.5
 
 [components.spancat.model.reducer]
 @layers = "spacy.mean_max_reducer.v1"
-hidden_size = 128
+hidden_size = 256
 
 [components.spancat.model.scorer]
 @layers = "spacy.LinearLogistic.v1"
@@ -90,16 +90,16 @@ limit = 0
 augmenter = null
 
 [training]
-accumulate_gradient = 3
+accumulate_gradient = 5
 dev_corpus = "corpora.dev"
 train_corpus = "corpora.train"
 frozen_components = []
-dropout = 0.1
+dropout = 0.0
+patience = 10000
 max_epochs = 0
 max_steps = 25000
 seed = ${system.seed}
 gpu_allocator = ${system.gpu_allocator}
-patience = 5000
 eval_frequency = 200
 annotating_components = []
 before_to_disk = null
@@ -128,8 +128,8 @@ use_averages = false
 [training.optimizer.learn_rate]
 @schedules = "warmup_linear.v1"
 warmup_steps = 250
-total_steps = 25000
-initial_rate = 0.00005
+total_steps = 20000
+initial_rate = 0.00003
 
 [training.score_weights]
 spans_sc_f = 1.0

en_pipeline-any-py3-none-any.whl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c454770b2f8026fdd2273147e62f921ca44902c66ae63b52840413b64aea77c5
-size 434870984
+oid sha256:1aa9b012ca42a49356294b686911d7ef0410e962cfc92d342ec65e42bc6c8924
+size 439235489

meta.json

@@ -1,9 +1,9 @@
 {
   "lang":"en",
   "name":"pipeline",
-  "version":"0.0.0",
-  "description":"",
-  "author":"",
+  "version":"0.0.1",
+  "description":"Span Categorizer for Company names",
+  "author":"oMateos2020",
   "email":"",
   "url":"",
   "license":"",
@@ -35,11 +35,11 @@
 
   ],
   "performance":{
-    "spans_sc_f":0.7530064755,
-    "spans_sc_p":0.7320143885,
-    "spans_sc_r":0.7752380952,
-    "transformer_loss":17.0580949257,
-    "spancat_loss":512.928282738
+    "spans_sc_f":0.7580340265,
+    "spans_sc_p":0.7523452158,
+    "spans_sc_r":0.7638095238,
+    "transformer_loss":2.7061057311,
+    "spancat_loss":174.07806898
   },
   "requirements":[
     "spacy-transformers>=1.1.8,<1.2.0"

spancat/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:69570e3c19d152de5c34e73dbe13980f3c07dcc4af4050b1ed287b235b1fe5f1
-size 4723491
+oid sha256:fe261449c76ac7a8f4082659963e41b8583f60ea0f838dd4a13cdf90d10d5561
+size 9445168

transformer/model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d7016162ccd3d0c214477c2dde839e5533b864f54b8d4e8c9ef12fa7183475ee
+oid sha256:968c50b0d5e64db1aca2096480f96164f0c2d29a961c175ac9fef4c4cd8a2990
 size 502030632