Saving weights and logs of step 5001

events.out.tfevents.1638603525.t1v-n-4e27a527-w-0.162184.0.v2 → events.out.tfevents.1638863641.t1v-n-4e27a527-w-0.401059.0.v2

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:066eb6e8b1b557aed8073085fcce8190bd506e82c502e797c14011237c9f3892
-size 6674060
+oid sha256:e76f35263e83fa73b2fbd6257c39113ea9b8520f70b1a428d549d4b2a2e03f51
+size 735136

flax_model.msgpack

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:441020af2b64901dc28919af7331a98efde9bd3cc6d49199f8ea311a5322788f
+oid sha256:eaa7a07462cab6b1259e745b10570b9d2914167560e16a494c6fbfcf4e6b5c17
 size 498796983

run_flax_stream_no_workers_no_adafactor.py

@@ -0,0 +1,20 @@
+./run_mlm_flax_stream.py \
+    --output_dir="./" \
+    --model_type="roberta" \
+    --config_name="./" \
+    --tokenizer_name="./" \
+    --dataset_name="NbAiLab/NCC_small" \
+    --max_seq_length="128" \
+    --weight_decay="0.01" \
+    --per_device_train_batch_size="232" \
+    --per_device_eval_batch_size="232" \
+    --learning_rate="6e-4" \
+    --warmup_steps="5000" \
+    --overwrite_output_dir \
+    --num_train_epochs="10" \
+    --adam_beta1="0.9" \
+    --adam_beta2="0.98" \
+    --logging_steps="5000" \
+    --save_steps="5000" \
+    --eval_steps="5000" \
+    --push_to_hub