Training in progress, epoch 1

config.json

@@ -1,5 +1,6 @@
 {
-  "_name_or_path": "google/bert_uncased_L-6_H-768_A-12",
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
   "adapters": {
     "adapters": {},
     "config_map": {},
@@ -7,34 +8,32 @@
     "fusions": {}
   },
   "architectures": [
-    "BertForSequenceClassification"
+    "DistilBertForSequenceClassification"
   ],
-  "attention_probs_dropout_prob": 0.1,
-  "classifier_dropout": null,
-  "hidden_act": "gelu",
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
   "id2label": {
     "0": "entailment",
     "1": "not_entailment"
   },
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
   "label2id": {
     "entailment": "0",
     "not_entailment": "1"
   },
-  "layer_norm_eps": 1e-12,
   "max_position_embeddings": 512,
-  "model_type": "bert",
-  "num_attention_heads": 12,
-  "num_hidden_layers": 6,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
   "pad_token_id": 0,
-  "position_embedding_type": "absolute",
   "problem_type": "single_label_classification",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
   "torch_dtype": "float32",
   "transformers_version": "4.21.3",
-  "type_vocab_size": 2,
-  "use_cache": true,
   "vocab_size": 30522
 }

logs/1663304779.8945827/events.out.tfevents.1663304779.cd02b94e3af8.5814.4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f2d269b5491445066f65ba4f705f7cd1d1b306577043233e33286a40b3fa5d3
+size 5437

logs/1663305699.3569372/events.out.tfevents.1663305699.cd02b94e3af8.5814.6

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a47acabe3a336cbef74ba2d3f54a8cf9cfde64db5fe0ba1de3d398f773450813
+size 5437

logs/events.out.tfevents.1663302864.cd02b94e3af8.5814.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e94be61ddd1d218cce0d09a76d1042bfe890598f246aca9e30466f056c2923c8
-size 4631
+oid sha256:256212836eec81d4f0dcdac4a464f8298fc88327d9c401038d52bc9e1477590c
+size 14087

logs/events.out.tfevents.1663303742.cd02b94e3af8.5814.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8d92f1e085e6affb8f7bb8e04c87b77e4d6adba4a49c23345263f53a80eeae4
+size 363

logs/events.out.tfevents.1663304773.cd02b94e3af8.5814.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9cd8c1b61017506c2b3b465fb28524e459fec81bc4a4cef144b7ced0e5176a1a
+size 4721

logs/events.out.tfevents.1663305696.cd02b94e3af8.5814.5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17d7e47f034bb29b015a104efba4d4cdaf9ab6aa948a47f77a7433a95e3282dd
+size 4586

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1aa1ae283b84f37efb07ebfc01393aa69d8e59ca68deb868b0103cd8c8856149
-size 267864813
+oid sha256:de8f16a0e2beeb1c437a68937553020bedfe3b7889b4f9404a6546bb071dddae
+size 267854321