Saving weights and logs of epoch 1
Browse files
events.out.tfevents.1625649517.t1v-n-b95d739e-w-0.10251.3.v2
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:1049dd09f81c1c4446d4258b96c037dc61968c1e94aa392f0e17259daeea6bff
|
3 |
-
size 19919742
|
|
|
|
|
|
|
|
events.out.tfevents.1625719131.t1v-n-b95d739e-w-0.74969.3.v2 → events.out.tfevents.1625741152.t1v-n-b95d739e-w-0.95396.3.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:a393b11affb3db2764adb97b61578f82378808344c246f16ecff4a67ebc47e77
|
3 |
+
size 3581817
|
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 sha256:8a309f76b746c4515f4ef732ac73e29dca90a8c6e074f7b46936375e1980fa30
|
3 |
size 498796983
|
nohup.out
CHANGED
@@ -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 sha256:2a8b1dee60938c59f43e6e76b676ef82307e1b911c45cb3a172607d950505880
|
3 |
+
size 2613881
|
run.sh
CHANGED
@@ -8,15 +8,12 @@ python3 run_mlm_flax.py \
|
|
8 |
--dataset_name="oscar" \
|
9 |
--dataset_config_name="unshuffled_deduplicated_id" \
|
10 |
--max_seq_length="128" \
|
11 |
-
--
|
12 |
--per_device_train_batch_size="128" \
|
13 |
--per_device_eval_batch_size="128" \
|
14 |
-
--learning_rate="
|
15 |
--warmup_steps="1000" \
|
16 |
--overwrite_output_dir \
|
17 |
-
--
|
18 |
-
--num_train_epochs="10" \
|
19 |
-
--adam_beta1="0.9" \
|
20 |
-
--adam_beta2="0.98" \
|
21 |
--push_to_hub
|
22 |
|
|
|
8 |
--dataset_name="oscar" \
|
9 |
--dataset_config_name="unshuffled_deduplicated_id" \
|
10 |
--max_seq_length="128" \
|
11 |
+
--preprocessing_num_workers="64" \
|
12 |
--per_device_train_batch_size="128" \
|
13 |
--per_device_eval_batch_size="128" \
|
14 |
+
--learning_rate="2e-4" \
|
15 |
--warmup_steps="1000" \
|
16 |
--overwrite_output_dir \
|
17 |
+
--num_train_epochs="8" \
|
|
|
|
|
|
|
18 |
--push_to_hub
|
19 |
|