Maltehb
commited on
Commit
•
59bb3dd
1
Parent(s):
6eb3e4b
continued 111001 training steps
Browse files- continue_run_mlm_flax_stream.sh +7 -7
- events.out.tfevents.1625999437.t1v-n-58ed2201-w-0.98932.3.v2 → events.out.tfevents.1626175943.t1v-n-58ed2201-w-0.155891.3.v2 +2 -2
- flax_model.msgpack +1 -1
- src/{scandinavian_run_mlm_flax.py → run_mlm_flax.py} +0 -0
- src/{run_mlm_flax_stream.py → scandinavian_run_mlm_flax_stream.py} +0 -0
continue_run_mlm_flax_stream.sh
CHANGED
@@ -1,22 +1,22 @@
|
|
1 |
-
export MODEL_DIR=/home/Z6HJB/roberta-
|
2 |
|
3 |
source /home/Z6HJB/test/bin/activate
|
4 |
|
5 |
-
python3 ./src/
|
6 |
--model_name_or_path="${MODEL_DIR}" \
|
7 |
--output_dir="${MODEL_DIR}" \
|
8 |
--tokenizer_name="${MODEL_DIR}" \
|
9 |
--dataset_name="mc4" \
|
10 |
--dataset_config_name="unshuffled_deduplicated_en" \
|
11 |
-
--max_seq_length="
|
12 |
-
--per_device_train_batch_size="
|
13 |
-
--per_device_eval_batch_size="
|
14 |
-
--learning_rate="
|
15 |
--warmup_steps="1000" \
|
16 |
--overwrite_output_dir \
|
17 |
--adam_beta1="0.9" \
|
18 |
--adam_beta2="0.98" \
|
19 |
-
--num_train_steps="
|
20 |
--num_eval_samples="5000" \
|
21 |
--save_steps="1000" \
|
22 |
--logging_steps="250" \
|
|
|
1 |
+
export MODEL_DIR=/home/Z6HJB/train-roberta-base-scandinavian/roberta-base-scandinavian/
|
2 |
|
3 |
source /home/Z6HJB/test/bin/activate
|
4 |
|
5 |
+
python3 ./src/scandinavian_run_mlm_flax_stream.py \
|
6 |
--model_name_or_path="${MODEL_DIR}" \
|
7 |
--output_dir="${MODEL_DIR}" \
|
8 |
--tokenizer_name="${MODEL_DIR}" \
|
9 |
--dataset_name="mc4" \
|
10 |
--dataset_config_name="unshuffled_deduplicated_en" \
|
11 |
+
--max_seq_length="128" \
|
12 |
+
--per_device_train_batch_size="128" \
|
13 |
+
--per_device_eval_batch_size="128" \
|
14 |
+
--learning_rate="1.087e-6" \
|
15 |
--warmup_steps="1000" \
|
16 |
--overwrite_output_dir \
|
17 |
--adam_beta1="0.9" \
|
18 |
--adam_beta2="0.98" \
|
19 |
+
--num_train_steps="200000" \
|
20 |
--num_eval_samples="5000" \
|
21 |
--save_steps="1000" \
|
22 |
--logging_steps="250" \
|
events.out.tfevents.1625999437.t1v-n-58ed2201-w-0.98932.3.v2 → events.out.tfevents.1626175943.t1v-n-58ed2201-w-0.155891.3.v2
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8ffed2c9c5c4999da79f6278e1e5747558ffbb3f39f236eb56271bd130c5447b
|
3 |
+
size 16627546
|
flax_model.msgpack
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 498796983
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7ef047be1c30de5fd9b499f1d217091eb0baf194be896c002d4c2e87939639c7
|
3 |
size 498796983
|
src/{scandinavian_run_mlm_flax.py → run_mlm_flax.py}
RENAMED
File without changes
|
src/{run_mlm_flax_stream.py → scandinavian_run_mlm_flax_stream.py}
RENAMED
File without changes
|