Step 19519

README.md

@@ -60,7 +60,7 @@ WIP
 
 ## Training data
 
-The model was initialized with pre-trained weights of [sahajBERT](https://huggingface.co/neuropark/sahajBERT-NER) at step 19169 and trained on the bengali split of [WikiANN ](https://huggingface.co/datasets/wikiann)
+The model was initialized with pre-trained weights of [sahajBERT](https://huggingface.co/neuropark/sahajBERT-NER) at step 19519 and trained on the bengali split of [WikiANN ](https://huggingface.co/datasets/wikiann)
 
 ## Training procedure
 
@@ -73,15 +73,15 @@ Coming soon!
 
 ## Eval results
 
-accuracy: 0.9728682170542635
+accuracy: 0.9756540697674418
 
-f1: 0.9515892420537897
+f1: 0.9570102589154861
 
-loss: 0.19455023109912872
+loss: 0.13705264031887054
 
-precision: 0.9474196689386563
+precision: 0.9518950437317785
 
-recall: 0.9557956777996071
+recall: 0.962180746561886
 
 
 

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "albertvillanova/autonlp-wikiann-entity_extraction-0d675eb-1281107",
+  "_name_or_path": "albertvillanova/autonlp-wikiann-entity_extraction-1e67664-1301123",
   "_num_labels": 7,
   "architectures": [
     "AlbertForTokenClassification"
@@ -36,7 +36,7 @@
     "6": 6
   },
   "layer_norm_eps": 1e-12,
-  "max_length": 256,
+  "max_length": 96,
   "max_position_embeddings": 512,
   "model_type": "albert",
   "net_structure_type": 0,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2d49012ab934ab252223f215d50f7117057f11f56fcc82874fc90b4259f63c4d
+oid sha256:6b4ef61f0bfac93755c2de19e8030ed55289ccb917aafbadf37ff66b5abbf682
 size 67605209