Training in progress, epoch 1

config.json

@@ -8,13 +8,13 @@
   "gradient_checkpointing": false,
   "hidden_act": "gelu",
   "hidden_dropout_prob": 0.1,
-  "hidden_size": 384,
+  "hidden_size": 512,
   "initializer_range": 0.02,
-  "intermediate_size": 1536,
+  "intermediate_size": 3072,
   "layer_norm_eps": 1e-12,
   "max_position_embeddings": 512,
   "model_type": "bert",
-  "num_attention_heads": 12,
+  "num_attention_heads": 16,
   "num_hidden_layers": 12,
   "pad_token_id": 0,
   "position_embedding_type": "absolute",

logs/events.out.tfevents.1727810100.spartan-gpgpu147.hpc.unimelb.edu.au.4989.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d77c90c3b8347f0b1c5d43d0f4f13fa960ef98581c7392d678194488f3a9bac9
-size 10684
+oid sha256:bf50add9e54d3037762e2a194fe6c728901e27985ed4e58c5cc80c9ff9d0ccd8
+size 11638

logs/events.out.tfevents.1727810427.spartan-gpgpu147.hpc.unimelb.edu.au.4989.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a8e6dbf7ca7e5f0081ae24d0af87765a51cdc8a5b2ab42eaee122545c3d7945
+size 477

logs/events.out.tfevents.1727810586.spartan-gpgpu147.hpc.unimelb.edu.au.14514.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64bcecbf9029ddc8d16b59633d0e9cea68724dc74dda74da008e6f369e21ab22
+size 5906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d0db97217893482d5bf195795abce29e626fe746e74f7be8ac292bad42a0b132
-size 133466376
+oid sha256:fcce87620ccb84e18c700b7609b996f76d2bdcc79e15120f1276be706c2d1fb6
+size 266339248