Training in progress, step 500

config.json

@@ -5,12 +5,9 @@
     "DistilBertForSequenceClassification"
   ],
   "attention_dropout": 0.1,
-  "attention_dropout_prob": 0.3,
-  "dim": 512,
+  "dim": 768,
   "dropout": 0.1,
-  "finetuning_task": "sequence-classification",
   "hidden_dim": 3072,
-  "hidden_dropout_prob": 0.3,
   "id2label": {
     "0": "LABEL_0",
     "1": "LABEL_1",
@@ -28,8 +25,8 @@
   },
   "max_position_embeddings": 512,
   "model_type": "distilbert",
-  "n_heads": 8,
-  "n_layers": 4,
+  "n_heads": 12,
+  "n_layers": 6,
   "pad_token_id": 0,
   "problem_type": "single_label_classification",
   "qa_dropout": 0.1,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f4b8aeecd3ac88089b512675c9779e7f0932a900cd228c37f46115ae83c5471
-size 131862908
+oid sha256:932fb44a384a41290a1751ec0b675febc84060fd8ab2b8befefb30a206d5ef48
+size 267841796

runs/Mar26_15-03-04_0a1f98dac3fc/events.out.tfevents.1711465418.0a1f98dac3fc.2285.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1d517dd3c38f9f870df6f0578334d2ec69f74ea83c905fdc8a85d3b0d4902e0
+size 4184

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:99be8c0cc4520633905a0e609f49d8cf7732be3e3036c8272ebb42980aee5683
+oid sha256:2c47cb7dc9f5aebc774b62ef788842f19095bff488cf4df93f668bccf2846966
 size 4984