pere commited on
Commit
2af672a
1 Parent(s): 493b305

Saving weights and logs of step 2500

Browse files
.create_config.py.un~ ADDED
Binary file (2.25 kB). View file
 
.rrain.sh.un~ ADDED
Binary file (523 Bytes). View file
 
.run_train.sh.un~ ADDED
Binary file (7.51 kB). View file
 
.run_train.un~ ADDED
Binary file (1.81 kB). View file
 
events.out.tfevents.1734078421.t1v-n-53cd541d-w-35.1080156.0.v2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0a7f6357896579784c2fd683291c679d3bd919af4c0329e41ecd08633ef460a
3
+ size 63038
flax_model.msgpack ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f1adcfdad8a681e08bfe5535021067e1c876db067eda92f40f92ccf9c4c5e63
3
+ size 1421658229
run_train.sh CHANGED
@@ -10,10 +10,10 @@ python run_mlm_flax.py \
10
  --weight_decay="0.01" \
11
  --per_device_train_batch_size="64" \
12
  --per_device_eval_batch_size="64" \
13
- --learning_rate="3e-4" \
14
  --warmup_steps="1000" \
15
  --overwrite_output_dir \
16
- --num_train_epochs="18" \
17
  --adam_beta1="0.9" \
18
  --adam_beta2="0.98" \
19
  --logging_steps="500" \
 
10
  --weight_decay="0.01" \
11
  --per_device_train_batch_size="64" \
12
  --per_device_eval_batch_size="64" \
13
+ --learning_rate="5e-4" \
14
  --warmup_steps="1000" \
15
  --overwrite_output_dir \
16
+ --num_train_epochs="100" \
17
  --adam_beta1="0.9" \
18
  --adam_beta2="0.98" \
19
  --logging_steps="500" \