kauffinger
commited on
Commit
•
86bc393
1
Parent(s):
16ffb15
initial
Browse files- README.md +84 -0
- all_results.json +22 -0
- config.json +37 -0
- eval_results.json +16 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +16 -0
- train_results.json +9 -0
- trainer_state.json +463 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
README.md
ADDED
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
datasets:
|
7 |
+
- mnli
|
8 |
+
metrics:
|
9 |
+
- accuracy
|
10 |
+
model-index:
|
11 |
+
- name: glue
|
12 |
+
results:
|
13 |
+
- task:
|
14 |
+
name: Text Classification
|
15 |
+
type: text-classification
|
16 |
+
dataset:
|
17 |
+
name: GLUE MNLI
|
18 |
+
type: glue
|
19 |
+
args: mnli
|
20 |
+
metrics:
|
21 |
+
- name: Accuracy
|
22 |
+
type: accuracy
|
23 |
+
value: 0.834519934906428
|
24 |
+
---
|
25 |
+
|
26 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
27 |
+
should probably proofread and complete it, then remove this comment. -->
|
28 |
+
|
29 |
+
# mnli
|
30 |
+
|
31 |
+
This model is a fine-tuned version of [allenai/scibert_scivocab_uncased](https://huggingface.co/allenai/scibert_scivocab_uncased) on the GLUE MNLI dataset.
|
32 |
+
It achieves the following results on the evaluation set:
|
33 |
+
- Loss: 0.4917
|
34 |
+
- Accuracy: 0.8345
|
35 |
+
|
36 |
+
## Model description
|
37 |
+
|
38 |
+
This is the pretrained model presented in [SciBERT: A Pretrained Language Model for Scientific Text](https://www.aclweb.org/anthology/D19-1371/), which is a BERT model trained on scientific text, then finetuned on GLUE MNLI for zero-shot classification.
|
39 |
+
|
40 |
+
The training corpus was papers taken from [Semantic Scholar](https://www.semanticscholar.org). Corpus size is 1.14M papers, 3.1B tokens. We use the full text of the papers in training, not just abstracts.
|
41 |
+
|
42 |
+
SciBERT has its own wordpiece vocabulary (scivocab) that's built to best match the training corpus.
|
43 |
+
|
44 |
+
## Intended uses & limitations
|
45 |
+
|
46 |
+
Zero-shot classification of scientific texts. Note that this model is outperformed by multiple models and was uploaded for research purposes. For actually classifying scientific text, I recommend looking into [Deberta v3 Large tuned on MNLI](https://huggingface.co/navteca/nli-deberta-v3-large) which according to my benchmark on abstracts performs best at current date (7/10/22).
|
47 |
+
|
48 |
+
## Training and evaluation data
|
49 |
+
|
50 |
+
GLUE MNLI
|
51 |
+
|
52 |
+
## Training procedure
|
53 |
+
|
54 |
+
### Training hyperparameters
|
55 |
+
|
56 |
+
The following hyperparameters were used during training:
|
57 |
+
- learning_rate: 2e-05
|
58 |
+
- train_batch_size: 32
|
59 |
+
- eval_batch_size: 8
|
60 |
+
- seed: 42
|
61 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
62 |
+
- lr_scheduler_type: linear
|
63 |
+
- num_epochs: 3.0
|
64 |
+
- mixed_precision_training: Native AMP
|
65 |
+
|
66 |
+
### Framework versions
|
67 |
+
|
68 |
+
- Transformers 4.22.2
|
69 |
+
- Pytorch 1.11.0+cu113
|
70 |
+
- Datasets 2.5.1
|
71 |
+
- Tokenizers 0.12.1
|
72 |
+
|
73 |
+
|
74 |
+
If using these models, please cite the following paper:
|
75 |
+
```
|
76 |
+
@inproceedings{beltagy-etal-2019-scibert,
|
77 |
+
title = "SciBERT: A Pretrained Language Model for Scientific Text",
|
78 |
+
author = "Beltagy, Iz and Lo, Kyle and Cohan, Arman",
|
79 |
+
booktitle = "EMNLP",
|
80 |
+
year = "2019",
|
81 |
+
publisher = "Association for Computational Linguistics",
|
82 |
+
url = "https://www.aclweb.org/anthology/D19-1371"
|
83 |
+
}
|
84 |
+
```
|
all_results.json
ADDED
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"epoch_mm": 3.0,
|
4 |
+
"eval_accuracy": 0.823841059602649,
|
5 |
+
"eval_accuracy_mm": 0.834519934906428,
|
6 |
+
"eval_loss": 0.5203812718391418,
|
7 |
+
"eval_loss_mm": 0.4916685223579407,
|
8 |
+
"eval_runtime": 17.8999,
|
9 |
+
"eval_runtime_mm": 20.4415,
|
10 |
+
"eval_samples": 9815,
|
11 |
+
"eval_samples_mm": 9832,
|
12 |
+
"eval_samples_per_second": 548.327,
|
13 |
+
"eval_samples_per_second_mm": 480.983,
|
14 |
+
"eval_steps_per_second": 68.548,
|
15 |
+
"eval_steps_per_second_mm": 60.123,
|
16 |
+
"total_flos": 7.74938740264658e+16,
|
17 |
+
"train_loss": 0.42257067142803223,
|
18 |
+
"train_runtime": 4332.4547,
|
19 |
+
"train_samples": 392702,
|
20 |
+
"train_samples_per_second": 271.926,
|
21 |
+
"train_steps_per_second": 8.498
|
22 |
+
}
|
config.json
ADDED
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "allenai/scibert_scivocab_uncased",
|
3 |
+
"architectures": [
|
4 |
+
"BertForSequenceClassification"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"classifier_dropout": null,
|
8 |
+
"finetuning_task": "mnli",
|
9 |
+
"hidden_act": "gelu",
|
10 |
+
"hidden_dropout_prob": 0.1,
|
11 |
+
"hidden_size": 768,
|
12 |
+
"id2label": {
|
13 |
+
"0": "entailment",
|
14 |
+
"1": "neutral",
|
15 |
+
"2": "contradiction"
|
16 |
+
},
|
17 |
+
"initializer_range": 0.02,
|
18 |
+
"intermediate_size": 3072,
|
19 |
+
"label2id": {
|
20 |
+
"contradiction": 2,
|
21 |
+
"entailment": 0,
|
22 |
+
"neutral": 1
|
23 |
+
},
|
24 |
+
"layer_norm_eps": 1e-12,
|
25 |
+
"max_position_embeddings": 512,
|
26 |
+
"model_type": "bert",
|
27 |
+
"num_attention_heads": 12,
|
28 |
+
"num_hidden_layers": 12,
|
29 |
+
"pad_token_id": 0,
|
30 |
+
"position_embedding_type": "absolute",
|
31 |
+
"problem_type": "single_label_classification",
|
32 |
+
"torch_dtype": "float32",
|
33 |
+
"transformers_version": "4.22.2",
|
34 |
+
"type_vocab_size": 2,
|
35 |
+
"use_cache": true,
|
36 |
+
"vocab_size": 31090
|
37 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"epoch_mm": 3.0,
|
4 |
+
"eval_accuracy": 0.823841059602649,
|
5 |
+
"eval_accuracy_mm": 0.834519934906428,
|
6 |
+
"eval_loss": 0.5203812718391418,
|
7 |
+
"eval_loss_mm": 0.4916685223579407,
|
8 |
+
"eval_runtime": 17.8999,
|
9 |
+
"eval_runtime_mm": 20.4415,
|
10 |
+
"eval_samples": 9815,
|
11 |
+
"eval_samples_mm": 9832,
|
12 |
+
"eval_samples_per_second": 548.327,
|
13 |
+
"eval_samples_per_second_mm": 480.983,
|
14 |
+
"eval_steps_per_second": 68.548,
|
15 |
+
"eval_steps_per_second_mm": 60.123
|
16 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1e640c10f6e574204b0cfd703c8d80805269a934b6ddba02321e7096b33b17b8
|
3 |
+
size 439754093
|
special_tokens_map.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cls_token": "[CLS]",
|
3 |
+
"mask_token": "[MASK]",
|
4 |
+
"pad_token": "[PAD]",
|
5 |
+
"sep_token": "[SEP]",
|
6 |
+
"unk_token": "[UNK]"
|
7 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cls_token": "[CLS]",
|
3 |
+
"do_basic_tokenize": true,
|
4 |
+
"do_lower_case": true,
|
5 |
+
"mask_token": "[MASK]",
|
6 |
+
"name_or_path": "allenai/scibert_scivocab_uncased",
|
7 |
+
"never_split": null,
|
8 |
+
"pad_token": "[PAD]",
|
9 |
+
"sep_token": "[SEP]",
|
10 |
+
"model_max_length": 512,
|
11 |
+
"special_tokens_map_file": null,
|
12 |
+
"strip_accents": null,
|
13 |
+
"tokenize_chinese_chars": true,
|
14 |
+
"tokenizer_class": "BertTokenizer",
|
15 |
+
"unk_token": "[UNK]"
|
16 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"total_flos": 7.74938740264658e+16,
|
4 |
+
"train_loss": 0.42257067142803223,
|
5 |
+
"train_runtime": 4332.4547,
|
6 |
+
"train_samples": 392702,
|
7 |
+
"train_samples_per_second": 271.926,
|
8 |
+
"train_steps_per_second": 8.498
|
9 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,463 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 3.0,
|
5 |
+
"global_step": 36816,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.04,
|
12 |
+
"learning_rate": 1.9728378965667103e-05,
|
13 |
+
"loss": 0.8341,
|
14 |
+
"step": 500
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.08,
|
18 |
+
"learning_rate": 1.9456757931334205e-05,
|
19 |
+
"loss": 0.6968,
|
20 |
+
"step": 1000
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.12,
|
24 |
+
"learning_rate": 1.9185136897001307e-05,
|
25 |
+
"loss": 0.6557,
|
26 |
+
"step": 1500
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.16,
|
30 |
+
"learning_rate": 1.8913515862668405e-05,
|
31 |
+
"loss": 0.6154,
|
32 |
+
"step": 2000
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.2,
|
36 |
+
"learning_rate": 1.8642438070404173e-05,
|
37 |
+
"loss": 0.6167,
|
38 |
+
"step": 2500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 0.24,
|
42 |
+
"learning_rate": 1.8370817036071274e-05,
|
43 |
+
"loss": 0.6043,
|
44 |
+
"step": 3000
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.29,
|
48 |
+
"learning_rate": 1.8099196001738376e-05,
|
49 |
+
"loss": 0.5908,
|
50 |
+
"step": 3500
|
51 |
+
},
|
52 |
+
{
|
53 |
+
"epoch": 0.33,
|
54 |
+
"learning_rate": 1.7828118209474144e-05,
|
55 |
+
"loss": 0.5746,
|
56 |
+
"step": 4000
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.37,
|
60 |
+
"learning_rate": 1.7556497175141245e-05,
|
61 |
+
"loss": 0.5532,
|
62 |
+
"step": 4500
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 0.41,
|
66 |
+
"learning_rate": 1.7284876140808347e-05,
|
67 |
+
"loss": 0.5657,
|
68 |
+
"step": 5000
|
69 |
+
},
|
70 |
+
{
|
71 |
+
"epoch": 0.45,
|
72 |
+
"learning_rate": 1.7013255106475445e-05,
|
73 |
+
"loss": 0.5564,
|
74 |
+
"step": 5500
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 0.49,
|
78 |
+
"learning_rate": 1.6741634072142547e-05,
|
79 |
+
"loss": 0.5406,
|
80 |
+
"step": 6000
|
81 |
+
},
|
82 |
+
{
|
83 |
+
"epoch": 0.53,
|
84 |
+
"learning_rate": 1.647001303780965e-05,
|
85 |
+
"loss": 0.5518,
|
86 |
+
"step": 6500
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.57,
|
90 |
+
"learning_rate": 1.619839200347675e-05,
|
91 |
+
"loss": 0.546,
|
92 |
+
"step": 7000
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 0.61,
|
96 |
+
"learning_rate": 1.5927314211212518e-05,
|
97 |
+
"loss": 0.5373,
|
98 |
+
"step": 7500
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 0.65,
|
102 |
+
"learning_rate": 1.565569317687962e-05,
|
103 |
+
"loss": 0.5262,
|
104 |
+
"step": 8000
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 0.69,
|
108 |
+
"learning_rate": 1.538407214254672e-05,
|
109 |
+
"loss": 0.5285,
|
110 |
+
"step": 8500
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 0.73,
|
114 |
+
"learning_rate": 1.5112451108213821e-05,
|
115 |
+
"loss": 0.5331,
|
116 |
+
"step": 9000
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 0.77,
|
120 |
+
"learning_rate": 1.4840830073880923e-05,
|
121 |
+
"loss": 0.5171,
|
122 |
+
"step": 9500
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 0.81,
|
126 |
+
"learning_rate": 1.4569209039548025e-05,
|
127 |
+
"loss": 0.5122,
|
128 |
+
"step": 10000
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.86,
|
132 |
+
"learning_rate": 1.4297588005215127e-05,
|
133 |
+
"loss": 0.5187,
|
134 |
+
"step": 10500
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 0.9,
|
138 |
+
"learning_rate": 1.4025966970882225e-05,
|
139 |
+
"loss": 0.5141,
|
140 |
+
"step": 11000
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 0.94,
|
144 |
+
"learning_rate": 1.3754889178617994e-05,
|
145 |
+
"loss": 0.5068,
|
146 |
+
"step": 11500
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 0.98,
|
150 |
+
"learning_rate": 1.3483268144285096e-05,
|
151 |
+
"loss": 0.5089,
|
152 |
+
"step": 12000
|
153 |
+
},
|
154 |
+
{
|
155 |
+
"epoch": 1.02,
|
156 |
+
"learning_rate": 1.3211647109952194e-05,
|
157 |
+
"loss": 0.4605,
|
158 |
+
"step": 12500
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"epoch": 1.06,
|
162 |
+
"learning_rate": 1.2940026075619296e-05,
|
163 |
+
"loss": 0.4191,
|
164 |
+
"step": 13000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 1.1,
|
168 |
+
"learning_rate": 1.2668948283355065e-05,
|
169 |
+
"loss": 0.4084,
|
170 |
+
"step": 13500
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 1.14,
|
174 |
+
"learning_rate": 1.2397327249022167e-05,
|
175 |
+
"loss": 0.4139,
|
176 |
+
"step": 14000
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 1.18,
|
180 |
+
"learning_rate": 1.2126249456757933e-05,
|
181 |
+
"loss": 0.3967,
|
182 |
+
"step": 14500
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 1.22,
|
186 |
+
"learning_rate": 1.1854628422425034e-05,
|
187 |
+
"loss": 0.4078,
|
188 |
+
"step": 15000
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 1.26,
|
192 |
+
"learning_rate": 1.1583007388092136e-05,
|
193 |
+
"loss": 0.4031,
|
194 |
+
"step": 15500
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 1.3,
|
198 |
+
"learning_rate": 1.1311386353759234e-05,
|
199 |
+
"loss": 0.4051,
|
200 |
+
"step": 16000
|
201 |
+
},
|
202 |
+
{
|
203 |
+
"epoch": 1.34,
|
204 |
+
"learning_rate": 1.1039765319426336e-05,
|
205 |
+
"loss": 0.4085,
|
206 |
+
"step": 16500
|
207 |
+
},
|
208 |
+
{
|
209 |
+
"epoch": 1.39,
|
210 |
+
"learning_rate": 1.0768144285093438e-05,
|
211 |
+
"loss": 0.4107,
|
212 |
+
"step": 17000
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 1.43,
|
216 |
+
"learning_rate": 1.049652325076054e-05,
|
217 |
+
"loss": 0.3993,
|
218 |
+
"step": 17500
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 1.47,
|
222 |
+
"learning_rate": 1.0224902216427641e-05,
|
223 |
+
"loss": 0.4082,
|
224 |
+
"step": 18000
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 1.51,
|
228 |
+
"learning_rate": 9.953824424163409e-06,
|
229 |
+
"loss": 0.3987,
|
230 |
+
"step": 18500
|
231 |
+
},
|
232 |
+
{
|
233 |
+
"epoch": 1.55,
|
234 |
+
"learning_rate": 9.682203389830509e-06,
|
235 |
+
"loss": 0.4062,
|
236 |
+
"step": 19000
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 1.59,
|
240 |
+
"learning_rate": 9.41058235549761e-06,
|
241 |
+
"loss": 0.4048,
|
242 |
+
"step": 19500
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 1.63,
|
246 |
+
"learning_rate": 9.138961321164712e-06,
|
247 |
+
"loss": 0.408,
|
248 |
+
"step": 20000
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 1.67,
|
252 |
+
"learning_rate": 8.867883528900478e-06,
|
253 |
+
"loss": 0.4056,
|
254 |
+
"step": 20500
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 1.71,
|
258 |
+
"learning_rate": 8.59626249456758e-06,
|
259 |
+
"loss": 0.3995,
|
260 |
+
"step": 21000
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 1.75,
|
264 |
+
"learning_rate": 8.325184702303347e-06,
|
265 |
+
"loss": 0.4005,
|
266 |
+
"step": 21500
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 1.79,
|
270 |
+
"learning_rate": 8.053563667970449e-06,
|
271 |
+
"loss": 0.3974,
|
272 |
+
"step": 22000
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 1.83,
|
276 |
+
"learning_rate": 7.781942633637549e-06,
|
277 |
+
"loss": 0.3992,
|
278 |
+
"step": 22500
|
279 |
+
},
|
280 |
+
{
|
281 |
+
"epoch": 1.87,
|
282 |
+
"learning_rate": 7.510321599304651e-06,
|
283 |
+
"loss": 0.405,
|
284 |
+
"step": 23000
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 1.91,
|
288 |
+
"learning_rate": 7.238700564971752e-06,
|
289 |
+
"loss": 0.3901,
|
290 |
+
"step": 23500
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 1.96,
|
294 |
+
"learning_rate": 6.967079530638852e-06,
|
295 |
+
"loss": 0.3989,
|
296 |
+
"step": 24000
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 2.0,
|
300 |
+
"learning_rate": 6.695458496305954e-06,
|
301 |
+
"loss": 0.4007,
|
302 |
+
"step": 24500
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 2.04,
|
306 |
+
"learning_rate": 6.423837461973056e-06,
|
307 |
+
"loss": 0.301,
|
308 |
+
"step": 25000
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"epoch": 2.08,
|
312 |
+
"learning_rate": 6.152759669708823e-06,
|
313 |
+
"loss": 0.3016,
|
314 |
+
"step": 25500
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 2.12,
|
318 |
+
"learning_rate": 5.881138635375923e-06,
|
319 |
+
"loss": 0.3001,
|
320 |
+
"step": 26000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 2.16,
|
324 |
+
"learning_rate": 5.609517601043025e-06,
|
325 |
+
"loss": 0.2919,
|
326 |
+
"step": 26500
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 2.2,
|
330 |
+
"learning_rate": 5.337896566710127e-06,
|
331 |
+
"loss": 0.2992,
|
332 |
+
"step": 27000
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 2.24,
|
336 |
+
"learning_rate": 5.066818774445893e-06,
|
337 |
+
"loss": 0.2961,
|
338 |
+
"step": 27500
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 2.28,
|
342 |
+
"learning_rate": 4.79574098218166e-06,
|
343 |
+
"loss": 0.3074,
|
344 |
+
"step": 28000
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 2.32,
|
348 |
+
"learning_rate": 4.524119947848762e-06,
|
349 |
+
"loss": 0.2926,
|
350 |
+
"step": 28500
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 2.36,
|
354 |
+
"learning_rate": 4.252498913515863e-06,
|
355 |
+
"loss": 0.2947,
|
356 |
+
"step": 29000
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 2.4,
|
360 |
+
"learning_rate": 3.980877879182964e-06,
|
361 |
+
"loss": 0.299,
|
362 |
+
"step": 29500
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"epoch": 2.44,
|
366 |
+
"learning_rate": 3.7092568448500653e-06,
|
367 |
+
"loss": 0.3009,
|
368 |
+
"step": 30000
|
369 |
+
},
|
370 |
+
{
|
371 |
+
"epoch": 2.49,
|
372 |
+
"learning_rate": 3.437635810517167e-06,
|
373 |
+
"loss": 0.3013,
|
374 |
+
"step": 30500
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 2.53,
|
378 |
+
"learning_rate": 3.1665580182529336e-06,
|
379 |
+
"loss": 0.2908,
|
380 |
+
"step": 31000
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 2.57,
|
384 |
+
"learning_rate": 2.894936983920035e-06,
|
385 |
+
"loss": 0.2921,
|
386 |
+
"step": 31500
|
387 |
+
},
|
388 |
+
{
|
389 |
+
"epoch": 2.61,
|
390 |
+
"learning_rate": 2.6233159495871362e-06,
|
391 |
+
"loss": 0.295,
|
392 |
+
"step": 32000
|
393 |
+
},
|
394 |
+
{
|
395 |
+
"epoch": 2.65,
|
396 |
+
"learning_rate": 2.3516949152542375e-06,
|
397 |
+
"loss": 0.2833,
|
398 |
+
"step": 32500
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 2.69,
|
402 |
+
"learning_rate": 2.080073880921339e-06,
|
403 |
+
"loss": 0.3009,
|
404 |
+
"step": 33000
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 2.73,
|
408 |
+
"learning_rate": 1.80845284658844e-06,
|
409 |
+
"loss": 0.2795,
|
410 |
+
"step": 33500
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 2.77,
|
414 |
+
"learning_rate": 1.5368318122555412e-06,
|
415 |
+
"loss": 0.2861,
|
416 |
+
"step": 34000
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 2.81,
|
420 |
+
"learning_rate": 1.2652107779226425e-06,
|
421 |
+
"loss": 0.2991,
|
422 |
+
"step": 34500
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 2.85,
|
426 |
+
"learning_rate": 9.941329856584096e-07,
|
427 |
+
"loss": 0.2836,
|
428 |
+
"step": 35000
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 2.89,
|
432 |
+
"learning_rate": 7.225119513255108e-07,
|
433 |
+
"loss": 0.2925,
|
434 |
+
"step": 35500
|
435 |
+
},
|
436 |
+
{
|
437 |
+
"epoch": 2.93,
|
438 |
+
"learning_rate": 4.5089091699261194e-07,
|
439 |
+
"loss": 0.2927,
|
440 |
+
"step": 36000
|
441 |
+
},
|
442 |
+
{
|
443 |
+
"epoch": 2.97,
|
444 |
+
"learning_rate": 1.7926988265971318e-07,
|
445 |
+
"loss": 0.2921,
|
446 |
+
"step": 36500
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 3.0,
|
450 |
+
"step": 36816,
|
451 |
+
"total_flos": 7.74938740264658e+16,
|
452 |
+
"train_loss": 0.42257067142803223,
|
453 |
+
"train_runtime": 4332.4547,
|
454 |
+
"train_samples_per_second": 271.926,
|
455 |
+
"train_steps_per_second": 8.498
|
456 |
+
}
|
457 |
+
],
|
458 |
+
"max_steps": 36816,
|
459 |
+
"num_train_epochs": 3,
|
460 |
+
"total_flos": 7.74938740264658e+16,
|
461 |
+
"trial_name": null,
|
462 |
+
"trial_params": null
|
463 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5f118d321b6ab58292c298c3ea2c75a3211f3fb0b13bb9da8c24b3fa647f22be
|
3 |
+
size 3439
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|