Training in progress, step 500

config.json

@@ -4,10 +4,13 @@
   "architectures": [
     "DistilBertForSequenceClassification"
   ],
-  "attention_dropout": 0.2,
-  "dim": 768,
-  "dropout": 0.4,
+  "attention_dropout": 0.1,
+  "attention_dropout_prob": 0.3,
+  "dim": 512,
+  "dropout": 0.1,
+  "finetuning_task": "sequence-classification",
   "hidden_dim": 3072,
+  "hidden_dropout_prob": 0.3,
   "id2label": {
     "0": "LABEL_0",
     "1": "LABEL_1",
@@ -25,8 +28,8 @@
   },
   "max_position_embeddings": 512,
   "model_type": "distilbert",
-  "n_heads": 12,
-  "n_layers": 6,
+  "n_heads": 8,
+  "n_layers": 4,
   "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 sha256:0cb055690242908997dc35e0eb1a0fd95e29ae4748758e570b6245d1c5b2e133
-size 267841796
+oid sha256:3a35f488ef6ddfbf78266827428713264c66ad25c359ca12312d064d2808af85
+size 131862908

runs/Mar26_14-57-25_0a1f98dac3fc/events.out.tfevents.1711465082.0a1f98dac3fc.2285.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de0ee52ec8252e356d772904f1b1606208d2c9443d31fb4f614bc67c2957101b
+size 4184

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9321b90eb3ba4ae162e30cbac61fa8cddc18259aaf85f84da06261ea55eacf00
+oid sha256:99be8c0cc4520633905a0e609f49d8cf7732be3e3036c8272ebb42980aee5683
 size 4984