Upload 16 files
Browse files- mICE.ipynb +0 -0
- preprocess.py +91 -0
- rICE/checkpoint-100000/config.json +26 -0
- rICE/checkpoint-100000/optimizer.pt +3 -0
- rICE/checkpoint-100000/pytorch_model.bin +3 -0
- rICE/checkpoint-100000/rng_state.pth +3 -0
- rICE/checkpoint-100000/scheduler.pt +3 -0
- rICE/checkpoint-100000/trainer_state.json +1216 -0
- rICE/checkpoint-100000/training_args.bin +3 -0
- rICE/checkpoint-90000/config.json +26 -0
- rICE/checkpoint-90000/optimizer.pt +3 -0
- rICE/checkpoint-90000/pytorch_model.bin +3 -0
- rICE/checkpoint-90000/rng_state.pth +3 -0
- rICE/checkpoint-90000/scheduler.pt +3 -0
- rICE/checkpoint-90000/trainer_state.json +1096 -0
- rICE/checkpoint-90000/training_args.bin +3 -0
mICE.ipynb
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
preprocess.py
ADDED
|
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#! usr/bin/python3
|
| 2 |
+
import xml.etree.ElementTree as ET
|
| 3 |
+
from os import listdir, walk, remove
|
| 4 |
+
from os.path import isfile, join
|
| 5 |
+
import sys
|
| 6 |
+
from glob import glob
|
| 7 |
+
import urllib.request
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def download_data():
|
| 11 |
+
urls = ["http://hdl.handle.net/20.500.12537/192", "https://repository.clarin.is/repository/xmlui/handle/20.500.12537/32/allzip", "https://repository.clarin.is/repository/xmlui/bitstream/handle/20.500.12537/96/icesum.json?sequence=1&isAllowed=y"]
|
| 12 |
+
for url in urls:
|
| 13 |
+
urllib.request.urlretrieve(url)
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
def get_contents(path='./{http://www.tei-c.org/ns/1.0}text/*/*/*', root=None):
|
| 17 |
+
results = ''
|
| 18 |
+
paragraphs = root.findall(path)
|
| 19 |
+
|
| 20 |
+
for p in paragraphs:
|
| 21 |
+
results += ''.join(t for t in p.itertext()) + '\n'
|
| 22 |
+
|
| 23 |
+
return results
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
def save(path, data):
|
| 27 |
+
f = open(path, 'w', encoding='utf8')
|
| 28 |
+
f.write(data)
|
| 29 |
+
f.close()
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
def delete_contents_of_dir(path='./data/train/*'):
|
| 33 |
+
files = glob(path)
|
| 34 |
+
for f in files:
|
| 35 |
+
remove(f)
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
def path_to_files(path='./CC_BY/IGC-News1-21.05.TEI/frettabladid_is'):
|
| 39 |
+
the_paths = []
|
| 40 |
+
|
| 41 |
+
for dirs in walk(path):
|
| 42 |
+
subdirectory = dirs[0]
|
| 43 |
+
|
| 44 |
+
# extract the path to the text files
|
| 45 |
+
for file in listdir(subdirectory):
|
| 46 |
+
# if the file does not exist pass
|
| 47 |
+
if isfile(join(subdirectory, file)):
|
| 48 |
+
the_paths.append(f'{subdirectory}/{file}')
|
| 49 |
+
return the_paths
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
def convert_n_format(paths):
|
| 53 |
+
# train test ratio
|
| 54 |
+
train = int(len(paths) * .8)
|
| 55 |
+
|
| 56 |
+
for n in range(len(paths)):
|
| 57 |
+
|
| 58 |
+
path = paths[n]
|
| 59 |
+
|
| 60 |
+
# parse the xml file
|
| 61 |
+
tree = ET.parse(path)
|
| 62 |
+
root = tree.getroot()
|
| 63 |
+
|
| 64 |
+
# get contents of the file
|
| 65 |
+
contents = get_contents(root=root)
|
| 66 |
+
|
| 67 |
+
# empty directory before saving to it, to ensure no leeking of data
|
| 68 |
+
# delete_contents_of_dir('./data/test/*')
|
| 69 |
+
|
| 70 |
+
# save the file as a .txt in either the train or test dir
|
| 71 |
+
id = 'news'
|
| 72 |
+
output_path = (f'./data/train/{n}_{id}.txt' if n < (train - 1)
|
| 73 |
+
else f'./data/test/{(n - train)}_{id}.txt')
|
| 74 |
+
|
| 75 |
+
# output_path = (f'./another/{n}-train.txt' if n < (train -1) else f'./another/{(n - len(paths) *.01)}-test.txt')
|
| 76 |
+
|
| 77 |
+
save(output_path, contents)
|
| 78 |
+
|
| 79 |
+
|
| 80 |
+
|
| 81 |
+
print("Beginning data preproccessing")
|
| 82 |
+
|
| 83 |
+
# get all the paths to the files
|
| 84 |
+
print("Collecting paths")
|
| 85 |
+
all_paths = path_to_files('./CC_BY/IGC-News1-21.05.TEI/frettabladid_is')
|
| 86 |
+
print("Found ", len(all_paths))
|
| 87 |
+
|
| 88 |
+
# convert and format all the files
|
| 89 |
+
print("converting files...")
|
| 90 |
+
convert_n_format(all_paths)
|
| 91 |
+
print("\nDone!")
|
rICE/checkpoint-100000/config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"RobertaForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 0,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 3072,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"max_position_embeddings": 514,
|
| 16 |
+
"model_type": "roberta",
|
| 17 |
+
"num_attention_heads": 12,
|
| 18 |
+
"num_hidden_layers": 6,
|
| 19 |
+
"pad_token_id": 1,
|
| 20 |
+
"position_embedding_type": "absolute",
|
| 21 |
+
"torch_dtype": "float32",
|
| 22 |
+
"transformers_version": "4.22.0.dev0",
|
| 23 |
+
"type_vocab_size": 1,
|
| 24 |
+
"use_cache": true,
|
| 25 |
+
"vocab_size": 52000
|
| 26 |
+
}
|
rICE/checkpoint-100000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c33d07878ce6b6fda3afe9d17cee0623b5f9a4cdc54c6e00b803385c51b8c28
|
| 3 |
+
size 668096665
|
rICE/checkpoint-100000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5ba1e0bb3f3c474dcf1f4af243a0a80792793f0f1ca7b9a06d4ff777ab5284e
|
| 3 |
+
size 334056939
|
rICE/checkpoint-100000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b3ce6aacaece8596dc3bae1b725d4c237234e59c2ef92aa9fd8c6b6ee02bf646
|
| 3 |
+
size 14503
|
rICE/checkpoint-100000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:96dc5b3077da2241ec8c56790a190c5712e8ad100fc2d252f29477cb4e133c47
|
| 3 |
+
size 623
|
rICE/checkpoint-100000/trainer_state.json
ADDED
|
@@ -0,0 +1,1216 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9637068018426074,
|
| 5 |
+
"global_step": 100000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 4.9759073299539347e-05,
|
| 13 |
+
"loss": 8.65,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 4.95181465990787e-05,
|
| 19 |
+
"loss": 8.2084,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 4.927721989861805e-05,
|
| 25 |
+
"loss": 8.1327,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.02,
|
| 30 |
+
"learning_rate": 4.90362931981574e-05,
|
| 31 |
+
"loss": 7.9749,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.02,
|
| 36 |
+
"learning_rate": 4.879536649769674e-05,
|
| 37 |
+
"loss": 7.9396,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.03,
|
| 42 |
+
"learning_rate": 4.855443979723609e-05,
|
| 43 |
+
"loss": 7.9153,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03,
|
| 48 |
+
"learning_rate": 4.831351309677544e-05,
|
| 49 |
+
"loss": 7.9056,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.04,
|
| 54 |
+
"learning_rate": 4.807258639631479e-05,
|
| 55 |
+
"loss": 7.8028,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.04,
|
| 60 |
+
"learning_rate": 4.7831659695854133e-05,
|
| 61 |
+
"loss": 7.7427,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.05,
|
| 66 |
+
"learning_rate": 4.7590732995393484e-05,
|
| 67 |
+
"loss": 7.7145,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.05,
|
| 72 |
+
"learning_rate": 4.734980629493283e-05,
|
| 73 |
+
"loss": 7.6656,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.06,
|
| 78 |
+
"learning_rate": 4.7108879594472186e-05,
|
| 79 |
+
"loss": 7.689,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.06,
|
| 84 |
+
"learning_rate": 4.686795289401153e-05,
|
| 85 |
+
"loss": 7.7125,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.07,
|
| 90 |
+
"learning_rate": 4.6627026193550874e-05,
|
| 91 |
+
"loss": 7.5884,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.07,
|
| 96 |
+
"learning_rate": 4.6386099493090225e-05,
|
| 97 |
+
"loss": 7.5532,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.08,
|
| 102 |
+
"learning_rate": 4.614517279262957e-05,
|
| 103 |
+
"loss": 7.5568,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.08,
|
| 108 |
+
"learning_rate": 4.590424609216892e-05,
|
| 109 |
+
"loss": 7.5296,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.09,
|
| 114 |
+
"learning_rate": 4.566331939170827e-05,
|
| 115 |
+
"loss": 7.5091,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.09,
|
| 120 |
+
"learning_rate": 4.5422392691247615e-05,
|
| 121 |
+
"loss": 7.575,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.1,
|
| 126 |
+
"learning_rate": 4.5181465990786966e-05,
|
| 127 |
+
"loss": 7.4049,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1,
|
| 132 |
+
"learning_rate": 4.494053929032631e-05,
|
| 133 |
+
"loss": 7.4642,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.11,
|
| 138 |
+
"learning_rate": 4.469961258986566e-05,
|
| 139 |
+
"loss": 7.4697,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.11,
|
| 144 |
+
"learning_rate": 4.445868588940501e-05,
|
| 145 |
+
"loss": 7.4979,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.12,
|
| 150 |
+
"learning_rate": 4.4217759188944356e-05,
|
| 151 |
+
"loss": 7.3755,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.12,
|
| 156 |
+
"learning_rate": 4.397683248848371e-05,
|
| 157 |
+
"loss": 7.4219,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.13,
|
| 162 |
+
"learning_rate": 4.373590578802305e-05,
|
| 163 |
+
"loss": 7.326,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.13,
|
| 168 |
+
"learning_rate": 4.34949790875624e-05,
|
| 169 |
+
"loss": 7.3567,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.13,
|
| 174 |
+
"learning_rate": 4.325405238710175e-05,
|
| 175 |
+
"loss": 7.3167,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.14,
|
| 180 |
+
"learning_rate": 4.30131256866411e-05,
|
| 181 |
+
"loss": 7.2823,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.14,
|
| 186 |
+
"learning_rate": 4.277219898618045e-05,
|
| 187 |
+
"loss": 7.3109,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.15,
|
| 192 |
+
"learning_rate": 4.253127228571979e-05,
|
| 193 |
+
"loss": 7.3226,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.15,
|
| 198 |
+
"learning_rate": 4.229034558525914e-05,
|
| 199 |
+
"loss": 7.2789,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.16,
|
| 204 |
+
"learning_rate": 4.2049418884798494e-05,
|
| 205 |
+
"loss": 7.2836,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.16,
|
| 210 |
+
"learning_rate": 4.180849218433784e-05,
|
| 211 |
+
"loss": 7.2753,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.17,
|
| 216 |
+
"learning_rate": 4.156756548387718e-05,
|
| 217 |
+
"loss": 7.2343,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.17,
|
| 222 |
+
"learning_rate": 4.132663878341654e-05,
|
| 223 |
+
"loss": 7.2357,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.18,
|
| 228 |
+
"learning_rate": 4.1085712082955884e-05,
|
| 229 |
+
"loss": 7.3118,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.18,
|
| 234 |
+
"learning_rate": 4.0844785382495235e-05,
|
| 235 |
+
"loss": 7.2337,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.19,
|
| 240 |
+
"learning_rate": 4.060385868203458e-05,
|
| 241 |
+
"loss": 7.1841,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.19,
|
| 246 |
+
"learning_rate": 4.036293198157392e-05,
|
| 247 |
+
"loss": 7.1875,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.2,
|
| 252 |
+
"learning_rate": 4.012200528111328e-05,
|
| 253 |
+
"loss": 7.1547,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.2,
|
| 258 |
+
"learning_rate": 3.9881078580652625e-05,
|
| 259 |
+
"loss": 7.1319,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.21,
|
| 264 |
+
"learning_rate": 3.964015188019197e-05,
|
| 265 |
+
"loss": 7.1787,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.21,
|
| 270 |
+
"learning_rate": 3.939922517973132e-05,
|
| 271 |
+
"loss": 7.0693,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.22,
|
| 276 |
+
"learning_rate": 3.9158298479270664e-05,
|
| 277 |
+
"loss": 7.1514,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.22,
|
| 282 |
+
"learning_rate": 3.891737177881002e-05,
|
| 283 |
+
"loss": 7.122,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.23,
|
| 288 |
+
"learning_rate": 3.8676445078349366e-05,
|
| 289 |
+
"loss": 7.0971,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.23,
|
| 294 |
+
"learning_rate": 3.843551837788871e-05,
|
| 295 |
+
"loss": 7.0787,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.24,
|
| 300 |
+
"learning_rate": 3.819459167742806e-05,
|
| 301 |
+
"loss": 7.0807,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.24,
|
| 306 |
+
"learning_rate": 3.7953664976967405e-05,
|
| 307 |
+
"loss": 7.0816,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.25,
|
| 312 |
+
"learning_rate": 3.7712738276506756e-05,
|
| 313 |
+
"loss": 7.0138,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.25,
|
| 318 |
+
"learning_rate": 3.747181157604611e-05,
|
| 319 |
+
"loss": 7.0621,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.26,
|
| 324 |
+
"learning_rate": 3.723088487558545e-05,
|
| 325 |
+
"loss": 6.9711,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.26,
|
| 330 |
+
"learning_rate": 3.69899581751248e-05,
|
| 331 |
+
"loss": 6.9892,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.27,
|
| 336 |
+
"learning_rate": 3.6749031474664146e-05,
|
| 337 |
+
"loss": 6.9754,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.27,
|
| 342 |
+
"learning_rate": 3.65081047742035e-05,
|
| 343 |
+
"loss": 6.9831,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.27,
|
| 348 |
+
"learning_rate": 3.626717807374285e-05,
|
| 349 |
+
"loss": 6.9777,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.28,
|
| 354 |
+
"learning_rate": 3.602625137328219e-05,
|
| 355 |
+
"loss": 6.9249,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.28,
|
| 360 |
+
"learning_rate": 3.578532467282154e-05,
|
| 361 |
+
"loss": 6.8832,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.29,
|
| 366 |
+
"learning_rate": 3.5544397972360894e-05,
|
| 367 |
+
"loss": 6.8826,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.29,
|
| 372 |
+
"learning_rate": 3.530347127190024e-05,
|
| 373 |
+
"loss": 6.8175,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.3,
|
| 378 |
+
"learning_rate": 3.506254457143959e-05,
|
| 379 |
+
"loss": 6.9092,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.3,
|
| 384 |
+
"learning_rate": 3.482161787097893e-05,
|
| 385 |
+
"loss": 6.7819,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.31,
|
| 390 |
+
"learning_rate": 3.4580691170518284e-05,
|
| 391 |
+
"loss": 6.8296,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.31,
|
| 396 |
+
"learning_rate": 3.4339764470057635e-05,
|
| 397 |
+
"loss": 6.8417,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.32,
|
| 402 |
+
"learning_rate": 3.409883776959698e-05,
|
| 403 |
+
"loss": 6.7656,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.32,
|
| 408 |
+
"learning_rate": 3.385791106913633e-05,
|
| 409 |
+
"loss": 6.7687,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.33,
|
| 414 |
+
"learning_rate": 3.3616984368675674e-05,
|
| 415 |
+
"loss": 6.7539,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.33,
|
| 420 |
+
"learning_rate": 3.337605766821502e-05,
|
| 421 |
+
"loss": 6.7836,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.34,
|
| 426 |
+
"learning_rate": 3.3135130967754376e-05,
|
| 427 |
+
"loss": 6.6956,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.34,
|
| 432 |
+
"learning_rate": 3.289420426729372e-05,
|
| 433 |
+
"loss": 6.6429,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.35,
|
| 438 |
+
"learning_rate": 3.265327756683307e-05,
|
| 439 |
+
"loss": 6.7027,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.35,
|
| 444 |
+
"learning_rate": 3.2412350866372415e-05,
|
| 445 |
+
"loss": 6.7303,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.36,
|
| 450 |
+
"learning_rate": 3.217142416591176e-05,
|
| 451 |
+
"loss": 6.6036,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.36,
|
| 456 |
+
"learning_rate": 3.193049746545112e-05,
|
| 457 |
+
"loss": 6.5406,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.37,
|
| 462 |
+
"learning_rate": 3.168957076499046e-05,
|
| 463 |
+
"loss": 6.5995,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.37,
|
| 468 |
+
"learning_rate": 3.144864406452981e-05,
|
| 469 |
+
"loss": 6.5864,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.38,
|
| 474 |
+
"learning_rate": 3.1207717364069156e-05,
|
| 475 |
+
"loss": 6.5612,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.38,
|
| 480 |
+
"learning_rate": 3.09667906636085e-05,
|
| 481 |
+
"loss": 6.5546,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.39,
|
| 486 |
+
"learning_rate": 3.072586396314786e-05,
|
| 487 |
+
"loss": 6.4666,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.39,
|
| 492 |
+
"learning_rate": 3.0484937262687202e-05,
|
| 493 |
+
"loss": 6.4376,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.4,
|
| 498 |
+
"learning_rate": 3.024401056222655e-05,
|
| 499 |
+
"loss": 6.4767,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.4,
|
| 504 |
+
"learning_rate": 3.0003083861765897e-05,
|
| 505 |
+
"loss": 6.4738,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.4,
|
| 510 |
+
"learning_rate": 2.9762157161305248e-05,
|
| 511 |
+
"loss": 6.4136,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.41,
|
| 516 |
+
"learning_rate": 2.9521230460844596e-05,
|
| 517 |
+
"loss": 6.4317,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.41,
|
| 522 |
+
"learning_rate": 2.9280303760383943e-05,
|
| 523 |
+
"loss": 6.4216,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.42,
|
| 528 |
+
"learning_rate": 2.903937705992329e-05,
|
| 529 |
+
"loss": 6.4294,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.42,
|
| 534 |
+
"learning_rate": 2.8798450359462638e-05,
|
| 535 |
+
"loss": 6.3409,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.43,
|
| 540 |
+
"learning_rate": 2.855752365900199e-05,
|
| 541 |
+
"loss": 6.3394,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.43,
|
| 546 |
+
"learning_rate": 2.8316596958541336e-05,
|
| 547 |
+
"loss": 6.3577,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.44,
|
| 552 |
+
"learning_rate": 2.8075670258080684e-05,
|
| 553 |
+
"loss": 6.2582,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.44,
|
| 558 |
+
"learning_rate": 2.783474355762003e-05,
|
| 559 |
+
"loss": 6.2806,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.45,
|
| 564 |
+
"learning_rate": 2.7593816857159376e-05,
|
| 565 |
+
"loss": 6.3391,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.45,
|
| 570 |
+
"learning_rate": 2.735289015669873e-05,
|
| 571 |
+
"loss": 6.3571,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.46,
|
| 576 |
+
"learning_rate": 2.7111963456238077e-05,
|
| 577 |
+
"loss": 6.2282,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.46,
|
| 582 |
+
"learning_rate": 2.6871036755777425e-05,
|
| 583 |
+
"loss": 6.3069,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.47,
|
| 588 |
+
"learning_rate": 2.663011005531677e-05,
|
| 589 |
+
"loss": 6.2606,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.47,
|
| 594 |
+
"learning_rate": 2.6389183354856117e-05,
|
| 595 |
+
"loss": 6.2315,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.48,
|
| 600 |
+
"learning_rate": 2.614825665439547e-05,
|
| 601 |
+
"loss": 6.2534,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.48,
|
| 606 |
+
"learning_rate": 2.590732995393482e-05,
|
| 607 |
+
"loss": 6.1819,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.49,
|
| 612 |
+
"learning_rate": 2.5666403253474163e-05,
|
| 613 |
+
"loss": 6.123,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.49,
|
| 618 |
+
"learning_rate": 2.542547655301351e-05,
|
| 619 |
+
"loss": 6.1913,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.5,
|
| 624 |
+
"learning_rate": 2.5184549852552858e-05,
|
| 625 |
+
"loss": 6.1996,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.5,
|
| 630 |
+
"learning_rate": 2.494362315209221e-05,
|
| 631 |
+
"loss": 6.175,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.51,
|
| 636 |
+
"learning_rate": 2.4702696451631556e-05,
|
| 637 |
+
"loss": 6.1644,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.51,
|
| 642 |
+
"learning_rate": 2.4461769751170904e-05,
|
| 643 |
+
"loss": 6.096,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.52,
|
| 648 |
+
"learning_rate": 2.4220843050710254e-05,
|
| 649 |
+
"loss": 6.0954,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.52,
|
| 654 |
+
"learning_rate": 2.3979916350249602e-05,
|
| 655 |
+
"loss": 6.0694,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.53,
|
| 660 |
+
"learning_rate": 2.373898964978895e-05,
|
| 661 |
+
"loss": 6.0671,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.53,
|
| 666 |
+
"learning_rate": 2.3498062949328297e-05,
|
| 667 |
+
"loss": 6.0728,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.53,
|
| 672 |
+
"learning_rate": 2.3257136248867644e-05,
|
| 673 |
+
"loss": 6.1125,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.54,
|
| 678 |
+
"learning_rate": 2.3016209548406995e-05,
|
| 679 |
+
"loss": 6.0758,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.54,
|
| 684 |
+
"learning_rate": 2.2775282847946343e-05,
|
| 685 |
+
"loss": 6.0084,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.55,
|
| 690 |
+
"learning_rate": 2.253435614748569e-05,
|
| 691 |
+
"loss": 6.06,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.55,
|
| 696 |
+
"learning_rate": 2.2293429447025038e-05,
|
| 697 |
+
"loss": 6.0424,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.56,
|
| 702 |
+
"learning_rate": 2.2052502746564385e-05,
|
| 703 |
+
"loss": 6.1225,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.56,
|
| 708 |
+
"learning_rate": 2.1811576046103736e-05,
|
| 709 |
+
"loss": 6.0063,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.57,
|
| 714 |
+
"learning_rate": 2.157064934564308e-05,
|
| 715 |
+
"loss": 5.9357,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.57,
|
| 720 |
+
"learning_rate": 2.132972264518243e-05,
|
| 721 |
+
"loss": 5.9604,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.58,
|
| 726 |
+
"learning_rate": 2.108879594472178e-05,
|
| 727 |
+
"loss": 6.0146,
|
| 728 |
+
"step": 60000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.58,
|
| 732 |
+
"learning_rate": 2.0847869244261126e-05,
|
| 733 |
+
"loss": 5.9849,
|
| 734 |
+
"step": 60500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.59,
|
| 738 |
+
"learning_rate": 2.0606942543800474e-05,
|
| 739 |
+
"loss": 6.0048,
|
| 740 |
+
"step": 61000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.59,
|
| 744 |
+
"learning_rate": 2.036601584333982e-05,
|
| 745 |
+
"loss": 5.9397,
|
| 746 |
+
"step": 61500
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.6,
|
| 750 |
+
"learning_rate": 2.0125089142879172e-05,
|
| 751 |
+
"loss": 6.0005,
|
| 752 |
+
"step": 62000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.6,
|
| 756 |
+
"learning_rate": 1.988416244241852e-05,
|
| 757 |
+
"loss": 5.9041,
|
| 758 |
+
"step": 62500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.61,
|
| 762 |
+
"learning_rate": 1.9643235741957867e-05,
|
| 763 |
+
"loss": 5.9352,
|
| 764 |
+
"step": 63000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.61,
|
| 768 |
+
"learning_rate": 1.9402309041497215e-05,
|
| 769 |
+
"loss": 5.9038,
|
| 770 |
+
"step": 63500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.62,
|
| 774 |
+
"learning_rate": 1.9161382341036562e-05,
|
| 775 |
+
"loss": 5.9472,
|
| 776 |
+
"step": 64000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.62,
|
| 780 |
+
"learning_rate": 1.8920455640575913e-05,
|
| 781 |
+
"loss": 5.9163,
|
| 782 |
+
"step": 64500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.63,
|
| 786 |
+
"learning_rate": 1.867952894011526e-05,
|
| 787 |
+
"loss": 5.9089,
|
| 788 |
+
"step": 65000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.63,
|
| 792 |
+
"learning_rate": 1.843860223965461e-05,
|
| 793 |
+
"loss": 5.9437,
|
| 794 |
+
"step": 65500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.64,
|
| 798 |
+
"learning_rate": 1.8197675539193956e-05,
|
| 799 |
+
"loss": 5.8682,
|
| 800 |
+
"step": 66000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.64,
|
| 804 |
+
"learning_rate": 1.7956748838733303e-05,
|
| 805 |
+
"loss": 5.8244,
|
| 806 |
+
"step": 66500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.65,
|
| 810 |
+
"learning_rate": 1.7715822138272654e-05,
|
| 811 |
+
"loss": 5.851,
|
| 812 |
+
"step": 67000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.65,
|
| 816 |
+
"learning_rate": 1.7474895437812e-05,
|
| 817 |
+
"loss": 5.8414,
|
| 818 |
+
"step": 67500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.66,
|
| 822 |
+
"learning_rate": 1.723396873735135e-05,
|
| 823 |
+
"loss": 5.8678,
|
| 824 |
+
"step": 68000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.66,
|
| 828 |
+
"learning_rate": 1.6993042036890697e-05,
|
| 829 |
+
"loss": 5.8232,
|
| 830 |
+
"step": 68500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.66,
|
| 834 |
+
"learning_rate": 1.6752115336430048e-05,
|
| 835 |
+
"loss": 5.8244,
|
| 836 |
+
"step": 69000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.67,
|
| 840 |
+
"learning_rate": 1.6511188635969392e-05,
|
| 841 |
+
"loss": 5.8314,
|
| 842 |
+
"step": 69500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.67,
|
| 846 |
+
"learning_rate": 1.627026193550874e-05,
|
| 847 |
+
"loss": 5.8577,
|
| 848 |
+
"step": 70000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.68,
|
| 852 |
+
"learning_rate": 1.602933523504809e-05,
|
| 853 |
+
"loss": 5.7637,
|
| 854 |
+
"step": 70500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.68,
|
| 858 |
+
"learning_rate": 1.5788408534587438e-05,
|
| 859 |
+
"loss": 5.83,
|
| 860 |
+
"step": 71000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.69,
|
| 864 |
+
"learning_rate": 1.5547481834126785e-05,
|
| 865 |
+
"loss": 5.8128,
|
| 866 |
+
"step": 71500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.69,
|
| 870 |
+
"learning_rate": 1.5306555133666133e-05,
|
| 871 |
+
"loss": 5.7518,
|
| 872 |
+
"step": 72000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.7,
|
| 876 |
+
"learning_rate": 1.5065628433205484e-05,
|
| 877 |
+
"loss": 5.8433,
|
| 878 |
+
"step": 72500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.7,
|
| 882 |
+
"learning_rate": 1.4824701732744831e-05,
|
| 883 |
+
"loss": 5.7624,
|
| 884 |
+
"step": 73000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.71,
|
| 888 |
+
"learning_rate": 1.4583775032284177e-05,
|
| 889 |
+
"loss": 5.7489,
|
| 890 |
+
"step": 73500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.71,
|
| 894 |
+
"learning_rate": 1.4342848331823528e-05,
|
| 895 |
+
"loss": 5.7542,
|
| 896 |
+
"step": 74000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.72,
|
| 900 |
+
"learning_rate": 1.4101921631362874e-05,
|
| 901 |
+
"loss": 5.7664,
|
| 902 |
+
"step": 74500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.72,
|
| 906 |
+
"learning_rate": 1.3860994930902225e-05,
|
| 907 |
+
"loss": 5.7242,
|
| 908 |
+
"step": 75000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.73,
|
| 912 |
+
"learning_rate": 1.362006823044157e-05,
|
| 913 |
+
"loss": 5.7448,
|
| 914 |
+
"step": 75500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.73,
|
| 918 |
+
"learning_rate": 1.3379141529980918e-05,
|
| 919 |
+
"loss": 5.6599,
|
| 920 |
+
"step": 76000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.74,
|
| 924 |
+
"learning_rate": 1.3138214829520267e-05,
|
| 925 |
+
"loss": 5.8041,
|
| 926 |
+
"step": 76500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.74,
|
| 930 |
+
"learning_rate": 1.2897288129059615e-05,
|
| 931 |
+
"loss": 5.7124,
|
| 932 |
+
"step": 77000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.75,
|
| 936 |
+
"learning_rate": 1.2656361428598964e-05,
|
| 937 |
+
"loss": 5.7687,
|
| 938 |
+
"step": 77500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.75,
|
| 942 |
+
"learning_rate": 1.2415434728138312e-05,
|
| 943 |
+
"loss": 5.7568,
|
| 944 |
+
"step": 78000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.76,
|
| 948 |
+
"learning_rate": 1.217450802767766e-05,
|
| 949 |
+
"loss": 5.729,
|
| 950 |
+
"step": 78500
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.76,
|
| 954 |
+
"learning_rate": 1.1933581327217008e-05,
|
| 955 |
+
"loss": 5.7368,
|
| 956 |
+
"step": 79000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.77,
|
| 960 |
+
"learning_rate": 1.1692654626756357e-05,
|
| 961 |
+
"loss": 5.7522,
|
| 962 |
+
"step": 79500
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.77,
|
| 966 |
+
"learning_rate": 1.1451727926295705e-05,
|
| 967 |
+
"loss": 5.7453,
|
| 968 |
+
"step": 80000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.78,
|
| 972 |
+
"learning_rate": 1.1210801225835052e-05,
|
| 973 |
+
"loss": 5.6756,
|
| 974 |
+
"step": 80500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.78,
|
| 978 |
+
"learning_rate": 1.09698745253744e-05,
|
| 979 |
+
"loss": 5.764,
|
| 980 |
+
"step": 81000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.79,
|
| 984 |
+
"learning_rate": 1.072894782491375e-05,
|
| 985 |
+
"loss": 5.6947,
|
| 986 |
+
"step": 81500
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.79,
|
| 990 |
+
"learning_rate": 1.0488021124453097e-05,
|
| 991 |
+
"loss": 5.6659,
|
| 992 |
+
"step": 82000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.8,
|
| 996 |
+
"learning_rate": 1.0247094423992446e-05,
|
| 997 |
+
"loss": 5.7086,
|
| 998 |
+
"step": 82500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.8,
|
| 1002 |
+
"learning_rate": 1.0006167723531793e-05,
|
| 1003 |
+
"loss": 5.6405,
|
| 1004 |
+
"step": 83000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.8,
|
| 1008 |
+
"learning_rate": 9.765241023071141e-06,
|
| 1009 |
+
"loss": 5.6981,
|
| 1010 |
+
"step": 83500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.81,
|
| 1014 |
+
"learning_rate": 9.524314322610489e-06,
|
| 1015 |
+
"loss": 5.732,
|
| 1016 |
+
"step": 84000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.81,
|
| 1020 |
+
"learning_rate": 9.283387622149838e-06,
|
| 1021 |
+
"loss": 5.6404,
|
| 1022 |
+
"step": 84500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.82,
|
| 1026 |
+
"learning_rate": 9.042460921689185e-06,
|
| 1027 |
+
"loss": 5.6796,
|
| 1028 |
+
"step": 85000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.82,
|
| 1032 |
+
"learning_rate": 8.801534221228534e-06,
|
| 1033 |
+
"loss": 5.7269,
|
| 1034 |
+
"step": 85500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.83,
|
| 1038 |
+
"learning_rate": 8.560607520767882e-06,
|
| 1039 |
+
"loss": 5.6806,
|
| 1040 |
+
"step": 86000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.83,
|
| 1044 |
+
"learning_rate": 8.31968082030723e-06,
|
| 1045 |
+
"loss": 5.5852,
|
| 1046 |
+
"step": 86500
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.84,
|
| 1050 |
+
"learning_rate": 8.078754119846579e-06,
|
| 1051 |
+
"loss": 5.6754,
|
| 1052 |
+
"step": 87000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.84,
|
| 1056 |
+
"learning_rate": 7.837827419385926e-06,
|
| 1057 |
+
"loss": 5.6727,
|
| 1058 |
+
"step": 87500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.85,
|
| 1062 |
+
"learning_rate": 7.5969007189252746e-06,
|
| 1063 |
+
"loss": 5.7247,
|
| 1064 |
+
"step": 88000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.85,
|
| 1068 |
+
"learning_rate": 7.355974018464623e-06,
|
| 1069 |
+
"loss": 5.6207,
|
| 1070 |
+
"step": 88500
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.86,
|
| 1074 |
+
"learning_rate": 7.115047318003971e-06,
|
| 1075 |
+
"loss": 5.6297,
|
| 1076 |
+
"step": 89000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.86,
|
| 1080 |
+
"learning_rate": 6.874120617543319e-06,
|
| 1081 |
+
"loss": 5.6616,
|
| 1082 |
+
"step": 89500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.87,
|
| 1086 |
+
"learning_rate": 6.633193917082667e-06,
|
| 1087 |
+
"loss": 5.6567,
|
| 1088 |
+
"step": 90000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.87,
|
| 1092 |
+
"learning_rate": 6.3922672166220155e-06,
|
| 1093 |
+
"loss": 5.6692,
|
| 1094 |
+
"step": 90500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.88,
|
| 1098 |
+
"learning_rate": 6.151340516161364e-06,
|
| 1099 |
+
"loss": 5.5955,
|
| 1100 |
+
"step": 91000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.88,
|
| 1104 |
+
"learning_rate": 5.910413815700711e-06,
|
| 1105 |
+
"loss": 5.6221,
|
| 1106 |
+
"step": 91500
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.89,
|
| 1110 |
+
"learning_rate": 5.66948711524006e-06,
|
| 1111 |
+
"loss": 5.6195,
|
| 1112 |
+
"step": 92000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.89,
|
| 1116 |
+
"learning_rate": 5.428560414779408e-06,
|
| 1117 |
+
"loss": 5.5157,
|
| 1118 |
+
"step": 92500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.9,
|
| 1122 |
+
"learning_rate": 5.187633714318756e-06,
|
| 1123 |
+
"loss": 5.5166,
|
| 1124 |
+
"step": 93000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.9,
|
| 1128 |
+
"learning_rate": 4.946707013858104e-06,
|
| 1129 |
+
"loss": 5.6297,
|
| 1130 |
+
"step": 93500
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.91,
|
| 1134 |
+
"learning_rate": 4.705780313397452e-06,
|
| 1135 |
+
"loss": 5.6392,
|
| 1136 |
+
"step": 94000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.91,
|
| 1140 |
+
"learning_rate": 4.4648536129368e-06,
|
| 1141 |
+
"loss": 5.6358,
|
| 1142 |
+
"step": 94500
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.92,
|
| 1146 |
+
"learning_rate": 4.223926912476148e-06,
|
| 1147 |
+
"loss": 5.6061,
|
| 1148 |
+
"step": 95000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.92,
|
| 1152 |
+
"learning_rate": 3.983000212015497e-06,
|
| 1153 |
+
"loss": 5.575,
|
| 1154 |
+
"step": 95500
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.93,
|
| 1158 |
+
"learning_rate": 3.7420735115548446e-06,
|
| 1159 |
+
"loss": 5.6051,
|
| 1160 |
+
"step": 96000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.93,
|
| 1164 |
+
"learning_rate": 3.501146811094193e-06,
|
| 1165 |
+
"loss": 5.6199,
|
| 1166 |
+
"step": 96500
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.93,
|
| 1170 |
+
"learning_rate": 3.2602201106335413e-06,
|
| 1171 |
+
"loss": 5.6355,
|
| 1172 |
+
"step": 97000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.94,
|
| 1176 |
+
"learning_rate": 3.0192934101728892e-06,
|
| 1177 |
+
"loss": 5.5511,
|
| 1178 |
+
"step": 97500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.94,
|
| 1182 |
+
"learning_rate": 2.778366709712237e-06,
|
| 1183 |
+
"loss": 5.5902,
|
| 1184 |
+
"step": 98000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.95,
|
| 1188 |
+
"learning_rate": 2.5374400092515855e-06,
|
| 1189 |
+
"loss": 5.6163,
|
| 1190 |
+
"step": 98500
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.95,
|
| 1194 |
+
"learning_rate": 2.2965133087909335e-06,
|
| 1195 |
+
"loss": 5.5923,
|
| 1196 |
+
"step": 99000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.96,
|
| 1200 |
+
"learning_rate": 2.0555866083302814e-06,
|
| 1201 |
+
"loss": 5.5735,
|
| 1202 |
+
"step": 99500
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.96,
|
| 1206 |
+
"learning_rate": 1.81465990786963e-06,
|
| 1207 |
+
"loss": 5.6391,
|
| 1208 |
+
"step": 100000
|
| 1209 |
+
}
|
| 1210 |
+
],
|
| 1211 |
+
"max_steps": 103766,
|
| 1212 |
+
"num_train_epochs": 1,
|
| 1213 |
+
"total_flos": 8240866910684928.0,
|
| 1214 |
+
"trial_name": null,
|
| 1215 |
+
"trial_params": null
|
| 1216 |
+
}
|
rICE/checkpoint-100000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9de5a40da776e5cbb26e7943f3fb17679c7665327d670d554ee58b0243cb6bd0
|
| 3 |
+
size 3311
|
rICE/checkpoint-90000/config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"RobertaForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 0,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 3072,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"max_position_embeddings": 514,
|
| 16 |
+
"model_type": "roberta",
|
| 17 |
+
"num_attention_heads": 12,
|
| 18 |
+
"num_hidden_layers": 6,
|
| 19 |
+
"pad_token_id": 1,
|
| 20 |
+
"position_embedding_type": "absolute",
|
| 21 |
+
"torch_dtype": "float32",
|
| 22 |
+
"transformers_version": "4.22.0.dev0",
|
| 23 |
+
"type_vocab_size": 1,
|
| 24 |
+
"use_cache": true,
|
| 25 |
+
"vocab_size": 52000
|
| 26 |
+
}
|
rICE/checkpoint-90000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:910d72be22c0bcc3fb6fa386cac602b2c152f2d27bce2326d9699e5372c97825
|
| 3 |
+
size 668096665
|
rICE/checkpoint-90000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e67b3221c8c90eeea5e106daeee4dfe26cb2d983d8e687d675c4d89f1419dc28
|
| 3 |
+
size 334056939
|
rICE/checkpoint-90000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0dca7b74bf48a15c8c724a3c2cb10c797905f0d0d3ae1ac3137ce917b49fd857
|
| 3 |
+
size 14503
|
rICE/checkpoint-90000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:532993adf55f8bc06883f3c60b85c063fbac643d94175a44cd1b886f851b168b
|
| 3 |
+
size 623
|
rICE/checkpoint-90000/trainer_state.json
ADDED
|
@@ -0,0 +1,1096 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.8673361216583466,
|
| 5 |
+
"global_step": 90000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 4.9759073299539347e-05,
|
| 13 |
+
"loss": 8.65,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 4.95181465990787e-05,
|
| 19 |
+
"loss": 8.2084,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 4.927721989861805e-05,
|
| 25 |
+
"loss": 8.1327,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.02,
|
| 30 |
+
"learning_rate": 4.90362931981574e-05,
|
| 31 |
+
"loss": 7.9749,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.02,
|
| 36 |
+
"learning_rate": 4.879536649769674e-05,
|
| 37 |
+
"loss": 7.9396,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.03,
|
| 42 |
+
"learning_rate": 4.855443979723609e-05,
|
| 43 |
+
"loss": 7.9153,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03,
|
| 48 |
+
"learning_rate": 4.831351309677544e-05,
|
| 49 |
+
"loss": 7.9056,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.04,
|
| 54 |
+
"learning_rate": 4.807258639631479e-05,
|
| 55 |
+
"loss": 7.8028,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.04,
|
| 60 |
+
"learning_rate": 4.7831659695854133e-05,
|
| 61 |
+
"loss": 7.7427,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.05,
|
| 66 |
+
"learning_rate": 4.7590732995393484e-05,
|
| 67 |
+
"loss": 7.7145,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.05,
|
| 72 |
+
"learning_rate": 4.734980629493283e-05,
|
| 73 |
+
"loss": 7.6656,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.06,
|
| 78 |
+
"learning_rate": 4.7108879594472186e-05,
|
| 79 |
+
"loss": 7.689,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.06,
|
| 84 |
+
"learning_rate": 4.686795289401153e-05,
|
| 85 |
+
"loss": 7.7125,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.07,
|
| 90 |
+
"learning_rate": 4.6627026193550874e-05,
|
| 91 |
+
"loss": 7.5884,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.07,
|
| 96 |
+
"learning_rate": 4.6386099493090225e-05,
|
| 97 |
+
"loss": 7.5532,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.08,
|
| 102 |
+
"learning_rate": 4.614517279262957e-05,
|
| 103 |
+
"loss": 7.5568,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.08,
|
| 108 |
+
"learning_rate": 4.590424609216892e-05,
|
| 109 |
+
"loss": 7.5296,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.09,
|
| 114 |
+
"learning_rate": 4.566331939170827e-05,
|
| 115 |
+
"loss": 7.5091,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.09,
|
| 120 |
+
"learning_rate": 4.5422392691247615e-05,
|
| 121 |
+
"loss": 7.575,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.1,
|
| 126 |
+
"learning_rate": 4.5181465990786966e-05,
|
| 127 |
+
"loss": 7.4049,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1,
|
| 132 |
+
"learning_rate": 4.494053929032631e-05,
|
| 133 |
+
"loss": 7.4642,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.11,
|
| 138 |
+
"learning_rate": 4.469961258986566e-05,
|
| 139 |
+
"loss": 7.4697,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.11,
|
| 144 |
+
"learning_rate": 4.445868588940501e-05,
|
| 145 |
+
"loss": 7.4979,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.12,
|
| 150 |
+
"learning_rate": 4.4217759188944356e-05,
|
| 151 |
+
"loss": 7.3755,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.12,
|
| 156 |
+
"learning_rate": 4.397683248848371e-05,
|
| 157 |
+
"loss": 7.4219,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.13,
|
| 162 |
+
"learning_rate": 4.373590578802305e-05,
|
| 163 |
+
"loss": 7.326,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.13,
|
| 168 |
+
"learning_rate": 4.34949790875624e-05,
|
| 169 |
+
"loss": 7.3567,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.13,
|
| 174 |
+
"learning_rate": 4.325405238710175e-05,
|
| 175 |
+
"loss": 7.3167,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.14,
|
| 180 |
+
"learning_rate": 4.30131256866411e-05,
|
| 181 |
+
"loss": 7.2823,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.14,
|
| 186 |
+
"learning_rate": 4.277219898618045e-05,
|
| 187 |
+
"loss": 7.3109,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.15,
|
| 192 |
+
"learning_rate": 4.253127228571979e-05,
|
| 193 |
+
"loss": 7.3226,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.15,
|
| 198 |
+
"learning_rate": 4.229034558525914e-05,
|
| 199 |
+
"loss": 7.2789,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.16,
|
| 204 |
+
"learning_rate": 4.2049418884798494e-05,
|
| 205 |
+
"loss": 7.2836,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.16,
|
| 210 |
+
"learning_rate": 4.180849218433784e-05,
|
| 211 |
+
"loss": 7.2753,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.17,
|
| 216 |
+
"learning_rate": 4.156756548387718e-05,
|
| 217 |
+
"loss": 7.2343,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.17,
|
| 222 |
+
"learning_rate": 4.132663878341654e-05,
|
| 223 |
+
"loss": 7.2357,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.18,
|
| 228 |
+
"learning_rate": 4.1085712082955884e-05,
|
| 229 |
+
"loss": 7.3118,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.18,
|
| 234 |
+
"learning_rate": 4.0844785382495235e-05,
|
| 235 |
+
"loss": 7.2337,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.19,
|
| 240 |
+
"learning_rate": 4.060385868203458e-05,
|
| 241 |
+
"loss": 7.1841,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.19,
|
| 246 |
+
"learning_rate": 4.036293198157392e-05,
|
| 247 |
+
"loss": 7.1875,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.2,
|
| 252 |
+
"learning_rate": 4.012200528111328e-05,
|
| 253 |
+
"loss": 7.1547,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.2,
|
| 258 |
+
"learning_rate": 3.9881078580652625e-05,
|
| 259 |
+
"loss": 7.1319,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.21,
|
| 264 |
+
"learning_rate": 3.964015188019197e-05,
|
| 265 |
+
"loss": 7.1787,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.21,
|
| 270 |
+
"learning_rate": 3.939922517973132e-05,
|
| 271 |
+
"loss": 7.0693,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.22,
|
| 276 |
+
"learning_rate": 3.9158298479270664e-05,
|
| 277 |
+
"loss": 7.1514,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.22,
|
| 282 |
+
"learning_rate": 3.891737177881002e-05,
|
| 283 |
+
"loss": 7.122,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.23,
|
| 288 |
+
"learning_rate": 3.8676445078349366e-05,
|
| 289 |
+
"loss": 7.0971,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.23,
|
| 294 |
+
"learning_rate": 3.843551837788871e-05,
|
| 295 |
+
"loss": 7.0787,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.24,
|
| 300 |
+
"learning_rate": 3.819459167742806e-05,
|
| 301 |
+
"loss": 7.0807,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.24,
|
| 306 |
+
"learning_rate": 3.7953664976967405e-05,
|
| 307 |
+
"loss": 7.0816,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.25,
|
| 312 |
+
"learning_rate": 3.7712738276506756e-05,
|
| 313 |
+
"loss": 7.0138,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.25,
|
| 318 |
+
"learning_rate": 3.747181157604611e-05,
|
| 319 |
+
"loss": 7.0621,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.26,
|
| 324 |
+
"learning_rate": 3.723088487558545e-05,
|
| 325 |
+
"loss": 6.9711,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.26,
|
| 330 |
+
"learning_rate": 3.69899581751248e-05,
|
| 331 |
+
"loss": 6.9892,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.27,
|
| 336 |
+
"learning_rate": 3.6749031474664146e-05,
|
| 337 |
+
"loss": 6.9754,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.27,
|
| 342 |
+
"learning_rate": 3.65081047742035e-05,
|
| 343 |
+
"loss": 6.9831,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.27,
|
| 348 |
+
"learning_rate": 3.626717807374285e-05,
|
| 349 |
+
"loss": 6.9777,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.28,
|
| 354 |
+
"learning_rate": 3.602625137328219e-05,
|
| 355 |
+
"loss": 6.9249,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.28,
|
| 360 |
+
"learning_rate": 3.578532467282154e-05,
|
| 361 |
+
"loss": 6.8832,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.29,
|
| 366 |
+
"learning_rate": 3.5544397972360894e-05,
|
| 367 |
+
"loss": 6.8826,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.29,
|
| 372 |
+
"learning_rate": 3.530347127190024e-05,
|
| 373 |
+
"loss": 6.8175,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.3,
|
| 378 |
+
"learning_rate": 3.506254457143959e-05,
|
| 379 |
+
"loss": 6.9092,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.3,
|
| 384 |
+
"learning_rate": 3.482161787097893e-05,
|
| 385 |
+
"loss": 6.7819,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.31,
|
| 390 |
+
"learning_rate": 3.4580691170518284e-05,
|
| 391 |
+
"loss": 6.8296,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.31,
|
| 396 |
+
"learning_rate": 3.4339764470057635e-05,
|
| 397 |
+
"loss": 6.8417,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.32,
|
| 402 |
+
"learning_rate": 3.409883776959698e-05,
|
| 403 |
+
"loss": 6.7656,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.32,
|
| 408 |
+
"learning_rate": 3.385791106913633e-05,
|
| 409 |
+
"loss": 6.7687,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.33,
|
| 414 |
+
"learning_rate": 3.3616984368675674e-05,
|
| 415 |
+
"loss": 6.7539,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.33,
|
| 420 |
+
"learning_rate": 3.337605766821502e-05,
|
| 421 |
+
"loss": 6.7836,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.34,
|
| 426 |
+
"learning_rate": 3.3135130967754376e-05,
|
| 427 |
+
"loss": 6.6956,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.34,
|
| 432 |
+
"learning_rate": 3.289420426729372e-05,
|
| 433 |
+
"loss": 6.6429,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.35,
|
| 438 |
+
"learning_rate": 3.265327756683307e-05,
|
| 439 |
+
"loss": 6.7027,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.35,
|
| 444 |
+
"learning_rate": 3.2412350866372415e-05,
|
| 445 |
+
"loss": 6.7303,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.36,
|
| 450 |
+
"learning_rate": 3.217142416591176e-05,
|
| 451 |
+
"loss": 6.6036,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.36,
|
| 456 |
+
"learning_rate": 3.193049746545112e-05,
|
| 457 |
+
"loss": 6.5406,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.37,
|
| 462 |
+
"learning_rate": 3.168957076499046e-05,
|
| 463 |
+
"loss": 6.5995,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.37,
|
| 468 |
+
"learning_rate": 3.144864406452981e-05,
|
| 469 |
+
"loss": 6.5864,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.38,
|
| 474 |
+
"learning_rate": 3.1207717364069156e-05,
|
| 475 |
+
"loss": 6.5612,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.38,
|
| 480 |
+
"learning_rate": 3.09667906636085e-05,
|
| 481 |
+
"loss": 6.5546,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.39,
|
| 486 |
+
"learning_rate": 3.072586396314786e-05,
|
| 487 |
+
"loss": 6.4666,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.39,
|
| 492 |
+
"learning_rate": 3.0484937262687202e-05,
|
| 493 |
+
"loss": 6.4376,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.4,
|
| 498 |
+
"learning_rate": 3.024401056222655e-05,
|
| 499 |
+
"loss": 6.4767,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.4,
|
| 504 |
+
"learning_rate": 3.0003083861765897e-05,
|
| 505 |
+
"loss": 6.4738,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.4,
|
| 510 |
+
"learning_rate": 2.9762157161305248e-05,
|
| 511 |
+
"loss": 6.4136,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.41,
|
| 516 |
+
"learning_rate": 2.9521230460844596e-05,
|
| 517 |
+
"loss": 6.4317,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.41,
|
| 522 |
+
"learning_rate": 2.9280303760383943e-05,
|
| 523 |
+
"loss": 6.4216,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.42,
|
| 528 |
+
"learning_rate": 2.903937705992329e-05,
|
| 529 |
+
"loss": 6.4294,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.42,
|
| 534 |
+
"learning_rate": 2.8798450359462638e-05,
|
| 535 |
+
"loss": 6.3409,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.43,
|
| 540 |
+
"learning_rate": 2.855752365900199e-05,
|
| 541 |
+
"loss": 6.3394,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.43,
|
| 546 |
+
"learning_rate": 2.8316596958541336e-05,
|
| 547 |
+
"loss": 6.3577,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.44,
|
| 552 |
+
"learning_rate": 2.8075670258080684e-05,
|
| 553 |
+
"loss": 6.2582,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.44,
|
| 558 |
+
"learning_rate": 2.783474355762003e-05,
|
| 559 |
+
"loss": 6.2806,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.45,
|
| 564 |
+
"learning_rate": 2.7593816857159376e-05,
|
| 565 |
+
"loss": 6.3391,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.45,
|
| 570 |
+
"learning_rate": 2.735289015669873e-05,
|
| 571 |
+
"loss": 6.3571,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.46,
|
| 576 |
+
"learning_rate": 2.7111963456238077e-05,
|
| 577 |
+
"loss": 6.2282,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.46,
|
| 582 |
+
"learning_rate": 2.6871036755777425e-05,
|
| 583 |
+
"loss": 6.3069,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.47,
|
| 588 |
+
"learning_rate": 2.663011005531677e-05,
|
| 589 |
+
"loss": 6.2606,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.47,
|
| 594 |
+
"learning_rate": 2.6389183354856117e-05,
|
| 595 |
+
"loss": 6.2315,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.48,
|
| 600 |
+
"learning_rate": 2.614825665439547e-05,
|
| 601 |
+
"loss": 6.2534,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.48,
|
| 606 |
+
"learning_rate": 2.590732995393482e-05,
|
| 607 |
+
"loss": 6.1819,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.49,
|
| 612 |
+
"learning_rate": 2.5666403253474163e-05,
|
| 613 |
+
"loss": 6.123,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.49,
|
| 618 |
+
"learning_rate": 2.542547655301351e-05,
|
| 619 |
+
"loss": 6.1913,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.5,
|
| 624 |
+
"learning_rate": 2.5184549852552858e-05,
|
| 625 |
+
"loss": 6.1996,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.5,
|
| 630 |
+
"learning_rate": 2.494362315209221e-05,
|
| 631 |
+
"loss": 6.175,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.51,
|
| 636 |
+
"learning_rate": 2.4702696451631556e-05,
|
| 637 |
+
"loss": 6.1644,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.51,
|
| 642 |
+
"learning_rate": 2.4461769751170904e-05,
|
| 643 |
+
"loss": 6.096,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.52,
|
| 648 |
+
"learning_rate": 2.4220843050710254e-05,
|
| 649 |
+
"loss": 6.0954,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.52,
|
| 654 |
+
"learning_rate": 2.3979916350249602e-05,
|
| 655 |
+
"loss": 6.0694,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.53,
|
| 660 |
+
"learning_rate": 2.373898964978895e-05,
|
| 661 |
+
"loss": 6.0671,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.53,
|
| 666 |
+
"learning_rate": 2.3498062949328297e-05,
|
| 667 |
+
"loss": 6.0728,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.53,
|
| 672 |
+
"learning_rate": 2.3257136248867644e-05,
|
| 673 |
+
"loss": 6.1125,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.54,
|
| 678 |
+
"learning_rate": 2.3016209548406995e-05,
|
| 679 |
+
"loss": 6.0758,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.54,
|
| 684 |
+
"learning_rate": 2.2775282847946343e-05,
|
| 685 |
+
"loss": 6.0084,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.55,
|
| 690 |
+
"learning_rate": 2.253435614748569e-05,
|
| 691 |
+
"loss": 6.06,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.55,
|
| 696 |
+
"learning_rate": 2.2293429447025038e-05,
|
| 697 |
+
"loss": 6.0424,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.56,
|
| 702 |
+
"learning_rate": 2.2052502746564385e-05,
|
| 703 |
+
"loss": 6.1225,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.56,
|
| 708 |
+
"learning_rate": 2.1811576046103736e-05,
|
| 709 |
+
"loss": 6.0063,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.57,
|
| 714 |
+
"learning_rate": 2.157064934564308e-05,
|
| 715 |
+
"loss": 5.9357,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.57,
|
| 720 |
+
"learning_rate": 2.132972264518243e-05,
|
| 721 |
+
"loss": 5.9604,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.58,
|
| 726 |
+
"learning_rate": 2.108879594472178e-05,
|
| 727 |
+
"loss": 6.0146,
|
| 728 |
+
"step": 60000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.58,
|
| 732 |
+
"learning_rate": 2.0847869244261126e-05,
|
| 733 |
+
"loss": 5.9849,
|
| 734 |
+
"step": 60500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.59,
|
| 738 |
+
"learning_rate": 2.0606942543800474e-05,
|
| 739 |
+
"loss": 6.0048,
|
| 740 |
+
"step": 61000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.59,
|
| 744 |
+
"learning_rate": 2.036601584333982e-05,
|
| 745 |
+
"loss": 5.9397,
|
| 746 |
+
"step": 61500
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.6,
|
| 750 |
+
"learning_rate": 2.0125089142879172e-05,
|
| 751 |
+
"loss": 6.0005,
|
| 752 |
+
"step": 62000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.6,
|
| 756 |
+
"learning_rate": 1.988416244241852e-05,
|
| 757 |
+
"loss": 5.9041,
|
| 758 |
+
"step": 62500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.61,
|
| 762 |
+
"learning_rate": 1.9643235741957867e-05,
|
| 763 |
+
"loss": 5.9352,
|
| 764 |
+
"step": 63000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.61,
|
| 768 |
+
"learning_rate": 1.9402309041497215e-05,
|
| 769 |
+
"loss": 5.9038,
|
| 770 |
+
"step": 63500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.62,
|
| 774 |
+
"learning_rate": 1.9161382341036562e-05,
|
| 775 |
+
"loss": 5.9472,
|
| 776 |
+
"step": 64000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.62,
|
| 780 |
+
"learning_rate": 1.8920455640575913e-05,
|
| 781 |
+
"loss": 5.9163,
|
| 782 |
+
"step": 64500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.63,
|
| 786 |
+
"learning_rate": 1.867952894011526e-05,
|
| 787 |
+
"loss": 5.9089,
|
| 788 |
+
"step": 65000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.63,
|
| 792 |
+
"learning_rate": 1.843860223965461e-05,
|
| 793 |
+
"loss": 5.9437,
|
| 794 |
+
"step": 65500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.64,
|
| 798 |
+
"learning_rate": 1.8197675539193956e-05,
|
| 799 |
+
"loss": 5.8682,
|
| 800 |
+
"step": 66000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.64,
|
| 804 |
+
"learning_rate": 1.7956748838733303e-05,
|
| 805 |
+
"loss": 5.8244,
|
| 806 |
+
"step": 66500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.65,
|
| 810 |
+
"learning_rate": 1.7715822138272654e-05,
|
| 811 |
+
"loss": 5.851,
|
| 812 |
+
"step": 67000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.65,
|
| 816 |
+
"learning_rate": 1.7474895437812e-05,
|
| 817 |
+
"loss": 5.8414,
|
| 818 |
+
"step": 67500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.66,
|
| 822 |
+
"learning_rate": 1.723396873735135e-05,
|
| 823 |
+
"loss": 5.8678,
|
| 824 |
+
"step": 68000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.66,
|
| 828 |
+
"learning_rate": 1.6993042036890697e-05,
|
| 829 |
+
"loss": 5.8232,
|
| 830 |
+
"step": 68500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.66,
|
| 834 |
+
"learning_rate": 1.6752115336430048e-05,
|
| 835 |
+
"loss": 5.8244,
|
| 836 |
+
"step": 69000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.67,
|
| 840 |
+
"learning_rate": 1.6511188635969392e-05,
|
| 841 |
+
"loss": 5.8314,
|
| 842 |
+
"step": 69500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.67,
|
| 846 |
+
"learning_rate": 1.627026193550874e-05,
|
| 847 |
+
"loss": 5.8577,
|
| 848 |
+
"step": 70000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.68,
|
| 852 |
+
"learning_rate": 1.602933523504809e-05,
|
| 853 |
+
"loss": 5.7637,
|
| 854 |
+
"step": 70500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.68,
|
| 858 |
+
"learning_rate": 1.5788408534587438e-05,
|
| 859 |
+
"loss": 5.83,
|
| 860 |
+
"step": 71000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.69,
|
| 864 |
+
"learning_rate": 1.5547481834126785e-05,
|
| 865 |
+
"loss": 5.8128,
|
| 866 |
+
"step": 71500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.69,
|
| 870 |
+
"learning_rate": 1.5306555133666133e-05,
|
| 871 |
+
"loss": 5.7518,
|
| 872 |
+
"step": 72000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.7,
|
| 876 |
+
"learning_rate": 1.5065628433205484e-05,
|
| 877 |
+
"loss": 5.8433,
|
| 878 |
+
"step": 72500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.7,
|
| 882 |
+
"learning_rate": 1.4824701732744831e-05,
|
| 883 |
+
"loss": 5.7624,
|
| 884 |
+
"step": 73000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.71,
|
| 888 |
+
"learning_rate": 1.4583775032284177e-05,
|
| 889 |
+
"loss": 5.7489,
|
| 890 |
+
"step": 73500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.71,
|
| 894 |
+
"learning_rate": 1.4342848331823528e-05,
|
| 895 |
+
"loss": 5.7542,
|
| 896 |
+
"step": 74000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.72,
|
| 900 |
+
"learning_rate": 1.4101921631362874e-05,
|
| 901 |
+
"loss": 5.7664,
|
| 902 |
+
"step": 74500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.72,
|
| 906 |
+
"learning_rate": 1.3860994930902225e-05,
|
| 907 |
+
"loss": 5.7242,
|
| 908 |
+
"step": 75000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.73,
|
| 912 |
+
"learning_rate": 1.362006823044157e-05,
|
| 913 |
+
"loss": 5.7448,
|
| 914 |
+
"step": 75500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.73,
|
| 918 |
+
"learning_rate": 1.3379141529980918e-05,
|
| 919 |
+
"loss": 5.6599,
|
| 920 |
+
"step": 76000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.74,
|
| 924 |
+
"learning_rate": 1.3138214829520267e-05,
|
| 925 |
+
"loss": 5.8041,
|
| 926 |
+
"step": 76500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.74,
|
| 930 |
+
"learning_rate": 1.2897288129059615e-05,
|
| 931 |
+
"loss": 5.7124,
|
| 932 |
+
"step": 77000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.75,
|
| 936 |
+
"learning_rate": 1.2656361428598964e-05,
|
| 937 |
+
"loss": 5.7687,
|
| 938 |
+
"step": 77500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.75,
|
| 942 |
+
"learning_rate": 1.2415434728138312e-05,
|
| 943 |
+
"loss": 5.7568,
|
| 944 |
+
"step": 78000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.76,
|
| 948 |
+
"learning_rate": 1.217450802767766e-05,
|
| 949 |
+
"loss": 5.729,
|
| 950 |
+
"step": 78500
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.76,
|
| 954 |
+
"learning_rate": 1.1933581327217008e-05,
|
| 955 |
+
"loss": 5.7368,
|
| 956 |
+
"step": 79000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.77,
|
| 960 |
+
"learning_rate": 1.1692654626756357e-05,
|
| 961 |
+
"loss": 5.7522,
|
| 962 |
+
"step": 79500
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.77,
|
| 966 |
+
"learning_rate": 1.1451727926295705e-05,
|
| 967 |
+
"loss": 5.7453,
|
| 968 |
+
"step": 80000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.78,
|
| 972 |
+
"learning_rate": 1.1210801225835052e-05,
|
| 973 |
+
"loss": 5.6756,
|
| 974 |
+
"step": 80500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.78,
|
| 978 |
+
"learning_rate": 1.09698745253744e-05,
|
| 979 |
+
"loss": 5.764,
|
| 980 |
+
"step": 81000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.79,
|
| 984 |
+
"learning_rate": 1.072894782491375e-05,
|
| 985 |
+
"loss": 5.6947,
|
| 986 |
+
"step": 81500
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.79,
|
| 990 |
+
"learning_rate": 1.0488021124453097e-05,
|
| 991 |
+
"loss": 5.6659,
|
| 992 |
+
"step": 82000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.8,
|
| 996 |
+
"learning_rate": 1.0247094423992446e-05,
|
| 997 |
+
"loss": 5.7086,
|
| 998 |
+
"step": 82500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.8,
|
| 1002 |
+
"learning_rate": 1.0006167723531793e-05,
|
| 1003 |
+
"loss": 5.6405,
|
| 1004 |
+
"step": 83000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.8,
|
| 1008 |
+
"learning_rate": 9.765241023071141e-06,
|
| 1009 |
+
"loss": 5.6981,
|
| 1010 |
+
"step": 83500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.81,
|
| 1014 |
+
"learning_rate": 9.524314322610489e-06,
|
| 1015 |
+
"loss": 5.732,
|
| 1016 |
+
"step": 84000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.81,
|
| 1020 |
+
"learning_rate": 9.283387622149838e-06,
|
| 1021 |
+
"loss": 5.6404,
|
| 1022 |
+
"step": 84500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.82,
|
| 1026 |
+
"learning_rate": 9.042460921689185e-06,
|
| 1027 |
+
"loss": 5.6796,
|
| 1028 |
+
"step": 85000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.82,
|
| 1032 |
+
"learning_rate": 8.801534221228534e-06,
|
| 1033 |
+
"loss": 5.7269,
|
| 1034 |
+
"step": 85500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.83,
|
| 1038 |
+
"learning_rate": 8.560607520767882e-06,
|
| 1039 |
+
"loss": 5.6806,
|
| 1040 |
+
"step": 86000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.83,
|
| 1044 |
+
"learning_rate": 8.31968082030723e-06,
|
| 1045 |
+
"loss": 5.5852,
|
| 1046 |
+
"step": 86500
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.84,
|
| 1050 |
+
"learning_rate": 8.078754119846579e-06,
|
| 1051 |
+
"loss": 5.6754,
|
| 1052 |
+
"step": 87000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.84,
|
| 1056 |
+
"learning_rate": 7.837827419385926e-06,
|
| 1057 |
+
"loss": 5.6727,
|
| 1058 |
+
"step": 87500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.85,
|
| 1062 |
+
"learning_rate": 7.5969007189252746e-06,
|
| 1063 |
+
"loss": 5.7247,
|
| 1064 |
+
"step": 88000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.85,
|
| 1068 |
+
"learning_rate": 7.355974018464623e-06,
|
| 1069 |
+
"loss": 5.6207,
|
| 1070 |
+
"step": 88500
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.86,
|
| 1074 |
+
"learning_rate": 7.115047318003971e-06,
|
| 1075 |
+
"loss": 5.6297,
|
| 1076 |
+
"step": 89000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.86,
|
| 1080 |
+
"learning_rate": 6.874120617543319e-06,
|
| 1081 |
+
"loss": 5.6616,
|
| 1082 |
+
"step": 89500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.87,
|
| 1086 |
+
"learning_rate": 6.633193917082667e-06,
|
| 1087 |
+
"loss": 5.6567,
|
| 1088 |
+
"step": 90000
|
| 1089 |
+
}
|
| 1090 |
+
],
|
| 1091 |
+
"max_steps": 103766,
|
| 1092 |
+
"num_train_epochs": 1,
|
| 1093 |
+
"total_flos": 7418391121250304.0,
|
| 1094 |
+
"trial_name": null,
|
| 1095 |
+
"trial_params": null
|
| 1096 |
+
}
|
rICE/checkpoint-90000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9de5a40da776e5cbb26e7943f3fb17679c7665327d670d554ee58b0243cb6bd0
|
| 3 |
+
size 3311
|