Hugging Face's logo

morenolq
/
bart-it-WITS

Transformers
PyTorch
Safetensors
Italian
bart
text2text-generation
Model card Files
xet
 Community
1
MorenoLaQuatra commited on
Commit
9a16cb0
·
1 Parent(s): fa0aced

Model Publication

Browse files
Files changed (13) hide show
  1. README.md +63 -0
  2. config.json +45 -0
  3. merges.txt +0 -0
  4. optimizer.pt +3 -0
  5. pytorch_model.bin +3 -0
  6. rng_state.pth +3 -0
  7. scaler.pt +3 -0
  8. scheduler.pt +3 -0
  9. special_tokens_map.json +51 -0
  10. tokenizer_config.json +61 -0
  11. trainer_state.json +3928 -0
  12. training_args.bin +3 -0
  13. vocab.json +0 -0
README.md CHANGED
@@ -1,3 +1,66 @@
1
  ---
 
2
  license: mit
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ language: "it"
3
  license: mit
4
+ datasets:
5
+ - Silvia/WITS
6
+ tags:
7
+ - bart
8
+ - pytorch
9
+ pipeline:
10
+ - summarization
11
  ---
12
+
13
+ # BART-IT - FanPage abstractive summarization
14
+
15
+ BART-IT is a sequence-to-sequence model, based on the BART architecture that is specifically tailored to the Italian language. The model is pre-trained on a [large corpus of Italian text](https://huggingface.co/datasets/gsarti/clean_mc4_it), and can be fine-tuned on a variety of tasks.
16
+
17
+ ## Model description
18
+
19
+ The model is a `base-`sized BART model, with a vocabulary size of 52,000 tokens. It has 140M parameters and can be used for any task that requires a sequence-to-sequence model. It is trained from scratch on a large corpus of Italian text, and can be fine-tuned on a variety of tasks.
20
+
21
+
22
+ ## Pre-training
23
+
24
+ The code used to pre-train BART-IT together with additional information on model parameters can be found [here](https://github.com/MorenoLaQuatra/bart-it).
25
+
26
+ ## Fine-tuning
27
+
28
+ The model has been fine-tuned for the abstractive summarization task on 3 different Italian datasets:
29
+
30
+ - [FanPage](https://huggingface.co/datasets/ARTeLab/fanpage) - finetuned model [here](https://huggingface.co/MorenoLaQuatra/bart-it-fanpage)
31
+ - [IlPost](https://huggingface.co/datasets/ARTeLab/ilpost) - finetuned model [here](https://huggingface.co/MorenoLaQuatra/bart-it-ilpost)
32
+ - **This model** [WITS](https://huggingface.co/datasets/Silvia/WITS) - finetuned model [here](https://huggingface.co/MorenoLaQuatra/bart-it-WITS)
33
+
34
+ ## Usage
35
+
36
+ In order to use the model, you can use the following code:
37
+
38
+ ```python
39
+ from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
40
+
41
+ tokenizer = AutoTokenizer.from_pretrained("morenolq/bart-it-WITS")
42
+ model = AutoModelForSeq2SeqLM.from_pretrained("morenolq/bart-it-WITS")
43
+
44
+ input_ids = tokenizer.encode("Il modello BART-IT è stato pre-addestrato su un corpus di testo italiano", return_tensors="pt")
45
+ outputs = model.generate(input_ids, max_length=40, num_beams=4, early_stopping=True)
46
+ print(tokenizer.decode(outputs[0], skip_special_tokens=True))
47
+ ```
48
+
49
+ # Citation
50
+
51
+ If you find this model useful for your research, please cite the following paper:
52
+
53
+ ```bibtex
54
+ @Article{BARTIT,
55
+ AUTHOR = {La Quatra, Moreno and Cagliero, Luca},
56
+ TITLE = {BART-IT: An Efficient Sequence-to-Sequence Model for Italian Text Summarization},
57
+ JOURNAL = {Future Internet},
58
+ VOLUME = {15},
59
+ YEAR = {2023},
60
+ NUMBER = {1},
61
+ ARTICLE-NUMBER = {15},
62
+ URL = {https://www.mdpi.com/1999-5903/15/1/15},
63
+ ISSN = {1999-5903},
64
+ DOI = {10.3390/fi15010015}
65
+ }
66
+ ```
config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "../bart-it-s",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "architectures": [
6
+ "BartForConditionalGeneration"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 0,
10
+ "classifier_dropout": 0.0,
11
+ "d_model": 768,
12
+ "decoder_attention_heads": 12,
13
+ "decoder_ffn_dim": 3072,
14
+ "decoder_layerdrop": 0.0,
15
+ "decoder_layers": 6,
16
+ "decoder_start_token_id": 2,
17
+ "dropout": 0.1,
18
+ "encoder_attention_heads": 12,
19
+ "encoder_ffn_dim": 3072,
20
+ "encoder_layerdrop": 0.0,
21
+ "encoder_layers": 6,
22
+ "eos_token_id": 2,
23
+ "forced_eos_token_id": 2,
24
+ "id2label": {
25
+ "0": "LABEL_0",
26
+ "1": "LABEL_1",
27
+ "2": "LABEL_2"
28
+ },
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "label2id": {
32
+ "LABEL_0": 0,
33
+ "LABEL_1": 1,
34
+ "LABEL_2": 2
35
+ },
36
+ "max_position_embeddings": 1024,
37
+ "model_type": "bart",
38
+ "num_hidden_layers": 6,
39
+ "pad_token_id": 1,
40
+ "scale_embedding": false,
41
+ "torch_dtype": "float32",
42
+ "transformers_version": "4.22.1",
43
+ "use_cache": true,
44
+ "vocab_size": 52000
45
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b3924336104a61fb6f40904a1db46000c9d241d0b04904f0f37554e2df5a1e6e
3
+ size 1126173557
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e8bba015730061419a62b9a252ef873aac1cc1cf9469ecb9c586c4230ae02207
3
+ size 563305977
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e39d7b0b4214e7b54f9d97aa04b6cf98efddf12863afcc858d62dd48e61539b4
3
+ size 15523
scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:065f8f26cceeb1f27fd509c28355e76259811bd58ce56cd862d4f918dd8697b9
3
+ size 559
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e46b00642581534b2891547eb2b6896d8f932ee5d42a6d884195aa83b7f20666
3
+ size 623
special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": true,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": true,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": true,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": true,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": {
4
+ "__type": "AddedToken",
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false
10
+ },
11
+ "cls_token": {
12
+ "__type": "AddedToken",
13
+ "content": "<s>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false
18
+ },
19
+ "eos_token": {
20
+ "__type": "AddedToken",
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "errors": "replace",
28
+ "mask_token": {
29
+ "__type": "AddedToken",
30
+ "content": "<mask>",
31
+ "lstrip": true,
32
+ "normalized": true,
33
+ "rstrip": false,
34
+ "single_word": false
35
+ },
36
+ "pad_token": {
37
+ "__type": "AddedToken",
38
+ "content": "<pad>",
39
+ "lstrip": false,
40
+ "normalized": true,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "sep_token": {
45
+ "__type": "AddedToken",
46
+ "content": "</s>",
47
+ "lstrip": false,
48
+ "normalized": true,
49
+ "rstrip": false,
50
+ "single_word": false
51
+ },
52
+ "tokenizer_class": "BartTokenizer",
53
+ "unk_token": {
54
+ "__type": "AddedToken",
55
+ "content": "<unk>",
56
+ "lstrip": false,
57
+ "normalized": true,
58
+ "rstrip": false,
59
+ "single_word": false
60
+ }
61
+ }
trainer_state.json ADDED
@@ -0,0 +1,3928 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.3479,
3
+ "best_model_checkpoint": "wits_checkpoints/checkpoint-63702",
4
+ "epoch": 6.0,
5
+ "global_step": 63702,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 1.9800000000000004e-05,
13
+ "loss": 3.3878,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 0.02,
18
+ "learning_rate": 3.9800000000000005e-05,
19
+ "loss": 2.6013,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 0.03,
24
+ "learning_rate": 5.9800000000000003e-05,
25
+ "loss": 2.4783,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 0.04,
30
+ "learning_rate": 7.98e-05,
31
+ "loss": 2.4172,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 0.05,
36
+ "learning_rate": 9.98e-05,
37
+ "loss": 2.3612,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 0.06,
42
+ "learning_rate": 9.990631210371913e-05,
43
+ "loss": 2.3416,
44
+ "step": 600
45
+ },
46
+ {
47
+ "epoch": 0.07,
48
+ "learning_rate": 9.981167786505158e-05,
49
+ "loss": 2.2736,
50
+ "step": 700
51
+ },
52
+ {
53
+ "epoch": 0.08,
54
+ "learning_rate": 9.971704362638403e-05,
55
+ "loss": 2.2925,
56
+ "step": 800
57
+ },
58
+ {
59
+ "epoch": 0.08,
60
+ "learning_rate": 9.962240938771648e-05,
61
+ "loss": 2.2689,
62
+ "step": 900
63
+ },
64
+ {
65
+ "epoch": 0.09,
66
+ "learning_rate": 9.952777514904894e-05,
67
+ "loss": 2.2445,
68
+ "step": 1000
69
+ },
70
+ {
71
+ "epoch": 0.1,
72
+ "learning_rate": 9.943314091038138e-05,
73
+ "loss": 2.203,
74
+ "step": 1100
75
+ },
76
+ {
77
+ "epoch": 0.11,
78
+ "learning_rate": 9.933850667171382e-05,
79
+ "loss": 2.1892,
80
+ "step": 1200
81
+ },
82
+ {
83
+ "epoch": 0.12,
84
+ "learning_rate": 9.924387243304628e-05,
85
+ "loss": 2.1888,
86
+ "step": 1300
87
+ },
88
+ {
89
+ "epoch": 0.13,
90
+ "learning_rate": 9.914923819437874e-05,
91
+ "loss": 2.1622,
92
+ "step": 1400
93
+ },
94
+ {
95
+ "epoch": 0.14,
96
+ "learning_rate": 9.905460395571118e-05,
97
+ "loss": 2.1601,
98
+ "step": 1500
99
+ },
100
+ {
101
+ "epoch": 0.15,
102
+ "learning_rate": 9.895996971704363e-05,
103
+ "loss": 2.1562,
104
+ "step": 1600
105
+ },
106
+ {
107
+ "epoch": 0.16,
108
+ "learning_rate": 9.886533547837609e-05,
109
+ "loss": 2.1557,
110
+ "step": 1700
111
+ },
112
+ {
113
+ "epoch": 0.17,
114
+ "learning_rate": 9.877070123970853e-05,
115
+ "loss": 2.1417,
116
+ "step": 1800
117
+ },
118
+ {
119
+ "epoch": 0.18,
120
+ "learning_rate": 9.867606700104099e-05,
121
+ "loss": 2.1268,
122
+ "step": 1900
123
+ },
124
+ {
125
+ "epoch": 0.19,
126
+ "learning_rate": 9.858143276237343e-05,
127
+ "loss": 2.1194,
128
+ "step": 2000
129
+ },
130
+ {
131
+ "epoch": 0.2,
132
+ "learning_rate": 9.848679852370588e-05,
133
+ "loss": 2.1199,
134
+ "step": 2100
135
+ },
136
+ {
137
+ "epoch": 0.21,
138
+ "learning_rate": 9.839216428503833e-05,
139
+ "loss": 2.1339,
140
+ "step": 2200
141
+ },
142
+ {
143
+ "epoch": 0.22,
144
+ "learning_rate": 9.829753004637078e-05,
145
+ "loss": 2.1011,
146
+ "step": 2300
147
+ },
148
+ {
149
+ "epoch": 0.23,
150
+ "learning_rate": 9.820289580770324e-05,
151
+ "loss": 2.1096,
152
+ "step": 2400
153
+ },
154
+ {
155
+ "epoch": 0.24,
156
+ "learning_rate": 9.810826156903568e-05,
157
+ "loss": 2.0886,
158
+ "step": 2500
159
+ },
160
+ {
161
+ "epoch": 0.24,
162
+ "learning_rate": 9.801362733036812e-05,
163
+ "loss": 2.0803,
164
+ "step": 2600
165
+ },
166
+ {
167
+ "epoch": 0.25,
168
+ "learning_rate": 9.791899309170058e-05,
169
+ "loss": 2.0882,
170
+ "step": 2700
171
+ },
172
+ {
173
+ "epoch": 0.26,
174
+ "learning_rate": 9.782435885303303e-05,
175
+ "loss": 2.0889,
176
+ "step": 2800
177
+ },
178
+ {
179
+ "epoch": 0.27,
180
+ "learning_rate": 9.772972461436548e-05,
181
+ "loss": 2.0908,
182
+ "step": 2900
183
+ },
184
+ {
185
+ "epoch": 0.28,
186
+ "learning_rate": 9.763509037569793e-05,
187
+ "loss": 2.0775,
188
+ "step": 3000
189
+ },
190
+ {
191
+ "epoch": 0.29,
192
+ "learning_rate": 9.754045613703037e-05,
193
+ "loss": 2.0718,
194
+ "step": 3100
195
+ },
196
+ {
197
+ "epoch": 0.3,
198
+ "learning_rate": 9.744582189836283e-05,
199
+ "loss": 2.0858,
200
+ "step": 3200
201
+ },
202
+ {
203
+ "epoch": 0.31,
204
+ "learning_rate": 9.735118765969529e-05,
205
+ "loss": 2.0511,
206
+ "step": 3300
207
+ },
208
+ {
209
+ "epoch": 0.32,
210
+ "learning_rate": 9.725655342102773e-05,
211
+ "loss": 2.0592,
212
+ "step": 3400
213
+ },
214
+ {
215
+ "epoch": 0.33,
216
+ "learning_rate": 9.716191918236018e-05,
217
+ "loss": 2.0444,
218
+ "step": 3500
219
+ },
220
+ {
221
+ "epoch": 0.34,
222
+ "learning_rate": 9.706728494369263e-05,
223
+ "loss": 2.0513,
224
+ "step": 3600
225
+ },
226
+ {
227
+ "epoch": 0.35,
228
+ "learning_rate": 9.697265070502508e-05,
229
+ "loss": 2.0333,
230
+ "step": 3700
231
+ },
232
+ {
233
+ "epoch": 0.36,
234
+ "learning_rate": 9.687801646635754e-05,
235
+ "loss": 2.033,
236
+ "step": 3800
237
+ },
238
+ {
239
+ "epoch": 0.37,
240
+ "learning_rate": 9.678338222768998e-05,
241
+ "loss": 2.0503,
242
+ "step": 3900
243
+ },
244
+ {
245
+ "epoch": 0.38,
246
+ "learning_rate": 9.668874798902244e-05,
247
+ "loss": 2.0088,
248
+ "step": 4000
249
+ },
250
+ {
251
+ "epoch": 0.39,
252
+ "learning_rate": 9.659411375035488e-05,
253
+ "loss": 2.0314,
254
+ "step": 4100
255
+ },
256
+ {
257
+ "epoch": 0.4,
258
+ "learning_rate": 9.649947951168733e-05,
259
+ "loss": 2.02,
260
+ "step": 4200
261
+ },
262
+ {
263
+ "epoch": 0.41,
264
+ "learning_rate": 9.640484527301979e-05,
265
+ "loss": 2.0104,
266
+ "step": 4300
267
+ },
268
+ {
269
+ "epoch": 0.41,
270
+ "learning_rate": 9.631021103435224e-05,
271
+ "loss": 2.0084,
272
+ "step": 4400
273
+ },
274
+ {
275
+ "epoch": 0.42,
276
+ "learning_rate": 9.621557679568469e-05,
277
+ "loss": 1.9997,
278
+ "step": 4500
279
+ },
280
+ {
281
+ "epoch": 0.43,
282
+ "learning_rate": 9.612094255701713e-05,
283
+ "loss": 1.9974,
284
+ "step": 4600
285
+ },
286
+ {
287
+ "epoch": 0.44,
288
+ "learning_rate": 9.602630831834958e-05,
289
+ "loss": 2.0081,
290
+ "step": 4700
291
+ },
292
+ {
293
+ "epoch": 0.45,
294
+ "learning_rate": 9.593167407968203e-05,
295
+ "loss": 2.0234,
296
+ "step": 4800
297
+ },
298
+ {
299
+ "epoch": 0.46,
300
+ "learning_rate": 9.583703984101449e-05,
301
+ "loss": 2.0019,
302
+ "step": 4900
303
+ },
304
+ {
305
+ "epoch": 0.47,
306
+ "learning_rate": 9.574240560234694e-05,
307
+ "loss": 1.9859,
308
+ "step": 5000
309
+ },
310
+ {
311
+ "epoch": 0.48,
312
+ "learning_rate": 9.564777136367938e-05,
313
+ "loss": 1.9965,
314
+ "step": 5100
315
+ },
316
+ {
317
+ "epoch": 0.49,
318
+ "learning_rate": 9.555313712501184e-05,
319
+ "loss": 2.0021,
320
+ "step": 5200
321
+ },
322
+ {
323
+ "epoch": 0.5,
324
+ "learning_rate": 9.545850288634428e-05,
325
+ "loss": 1.9874,
326
+ "step": 5300
327
+ },
328
+ {
329
+ "epoch": 0.51,
330
+ "learning_rate": 9.536386864767674e-05,
331
+ "loss": 1.9913,
332
+ "step": 5400
333
+ },
334
+ {
335
+ "epoch": 0.52,
336
+ "learning_rate": 9.526923440900918e-05,
337
+ "loss": 1.9853,
338
+ "step": 5500
339
+ },
340
+ {
341
+ "epoch": 0.53,
342
+ "learning_rate": 9.517460017034163e-05,
343
+ "loss": 1.9815,
344
+ "step": 5600
345
+ },
346
+ {
347
+ "epoch": 0.54,
348
+ "learning_rate": 9.507996593167409e-05,
349
+ "loss": 1.9733,
350
+ "step": 5700
351
+ },
352
+ {
353
+ "epoch": 0.55,
354
+ "learning_rate": 9.498533169300653e-05,
355
+ "loss": 1.9825,
356
+ "step": 5800
357
+ },
358
+ {
359
+ "epoch": 0.56,
360
+ "learning_rate": 9.489069745433899e-05,
361
+ "loss": 1.9693,
362
+ "step": 5900
363
+ },
364
+ {
365
+ "epoch": 0.57,
366
+ "learning_rate": 9.479606321567143e-05,
367
+ "loss": 1.981,
368
+ "step": 6000
369
+ },
370
+ {
371
+ "epoch": 0.57,
372
+ "learning_rate": 9.470142897700388e-05,
373
+ "loss": 1.9738,
374
+ "step": 6100
375
+ },
376
+ {
377
+ "epoch": 0.58,
378
+ "learning_rate": 9.460679473833633e-05,
379
+ "loss": 1.9547,
380
+ "step": 6200
381
+ },
382
+ {
383
+ "epoch": 0.59,
384
+ "learning_rate": 9.451216049966879e-05,
385
+ "loss": 1.9699,
386
+ "step": 6300
387
+ },
388
+ {
389
+ "epoch": 0.6,
390
+ "learning_rate": 9.441752626100124e-05,
391
+ "loss": 1.9392,
392
+ "step": 6400
393
+ },
394
+ {
395
+ "epoch": 0.61,
396
+ "learning_rate": 9.43228920223337e-05,
397
+ "loss": 1.971,
398
+ "step": 6500
399
+ },
400
+ {
401
+ "epoch": 0.62,
402
+ "learning_rate": 9.422920412605281e-05,
403
+ "loss": 1.9549,
404
+ "step": 6600
405
+ },
406
+ {
407
+ "epoch": 0.63,
408
+ "learning_rate": 9.413456988738526e-05,
409
+ "loss": 1.96,
410
+ "step": 6700
411
+ },
412
+ {
413
+ "epoch": 0.64,
414
+ "learning_rate": 9.403993564871771e-05,
415
+ "loss": 1.9601,
416
+ "step": 6800
417
+ },
418
+ {
419
+ "epoch": 0.65,
420
+ "learning_rate": 9.394530141005015e-05,
421
+ "loss": 1.9535,
422
+ "step": 6900
423
+ },
424
+ {
425
+ "epoch": 0.66,
426
+ "learning_rate": 9.385066717138261e-05,
427
+ "loss": 1.9697,
428
+ "step": 7000
429
+ },
430
+ {
431
+ "epoch": 0.67,
432
+ "learning_rate": 9.375603293271507e-05,
433
+ "loss": 1.9383,
434
+ "step": 7100
435
+ },
436
+ {
437
+ "epoch": 0.68,
438
+ "learning_rate": 9.366139869404751e-05,
439
+ "loss": 1.9382,
440
+ "step": 7200
441
+ },
442
+ {
443
+ "epoch": 0.69,
444
+ "learning_rate": 9.356676445537996e-05,
445
+ "loss": 1.9561,
446
+ "step": 7300
447
+ },
448
+ {
449
+ "epoch": 0.7,
450
+ "learning_rate": 9.34721302167124e-05,
451
+ "loss": 1.9288,
452
+ "step": 7400
453
+ },
454
+ {
455
+ "epoch": 0.71,
456
+ "learning_rate": 9.337749597804486e-05,
457
+ "loss": 1.9319,
458
+ "step": 7500
459
+ },
460
+ {
461
+ "epoch": 0.72,
462
+ "learning_rate": 9.328286173937732e-05,
463
+ "loss": 1.944,
464
+ "step": 7600
465
+ },
466
+ {
467
+ "epoch": 0.73,
468
+ "learning_rate": 9.318822750070976e-05,
469
+ "loss": 1.9289,
470
+ "step": 7700
471
+ },
472
+ {
473
+ "epoch": 0.73,
474
+ "learning_rate": 9.30935932620422e-05,
475
+ "loss": 1.9314,
476
+ "step": 7800
477
+ },
478
+ {
479
+ "epoch": 0.74,
480
+ "learning_rate": 9.299895902337466e-05,
481
+ "loss": 1.9394,
482
+ "step": 7900
483
+ },
484
+ {
485
+ "epoch": 0.75,
486
+ "learning_rate": 9.290432478470711e-05,
487
+ "loss": 1.9239,
488
+ "step": 8000
489
+ },
490
+ {
491
+ "epoch": 0.76,
492
+ "learning_rate": 9.280969054603957e-05,
493
+ "loss": 1.9335,
494
+ "step": 8100
495
+ },
496
+ {
497
+ "epoch": 0.77,
498
+ "learning_rate": 9.271505630737202e-05,
499
+ "loss": 1.9334,
500
+ "step": 8200
501
+ },
502
+ {
503
+ "epoch": 0.78,
504
+ "learning_rate": 9.262042206870445e-05,
505
+ "loss": 1.9209,
506
+ "step": 8300
507
+ },
508
+ {
509
+ "epoch": 0.79,
510
+ "learning_rate": 9.252578783003691e-05,
511
+ "loss": 1.9132,
512
+ "step": 8400
513
+ },
514
+ {
515
+ "epoch": 0.8,
516
+ "learning_rate": 9.243115359136936e-05,
517
+ "loss": 1.9356,
518
+ "step": 8500
519
+ },
520
+ {
521
+ "epoch": 0.81,
522
+ "learning_rate": 9.233651935270181e-05,
523
+ "loss": 1.9078,
524
+ "step": 8600
525
+ },
526
+ {
527
+ "epoch": 0.82,
528
+ "learning_rate": 9.224188511403427e-05,
529
+ "loss": 1.9389,
530
+ "step": 8700
531
+ },
532
+ {
533
+ "epoch": 0.83,
534
+ "learning_rate": 9.21472508753667e-05,
535
+ "loss": 1.9438,
536
+ "step": 8800
537
+ },
538
+ {
539
+ "epoch": 0.84,
540
+ "learning_rate": 9.205261663669916e-05,
541
+ "loss": 1.939,
542
+ "step": 8900
543
+ },
544
+ {
545
+ "epoch": 0.85,
546
+ "learning_rate": 9.195798239803162e-05,
547
+ "loss": 1.9108,
548
+ "step": 9000
549
+ },
550
+ {
551
+ "epoch": 0.86,
552
+ "learning_rate": 9.186334815936406e-05,
553
+ "loss": 1.932,
554
+ "step": 9100
555
+ },
556
+ {
557
+ "epoch": 0.87,
558
+ "learning_rate": 9.176871392069652e-05,
559
+ "loss": 1.9119,
560
+ "step": 9200
561
+ },
562
+ {
563
+ "epoch": 0.88,
564
+ "learning_rate": 9.167407968202896e-05,
565
+ "loss": 1.9181,
566
+ "step": 9300
567
+ },
568
+ {
569
+ "epoch": 0.89,
570
+ "learning_rate": 9.157944544336141e-05,
571
+ "loss": 1.9145,
572
+ "step": 9400
573
+ },
574
+ {
575
+ "epoch": 0.89,
576
+ "learning_rate": 9.148481120469387e-05,
577
+ "loss": 1.9015,
578
+ "step": 9500
579
+ },
580
+ {
581
+ "epoch": 0.9,
582
+ "learning_rate": 9.139017696602631e-05,
583
+ "loss": 1.8919,
584
+ "step": 9600
585
+ },
586
+ {
587
+ "epoch": 0.91,
588
+ "learning_rate": 9.129554272735877e-05,
589
+ "loss": 1.9034,
590
+ "step": 9700
591
+ },
592
+ {
593
+ "epoch": 0.92,
594
+ "learning_rate": 9.120090848869121e-05,
595
+ "loss": 1.91,
596
+ "step": 9800
597
+ },
598
+ {
599
+ "epoch": 0.93,
600
+ "learning_rate": 9.110627425002366e-05,
601
+ "loss": 1.9128,
602
+ "step": 9900
603
+ },
604
+ {
605
+ "epoch": 0.94,
606
+ "learning_rate": 9.101164001135611e-05,
607
+ "loss": 1.902,
608
+ "step": 10000
609
+ },
610
+ {
611
+ "epoch": 0.95,
612
+ "learning_rate": 9.091700577268857e-05,
613
+ "loss": 1.8826,
614
+ "step": 10100
615
+ },
616
+ {
617
+ "epoch": 0.96,
618
+ "learning_rate": 9.082237153402102e-05,
619
+ "loss": 1.8919,
620
+ "step": 10200
621
+ },
622
+ {
623
+ "epoch": 0.97,
624
+ "learning_rate": 9.072773729535346e-05,
625
+ "loss": 1.9028,
626
+ "step": 10300
627
+ },
628
+ {
629
+ "epoch": 0.98,
630
+ "learning_rate": 9.06331030566859e-05,
631
+ "loss": 1.8857,
632
+ "step": 10400
633
+ },
634
+ {
635
+ "epoch": 0.99,
636
+ "learning_rate": 9.053846881801836e-05,
637
+ "loss": 1.9101,
638
+ "step": 10500
639
+ },
640
+ {
641
+ "epoch": 1.0,
642
+ "learning_rate": 9.044383457935082e-05,
643
+ "loss": 1.8769,
644
+ "step": 10600
645
+ },
646
+ {
647
+ "epoch": 1.0,
648
+ "eval_BERT_F": 0.8193,
649
+ "eval_BERT_P": 0.8208,
650
+ "eval_BERT_R": 0.8185,
651
+ "eval_R1": 0.5426,
652
+ "eval_R2": 0.321,
653
+ "eval_RL": 0.5097,
654
+ "eval_RLsum": 0.5186,
655
+ "eval_loss": 1.827642560005188,
656
+ "eval_runtime": 113.4614,
657
+ "eval_samples_per_second": 88.136,
658
+ "eval_steps_per_second": 1.384,
659
+ "step": 10617
660
+ },
661
+ {
662
+ "epoch": 1.01,
663
+ "learning_rate": 9.035109302545662e-05,
664
+ "loss": 1.7841,
665
+ "step": 10700
666
+ },
667
+ {
668
+ "epoch": 1.02,
669
+ "learning_rate": 9.025645878678906e-05,
670
+ "loss": 1.7802,
671
+ "step": 10800
672
+ },
673
+ {
674
+ "epoch": 1.03,
675
+ "learning_rate": 9.01618245481215e-05,
676
+ "loss": 1.7863,
677
+ "step": 10900
678
+ },
679
+ {
680
+ "epoch": 1.04,
681
+ "learning_rate": 9.006719030945396e-05,
682
+ "loss": 1.7706,
683
+ "step": 11000
684
+ },
685
+ {
686
+ "epoch": 1.05,
687
+ "learning_rate": 8.997255607078642e-05,
688
+ "loss": 1.7725,
689
+ "step": 11100
690
+ },
691
+ {
692
+ "epoch": 1.05,
693
+ "learning_rate": 8.987792183211887e-05,
694
+ "loss": 1.795,
695
+ "step": 11200
696
+ },
697
+ {
698
+ "epoch": 1.06,
699
+ "learning_rate": 8.978328759345131e-05,
700
+ "loss": 1.7763,
701
+ "step": 11300
702
+ },
703
+ {
704
+ "epoch": 1.07,
705
+ "learning_rate": 8.968865335478377e-05,
706
+ "loss": 1.7987,
707
+ "step": 11400
708
+ },
709
+ {
710
+ "epoch": 1.08,
711
+ "learning_rate": 8.959401911611621e-05,
712
+ "loss": 1.7734,
713
+ "step": 11500
714
+ },
715
+ {
716
+ "epoch": 1.09,
717
+ "learning_rate": 8.949938487744867e-05,
718
+ "loss": 1.786,
719
+ "step": 11600
720
+ },
721
+ {
722
+ "epoch": 1.1,
723
+ "learning_rate": 8.940475063878113e-05,
724
+ "loss": 1.7956,
725
+ "step": 11700
726
+ },
727
+ {
728
+ "epoch": 1.11,
729
+ "learning_rate": 8.931011640011356e-05,
730
+ "loss": 1.765,
731
+ "step": 11800
732
+ },
733
+ {
734
+ "epoch": 1.12,
735
+ "learning_rate": 8.921548216144602e-05,
736
+ "loss": 1.7879,
737
+ "step": 11900
738
+ },
739
+ {
740
+ "epoch": 1.13,
741
+ "learning_rate": 8.912084792277846e-05,
742
+ "loss": 1.7747,
743
+ "step": 12000
744
+ },
745
+ {
746
+ "epoch": 1.14,
747
+ "learning_rate": 8.902621368411092e-05,
748
+ "loss": 1.7789,
749
+ "step": 12100
750
+ },
751
+ {
752
+ "epoch": 1.15,
753
+ "learning_rate": 8.893157944544337e-05,
754
+ "loss": 1.7897,
755
+ "step": 12200
756
+ },
757
+ {
758
+ "epoch": 1.16,
759
+ "learning_rate": 8.88369452067758e-05,
760
+ "loss": 1.8051,
761
+ "step": 12300
762
+ },
763
+ {
764
+ "epoch": 1.17,
765
+ "learning_rate": 8.874231096810826e-05,
766
+ "loss": 1.7807,
767
+ "step": 12400
768
+ },
769
+ {
770
+ "epoch": 1.18,
771
+ "learning_rate": 8.864767672944072e-05,
772
+ "loss": 1.7866,
773
+ "step": 12500
774
+ },
775
+ {
776
+ "epoch": 1.19,
777
+ "learning_rate": 8.855304249077317e-05,
778
+ "loss": 1.7999,
779
+ "step": 12600
780
+ },
781
+ {
782
+ "epoch": 1.2,
783
+ "learning_rate": 8.845840825210562e-05,
784
+ "loss": 1.7748,
785
+ "step": 12700
786
+ },
787
+ {
788
+ "epoch": 1.21,
789
+ "learning_rate": 8.836377401343807e-05,
790
+ "loss": 1.7772,
791
+ "step": 12800
792
+ },
793
+ {
794
+ "epoch": 1.22,
795
+ "learning_rate": 8.826913977477051e-05,
796
+ "loss": 1.8056,
797
+ "step": 12900
798
+ },
799
+ {
800
+ "epoch": 1.22,
801
+ "learning_rate": 8.817450553610297e-05,
802
+ "loss": 1.7672,
803
+ "step": 13000
804
+ },
805
+ {
806
+ "epoch": 1.23,
807
+ "learning_rate": 8.807987129743541e-05,
808
+ "loss": 1.8047,
809
+ "step": 13100
810
+ },
811
+ {
812
+ "epoch": 1.24,
813
+ "learning_rate": 8.798523705876787e-05,
814
+ "loss": 1.7894,
815
+ "step": 13200
816
+ },
817
+ {
818
+ "epoch": 1.25,
819
+ "learning_rate": 8.789060282010032e-05,
820
+ "loss": 1.7757,
821
+ "step": 13300
822
+ },
823
+ {
824
+ "epoch": 1.26,
825
+ "learning_rate": 8.779596858143276e-05,
826
+ "loss": 1.7827,
827
+ "step": 13400
828
+ },
829
+ {
830
+ "epoch": 1.27,
831
+ "learning_rate": 8.770133434276522e-05,
832
+ "loss": 1.7911,
833
+ "step": 13500
834
+ },
835
+ {
836
+ "epoch": 1.28,
837
+ "learning_rate": 8.760670010409768e-05,
838
+ "loss": 1.8014,
839
+ "step": 13600
840
+ },
841
+ {
842
+ "epoch": 1.29,
843
+ "learning_rate": 8.751206586543012e-05,
844
+ "loss": 1.778,
845
+ "step": 13700
846
+ },
847
+ {
848
+ "epoch": 1.3,
849
+ "learning_rate": 8.741743162676256e-05,
850
+ "loss": 1.791,
851
+ "step": 13800
852
+ },
853
+ {
854
+ "epoch": 1.31,
855
+ "learning_rate": 8.732279738809501e-05,
856
+ "loss": 1.7901,
857
+ "step": 13900
858
+ },
859
+ {
860
+ "epoch": 1.32,
861
+ "learning_rate": 8.722816314942747e-05,
862
+ "loss": 1.7916,
863
+ "step": 14000
864
+ },
865
+ {
866
+ "epoch": 1.33,
867
+ "learning_rate": 8.713352891075992e-05,
868
+ "loss": 1.7922,
869
+ "step": 14100
870
+ },
871
+ {
872
+ "epoch": 1.34,
873
+ "learning_rate": 8.703889467209237e-05,
874
+ "loss": 1.7845,
875
+ "step": 14200
876
+ },
877
+ {
878
+ "epoch": 1.35,
879
+ "learning_rate": 8.694426043342481e-05,
880
+ "loss": 1.8001,
881
+ "step": 14300
882
+ },
883
+ {
884
+ "epoch": 1.36,
885
+ "learning_rate": 8.684962619475727e-05,
886
+ "loss": 1.785,
887
+ "step": 14400
888
+ },
889
+ {
890
+ "epoch": 1.37,
891
+ "learning_rate": 8.675499195608971e-05,
892
+ "loss": 1.802,
893
+ "step": 14500
894
+ },
895
+ {
896
+ "epoch": 1.38,
897
+ "learning_rate": 8.666035771742217e-05,
898
+ "loss": 1.7882,
899
+ "step": 14600
900
+ },
901
+ {
902
+ "epoch": 1.38,
903
+ "learning_rate": 8.656666982114129e-05,
904
+ "loss": 1.7933,
905
+ "step": 14700
906
+ },
907
+ {
908
+ "epoch": 1.39,
909
+ "learning_rate": 8.647203558247374e-05,
910
+ "loss": 1.7917,
911
+ "step": 14800
912
+ },
913
+ {
914
+ "epoch": 1.4,
915
+ "learning_rate": 8.63774013438062e-05,
916
+ "loss": 1.7825,
917
+ "step": 14900
918
+ },
919
+ {
920
+ "epoch": 1.41,
921
+ "learning_rate": 8.628276710513865e-05,
922
+ "loss": 1.7712,
923
+ "step": 15000
924
+ },
925
+ {
926
+ "epoch": 1.42,
927
+ "learning_rate": 8.618813286647109e-05,
928
+ "loss": 1.7599,
929
+ "step": 15100
930
+ },
931
+ {
932
+ "epoch": 1.43,
933
+ "learning_rate": 8.609349862780355e-05,
934
+ "loss": 1.7827,
935
+ "step": 15200
936
+ },
937
+ {
938
+ "epoch": 1.44,
939
+ "learning_rate": 8.599886438913599e-05,
940
+ "loss": 1.7681,
941
+ "step": 15300
942
+ },
943
+ {
944
+ "epoch": 1.45,
945
+ "learning_rate": 8.590423015046845e-05,
946
+ "loss": 1.7745,
947
+ "step": 15400
948
+ },
949
+ {
950
+ "epoch": 1.46,
951
+ "learning_rate": 8.58095959118009e-05,
952
+ "loss": 1.771,
953
+ "step": 15500
954
+ },
955
+ {
956
+ "epoch": 1.47,
957
+ "learning_rate": 8.571496167313334e-05,
958
+ "loss": 1.7908,
959
+ "step": 15600
960
+ },
961
+ {
962
+ "epoch": 1.48,
963
+ "learning_rate": 8.56203274344658e-05,
964
+ "loss": 1.7892,
965
+ "step": 15700
966
+ },
967
+ {
968
+ "epoch": 1.49,
969
+ "learning_rate": 8.552569319579824e-05,
970
+ "loss": 1.789,
971
+ "step": 15800
972
+ },
973
+ {
974
+ "epoch": 1.5,
975
+ "learning_rate": 8.54310589571307e-05,
976
+ "loss": 1.7901,
977
+ "step": 15900
978
+ },
979
+ {
980
+ "epoch": 1.51,
981
+ "learning_rate": 8.533642471846314e-05,
982
+ "loss": 1.7715,
983
+ "step": 16000
984
+ },
985
+ {
986
+ "epoch": 1.52,
987
+ "learning_rate": 8.524179047979559e-05,
988
+ "loss": 1.7816,
989
+ "step": 16100
990
+ },
991
+ {
992
+ "epoch": 1.53,
993
+ "learning_rate": 8.514715624112804e-05,
994
+ "loss": 1.7694,
995
+ "step": 16200
996
+ },
997
+ {
998
+ "epoch": 1.54,
999
+ "learning_rate": 8.50525220024605e-05,
1000
+ "loss": 1.7776,
1001
+ "step": 16300
1002
+ },
1003
+ {
1004
+ "epoch": 1.54,
1005
+ "learning_rate": 8.495788776379295e-05,
1006
+ "loss": 1.7934,
1007
+ "step": 16400
1008
+ },
1009
+ {
1010
+ "epoch": 1.55,
1011
+ "learning_rate": 8.48632535251254e-05,
1012
+ "loss": 1.755,
1013
+ "step": 16500
1014
+ },
1015
+ {
1016
+ "epoch": 1.56,
1017
+ "learning_rate": 8.476861928645783e-05,
1018
+ "loss": 1.7771,
1019
+ "step": 16600
1020
+ },
1021
+ {
1022
+ "epoch": 1.57,
1023
+ "learning_rate": 8.467398504779029e-05,
1024
+ "loss": 1.7807,
1025
+ "step": 16700
1026
+ },
1027
+ {
1028
+ "epoch": 1.58,
1029
+ "learning_rate": 8.457935080912275e-05,
1030
+ "loss": 1.7867,
1031
+ "step": 16800
1032
+ },
1033
+ {
1034
+ "epoch": 1.59,
1035
+ "learning_rate": 8.448566291284186e-05,
1036
+ "loss": 1.7787,
1037
+ "step": 16900
1038
+ },
1039
+ {
1040
+ "epoch": 1.6,
1041
+ "learning_rate": 8.439102867417432e-05,
1042
+ "loss": 1.7719,
1043
+ "step": 17000
1044
+ },
1045
+ {
1046
+ "epoch": 1.61,
1047
+ "learning_rate": 8.429639443550678e-05,
1048
+ "loss": 1.7861,
1049
+ "step": 17100
1050
+ },
1051
+ {
1052
+ "epoch": 1.62,
1053
+ "learning_rate": 8.420176019683922e-05,
1054
+ "loss": 1.7558,
1055
+ "step": 17200
1056
+ },
1057
+ {
1058
+ "epoch": 1.63,
1059
+ "learning_rate": 8.410712595817167e-05,
1060
+ "loss": 1.7663,
1061
+ "step": 17300
1062
+ },
1063
+ {
1064
+ "epoch": 1.64,
1065
+ "learning_rate": 8.401249171950411e-05,
1066
+ "loss": 1.7819,
1067
+ "step": 17400
1068
+ },
1069
+ {
1070
+ "epoch": 1.65,
1071
+ "learning_rate": 8.391785748083657e-05,
1072
+ "loss": 1.7636,
1073
+ "step": 17500
1074
+ },
1075
+ {
1076
+ "epoch": 1.66,
1077
+ "learning_rate": 8.382322324216903e-05,
1078
+ "loss": 1.7582,
1079
+ "step": 17600
1080
+ },
1081
+ {
1082
+ "epoch": 1.67,
1083
+ "learning_rate": 8.372858900350147e-05,
1084
+ "loss": 1.7835,
1085
+ "step": 17700
1086
+ },
1087
+ {
1088
+ "epoch": 1.68,
1089
+ "learning_rate": 8.363395476483392e-05,
1090
+ "loss": 1.7795,
1091
+ "step": 17800
1092
+ },
1093
+ {
1094
+ "epoch": 1.69,
1095
+ "learning_rate": 8.353932052616637e-05,
1096
+ "loss": 1.7823,
1097
+ "step": 17900
1098
+ },
1099
+ {
1100
+ "epoch": 1.7,
1101
+ "learning_rate": 8.344468628749882e-05,
1102
+ "loss": 1.7758,
1103
+ "step": 18000
1104
+ },
1105
+ {
1106
+ "epoch": 1.7,
1107
+ "learning_rate": 8.335005204883128e-05,
1108
+ "loss": 1.7764,
1109
+ "step": 18100
1110
+ },
1111
+ {
1112
+ "epoch": 1.71,
1113
+ "learning_rate": 8.325541781016372e-05,
1114
+ "loss": 1.7771,
1115
+ "step": 18200
1116
+ },
1117
+ {
1118
+ "epoch": 1.72,
1119
+ "learning_rate": 8.316078357149616e-05,
1120
+ "loss": 1.7762,
1121
+ "step": 18300
1122
+ },
1123
+ {
1124
+ "epoch": 1.73,
1125
+ "learning_rate": 8.306614933282862e-05,
1126
+ "loss": 1.7685,
1127
+ "step": 18400
1128
+ },
1129
+ {
1130
+ "epoch": 1.74,
1131
+ "learning_rate": 8.297151509416107e-05,
1132
+ "loss": 1.7567,
1133
+ "step": 18500
1134
+ },
1135
+ {
1136
+ "epoch": 1.75,
1137
+ "learning_rate": 8.287688085549352e-05,
1138
+ "loss": 1.7808,
1139
+ "step": 18600
1140
+ },
1141
+ {
1142
+ "epoch": 1.76,
1143
+ "learning_rate": 8.278224661682598e-05,
1144
+ "loss": 1.7615,
1145
+ "step": 18700
1146
+ },
1147
+ {
1148
+ "epoch": 1.77,
1149
+ "learning_rate": 8.268761237815843e-05,
1150
+ "loss": 1.7585,
1151
+ "step": 18800
1152
+ },
1153
+ {
1154
+ "epoch": 1.78,
1155
+ "learning_rate": 8.259297813949087e-05,
1156
+ "loss": 1.7915,
1157
+ "step": 18900
1158
+ },
1159
+ {
1160
+ "epoch": 1.79,
1161
+ "learning_rate": 8.249929024321e-05,
1162
+ "loss": 1.7715,
1163
+ "step": 19000
1164
+ },
1165
+ {
1166
+ "epoch": 1.8,
1167
+ "learning_rate": 8.240465600454244e-05,
1168
+ "loss": 1.7666,
1169
+ "step": 19100
1170
+ },
1171
+ {
1172
+ "epoch": 1.81,
1173
+ "learning_rate": 8.23100217658749e-05,
1174
+ "loss": 1.7823,
1175
+ "step": 19200
1176
+ },
1177
+ {
1178
+ "epoch": 1.82,
1179
+ "learning_rate": 8.221538752720734e-05,
1180
+ "loss": 1.7735,
1181
+ "step": 19300
1182
+ },
1183
+ {
1184
+ "epoch": 1.83,
1185
+ "learning_rate": 8.21207532885398e-05,
1186
+ "loss": 1.7639,
1187
+ "step": 19400
1188
+ },
1189
+ {
1190
+ "epoch": 1.84,
1191
+ "learning_rate": 8.202611904987225e-05,
1192
+ "loss": 1.7581,
1193
+ "step": 19500
1194
+ },
1195
+ {
1196
+ "epoch": 1.85,
1197
+ "learning_rate": 8.193148481120469e-05,
1198
+ "loss": 1.7402,
1199
+ "step": 19600
1200
+ },
1201
+ {
1202
+ "epoch": 1.86,
1203
+ "learning_rate": 8.183685057253715e-05,
1204
+ "loss": 1.768,
1205
+ "step": 19700
1206
+ },
1207
+ {
1208
+ "epoch": 1.86,
1209
+ "learning_rate": 8.17422163338696e-05,
1210
+ "loss": 1.7646,
1211
+ "step": 19800
1212
+ },
1213
+ {
1214
+ "epoch": 1.87,
1215
+ "learning_rate": 8.164758209520205e-05,
1216
+ "loss": 1.7536,
1217
+ "step": 19900
1218
+ },
1219
+ {
1220
+ "epoch": 1.88,
1221
+ "learning_rate": 8.155294785653451e-05,
1222
+ "loss": 1.7421,
1223
+ "step": 20000
1224
+ },
1225
+ {
1226
+ "epoch": 1.89,
1227
+ "learning_rate": 8.145831361786694e-05,
1228
+ "loss": 1.7618,
1229
+ "step": 20100
1230
+ },
1231
+ {
1232
+ "epoch": 1.9,
1233
+ "learning_rate": 8.13636793791994e-05,
1234
+ "loss": 1.778,
1235
+ "step": 20200
1236
+ },
1237
+ {
1238
+ "epoch": 1.91,
1239
+ "learning_rate": 8.126904514053185e-05,
1240
+ "loss": 1.7714,
1241
+ "step": 20300
1242
+ },
1243
+ {
1244
+ "epoch": 1.92,
1245
+ "learning_rate": 8.11744109018643e-05,
1246
+ "loss": 1.7724,
1247
+ "step": 20400
1248
+ },
1249
+ {
1250
+ "epoch": 1.93,
1251
+ "learning_rate": 8.107977666319676e-05,
1252
+ "loss": 1.7576,
1253
+ "step": 20500
1254
+ },
1255
+ {
1256
+ "epoch": 1.94,
1257
+ "learning_rate": 8.09851424245292e-05,
1258
+ "loss": 1.7465,
1259
+ "step": 20600
1260
+ },
1261
+ {
1262
+ "epoch": 1.95,
1263
+ "learning_rate": 8.089050818586164e-05,
1264
+ "loss": 1.7562,
1265
+ "step": 20700
1266
+ },
1267
+ {
1268
+ "epoch": 1.96,
1269
+ "learning_rate": 8.07958739471941e-05,
1270
+ "loss": 1.7439,
1271
+ "step": 20800
1272
+ },
1273
+ {
1274
+ "epoch": 1.97,
1275
+ "learning_rate": 8.070123970852655e-05,
1276
+ "loss": 1.7543,
1277
+ "step": 20900
1278
+ },
1279
+ {
1280
+ "epoch": 1.98,
1281
+ "learning_rate": 8.0606605469859e-05,
1282
+ "loss": 1.7432,
1283
+ "step": 21000
1284
+ },
1285
+ {
1286
+ "epoch": 1.99,
1287
+ "learning_rate": 8.051197123119145e-05,
1288
+ "loss": 1.7447,
1289
+ "step": 21100
1290
+ },
1291
+ {
1292
+ "epoch": 2.0,
1293
+ "learning_rate": 8.041828333491058e-05,
1294
+ "loss": 1.7422,
1295
+ "step": 21200
1296
+ },
1297
+ {
1298
+ "epoch": 2.0,
1299
+ "eval_BERT_F": 0.8142,
1300
+ "eval_BERT_P": 0.8118,
1301
+ "eval_BERT_R": 0.8175,
1302
+ "eval_R1": 0.5484,
1303
+ "eval_R2": 0.3326,
1304
+ "eval_RL": 0.5142,
1305
+ "eval_RLsum": 0.526,
1306
+ "eval_loss": 1.7479418516159058,
1307
+ "eval_runtime": 111.6247,
1308
+ "eval_samples_per_second": 89.586,
1309
+ "eval_steps_per_second": 1.406,
1310
+ "step": 21234
1311
+ },
1312
+ {
1313
+ "epoch": 2.01,
1314
+ "learning_rate": 8.032364909624302e-05,
1315
+ "loss": 1.6542,
1316
+ "step": 21300
1317
+ },
1318
+ {
1319
+ "epoch": 2.02,
1320
+ "learning_rate": 8.022901485757548e-05,
1321
+ "loss": 1.6407,
1322
+ "step": 21400
1323
+ },
1324
+ {
1325
+ "epoch": 2.03,
1326
+ "learning_rate": 8.013438061890792e-05,
1327
+ "loss": 1.6324,
1328
+ "step": 21500
1329
+ },
1330
+ {
1331
+ "epoch": 2.03,
1332
+ "learning_rate": 8.003974638024038e-05,
1333
+ "loss": 1.6509,
1334
+ "step": 21600
1335
+ },
1336
+ {
1337
+ "epoch": 2.04,
1338
+ "learning_rate": 7.994511214157282e-05,
1339
+ "loss": 1.6328,
1340
+ "step": 21700
1341
+ },
1342
+ {
1343
+ "epoch": 2.05,
1344
+ "learning_rate": 7.985047790290527e-05,
1345
+ "loss": 1.6416,
1346
+ "step": 21800
1347
+ },
1348
+ {
1349
+ "epoch": 2.06,
1350
+ "learning_rate": 7.975584366423773e-05,
1351
+ "loss": 1.6527,
1352
+ "step": 21900
1353
+ },
1354
+ {
1355
+ "epoch": 2.07,
1356
+ "learning_rate": 7.966120942557017e-05,
1357
+ "loss": 1.6546,
1358
+ "step": 22000
1359
+ },
1360
+ {
1361
+ "epoch": 2.08,
1362
+ "learning_rate": 7.956657518690263e-05,
1363
+ "loss": 1.6378,
1364
+ "step": 22100
1365
+ },
1366
+ {
1367
+ "epoch": 2.09,
1368
+ "learning_rate": 7.947194094823509e-05,
1369
+ "loss": 1.6581,
1370
+ "step": 22200
1371
+ },
1372
+ {
1373
+ "epoch": 2.1,
1374
+ "learning_rate": 7.937730670956753e-05,
1375
+ "loss": 1.6555,
1376
+ "step": 22300
1377
+ },
1378
+ {
1379
+ "epoch": 2.11,
1380
+ "learning_rate": 7.928267247089997e-05,
1381
+ "loss": 1.6567,
1382
+ "step": 22400
1383
+ },
1384
+ {
1385
+ "epoch": 2.12,
1386
+ "learning_rate": 7.918803823223243e-05,
1387
+ "loss": 1.654,
1388
+ "step": 22500
1389
+ },
1390
+ {
1391
+ "epoch": 2.13,
1392
+ "learning_rate": 7.909340399356488e-05,
1393
+ "loss": 1.6529,
1394
+ "step": 22600
1395
+ },
1396
+ {
1397
+ "epoch": 2.14,
1398
+ "learning_rate": 7.899876975489733e-05,
1399
+ "loss": 1.6621,
1400
+ "step": 22700
1401
+ },
1402
+ {
1403
+ "epoch": 2.15,
1404
+ "learning_rate": 7.890413551622978e-05,
1405
+ "loss": 1.668,
1406
+ "step": 22800
1407
+ },
1408
+ {
1409
+ "epoch": 2.16,
1410
+ "learning_rate": 7.880950127756222e-05,
1411
+ "loss": 1.6415,
1412
+ "step": 22900
1413
+ },
1414
+ {
1415
+ "epoch": 2.17,
1416
+ "learning_rate": 7.871486703889468e-05,
1417
+ "loss": 1.6564,
1418
+ "step": 23000
1419
+ },
1420
+ {
1421
+ "epoch": 2.18,
1422
+ "learning_rate": 7.862023280022712e-05,
1423
+ "loss": 1.653,
1424
+ "step": 23100
1425
+ },
1426
+ {
1427
+ "epoch": 2.19,
1428
+ "learning_rate": 7.852559856155958e-05,
1429
+ "loss": 1.6562,
1430
+ "step": 23200
1431
+ },
1432
+ {
1433
+ "epoch": 2.19,
1434
+ "learning_rate": 7.843096432289203e-05,
1435
+ "loss": 1.662,
1436
+ "step": 23300
1437
+ },
1438
+ {
1439
+ "epoch": 2.2,
1440
+ "learning_rate": 7.833633008422447e-05,
1441
+ "loss": 1.6417,
1442
+ "step": 23400
1443
+ },
1444
+ {
1445
+ "epoch": 2.21,
1446
+ "learning_rate": 7.824169584555693e-05,
1447
+ "loss": 1.6693,
1448
+ "step": 23500
1449
+ },
1450
+ {
1451
+ "epoch": 2.22,
1452
+ "learning_rate": 7.814800794927604e-05,
1453
+ "loss": 1.6486,
1454
+ "step": 23600
1455
+ },
1456
+ {
1457
+ "epoch": 2.23,
1458
+ "learning_rate": 7.80533737106085e-05,
1459
+ "loss": 1.6592,
1460
+ "step": 23700
1461
+ },
1462
+ {
1463
+ "epoch": 2.24,
1464
+ "learning_rate": 7.795873947194096e-05,
1465
+ "loss": 1.6648,
1466
+ "step": 23800
1467
+ },
1468
+ {
1469
+ "epoch": 2.25,
1470
+ "learning_rate": 7.78641052332734e-05,
1471
+ "loss": 1.6559,
1472
+ "step": 23900
1473
+ },
1474
+ {
1475
+ "epoch": 2.26,
1476
+ "learning_rate": 7.776947099460586e-05,
1477
+ "loss": 1.6726,
1478
+ "step": 24000
1479
+ },
1480
+ {
1481
+ "epoch": 2.27,
1482
+ "learning_rate": 7.76748367559383e-05,
1483
+ "loss": 1.6576,
1484
+ "step": 24100
1485
+ },
1486
+ {
1487
+ "epoch": 2.28,
1488
+ "learning_rate": 7.758020251727075e-05,
1489
+ "loss": 1.6655,
1490
+ "step": 24200
1491
+ },
1492
+ {
1493
+ "epoch": 2.29,
1494
+ "learning_rate": 7.74855682786032e-05,
1495
+ "loss": 1.6511,
1496
+ "step": 24300
1497
+ },
1498
+ {
1499
+ "epoch": 2.3,
1500
+ "learning_rate": 7.739093403993565e-05,
1501
+ "loss": 1.6536,
1502
+ "step": 24400
1503
+ },
1504
+ {
1505
+ "epoch": 2.31,
1506
+ "learning_rate": 7.729629980126811e-05,
1507
+ "loss": 1.6646,
1508
+ "step": 24500
1509
+ },
1510
+ {
1511
+ "epoch": 2.32,
1512
+ "learning_rate": 7.720166556260055e-05,
1513
+ "loss": 1.6448,
1514
+ "step": 24600
1515
+ },
1516
+ {
1517
+ "epoch": 2.33,
1518
+ "learning_rate": 7.7107031323933e-05,
1519
+ "loss": 1.6614,
1520
+ "step": 24700
1521
+ },
1522
+ {
1523
+ "epoch": 2.34,
1524
+ "learning_rate": 7.701334342765212e-05,
1525
+ "loss": 1.6532,
1526
+ "step": 24800
1527
+ },
1528
+ {
1529
+ "epoch": 2.35,
1530
+ "learning_rate": 7.691870918898458e-05,
1531
+ "loss": 1.671,
1532
+ "step": 24900
1533
+ },
1534
+ {
1535
+ "epoch": 2.35,
1536
+ "learning_rate": 7.682407495031703e-05,
1537
+ "loss": 1.6616,
1538
+ "step": 25000
1539
+ },
1540
+ {
1541
+ "epoch": 2.36,
1542
+ "learning_rate": 7.672944071164948e-05,
1543
+ "loss": 1.6526,
1544
+ "step": 25100
1545
+ },
1546
+ {
1547
+ "epoch": 2.37,
1548
+ "learning_rate": 7.663480647298193e-05,
1549
+ "loss": 1.665,
1550
+ "step": 25200
1551
+ },
1552
+ {
1553
+ "epoch": 2.38,
1554
+ "learning_rate": 7.654017223431437e-05,
1555
+ "loss": 1.6594,
1556
+ "step": 25300
1557
+ },
1558
+ {
1559
+ "epoch": 2.39,
1560
+ "learning_rate": 7.644553799564683e-05,
1561
+ "loss": 1.665,
1562
+ "step": 25400
1563
+ },
1564
+ {
1565
+ "epoch": 2.4,
1566
+ "learning_rate": 7.635090375697927e-05,
1567
+ "loss": 1.6581,
1568
+ "step": 25500
1569
+ },
1570
+ {
1571
+ "epoch": 2.41,
1572
+ "learning_rate": 7.625626951831173e-05,
1573
+ "loss": 1.6541,
1574
+ "step": 25600
1575
+ },
1576
+ {
1577
+ "epoch": 2.42,
1578
+ "learning_rate": 7.616163527964419e-05,
1579
+ "loss": 1.6697,
1580
+ "step": 25700
1581
+ },
1582
+ {
1583
+ "epoch": 2.43,
1584
+ "learning_rate": 7.606700104097663e-05,
1585
+ "loss": 1.6669,
1586
+ "step": 25800
1587
+ },
1588
+ {
1589
+ "epoch": 2.44,
1590
+ "learning_rate": 7.597236680230908e-05,
1591
+ "loss": 1.6664,
1592
+ "step": 25900
1593
+ },
1594
+ {
1595
+ "epoch": 2.45,
1596
+ "learning_rate": 7.587773256364152e-05,
1597
+ "loss": 1.6828,
1598
+ "step": 26000
1599
+ },
1600
+ {
1601
+ "epoch": 2.46,
1602
+ "learning_rate": 7.578309832497398e-05,
1603
+ "loss": 1.6643,
1604
+ "step": 26100
1605
+ },
1606
+ {
1607
+ "epoch": 2.47,
1608
+ "learning_rate": 7.568846408630644e-05,
1609
+ "loss": 1.6883,
1610
+ "step": 26200
1611
+ },
1612
+ {
1613
+ "epoch": 2.48,
1614
+ "learning_rate": 7.559382984763888e-05,
1615
+ "loss": 1.6593,
1616
+ "step": 26300
1617
+ },
1618
+ {
1619
+ "epoch": 2.49,
1620
+ "learning_rate": 7.549919560897133e-05,
1621
+ "loss": 1.6561,
1622
+ "step": 26400
1623
+ },
1624
+ {
1625
+ "epoch": 2.5,
1626
+ "learning_rate": 7.540456137030378e-05,
1627
+ "loss": 1.6558,
1628
+ "step": 26500
1629
+ },
1630
+ {
1631
+ "epoch": 2.51,
1632
+ "learning_rate": 7.530992713163623e-05,
1633
+ "loss": 1.6527,
1634
+ "step": 26600
1635
+ },
1636
+ {
1637
+ "epoch": 2.51,
1638
+ "learning_rate": 7.521529289296869e-05,
1639
+ "loss": 1.6563,
1640
+ "step": 26700
1641
+ },
1642
+ {
1643
+ "epoch": 2.52,
1644
+ "learning_rate": 7.512065865430113e-05,
1645
+ "loss": 1.6588,
1646
+ "step": 26800
1647
+ },
1648
+ {
1649
+ "epoch": 2.53,
1650
+ "learning_rate": 7.502602441563357e-05,
1651
+ "loss": 1.6759,
1652
+ "step": 26900
1653
+ },
1654
+ {
1655
+ "epoch": 2.54,
1656
+ "learning_rate": 7.493139017696603e-05,
1657
+ "loss": 1.6517,
1658
+ "step": 27000
1659
+ },
1660
+ {
1661
+ "epoch": 2.55,
1662
+ "learning_rate": 7.483675593829848e-05,
1663
+ "loss": 1.6702,
1664
+ "step": 27100
1665
+ },
1666
+ {
1667
+ "epoch": 2.56,
1668
+ "learning_rate": 7.474212169963093e-05,
1669
+ "loss": 1.6555,
1670
+ "step": 27200
1671
+ },
1672
+ {
1673
+ "epoch": 2.57,
1674
+ "learning_rate": 7.464748746096338e-05,
1675
+ "loss": 1.6657,
1676
+ "step": 27300
1677
+ },
1678
+ {
1679
+ "epoch": 2.58,
1680
+ "learning_rate": 7.455285322229582e-05,
1681
+ "loss": 1.6538,
1682
+ "step": 27400
1683
+ },
1684
+ {
1685
+ "epoch": 2.59,
1686
+ "learning_rate": 7.445821898362828e-05,
1687
+ "loss": 1.672,
1688
+ "step": 27500
1689
+ },
1690
+ {
1691
+ "epoch": 2.6,
1692
+ "learning_rate": 7.436358474496074e-05,
1693
+ "loss": 1.6703,
1694
+ "step": 27600
1695
+ },
1696
+ {
1697
+ "epoch": 2.61,
1698
+ "learning_rate": 7.426895050629318e-05,
1699
+ "loss": 1.6756,
1700
+ "step": 27700
1701
+ },
1702
+ {
1703
+ "epoch": 2.62,
1704
+ "learning_rate": 7.417431626762563e-05,
1705
+ "loss": 1.6694,
1706
+ "step": 27800
1707
+ },
1708
+ {
1709
+ "epoch": 2.63,
1710
+ "learning_rate": 7.407968202895807e-05,
1711
+ "loss": 1.6616,
1712
+ "step": 27900
1713
+ },
1714
+ {
1715
+ "epoch": 2.64,
1716
+ "learning_rate": 7.398504779029053e-05,
1717
+ "loss": 1.6559,
1718
+ "step": 28000
1719
+ },
1720
+ {
1721
+ "epoch": 2.65,
1722
+ "learning_rate": 7.389041355162299e-05,
1723
+ "loss": 1.6677,
1724
+ "step": 28100
1725
+ },
1726
+ {
1727
+ "epoch": 2.66,
1728
+ "learning_rate": 7.379577931295543e-05,
1729
+ "loss": 1.668,
1730
+ "step": 28200
1731
+ },
1732
+ {
1733
+ "epoch": 2.67,
1734
+ "learning_rate": 7.370114507428787e-05,
1735
+ "loss": 1.6657,
1736
+ "step": 28300
1737
+ },
1738
+ {
1739
+ "epoch": 2.67,
1740
+ "learning_rate": 7.360651083562033e-05,
1741
+ "loss": 1.6685,
1742
+ "step": 28400
1743
+ },
1744
+ {
1745
+ "epoch": 2.68,
1746
+ "learning_rate": 7.351187659695278e-05,
1747
+ "loss": 1.6651,
1748
+ "step": 28500
1749
+ },
1750
+ {
1751
+ "epoch": 2.69,
1752
+ "learning_rate": 7.341724235828523e-05,
1753
+ "loss": 1.6531,
1754
+ "step": 28600
1755
+ },
1756
+ {
1757
+ "epoch": 2.7,
1758
+ "learning_rate": 7.332260811961769e-05,
1759
+ "loss": 1.6755,
1760
+ "step": 28700
1761
+ },
1762
+ {
1763
+ "epoch": 2.71,
1764
+ "learning_rate": 7.322797388095014e-05,
1765
+ "loss": 1.6357,
1766
+ "step": 28800
1767
+ },
1768
+ {
1769
+ "epoch": 2.72,
1770
+ "learning_rate": 7.313333964228258e-05,
1771
+ "loss": 1.6642,
1772
+ "step": 28900
1773
+ },
1774
+ {
1775
+ "epoch": 2.73,
1776
+ "learning_rate": 7.303965174600171e-05,
1777
+ "loss": 1.6544,
1778
+ "step": 29000
1779
+ },
1780
+ {
1781
+ "epoch": 2.74,
1782
+ "learning_rate": 7.294501750733415e-05,
1783
+ "loss": 1.6557,
1784
+ "step": 29100
1785
+ },
1786
+ {
1787
+ "epoch": 2.75,
1788
+ "learning_rate": 7.285038326866661e-05,
1789
+ "loss": 1.6537,
1790
+ "step": 29200
1791
+ },
1792
+ {
1793
+ "epoch": 2.76,
1794
+ "learning_rate": 7.275574902999905e-05,
1795
+ "loss": 1.6657,
1796
+ "step": 29300
1797
+ },
1798
+ {
1799
+ "epoch": 2.77,
1800
+ "learning_rate": 7.266111479133151e-05,
1801
+ "loss": 1.6748,
1802
+ "step": 29400
1803
+ },
1804
+ {
1805
+ "epoch": 2.78,
1806
+ "learning_rate": 7.256648055266397e-05,
1807
+ "loss": 1.6815,
1808
+ "step": 29500
1809
+ },
1810
+ {
1811
+ "epoch": 2.79,
1812
+ "learning_rate": 7.24718463139964e-05,
1813
+ "loss": 1.6541,
1814
+ "step": 29600
1815
+ },
1816
+ {
1817
+ "epoch": 2.8,
1818
+ "learning_rate": 7.237721207532886e-05,
1819
+ "loss": 1.6779,
1820
+ "step": 29700
1821
+ },
1822
+ {
1823
+ "epoch": 2.81,
1824
+ "learning_rate": 7.22825778366613e-05,
1825
+ "loss": 1.6538,
1826
+ "step": 29800
1827
+ },
1828
+ {
1829
+ "epoch": 2.82,
1830
+ "learning_rate": 7.218794359799376e-05,
1831
+ "loss": 1.6755,
1832
+ "step": 29900
1833
+ },
1834
+ {
1835
+ "epoch": 2.83,
1836
+ "learning_rate": 7.209330935932622e-05,
1837
+ "loss": 1.6481,
1838
+ "step": 30000
1839
+ },
1840
+ {
1841
+ "epoch": 2.84,
1842
+ "learning_rate": 7.199867512065865e-05,
1843
+ "loss": 1.6627,
1844
+ "step": 30100
1845
+ },
1846
+ {
1847
+ "epoch": 2.84,
1848
+ "learning_rate": 7.19040408819911e-05,
1849
+ "loss": 1.6678,
1850
+ "step": 30200
1851
+ },
1852
+ {
1853
+ "epoch": 2.85,
1854
+ "learning_rate": 7.180940664332356e-05,
1855
+ "loss": 1.6592,
1856
+ "step": 30300
1857
+ },
1858
+ {
1859
+ "epoch": 2.86,
1860
+ "learning_rate": 7.171477240465601e-05,
1861
+ "loss": 1.6461,
1862
+ "step": 30400
1863
+ },
1864
+ {
1865
+ "epoch": 2.87,
1866
+ "learning_rate": 7.162013816598847e-05,
1867
+ "loss": 1.6494,
1868
+ "step": 30500
1869
+ },
1870
+ {
1871
+ "epoch": 2.88,
1872
+ "learning_rate": 7.15255039273209e-05,
1873
+ "loss": 1.6681,
1874
+ "step": 30600
1875
+ },
1876
+ {
1877
+ "epoch": 2.89,
1878
+ "learning_rate": 7.143086968865335e-05,
1879
+ "loss": 1.6576,
1880
+ "step": 30700
1881
+ },
1882
+ {
1883
+ "epoch": 2.9,
1884
+ "learning_rate": 7.133623544998581e-05,
1885
+ "loss": 1.659,
1886
+ "step": 30800
1887
+ },
1888
+ {
1889
+ "epoch": 2.91,
1890
+ "learning_rate": 7.124160121131826e-05,
1891
+ "loss": 1.6453,
1892
+ "step": 30900
1893
+ },
1894
+ {
1895
+ "epoch": 2.92,
1896
+ "learning_rate": 7.114696697265071e-05,
1897
+ "loss": 1.6817,
1898
+ "step": 31000
1899
+ },
1900
+ {
1901
+ "epoch": 2.93,
1902
+ "learning_rate": 7.105233273398316e-05,
1903
+ "loss": 1.6541,
1904
+ "step": 31100
1905
+ },
1906
+ {
1907
+ "epoch": 2.94,
1908
+ "learning_rate": 7.09576984953156e-05,
1909
+ "loss": 1.67,
1910
+ "step": 31200
1911
+ },
1912
+ {
1913
+ "epoch": 2.95,
1914
+ "learning_rate": 7.086306425664806e-05,
1915
+ "loss": 1.6668,
1916
+ "step": 31300
1917
+ },
1918
+ {
1919
+ "epoch": 2.96,
1920
+ "learning_rate": 7.076843001798052e-05,
1921
+ "loss": 1.6509,
1922
+ "step": 31400
1923
+ },
1924
+ {
1925
+ "epoch": 2.97,
1926
+ "learning_rate": 7.067474212169963e-05,
1927
+ "loss": 1.6596,
1928
+ "step": 31500
1929
+ },
1930
+ {
1931
+ "epoch": 2.98,
1932
+ "learning_rate": 7.058010788303209e-05,
1933
+ "loss": 1.6601,
1934
+ "step": 31600
1935
+ },
1936
+ {
1937
+ "epoch": 2.99,
1938
+ "learning_rate": 7.048547364436453e-05,
1939
+ "loss": 1.6559,
1940
+ "step": 31700
1941
+ },
1942
+ {
1943
+ "epoch": 3.0,
1944
+ "learning_rate": 7.039083940569698e-05,
1945
+ "loss": 1.6738,
1946
+ "step": 31800
1947
+ },
1948
+ {
1949
+ "epoch": 3.0,
1950
+ "eval_BERT_F": 0.7989,
1951
+ "eval_BERT_P": 0.786,
1952
+ "eval_BERT_R": 0.8157,
1953
+ "eval_R1": 0.5408,
1954
+ "eval_R2": 0.3319,
1955
+ "eval_RL": 0.5074,
1956
+ "eval_RLsum": 0.5192,
1957
+ "eval_loss": 1.7132327556610107,
1958
+ "eval_runtime": 115.0682,
1959
+ "eval_samples_per_second": 86.905,
1960
+ "eval_steps_per_second": 1.364,
1961
+ "step": 31851
1962
+ },
1963
+ {
1964
+ "epoch": 3.0,
1965
+ "learning_rate": 7.029715150941612e-05,
1966
+ "loss": 1.586,
1967
+ "step": 31900
1968
+ },
1969
+ {
1970
+ "epoch": 3.01,
1971
+ "learning_rate": 7.020251727074856e-05,
1972
+ "loss": 1.5578,
1973
+ "step": 32000
1974
+ },
1975
+ {
1976
+ "epoch": 3.02,
1977
+ "learning_rate": 7.010788303208101e-05,
1978
+ "loss": 1.5402,
1979
+ "step": 32100
1980
+ },
1981
+ {
1982
+ "epoch": 3.03,
1983
+ "learning_rate": 7.001324879341345e-05,
1984
+ "loss": 1.5364,
1985
+ "step": 32200
1986
+ },
1987
+ {
1988
+ "epoch": 3.04,
1989
+ "learning_rate": 6.991861455474591e-05,
1990
+ "loss": 1.546,
1991
+ "step": 32300
1992
+ },
1993
+ {
1994
+ "epoch": 3.05,
1995
+ "learning_rate": 6.982398031607837e-05,
1996
+ "loss": 1.5478,
1997
+ "step": 32400
1998
+ },
1999
+ {
2000
+ "epoch": 3.06,
2001
+ "learning_rate": 6.972934607741081e-05,
2002
+ "loss": 1.5341,
2003
+ "step": 32500
2004
+ },
2005
+ {
2006
+ "epoch": 3.07,
2007
+ "learning_rate": 6.963471183874326e-05,
2008
+ "loss": 1.5632,
2009
+ "step": 32600
2010
+ },
2011
+ {
2012
+ "epoch": 3.08,
2013
+ "learning_rate": 6.954007760007571e-05,
2014
+ "loss": 1.5556,
2015
+ "step": 32700
2016
+ },
2017
+ {
2018
+ "epoch": 3.09,
2019
+ "learning_rate": 6.944544336140816e-05,
2020
+ "loss": 1.5686,
2021
+ "step": 32800
2022
+ },
2023
+ {
2024
+ "epoch": 3.1,
2025
+ "learning_rate": 6.935080912274062e-05,
2026
+ "loss": 1.5604,
2027
+ "step": 32900
2028
+ },
2029
+ {
2030
+ "epoch": 3.11,
2031
+ "learning_rate": 6.925617488407307e-05,
2032
+ "loss": 1.5557,
2033
+ "step": 33000
2034
+ },
2035
+ {
2036
+ "epoch": 3.12,
2037
+ "learning_rate": 6.91615406454055e-05,
2038
+ "loss": 1.5707,
2039
+ "step": 33100
2040
+ },
2041
+ {
2042
+ "epoch": 3.13,
2043
+ "learning_rate": 6.906690640673796e-05,
2044
+ "loss": 1.5569,
2045
+ "step": 33200
2046
+ },
2047
+ {
2048
+ "epoch": 3.14,
2049
+ "learning_rate": 6.89722721680704e-05,
2050
+ "loss": 1.5547,
2051
+ "step": 33300
2052
+ },
2053
+ {
2054
+ "epoch": 3.15,
2055
+ "learning_rate": 6.887763792940286e-05,
2056
+ "loss": 1.5618,
2057
+ "step": 33400
2058
+ },
2059
+ {
2060
+ "epoch": 3.16,
2061
+ "learning_rate": 6.878300369073532e-05,
2062
+ "loss": 1.5604,
2063
+ "step": 33500
2064
+ },
2065
+ {
2066
+ "epoch": 3.16,
2067
+ "learning_rate": 6.868836945206775e-05,
2068
+ "loss": 1.5614,
2069
+ "step": 33600
2070
+ },
2071
+ {
2072
+ "epoch": 3.17,
2073
+ "learning_rate": 6.859373521340021e-05,
2074
+ "loss": 1.565,
2075
+ "step": 33700
2076
+ },
2077
+ {
2078
+ "epoch": 3.18,
2079
+ "learning_rate": 6.849910097473267e-05,
2080
+ "loss": 1.5578,
2081
+ "step": 33800
2082
+ },
2083
+ {
2084
+ "epoch": 3.19,
2085
+ "learning_rate": 6.840446673606511e-05,
2086
+ "loss": 1.5507,
2087
+ "step": 33900
2088
+ },
2089
+ {
2090
+ "epoch": 3.2,
2091
+ "learning_rate": 6.830983249739757e-05,
2092
+ "loss": 1.5678,
2093
+ "step": 34000
2094
+ },
2095
+ {
2096
+ "epoch": 3.21,
2097
+ "learning_rate": 6.821519825873e-05,
2098
+ "loss": 1.5675,
2099
+ "step": 34100
2100
+ },
2101
+ {
2102
+ "epoch": 3.22,
2103
+ "learning_rate": 6.812056402006246e-05,
2104
+ "loss": 1.5484,
2105
+ "step": 34200
2106
+ },
2107
+ {
2108
+ "epoch": 3.23,
2109
+ "learning_rate": 6.802592978139492e-05,
2110
+ "loss": 1.5619,
2111
+ "step": 34300
2112
+ },
2113
+ {
2114
+ "epoch": 3.24,
2115
+ "learning_rate": 6.793129554272736e-05,
2116
+ "loss": 1.5572,
2117
+ "step": 34400
2118
+ },
2119
+ {
2120
+ "epoch": 3.25,
2121
+ "learning_rate": 6.783666130405982e-05,
2122
+ "loss": 1.5702,
2123
+ "step": 34500
2124
+ },
2125
+ {
2126
+ "epoch": 3.26,
2127
+ "learning_rate": 6.774202706539226e-05,
2128
+ "loss": 1.5541,
2129
+ "step": 34600
2130
+ },
2131
+ {
2132
+ "epoch": 3.27,
2133
+ "learning_rate": 6.76473928267247e-05,
2134
+ "loss": 1.5692,
2135
+ "step": 34700
2136
+ },
2137
+ {
2138
+ "epoch": 3.28,
2139
+ "learning_rate": 6.755275858805716e-05,
2140
+ "loss": 1.5601,
2141
+ "step": 34800
2142
+ },
2143
+ {
2144
+ "epoch": 3.29,
2145
+ "learning_rate": 6.745812434938962e-05,
2146
+ "loss": 1.5777,
2147
+ "step": 34900
2148
+ },
2149
+ {
2150
+ "epoch": 3.3,
2151
+ "learning_rate": 6.736349011072207e-05,
2152
+ "loss": 1.5749,
2153
+ "step": 35000
2154
+ },
2155
+ {
2156
+ "epoch": 3.31,
2157
+ "learning_rate": 6.726885587205451e-05,
2158
+ "loss": 1.5715,
2159
+ "step": 35100
2160
+ },
2161
+ {
2162
+ "epoch": 3.32,
2163
+ "learning_rate": 6.717422163338695e-05,
2164
+ "loss": 1.5619,
2165
+ "step": 35200
2166
+ },
2167
+ {
2168
+ "epoch": 3.32,
2169
+ "learning_rate": 6.707958739471941e-05,
2170
+ "loss": 1.5763,
2171
+ "step": 35300
2172
+ },
2173
+ {
2174
+ "epoch": 3.33,
2175
+ "learning_rate": 6.698495315605187e-05,
2176
+ "loss": 1.5734,
2177
+ "step": 35400
2178
+ },
2179
+ {
2180
+ "epoch": 3.34,
2181
+ "learning_rate": 6.689031891738431e-05,
2182
+ "loss": 1.5816,
2183
+ "step": 35500
2184
+ },
2185
+ {
2186
+ "epoch": 3.35,
2187
+ "learning_rate": 6.679568467871676e-05,
2188
+ "loss": 1.5643,
2189
+ "step": 35600
2190
+ },
2191
+ {
2192
+ "epoch": 3.36,
2193
+ "learning_rate": 6.670105044004922e-05,
2194
+ "loss": 1.5756,
2195
+ "step": 35700
2196
+ },
2197
+ {
2198
+ "epoch": 3.37,
2199
+ "learning_rate": 6.660641620138166e-05,
2200
+ "loss": 1.5827,
2201
+ "step": 35800
2202
+ },
2203
+ {
2204
+ "epoch": 3.38,
2205
+ "learning_rate": 6.651178196271412e-05,
2206
+ "loss": 1.5756,
2207
+ "step": 35900
2208
+ },
2209
+ {
2210
+ "epoch": 3.39,
2211
+ "learning_rate": 6.641714772404656e-05,
2212
+ "loss": 1.5854,
2213
+ "step": 36000
2214
+ },
2215
+ {
2216
+ "epoch": 3.4,
2217
+ "learning_rate": 6.6322513485379e-05,
2218
+ "loss": 1.5782,
2219
+ "step": 36100
2220
+ },
2221
+ {
2222
+ "epoch": 3.41,
2223
+ "learning_rate": 6.622787924671146e-05,
2224
+ "loss": 1.612,
2225
+ "step": 36200
2226
+ },
2227
+ {
2228
+ "epoch": 3.42,
2229
+ "learning_rate": 6.613324500804391e-05,
2230
+ "loss": 1.5981,
2231
+ "step": 36300
2232
+ },
2233
+ {
2234
+ "epoch": 3.43,
2235
+ "learning_rate": 6.603861076937637e-05,
2236
+ "loss": 1.5692,
2237
+ "step": 36400
2238
+ },
2239
+ {
2240
+ "epoch": 3.44,
2241
+ "learning_rate": 6.594397653070882e-05,
2242
+ "loss": 1.58,
2243
+ "step": 36500
2244
+ },
2245
+ {
2246
+ "epoch": 3.45,
2247
+ "learning_rate": 6.585028863442794e-05,
2248
+ "loss": 1.5951,
2249
+ "step": 36600
2250
+ },
2251
+ {
2252
+ "epoch": 3.46,
2253
+ "learning_rate": 6.57556543957604e-05,
2254
+ "loss": 1.5739,
2255
+ "step": 36700
2256
+ },
2257
+ {
2258
+ "epoch": 3.47,
2259
+ "learning_rate": 6.566102015709284e-05,
2260
+ "loss": 1.573,
2261
+ "step": 36800
2262
+ },
2263
+ {
2264
+ "epoch": 3.48,
2265
+ "learning_rate": 6.556638591842528e-05,
2266
+ "loss": 1.5624,
2267
+ "step": 36900
2268
+ },
2269
+ {
2270
+ "epoch": 3.48,
2271
+ "learning_rate": 6.547175167975774e-05,
2272
+ "loss": 1.5822,
2273
+ "step": 37000
2274
+ },
2275
+ {
2276
+ "epoch": 3.49,
2277
+ "learning_rate": 6.537711744109019e-05,
2278
+ "loss": 1.5994,
2279
+ "step": 37100
2280
+ },
2281
+ {
2282
+ "epoch": 3.5,
2283
+ "learning_rate": 6.528248320242264e-05,
2284
+ "loss": 1.5761,
2285
+ "step": 37200
2286
+ },
2287
+ {
2288
+ "epoch": 3.51,
2289
+ "learning_rate": 6.518784896375509e-05,
2290
+ "loss": 1.5654,
2291
+ "step": 37300
2292
+ },
2293
+ {
2294
+ "epoch": 3.52,
2295
+ "learning_rate": 6.509321472508753e-05,
2296
+ "loss": 1.5765,
2297
+ "step": 37400
2298
+ },
2299
+ {
2300
+ "epoch": 3.53,
2301
+ "learning_rate": 6.499858048641999e-05,
2302
+ "loss": 1.5618,
2303
+ "step": 37500
2304
+ },
2305
+ {
2306
+ "epoch": 3.54,
2307
+ "learning_rate": 6.490394624775245e-05,
2308
+ "loss": 1.5787,
2309
+ "step": 37600
2310
+ },
2311
+ {
2312
+ "epoch": 3.55,
2313
+ "learning_rate": 6.480931200908489e-05,
2314
+ "loss": 1.568,
2315
+ "step": 37700
2316
+ },
2317
+ {
2318
+ "epoch": 3.56,
2319
+ "learning_rate": 6.471467777041734e-05,
2320
+ "loss": 1.5724,
2321
+ "step": 37800
2322
+ },
2323
+ {
2324
+ "epoch": 3.57,
2325
+ "learning_rate": 6.462004353174978e-05,
2326
+ "loss": 1.585,
2327
+ "step": 37900
2328
+ },
2329
+ {
2330
+ "epoch": 3.58,
2331
+ "learning_rate": 6.452540929308224e-05,
2332
+ "loss": 1.595,
2333
+ "step": 38000
2334
+ },
2335
+ {
2336
+ "epoch": 3.59,
2337
+ "learning_rate": 6.44307750544147e-05,
2338
+ "loss": 1.5883,
2339
+ "step": 38100
2340
+ },
2341
+ {
2342
+ "epoch": 3.6,
2343
+ "learning_rate": 6.433614081574714e-05,
2344
+ "loss": 1.5744,
2345
+ "step": 38200
2346
+ },
2347
+ {
2348
+ "epoch": 3.61,
2349
+ "learning_rate": 6.42415065770796e-05,
2350
+ "loss": 1.5909,
2351
+ "step": 38300
2352
+ },
2353
+ {
2354
+ "epoch": 3.62,
2355
+ "learning_rate": 6.414687233841204e-05,
2356
+ "loss": 1.6012,
2357
+ "step": 38400
2358
+ },
2359
+ {
2360
+ "epoch": 3.63,
2361
+ "learning_rate": 6.405223809974449e-05,
2362
+ "loss": 1.5996,
2363
+ "step": 38500
2364
+ },
2365
+ {
2366
+ "epoch": 3.64,
2367
+ "learning_rate": 6.395760386107694e-05,
2368
+ "loss": 1.5837,
2369
+ "step": 38600
2370
+ },
2371
+ {
2372
+ "epoch": 3.65,
2373
+ "learning_rate": 6.386296962240939e-05,
2374
+ "loss": 1.5804,
2375
+ "step": 38700
2376
+ },
2377
+ {
2378
+ "epoch": 3.65,
2379
+ "learning_rate": 6.376833538374185e-05,
2380
+ "loss": 1.5834,
2381
+ "step": 38800
2382
+ },
2383
+ {
2384
+ "epoch": 3.66,
2385
+ "learning_rate": 6.367370114507429e-05,
2386
+ "loss": 1.5674,
2387
+ "step": 38900
2388
+ },
2389
+ {
2390
+ "epoch": 3.67,
2391
+ "learning_rate": 6.357906690640673e-05,
2392
+ "loss": 1.5805,
2393
+ "step": 39000
2394
+ },
2395
+ {
2396
+ "epoch": 3.68,
2397
+ "learning_rate": 6.348443266773919e-05,
2398
+ "loss": 1.5676,
2399
+ "step": 39100
2400
+ },
2401
+ {
2402
+ "epoch": 3.69,
2403
+ "learning_rate": 6.338979842907165e-05,
2404
+ "loss": 1.5772,
2405
+ "step": 39200
2406
+ },
2407
+ {
2408
+ "epoch": 3.7,
2409
+ "learning_rate": 6.32951641904041e-05,
2410
+ "loss": 1.5804,
2411
+ "step": 39300
2412
+ },
2413
+ {
2414
+ "epoch": 3.71,
2415
+ "learning_rate": 6.320052995173654e-05,
2416
+ "loss": 1.5923,
2417
+ "step": 39400
2418
+ },
2419
+ {
2420
+ "epoch": 3.72,
2421
+ "learning_rate": 6.3105895713069e-05,
2422
+ "loss": 1.5926,
2423
+ "step": 39500
2424
+ },
2425
+ {
2426
+ "epoch": 3.73,
2427
+ "learning_rate": 6.301126147440144e-05,
2428
+ "loss": 1.5947,
2429
+ "step": 39600
2430
+ },
2431
+ {
2432
+ "epoch": 3.74,
2433
+ "learning_rate": 6.29166272357339e-05,
2434
+ "loss": 1.5765,
2435
+ "step": 39700
2436
+ },
2437
+ {
2438
+ "epoch": 3.75,
2439
+ "learning_rate": 6.282199299706634e-05,
2440
+ "loss": 1.58,
2441
+ "step": 39800
2442
+ },
2443
+ {
2444
+ "epoch": 3.76,
2445
+ "learning_rate": 6.272735875839879e-05,
2446
+ "loss": 1.5761,
2447
+ "step": 39900
2448
+ },
2449
+ {
2450
+ "epoch": 3.77,
2451
+ "learning_rate": 6.263367086211793e-05,
2452
+ "loss": 1.5837,
2453
+ "step": 40000
2454
+ },
2455
+ {
2456
+ "epoch": 3.78,
2457
+ "learning_rate": 6.253903662345036e-05,
2458
+ "loss": 1.5883,
2459
+ "step": 40100
2460
+ },
2461
+ {
2462
+ "epoch": 3.79,
2463
+ "learning_rate": 6.244440238478282e-05,
2464
+ "loss": 1.5805,
2465
+ "step": 40200
2466
+ },
2467
+ {
2468
+ "epoch": 3.8,
2469
+ "learning_rate": 6.234976814611527e-05,
2470
+ "loss": 1.5897,
2471
+ "step": 40300
2472
+ },
2473
+ {
2474
+ "epoch": 3.81,
2475
+ "learning_rate": 6.225513390744772e-05,
2476
+ "loss": 1.5846,
2477
+ "step": 40400
2478
+ },
2479
+ {
2480
+ "epoch": 3.81,
2481
+ "learning_rate": 6.216049966878018e-05,
2482
+ "loss": 1.5672,
2483
+ "step": 40500
2484
+ },
2485
+ {
2486
+ "epoch": 3.82,
2487
+ "learning_rate": 6.20658654301126e-05,
2488
+ "loss": 1.5917,
2489
+ "step": 40600
2490
+ },
2491
+ {
2492
+ "epoch": 3.83,
2493
+ "learning_rate": 6.197123119144506e-05,
2494
+ "loss": 1.5861,
2495
+ "step": 40700
2496
+ },
2497
+ {
2498
+ "epoch": 3.84,
2499
+ "learning_rate": 6.187659695277752e-05,
2500
+ "loss": 1.5601,
2501
+ "step": 40800
2502
+ },
2503
+ {
2504
+ "epoch": 3.85,
2505
+ "learning_rate": 6.178196271410997e-05,
2506
+ "loss": 1.5743,
2507
+ "step": 40900
2508
+ },
2509
+ {
2510
+ "epoch": 3.86,
2511
+ "learning_rate": 6.168732847544242e-05,
2512
+ "loss": 1.564,
2513
+ "step": 41000
2514
+ },
2515
+ {
2516
+ "epoch": 3.87,
2517
+ "learning_rate": 6.159269423677487e-05,
2518
+ "loss": 1.5932,
2519
+ "step": 41100
2520
+ },
2521
+ {
2522
+ "epoch": 3.88,
2523
+ "learning_rate": 6.149805999810731e-05,
2524
+ "loss": 1.5772,
2525
+ "step": 41200
2526
+ },
2527
+ {
2528
+ "epoch": 3.89,
2529
+ "learning_rate": 6.140342575943977e-05,
2530
+ "loss": 1.5885,
2531
+ "step": 41300
2532
+ },
2533
+ {
2534
+ "epoch": 3.9,
2535
+ "learning_rate": 6.130879152077221e-05,
2536
+ "loss": 1.5848,
2537
+ "step": 41400
2538
+ },
2539
+ {
2540
+ "epoch": 3.91,
2541
+ "learning_rate": 6.121415728210467e-05,
2542
+ "loss": 1.5827,
2543
+ "step": 41500
2544
+ },
2545
+ {
2546
+ "epoch": 3.92,
2547
+ "learning_rate": 6.111952304343712e-05,
2548
+ "loss": 1.5894,
2549
+ "step": 41600
2550
+ },
2551
+ {
2552
+ "epoch": 3.93,
2553
+ "learning_rate": 6.102488880476957e-05,
2554
+ "loss": 1.5962,
2555
+ "step": 41700
2556
+ },
2557
+ {
2558
+ "epoch": 3.94,
2559
+ "learning_rate": 6.093025456610202e-05,
2560
+ "loss": 1.5792,
2561
+ "step": 41800
2562
+ },
2563
+ {
2564
+ "epoch": 3.95,
2565
+ "learning_rate": 6.0835620327434476e-05,
2566
+ "loss": 1.5685,
2567
+ "step": 41900
2568
+ },
2569
+ {
2570
+ "epoch": 3.96,
2571
+ "learning_rate": 6.074098608876691e-05,
2572
+ "loss": 1.6087,
2573
+ "step": 42000
2574
+ },
2575
+ {
2576
+ "epoch": 3.97,
2577
+ "learning_rate": 6.064635185009937e-05,
2578
+ "loss": 1.5927,
2579
+ "step": 42100
2580
+ },
2581
+ {
2582
+ "epoch": 3.97,
2583
+ "learning_rate": 6.055171761143182e-05,
2584
+ "loss": 1.5878,
2585
+ "step": 42200
2586
+ },
2587
+ {
2588
+ "epoch": 3.98,
2589
+ "learning_rate": 6.0457083372764266e-05,
2590
+ "loss": 1.5851,
2591
+ "step": 42300
2592
+ },
2593
+ {
2594
+ "epoch": 3.99,
2595
+ "learning_rate": 6.03633954764834e-05,
2596
+ "loss": 1.5899,
2597
+ "step": 42400
2598
+ },
2599
+ {
2600
+ "epoch": 4.0,
2601
+ "eval_BERT_F": 0.7955,
2602
+ "eval_BERT_P": 0.7803,
2603
+ "eval_BERT_R": 0.8157,
2604
+ "eval_R1": 0.5558,
2605
+ "eval_R2": 0.3476,
2606
+ "eval_RL": 0.5232,
2607
+ "eval_RLsum": 0.5346,
2608
+ "eval_loss": 1.6942824125289917,
2609
+ "eval_runtime": 117.0149,
2610
+ "eval_samples_per_second": 85.459,
2611
+ "eval_steps_per_second": 1.342,
2612
+ "step": 42468
2613
+ },
2614
+ {
2615
+ "epoch": 4.0,
2616
+ "learning_rate": 6.026876123781584e-05,
2617
+ "loss": 1.5358,
2618
+ "step": 42500
2619
+ },
2620
+ {
2621
+ "epoch": 4.01,
2622
+ "learning_rate": 6.0174126999148296e-05,
2623
+ "loss": 1.4762,
2624
+ "step": 42600
2625
+ },
2626
+ {
2627
+ "epoch": 4.02,
2628
+ "learning_rate": 6.007949276048075e-05,
2629
+ "loss": 1.4729,
2630
+ "step": 42700
2631
+ },
2632
+ {
2633
+ "epoch": 4.03,
2634
+ "learning_rate": 5.998485852181319e-05,
2635
+ "loss": 1.4913,
2636
+ "step": 42800
2637
+ },
2638
+ {
2639
+ "epoch": 4.04,
2640
+ "learning_rate": 5.989022428314565e-05,
2641
+ "loss": 1.4785,
2642
+ "step": 42900
2643
+ },
2644
+ {
2645
+ "epoch": 4.05,
2646
+ "learning_rate": 5.97955900444781e-05,
2647
+ "loss": 1.4794,
2648
+ "step": 43000
2649
+ },
2650
+ {
2651
+ "epoch": 4.06,
2652
+ "learning_rate": 5.9700955805810544e-05,
2653
+ "loss": 1.5019,
2654
+ "step": 43100
2655
+ },
2656
+ {
2657
+ "epoch": 4.07,
2658
+ "learning_rate": 5.9606321567142995e-05,
2659
+ "loss": 1.4861,
2660
+ "step": 43200
2661
+ },
2662
+ {
2663
+ "epoch": 4.08,
2664
+ "learning_rate": 5.951168732847544e-05,
2665
+ "loss": 1.4744,
2666
+ "step": 43300
2667
+ },
2668
+ {
2669
+ "epoch": 4.09,
2670
+ "learning_rate": 5.94170530898079e-05,
2671
+ "loss": 1.4771,
2672
+ "step": 43400
2673
+ },
2674
+ {
2675
+ "epoch": 4.1,
2676
+ "learning_rate": 5.932241885114035e-05,
2677
+ "loss": 1.5019,
2678
+ "step": 43500
2679
+ },
2680
+ {
2681
+ "epoch": 4.11,
2682
+ "learning_rate": 5.922778461247279e-05,
2683
+ "loss": 1.4813,
2684
+ "step": 43600
2685
+ },
2686
+ {
2687
+ "epoch": 4.12,
2688
+ "learning_rate": 5.913315037380524e-05,
2689
+ "loss": 1.5028,
2690
+ "step": 43700
2691
+ },
2692
+ {
2693
+ "epoch": 4.13,
2694
+ "learning_rate": 5.90385161351377e-05,
2695
+ "loss": 1.4876,
2696
+ "step": 43800
2697
+ },
2698
+ {
2699
+ "epoch": 4.13,
2700
+ "learning_rate": 5.8943881896470145e-05,
2701
+ "loss": 1.4936,
2702
+ "step": 43900
2703
+ },
2704
+ {
2705
+ "epoch": 4.14,
2706
+ "learning_rate": 5.8849247657802596e-05,
2707
+ "loss": 1.4916,
2708
+ "step": 44000
2709
+ },
2710
+ {
2711
+ "epoch": 4.15,
2712
+ "learning_rate": 5.875461341913504e-05,
2713
+ "loss": 1.5021,
2714
+ "step": 44100
2715
+ },
2716
+ {
2717
+ "epoch": 4.16,
2718
+ "learning_rate": 5.86599791804675e-05,
2719
+ "loss": 1.4823,
2720
+ "step": 44200
2721
+ },
2722
+ {
2723
+ "epoch": 4.17,
2724
+ "learning_rate": 5.856534494179995e-05,
2725
+ "loss": 1.4936,
2726
+ "step": 44300
2727
+ },
2728
+ {
2729
+ "epoch": 4.18,
2730
+ "learning_rate": 5.847071070313239e-05,
2731
+ "loss": 1.4905,
2732
+ "step": 44400
2733
+ },
2734
+ {
2735
+ "epoch": 4.19,
2736
+ "learning_rate": 5.8376076464464844e-05,
2737
+ "loss": 1.4979,
2738
+ "step": 44500
2739
+ },
2740
+ {
2741
+ "epoch": 4.2,
2742
+ "learning_rate": 5.82814422257973e-05,
2743
+ "loss": 1.4852,
2744
+ "step": 44600
2745
+ },
2746
+ {
2747
+ "epoch": 4.21,
2748
+ "learning_rate": 5.8186807987129746e-05,
2749
+ "loss": 1.4889,
2750
+ "step": 44700
2751
+ },
2752
+ {
2753
+ "epoch": 4.22,
2754
+ "learning_rate": 5.80921737484622e-05,
2755
+ "loss": 1.5144,
2756
+ "step": 44800
2757
+ },
2758
+ {
2759
+ "epoch": 4.23,
2760
+ "learning_rate": 5.7997539509794655e-05,
2761
+ "loss": 1.49,
2762
+ "step": 44900
2763
+ },
2764
+ {
2765
+ "epoch": 4.24,
2766
+ "learning_rate": 5.790290527112709e-05,
2767
+ "loss": 1.5009,
2768
+ "step": 45000
2769
+ },
2770
+ {
2771
+ "epoch": 4.25,
2772
+ "learning_rate": 5.780921737484623e-05,
2773
+ "loss": 1.508,
2774
+ "step": 45100
2775
+ },
2776
+ {
2777
+ "epoch": 4.26,
2778
+ "learning_rate": 5.771458313617867e-05,
2779
+ "loss": 1.4978,
2780
+ "step": 45200
2781
+ },
2782
+ {
2783
+ "epoch": 4.27,
2784
+ "learning_rate": 5.761994889751112e-05,
2785
+ "loss": 1.4981,
2786
+ "step": 45300
2787
+ },
2788
+ {
2789
+ "epoch": 4.28,
2790
+ "learning_rate": 5.752531465884358e-05,
2791
+ "loss": 1.4941,
2792
+ "step": 45400
2793
+ },
2794
+ {
2795
+ "epoch": 4.29,
2796
+ "learning_rate": 5.743068042017602e-05,
2797
+ "loss": 1.4977,
2798
+ "step": 45500
2799
+ },
2800
+ {
2801
+ "epoch": 4.29,
2802
+ "learning_rate": 5.7336046181508475e-05,
2803
+ "loss": 1.4798,
2804
+ "step": 45600
2805
+ },
2806
+ {
2807
+ "epoch": 4.3,
2808
+ "learning_rate": 5.7241411942840926e-05,
2809
+ "loss": 1.5163,
2810
+ "step": 45700
2811
+ },
2812
+ {
2813
+ "epoch": 4.31,
2814
+ "learning_rate": 5.714677770417337e-05,
2815
+ "loss": 1.4993,
2816
+ "step": 45800
2817
+ },
2818
+ {
2819
+ "epoch": 4.32,
2820
+ "learning_rate": 5.705214346550583e-05,
2821
+ "loss": 1.5128,
2822
+ "step": 45900
2823
+ },
2824
+ {
2825
+ "epoch": 4.33,
2826
+ "learning_rate": 5.6957509226838265e-05,
2827
+ "loss": 1.5052,
2828
+ "step": 46000
2829
+ },
2830
+ {
2831
+ "epoch": 4.34,
2832
+ "learning_rate": 5.686287498817072e-05,
2833
+ "loss": 1.4949,
2834
+ "step": 46100
2835
+ },
2836
+ {
2837
+ "epoch": 4.35,
2838
+ "learning_rate": 5.6768240749503174e-05,
2839
+ "loss": 1.5059,
2840
+ "step": 46200
2841
+ },
2842
+ {
2843
+ "epoch": 4.36,
2844
+ "learning_rate": 5.667360651083562e-05,
2845
+ "loss": 1.5189,
2846
+ "step": 46300
2847
+ },
2848
+ {
2849
+ "epoch": 4.37,
2850
+ "learning_rate": 5.6578972272168076e-05,
2851
+ "loss": 1.5086,
2852
+ "step": 46400
2853
+ },
2854
+ {
2855
+ "epoch": 4.38,
2856
+ "learning_rate": 5.648433803350053e-05,
2857
+ "loss": 1.5097,
2858
+ "step": 46500
2859
+ },
2860
+ {
2861
+ "epoch": 4.39,
2862
+ "learning_rate": 5.638970379483297e-05,
2863
+ "loss": 1.5071,
2864
+ "step": 46600
2865
+ },
2866
+ {
2867
+ "epoch": 4.4,
2868
+ "learning_rate": 5.629506955616542e-05,
2869
+ "loss": 1.4937,
2870
+ "step": 46700
2871
+ },
2872
+ {
2873
+ "epoch": 4.41,
2874
+ "learning_rate": 5.6200435317497866e-05,
2875
+ "loss": 1.4931,
2876
+ "step": 46800
2877
+ },
2878
+ {
2879
+ "epoch": 4.42,
2880
+ "learning_rate": 5.6105801078830324e-05,
2881
+ "loss": 1.5177,
2882
+ "step": 46900
2883
+ },
2884
+ {
2885
+ "epoch": 4.43,
2886
+ "learning_rate": 5.6011166840162775e-05,
2887
+ "loss": 1.5191,
2888
+ "step": 47000
2889
+ },
2890
+ {
2891
+ "epoch": 4.44,
2892
+ "learning_rate": 5.5917478943881896e-05,
2893
+ "loss": 1.5005,
2894
+ "step": 47100
2895
+ },
2896
+ {
2897
+ "epoch": 4.45,
2898
+ "learning_rate": 5.582284470521435e-05,
2899
+ "loss": 1.4894,
2900
+ "step": 47200
2901
+ },
2902
+ {
2903
+ "epoch": 4.46,
2904
+ "learning_rate": 5.5728210466546805e-05,
2905
+ "loss": 1.4984,
2906
+ "step": 47300
2907
+ },
2908
+ {
2909
+ "epoch": 4.46,
2910
+ "learning_rate": 5.563357622787925e-05,
2911
+ "loss": 1.5259,
2912
+ "step": 47400
2913
+ },
2914
+ {
2915
+ "epoch": 4.47,
2916
+ "learning_rate": 5.55389419892117e-05,
2917
+ "loss": 1.5046,
2918
+ "step": 47500
2919
+ },
2920
+ {
2921
+ "epoch": 4.48,
2922
+ "learning_rate": 5.5444307750544144e-05,
2923
+ "loss": 1.5143,
2924
+ "step": 47600
2925
+ },
2926
+ {
2927
+ "epoch": 4.49,
2928
+ "learning_rate": 5.53496735118766e-05,
2929
+ "loss": 1.4972,
2930
+ "step": 47700
2931
+ },
2932
+ {
2933
+ "epoch": 4.5,
2934
+ "learning_rate": 5.525503927320905e-05,
2935
+ "loss": 1.5012,
2936
+ "step": 47800
2937
+ },
2938
+ {
2939
+ "epoch": 4.51,
2940
+ "learning_rate": 5.51604050345415e-05,
2941
+ "loss": 1.5013,
2942
+ "step": 47900
2943
+ },
2944
+ {
2945
+ "epoch": 4.52,
2946
+ "learning_rate": 5.506577079587395e-05,
2947
+ "loss": 1.5109,
2948
+ "step": 48000
2949
+ },
2950
+ {
2951
+ "epoch": 4.53,
2952
+ "learning_rate": 5.4971136557206406e-05,
2953
+ "loss": 1.5002,
2954
+ "step": 48100
2955
+ },
2956
+ {
2957
+ "epoch": 4.54,
2958
+ "learning_rate": 5.487650231853885e-05,
2959
+ "loss": 1.5198,
2960
+ "step": 48200
2961
+ },
2962
+ {
2963
+ "epoch": 4.55,
2964
+ "learning_rate": 5.47818680798713e-05,
2965
+ "loss": 1.5106,
2966
+ "step": 48300
2967
+ },
2968
+ {
2969
+ "epoch": 4.56,
2970
+ "learning_rate": 5.468723384120376e-05,
2971
+ "loss": 1.5139,
2972
+ "step": 48400
2973
+ },
2974
+ {
2975
+ "epoch": 4.57,
2976
+ "learning_rate": 5.4592599602536196e-05,
2977
+ "loss": 1.5,
2978
+ "step": 48500
2979
+ },
2980
+ {
2981
+ "epoch": 4.58,
2982
+ "learning_rate": 5.4497965363868654e-05,
2983
+ "loss": 1.5069,
2984
+ "step": 48600
2985
+ },
2986
+ {
2987
+ "epoch": 4.59,
2988
+ "learning_rate": 5.44033311252011e-05,
2989
+ "loss": 1.5165,
2990
+ "step": 48700
2991
+ },
2992
+ {
2993
+ "epoch": 4.6,
2994
+ "learning_rate": 5.430869688653355e-05,
2995
+ "loss": 1.5096,
2996
+ "step": 48800
2997
+ },
2998
+ {
2999
+ "epoch": 4.61,
3000
+ "learning_rate": 5.421406264786601e-05,
3001
+ "loss": 1.5159,
3002
+ "step": 48900
3003
+ },
3004
+ {
3005
+ "epoch": 4.62,
3006
+ "learning_rate": 5.4119428409198444e-05,
3007
+ "loss": 1.5074,
3008
+ "step": 49000
3009
+ },
3010
+ {
3011
+ "epoch": 4.62,
3012
+ "learning_rate": 5.40247941705309e-05,
3013
+ "loss": 1.5173,
3014
+ "step": 49100
3015
+ },
3016
+ {
3017
+ "epoch": 4.63,
3018
+ "learning_rate": 5.393015993186335e-05,
3019
+ "loss": 1.5039,
3020
+ "step": 49200
3021
+ },
3022
+ {
3023
+ "epoch": 4.64,
3024
+ "learning_rate": 5.38355256931958e-05,
3025
+ "loss": 1.5085,
3026
+ "step": 49300
3027
+ },
3028
+ {
3029
+ "epoch": 4.65,
3030
+ "learning_rate": 5.3740891454528255e-05,
3031
+ "loss": 1.5142,
3032
+ "step": 49400
3033
+ },
3034
+ {
3035
+ "epoch": 4.66,
3036
+ "learning_rate": 5.364625721586069e-05,
3037
+ "loss": 1.509,
3038
+ "step": 49500
3039
+ },
3040
+ {
3041
+ "epoch": 4.67,
3042
+ "learning_rate": 5.355162297719315e-05,
3043
+ "loss": 1.5064,
3044
+ "step": 49600
3045
+ },
3046
+ {
3047
+ "epoch": 4.68,
3048
+ "learning_rate": 5.34569887385256e-05,
3049
+ "loss": 1.5169,
3050
+ "step": 49700
3051
+ },
3052
+ {
3053
+ "epoch": 4.69,
3054
+ "learning_rate": 5.3362354499858045e-05,
3055
+ "loss": 1.523,
3056
+ "step": 49800
3057
+ },
3058
+ {
3059
+ "epoch": 4.7,
3060
+ "learning_rate": 5.32677202611905e-05,
3061
+ "loss": 1.5003,
3062
+ "step": 49900
3063
+ },
3064
+ {
3065
+ "epoch": 4.71,
3066
+ "learning_rate": 5.3173086022522954e-05,
3067
+ "loss": 1.4882,
3068
+ "step": 50000
3069
+ },
3070
+ {
3071
+ "epoch": 4.72,
3072
+ "learning_rate": 5.30784517838554e-05,
3073
+ "loss": 1.5001,
3074
+ "step": 50100
3075
+ },
3076
+ {
3077
+ "epoch": 4.73,
3078
+ "learning_rate": 5.298381754518785e-05,
3079
+ "loss": 1.5085,
3080
+ "step": 50200
3081
+ },
3082
+ {
3083
+ "epoch": 4.74,
3084
+ "learning_rate": 5.288918330652031e-05,
3085
+ "loss": 1.5144,
3086
+ "step": 50300
3087
+ },
3088
+ {
3089
+ "epoch": 4.75,
3090
+ "learning_rate": 5.279454906785275e-05,
3091
+ "loss": 1.515,
3092
+ "step": 50400
3093
+ },
3094
+ {
3095
+ "epoch": 4.76,
3096
+ "learning_rate": 5.26999148291852e-05,
3097
+ "loss": 1.5394,
3098
+ "step": 50500
3099
+ },
3100
+ {
3101
+ "epoch": 4.77,
3102
+ "learning_rate": 5.260528059051765e-05,
3103
+ "loss": 1.5241,
3104
+ "step": 50600
3105
+ },
3106
+ {
3107
+ "epoch": 4.78,
3108
+ "learning_rate": 5.2511592694236774e-05,
3109
+ "loss": 1.5094,
3110
+ "step": 50700
3111
+ },
3112
+ {
3113
+ "epoch": 4.78,
3114
+ "learning_rate": 5.241695845556923e-05,
3115
+ "loss": 1.5208,
3116
+ "step": 50800
3117
+ },
3118
+ {
3119
+ "epoch": 4.79,
3120
+ "learning_rate": 5.2322324216901676e-05,
3121
+ "loss": 1.5185,
3122
+ "step": 50900
3123
+ },
3124
+ {
3125
+ "epoch": 4.8,
3126
+ "learning_rate": 5.222768997823413e-05,
3127
+ "loss": 1.5055,
3128
+ "step": 51000
3129
+ },
3130
+ {
3131
+ "epoch": 4.81,
3132
+ "learning_rate": 5.2133055739566585e-05,
3133
+ "loss": 1.5276,
3134
+ "step": 51100
3135
+ },
3136
+ {
3137
+ "epoch": 4.82,
3138
+ "learning_rate": 5.203842150089903e-05,
3139
+ "loss": 1.5167,
3140
+ "step": 51200
3141
+ },
3142
+ {
3143
+ "epoch": 4.83,
3144
+ "learning_rate": 5.194378726223148e-05,
3145
+ "loss": 1.5258,
3146
+ "step": 51300
3147
+ },
3148
+ {
3149
+ "epoch": 4.84,
3150
+ "learning_rate": 5.1849153023563924e-05,
3151
+ "loss": 1.5197,
3152
+ "step": 51400
3153
+ },
3154
+ {
3155
+ "epoch": 4.85,
3156
+ "learning_rate": 5.1754518784896375e-05,
3157
+ "loss": 1.5246,
3158
+ "step": 51500
3159
+ },
3160
+ {
3161
+ "epoch": 4.86,
3162
+ "learning_rate": 5.165988454622883e-05,
3163
+ "loss": 1.517,
3164
+ "step": 51600
3165
+ },
3166
+ {
3167
+ "epoch": 4.87,
3168
+ "learning_rate": 5.156525030756128e-05,
3169
+ "loss": 1.5076,
3170
+ "step": 51700
3171
+ },
3172
+ {
3173
+ "epoch": 4.88,
3174
+ "learning_rate": 5.147061606889373e-05,
3175
+ "loss": 1.5058,
3176
+ "step": 51800
3177
+ },
3178
+ {
3179
+ "epoch": 4.89,
3180
+ "learning_rate": 5.1375981830226186e-05,
3181
+ "loss": 1.5357,
3182
+ "step": 51900
3183
+ },
3184
+ {
3185
+ "epoch": 4.9,
3186
+ "learning_rate": 5.1281347591558623e-05,
3187
+ "loss": 1.5151,
3188
+ "step": 52000
3189
+ },
3190
+ {
3191
+ "epoch": 4.91,
3192
+ "learning_rate": 5.118671335289108e-05,
3193
+ "loss": 1.5184,
3194
+ "step": 52100
3195
+ },
3196
+ {
3197
+ "epoch": 4.92,
3198
+ "learning_rate": 5.1092079114223525e-05,
3199
+ "loss": 1.5095,
3200
+ "step": 52200
3201
+ },
3202
+ {
3203
+ "epoch": 4.93,
3204
+ "learning_rate": 5.0997444875555976e-05,
3205
+ "loss": 1.5093,
3206
+ "step": 52300
3207
+ },
3208
+ {
3209
+ "epoch": 4.94,
3210
+ "learning_rate": 5.0902810636888434e-05,
3211
+ "loss": 1.52,
3212
+ "step": 52400
3213
+ },
3214
+ {
3215
+ "epoch": 4.94,
3216
+ "learning_rate": 5.080817639822087e-05,
3217
+ "loss": 1.5037,
3218
+ "step": 52500
3219
+ },
3220
+ {
3221
+ "epoch": 4.95,
3222
+ "learning_rate": 5.071354215955333e-05,
3223
+ "loss": 1.5203,
3224
+ "step": 52600
3225
+ },
3226
+ {
3227
+ "epoch": 4.96,
3228
+ "learning_rate": 5.061985426327246e-05,
3229
+ "loss": 1.5042,
3230
+ "step": 52700
3231
+ },
3232
+ {
3233
+ "epoch": 4.97,
3234
+ "learning_rate": 5.05252200246049e-05,
3235
+ "loss": 1.5167,
3236
+ "step": 52800
3237
+ },
3238
+ {
3239
+ "epoch": 4.98,
3240
+ "learning_rate": 5.043058578593736e-05,
3241
+ "loss": 1.512,
3242
+ "step": 52900
3243
+ },
3244
+ {
3245
+ "epoch": 4.99,
3246
+ "learning_rate": 5.0335951547269796e-05,
3247
+ "loss": 1.5099,
3248
+ "step": 53000
3249
+ },
3250
+ {
3251
+ "epoch": 5.0,
3252
+ "eval_BERT_F": 0.7784,
3253
+ "eval_BERT_P": 0.7522,
3254
+ "eval_BERT_R": 0.813,
3255
+ "eval_R1": 0.5194,
3256
+ "eval_R2": 0.3241,
3257
+ "eval_RL": 0.4874,
3258
+ "eval_RLsum": 0.4989,
3259
+ "eval_loss": 1.6852939128875732,
3260
+ "eval_runtime": 118.2076,
3261
+ "eval_samples_per_second": 84.597,
3262
+ "eval_steps_per_second": 1.328,
3263
+ "step": 53085
3264
+ },
3265
+ {
3266
+ "epoch": 5.0,
3267
+ "learning_rate": 5.024226365098893e-05,
3268
+ "loss": 1.4963,
3269
+ "step": 53100
3270
+ },
3271
+ {
3272
+ "epoch": 5.01,
3273
+ "learning_rate": 5.014762941232138e-05,
3274
+ "loss": 1.4269,
3275
+ "step": 53200
3276
+ },
3277
+ {
3278
+ "epoch": 5.02,
3279
+ "learning_rate": 5.0052995173653826e-05,
3280
+ "loss": 1.4098,
3281
+ "step": 53300
3282
+ },
3283
+ {
3284
+ "epoch": 5.03,
3285
+ "learning_rate": 4.9958360934986284e-05,
3286
+ "loss": 1.3929,
3287
+ "step": 53400
3288
+ },
3289
+ {
3290
+ "epoch": 5.04,
3291
+ "learning_rate": 4.986372669631873e-05,
3292
+ "loss": 1.4166,
3293
+ "step": 53500
3294
+ },
3295
+ {
3296
+ "epoch": 5.05,
3297
+ "learning_rate": 4.976909245765118e-05,
3298
+ "loss": 1.4201,
3299
+ "step": 53600
3300
+ },
3301
+ {
3302
+ "epoch": 5.06,
3303
+ "learning_rate": 4.967445821898363e-05,
3304
+ "loss": 1.443,
3305
+ "step": 53700
3306
+ },
3307
+ {
3308
+ "epoch": 5.07,
3309
+ "learning_rate": 4.957982398031608e-05,
3310
+ "loss": 1.4142,
3311
+ "step": 53800
3312
+ },
3313
+ {
3314
+ "epoch": 5.08,
3315
+ "learning_rate": 4.948518974164853e-05,
3316
+ "loss": 1.4207,
3317
+ "step": 53900
3318
+ },
3319
+ {
3320
+ "epoch": 5.09,
3321
+ "learning_rate": 4.9390555502980976e-05,
3322
+ "loss": 1.415,
3323
+ "step": 54000
3324
+ },
3325
+ {
3326
+ "epoch": 5.1,
3327
+ "learning_rate": 4.9295921264313434e-05,
3328
+ "loss": 1.4143,
3329
+ "step": 54100
3330
+ },
3331
+ {
3332
+ "epoch": 5.11,
3333
+ "learning_rate": 4.9201287025645885e-05,
3334
+ "loss": 1.4516,
3335
+ "step": 54200
3336
+ },
3337
+ {
3338
+ "epoch": 5.11,
3339
+ "learning_rate": 4.910665278697833e-05,
3340
+ "loss": 1.422,
3341
+ "step": 54300
3342
+ },
3343
+ {
3344
+ "epoch": 5.12,
3345
+ "learning_rate": 4.901201854831078e-05,
3346
+ "loss": 1.4181,
3347
+ "step": 54400
3348
+ },
3349
+ {
3350
+ "epoch": 5.13,
3351
+ "learning_rate": 4.891738430964323e-05,
3352
+ "loss": 1.4269,
3353
+ "step": 54500
3354
+ },
3355
+ {
3356
+ "epoch": 5.14,
3357
+ "learning_rate": 4.882275007097568e-05,
3358
+ "loss": 1.4393,
3359
+ "step": 54600
3360
+ },
3361
+ {
3362
+ "epoch": 5.15,
3363
+ "learning_rate": 4.872811583230813e-05,
3364
+ "loss": 1.4353,
3365
+ "step": 54700
3366
+ },
3367
+ {
3368
+ "epoch": 5.16,
3369
+ "learning_rate": 4.8633481593640584e-05,
3370
+ "loss": 1.4378,
3371
+ "step": 54800
3372
+ },
3373
+ {
3374
+ "epoch": 5.17,
3375
+ "learning_rate": 4.8538847354973035e-05,
3376
+ "loss": 1.4281,
3377
+ "step": 54900
3378
+ },
3379
+ {
3380
+ "epoch": 5.18,
3381
+ "learning_rate": 4.844421311630548e-05,
3382
+ "loss": 1.4364,
3383
+ "step": 55000
3384
+ },
3385
+ {
3386
+ "epoch": 5.19,
3387
+ "learning_rate": 4.834957887763793e-05,
3388
+ "loss": 1.4453,
3389
+ "step": 55100
3390
+ },
3391
+ {
3392
+ "epoch": 5.2,
3393
+ "learning_rate": 4.825494463897038e-05,
3394
+ "loss": 1.428,
3395
+ "step": 55200
3396
+ },
3397
+ {
3398
+ "epoch": 5.21,
3399
+ "learning_rate": 4.816031040030283e-05,
3400
+ "loss": 1.4388,
3401
+ "step": 55300
3402
+ },
3403
+ {
3404
+ "epoch": 5.22,
3405
+ "learning_rate": 4.806567616163528e-05,
3406
+ "loss": 1.4495,
3407
+ "step": 55400
3408
+ },
3409
+ {
3410
+ "epoch": 5.23,
3411
+ "learning_rate": 4.797104192296773e-05,
3412
+ "loss": 1.4288,
3413
+ "step": 55500
3414
+ },
3415
+ {
3416
+ "epoch": 5.24,
3417
+ "learning_rate": 4.7876407684300185e-05,
3418
+ "loss": 1.4165,
3419
+ "step": 55600
3420
+ },
3421
+ {
3422
+ "epoch": 5.25,
3423
+ "learning_rate": 4.7781773445632636e-05,
3424
+ "loss": 1.4271,
3425
+ "step": 55700
3426
+ },
3427
+ {
3428
+ "epoch": 5.26,
3429
+ "learning_rate": 4.768713920696508e-05,
3430
+ "loss": 1.4435,
3431
+ "step": 55800
3432
+ },
3433
+ {
3434
+ "epoch": 5.27,
3435
+ "learning_rate": 4.759250496829753e-05,
3436
+ "loss": 1.459,
3437
+ "step": 55900
3438
+ },
3439
+ {
3440
+ "epoch": 5.27,
3441
+ "learning_rate": 4.749787072962998e-05,
3442
+ "loss": 1.4589,
3443
+ "step": 56000
3444
+ },
3445
+ {
3446
+ "epoch": 5.28,
3447
+ "learning_rate": 4.740323649096243e-05,
3448
+ "loss": 1.4366,
3449
+ "step": 56100
3450
+ },
3451
+ {
3452
+ "epoch": 5.29,
3453
+ "learning_rate": 4.7308602252294884e-05,
3454
+ "loss": 1.4401,
3455
+ "step": 56200
3456
+ },
3457
+ {
3458
+ "epoch": 5.3,
3459
+ "learning_rate": 4.721396801362733e-05,
3460
+ "loss": 1.4442,
3461
+ "step": 56300
3462
+ },
3463
+ {
3464
+ "epoch": 5.31,
3465
+ "learning_rate": 4.7119333774959786e-05,
3466
+ "loss": 1.4414,
3467
+ "step": 56400
3468
+ },
3469
+ {
3470
+ "epoch": 5.32,
3471
+ "learning_rate": 4.702469953629223e-05,
3472
+ "loss": 1.4331,
3473
+ "step": 56500
3474
+ },
3475
+ {
3476
+ "epoch": 5.33,
3477
+ "learning_rate": 4.693006529762468e-05,
3478
+ "loss": 1.4445,
3479
+ "step": 56600
3480
+ },
3481
+ {
3482
+ "epoch": 5.34,
3483
+ "learning_rate": 4.683543105895713e-05,
3484
+ "loss": 1.4418,
3485
+ "step": 56700
3486
+ },
3487
+ {
3488
+ "epoch": 5.35,
3489
+ "learning_rate": 4.674079682028958e-05,
3490
+ "loss": 1.4575,
3491
+ "step": 56800
3492
+ },
3493
+ {
3494
+ "epoch": 5.36,
3495
+ "learning_rate": 4.6646162581622034e-05,
3496
+ "loss": 1.4358,
3497
+ "step": 56900
3498
+ },
3499
+ {
3500
+ "epoch": 5.37,
3501
+ "learning_rate": 4.655152834295448e-05,
3502
+ "loss": 1.447,
3503
+ "step": 57000
3504
+ },
3505
+ {
3506
+ "epoch": 5.38,
3507
+ "learning_rate": 4.6456894104286936e-05,
3508
+ "loss": 1.4421,
3509
+ "step": 57100
3510
+ },
3511
+ {
3512
+ "epoch": 5.39,
3513
+ "learning_rate": 4.636225986561939e-05,
3514
+ "loss": 1.448,
3515
+ "step": 57200
3516
+ },
3517
+ {
3518
+ "epoch": 5.4,
3519
+ "learning_rate": 4.626762562695183e-05,
3520
+ "loss": 1.4538,
3521
+ "step": 57300
3522
+ },
3523
+ {
3524
+ "epoch": 5.41,
3525
+ "learning_rate": 4.617299138828428e-05,
3526
+ "loss": 1.441,
3527
+ "step": 57400
3528
+ },
3529
+ {
3530
+ "epoch": 5.42,
3531
+ "learning_rate": 4.6078357149616733e-05,
3532
+ "loss": 1.4521,
3533
+ "step": 57500
3534
+ },
3535
+ {
3536
+ "epoch": 5.43,
3537
+ "learning_rate": 4.5983722910949184e-05,
3538
+ "loss": 1.4388,
3539
+ "step": 57600
3540
+ },
3541
+ {
3542
+ "epoch": 5.43,
3543
+ "learning_rate": 4.5889088672281635e-05,
3544
+ "loss": 1.4403,
3545
+ "step": 57700
3546
+ },
3547
+ {
3548
+ "epoch": 5.44,
3549
+ "learning_rate": 4.579445443361408e-05,
3550
+ "loss": 1.434,
3551
+ "step": 57800
3552
+ },
3553
+ {
3554
+ "epoch": 5.45,
3555
+ "learning_rate": 4.569982019494654e-05,
3556
+ "loss": 1.4614,
3557
+ "step": 57900
3558
+ },
3559
+ {
3560
+ "epoch": 5.46,
3561
+ "learning_rate": 4.560518595627898e-05,
3562
+ "loss": 1.4574,
3563
+ "step": 58000
3564
+ },
3565
+ {
3566
+ "epoch": 5.47,
3567
+ "learning_rate": 4.551055171761143e-05,
3568
+ "loss": 1.441,
3569
+ "step": 58100
3570
+ },
3571
+ {
3572
+ "epoch": 5.48,
3573
+ "learning_rate": 4.5415917478943884e-05,
3574
+ "loss": 1.4481,
3575
+ "step": 58200
3576
+ },
3577
+ {
3578
+ "epoch": 5.49,
3579
+ "learning_rate": 4.5321283240276335e-05,
3580
+ "loss": 1.4533,
3581
+ "step": 58300
3582
+ },
3583
+ {
3584
+ "epoch": 5.5,
3585
+ "learning_rate": 4.5226649001608786e-05,
3586
+ "loss": 1.4491,
3587
+ "step": 58400
3588
+ },
3589
+ {
3590
+ "epoch": 5.51,
3591
+ "learning_rate": 4.5132961105327906e-05,
3592
+ "loss": 1.4438,
3593
+ "step": 58500
3594
+ },
3595
+ {
3596
+ "epoch": 5.52,
3597
+ "learning_rate": 4.503832686666036e-05,
3598
+ "loss": 1.4657,
3599
+ "step": 58600
3600
+ },
3601
+ {
3602
+ "epoch": 5.53,
3603
+ "learning_rate": 4.494369262799281e-05,
3604
+ "loss": 1.4453,
3605
+ "step": 58700
3606
+ },
3607
+ {
3608
+ "epoch": 5.54,
3609
+ "learning_rate": 4.484905838932526e-05,
3610
+ "loss": 1.4498,
3611
+ "step": 58800
3612
+ },
3613
+ {
3614
+ "epoch": 5.55,
3615
+ "learning_rate": 4.475442415065771e-05,
3616
+ "loss": 1.4508,
3617
+ "step": 58900
3618
+ },
3619
+ {
3620
+ "epoch": 5.56,
3621
+ "learning_rate": 4.4659789911990154e-05,
3622
+ "loss": 1.4477,
3623
+ "step": 59000
3624
+ },
3625
+ {
3626
+ "epoch": 5.57,
3627
+ "learning_rate": 4.456515567332261e-05,
3628
+ "loss": 1.4328,
3629
+ "step": 59100
3630
+ },
3631
+ {
3632
+ "epoch": 5.58,
3633
+ "learning_rate": 4.447052143465506e-05,
3634
+ "loss": 1.4432,
3635
+ "step": 59200
3636
+ },
3637
+ {
3638
+ "epoch": 5.59,
3639
+ "learning_rate": 4.437588719598751e-05,
3640
+ "loss": 1.4596,
3641
+ "step": 59300
3642
+ },
3643
+ {
3644
+ "epoch": 5.59,
3645
+ "learning_rate": 4.4281252957319965e-05,
3646
+ "loss": 1.4558,
3647
+ "step": 59400
3648
+ },
3649
+ {
3650
+ "epoch": 5.6,
3651
+ "learning_rate": 4.418661871865241e-05,
3652
+ "loss": 1.4636,
3653
+ "step": 59500
3654
+ },
3655
+ {
3656
+ "epoch": 5.61,
3657
+ "learning_rate": 4.409198447998486e-05,
3658
+ "loss": 1.4446,
3659
+ "step": 59600
3660
+ },
3661
+ {
3662
+ "epoch": 5.62,
3663
+ "learning_rate": 4.399735024131731e-05,
3664
+ "loss": 1.4803,
3665
+ "step": 59700
3666
+ },
3667
+ {
3668
+ "epoch": 5.63,
3669
+ "learning_rate": 4.390271600264976e-05,
3670
+ "loss": 1.4664,
3671
+ "step": 59800
3672
+ },
3673
+ {
3674
+ "epoch": 5.64,
3675
+ "learning_rate": 4.3808081763982213e-05,
3676
+ "loss": 1.4365,
3677
+ "step": 59900
3678
+ },
3679
+ {
3680
+ "epoch": 5.65,
3681
+ "learning_rate": 4.371344752531466e-05,
3682
+ "loss": 1.4544,
3683
+ "step": 60000
3684
+ },
3685
+ {
3686
+ "epoch": 5.66,
3687
+ "learning_rate": 4.361881328664711e-05,
3688
+ "loss": 1.4629,
3689
+ "step": 60100
3690
+ },
3691
+ {
3692
+ "epoch": 5.67,
3693
+ "learning_rate": 4.3524179047979566e-05,
3694
+ "loss": 1.4479,
3695
+ "step": 60200
3696
+ },
3697
+ {
3698
+ "epoch": 5.68,
3699
+ "learning_rate": 4.342954480931201e-05,
3700
+ "loss": 1.4689,
3701
+ "step": 60300
3702
+ },
3703
+ {
3704
+ "epoch": 5.69,
3705
+ "learning_rate": 4.333491057064446e-05,
3706
+ "loss": 1.4503,
3707
+ "step": 60400
3708
+ },
3709
+ {
3710
+ "epoch": 5.7,
3711
+ "learning_rate": 4.3240276331976906e-05,
3712
+ "loss": 1.4436,
3713
+ "step": 60500
3714
+ },
3715
+ {
3716
+ "epoch": 5.71,
3717
+ "learning_rate": 4.3145642093309364e-05,
3718
+ "loss": 1.4532,
3719
+ "step": 60600
3720
+ },
3721
+ {
3722
+ "epoch": 5.72,
3723
+ "learning_rate": 4.3051007854641815e-05,
3724
+ "loss": 1.4579,
3725
+ "step": 60700
3726
+ },
3727
+ {
3728
+ "epoch": 5.73,
3729
+ "learning_rate": 4.295637361597426e-05,
3730
+ "loss": 1.4457,
3731
+ "step": 60800
3732
+ },
3733
+ {
3734
+ "epoch": 5.74,
3735
+ "learning_rate": 4.286173937730671e-05,
3736
+ "loss": 1.4546,
3737
+ "step": 60900
3738
+ },
3739
+ {
3740
+ "epoch": 5.75,
3741
+ "learning_rate": 4.276710513863916e-05,
3742
+ "loss": 1.4429,
3743
+ "step": 61000
3744
+ },
3745
+ {
3746
+ "epoch": 5.75,
3747
+ "learning_rate": 4.267247089997161e-05,
3748
+ "loss": 1.4435,
3749
+ "step": 61100
3750
+ },
3751
+ {
3752
+ "epoch": 5.76,
3753
+ "learning_rate": 4.257783666130406e-05,
3754
+ "loss": 1.4409,
3755
+ "step": 61200
3756
+ },
3757
+ {
3758
+ "epoch": 5.77,
3759
+ "learning_rate": 4.2483202422636514e-05,
3760
+ "loss": 1.4432,
3761
+ "step": 61300
3762
+ },
3763
+ {
3764
+ "epoch": 5.78,
3765
+ "learning_rate": 4.2388568183968965e-05,
3766
+ "loss": 1.442,
3767
+ "step": 61400
3768
+ },
3769
+ {
3770
+ "epoch": 5.79,
3771
+ "learning_rate": 4.229393394530141e-05,
3772
+ "loss": 1.4587,
3773
+ "step": 61500
3774
+ },
3775
+ {
3776
+ "epoch": 5.8,
3777
+ "learning_rate": 4.219929970663386e-05,
3778
+ "loss": 1.4599,
3779
+ "step": 61600
3780
+ },
3781
+ {
3782
+ "epoch": 5.81,
3783
+ "learning_rate": 4.210466546796632e-05,
3784
+ "loss": 1.4511,
3785
+ "step": 61700
3786
+ },
3787
+ {
3788
+ "epoch": 5.82,
3789
+ "learning_rate": 4.201003122929876e-05,
3790
+ "loss": 1.4434,
3791
+ "step": 61800
3792
+ },
3793
+ {
3794
+ "epoch": 5.83,
3795
+ "learning_rate": 4.191539699063121e-05,
3796
+ "loss": 1.4735,
3797
+ "step": 61900
3798
+ },
3799
+ {
3800
+ "epoch": 5.84,
3801
+ "learning_rate": 4.182076275196366e-05,
3802
+ "loss": 1.4462,
3803
+ "step": 62000
3804
+ },
3805
+ {
3806
+ "epoch": 5.85,
3807
+ "learning_rate": 4.1726128513296115e-05,
3808
+ "loss": 1.4536,
3809
+ "step": 62100
3810
+ },
3811
+ {
3812
+ "epoch": 5.86,
3813
+ "learning_rate": 4.1631494274628566e-05,
3814
+ "loss": 1.4692,
3815
+ "step": 62200
3816
+ },
3817
+ {
3818
+ "epoch": 5.87,
3819
+ "learning_rate": 4.153686003596101e-05,
3820
+ "loss": 1.468,
3821
+ "step": 62300
3822
+ },
3823
+ {
3824
+ "epoch": 5.88,
3825
+ "learning_rate": 4.144222579729346e-05,
3826
+ "loss": 1.4483,
3827
+ "step": 62400
3828
+ },
3829
+ {
3830
+ "epoch": 5.89,
3831
+ "learning_rate": 4.134853790101259e-05,
3832
+ "loss": 1.4462,
3833
+ "step": 62500
3834
+ },
3835
+ {
3836
+ "epoch": 5.9,
3837
+ "learning_rate": 4.125390366234504e-05,
3838
+ "loss": 1.4618,
3839
+ "step": 62600
3840
+ },
3841
+ {
3842
+ "epoch": 5.91,
3843
+ "learning_rate": 4.115926942367749e-05,
3844
+ "loss": 1.4566,
3845
+ "step": 62700
3846
+ },
3847
+ {
3848
+ "epoch": 5.92,
3849
+ "learning_rate": 4.1064635185009935e-05,
3850
+ "loss": 1.4602,
3851
+ "step": 62800
3852
+ },
3853
+ {
3854
+ "epoch": 5.92,
3855
+ "learning_rate": 4.097000094634239e-05,
3856
+ "loss": 1.4431,
3857
+ "step": 62900
3858
+ },
3859
+ {
3860
+ "epoch": 5.93,
3861
+ "learning_rate": 4.087536670767484e-05,
3862
+ "loss": 1.4559,
3863
+ "step": 63000
3864
+ },
3865
+ {
3866
+ "epoch": 5.94,
3867
+ "learning_rate": 4.078073246900729e-05,
3868
+ "loss": 1.4718,
3869
+ "step": 63100
3870
+ },
3871
+ {
3872
+ "epoch": 5.95,
3873
+ "learning_rate": 4.068609823033974e-05,
3874
+ "loss": 1.4524,
3875
+ "step": 63200
3876
+ },
3877
+ {
3878
+ "epoch": 5.96,
3879
+ "learning_rate": 4.059146399167219e-05,
3880
+ "loss": 1.4366,
3881
+ "step": 63300
3882
+ },
3883
+ {
3884
+ "epoch": 5.97,
3885
+ "learning_rate": 4.049682975300464e-05,
3886
+ "loss": 1.4507,
3887
+ "step": 63400
3888
+ },
3889
+ {
3890
+ "epoch": 5.98,
3891
+ "learning_rate": 4.0402195514337085e-05,
3892
+ "loss": 1.4674,
3893
+ "step": 63500
3894
+ },
3895
+ {
3896
+ "epoch": 5.99,
3897
+ "learning_rate": 4.0307561275669536e-05,
3898
+ "loss": 1.4687,
3899
+ "step": 63600
3900
+ },
3901
+ {
3902
+ "epoch": 6.0,
3903
+ "learning_rate": 4.0212927037001994e-05,
3904
+ "loss": 1.4647,
3905
+ "step": 63700
3906
+ },
3907
+ {
3908
+ "epoch": 6.0,
3909
+ "eval_BERT_F": 0.7857,
3910
+ "eval_BERT_P": 0.7629,
3911
+ "eval_BERT_R": 0.8159,
3912
+ "eval_R1": 0.5521,
3913
+ "eval_R2": 0.3479,
3914
+ "eval_RL": 0.5202,
3915
+ "eval_RLsum": 0.5309,
3916
+ "eval_loss": 1.6809284687042236,
3917
+ "eval_runtime": 117.0136,
3918
+ "eval_samples_per_second": 85.46,
3919
+ "eval_steps_per_second": 1.342,
3920
+ "step": 63702
3921
+ }
3922
+ ],
3923
+ "max_steps": 106170,
3924
+ "num_train_epochs": 10,
3925
+ "total_flos": 2.4856248172570214e+18,
3926
+ "trial_name": null,
3927
+ "trial_params": null
3928
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:de28c8d1cf723377a02d98ec974d51b80ec1d5db90b7d3b0fb880cc21726456a
3
+ size 3375
vocab.json ADDED
The diff for this file is too large to render. See raw diff