Hugging Face's logo

vocab-transformers
/
msmarco-distilbert-word2vec256k-MLM_210k_emb_updated

Fill-Mask
Transformers
PyTorch
distilbert
Inference Endpoints
Model card Files Community
nreimers commited on
Commit
b8115aa
·
1 Parent(s): 9658e26

upload

Browse files
Files changed (13) hide show
  1. README.md +4 -0
  2. config.json +24 -0
  3. optimizer.pt +3 -0
  4. pytorch_model.bin +3 -0
  5. rng_state.pth +3 -0
  6. scaler.pt +3 -0
  7. scheduler.pt +3 -0
  8. special_tokens_map.json +1 -0
  9. tokenizer.json +0 -0
  10. tokenizer_config.json +1 -0
  11. train_mlm.py +120 -0
  12. trainer_state.json +4216 -0
  13. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ # Model
2
+ This model is based on [nicoladecao/msmarco-word2vec256000-distilbert-base-uncased](https://huggingface.co/nicoladecao/msmarco-word2vec256000-distilbert-base-uncased) with a 256k sized vocabulary initialized with word2vec.
3
+
4
+ This model has been trained with MLM on the MS MARCO corpus collection for 210k steps. See train_mlm.py for the train script. It was run on 2x V100 GPUs.
config.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "nicoladecao/msmarco-word2vec256000-distilbert-base-uncased",
3
+ "activation": "gelu",
4
+ "architectures": [
5
+ "DistilBertForMaskedLM"
6
+ ],
7
+ "attention_dropout": 0.1,
8
+ "dim": 768,
9
+ "dropout": 0.1,
10
+ "hidden_dim": 3072,
11
+ "initializer_range": 0.02,
12
+ "max_position_embeddings": 512,
13
+ "model_type": "distilbert",
14
+ "n_heads": 12,
15
+ "n_layers": 6,
16
+ "pad_token_id": 0,
17
+ "qa_dropout": 0.1,
18
+ "seq_classif_dropout": 0.2,
19
+ "sinusoidal_pos_embds": false,
20
+ "tie_weights_": true,
21
+ "torch_dtype": "float32",
22
+ "transformers_version": "4.11.3",
23
+ "vocab_size": 256000
24
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9618397ace5f1e250db89c4a08d1cbaf10d400de1f2e2bdb6c2c6c5b1aaed63
3
+ size 1923091089
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:864b2c91e4c3a2ba615418e016df3f526c7f639c08b6e4949a9db967597e5072
3
+ size 961556128
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49e8bc1f2bc5dbad22f15a47e3834432d4cc01a3830fb82965cfec2fd9cc90fa
3
+ size 15627
scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c43847b549b235cd400c368891008a622f3ed088911686975a0ee89e18836821
3
+ size 559
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c55f63d9b6953e879f913b188e3a7d6bee07b8e825f67f55f7a29fd41363e839
3
+ size 623
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"model_max_length": 512, "unk_token": "[UNK]", "cls_token": "[CLS]", "sep_token": "[SEP]", "pad_token": "[PAD]", "mask_token": "[MASK]", "model_input_names": ["input_ids", "attention_mask"], "special_tokens_map_file": "/root/.cache/huggingface/transformers/fe09c361189d8238b9e387f10a088e93f70620bfe74b82036baff1fed512a153.dd8bd9bfd3664b530ea4e645105f557769387b3da9f79bdb55ed556bdd80611d", "name_or_path": "nicoladecao/msmarco-word2vec256000-distilbert-base-uncased", "tokenizer_class": "PreTrainedTokenizerFast"}
train_mlm.py ADDED
@@ -0,0 +1,120 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """
2
+ This file runs Masked Language Model. You provide a training file. Each line is interpreted as a sentence / paragraph.
3
+ Optionally, you can also provide a dev file.
4
+
5
+ The fine-tuned model is stored in the output/model_name folder.
6
+
7
+ python train_mlm.py model_name data/train_sentences.txt [data/dev_sentences.txt]
8
+ """
9
+
10
+ from transformers import AutoModelForMaskedLM, AutoTokenizer
11
+ from transformers import DataCollatorForLanguageModeling, DataCollatorForWholeWordMask
12
+ from transformers import Trainer, TrainingArguments
13
+ import sys
14
+ import gzip
15
+ from datetime import datetime
16
+ import wandb
17
+
18
+ wandb.init(project="bert-word2vec")
19
+
20
+ model_name = "nicoladecao/msmarco-word2vec256000-distilbert-base-uncased"
21
+ per_device_train_batch_size = 16
22
+ save_steps = 5000
23
+ eval_steps = 1000
24
+ num_train_epochs = 3
25
+ use_fp16 = True #Set to True, if your GPU supports FP16 operations
26
+ max_length = 250 #Max length for a text input
27
+ do_whole_word_mask = True #If set to true, whole words are masked
28
+ mlm_prob = 15 #Probability that a word is replaced by a [MASK] token
29
+
30
+ model = AutoModelForMaskedLM.from_pretrained(model_name)
31
+ tokenizer = AutoTokenizer.from_pretrained(model_name)
32
+
33
+ ## Freeze embedding layer
34
+ model.distilbert.embeddings.requires_grad = False
35
+
36
+ output_dir = "output/{}-{}".format(model_name.replace("/", "_"), datetime.now().strftime("%Y-%m-%d_%H-%M-%S"))
37
+ print("Save checkpoints to:", output_dir)
38
+
39
+
40
+ ##### Load our training datasets
41
+
42
+ train_sentences = []
43
+ train_path = 'data/train.txt'
44
+ with gzip.open(train_path, 'rt', encoding='utf8') if train_path.endswith('.gz') else open(train_path, 'r', encoding='utf8') as fIn:
45
+ for line in fIn:
46
+ line = line.strip()
47
+ if len(line) >= 10:
48
+ train_sentences.append(line)
49
+
50
+ print("Train sentences:", len(train_sentences))
51
+
52
+ dev_sentences = []
53
+
54
+ dev_path = 'data/dev.txt'
55
+ with gzip.open(dev_path, 'rt', encoding='utf8') if dev_path.endswith('.gz') else open(dev_path, 'r', encoding='utf8') as fIn:
56
+ for line in fIn:
57
+ line = line.strip()
58
+ if len(line) >= 10:
59
+ dev_sentences.append(line)
60
+
61
+ print("Dev sentences:", len(dev_sentences))
62
+
63
+ #A dataset wrapper, that tokenizes our data on-the-fly
64
+ class TokenizedSentencesDataset:
65
+ def __init__(self, sentences, tokenizer, max_length, cache_tokenization=False):
66
+ self.tokenizer = tokenizer
67
+ self.sentences = sentences
68
+ self.max_length = max_length
69
+ self.cache_tokenization = cache_tokenization
70
+
71
+ def __getitem__(self, item):
72
+ if not self.cache_tokenization:
73
+ return self.tokenizer(self.sentences[item], add_special_tokens=True, truncation=True, max_length=self.max_length, return_special_tokens_mask=True)
74
+
75
+ if isinstance(self.sentences[item], str):
76
+ self.sentences[item] = self.tokenizer(self.sentences[item], add_special_tokens=True, truncation=True, max_length=self.max_length, return_special_tokens_mask=True)
77
+ return self.sentences[item]
78
+
79
+ def __len__(self):
80
+ return len(self.sentences)
81
+
82
+ train_dataset = TokenizedSentencesDataset(train_sentences, tokenizer, max_length)
83
+ dev_dataset = TokenizedSentencesDataset(dev_sentences, tokenizer, max_length, cache_tokenization=True) if len(dev_sentences) > 0 else None
84
+
85
+
86
+ ##### Training arguments
87
+
88
+ if do_whole_word_mask:
89
+ data_collator = DataCollatorForWholeWordMask(tokenizer=tokenizer, mlm=True, mlm_probability=mlm_prob)
90
+ else:
91
+ data_collator = DataCollatorForLanguageModeling(tokenizer=tokenizer, mlm=True, mlm_probability=mlm_prob)
92
+
93
+ training_args = TrainingArguments(
94
+ output_dir=output_dir,
95
+ overwrite_output_dir=True,
96
+ num_train_epochs=num_train_epochs,
97
+ evaluation_strategy="steps" if dev_dataset is not None else "no",
98
+ per_device_train_batch_size=per_device_train_batch_size,
99
+ eval_steps=eval_steps,
100
+ save_steps=save_steps,
101
+ save_total_limit=1,
102
+ prediction_loss_only=True,
103
+ fp16=use_fp16
104
+ )
105
+
106
+ trainer = Trainer(
107
+ model=model,
108
+ args=training_args,
109
+ data_collator=data_collator,
110
+ train_dataset=train_dataset,
111
+ eval_dataset=dev_dataset
112
+ )
113
+
114
+ trainer.train()
115
+
116
+ print("Save model to:", output_dir)
117
+ model.save_pretrained(output_dir)
118
+ tokenizer.save_pretrained(output_dir)
119
+
120
+ print("Training done")
trainer_state.json ADDED
@@ -0,0 +1,4216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.7601094557616297,
5
+ "global_step": 210000,
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": 4.996983692635866e-05,
13
+ "loss": 7.2434,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 4.9939673852717336e-05,
19
+ "loss": 6.3951,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "eval_loss": 6.350970268249512,
25
+ "eval_runtime": 3.8385,
26
+ "eval_samples_per_second": 260.78,
27
+ "eval_steps_per_second": 16.413,
28
+ "step": 1000
29
+ },
30
+ {
31
+ "epoch": 0.01,
32
+ "learning_rate": 4.9909510779075997e-05,
33
+ "loss": 6.2791,
34
+ "step": 1500
35
+ },
36
+ {
37
+ "epoch": 0.01,
38
+ "learning_rate": 4.9879347705434663e-05,
39
+ "loss": 6.2011,
40
+ "step": 2000
41
+ },
42
+ {
43
+ "epoch": 0.01,
44
+ "eval_loss": 6.223337173461914,
45
+ "eval_runtime": 3.7358,
46
+ "eval_samples_per_second": 267.946,
47
+ "eval_steps_per_second": 16.864,
48
+ "step": 2000
49
+ },
50
+ {
51
+ "epoch": 0.01,
52
+ "learning_rate": 4.984918463179333e-05,
53
+ "loss": 6.15,
54
+ "step": 2500
55
+ },
56
+ {
57
+ "epoch": 0.01,
58
+ "learning_rate": 4.9819021558152e-05,
59
+ "loss": 6.121,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.01,
64
+ "eval_loss": 6.134737014770508,
65
+ "eval_runtime": 3.7442,
66
+ "eval_samples_per_second": 267.346,
67
+ "eval_steps_per_second": 16.826,
68
+ "step": 3000
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 4.978885848451066e-05,
73
+ "loss": 6.0839,
74
+ "step": 3500
75
+ },
76
+ {
77
+ "epoch": 0.01,
78
+ "learning_rate": 4.9758695410869325e-05,
79
+ "loss": 6.0626,
80
+ "step": 4000
81
+ },
82
+ {
83
+ "epoch": 0.01,
84
+ "eval_loss": 6.063777923583984,
85
+ "eval_runtime": 4.0904,
86
+ "eval_samples_per_second": 244.72,
87
+ "eval_steps_per_second": 15.402,
88
+ "step": 4000
89
+ },
90
+ {
91
+ "epoch": 0.02,
92
+ "learning_rate": 4.972853233722799e-05,
93
+ "loss": 6.0263,
94
+ "step": 4500
95
+ },
96
+ {
97
+ "epoch": 0.02,
98
+ "learning_rate": 4.969836926358666e-05,
99
+ "loss": 6.0177,
100
+ "step": 5000
101
+ },
102
+ {
103
+ "epoch": 0.02,
104
+ "eval_loss": 6.02394962310791,
105
+ "eval_runtime": 3.9573,
106
+ "eval_samples_per_second": 252.948,
107
+ "eval_steps_per_second": 15.92,
108
+ "step": 5000
109
+ },
110
+ {
111
+ "epoch": 0.02,
112
+ "learning_rate": 4.966820618994532e-05,
113
+ "loss": 5.9988,
114
+ "step": 5500
115
+ },
116
+ {
117
+ "epoch": 0.02,
118
+ "learning_rate": 4.963804311630399e-05,
119
+ "loss": 5.9921,
120
+ "step": 6000
121
+ },
122
+ {
123
+ "epoch": 0.02,
124
+ "eval_loss": 6.0168046951293945,
125
+ "eval_runtime": 4.0579,
126
+ "eval_samples_per_second": 246.677,
127
+ "eval_steps_per_second": 15.525,
128
+ "step": 6000
129
+ },
130
+ {
131
+ "epoch": 0.02,
132
+ "learning_rate": 4.9607940368809935e-05,
133
+ "loss": 5.9751,
134
+ "step": 6500
135
+ },
136
+ {
137
+ "epoch": 0.03,
138
+ "learning_rate": 4.9577837621315885e-05,
139
+ "loss": 5.9632,
140
+ "step": 7000
141
+ },
142
+ {
143
+ "epoch": 0.03,
144
+ "eval_loss": 5.997592926025391,
145
+ "eval_runtime": 4.2019,
146
+ "eval_samples_per_second": 238.224,
147
+ "eval_steps_per_second": 14.993,
148
+ "step": 7000
149
+ },
150
+ {
151
+ "epoch": 0.03,
152
+ "learning_rate": 4.954767454767455e-05,
153
+ "loss": 5.9386,
154
+ "step": 7500
155
+ },
156
+ {
157
+ "epoch": 0.03,
158
+ "learning_rate": 4.951751147403322e-05,
159
+ "loss": 5.9524,
160
+ "step": 8000
161
+ },
162
+ {
163
+ "epoch": 0.03,
164
+ "eval_loss": 5.969239711761475,
165
+ "eval_runtime": 4.0231,
166
+ "eval_samples_per_second": 248.812,
167
+ "eval_steps_per_second": 15.659,
168
+ "step": 8000
169
+ },
170
+ {
171
+ "epoch": 0.03,
172
+ "learning_rate": 4.948734840039188e-05,
173
+ "loss": 5.9376,
174
+ "step": 8500
175
+ },
176
+ {
177
+ "epoch": 0.03,
178
+ "learning_rate": 4.9457185326750546e-05,
179
+ "loss": 5.9328,
180
+ "step": 9000
181
+ },
182
+ {
183
+ "epoch": 0.03,
184
+ "eval_loss": 5.940969944000244,
185
+ "eval_runtime": 3.8844,
186
+ "eval_samples_per_second": 257.697,
187
+ "eval_steps_per_second": 16.219,
188
+ "step": 9000
189
+ },
190
+ {
191
+ "epoch": 0.03,
192
+ "learning_rate": 4.942702225310921e-05,
193
+ "loss": 5.9177,
194
+ "step": 9500
195
+ },
196
+ {
197
+ "epoch": 0.04,
198
+ "learning_rate": 4.939685917946787e-05,
199
+ "loss": 5.9162,
200
+ "step": 10000
201
+ },
202
+ {
203
+ "epoch": 0.04,
204
+ "eval_loss": 5.932608604431152,
205
+ "eval_runtime": 4.0642,
206
+ "eval_samples_per_second": 246.298,
207
+ "eval_steps_per_second": 15.501,
208
+ "step": 10000
209
+ },
210
+ {
211
+ "epoch": 0.04,
212
+ "learning_rate": 4.936669610582654e-05,
213
+ "loss": 5.9054,
214
+ "step": 10500
215
+ },
216
+ {
217
+ "epoch": 0.04,
218
+ "learning_rate": 4.933659335833249e-05,
219
+ "loss": 5.8906,
220
+ "step": 11000
221
+ },
222
+ {
223
+ "epoch": 0.04,
224
+ "eval_loss": 5.921965599060059,
225
+ "eval_runtime": 4.2691,
226
+ "eval_samples_per_second": 234.477,
227
+ "eval_steps_per_second": 14.757,
228
+ "step": 11000
229
+ },
230
+ {
231
+ "epoch": 0.04,
232
+ "learning_rate": 4.9306430284691156e-05,
233
+ "loss": 5.8952,
234
+ "step": 11500
235
+ },
236
+ {
237
+ "epoch": 0.04,
238
+ "learning_rate": 4.9276267211049823e-05,
239
+ "loss": 5.8868,
240
+ "step": 12000
241
+ },
242
+ {
243
+ "epoch": 0.04,
244
+ "eval_loss": 5.910747528076172,
245
+ "eval_runtime": 4.3284,
246
+ "eval_samples_per_second": 231.264,
247
+ "eval_steps_per_second": 14.555,
248
+ "step": 12000
249
+ },
250
+ {
251
+ "epoch": 0.05,
252
+ "learning_rate": 4.924610413740849e-05,
253
+ "loss": 5.8829,
254
+ "step": 12500
255
+ },
256
+ {
257
+ "epoch": 0.05,
258
+ "learning_rate": 4.921600138991443e-05,
259
+ "loss": 5.8738,
260
+ "step": 13000
261
+ },
262
+ {
263
+ "epoch": 0.05,
264
+ "eval_loss": 5.895788669586182,
265
+ "eval_runtime": 4.2991,
266
+ "eval_samples_per_second": 232.842,
267
+ "eval_steps_per_second": 14.654,
268
+ "step": 13000
269
+ },
270
+ {
271
+ "epoch": 0.05,
272
+ "learning_rate": 4.91858383162731e-05,
273
+ "loss": 5.8717,
274
+ "step": 13500
275
+ },
276
+ {
277
+ "epoch": 0.05,
278
+ "learning_rate": 4.915567524263177e-05,
279
+ "loss": 5.8642,
280
+ "step": 14000
281
+ },
282
+ {
283
+ "epoch": 0.05,
284
+ "eval_loss": 5.891711235046387,
285
+ "eval_runtime": 4.3228,
286
+ "eval_samples_per_second": 231.564,
287
+ "eval_steps_per_second": 14.574,
288
+ "step": 14000
289
+ },
290
+ {
291
+ "epoch": 0.05,
292
+ "learning_rate": 4.9125512168990434e-05,
293
+ "loss": 5.8667,
294
+ "step": 14500
295
+ },
296
+ {
297
+ "epoch": 0.05,
298
+ "learning_rate": 4.909540942149638e-05,
299
+ "loss": 5.8494,
300
+ "step": 15000
301
+ },
302
+ {
303
+ "epoch": 0.05,
304
+ "eval_loss": 5.886107921600342,
305
+ "eval_runtime": 4.4064,
306
+ "eval_samples_per_second": 227.167,
307
+ "eval_steps_per_second": 14.297,
308
+ "step": 15000
309
+ },
310
+ {
311
+ "epoch": 0.06,
312
+ "learning_rate": 4.906524634785505e-05,
313
+ "loss": 5.8369,
314
+ "step": 15500
315
+ },
316
+ {
317
+ "epoch": 0.06,
318
+ "learning_rate": 4.903508327421371e-05,
319
+ "loss": 5.8414,
320
+ "step": 16000
321
+ },
322
+ {
323
+ "epoch": 0.06,
324
+ "eval_loss": 5.89373254776001,
325
+ "eval_runtime": 4.3188,
326
+ "eval_samples_per_second": 231.776,
327
+ "eval_steps_per_second": 14.587,
328
+ "step": 16000
329
+ },
330
+ {
331
+ "epoch": 0.06,
332
+ "learning_rate": 4.900492020057238e-05,
333
+ "loss": 5.851,
334
+ "step": 16500
335
+ },
336
+ {
337
+ "epoch": 0.06,
338
+ "learning_rate": 4.8974757126931045e-05,
339
+ "loss": 5.8512,
340
+ "step": 17000
341
+ },
342
+ {
343
+ "epoch": 0.06,
344
+ "eval_loss": 5.9001264572143555,
345
+ "eval_runtime": 4.3173,
346
+ "eval_samples_per_second": 231.857,
347
+ "eval_steps_per_second": 14.592,
348
+ "step": 17000
349
+ },
350
+ {
351
+ "epoch": 0.06,
352
+ "learning_rate": 4.8944594053289705e-05,
353
+ "loss": 5.8471,
354
+ "step": 17500
355
+ },
356
+ {
357
+ "epoch": 0.07,
358
+ "learning_rate": 4.8914491305795654e-05,
359
+ "loss": 5.8319,
360
+ "step": 18000
361
+ },
362
+ {
363
+ "epoch": 0.07,
364
+ "eval_loss": 5.870309829711914,
365
+ "eval_runtime": 4.2282,
366
+ "eval_samples_per_second": 236.743,
367
+ "eval_steps_per_second": 14.9,
368
+ "step": 18000
369
+ },
370
+ {
371
+ "epoch": 0.07,
372
+ "learning_rate": 4.888432823215432e-05,
373
+ "loss": 5.8285,
374
+ "step": 18500
375
+ },
376
+ {
377
+ "epoch": 0.07,
378
+ "learning_rate": 4.885416515851299e-05,
379
+ "loss": 5.8282,
380
+ "step": 19000
381
+ },
382
+ {
383
+ "epoch": 0.07,
384
+ "eval_loss": 5.834588050842285,
385
+ "eval_runtime": 4.2534,
386
+ "eval_samples_per_second": 235.344,
387
+ "eval_steps_per_second": 14.812,
388
+ "step": 19000
389
+ },
390
+ {
391
+ "epoch": 0.07,
392
+ "learning_rate": 4.882400208487165e-05,
393
+ "loss": 5.8305,
394
+ "step": 19500
395
+ },
396
+ {
397
+ "epoch": 0.07,
398
+ "learning_rate": 4.879383901123032e-05,
399
+ "loss": 5.8159,
400
+ "step": 20000
401
+ },
402
+ {
403
+ "epoch": 0.07,
404
+ "eval_loss": 5.825201034545898,
405
+ "eval_runtime": 4.2051,
406
+ "eval_samples_per_second": 238.046,
407
+ "eval_steps_per_second": 14.982,
408
+ "step": 20000
409
+ },
410
+ {
411
+ "epoch": 0.07,
412
+ "learning_rate": 4.876367593758898e-05,
413
+ "loss": 5.8109,
414
+ "step": 20500
415
+ },
416
+ {
417
+ "epoch": 0.08,
418
+ "learning_rate": 4.873357319009493e-05,
419
+ "loss": 5.8206,
420
+ "step": 21000
421
+ },
422
+ {
423
+ "epoch": 0.08,
424
+ "eval_loss": 5.839576244354248,
425
+ "eval_runtime": 4.2748,
426
+ "eval_samples_per_second": 234.162,
427
+ "eval_steps_per_second": 14.737,
428
+ "step": 21000
429
+ },
430
+ {
431
+ "epoch": 0.08,
432
+ "learning_rate": 4.87034101164536e-05,
433
+ "loss": 5.8187,
434
+ "step": 21500
435
+ },
436
+ {
437
+ "epoch": 0.08,
438
+ "learning_rate": 4.8673247042812266e-05,
439
+ "loss": 5.826,
440
+ "step": 22000
441
+ },
442
+ {
443
+ "epoch": 0.08,
444
+ "eval_loss": 5.854990482330322,
445
+ "eval_runtime": 6.3825,
446
+ "eval_samples_per_second": 156.835,
447
+ "eval_steps_per_second": 9.871,
448
+ "step": 22000
449
+ },
450
+ {
451
+ "epoch": 0.08,
452
+ "learning_rate": 4.8643083969170926e-05,
453
+ "loss": 5.8221,
454
+ "step": 22500
455
+ },
456
+ {
457
+ "epoch": 0.08,
458
+ "learning_rate": 4.861292089552959e-05,
459
+ "loss": 5.8109,
460
+ "step": 23000
461
+ },
462
+ {
463
+ "epoch": 0.08,
464
+ "eval_loss": 5.85267448425293,
465
+ "eval_runtime": 4.2815,
466
+ "eval_samples_per_second": 233.795,
467
+ "eval_steps_per_second": 14.714,
468
+ "step": 23000
469
+ },
470
+ {
471
+ "epoch": 0.09,
472
+ "learning_rate": 4.858275782188826e-05,
473
+ "loss": 5.8124,
474
+ "step": 23500
475
+ },
476
+ {
477
+ "epoch": 0.09,
478
+ "learning_rate": 4.855259474824693e-05,
479
+ "loss": 5.8002,
480
+ "step": 24000
481
+ },
482
+ {
483
+ "epoch": 0.09,
484
+ "eval_loss": 5.857957363128662,
485
+ "eval_runtime": 4.2472,
486
+ "eval_samples_per_second": 235.683,
487
+ "eval_steps_per_second": 14.833,
488
+ "step": 24000
489
+ },
490
+ {
491
+ "epoch": 0.09,
492
+ "learning_rate": 4.852243167460559e-05,
493
+ "loss": 5.7933,
494
+ "step": 24500
495
+ },
496
+ {
497
+ "epoch": 0.09,
498
+ "learning_rate": 4.849232892711154e-05,
499
+ "loss": 5.7957,
500
+ "step": 25000
501
+ },
502
+ {
503
+ "epoch": 0.09,
504
+ "eval_loss": 5.843862533569336,
505
+ "eval_runtime": 4.2251,
506
+ "eval_samples_per_second": 236.92,
507
+ "eval_steps_per_second": 14.911,
508
+ "step": 25000
509
+ },
510
+ {
511
+ "epoch": 0.09,
512
+ "learning_rate": 4.8462165853470204e-05,
513
+ "loss": 5.7919,
514
+ "step": 25500
515
+ },
516
+ {
517
+ "epoch": 0.09,
518
+ "learning_rate": 4.8432002779828864e-05,
519
+ "loss": 5.8001,
520
+ "step": 26000
521
+ },
522
+ {
523
+ "epoch": 0.09,
524
+ "eval_loss": 5.8175740242004395,
525
+ "eval_runtime": 4.5898,
526
+ "eval_samples_per_second": 218.094,
527
+ "eval_steps_per_second": 13.726,
528
+ "step": 26000
529
+ },
530
+ {
531
+ "epoch": 0.1,
532
+ "learning_rate": 4.840183970618754e-05,
533
+ "loss": 5.7874,
534
+ "step": 26500
535
+ },
536
+ {
537
+ "epoch": 0.1,
538
+ "learning_rate": 4.837173695869348e-05,
539
+ "loss": 5.7836,
540
+ "step": 27000
541
+ },
542
+ {
543
+ "epoch": 0.1,
544
+ "eval_loss": 5.7963480949401855,
545
+ "eval_runtime": 4.2794,
546
+ "eval_samples_per_second": 233.909,
547
+ "eval_steps_per_second": 14.722,
548
+ "step": 27000
549
+ },
550
+ {
551
+ "epoch": 0.1,
552
+ "learning_rate": 4.834157388505215e-05,
553
+ "loss": 5.7915,
554
+ "step": 27500
555
+ },
556
+ {
557
+ "epoch": 0.1,
558
+ "learning_rate": 4.8311410811410814e-05,
559
+ "loss": 5.7797,
560
+ "step": 28000
561
+ },
562
+ {
563
+ "epoch": 0.1,
564
+ "eval_loss": 5.796699047088623,
565
+ "eval_runtime": 4.2933,
566
+ "eval_samples_per_second": 233.155,
567
+ "eval_steps_per_second": 14.674,
568
+ "step": 28000
569
+ },
570
+ {
571
+ "epoch": 0.1,
572
+ "learning_rate": 4.828124773776948e-05,
573
+ "loss": 5.7768,
574
+ "step": 28500
575
+ },
576
+ {
577
+ "epoch": 0.1,
578
+ "learning_rate": 4.8251144990275424e-05,
579
+ "loss": 5.7876,
580
+ "step": 29000
581
+ },
582
+ {
583
+ "epoch": 0.1,
584
+ "eval_loss": 5.799530506134033,
585
+ "eval_runtime": 4.2207,
586
+ "eval_samples_per_second": 237.163,
587
+ "eval_steps_per_second": 14.926,
588
+ "step": 29000
589
+ },
590
+ {
591
+ "epoch": 0.11,
592
+ "learning_rate": 4.82209819166341e-05,
593
+ "loss": 5.7904,
594
+ "step": 29500
595
+ },
596
+ {
597
+ "epoch": 0.11,
598
+ "learning_rate": 4.819081884299276e-05,
599
+ "loss": 5.7837,
600
+ "step": 30000
601
+ },
602
+ {
603
+ "epoch": 0.11,
604
+ "eval_loss": 5.798836708068848,
605
+ "eval_runtime": 4.2831,
606
+ "eval_samples_per_second": 233.71,
607
+ "eval_steps_per_second": 14.709,
608
+ "step": 30000
609
+ },
610
+ {
611
+ "epoch": 0.11,
612
+ "learning_rate": 4.8160655769351425e-05,
613
+ "loss": 5.778,
614
+ "step": 30500
615
+ },
616
+ {
617
+ "epoch": 0.11,
618
+ "learning_rate": 4.813049269571009e-05,
619
+ "loss": 5.7681,
620
+ "step": 31000
621
+ },
622
+ {
623
+ "epoch": 0.11,
624
+ "eval_loss": 5.811764240264893,
625
+ "eval_runtime": 4.3287,
626
+ "eval_samples_per_second": 231.246,
627
+ "eval_steps_per_second": 14.554,
628
+ "step": 31000
629
+ },
630
+ {
631
+ "epoch": 0.11,
632
+ "learning_rate": 4.810032962206876e-05,
633
+ "loss": 5.7695,
634
+ "step": 31500
635
+ },
636
+ {
637
+ "epoch": 0.12,
638
+ "learning_rate": 4.80702268745747e-05,
639
+ "loss": 5.7698,
640
+ "step": 32000
641
+ },
642
+ {
643
+ "epoch": 0.12,
644
+ "eval_loss": 5.799587726593018,
645
+ "eval_runtime": 6.2437,
646
+ "eval_samples_per_second": 160.322,
647
+ "eval_steps_per_second": 10.09,
648
+ "step": 32000
649
+ },
650
+ {
651
+ "epoch": 0.12,
652
+ "learning_rate": 4.804006380093337e-05,
653
+ "loss": 5.7613,
654
+ "step": 32500
655
+ },
656
+ {
657
+ "epoch": 0.12,
658
+ "learning_rate": 4.8009900727292036e-05,
659
+ "loss": 5.7875,
660
+ "step": 33000
661
+ },
662
+ {
663
+ "epoch": 0.12,
664
+ "eval_loss": 5.817694664001465,
665
+ "eval_runtime": 4.2624,
666
+ "eval_samples_per_second": 234.843,
667
+ "eval_steps_per_second": 14.78,
668
+ "step": 33000
669
+ },
670
+ {
671
+ "epoch": 0.12,
672
+ "learning_rate": 4.7979737653650696e-05,
673
+ "loss": 5.7868,
674
+ "step": 33500
675
+ },
676
+ {
677
+ "epoch": 0.12,
678
+ "learning_rate": 4.794957458000936e-05,
679
+ "loss": 5.7672,
680
+ "step": 34000
681
+ },
682
+ {
683
+ "epoch": 0.12,
684
+ "eval_loss": 5.817095756530762,
685
+ "eval_runtime": 4.3001,
686
+ "eval_samples_per_second": 232.788,
687
+ "eval_steps_per_second": 14.651,
688
+ "step": 34000
689
+ },
690
+ {
691
+ "epoch": 0.12,
692
+ "learning_rate": 4.791947183251531e-05,
693
+ "loss": 5.7585,
694
+ "step": 34500
695
+ },
696
+ {
697
+ "epoch": 0.13,
698
+ "learning_rate": 4.788930875887398e-05,
699
+ "loss": 5.77,
700
+ "step": 35000
701
+ },
702
+ {
703
+ "epoch": 0.13,
704
+ "eval_loss": 5.829194068908691,
705
+ "eval_runtime": 4.4257,
706
+ "eval_samples_per_second": 226.177,
707
+ "eval_steps_per_second": 14.235,
708
+ "step": 35000
709
+ },
710
+ {
711
+ "epoch": 0.13,
712
+ "learning_rate": 4.785914568523264e-05,
713
+ "loss": 5.7525,
714
+ "step": 35500
715
+ },
716
+ {
717
+ "epoch": 0.13,
718
+ "learning_rate": 4.782898261159131e-05,
719
+ "loss": 5.7534,
720
+ "step": 36000
721
+ },
722
+ {
723
+ "epoch": 0.13,
724
+ "eval_loss": 5.81154203414917,
725
+ "eval_runtime": 4.3328,
726
+ "eval_samples_per_second": 231.028,
727
+ "eval_steps_per_second": 14.54,
728
+ "step": 36000
729
+ },
730
+ {
731
+ "epoch": 0.13,
732
+ "learning_rate": 4.7798819537949973e-05,
733
+ "loss": 5.7679,
734
+ "step": 36500
735
+ },
736
+ {
737
+ "epoch": 0.13,
738
+ "learning_rate": 4.776871679045592e-05,
739
+ "loss": 5.7461,
740
+ "step": 37000
741
+ },
742
+ {
743
+ "epoch": 0.13,
744
+ "eval_loss": 5.76007604598999,
745
+ "eval_runtime": 4.2778,
746
+ "eval_samples_per_second": 233.997,
747
+ "eval_steps_per_second": 14.727,
748
+ "step": 37000
749
+ },
750
+ {
751
+ "epoch": 0.14,
752
+ "learning_rate": 4.773855371681459e-05,
753
+ "loss": 5.7546,
754
+ "step": 37500
755
+ },
756
+ {
757
+ "epoch": 0.14,
758
+ "learning_rate": 4.770839064317326e-05,
759
+ "loss": 5.7542,
760
+ "step": 38000
761
+ },
762
+ {
763
+ "epoch": 0.14,
764
+ "eval_loss": 5.81342077255249,
765
+ "eval_runtime": 4.3112,
766
+ "eval_samples_per_second": 232.186,
767
+ "eval_steps_per_second": 14.613,
768
+ "step": 38000
769
+ },
770
+ {
771
+ "epoch": 0.14,
772
+ "learning_rate": 4.767822756953192e-05,
773
+ "loss": 5.7467,
774
+ "step": 38500
775
+ },
776
+ {
777
+ "epoch": 0.14,
778
+ "learning_rate": 4.764806449589059e-05,
779
+ "loss": 5.7487,
780
+ "step": 39000
781
+ },
782
+ {
783
+ "epoch": 0.14,
784
+ "eval_loss": 5.810147762298584,
785
+ "eval_runtime": 4.2762,
786
+ "eval_samples_per_second": 234.084,
787
+ "eval_steps_per_second": 14.733,
788
+ "step": 39000
789
+ },
790
+ {
791
+ "epoch": 0.14,
792
+ "learning_rate": 4.7617961748396534e-05,
793
+ "loss": 5.7516,
794
+ "step": 39500
795
+ },
796
+ {
797
+ "epoch": 0.14,
798
+ "learning_rate": 4.7587798674755194e-05,
799
+ "loss": 5.7464,
800
+ "step": 40000
801
+ },
802
+ {
803
+ "epoch": 0.14,
804
+ "eval_loss": 5.790091514587402,
805
+ "eval_runtime": 6.3564,
806
+ "eval_samples_per_second": 157.479,
807
+ "eval_steps_per_second": 9.911,
808
+ "step": 40000
809
+ },
810
+ {
811
+ "epoch": 0.15,
812
+ "learning_rate": 4.755763560111387e-05,
813
+ "loss": 5.7518,
814
+ "step": 40500
815
+ },
816
+ {
817
+ "epoch": 0.15,
818
+ "learning_rate": 4.752747252747253e-05,
819
+ "loss": 5.7573,
820
+ "step": 41000
821
+ },
822
+ {
823
+ "epoch": 0.15,
824
+ "eval_loss": 5.804274082183838,
825
+ "eval_runtime": 4.355,
826
+ "eval_samples_per_second": 229.852,
827
+ "eval_steps_per_second": 14.466,
828
+ "step": 41000
829
+ },
830
+ {
831
+ "epoch": 0.15,
832
+ "learning_rate": 4.7497309453831195e-05,
833
+ "loss": 5.7512,
834
+ "step": 41500
835
+ },
836
+ {
837
+ "epoch": 0.15,
838
+ "learning_rate": 4.746714638018986e-05,
839
+ "loss": 5.7431,
840
+ "step": 42000
841
+ },
842
+ {
843
+ "epoch": 0.15,
844
+ "eval_loss": 5.761765480041504,
845
+ "eval_runtime": 4.3203,
846
+ "eval_samples_per_second": 231.698,
847
+ "eval_steps_per_second": 14.582,
848
+ "step": 42000
849
+ },
850
+ {
851
+ "epoch": 0.15,
852
+ "learning_rate": 4.743704363269581e-05,
853
+ "loss": 5.7573,
854
+ "step": 42500
855
+ },
856
+ {
857
+ "epoch": 0.16,
858
+ "learning_rate": 4.740688055905447e-05,
859
+ "loss": 5.7202,
860
+ "step": 43000
861
+ },
862
+ {
863
+ "epoch": 0.16,
864
+ "eval_loss": 5.7786478996276855,
865
+ "eval_runtime": 4.3352,
866
+ "eval_samples_per_second": 230.9,
867
+ "eval_steps_per_second": 14.532,
868
+ "step": 43000
869
+ },
870
+ {
871
+ "epoch": 0.16,
872
+ "learning_rate": 4.737671748541314e-05,
873
+ "loss": 5.7353,
874
+ "step": 43500
875
+ },
876
+ {
877
+ "epoch": 0.16,
878
+ "learning_rate": 4.7346554411771805e-05,
879
+ "loss": 5.7476,
880
+ "step": 44000
881
+ },
882
+ {
883
+ "epoch": 0.16,
884
+ "eval_loss": 5.780131816864014,
885
+ "eval_runtime": 4.4034,
886
+ "eval_samples_per_second": 227.326,
887
+ "eval_steps_per_second": 14.307,
888
+ "step": 44000
889
+ },
890
+ {
891
+ "epoch": 0.16,
892
+ "learning_rate": 4.731639133813047e-05,
893
+ "loss": 5.7345,
894
+ "step": 44500
895
+ },
896
+ {
897
+ "epoch": 0.16,
898
+ "learning_rate": 4.7286288590636415e-05,
899
+ "loss": 5.7416,
900
+ "step": 45000
901
+ },
902
+ {
903
+ "epoch": 0.16,
904
+ "eval_loss": 5.7368245124816895,
905
+ "eval_runtime": 4.2102,
906
+ "eval_samples_per_second": 237.757,
907
+ "eval_steps_per_second": 14.964,
908
+ "step": 45000
909
+ },
910
+ {
911
+ "epoch": 0.16,
912
+ "learning_rate": 4.725612551699509e-05,
913
+ "loss": 5.7451,
914
+ "step": 45500
915
+ },
916
+ {
917
+ "epoch": 0.17,
918
+ "learning_rate": 4.722596244335375e-05,
919
+ "loss": 5.7401,
920
+ "step": 46000
921
+ },
922
+ {
923
+ "epoch": 0.17,
924
+ "eval_loss": 5.775325298309326,
925
+ "eval_runtime": 4.2649,
926
+ "eval_samples_per_second": 234.709,
927
+ "eval_steps_per_second": 14.772,
928
+ "step": 46000
929
+ },
930
+ {
931
+ "epoch": 0.17,
932
+ "learning_rate": 4.7195799369712416e-05,
933
+ "loss": 5.7308,
934
+ "step": 46500
935
+ },
936
+ {
937
+ "epoch": 0.17,
938
+ "learning_rate": 4.716563629607108e-05,
939
+ "loss": 5.7401,
940
+ "step": 47000
941
+ },
942
+ {
943
+ "epoch": 0.17,
944
+ "eval_loss": 5.7712297439575195,
945
+ "eval_runtime": 4.3345,
946
+ "eval_samples_per_second": 230.935,
947
+ "eval_steps_per_second": 14.534,
948
+ "step": 47000
949
+ },
950
+ {
951
+ "epoch": 0.17,
952
+ "learning_rate": 4.7135533548577026e-05,
953
+ "loss": 5.7317,
954
+ "step": 47500
955
+ },
956
+ {
957
+ "epoch": 0.17,
958
+ "learning_rate": 4.710537047493569e-05,
959
+ "loss": 5.7336,
960
+ "step": 48000
961
+ },
962
+ {
963
+ "epoch": 0.17,
964
+ "eval_loss": 5.742641925811768,
965
+ "eval_runtime": 4.2969,
966
+ "eval_samples_per_second": 232.958,
967
+ "eval_steps_per_second": 14.662,
968
+ "step": 48000
969
+ },
970
+ {
971
+ "epoch": 0.18,
972
+ "learning_rate": 4.707520740129436e-05,
973
+ "loss": 5.7449,
974
+ "step": 48500
975
+ },
976
+ {
977
+ "epoch": 0.18,
978
+ "learning_rate": 4.704510465380031e-05,
979
+ "loss": 5.7297,
980
+ "step": 49000
981
+ },
982
+ {
983
+ "epoch": 0.18,
984
+ "eval_loss": 5.769224643707275,
985
+ "eval_runtime": 4.3211,
986
+ "eval_samples_per_second": 231.651,
987
+ "eval_steps_per_second": 14.579,
988
+ "step": 49000
989
+ },
990
+ {
991
+ "epoch": 0.18,
992
+ "learning_rate": 4.701494158015897e-05,
993
+ "loss": 5.7413,
994
+ "step": 49500
995
+ },
996
+ {
997
+ "epoch": 0.18,
998
+ "learning_rate": 4.698477850651764e-05,
999
+ "loss": 5.7347,
1000
+ "step": 50000
1001
+ },
1002
+ {
1003
+ "epoch": 0.18,
1004
+ "eval_loss": 5.765679359436035,
1005
+ "eval_runtime": 4.2691,
1006
+ "eval_samples_per_second": 234.477,
1007
+ "eval_steps_per_second": 14.757,
1008
+ "step": 50000
1009
+ },
1010
+ {
1011
+ "epoch": 0.18,
1012
+ "learning_rate": 4.69546154328763e-05,
1013
+ "loss": 5.7256,
1014
+ "step": 50500
1015
+ },
1016
+ {
1017
+ "epoch": 0.18,
1018
+ "learning_rate": 4.692445235923497e-05,
1019
+ "loss": 5.7183,
1020
+ "step": 51000
1021
+ },
1022
+ {
1023
+ "epoch": 0.18,
1024
+ "eval_loss": 5.7880048751831055,
1025
+ "eval_runtime": 8.1705,
1026
+ "eval_samples_per_second": 122.514,
1027
+ "eval_steps_per_second": 7.711,
1028
+ "step": 51000
1029
+ },
1030
+ {
1031
+ "epoch": 0.19,
1032
+ "learning_rate": 4.689428928559364e-05,
1033
+ "loss": 5.7377,
1034
+ "step": 51500
1035
+ },
1036
+ {
1037
+ "epoch": 0.19,
1038
+ "learning_rate": 4.6864126211952304e-05,
1039
+ "loss": 5.7239,
1040
+ "step": 52000
1041
+ },
1042
+ {
1043
+ "epoch": 0.19,
1044
+ "eval_loss": 5.784894943237305,
1045
+ "eval_runtime": 4.3979,
1046
+ "eval_samples_per_second": 227.609,
1047
+ "eval_steps_per_second": 14.325,
1048
+ "step": 52000
1049
+ },
1050
+ {
1051
+ "epoch": 0.19,
1052
+ "learning_rate": 4.6833963138310964e-05,
1053
+ "loss": 5.7332,
1054
+ "step": 52500
1055
+ },
1056
+ {
1057
+ "epoch": 0.19,
1058
+ "learning_rate": 4.680380006466963e-05,
1059
+ "loss": 5.7316,
1060
+ "step": 53000
1061
+ },
1062
+ {
1063
+ "epoch": 0.19,
1064
+ "eval_loss": 5.74093770980835,
1065
+ "eval_runtime": 4.3725,
1066
+ "eval_samples_per_second": 228.931,
1067
+ "eval_steps_per_second": 14.408,
1068
+ "step": 53000
1069
+ },
1070
+ {
1071
+ "epoch": 0.19,
1072
+ "learning_rate": 4.67736369910283e-05,
1073
+ "loss": 5.7227,
1074
+ "step": 53500
1075
+ },
1076
+ {
1077
+ "epoch": 0.2,
1078
+ "learning_rate": 4.6743473917386965e-05,
1079
+ "loss": 5.7202,
1080
+ "step": 54000
1081
+ },
1082
+ {
1083
+ "epoch": 0.2,
1084
+ "eval_loss": 5.766174793243408,
1085
+ "eval_runtime": 4.2835,
1086
+ "eval_samples_per_second": 233.689,
1087
+ "eval_steps_per_second": 14.708,
1088
+ "step": 54000
1089
+ },
1090
+ {
1091
+ "epoch": 0.2,
1092
+ "learning_rate": 4.671337116989291e-05,
1093
+ "loss": 5.7194,
1094
+ "step": 54500
1095
+ },
1096
+ {
1097
+ "epoch": 0.2,
1098
+ "learning_rate": 4.668326842239886e-05,
1099
+ "loss": 5.7073,
1100
+ "step": 55000
1101
+ },
1102
+ {
1103
+ "epoch": 0.2,
1104
+ "eval_loss": 5.740816593170166,
1105
+ "eval_runtime": 4.2663,
1106
+ "eval_samples_per_second": 234.632,
1107
+ "eval_steps_per_second": 14.767,
1108
+ "step": 55000
1109
+ },
1110
+ {
1111
+ "epoch": 0.2,
1112
+ "learning_rate": 4.6653105348757525e-05,
1113
+ "loss": 5.7176,
1114
+ "step": 55500
1115
+ },
1116
+ {
1117
+ "epoch": 0.2,
1118
+ "learning_rate": 4.662294227511619e-05,
1119
+ "loss": 5.7113,
1120
+ "step": 56000
1121
+ },
1122
+ {
1123
+ "epoch": 0.2,
1124
+ "eval_loss": 5.764217376708984,
1125
+ "eval_runtime": 4.4652,
1126
+ "eval_samples_per_second": 224.178,
1127
+ "eval_steps_per_second": 14.109,
1128
+ "step": 56000
1129
+ },
1130
+ {
1131
+ "epoch": 0.2,
1132
+ "learning_rate": 4.659277920147486e-05,
1133
+ "loss": 5.7094,
1134
+ "step": 56500
1135
+ },
1136
+ {
1137
+ "epoch": 0.21,
1138
+ "learning_rate": 4.656261612783352e-05,
1139
+ "loss": 5.722,
1140
+ "step": 57000
1141
+ },
1142
+ {
1143
+ "epoch": 0.21,
1144
+ "eval_loss": 5.7554473876953125,
1145
+ "eval_runtime": 4.3539,
1146
+ "eval_samples_per_second": 229.907,
1147
+ "eval_steps_per_second": 14.47,
1148
+ "step": 57000
1149
+ },
1150
+ {
1151
+ "epoch": 0.21,
1152
+ "learning_rate": 4.6532453054192186e-05,
1153
+ "loss": 5.7145,
1154
+ "step": 57500
1155
+ },
1156
+ {
1157
+ "epoch": 0.21,
1158
+ "learning_rate": 4.650228998055085e-05,
1159
+ "loss": 5.7077,
1160
+ "step": 58000
1161
+ },
1162
+ {
1163
+ "epoch": 0.21,
1164
+ "eval_loss": 5.74335241317749,
1165
+ "eval_runtime": 4.3356,
1166
+ "eval_samples_per_second": 230.882,
1167
+ "eval_steps_per_second": 14.531,
1168
+ "step": 58000
1169
+ },
1170
+ {
1171
+ "epoch": 0.21,
1172
+ "learning_rate": 4.64721872330568e-05,
1173
+ "loss": 5.7123,
1174
+ "step": 58500
1175
+ },
1176
+ {
1177
+ "epoch": 0.21,
1178
+ "learning_rate": 4.644202415941546e-05,
1179
+ "loss": 5.7163,
1180
+ "step": 59000
1181
+ },
1182
+ {
1183
+ "epoch": 0.21,
1184
+ "eval_loss": 5.77748441696167,
1185
+ "eval_runtime": 4.4579,
1186
+ "eval_samples_per_second": 224.543,
1187
+ "eval_steps_per_second": 14.132,
1188
+ "step": 59000
1189
+ },
1190
+ {
1191
+ "epoch": 0.22,
1192
+ "learning_rate": 4.6411861085774136e-05,
1193
+ "loss": 5.6978,
1194
+ "step": 59500
1195
+ },
1196
+ {
1197
+ "epoch": 0.22,
1198
+ "learning_rate": 4.6381698012132796e-05,
1199
+ "loss": 5.7185,
1200
+ "step": 60000
1201
+ },
1202
+ {
1203
+ "epoch": 0.22,
1204
+ "eval_loss": 5.755308151245117,
1205
+ "eval_runtime": 4.4657,
1206
+ "eval_samples_per_second": 224.155,
1207
+ "eval_steps_per_second": 14.108,
1208
+ "step": 60000
1209
+ },
1210
+ {
1211
+ "epoch": 0.22,
1212
+ "learning_rate": 4.635153493849146e-05,
1213
+ "loss": 5.7133,
1214
+ "step": 60500
1215
+ },
1216
+ {
1217
+ "epoch": 0.22,
1218
+ "learning_rate": 4.632137186485013e-05,
1219
+ "loss": 5.7126,
1220
+ "step": 61000
1221
+ },
1222
+ {
1223
+ "epoch": 0.22,
1224
+ "eval_loss": 5.742791175842285,
1225
+ "eval_runtime": 4.4874,
1226
+ "eval_samples_per_second": 223.069,
1227
+ "eval_steps_per_second": 14.039,
1228
+ "step": 61000
1229
+ },
1230
+ {
1231
+ "epoch": 0.22,
1232
+ "learning_rate": 4.62912087912088e-05,
1233
+ "loss": 5.7105,
1234
+ "step": 61500
1235
+ },
1236
+ {
1237
+ "epoch": 0.22,
1238
+ "learning_rate": 4.626104571756746e-05,
1239
+ "loss": 5.6955,
1240
+ "step": 62000
1241
+ },
1242
+ {
1243
+ "epoch": 0.22,
1244
+ "eval_loss": 5.750290870666504,
1245
+ "eval_runtime": 4.4288,
1246
+ "eval_samples_per_second": 226.023,
1247
+ "eval_steps_per_second": 14.225,
1248
+ "step": 62000
1249
+ },
1250
+ {
1251
+ "epoch": 0.23,
1252
+ "learning_rate": 4.6230882643926124e-05,
1253
+ "loss": 5.7056,
1254
+ "step": 62500
1255
+ },
1256
+ {
1257
+ "epoch": 0.23,
1258
+ "learning_rate": 4.6200779896432074e-05,
1259
+ "loss": 5.7177,
1260
+ "step": 63000
1261
+ },
1262
+ {
1263
+ "epoch": 0.23,
1264
+ "eval_loss": 5.7632975578308105,
1265
+ "eval_runtime": 4.4407,
1266
+ "eval_samples_per_second": 225.416,
1267
+ "eval_steps_per_second": 14.187,
1268
+ "step": 63000
1269
+ },
1270
+ {
1271
+ "epoch": 0.23,
1272
+ "learning_rate": 4.6170616822790734e-05,
1273
+ "loss": 5.6985,
1274
+ "step": 63500
1275
+ },
1276
+ {
1277
+ "epoch": 0.23,
1278
+ "learning_rate": 4.61404537491494e-05,
1279
+ "loss": 5.7103,
1280
+ "step": 64000
1281
+ },
1282
+ {
1283
+ "epoch": 0.23,
1284
+ "eval_loss": 5.735149383544922,
1285
+ "eval_runtime": 4.3259,
1286
+ "eval_samples_per_second": 231.399,
1287
+ "eval_steps_per_second": 14.564,
1288
+ "step": 64000
1289
+ },
1290
+ {
1291
+ "epoch": 0.23,
1292
+ "learning_rate": 4.611029067550807e-05,
1293
+ "loss": 5.7056,
1294
+ "step": 64500
1295
+ },
1296
+ {
1297
+ "epoch": 0.24,
1298
+ "learning_rate": 4.6080127601866735e-05,
1299
+ "loss": 5.7036,
1300
+ "step": 65000
1301
+ },
1302
+ {
1303
+ "epoch": 0.24,
1304
+ "eval_loss": 5.733882427215576,
1305
+ "eval_runtime": 4.2423,
1306
+ "eval_samples_per_second": 235.959,
1307
+ "eval_steps_per_second": 14.851,
1308
+ "step": 65000
1309
+ },
1310
+ {
1311
+ "epoch": 0.24,
1312
+ "learning_rate": 4.605002485437268e-05,
1313
+ "loss": 5.7037,
1314
+ "step": 65500
1315
+ },
1316
+ {
1317
+ "epoch": 0.24,
1318
+ "learning_rate": 4.601986178073135e-05,
1319
+ "loss": 5.7035,
1320
+ "step": 66000
1321
+ },
1322
+ {
1323
+ "epoch": 0.24,
1324
+ "eval_loss": 5.746898651123047,
1325
+ "eval_runtime": 4.2969,
1326
+ "eval_samples_per_second": 232.956,
1327
+ "eval_steps_per_second": 14.662,
1328
+ "step": 66000
1329
+ },
1330
+ {
1331
+ "epoch": 0.24,
1332
+ "learning_rate": 4.598969870709001e-05,
1333
+ "loss": 5.6979,
1334
+ "step": 66500
1335
+ },
1336
+ {
1337
+ "epoch": 0.24,
1338
+ "learning_rate": 4.595953563344868e-05,
1339
+ "loss": 5.6953,
1340
+ "step": 67000
1341
+ },
1342
+ {
1343
+ "epoch": 0.24,
1344
+ "eval_loss": 5.743001461029053,
1345
+ "eval_runtime": 6.5732,
1346
+ "eval_samples_per_second": 152.285,
1347
+ "eval_steps_per_second": 9.584,
1348
+ "step": 67000
1349
+ },
1350
+ {
1351
+ "epoch": 0.24,
1352
+ "learning_rate": 4.592943288595463e-05,
1353
+ "loss": 5.7201,
1354
+ "step": 67500
1355
+ },
1356
+ {
1357
+ "epoch": 0.25,
1358
+ "learning_rate": 4.5899269812313295e-05,
1359
+ "loss": 5.704,
1360
+ "step": 68000
1361
+ },
1362
+ {
1363
+ "epoch": 0.25,
1364
+ "eval_loss": 5.756634712219238,
1365
+ "eval_runtime": 4.4285,
1366
+ "eval_samples_per_second": 226.037,
1367
+ "eval_steps_per_second": 14.226,
1368
+ "step": 68000
1369
+ },
1370
+ {
1371
+ "epoch": 0.25,
1372
+ "learning_rate": 4.5869106738671955e-05,
1373
+ "loss": 5.7095,
1374
+ "step": 68500
1375
+ },
1376
+ {
1377
+ "epoch": 0.25,
1378
+ "learning_rate": 4.583894366503063e-05,
1379
+ "loss": 5.712,
1380
+ "step": 69000
1381
+ },
1382
+ {
1383
+ "epoch": 0.25,
1384
+ "eval_loss": 5.764980316162109,
1385
+ "eval_runtime": 4.3235,
1386
+ "eval_samples_per_second": 231.523,
1387
+ "eval_steps_per_second": 14.571,
1388
+ "step": 69000
1389
+ },
1390
+ {
1391
+ "epoch": 0.25,
1392
+ "learning_rate": 4.580884091753657e-05,
1393
+ "loss": 5.6976,
1394
+ "step": 69500
1395
+ },
1396
+ {
1397
+ "epoch": 0.25,
1398
+ "learning_rate": 4.577867784389523e-05,
1399
+ "loss": 5.7046,
1400
+ "step": 70000
1401
+ },
1402
+ {
1403
+ "epoch": 0.25,
1404
+ "eval_loss": 5.744942665100098,
1405
+ "eval_runtime": 4.3426,
1406
+ "eval_samples_per_second": 230.505,
1407
+ "eval_steps_per_second": 14.507,
1408
+ "step": 70000
1409
+ },
1410
+ {
1411
+ "epoch": 0.26,
1412
+ "learning_rate": 4.5748514770253906e-05,
1413
+ "loss": 5.7025,
1414
+ "step": 70500
1415
+ },
1416
+ {
1417
+ "epoch": 0.26,
1418
+ "learning_rate": 4.5718351696612566e-05,
1419
+ "loss": 5.7031,
1420
+ "step": 71000
1421
+ },
1422
+ {
1423
+ "epoch": 0.26,
1424
+ "eval_loss": 5.732196807861328,
1425
+ "eval_runtime": 4.4018,
1426
+ "eval_samples_per_second": 227.407,
1427
+ "eval_steps_per_second": 14.312,
1428
+ "step": 71000
1429
+ },
1430
+ {
1431
+ "epoch": 0.26,
1432
+ "learning_rate": 4.568818862297123e-05,
1433
+ "loss": 5.6871,
1434
+ "step": 71500
1435
+ },
1436
+ {
1437
+ "epoch": 0.26,
1438
+ "learning_rate": 4.56580255493299e-05,
1439
+ "loss": 5.6842,
1440
+ "step": 72000
1441
+ },
1442
+ {
1443
+ "epoch": 0.26,
1444
+ "eval_loss": 5.773156642913818,
1445
+ "eval_runtime": 4.3306,
1446
+ "eval_samples_per_second": 231.146,
1447
+ "eval_steps_per_second": 14.548,
1448
+ "step": 72000
1449
+ },
1450
+ {
1451
+ "epoch": 0.26,
1452
+ "learning_rate": 4.562786247568857e-05,
1453
+ "loss": 5.6853,
1454
+ "step": 72500
1455
+ },
1456
+ {
1457
+ "epoch": 0.26,
1458
+ "learning_rate": 4.559769940204723e-05,
1459
+ "loss": 5.7022,
1460
+ "step": 73000
1461
+ },
1462
+ {
1463
+ "epoch": 0.26,
1464
+ "eval_loss": 5.724014759063721,
1465
+ "eval_runtime": 4.268,
1466
+ "eval_samples_per_second": 234.536,
1467
+ "eval_steps_per_second": 14.761,
1468
+ "step": 73000
1469
+ },
1470
+ {
1471
+ "epoch": 0.27,
1472
+ "learning_rate": 4.5567536328405894e-05,
1473
+ "loss": 5.7029,
1474
+ "step": 73500
1475
+ },
1476
+ {
1477
+ "epoch": 0.27,
1478
+ "learning_rate": 4.553737325476456e-05,
1479
+ "loss": 5.707,
1480
+ "step": 74000
1481
+ },
1482
+ {
1483
+ "epoch": 0.27,
1484
+ "eval_loss": 5.71505880355835,
1485
+ "eval_runtime": 4.2511,
1486
+ "eval_samples_per_second": 235.471,
1487
+ "eval_steps_per_second": 14.82,
1488
+ "step": 74000
1489
+ },
1490
+ {
1491
+ "epoch": 0.27,
1492
+ "learning_rate": 4.550727050727051e-05,
1493
+ "loss": 5.7014,
1494
+ "step": 74500
1495
+ },
1496
+ {
1497
+ "epoch": 0.27,
1498
+ "learning_rate": 4.547710743362917e-05,
1499
+ "loss": 5.7068,
1500
+ "step": 75000
1501
+ },
1502
+ {
1503
+ "epoch": 0.27,
1504
+ "eval_loss": 5.703161239624023,
1505
+ "eval_runtime": 4.3019,
1506
+ "eval_samples_per_second": 232.687,
1507
+ "eval_steps_per_second": 14.645,
1508
+ "step": 75000
1509
+ },
1510
+ {
1511
+ "epoch": 0.27,
1512
+ "learning_rate": 4.5446944359987844e-05,
1513
+ "loss": 5.6946,
1514
+ "step": 75500
1515
+ },
1516
+ {
1517
+ "epoch": 0.28,
1518
+ "learning_rate": 4.5416781286346505e-05,
1519
+ "loss": 5.6892,
1520
+ "step": 76000
1521
+ },
1522
+ {
1523
+ "epoch": 0.28,
1524
+ "eval_loss": 5.7333149909973145,
1525
+ "eval_runtime": 4.2526,
1526
+ "eval_samples_per_second": 235.387,
1527
+ "eval_steps_per_second": 14.815,
1528
+ "step": 76000
1529
+ },
1530
+ {
1531
+ "epoch": 0.28,
1532
+ "learning_rate": 4.5386678538852454e-05,
1533
+ "loss": 5.6814,
1534
+ "step": 76500
1535
+ },
1536
+ {
1537
+ "epoch": 0.28,
1538
+ "learning_rate": 4.535651546521112e-05,
1539
+ "loss": 5.6979,
1540
+ "step": 77000
1541
+ },
1542
+ {
1543
+ "epoch": 0.28,
1544
+ "eval_loss": 5.712403297424316,
1545
+ "eval_runtime": 4.3234,
1546
+ "eval_samples_per_second": 231.532,
1547
+ "eval_steps_per_second": 14.572,
1548
+ "step": 77000
1549
+ },
1550
+ {
1551
+ "epoch": 0.28,
1552
+ "learning_rate": 4.532635239156979e-05,
1553
+ "loss": 5.6959,
1554
+ "step": 77500
1555
+ },
1556
+ {
1557
+ "epoch": 0.28,
1558
+ "learning_rate": 4.529618931792845e-05,
1559
+ "loss": 5.6791,
1560
+ "step": 78000
1561
+ },
1562
+ {
1563
+ "epoch": 0.28,
1564
+ "eval_loss": 5.754906177520752,
1565
+ "eval_runtime": 4.3238,
1566
+ "eval_samples_per_second": 231.508,
1567
+ "eval_steps_per_second": 14.57,
1568
+ "step": 78000
1569
+ },
1570
+ {
1571
+ "epoch": 0.28,
1572
+ "learning_rate": 4.5266026244287115e-05,
1573
+ "loss": 5.6984,
1574
+ "step": 78500
1575
+ },
1576
+ {
1577
+ "epoch": 0.29,
1578
+ "learning_rate": 4.5235923496793065e-05,
1579
+ "loss": 5.6953,
1580
+ "step": 79000
1581
+ },
1582
+ {
1583
+ "epoch": 0.29,
1584
+ "eval_loss": 5.720048427581787,
1585
+ "eval_runtime": 4.3352,
1586
+ "eval_samples_per_second": 230.899,
1587
+ "eval_steps_per_second": 14.532,
1588
+ "step": 79000
1589
+ },
1590
+ {
1591
+ "epoch": 0.29,
1592
+ "learning_rate": 4.5205760423151725e-05,
1593
+ "loss": 5.6924,
1594
+ "step": 79500
1595
+ },
1596
+ {
1597
+ "epoch": 0.29,
1598
+ "learning_rate": 4.51755973495104e-05,
1599
+ "loss": 5.6845,
1600
+ "step": 80000
1601
+ },
1602
+ {
1603
+ "epoch": 0.29,
1604
+ "eval_loss": 5.744686126708984,
1605
+ "eval_runtime": 4.3839,
1606
+ "eval_samples_per_second": 228.335,
1607
+ "eval_steps_per_second": 14.371,
1608
+ "step": 80000
1609
+ },
1610
+ {
1611
+ "epoch": 0.29,
1612
+ "learning_rate": 4.514543427586906e-05,
1613
+ "loss": 5.695,
1614
+ "step": 80500
1615
+ },
1616
+ {
1617
+ "epoch": 0.29,
1618
+ "learning_rate": 4.511533152837501e-05,
1619
+ "loss": 5.6872,
1620
+ "step": 81000
1621
+ },
1622
+ {
1623
+ "epoch": 0.29,
1624
+ "eval_loss": 5.723191738128662,
1625
+ "eval_runtime": 4.2942,
1626
+ "eval_samples_per_second": 233.107,
1627
+ "eval_steps_per_second": 14.671,
1628
+ "step": 81000
1629
+ },
1630
+ {
1631
+ "epoch": 0.29,
1632
+ "learning_rate": 4.5085168454733675e-05,
1633
+ "loss": 5.6841,
1634
+ "step": 81500
1635
+ },
1636
+ {
1637
+ "epoch": 0.3,
1638
+ "learning_rate": 4.505500538109234e-05,
1639
+ "loss": 5.687,
1640
+ "step": 82000
1641
+ },
1642
+ {
1643
+ "epoch": 0.3,
1644
+ "eval_loss": 5.738052845001221,
1645
+ "eval_runtime": 4.282,
1646
+ "eval_samples_per_second": 233.769,
1647
+ "eval_steps_per_second": 14.713,
1648
+ "step": 82000
1649
+ },
1650
+ {
1651
+ "epoch": 0.3,
1652
+ "learning_rate": 4.5024842307451e-05,
1653
+ "loss": 5.6913,
1654
+ "step": 82500
1655
+ },
1656
+ {
1657
+ "epoch": 0.3,
1658
+ "learning_rate": 4.499467923380967e-05,
1659
+ "loss": 5.69,
1660
+ "step": 83000
1661
+ },
1662
+ {
1663
+ "epoch": 0.3,
1664
+ "eval_loss": 5.709909915924072,
1665
+ "eval_runtime": 4.3445,
1666
+ "eval_samples_per_second": 230.406,
1667
+ "eval_steps_per_second": 14.501,
1668
+ "step": 83000
1669
+ },
1670
+ {
1671
+ "epoch": 0.3,
1672
+ "learning_rate": 4.496457648631562e-05,
1673
+ "loss": 5.6785,
1674
+ "step": 83500
1675
+ },
1676
+ {
1677
+ "epoch": 0.3,
1678
+ "learning_rate": 4.4934413412674286e-05,
1679
+ "loss": 5.6831,
1680
+ "step": 84000
1681
+ },
1682
+ {
1683
+ "epoch": 0.3,
1684
+ "eval_loss": 5.698389053344727,
1685
+ "eval_runtime": 6.3376,
1686
+ "eval_samples_per_second": 157.947,
1687
+ "eval_steps_per_second": 9.941,
1688
+ "step": 84000
1689
+ },
1690
+ {
1691
+ "epoch": 0.31,
1692
+ "learning_rate": 4.4904250339032946e-05,
1693
+ "loss": 5.6893,
1694
+ "step": 84500
1695
+ },
1696
+ {
1697
+ "epoch": 0.31,
1698
+ "learning_rate": 4.4874147591538896e-05,
1699
+ "loss": 5.682,
1700
+ "step": 85000
1701
+ },
1702
+ {
1703
+ "epoch": 0.31,
1704
+ "eval_loss": 5.747517108917236,
1705
+ "eval_runtime": 4.4117,
1706
+ "eval_samples_per_second": 226.897,
1707
+ "eval_steps_per_second": 14.28,
1708
+ "step": 85000
1709
+ },
1710
+ {
1711
+ "epoch": 0.31,
1712
+ "learning_rate": 4.484398451789756e-05,
1713
+ "loss": 5.6907,
1714
+ "step": 85500
1715
+ },
1716
+ {
1717
+ "epoch": 0.31,
1718
+ "learning_rate": 4.481382144425623e-05,
1719
+ "loss": 5.6771,
1720
+ "step": 86000
1721
+ },
1722
+ {
1723
+ "epoch": 0.31,
1724
+ "eval_loss": 5.721287727355957,
1725
+ "eval_runtime": 4.3352,
1726
+ "eval_samples_per_second": 230.902,
1727
+ "eval_steps_per_second": 14.532,
1728
+ "step": 86000
1729
+ },
1730
+ {
1731
+ "epoch": 0.31,
1732
+ "learning_rate": 4.47836583706149e-05,
1733
+ "loss": 5.6743,
1734
+ "step": 86500
1735
+ },
1736
+ {
1737
+ "epoch": 0.31,
1738
+ "learning_rate": 4.475349529697356e-05,
1739
+ "loss": 5.6818,
1740
+ "step": 87000
1741
+ },
1742
+ {
1743
+ "epoch": 0.31,
1744
+ "eval_loss": 5.736617088317871,
1745
+ "eval_runtime": 4.3397,
1746
+ "eval_samples_per_second": 230.66,
1747
+ "eval_steps_per_second": 14.517,
1748
+ "step": 87000
1749
+ },
1750
+ {
1751
+ "epoch": 0.32,
1752
+ "learning_rate": 4.4723332223332224e-05,
1753
+ "loss": 5.6836,
1754
+ "step": 87500
1755
+ },
1756
+ {
1757
+ "epoch": 0.32,
1758
+ "learning_rate": 4.469316914969089e-05,
1759
+ "loss": 5.6862,
1760
+ "step": 88000
1761
+ },
1762
+ {
1763
+ "epoch": 0.32,
1764
+ "eval_loss": 5.706295490264893,
1765
+ "eval_runtime": 4.3399,
1766
+ "eval_samples_per_second": 230.65,
1767
+ "eval_steps_per_second": 14.516,
1768
+ "step": 88000
1769
+ },
1770
+ {
1771
+ "epoch": 0.32,
1772
+ "learning_rate": 4.466300607604956e-05,
1773
+ "loss": 5.6792,
1774
+ "step": 88500
1775
+ },
1776
+ {
1777
+ "epoch": 0.32,
1778
+ "learning_rate": 4.46329033285555e-05,
1779
+ "loss": 5.6706,
1780
+ "step": 89000
1781
+ },
1782
+ {
1783
+ "epoch": 0.32,
1784
+ "eval_loss": 5.7102370262146,
1785
+ "eval_runtime": 4.2947,
1786
+ "eval_samples_per_second": 233.076,
1787
+ "eval_steps_per_second": 14.669,
1788
+ "step": 89000
1789
+ },
1790
+ {
1791
+ "epoch": 0.32,
1792
+ "learning_rate": 4.4602740254914174e-05,
1793
+ "loss": 5.6775,
1794
+ "step": 89500
1795
+ },
1796
+ {
1797
+ "epoch": 0.33,
1798
+ "learning_rate": 4.4572577181272834e-05,
1799
+ "loss": 5.6662,
1800
+ "step": 90000
1801
+ },
1802
+ {
1803
+ "epoch": 0.33,
1804
+ "eval_loss": 5.69236421585083,
1805
+ "eval_runtime": 6.3959,
1806
+ "eval_samples_per_second": 156.506,
1807
+ "eval_steps_per_second": 9.85,
1808
+ "step": 90000
1809
+ },
1810
+ {
1811
+ "epoch": 0.33,
1812
+ "learning_rate": 4.45424141076315e-05,
1813
+ "loss": 5.6691,
1814
+ "step": 90500
1815
+ },
1816
+ {
1817
+ "epoch": 0.33,
1818
+ "learning_rate": 4.4512371686284734e-05,
1819
+ "loss": 5.6839,
1820
+ "step": 91000
1821
+ },
1822
+ {
1823
+ "epoch": 0.33,
1824
+ "eval_loss": 5.711584091186523,
1825
+ "eval_runtime": 4.2918,
1826
+ "eval_samples_per_second": 233.234,
1827
+ "eval_steps_per_second": 14.679,
1828
+ "step": 91000
1829
+ },
1830
+ {
1831
+ "epoch": 0.33,
1832
+ "learning_rate": 4.44822086126434e-05,
1833
+ "loss": 5.6725,
1834
+ "step": 91500
1835
+ },
1836
+ {
1837
+ "epoch": 0.33,
1838
+ "learning_rate": 4.445204553900206e-05,
1839
+ "loss": 5.6789,
1840
+ "step": 92000
1841
+ },
1842
+ {
1843
+ "epoch": 0.33,
1844
+ "eval_loss": 5.663390159606934,
1845
+ "eval_runtime": 4.3322,
1846
+ "eval_samples_per_second": 231.063,
1847
+ "eval_steps_per_second": 14.542,
1848
+ "step": 92000
1849
+ },
1850
+ {
1851
+ "epoch": 0.33,
1852
+ "learning_rate": 4.442188246536073e-05,
1853
+ "loss": 5.6765,
1854
+ "step": 92500
1855
+ },
1856
+ {
1857
+ "epoch": 0.34,
1858
+ "learning_rate": 4.4391719391719395e-05,
1859
+ "loss": 5.6618,
1860
+ "step": 93000
1861
+ },
1862
+ {
1863
+ "epoch": 0.34,
1864
+ "eval_loss": 5.73824405670166,
1865
+ "eval_runtime": 4.2539,
1866
+ "eval_samples_per_second": 235.311,
1867
+ "eval_steps_per_second": 14.81,
1868
+ "step": 93000
1869
+ },
1870
+ {
1871
+ "epoch": 0.34,
1872
+ "learning_rate": 4.4361556318078055e-05,
1873
+ "loss": 5.6636,
1874
+ "step": 93500
1875
+ },
1876
+ {
1877
+ "epoch": 0.34,
1878
+ "learning_rate": 4.433139324443673e-05,
1879
+ "loss": 5.6733,
1880
+ "step": 94000
1881
+ },
1882
+ {
1883
+ "epoch": 0.34,
1884
+ "eval_loss": 5.712129592895508,
1885
+ "eval_runtime": 4.3264,
1886
+ "eval_samples_per_second": 231.369,
1887
+ "eval_steps_per_second": 14.562,
1888
+ "step": 94000
1889
+ },
1890
+ {
1891
+ "epoch": 0.34,
1892
+ "learning_rate": 4.430123017079539e-05,
1893
+ "loss": 5.6787,
1894
+ "step": 94500
1895
+ },
1896
+ {
1897
+ "epoch": 0.34,
1898
+ "learning_rate": 4.427112742330134e-05,
1899
+ "loss": 5.6709,
1900
+ "step": 95000
1901
+ },
1902
+ {
1903
+ "epoch": 0.34,
1904
+ "eval_loss": 5.736870288848877,
1905
+ "eval_runtime": 4.381,
1906
+ "eval_samples_per_second": 228.487,
1907
+ "eval_steps_per_second": 14.38,
1908
+ "step": 95000
1909
+ },
1910
+ {
1911
+ "epoch": 0.35,
1912
+ "learning_rate": 4.4240964349660005e-05,
1913
+ "loss": 5.6618,
1914
+ "step": 95500
1915
+ },
1916
+ {
1917
+ "epoch": 0.35,
1918
+ "learning_rate": 4.421080127601867e-05,
1919
+ "loss": 5.6627,
1920
+ "step": 96000
1921
+ },
1922
+ {
1923
+ "epoch": 0.35,
1924
+ "eval_loss": 5.6889142990112305,
1925
+ "eval_runtime": 4.2821,
1926
+ "eval_samples_per_second": 233.762,
1927
+ "eval_steps_per_second": 14.712,
1928
+ "step": 96000
1929
+ },
1930
+ {
1931
+ "epoch": 0.35,
1932
+ "learning_rate": 4.418063820237733e-05,
1933
+ "loss": 5.6751,
1934
+ "step": 96500
1935
+ },
1936
+ {
1937
+ "epoch": 0.35,
1938
+ "learning_rate": 4.415053545488328e-05,
1939
+ "loss": 5.655,
1940
+ "step": 97000
1941
+ },
1942
+ {
1943
+ "epoch": 0.35,
1944
+ "eval_loss": 5.685154438018799,
1945
+ "eval_runtime": 6.9191,
1946
+ "eval_samples_per_second": 144.671,
1947
+ "eval_steps_per_second": 9.105,
1948
+ "step": 97000
1949
+ },
1950
+ {
1951
+ "epoch": 0.35,
1952
+ "learning_rate": 4.412037238124195e-05,
1953
+ "loss": 5.6734,
1954
+ "step": 97500
1955
+ },
1956
+ {
1957
+ "epoch": 0.35,
1958
+ "learning_rate": 4.4090209307600616e-05,
1959
+ "loss": 5.6649,
1960
+ "step": 98000
1961
+ },
1962
+ {
1963
+ "epoch": 0.35,
1964
+ "eval_loss": 5.713215351104736,
1965
+ "eval_runtime": 4.2937,
1966
+ "eval_samples_per_second": 233.132,
1967
+ "eval_steps_per_second": 14.673,
1968
+ "step": 98000
1969
+ },
1970
+ {
1971
+ "epoch": 0.36,
1972
+ "learning_rate": 4.4060046233959276e-05,
1973
+ "loss": 5.6724,
1974
+ "step": 98500
1975
+ },
1976
+ {
1977
+ "epoch": 0.36,
1978
+ "learning_rate": 4.4029943486465226e-05,
1979
+ "loss": 5.6634,
1980
+ "step": 99000
1981
+ },
1982
+ {
1983
+ "epoch": 0.36,
1984
+ "eval_loss": 5.674860000610352,
1985
+ "eval_runtime": 4.3043,
1986
+ "eval_samples_per_second": 232.557,
1987
+ "eval_steps_per_second": 14.636,
1988
+ "step": 99000
1989
+ },
1990
+ {
1991
+ "epoch": 0.36,
1992
+ "learning_rate": 4.399978041282389e-05,
1993
+ "loss": 5.6674,
1994
+ "step": 99500
1995
+ },
1996
+ {
1997
+ "epoch": 0.36,
1998
+ "learning_rate": 4.396961733918255e-05,
1999
+ "loss": 5.6625,
2000
+ "step": 100000
2001
+ },
2002
+ {
2003
+ "epoch": 0.36,
2004
+ "eval_loss": 5.694820880889893,
2005
+ "eval_runtime": 4.396,
2006
+ "eval_samples_per_second": 227.709,
2007
+ "eval_steps_per_second": 14.331,
2008
+ "step": 100000
2009
+ },
2010
+ {
2011
+ "epoch": 0.36,
2012
+ "learning_rate": 4.393951459168851e-05,
2013
+ "loss": 5.6711,
2014
+ "step": 100500
2015
+ },
2016
+ {
2017
+ "epoch": 0.37,
2018
+ "learning_rate": 4.390935151804717e-05,
2019
+ "loss": 5.6721,
2020
+ "step": 101000
2021
+ },
2022
+ {
2023
+ "epoch": 0.37,
2024
+ "eval_loss": 5.711126804351807,
2025
+ "eval_runtime": 6.4814,
2026
+ "eval_samples_per_second": 154.443,
2027
+ "eval_steps_per_second": 9.72,
2028
+ "step": 101000
2029
+ },
2030
+ {
2031
+ "epoch": 0.37,
2032
+ "learning_rate": 4.3879188444405836e-05,
2033
+ "loss": 5.6574,
2034
+ "step": 101500
2035
+ },
2036
+ {
2037
+ "epoch": 0.37,
2038
+ "learning_rate": 4.38490253707645e-05,
2039
+ "loss": 5.6608,
2040
+ "step": 102000
2041
+ },
2042
+ {
2043
+ "epoch": 0.37,
2044
+ "eval_loss": 5.73061990737915,
2045
+ "eval_runtime": 4.2517,
2046
+ "eval_samples_per_second": 235.437,
2047
+ "eval_steps_per_second": 14.818,
2048
+ "step": 102000
2049
+ },
2050
+ {
2051
+ "epoch": 0.37,
2052
+ "learning_rate": 4.381886229712317e-05,
2053
+ "loss": 5.6663,
2054
+ "step": 102500
2055
+ },
2056
+ {
2057
+ "epoch": 0.37,
2058
+ "learning_rate": 4.378869922348183e-05,
2059
+ "loss": 5.6685,
2060
+ "step": 103000
2061
+ },
2062
+ {
2063
+ "epoch": 0.37,
2064
+ "eval_loss": 5.769214630126953,
2065
+ "eval_runtime": 4.3149,
2066
+ "eval_samples_per_second": 231.988,
2067
+ "eval_steps_per_second": 14.601,
2068
+ "step": 103000
2069
+ },
2070
+ {
2071
+ "epoch": 0.37,
2072
+ "learning_rate": 4.3758536149840504e-05,
2073
+ "loss": 5.6631,
2074
+ "step": 103500
2075
+ },
2076
+ {
2077
+ "epoch": 0.38,
2078
+ "learning_rate": 4.3728373076199164e-05,
2079
+ "loss": 5.6557,
2080
+ "step": 104000
2081
+ },
2082
+ {
2083
+ "epoch": 0.38,
2084
+ "eval_loss": 5.712480545043945,
2085
+ "eval_runtime": 4.3273,
2086
+ "eval_samples_per_second": 231.325,
2087
+ "eval_steps_per_second": 14.559,
2088
+ "step": 104000
2089
+ },
2090
+ {
2091
+ "epoch": 0.38,
2092
+ "learning_rate": 4.3698270328705114e-05,
2093
+ "loss": 5.6767,
2094
+ "step": 104500
2095
+ },
2096
+ {
2097
+ "epoch": 0.38,
2098
+ "learning_rate": 4.3668167581211063e-05,
2099
+ "loss": 5.6665,
2100
+ "step": 105000
2101
+ },
2102
+ {
2103
+ "epoch": 0.38,
2104
+ "eval_loss": 5.739169120788574,
2105
+ "eval_runtime": 4.2992,
2106
+ "eval_samples_per_second": 232.836,
2107
+ "eval_steps_per_second": 14.654,
2108
+ "step": 105000
2109
+ },
2110
+ {
2111
+ "epoch": 0.38,
2112
+ "learning_rate": 4.363800450756973e-05,
2113
+ "loss": 5.6699,
2114
+ "step": 105500
2115
+ },
2116
+ {
2117
+ "epoch": 0.38,
2118
+ "learning_rate": 4.360784143392839e-05,
2119
+ "loss": 5.6603,
2120
+ "step": 106000
2121
+ },
2122
+ {
2123
+ "epoch": 0.38,
2124
+ "eval_loss": 5.6816325187683105,
2125
+ "eval_runtime": 4.353,
2126
+ "eval_samples_per_second": 229.954,
2127
+ "eval_steps_per_second": 14.473,
2128
+ "step": 106000
2129
+ },
2130
+ {
2131
+ "epoch": 0.39,
2132
+ "learning_rate": 4.357767836028706e-05,
2133
+ "loss": 5.6717,
2134
+ "step": 106500
2135
+ },
2136
+ {
2137
+ "epoch": 0.39,
2138
+ "learning_rate": 4.3547515286645725e-05,
2139
+ "loss": 5.6712,
2140
+ "step": 107000
2141
+ },
2142
+ {
2143
+ "epoch": 0.39,
2144
+ "eval_loss": 5.7074737548828125,
2145
+ "eval_runtime": 4.3743,
2146
+ "eval_samples_per_second": 228.834,
2147
+ "eval_steps_per_second": 14.402,
2148
+ "step": 107000
2149
+ },
2150
+ {
2151
+ "epoch": 0.39,
2152
+ "learning_rate": 4.3517352213004385e-05,
2153
+ "loss": 5.6676,
2154
+ "step": 107500
2155
+ },
2156
+ {
2157
+ "epoch": 0.39,
2158
+ "learning_rate": 4.348718913936305e-05,
2159
+ "loss": 5.6547,
2160
+ "step": 108000
2161
+ },
2162
+ {
2163
+ "epoch": 0.39,
2164
+ "eval_loss": 5.6991777420043945,
2165
+ "eval_runtime": 4.4989,
2166
+ "eval_samples_per_second": 222.5,
2167
+ "eval_steps_per_second": 14.004,
2168
+ "step": 108000
2169
+ },
2170
+ {
2171
+ "epoch": 0.39,
2172
+ "learning_rate": 4.3457086391869e-05,
2173
+ "loss": 5.6638,
2174
+ "step": 108500
2175
+ },
2176
+ {
2177
+ "epoch": 0.39,
2178
+ "learning_rate": 4.342692331822767e-05,
2179
+ "loss": 5.6511,
2180
+ "step": 109000
2181
+ },
2182
+ {
2183
+ "epoch": 0.39,
2184
+ "eval_loss": 5.679354190826416,
2185
+ "eval_runtime": 4.2833,
2186
+ "eval_samples_per_second": 233.696,
2187
+ "eval_steps_per_second": 14.708,
2188
+ "step": 109000
2189
+ },
2190
+ {
2191
+ "epoch": 0.4,
2192
+ "learning_rate": 4.3396760244586335e-05,
2193
+ "loss": 5.6673,
2194
+ "step": 109500
2195
+ },
2196
+ {
2197
+ "epoch": 0.4,
2198
+ "learning_rate": 4.3366597170945e-05,
2199
+ "loss": 5.6663,
2200
+ "step": 110000
2201
+ },
2202
+ {
2203
+ "epoch": 0.4,
2204
+ "eval_loss": 5.680385589599609,
2205
+ "eval_runtime": 4.3563,
2206
+ "eval_samples_per_second": 229.783,
2207
+ "eval_steps_per_second": 14.462,
2208
+ "step": 110000
2209
+ },
2210
+ {
2211
+ "epoch": 0.4,
2212
+ "learning_rate": 4.333643409730366e-05,
2213
+ "loss": 5.6682,
2214
+ "step": 110500
2215
+ },
2216
+ {
2217
+ "epoch": 0.4,
2218
+ "learning_rate": 4.330627102366233e-05,
2219
+ "loss": 5.6562,
2220
+ "step": 111000
2221
+ },
2222
+ {
2223
+ "epoch": 0.4,
2224
+ "eval_loss": 5.730945587158203,
2225
+ "eval_runtime": 6.374,
2226
+ "eval_samples_per_second": 157.044,
2227
+ "eval_steps_per_second": 9.884,
2228
+ "step": 111000
2229
+ },
2230
+ {
2231
+ "epoch": 0.4,
2232
+ "learning_rate": 4.3276107950020996e-05,
2233
+ "loss": 5.6656,
2234
+ "step": 111500
2235
+ },
2236
+ {
2237
+ "epoch": 0.41,
2238
+ "learning_rate": 4.3246005202526946e-05,
2239
+ "loss": 5.6546,
2240
+ "step": 112000
2241
+ },
2242
+ {
2243
+ "epoch": 0.41,
2244
+ "eval_loss": 5.708312034606934,
2245
+ "eval_runtime": 4.2353,
2246
+ "eval_samples_per_second": 236.345,
2247
+ "eval_steps_per_second": 14.875,
2248
+ "step": 112000
2249
+ },
2250
+ {
2251
+ "epoch": 0.41,
2252
+ "learning_rate": 4.3215842128885606e-05,
2253
+ "loss": 5.6529,
2254
+ "step": 112500
2255
+ },
2256
+ {
2257
+ "epoch": 0.41,
2258
+ "learning_rate": 4.318567905524428e-05,
2259
+ "loss": 5.656,
2260
+ "step": 113000
2261
+ },
2262
+ {
2263
+ "epoch": 0.41,
2264
+ "eval_loss": 5.711429119110107,
2265
+ "eval_runtime": 4.3481,
2266
+ "eval_samples_per_second": 230.216,
2267
+ "eval_steps_per_second": 14.489,
2268
+ "step": 113000
2269
+ },
2270
+ {
2271
+ "epoch": 0.41,
2272
+ "learning_rate": 4.315551598160294e-05,
2273
+ "loss": 5.6731,
2274
+ "step": 113500
2275
+ },
2276
+ {
2277
+ "epoch": 0.41,
2278
+ "learning_rate": 4.312535290796161e-05,
2279
+ "loss": 5.6712,
2280
+ "step": 114000
2281
+ },
2282
+ {
2283
+ "epoch": 0.41,
2284
+ "eval_loss": 5.6710896492004395,
2285
+ "eval_runtime": 4.2998,
2286
+ "eval_samples_per_second": 232.804,
2287
+ "eval_steps_per_second": 14.652,
2288
+ "step": 114000
2289
+ },
2290
+ {
2291
+ "epoch": 0.41,
2292
+ "learning_rate": 4.3095189834320274e-05,
2293
+ "loss": 5.6619,
2294
+ "step": 114500
2295
+ },
2296
+ {
2297
+ "epoch": 0.42,
2298
+ "learning_rate": 4.306502676067894e-05,
2299
+ "loss": 5.6473,
2300
+ "step": 115000
2301
+ },
2302
+ {
2303
+ "epoch": 0.42,
2304
+ "eval_loss": 5.6910600662231445,
2305
+ "eval_runtime": 4.3442,
2306
+ "eval_samples_per_second": 230.421,
2307
+ "eval_steps_per_second": 14.502,
2308
+ "step": 115000
2309
+ },
2310
+ {
2311
+ "epoch": 0.42,
2312
+ "learning_rate": 4.3034924013184884e-05,
2313
+ "loss": 5.6577,
2314
+ "step": 115500
2315
+ },
2316
+ {
2317
+ "epoch": 0.42,
2318
+ "learning_rate": 4.300476093954355e-05,
2319
+ "loss": 5.6352,
2320
+ "step": 116000
2321
+ },
2322
+ {
2323
+ "epoch": 0.42,
2324
+ "eval_loss": 5.7348408699035645,
2325
+ "eval_runtime": 4.3322,
2326
+ "eval_samples_per_second": 231.062,
2327
+ "eval_steps_per_second": 14.542,
2328
+ "step": 116000
2329
+ },
2330
+ {
2331
+ "epoch": 0.42,
2332
+ "learning_rate": 4.297459786590222e-05,
2333
+ "loss": 5.647,
2334
+ "step": 116500
2335
+ },
2336
+ {
2337
+ "epoch": 0.42,
2338
+ "learning_rate": 4.294443479226088e-05,
2339
+ "loss": 5.6602,
2340
+ "step": 117000
2341
+ },
2342
+ {
2343
+ "epoch": 0.42,
2344
+ "eval_loss": 5.7037835121154785,
2345
+ "eval_runtime": 6.3915,
2346
+ "eval_samples_per_second": 156.614,
2347
+ "eval_steps_per_second": 9.857,
2348
+ "step": 117000
2349
+ },
2350
+ {
2351
+ "epoch": 0.43,
2352
+ "learning_rate": 4.2914271718619545e-05,
2353
+ "loss": 5.6598,
2354
+ "step": 117500
2355
+ },
2356
+ {
2357
+ "epoch": 0.43,
2358
+ "learning_rate": 4.288410864497821e-05,
2359
+ "loss": 5.645,
2360
+ "step": 118000
2361
+ },
2362
+ {
2363
+ "epoch": 0.43,
2364
+ "eval_loss": 5.670388698577881,
2365
+ "eval_runtime": 4.2976,
2366
+ "eval_samples_per_second": 232.922,
2367
+ "eval_steps_per_second": 14.659,
2368
+ "step": 118000
2369
+ },
2370
+ {
2371
+ "epoch": 0.43,
2372
+ "learning_rate": 4.285394557133688e-05,
2373
+ "loss": 5.637,
2374
+ "step": 118500
2375
+ },
2376
+ {
2377
+ "epoch": 0.43,
2378
+ "learning_rate": 4.282378249769554e-05,
2379
+ "loss": 5.6611,
2380
+ "step": 119000
2381
+ },
2382
+ {
2383
+ "epoch": 0.43,
2384
+ "eval_loss": 5.681214332580566,
2385
+ "eval_runtime": 4.3365,
2386
+ "eval_samples_per_second": 230.833,
2387
+ "eval_steps_per_second": 14.528,
2388
+ "step": 119000
2389
+ },
2390
+ {
2391
+ "epoch": 0.43,
2392
+ "learning_rate": 4.2793679750201495e-05,
2393
+ "loss": 5.6413,
2394
+ "step": 119500
2395
+ },
2396
+ {
2397
+ "epoch": 0.43,
2398
+ "learning_rate": 4.276357700270744e-05,
2399
+ "loss": 5.6442,
2400
+ "step": 120000
2401
+ },
2402
+ {
2403
+ "epoch": 0.43,
2404
+ "eval_loss": 5.678561687469482,
2405
+ "eval_runtime": 4.2838,
2406
+ "eval_samples_per_second": 233.671,
2407
+ "eval_steps_per_second": 14.707,
2408
+ "step": 120000
2409
+ },
2410
+ {
2411
+ "epoch": 0.44,
2412
+ "learning_rate": 4.2733413929066105e-05,
2413
+ "loss": 5.6429,
2414
+ "step": 120500
2415
+ },
2416
+ {
2417
+ "epoch": 0.44,
2418
+ "learning_rate": 4.270325085542477e-05,
2419
+ "loss": 5.6572,
2420
+ "step": 121000
2421
+ },
2422
+ {
2423
+ "epoch": 0.44,
2424
+ "eval_loss": 5.70313024520874,
2425
+ "eval_runtime": 4.2374,
2426
+ "eval_samples_per_second": 236.231,
2427
+ "eval_steps_per_second": 14.868,
2428
+ "step": 121000
2429
+ },
2430
+ {
2431
+ "epoch": 0.44,
2432
+ "learning_rate": 4.267308778178344e-05,
2433
+ "loss": 5.6563,
2434
+ "step": 121500
2435
+ },
2436
+ {
2437
+ "epoch": 0.44,
2438
+ "learning_rate": 4.26429247081421e-05,
2439
+ "loss": 5.6631,
2440
+ "step": 122000
2441
+ },
2442
+ {
2443
+ "epoch": 0.44,
2444
+ "eval_loss": 5.692766189575195,
2445
+ "eval_runtime": 4.2623,
2446
+ "eval_samples_per_second": 234.85,
2447
+ "eval_steps_per_second": 14.781,
2448
+ "step": 122000
2449
+ },
2450
+ {
2451
+ "epoch": 0.44,
2452
+ "learning_rate": 4.261276163450077e-05,
2453
+ "loss": 5.6386,
2454
+ "step": 122500
2455
+ },
2456
+ {
2457
+ "epoch": 0.45,
2458
+ "learning_rate": 4.258259856085943e-05,
2459
+ "loss": 5.6553,
2460
+ "step": 123000
2461
+ },
2462
+ {
2463
+ "epoch": 0.45,
2464
+ "eval_loss": 5.708261013031006,
2465
+ "eval_runtime": 4.2737,
2466
+ "eval_samples_per_second": 234.224,
2467
+ "eval_steps_per_second": 14.741,
2468
+ "step": 123000
2469
+ },
2470
+ {
2471
+ "epoch": 0.45,
2472
+ "learning_rate": 4.2552495813365376e-05,
2473
+ "loss": 5.6524,
2474
+ "step": 123500
2475
+ },
2476
+ {
2477
+ "epoch": 0.45,
2478
+ "learning_rate": 4.252233273972405e-05,
2479
+ "loss": 5.6521,
2480
+ "step": 124000
2481
+ },
2482
+ {
2483
+ "epoch": 0.45,
2484
+ "eval_loss": 5.687132358551025,
2485
+ "eval_runtime": 4.4286,
2486
+ "eval_samples_per_second": 226.029,
2487
+ "eval_steps_per_second": 14.226,
2488
+ "step": 124000
2489
+ },
2490
+ {
2491
+ "epoch": 0.45,
2492
+ "learning_rate": 4.249216966608271e-05,
2493
+ "loss": 5.6559,
2494
+ "step": 124500
2495
+ },
2496
+ {
2497
+ "epoch": 0.45,
2498
+ "learning_rate": 4.2462006592441377e-05,
2499
+ "loss": 5.6644,
2500
+ "step": 125000
2501
+ },
2502
+ {
2503
+ "epoch": 0.45,
2504
+ "eval_loss": 5.712847709655762,
2505
+ "eval_runtime": 4.3174,
2506
+ "eval_samples_per_second": 231.852,
2507
+ "eval_steps_per_second": 14.592,
2508
+ "step": 125000
2509
+ },
2510
+ {
2511
+ "epoch": 0.45,
2512
+ "learning_rate": 4.2431843518800044e-05,
2513
+ "loss": 5.6407,
2514
+ "step": 125500
2515
+ },
2516
+ {
2517
+ "epoch": 0.46,
2518
+ "learning_rate": 4.240174077130599e-05,
2519
+ "loss": 5.6421,
2520
+ "step": 126000
2521
+ },
2522
+ {
2523
+ "epoch": 0.46,
2524
+ "eval_loss": 5.716492652893066,
2525
+ "eval_runtime": 4.2976,
2526
+ "eval_samples_per_second": 232.92,
2527
+ "eval_steps_per_second": 14.659,
2528
+ "step": 126000
2529
+ },
2530
+ {
2531
+ "epoch": 0.46,
2532
+ "learning_rate": 4.237157769766465e-05,
2533
+ "loss": 5.6653,
2534
+ "step": 126500
2535
+ },
2536
+ {
2537
+ "epoch": 0.46,
2538
+ "learning_rate": 4.234141462402332e-05,
2539
+ "loss": 5.6465,
2540
+ "step": 127000
2541
+ },
2542
+ {
2543
+ "epoch": 0.46,
2544
+ "eval_loss": 5.711600303649902,
2545
+ "eval_runtime": 4.3593,
2546
+ "eval_samples_per_second": 229.622,
2547
+ "eval_steps_per_second": 14.452,
2548
+ "step": 127000
2549
+ },
2550
+ {
2551
+ "epoch": 0.46,
2552
+ "learning_rate": 4.231125155038199e-05,
2553
+ "loss": 5.6545,
2554
+ "step": 127500
2555
+ },
2556
+ {
2557
+ "epoch": 0.46,
2558
+ "learning_rate": 4.2281088476740654e-05,
2559
+ "loss": 5.653,
2560
+ "step": 128000
2561
+ },
2562
+ {
2563
+ "epoch": 0.46,
2564
+ "eval_loss": 5.694251537322998,
2565
+ "eval_runtime": 6.4453,
2566
+ "eval_samples_per_second": 155.307,
2567
+ "eval_steps_per_second": 9.775,
2568
+ "step": 128000
2569
+ },
2570
+ {
2571
+ "epoch": 0.47,
2572
+ "learning_rate": 4.2250925403099314e-05,
2573
+ "loss": 5.6472,
2574
+ "step": 128500
2575
+ },
2576
+ {
2577
+ "epoch": 0.47,
2578
+ "learning_rate": 4.222076232945799e-05,
2579
+ "loss": 5.6546,
2580
+ "step": 129000
2581
+ },
2582
+ {
2583
+ "epoch": 0.47,
2584
+ "eval_loss": 5.6814446449279785,
2585
+ "eval_runtime": 4.3371,
2586
+ "eval_samples_per_second": 230.802,
2587
+ "eval_steps_per_second": 14.526,
2588
+ "step": 129000
2589
+ },
2590
+ {
2591
+ "epoch": 0.47,
2592
+ "learning_rate": 4.219059925581665e-05,
2593
+ "loss": 5.6483,
2594
+ "step": 129500
2595
+ },
2596
+ {
2597
+ "epoch": 0.47,
2598
+ "learning_rate": 4.216055683446988e-05,
2599
+ "loss": 5.654,
2600
+ "step": 130000
2601
+ },
2602
+ {
2603
+ "epoch": 0.47,
2604
+ "eval_loss": 5.7155609130859375,
2605
+ "eval_runtime": 4.3569,
2606
+ "eval_samples_per_second": 229.748,
2607
+ "eval_steps_per_second": 14.46,
2608
+ "step": 130000
2609
+ },
2610
+ {
2611
+ "epoch": 0.47,
2612
+ "learning_rate": 4.213039376082855e-05,
2613
+ "loss": 5.6427,
2614
+ "step": 130500
2615
+ },
2616
+ {
2617
+ "epoch": 0.47,
2618
+ "learning_rate": 4.210023068718721e-05,
2619
+ "loss": 5.6526,
2620
+ "step": 131000
2621
+ },
2622
+ {
2623
+ "epoch": 0.47,
2624
+ "eval_loss": 5.657435417175293,
2625
+ "eval_runtime": 4.288,
2626
+ "eval_samples_per_second": 233.441,
2627
+ "eval_steps_per_second": 14.692,
2628
+ "step": 131000
2629
+ },
2630
+ {
2631
+ "epoch": 0.48,
2632
+ "learning_rate": 4.2070067613545875e-05,
2633
+ "loss": 5.6495,
2634
+ "step": 131500
2635
+ },
2636
+ {
2637
+ "epoch": 0.48,
2638
+ "learning_rate": 4.203990453990454e-05,
2639
+ "loss": 5.649,
2640
+ "step": 132000
2641
+ },
2642
+ {
2643
+ "epoch": 0.48,
2644
+ "eval_loss": 5.6852569580078125,
2645
+ "eval_runtime": 4.3096,
2646
+ "eval_samples_per_second": 232.275,
2647
+ "eval_steps_per_second": 14.619,
2648
+ "step": 132000
2649
+ },
2650
+ {
2651
+ "epoch": 0.48,
2652
+ "learning_rate": 4.200980179241049e-05,
2653
+ "loss": 5.6591,
2654
+ "step": 132500
2655
+ },
2656
+ {
2657
+ "epoch": 0.48,
2658
+ "learning_rate": 4.197963871876915e-05,
2659
+ "loss": 5.6427,
2660
+ "step": 133000
2661
+ },
2662
+ {
2663
+ "epoch": 0.48,
2664
+ "eval_loss": 5.6808366775512695,
2665
+ "eval_runtime": 4.2828,
2666
+ "eval_samples_per_second": 233.726,
2667
+ "eval_steps_per_second": 14.71,
2668
+ "step": 133000
2669
+ },
2670
+ {
2671
+ "epoch": 0.48,
2672
+ "learning_rate": 4.1949475645127825e-05,
2673
+ "loss": 5.6409,
2674
+ "step": 133500
2675
+ },
2676
+ {
2677
+ "epoch": 0.49,
2678
+ "learning_rate": 4.1919312571486485e-05,
2679
+ "loss": 5.6436,
2680
+ "step": 134000
2681
+ },
2682
+ {
2683
+ "epoch": 0.49,
2684
+ "eval_loss": 5.664061546325684,
2685
+ "eval_runtime": 4.2839,
2686
+ "eval_samples_per_second": 233.663,
2687
+ "eval_steps_per_second": 14.706,
2688
+ "step": 134000
2689
+ },
2690
+ {
2691
+ "epoch": 0.49,
2692
+ "learning_rate": 4.188914949784515e-05,
2693
+ "loss": 5.6421,
2694
+ "step": 134500
2695
+ },
2696
+ {
2697
+ "epoch": 0.49,
2698
+ "learning_rate": 4.185898642420382e-05,
2699
+ "loss": 5.6319,
2700
+ "step": 135000
2701
+ },
2702
+ {
2703
+ "epoch": 0.49,
2704
+ "eval_loss": 5.68691349029541,
2705
+ "eval_runtime": 4.3423,
2706
+ "eval_samples_per_second": 230.522,
2707
+ "eval_steps_per_second": 14.508,
2708
+ "step": 135000
2709
+ },
2710
+ {
2711
+ "epoch": 0.49,
2712
+ "learning_rate": 4.1828823350562486e-05,
2713
+ "loss": 5.6542,
2714
+ "step": 135500
2715
+ },
2716
+ {
2717
+ "epoch": 0.49,
2718
+ "learning_rate": 4.1798660276921146e-05,
2719
+ "loss": 5.6393,
2720
+ "step": 136000
2721
+ },
2722
+ {
2723
+ "epoch": 0.49,
2724
+ "eval_loss": 5.684676647186279,
2725
+ "eval_runtime": 4.3763,
2726
+ "eval_samples_per_second": 228.73,
2727
+ "eval_steps_per_second": 14.396,
2728
+ "step": 136000
2729
+ },
2730
+ {
2731
+ "epoch": 0.49,
2732
+ "learning_rate": 4.176849720327981e-05,
2733
+ "loss": 5.6442,
2734
+ "step": 136500
2735
+ },
2736
+ {
2737
+ "epoch": 0.5,
2738
+ "learning_rate": 4.173839445578576e-05,
2739
+ "loss": 5.6363,
2740
+ "step": 137000
2741
+ },
2742
+ {
2743
+ "epoch": 0.5,
2744
+ "eval_loss": 5.680596828460693,
2745
+ "eval_runtime": 4.349,
2746
+ "eval_samples_per_second": 230.166,
2747
+ "eval_steps_per_second": 14.486,
2748
+ "step": 137000
2749
+ },
2750
+ {
2751
+ "epoch": 0.5,
2752
+ "learning_rate": 4.170823138214443e-05,
2753
+ "loss": 5.6416,
2754
+ "step": 137500
2755
+ },
2756
+ {
2757
+ "epoch": 0.5,
2758
+ "learning_rate": 4.167812863465038e-05,
2759
+ "loss": 5.648,
2760
+ "step": 138000
2761
+ },
2762
+ {
2763
+ "epoch": 0.5,
2764
+ "eval_loss": 5.648463249206543,
2765
+ "eval_runtime": 4.3014,
2766
+ "eval_samples_per_second": 232.713,
2767
+ "eval_steps_per_second": 14.646,
2768
+ "step": 138000
2769
+ },
2770
+ {
2771
+ "epoch": 0.5,
2772
+ "learning_rate": 4.164796556100904e-05,
2773
+ "loss": 5.6414,
2774
+ "step": 138500
2775
+ },
2776
+ {
2777
+ "epoch": 0.5,
2778
+ "learning_rate": 4.1617802487367706e-05,
2779
+ "loss": 5.6297,
2780
+ "step": 139000
2781
+ },
2782
+ {
2783
+ "epoch": 0.5,
2784
+ "eval_loss": 5.679803371429443,
2785
+ "eval_runtime": 4.2827,
2786
+ "eval_samples_per_second": 233.729,
2787
+ "eval_steps_per_second": 14.71,
2788
+ "step": 139000
2789
+ },
2790
+ {
2791
+ "epoch": 0.5,
2792
+ "learning_rate": 4.1587639413726373e-05,
2793
+ "loss": 5.6411,
2794
+ "step": 139500
2795
+ },
2796
+ {
2797
+ "epoch": 0.51,
2798
+ "learning_rate": 4.155747634008504e-05,
2799
+ "loss": 5.6442,
2800
+ "step": 140000
2801
+ },
2802
+ {
2803
+ "epoch": 0.51,
2804
+ "eval_loss": 5.673394203186035,
2805
+ "eval_runtime": 4.3412,
2806
+ "eval_samples_per_second": 230.582,
2807
+ "eval_steps_per_second": 14.512,
2808
+ "step": 140000
2809
+ },
2810
+ {
2811
+ "epoch": 0.51,
2812
+ "learning_rate": 4.15273132664437e-05,
2813
+ "loss": 5.6313,
2814
+ "step": 140500
2815
+ },
2816
+ {
2817
+ "epoch": 0.51,
2818
+ "learning_rate": 4.149715019280237e-05,
2819
+ "loss": 5.6546,
2820
+ "step": 141000
2821
+ },
2822
+ {
2823
+ "epoch": 0.51,
2824
+ "eval_loss": 5.6783528327941895,
2825
+ "eval_runtime": 4.4144,
2826
+ "eval_samples_per_second": 226.756,
2827
+ "eval_steps_per_second": 14.271,
2828
+ "step": 141000
2829
+ },
2830
+ {
2831
+ "epoch": 0.51,
2832
+ "learning_rate": 4.146704744530832e-05,
2833
+ "loss": 5.6374,
2834
+ "step": 141500
2835
+ },
2836
+ {
2837
+ "epoch": 0.51,
2838
+ "learning_rate": 4.1436884371666984e-05,
2839
+ "loss": 5.6325,
2840
+ "step": 142000
2841
+ },
2842
+ {
2843
+ "epoch": 0.51,
2844
+ "eval_loss": 5.645352840423584,
2845
+ "eval_runtime": 4.3033,
2846
+ "eval_samples_per_second": 232.612,
2847
+ "eval_steps_per_second": 14.64,
2848
+ "step": 142000
2849
+ },
2850
+ {
2851
+ "epoch": 0.52,
2852
+ "learning_rate": 4.1406721298025644e-05,
2853
+ "loss": 5.643,
2854
+ "step": 142500
2855
+ },
2856
+ {
2857
+ "epoch": 0.52,
2858
+ "learning_rate": 4.137655822438432e-05,
2859
+ "loss": 5.6386,
2860
+ "step": 143000
2861
+ },
2862
+ {
2863
+ "epoch": 0.52,
2864
+ "eval_loss": 5.695677757263184,
2865
+ "eval_runtime": 4.2949,
2866
+ "eval_samples_per_second": 233.065,
2867
+ "eval_steps_per_second": 14.668,
2868
+ "step": 143000
2869
+ },
2870
+ {
2871
+ "epoch": 0.52,
2872
+ "learning_rate": 4.134639515074298e-05,
2873
+ "loss": 5.6337,
2874
+ "step": 143500
2875
+ },
2876
+ {
2877
+ "epoch": 0.52,
2878
+ "learning_rate": 4.131629240324893e-05,
2879
+ "loss": 5.6236,
2880
+ "step": 144000
2881
+ },
2882
+ {
2883
+ "epoch": 0.52,
2884
+ "eval_loss": 5.678653240203857,
2885
+ "eval_runtime": 4.4357,
2886
+ "eval_samples_per_second": 225.669,
2887
+ "eval_steps_per_second": 14.203,
2888
+ "step": 144000
2889
+ },
2890
+ {
2891
+ "epoch": 0.52,
2892
+ "learning_rate": 4.1286129329607595e-05,
2893
+ "loss": 5.639,
2894
+ "step": 144500
2895
+ },
2896
+ {
2897
+ "epoch": 0.52,
2898
+ "learning_rate": 4.1255966255966255e-05,
2899
+ "loss": 5.6269,
2900
+ "step": 145000
2901
+ },
2902
+ {
2903
+ "epoch": 0.52,
2904
+ "eval_loss": 5.681842803955078,
2905
+ "eval_runtime": 4.3725,
2906
+ "eval_samples_per_second": 228.929,
2907
+ "eval_steps_per_second": 14.408,
2908
+ "step": 145000
2909
+ },
2910
+ {
2911
+ "epoch": 0.53,
2912
+ "learning_rate": 4.122580318232492e-05,
2913
+ "loss": 5.635,
2914
+ "step": 145500
2915
+ },
2916
+ {
2917
+ "epoch": 0.53,
2918
+ "learning_rate": 4.119564010868359e-05,
2919
+ "loss": 5.6285,
2920
+ "step": 146000
2921
+ },
2922
+ {
2923
+ "epoch": 0.53,
2924
+ "eval_loss": 5.673871040344238,
2925
+ "eval_runtime": 4.3771,
2926
+ "eval_samples_per_second": 228.689,
2927
+ "eval_steps_per_second": 14.393,
2928
+ "step": 146000
2929
+ },
2930
+ {
2931
+ "epoch": 0.53,
2932
+ "learning_rate": 4.1165477035042256e-05,
2933
+ "loss": 5.6383,
2934
+ "step": 146500
2935
+ },
2936
+ {
2937
+ "epoch": 0.53,
2938
+ "learning_rate": 4.11353742875482e-05,
2939
+ "loss": 5.65,
2940
+ "step": 147000
2941
+ },
2942
+ {
2943
+ "epoch": 0.53,
2944
+ "eval_loss": 5.630629539489746,
2945
+ "eval_runtime": 4.41,
2946
+ "eval_samples_per_second": 226.983,
2947
+ "eval_steps_per_second": 14.286,
2948
+ "step": 147000
2949
+ },
2950
+ {
2951
+ "epoch": 0.53,
2952
+ "learning_rate": 4.110521121390687e-05,
2953
+ "loss": 5.6381,
2954
+ "step": 147500
2955
+ },
2956
+ {
2957
+ "epoch": 0.54,
2958
+ "learning_rate": 4.107504814026553e-05,
2959
+ "loss": 5.6313,
2960
+ "step": 148000
2961
+ },
2962
+ {
2963
+ "epoch": 0.54,
2964
+ "eval_loss": 5.6462812423706055,
2965
+ "eval_runtime": 4.3445,
2966
+ "eval_samples_per_second": 230.407,
2967
+ "eval_steps_per_second": 14.501,
2968
+ "step": 148000
2969
+ },
2970
+ {
2971
+ "epoch": 0.54,
2972
+ "learning_rate": 4.10448850666242e-05,
2973
+ "loss": 5.6347,
2974
+ "step": 148500
2975
+ },
2976
+ {
2977
+ "epoch": 0.54,
2978
+ "learning_rate": 4.1014721992982866e-05,
2979
+ "loss": 5.6412,
2980
+ "step": 149000
2981
+ },
2982
+ {
2983
+ "epoch": 0.54,
2984
+ "eval_loss": 5.668566703796387,
2985
+ "eval_runtime": 4.3418,
2986
+ "eval_samples_per_second": 230.551,
2987
+ "eval_steps_per_second": 14.51,
2988
+ "step": 149000
2989
+ },
2990
+ {
2991
+ "epoch": 0.54,
2992
+ "learning_rate": 4.098455891934153e-05,
2993
+ "loss": 5.639,
2994
+ "step": 149500
2995
+ },
2996
+ {
2997
+ "epoch": 0.54,
2998
+ "learning_rate": 4.0954456171847476e-05,
2999
+ "loss": 5.6278,
3000
+ "step": 150000
3001
+ },
3002
+ {
3003
+ "epoch": 0.54,
3004
+ "eval_loss": 5.6881489753723145,
3005
+ "eval_runtime": 4.455,
3006
+ "eval_samples_per_second": 224.692,
3007
+ "eval_steps_per_second": 14.141,
3008
+ "step": 150000
3009
+ },
3010
+ {
3011
+ "epoch": 0.54,
3012
+ "learning_rate": 4.092429309820614e-05,
3013
+ "loss": 5.6408,
3014
+ "step": 150500
3015
+ },
3016
+ {
3017
+ "epoch": 0.55,
3018
+ "learning_rate": 4.089413002456481e-05,
3019
+ "loss": 5.637,
3020
+ "step": 151000
3021
+ },
3022
+ {
3023
+ "epoch": 0.55,
3024
+ "eval_loss": 5.688764572143555,
3025
+ "eval_runtime": 6.2867,
3026
+ "eval_samples_per_second": 159.224,
3027
+ "eval_steps_per_second": 10.021,
3028
+ "step": 151000
3029
+ },
3030
+ {
3031
+ "epoch": 0.55,
3032
+ "learning_rate": 4.086396695092348e-05,
3033
+ "loss": 5.646,
3034
+ "step": 151500
3035
+ },
3036
+ {
3037
+ "epoch": 0.55,
3038
+ "learning_rate": 4.083380387728214e-05,
3039
+ "loss": 5.626,
3040
+ "step": 152000
3041
+ },
3042
+ {
3043
+ "epoch": 0.55,
3044
+ "eval_loss": 5.671611785888672,
3045
+ "eval_runtime": 4.3616,
3046
+ "eval_samples_per_second": 229.503,
3047
+ "eval_steps_per_second": 14.444,
3048
+ "step": 152000
3049
+ },
3050
+ {
3051
+ "epoch": 0.55,
3052
+ "learning_rate": 4.080364080364081e-05,
3053
+ "loss": 5.6378,
3054
+ "step": 152500
3055
+ },
3056
+ {
3057
+ "epoch": 0.55,
3058
+ "learning_rate": 4.0773538056146754e-05,
3059
+ "loss": 5.6338,
3060
+ "step": 153000
3061
+ },
3062
+ {
3063
+ "epoch": 0.55,
3064
+ "eval_loss": 5.616012096405029,
3065
+ "eval_runtime": 4.4085,
3066
+ "eval_samples_per_second": 227.06,
3067
+ "eval_steps_per_second": 14.29,
3068
+ "step": 153000
3069
+ },
3070
+ {
3071
+ "epoch": 0.56,
3072
+ "learning_rate": 4.0743374982505414e-05,
3073
+ "loss": 5.6318,
3074
+ "step": 153500
3075
+ },
3076
+ {
3077
+ "epoch": 0.56,
3078
+ "learning_rate": 4.071321190886409e-05,
3079
+ "loss": 5.6361,
3080
+ "step": 154000
3081
+ },
3082
+ {
3083
+ "epoch": 0.56,
3084
+ "eval_loss": 5.667599201202393,
3085
+ "eval_runtime": 6.2583,
3086
+ "eval_samples_per_second": 159.947,
3087
+ "eval_steps_per_second": 10.067,
3088
+ "step": 154000
3089
+ },
3090
+ {
3091
+ "epoch": 0.56,
3092
+ "learning_rate": 4.068304883522275e-05,
3093
+ "loss": 5.625,
3094
+ "step": 154500
3095
+ },
3096
+ {
3097
+ "epoch": 0.56,
3098
+ "learning_rate": 4.0652885761581415e-05,
3099
+ "loss": 5.6336,
3100
+ "step": 155000
3101
+ },
3102
+ {
3103
+ "epoch": 0.56,
3104
+ "eval_loss": 5.684675216674805,
3105
+ "eval_runtime": 4.3154,
3106
+ "eval_samples_per_second": 231.962,
3107
+ "eval_steps_per_second": 14.599,
3108
+ "step": 155000
3109
+ },
3110
+ {
3111
+ "epoch": 0.56,
3112
+ "learning_rate": 4.062272268794008e-05,
3113
+ "loss": 5.6388,
3114
+ "step": 155500
3115
+ },
3116
+ {
3117
+ "epoch": 0.56,
3118
+ "learning_rate": 4.059255961429875e-05,
3119
+ "loss": 5.6351,
3120
+ "step": 156000
3121
+ },
3122
+ {
3123
+ "epoch": 0.56,
3124
+ "eval_loss": 5.675138473510742,
3125
+ "eval_runtime": 4.3055,
3126
+ "eval_samples_per_second": 232.494,
3127
+ "eval_steps_per_second": 14.632,
3128
+ "step": 156000
3129
+ },
3130
+ {
3131
+ "epoch": 0.57,
3132
+ "learning_rate": 4.056239654065741e-05,
3133
+ "loss": 5.6266,
3134
+ "step": 156500
3135
+ },
3136
+ {
3137
+ "epoch": 0.57,
3138
+ "learning_rate": 4.053229379316336e-05,
3139
+ "loss": 5.6408,
3140
+ "step": 157000
3141
+ },
3142
+ {
3143
+ "epoch": 0.57,
3144
+ "eval_loss": 5.615994453430176,
3145
+ "eval_runtime": 4.3074,
3146
+ "eval_samples_per_second": 232.39,
3147
+ "eval_steps_per_second": 14.626,
3148
+ "step": 157000
3149
+ },
3150
+ {
3151
+ "epoch": 0.57,
3152
+ "learning_rate": 4.0502130719522025e-05,
3153
+ "loss": 5.6399,
3154
+ "step": 157500
3155
+ },
3156
+ {
3157
+ "epoch": 0.57,
3158
+ "learning_rate": 4.047196764588069e-05,
3159
+ "loss": 5.6232,
3160
+ "step": 158000
3161
+ },
3162
+ {
3163
+ "epoch": 0.57,
3164
+ "eval_loss": 5.652904033660889,
3165
+ "eval_runtime": 4.3343,
3166
+ "eval_samples_per_second": 230.949,
3167
+ "eval_steps_per_second": 14.535,
3168
+ "step": 158000
3169
+ },
3170
+ {
3171
+ "epoch": 0.57,
3172
+ "learning_rate": 4.044186489838664e-05,
3173
+ "loss": 5.6249,
3174
+ "step": 158500
3175
+ },
3176
+ {
3177
+ "epoch": 0.58,
3178
+ "learning_rate": 4.041170182474531e-05,
3179
+ "loss": 5.6319,
3180
+ "step": 159000
3181
+ },
3182
+ {
3183
+ "epoch": 0.58,
3184
+ "eval_loss": 5.64531946182251,
3185
+ "eval_runtime": 5.6252,
3186
+ "eval_samples_per_second": 177.948,
3187
+ "eval_steps_per_second": 11.2,
3188
+ "step": 159000
3189
+ },
3190
+ {
3191
+ "epoch": 0.58,
3192
+ "learning_rate": 4.038153875110397e-05,
3193
+ "loss": 5.6244,
3194
+ "step": 159500
3195
+ },
3196
+ {
3197
+ "epoch": 0.58,
3198
+ "learning_rate": 4.0351375677462636e-05,
3199
+ "loss": 5.6204,
3200
+ "step": 160000
3201
+ },
3202
+ {
3203
+ "epoch": 0.58,
3204
+ "eval_loss": 5.661260604858398,
3205
+ "eval_runtime": 4.3573,
3206
+ "eval_samples_per_second": 229.732,
3207
+ "eval_steps_per_second": 14.459,
3208
+ "step": 160000
3209
+ },
3210
+ {
3211
+ "epoch": 0.58,
3212
+ "learning_rate": 4.03212126038213e-05,
3213
+ "loss": 5.6353,
3214
+ "step": 160500
3215
+ },
3216
+ {
3217
+ "epoch": 0.58,
3218
+ "learning_rate": 4.0291109856327246e-05,
3219
+ "loss": 5.6231,
3220
+ "step": 161000
3221
+ },
3222
+ {
3223
+ "epoch": 0.58,
3224
+ "eval_loss": 5.666103363037109,
3225
+ "eval_runtime": 4.5226,
3226
+ "eval_samples_per_second": 221.333,
3227
+ "eval_steps_per_second": 13.93,
3228
+ "step": 161000
3229
+ },
3230
+ {
3231
+ "epoch": 0.58,
3232
+ "learning_rate": 4.026094678268591e-05,
3233
+ "loss": 5.6318,
3234
+ "step": 161500
3235
+ },
3236
+ {
3237
+ "epoch": 0.59,
3238
+ "learning_rate": 4.023078370904458e-05,
3239
+ "loss": 5.6322,
3240
+ "step": 162000
3241
+ },
3242
+ {
3243
+ "epoch": 0.59,
3244
+ "eval_loss": 5.640702247619629,
3245
+ "eval_runtime": 4.3357,
3246
+ "eval_samples_per_second": 230.872,
3247
+ "eval_steps_per_second": 14.53,
3248
+ "step": 162000
3249
+ },
3250
+ {
3251
+ "epoch": 0.59,
3252
+ "learning_rate": 4.020062063540325e-05,
3253
+ "loss": 5.6135,
3254
+ "step": 162500
3255
+ },
3256
+ {
3257
+ "epoch": 0.59,
3258
+ "learning_rate": 4.017045756176191e-05,
3259
+ "loss": 5.6369,
3260
+ "step": 163000
3261
+ },
3262
+ {
3263
+ "epoch": 0.59,
3264
+ "eval_loss": 5.658245086669922,
3265
+ "eval_runtime": 4.353,
3266
+ "eval_samples_per_second": 229.957,
3267
+ "eval_steps_per_second": 14.473,
3268
+ "step": 163000
3269
+ },
3270
+ {
3271
+ "epoch": 0.59,
3272
+ "learning_rate": 4.014029448812058e-05,
3273
+ "loss": 5.6445,
3274
+ "step": 163500
3275
+ },
3276
+ {
3277
+ "epoch": 0.59,
3278
+ "learning_rate": 4.0110191740626523e-05,
3279
+ "loss": 5.6337,
3280
+ "step": 164000
3281
+ },
3282
+ {
3283
+ "epoch": 0.59,
3284
+ "eval_loss": 5.672937393188477,
3285
+ "eval_runtime": 4.4799,
3286
+ "eval_samples_per_second": 223.441,
3287
+ "eval_steps_per_second": 14.063,
3288
+ "step": 164000
3289
+ },
3290
+ {
3291
+ "epoch": 0.6,
3292
+ "learning_rate": 4.008002866698519e-05,
3293
+ "loss": 5.6349,
3294
+ "step": 164500
3295
+ },
3296
+ {
3297
+ "epoch": 0.6,
3298
+ "learning_rate": 4.004986559334386e-05,
3299
+ "loss": 5.6278,
3300
+ "step": 165000
3301
+ },
3302
+ {
3303
+ "epoch": 0.6,
3304
+ "eval_loss": 5.6635026931762695,
3305
+ "eval_runtime": 6.1168,
3306
+ "eval_samples_per_second": 163.647,
3307
+ "eval_steps_per_second": 10.299,
3308
+ "step": 165000
3309
+ },
3310
+ {
3311
+ "epoch": 0.6,
3312
+ "learning_rate": 4.0019702519702524e-05,
3313
+ "loss": 5.6318,
3314
+ "step": 165500
3315
+ },
3316
+ {
3317
+ "epoch": 0.6,
3318
+ "learning_rate": 3.9989539446061184e-05,
3319
+ "loss": 5.6371,
3320
+ "step": 166000
3321
+ },
3322
+ {
3323
+ "epoch": 0.6,
3324
+ "eval_loss": 5.643805503845215,
3325
+ "eval_runtime": 4.2716,
3326
+ "eval_samples_per_second": 234.336,
3327
+ "eval_steps_per_second": 14.748,
3328
+ "step": 166000
3329
+ },
3330
+ {
3331
+ "epoch": 0.6,
3332
+ "learning_rate": 3.995943669856714e-05,
3333
+ "loss": 5.6145,
3334
+ "step": 166500
3335
+ },
3336
+ {
3337
+ "epoch": 0.6,
3338
+ "learning_rate": 3.99292736249258e-05,
3339
+ "loss": 5.624,
3340
+ "step": 167000
3341
+ },
3342
+ {
3343
+ "epoch": 0.6,
3344
+ "eval_loss": 5.633095741271973,
3345
+ "eval_runtime": 4.3138,
3346
+ "eval_samples_per_second": 232.048,
3347
+ "eval_steps_per_second": 14.604,
3348
+ "step": 167000
3349
+ },
3350
+ {
3351
+ "epoch": 0.61,
3352
+ "learning_rate": 3.989911055128447e-05,
3353
+ "loss": 5.619,
3354
+ "step": 167500
3355
+ },
3356
+ {
3357
+ "epoch": 0.61,
3358
+ "learning_rate": 3.9868947477643135e-05,
3359
+ "loss": 5.6198,
3360
+ "step": 168000
3361
+ },
3362
+ {
3363
+ "epoch": 0.61,
3364
+ "eval_loss": 5.6540656089782715,
3365
+ "eval_runtime": 4.3595,
3366
+ "eval_samples_per_second": 229.615,
3367
+ "eval_steps_per_second": 14.451,
3368
+ "step": 168000
3369
+ },
3370
+ {
3371
+ "epoch": 0.61,
3372
+ "learning_rate": 3.98387844040018e-05,
3373
+ "loss": 5.624,
3374
+ "step": 168500
3375
+ },
3376
+ {
3377
+ "epoch": 0.61,
3378
+ "learning_rate": 3.9808681656507745e-05,
3379
+ "loss": 5.6243,
3380
+ "step": 169000
3381
+ },
3382
+ {
3383
+ "epoch": 0.61,
3384
+ "eval_loss": 5.651320457458496,
3385
+ "eval_runtime": 4.3133,
3386
+ "eval_samples_per_second": 232.072,
3387
+ "eval_steps_per_second": 14.606,
3388
+ "step": 169000
3389
+ },
3390
+ {
3391
+ "epoch": 0.61,
3392
+ "learning_rate": 3.977851858286641e-05,
3393
+ "loss": 5.6133,
3394
+ "step": 169500
3395
+ },
3396
+ {
3397
+ "epoch": 0.62,
3398
+ "learning_rate": 3.974835550922508e-05,
3399
+ "loss": 5.6118,
3400
+ "step": 170000
3401
+ },
3402
+ {
3403
+ "epoch": 0.62,
3404
+ "eval_loss": 5.682806491851807,
3405
+ "eval_runtime": 4.4555,
3406
+ "eval_samples_per_second": 224.668,
3407
+ "eval_steps_per_second": 14.14,
3408
+ "step": 170000
3409
+ },
3410
+ {
3411
+ "epoch": 0.62,
3412
+ "learning_rate": 3.971819243558374e-05,
3413
+ "loss": 5.6085,
3414
+ "step": 170500
3415
+ },
3416
+ {
3417
+ "epoch": 0.62,
3418
+ "learning_rate": 3.9688029361942406e-05,
3419
+ "loss": 5.6315,
3420
+ "step": 171000
3421
+ },
3422
+ {
3423
+ "epoch": 0.62,
3424
+ "eval_loss": 5.685708045959473,
3425
+ "eval_runtime": 4.3306,
3426
+ "eval_samples_per_second": 231.143,
3427
+ "eval_steps_per_second": 14.547,
3428
+ "step": 171000
3429
+ },
3430
+ {
3431
+ "epoch": 0.62,
3432
+ "learning_rate": 3.965786628830107e-05,
3433
+ "loss": 5.6271,
3434
+ "step": 171500
3435
+ },
3436
+ {
3437
+ "epoch": 0.62,
3438
+ "learning_rate": 3.962770321465974e-05,
3439
+ "loss": 5.6208,
3440
+ "step": 172000
3441
+ },
3442
+ {
3443
+ "epoch": 0.62,
3444
+ "eval_loss": 5.64992618560791,
3445
+ "eval_runtime": 4.409,
3446
+ "eval_samples_per_second": 227.033,
3447
+ "eval_steps_per_second": 14.289,
3448
+ "step": 172000
3449
+ },
3450
+ {
3451
+ "epoch": 0.62,
3452
+ "learning_rate": 3.95975401410184e-05,
3453
+ "loss": 5.6218,
3454
+ "step": 172500
3455
+ },
3456
+ {
3457
+ "epoch": 0.63,
3458
+ "learning_rate": 3.9567437393524356e-05,
3459
+ "loss": 5.6261,
3460
+ "step": 173000
3461
+ },
3462
+ {
3463
+ "epoch": 0.63,
3464
+ "eval_loss": 5.64516544342041,
3465
+ "eval_runtime": 4.4391,
3466
+ "eval_samples_per_second": 225.497,
3467
+ "eval_steps_per_second": 14.192,
3468
+ "step": 173000
3469
+ },
3470
+ {
3471
+ "epoch": 0.63,
3472
+ "learning_rate": 3.9537274319883016e-05,
3473
+ "loss": 5.6255,
3474
+ "step": 173500
3475
+ },
3476
+ {
3477
+ "epoch": 0.63,
3478
+ "learning_rate": 3.950711124624168e-05,
3479
+ "loss": 5.6247,
3480
+ "step": 174000
3481
+ },
3482
+ {
3483
+ "epoch": 0.63,
3484
+ "eval_loss": 5.677004814147949,
3485
+ "eval_runtime": 4.3652,
3486
+ "eval_samples_per_second": 229.316,
3487
+ "eval_steps_per_second": 14.432,
3488
+ "step": 174000
3489
+ },
3490
+ {
3491
+ "epoch": 0.63,
3492
+ "learning_rate": 3.947700849874763e-05,
3493
+ "loss": 5.6099,
3494
+ "step": 174500
3495
+ },
3496
+ {
3497
+ "epoch": 0.63,
3498
+ "learning_rate": 3.94468454251063e-05,
3499
+ "loss": 5.6204,
3500
+ "step": 175000
3501
+ },
3502
+ {
3503
+ "epoch": 0.63,
3504
+ "eval_loss": 5.616661548614502,
3505
+ "eval_runtime": 4.3551,
3506
+ "eval_samples_per_second": 229.845,
3507
+ "eval_steps_per_second": 14.466,
3508
+ "step": 175000
3509
+ },
3510
+ {
3511
+ "epoch": 0.64,
3512
+ "learning_rate": 3.941668235146496e-05,
3513
+ "loss": 5.6196,
3514
+ "step": 175500
3515
+ },
3516
+ {
3517
+ "epoch": 0.64,
3518
+ "learning_rate": 3.938651927782363e-05,
3519
+ "loss": 5.6166,
3520
+ "step": 176000
3521
+ },
3522
+ {
3523
+ "epoch": 0.64,
3524
+ "eval_loss": 5.683280944824219,
3525
+ "eval_runtime": 4.3154,
3526
+ "eval_samples_per_second": 231.959,
3527
+ "eval_steps_per_second": 14.599,
3528
+ "step": 176000
3529
+ },
3530
+ {
3531
+ "epoch": 0.64,
3532
+ "learning_rate": 3.9356356204182294e-05,
3533
+ "loss": 5.6303,
3534
+ "step": 176500
3535
+ },
3536
+ {
3537
+ "epoch": 0.64,
3538
+ "learning_rate": 3.932625345668824e-05,
3539
+ "loss": 5.6145,
3540
+ "step": 177000
3541
+ },
3542
+ {
3543
+ "epoch": 0.64,
3544
+ "eval_loss": 5.708868980407715,
3545
+ "eval_runtime": 4.3953,
3546
+ "eval_samples_per_second": 227.744,
3547
+ "eval_steps_per_second": 14.334,
3548
+ "step": 177000
3549
+ },
3550
+ {
3551
+ "epoch": 0.64,
3552
+ "learning_rate": 3.929609038304691e-05,
3553
+ "loss": 5.6208,
3554
+ "step": 177500
3555
+ },
3556
+ {
3557
+ "epoch": 0.64,
3558
+ "learning_rate": 3.926592730940557e-05,
3559
+ "loss": 5.6155,
3560
+ "step": 178000
3561
+ },
3562
+ {
3563
+ "epoch": 0.64,
3564
+ "eval_loss": 5.643332481384277,
3565
+ "eval_runtime": 4.3751,
3566
+ "eval_samples_per_second": 228.794,
3567
+ "eval_steps_per_second": 14.4,
3568
+ "step": 178000
3569
+ },
3570
+ {
3571
+ "epoch": 0.65,
3572
+ "learning_rate": 3.923576423576424e-05,
3573
+ "loss": 5.6361,
3574
+ "step": 178500
3575
+ },
3576
+ {
3577
+ "epoch": 0.65,
3578
+ "learning_rate": 3.9205601162122905e-05,
3579
+ "loss": 5.6162,
3580
+ "step": 179000
3581
+ },
3582
+ {
3583
+ "epoch": 0.65,
3584
+ "eval_loss": 5.648618221282959,
3585
+ "eval_runtime": 4.3904,
3586
+ "eval_samples_per_second": 227.999,
3587
+ "eval_steps_per_second": 14.35,
3588
+ "step": 179000
3589
+ },
3590
+ {
3591
+ "epoch": 0.65,
3592
+ "learning_rate": 3.917543808848157e-05,
3593
+ "loss": 5.6344,
3594
+ "step": 179500
3595
+ },
3596
+ {
3597
+ "epoch": 0.65,
3598
+ "learning_rate": 3.914527501484023e-05,
3599
+ "loss": 5.6144,
3600
+ "step": 180000
3601
+ },
3602
+ {
3603
+ "epoch": 0.65,
3604
+ "eval_loss": 5.633224964141846,
3605
+ "eval_runtime": 4.3929,
3606
+ "eval_samples_per_second": 227.865,
3607
+ "eval_steps_per_second": 14.341,
3608
+ "step": 180000
3609
+ },
3610
+ {
3611
+ "epoch": 0.65,
3612
+ "learning_rate": 3.911517226734618e-05,
3613
+ "loss": 5.6285,
3614
+ "step": 180500
3615
+ },
3616
+ {
3617
+ "epoch": 0.66,
3618
+ "learning_rate": 3.908500919370485e-05,
3619
+ "loss": 5.6198,
3620
+ "step": 181000
3621
+ },
3622
+ {
3623
+ "epoch": 0.66,
3624
+ "eval_loss": 5.645481109619141,
3625
+ "eval_runtime": 4.3967,
3626
+ "eval_samples_per_second": 227.672,
3627
+ "eval_steps_per_second": 14.329,
3628
+ "step": 181000
3629
+ },
3630
+ {
3631
+ "epoch": 0.66,
3632
+ "learning_rate": 3.9054846120063515e-05,
3633
+ "loss": 5.6095,
3634
+ "step": 181500
3635
+ },
3636
+ {
3637
+ "epoch": 0.66,
3638
+ "learning_rate": 3.9024683046422175e-05,
3639
+ "loss": 5.6231,
3640
+ "step": 182000
3641
+ },
3642
+ {
3643
+ "epoch": 0.66,
3644
+ "eval_loss": 5.6638264656066895,
3645
+ "eval_runtime": 4.4238,
3646
+ "eval_samples_per_second": 226.274,
3647
+ "eval_steps_per_second": 14.241,
3648
+ "step": 182000
3649
+ },
3650
+ {
3651
+ "epoch": 0.66,
3652
+ "learning_rate": 3.899451997278085e-05,
3653
+ "loss": 5.6158,
3654
+ "step": 182500
3655
+ },
3656
+ {
3657
+ "epoch": 0.66,
3658
+ "learning_rate": 3.896441722528679e-05,
3659
+ "loss": 5.61,
3660
+ "step": 183000
3661
+ },
3662
+ {
3663
+ "epoch": 0.66,
3664
+ "eval_loss": 5.614964962005615,
3665
+ "eval_runtime": 4.3378,
3666
+ "eval_samples_per_second": 230.76,
3667
+ "eval_steps_per_second": 14.523,
3668
+ "step": 183000
3669
+ },
3670
+ {
3671
+ "epoch": 0.66,
3672
+ "learning_rate": 3.893425415164545e-05,
3673
+ "loss": 5.5975,
3674
+ "step": 183500
3675
+ },
3676
+ {
3677
+ "epoch": 0.67,
3678
+ "learning_rate": 3.8904091078004126e-05,
3679
+ "loss": 5.614,
3680
+ "step": 184000
3681
+ },
3682
+ {
3683
+ "epoch": 0.67,
3684
+ "eval_loss": 5.670050621032715,
3685
+ "eval_runtime": 4.3122,
3686
+ "eval_samples_per_second": 232.134,
3687
+ "eval_steps_per_second": 14.61,
3688
+ "step": 184000
3689
+ },
3690
+ {
3691
+ "epoch": 0.67,
3692
+ "learning_rate": 3.8873928004362786e-05,
3693
+ "loss": 5.6029,
3694
+ "step": 184500
3695
+ },
3696
+ {
3697
+ "epoch": 0.67,
3698
+ "learning_rate": 3.8843825256868736e-05,
3699
+ "loss": 5.6158,
3700
+ "step": 185000
3701
+ },
3702
+ {
3703
+ "epoch": 0.67,
3704
+ "eval_loss": 5.641917705535889,
3705
+ "eval_runtime": 4.3145,
3706
+ "eval_samples_per_second": 232.008,
3707
+ "eval_steps_per_second": 14.602,
3708
+ "step": 185000
3709
+ },
3710
+ {
3711
+ "epoch": 0.67,
3712
+ "learning_rate": 3.88136621832274e-05,
3713
+ "loss": 5.6351,
3714
+ "step": 185500
3715
+ },
3716
+ {
3717
+ "epoch": 0.67,
3718
+ "learning_rate": 3.878349910958607e-05,
3719
+ "loss": 5.6163,
3720
+ "step": 186000
3721
+ },
3722
+ {
3723
+ "epoch": 0.67,
3724
+ "eval_loss": 5.651737213134766,
3725
+ "eval_runtime": 4.3373,
3726
+ "eval_samples_per_second": 230.79,
3727
+ "eval_steps_per_second": 14.525,
3728
+ "step": 186000
3729
+ },
3730
+ {
3731
+ "epoch": 0.68,
3732
+ "learning_rate": 3.875333603594473e-05,
3733
+ "loss": 5.5979,
3734
+ "step": 186500
3735
+ },
3736
+ {
3737
+ "epoch": 0.68,
3738
+ "learning_rate": 3.8723172962303403e-05,
3739
+ "loss": 5.6151,
3740
+ "step": 187000
3741
+ },
3742
+ {
3743
+ "epoch": 0.68,
3744
+ "eval_loss": 5.646933078765869,
3745
+ "eval_runtime": 4.2964,
3746
+ "eval_samples_per_second": 232.987,
3747
+ "eval_steps_per_second": 14.664,
3748
+ "step": 187000
3749
+ },
3750
+ {
3751
+ "epoch": 0.68,
3752
+ "learning_rate": 3.8693070214809346e-05,
3753
+ "loss": 5.6236,
3754
+ "step": 187500
3755
+ },
3756
+ {
3757
+ "epoch": 0.68,
3758
+ "learning_rate": 3.866290714116801e-05,
3759
+ "loss": 5.6251,
3760
+ "step": 188000
3761
+ },
3762
+ {
3763
+ "epoch": 0.68,
3764
+ "eval_loss": 5.671980381011963,
3765
+ "eval_runtime": 4.4739,
3766
+ "eval_samples_per_second": 223.743,
3767
+ "eval_steps_per_second": 14.082,
3768
+ "step": 188000
3769
+ },
3770
+ {
3771
+ "epoch": 0.68,
3772
+ "learning_rate": 3.863274406752668e-05,
3773
+ "loss": 5.6018,
3774
+ "step": 188500
3775
+ },
3776
+ {
3777
+ "epoch": 0.68,
3778
+ "learning_rate": 3.860258099388535e-05,
3779
+ "loss": 5.6272,
3780
+ "step": 189000
3781
+ },
3782
+ {
3783
+ "epoch": 0.68,
3784
+ "eval_loss": 5.603824138641357,
3785
+ "eval_runtime": 4.3496,
3786
+ "eval_samples_per_second": 230.135,
3787
+ "eval_steps_per_second": 14.484,
3788
+ "step": 189000
3789
+ },
3790
+ {
3791
+ "epoch": 0.69,
3792
+ "learning_rate": 3.857247824639129e-05,
3793
+ "loss": 5.6282,
3794
+ "step": 189500
3795
+ },
3796
+ {
3797
+ "epoch": 0.69,
3798
+ "learning_rate": 3.854231517274996e-05,
3799
+ "loss": 5.6291,
3800
+ "step": 190000
3801
+ },
3802
+ {
3803
+ "epoch": 0.69,
3804
+ "eval_loss": 5.6158552169799805,
3805
+ "eval_runtime": 6.4339,
3806
+ "eval_samples_per_second": 155.582,
3807
+ "eval_steps_per_second": 9.792,
3808
+ "step": 190000
3809
+ },
3810
+ {
3811
+ "epoch": 0.69,
3812
+ "learning_rate": 3.8512152099108624e-05,
3813
+ "loss": 5.6017,
3814
+ "step": 190500
3815
+ },
3816
+ {
3817
+ "epoch": 0.69,
3818
+ "learning_rate": 3.8481989025467284e-05,
3819
+ "loss": 5.6114,
3820
+ "step": 191000
3821
+ },
3822
+ {
3823
+ "epoch": 0.69,
3824
+ "eval_loss": 5.6429595947265625,
3825
+ "eval_runtime": 4.3858,
3826
+ "eval_samples_per_second": 228.235,
3827
+ "eval_steps_per_second": 14.364,
3828
+ "step": 191000
3829
+ },
3830
+ {
3831
+ "epoch": 0.69,
3832
+ "learning_rate": 3.845182595182595e-05,
3833
+ "loss": 5.6194,
3834
+ "step": 191500
3835
+ },
3836
+ {
3837
+ "epoch": 0.69,
3838
+ "learning_rate": 3.842166287818462e-05,
3839
+ "loss": 5.6128,
3840
+ "step": 192000
3841
+ },
3842
+ {
3843
+ "epoch": 0.69,
3844
+ "eval_loss": 5.634527683258057,
3845
+ "eval_runtime": 4.3652,
3846
+ "eval_samples_per_second": 229.313,
3847
+ "eval_steps_per_second": 14.432,
3848
+ "step": 192000
3849
+ },
3850
+ {
3851
+ "epoch": 0.7,
3852
+ "learning_rate": 3.8391499804543285e-05,
3853
+ "loss": 5.62,
3854
+ "step": 192500
3855
+ },
3856
+ {
3857
+ "epoch": 0.7,
3858
+ "learning_rate": 3.836139705704923e-05,
3859
+ "loss": 5.6213,
3860
+ "step": 193000
3861
+ },
3862
+ {
3863
+ "epoch": 0.7,
3864
+ "eval_loss": 5.641001224517822,
3865
+ "eval_runtime": 4.355,
3866
+ "eval_samples_per_second": 229.851,
3867
+ "eval_steps_per_second": 14.466,
3868
+ "step": 193000
3869
+ },
3870
+ {
3871
+ "epoch": 0.7,
3872
+ "learning_rate": 3.83312339834079e-05,
3873
+ "loss": 5.6182,
3874
+ "step": 193500
3875
+ },
3876
+ {
3877
+ "epoch": 0.7,
3878
+ "learning_rate": 3.830107090976656e-05,
3879
+ "loss": 5.6104,
3880
+ "step": 194000
3881
+ },
3882
+ {
3883
+ "epoch": 0.7,
3884
+ "eval_loss": 5.6409525871276855,
3885
+ "eval_runtime": 4.3644,
3886
+ "eval_samples_per_second": 229.353,
3887
+ "eval_steps_per_second": 14.435,
3888
+ "step": 194000
3889
+ },
3890
+ {
3891
+ "epoch": 0.7,
3892
+ "learning_rate": 3.827090783612523e-05,
3893
+ "loss": 5.6061,
3894
+ "step": 194500
3895
+ },
3896
+ {
3897
+ "epoch": 0.71,
3898
+ "learning_rate": 3.8240744762483896e-05,
3899
+ "loss": 5.6081,
3900
+ "step": 195000
3901
+ },
3902
+ {
3903
+ "epoch": 0.71,
3904
+ "eval_loss": 5.624050617218018,
3905
+ "eval_runtime": 4.463,
3906
+ "eval_samples_per_second": 224.291,
3907
+ "eval_steps_per_second": 14.116,
3908
+ "step": 195000
3909
+ },
3910
+ {
3911
+ "epoch": 0.71,
3912
+ "learning_rate": 3.821058168884256e-05,
3913
+ "loss": 5.6234,
3914
+ "step": 195500
3915
+ },
3916
+ {
3917
+ "epoch": 0.71,
3918
+ "learning_rate": 3.8180478941348505e-05,
3919
+ "loss": 5.6288,
3920
+ "step": 196000
3921
+ },
3922
+ {
3923
+ "epoch": 0.71,
3924
+ "eval_loss": 5.689996242523193,
3925
+ "eval_runtime": 6.3147,
3926
+ "eval_samples_per_second": 158.518,
3927
+ "eval_steps_per_second": 9.977,
3928
+ "step": 196000
3929
+ },
3930
+ {
3931
+ "epoch": 0.71,
3932
+ "learning_rate": 3.815031586770718e-05,
3933
+ "loss": 5.6253,
3934
+ "step": 196500
3935
+ },
3936
+ {
3937
+ "epoch": 0.71,
3938
+ "learning_rate": 3.812015279406584e-05,
3939
+ "loss": 5.607,
3940
+ "step": 197000
3941
+ },
3942
+ {
3943
+ "epoch": 0.71,
3944
+ "eval_loss": 5.656425952911377,
3945
+ "eval_runtime": 4.3871,
3946
+ "eval_samples_per_second": 228.167,
3947
+ "eval_steps_per_second": 14.36,
3948
+ "step": 197000
3949
+ },
3950
+ {
3951
+ "epoch": 0.71,
3952
+ "learning_rate": 3.8089989720424506e-05,
3953
+ "loss": 5.6259,
3954
+ "step": 197500
3955
+ },
3956
+ {
3957
+ "epoch": 0.72,
3958
+ "learning_rate": 3.805982664678317e-05,
3959
+ "loss": 5.605,
3960
+ "step": 198000
3961
+ },
3962
+ {
3963
+ "epoch": 0.72,
3964
+ "eval_loss": 5.651924133300781,
3965
+ "eval_runtime": 4.362,
3966
+ "eval_samples_per_second": 229.484,
3967
+ "eval_steps_per_second": 14.443,
3968
+ "step": 198000
3969
+ },
3970
+ {
3971
+ "epoch": 0.72,
3972
+ "learning_rate": 3.8029723899289116e-05,
3973
+ "loss": 5.6153,
3974
+ "step": 198500
3975
+ },
3976
+ {
3977
+ "epoch": 0.72,
3978
+ "learning_rate": 3.799956082564778e-05,
3979
+ "loss": 5.6109,
3980
+ "step": 199000
3981
+ },
3982
+ {
3983
+ "epoch": 0.72,
3984
+ "eval_loss": 5.639188766479492,
3985
+ "eval_runtime": 4.301,
3986
+ "eval_samples_per_second": 232.739,
3987
+ "eval_steps_per_second": 14.648,
3988
+ "step": 199000
3989
+ },
3990
+ {
3991
+ "epoch": 0.72,
3992
+ "learning_rate": 3.796939775200645e-05,
3993
+ "loss": 5.6064,
3994
+ "step": 199500
3995
+ },
3996
+ {
3997
+ "epoch": 0.72,
3998
+ "learning_rate": 3.793923467836512e-05,
3999
+ "loss": 5.619,
4000
+ "step": 200000
4001
+ },
4002
+ {
4003
+ "epoch": 0.72,
4004
+ "eval_loss": 5.662364959716797,
4005
+ "eval_runtime": 4.3023,
4006
+ "eval_samples_per_second": 232.664,
4007
+ "eval_steps_per_second": 14.643,
4008
+ "step": 200000
4009
+ },
4010
+ {
4011
+ "epoch": 0.73,
4012
+ "learning_rate": 3.790913193087106e-05,
4013
+ "loss": 5.6079,
4014
+ "step": 200500
4015
+ },
4016
+ {
4017
+ "epoch": 0.73,
4018
+ "learning_rate": 3.7878968857229727e-05,
4019
+ "loss": 5.6019,
4020
+ "step": 201000
4021
+ },
4022
+ {
4023
+ "epoch": 0.73,
4024
+ "eval_loss": 5.653491973876953,
4025
+ "eval_runtime": 4.2771,
4026
+ "eval_samples_per_second": 234.036,
4027
+ "eval_steps_per_second": 14.73,
4028
+ "step": 201000
4029
+ },
4030
+ {
4031
+ "epoch": 0.73,
4032
+ "learning_rate": 3.7848805783588394e-05,
4033
+ "loss": 5.6022,
4034
+ "step": 201500
4035
+ },
4036
+ {
4037
+ "epoch": 0.73,
4038
+ "learning_rate": 3.781864270994706e-05,
4039
+ "loss": 5.6133,
4040
+ "step": 202000
4041
+ },
4042
+ {
4043
+ "epoch": 0.73,
4044
+ "eval_loss": 5.676525115966797,
4045
+ "eval_runtime": 4.3425,
4046
+ "eval_samples_per_second": 230.512,
4047
+ "eval_steps_per_second": 14.508,
4048
+ "step": 202000
4049
+ },
4050
+ {
4051
+ "epoch": 0.73,
4052
+ "learning_rate": 3.778847963630572e-05,
4053
+ "loss": 5.6117,
4054
+ "step": 202500
4055
+ },
4056
+ {
4057
+ "epoch": 0.73,
4058
+ "learning_rate": 3.775837688881168e-05,
4059
+ "loss": 5.5927,
4060
+ "step": 203000
4061
+ },
4062
+ {
4063
+ "epoch": 0.73,
4064
+ "eval_loss": 5.636429786682129,
4065
+ "eval_runtime": 4.3367,
4066
+ "eval_samples_per_second": 230.822,
4067
+ "eval_steps_per_second": 14.527,
4068
+ "step": 203000
4069
+ },
4070
+ {
4071
+ "epoch": 0.74,
4072
+ "learning_rate": 3.772821381517034e-05,
4073
+ "loss": 5.6198,
4074
+ "step": 203500
4075
+ },
4076
+ {
4077
+ "epoch": 0.74,
4078
+ "learning_rate": 3.7698050741529004e-05,
4079
+ "loss": 5.6119,
4080
+ "step": 204000
4081
+ },
4082
+ {
4083
+ "epoch": 0.74,
4084
+ "eval_loss": 5.666501045227051,
4085
+ "eval_runtime": 4.3287,
4086
+ "eval_samples_per_second": 231.246,
4087
+ "eval_steps_per_second": 14.554,
4088
+ "step": 204000
4089
+ },
4090
+ {
4091
+ "epoch": 0.74,
4092
+ "learning_rate": 3.766788766788767e-05,
4093
+ "loss": 5.6199,
4094
+ "step": 204500
4095
+ },
4096
+ {
4097
+ "epoch": 0.74,
4098
+ "learning_rate": 3.763778492039362e-05,
4099
+ "loss": 5.602,
4100
+ "step": 205000
4101
+ },
4102
+ {
4103
+ "epoch": 0.74,
4104
+ "eval_loss": 5.631441116333008,
4105
+ "eval_runtime": 4.3578,
4106
+ "eval_samples_per_second": 229.703,
4107
+ "eval_steps_per_second": 14.457,
4108
+ "step": 205000
4109
+ },
4110
+ {
4111
+ "epoch": 0.74,
4112
+ "learning_rate": 3.760762184675228e-05,
4113
+ "loss": 5.5997,
4114
+ "step": 205500
4115
+ },
4116
+ {
4117
+ "epoch": 0.75,
4118
+ "learning_rate": 3.757745877311095e-05,
4119
+ "loss": 5.605,
4120
+ "step": 206000
4121
+ },
4122
+ {
4123
+ "epoch": 0.75,
4124
+ "eval_loss": 5.630640983581543,
4125
+ "eval_runtime": 4.2859,
4126
+ "eval_samples_per_second": 233.558,
4127
+ "eval_steps_per_second": 14.699,
4128
+ "step": 206000
4129
+ },
4130
+ {
4131
+ "epoch": 0.75,
4132
+ "learning_rate": 3.7547295699469615e-05,
4133
+ "loss": 5.6155,
4134
+ "step": 206500
4135
+ },
4136
+ {
4137
+ "epoch": 0.75,
4138
+ "learning_rate": 3.7517132625828275e-05,
4139
+ "loss": 5.612,
4140
+ "step": 207000
4141
+ },
4142
+ {
4143
+ "epoch": 0.75,
4144
+ "eval_loss": 5.597905158996582,
4145
+ "eval_runtime": 4.3535,
4146
+ "eval_samples_per_second": 229.93,
4147
+ "eval_steps_per_second": 14.471,
4148
+ "step": 207000
4149
+ },
4150
+ {
4151
+ "epoch": 0.75,
4152
+ "learning_rate": 3.748696955218695e-05,
4153
+ "loss": 5.6189,
4154
+ "step": 207500
4155
+ },
4156
+ {
4157
+ "epoch": 0.75,
4158
+ "learning_rate": 3.745680647854561e-05,
4159
+ "loss": 5.6184,
4160
+ "step": 208000
4161
+ },
4162
+ {
4163
+ "epoch": 0.75,
4164
+ "eval_loss": 5.636325359344482,
4165
+ "eval_runtime": 4.3413,
4166
+ "eval_samples_per_second": 230.575,
4167
+ "eval_steps_per_second": 14.512,
4168
+ "step": 208000
4169
+ },
4170
+ {
4171
+ "epoch": 0.75,
4172
+ "learning_rate": 3.742670373105156e-05,
4173
+ "loss": 5.6078,
4174
+ "step": 208500
4175
+ },
4176
+ {
4177
+ "epoch": 0.76,
4178
+ "learning_rate": 3.7396540657410225e-05,
4179
+ "loss": 5.6131,
4180
+ "step": 209000
4181
+ },
4182
+ {
4183
+ "epoch": 0.76,
4184
+ "eval_loss": 5.635697364807129,
4185
+ "eval_runtime": 4.3624,
4186
+ "eval_samples_per_second": 229.461,
4187
+ "eval_steps_per_second": 14.442,
4188
+ "step": 209000
4189
+ },
4190
+ {
4191
+ "epoch": 0.76,
4192
+ "learning_rate": 3.736637758376889e-05,
4193
+ "loss": 5.6102,
4194
+ "step": 209500
4195
+ },
4196
+ {
4197
+ "epoch": 0.76,
4198
+ "learning_rate": 3.733621451012755e-05,
4199
+ "loss": 5.6063,
4200
+ "step": 210000
4201
+ },
4202
+ {
4203
+ "epoch": 0.76,
4204
+ "eval_loss": 5.6277570724487305,
4205
+ "eval_runtime": 4.4064,
4206
+ "eval_samples_per_second": 227.169,
4207
+ "eval_steps_per_second": 14.297,
4208
+ "step": 210000
4209
+ }
4210
+ ],
4211
+ "max_steps": 828828,
4212
+ "num_train_epochs": 3,
4213
+ "total_flos": 2.574739225286738e+17,
4214
+ "trial_name": null,
4215
+ "trial_params": null
4216
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5f6c16d5e3fb714d6d4b43f9e3d9077909eaef9723b3c5411687a5655ef2ca4
3
+ size 2991