Saving weights and logs of step 10000

config.json

@@ -19,7 +19,7 @@
   "num_hidden_layers": 12,
   "pad_token_id": 1,
   "position_embedding_type": "absolute",
-  "transformers_version": "4.15.0.dev0",
+  "transformers_version": "4.17.0.dev0",
   "type_vocab_size": 1,
   "use_cache": true,
   "vocab_size": 50265

events.out.tfevents.1645234163.t1v-n-c7babd08-w-0.12401.0.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7cc932fe59bb4e3ac79d762900dfba08b283e25c9a55f10def02bea6d991fb9a
+size 1470136

flax_model.msgpack

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:94ea9cc438cd8643a14be164330250e50a1680d61abb19853c66bb239d50ff52
+oid sha256:867f735869af4b211d6a8c011e0a616b2c9c323b7ebd7459c5667e4ec8019514
 size 498796983

run_512.sh

@@ -14,7 +14,7 @@ python run_mlm_flax.py \
     --learning_rate="0.00005" \
     --warmup_steps="150000" \
     --overwrite_output_dir \
-    --num_train_epochs="10" \
+    --num_train_epochs="20" \
     --adam_beta1="0.9" \
     --adam_beta2="0.98" \
     --adam_epsilon="1e-6" \