adjust script
Browse files
run.sh
CHANGED
|
@@ -1,5 +1,6 @@
|
|
| 1 |
#!/usr/bin/env bash
|
| 2 |
python3 run_mlm_flax.py \
|
|
|
|
| 3 |
--output_dir="./" \
|
| 4 |
--model_type="roberta" \
|
| 5 |
--config_name="./" \
|
|
@@ -7,8 +8,9 @@ python3 run_mlm_flax.py \
|
|
| 7 |
--dataset_name="oscar" \
|
| 8 |
--dataset_config_name="unshuffled_deduplicated_th" \
|
| 9 |
--max_seq_length="128" \
|
| 10 |
-
--
|
| 11 |
-
--
|
|
|
|
| 12 |
--learning_rate="3e-4" \
|
| 13 |
--warmup_steps="1000" \
|
| 14 |
--overwrite_output_dir \
|
|
|
|
| 1 |
#!/usr/bin/env bash
|
| 2 |
python3 run_mlm_flax.py \
|
| 3 |
+
--model_name_or_path="flax_model.msgpack" \
|
| 4 |
--output_dir="./" \
|
| 5 |
--model_type="roberta" \
|
| 6 |
--config_name="./" \
|
|
|
|
| 8 |
--dataset_name="oscar" \
|
| 9 |
--dataset_config_name="unshuffled_deduplicated_th" \
|
| 10 |
--max_seq_length="128" \
|
| 11 |
+
--preprocessing_num_workers="64" \
|
| 12 |
+
--per_device_train_batch_size="64" \
|
| 13 |
+
--per_device_eval_batch_size="64" \
|
| 14 |
--learning_rate="3e-4" \
|
| 15 |
--warmup_steps="1000" \
|
| 16 |
--overwrite_output_dir \
|