Saving weights of epoch 2 at step 74

flax_model.msgpack

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:026a9da1385727ad25b118be45961965cf77faaf56bfa6193d1b1105fd73fc2c
+oid sha256:f62dbb35840c3ee0b6e92eefbf42206f23b69c4d9bb70730914a09033cf3313b
 size 5262347427

results_tensorboard/{events.out.tfevents.1627575327.t1v-n-cb40e504-w-0.12731.3.v2 → events.out.tfevents.1627575762.t1v-n-cb40e504-w-0.14227.3.v2}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8da59be43f840d393332e6be49c905a9a5e49e95cfb3aa2d172593aa454e3283
-size 1160
+oid sha256:589f97a8921bac4055e7af9db80f84ebc5d41840ce4ffd6839b04da359c9de3e
+size 13044

src/gptneo_story.py

@@ -44,8 +44,8 @@ def preprocess(example):
 train_dataset=train_dataset.map(preprocess)
 validation_dataset=validation_dataset.map(preprocess)
 
-len_train_dataset=100
-len_validation_dataset=100
+len_train_dataset=2456
+len_validation_dataset=1200
 
 train_dataset=train_dataset.select(range(len_train_dataset))
 validation_dataset=validation_dataset.select(range(len_validation_dataset))
@@ -53,7 +53,7 @@ validation_dataset=validation_dataset.select(range(len_validation_dataset))
 remove_col=train_dataset.column_names
 
 def tokenize(examples):
-    tokenized_examples=tokenizer(examples['first_sentence'],examples['second_sentence'],padding='max_length',truncation=True,max_length=256,return_tensors='jax')
+    tokenized_examples=tokenizer(examples['first_sentence'],examples['second_sentence'],padding='max_length',truncation=True,max_length=339,return_tensors='jax')
     tokenized_examples['labels']=int(examples['right_answer'])
     return tokenized_examples