Hugging Face's logo

cstorm125
/
wangchanberta-base-att-spm-uncased-finetune-qa

Question Answering
Transformers
PyTorch
camembert
Model card Files
xet
Community
1
cstorm125 commited on
Commit
40a7850
·
1 Parent(s): 7318d0c

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +20 -29
README.md CHANGED
@@ -7,44 +7,35 @@ widget:
7
 
8
  Finetuning `wangchanberta-base-att-spm-uncased` with the training set of `iapp_wiki_qa_squad` and `thaiqa` (removed examples which have cosine similarity with validation and test examples over 0.8). Benchmarks shared on [wandb](https://wandb.ai/cstorm125/wangchanberta-qa) using validation and test sets of `iapp_wiki_qa_squad`.
9
 
10
- Trained with
11
  ```
12
  export WANDB_PROJECT=wangchanberta-qa
13
 
14
  export MODEL_NAME=wangchanberta-base-att-spm-uncased
15
- python train_question_answering_lm_finetuning.py \\\\
16
- --model_name $MODEL_NAME \\\\
17
- --dataset_name iapp_thaiqa \\\\
18
- --output_dir $MODEL_NAME-finetune-iapp_thaiqa-model \\\\
19
- --log_dir $MODEL_NAME-finetune-iapp_thaiqa-log \\\\
20
- --lowercase \\\\
21
- --pad_on_right \\\\
22
  --fp16
23
 
24
  export MODEL_NAME=xlm-roberta-base
25
- python train_question_answering_lm_finetuning.py \\\\
26
- --model_name $MODEL_NAME \\\\
27
- --dataset_name iapp_thaiqa \\\\
28
- --output_dir $MODEL_NAME-finetune-iapp_thaiqa-model \\\\
29
- --log_dir $MODEL_NAME-finetune-iapp_thaiqa-log \\\\
30
- --pad_on_right \\\\
31
  --fp16
32
 
33
  export MODEL_NAME=bert-base-multilingual-cased
34
- python train_question_answering_lm_finetuning.py \\\\
35
- --model_name $MODEL_NAME \\\\
36
- --dataset_name iapp_thaiqa \\\\
37
- --output_dir $MODEL_NAME-finetune-iapp_thaiqa-model \\\\
38
- --log_dir $MODEL_NAME-finetune-iapp_thaiqa-log \\\\
39
- --pad_on_right \\\\
40
- --fp16
41
-
42
- export MODEL_NAME=wangchanberta-base-wiki-spm
43
- python train_question_answering_lm_finetuning.py \\\\
44
- --model_name $MODEL_NAME \\\\
45
- --dataset_name iapp_thaiqa \\\\
46
- --output_dir $MODEL_NAME-finetune-iapp_thaiqa-model \\\\
47
- --log_dir $MODEL_NAME-finetune-iapp_thaiqa-log \\\\
48
- --pad_on_right \\\\
49
  --fp16
50
  ```
 
7
 
8
  Finetuning `wangchanberta-base-att-spm-uncased` with the training set of `iapp_wiki_qa_squad` and `thaiqa` (removed examples which have cosine similarity with validation and test examples over 0.8). Benchmarks shared on [wandb](https://wandb.ai/cstorm125/wangchanberta-qa) using validation and test sets of `iapp_wiki_qa_squad`.
9
 
10
+ Trained with [thai2transformers](https://github.com/vistec-AI/thai2transformers/blob/dev/scripts/downstream/train_question_answering_lm_finetuning.py).
11
  ```
12
  export WANDB_PROJECT=wangchanberta-qa
13
 
14
  export MODEL_NAME=wangchanberta-base-att-spm-uncased
15
+ python train_question_answering_lm_finetuning.py \\\\\\\\
16
+ --model_name $MODEL_NAME \\\\\\\\
17
+ --dataset_name iapp_thaiqa \\\\\\\\
18
+ --output_dir $MODEL_NAME-finetune-iapp_thaiqa-model \\\\\\\\
19
+ --log_dir $MODEL_NAME-finetune-iapp_thaiqa-log \\\\\\\\
20
+ --lowercase \\\\\\\\
21
+ --pad_on_right \\\\\\\\
22
  --fp16
23
 
24
  export MODEL_NAME=xlm-roberta-base
25
+ python train_question_answering_lm_finetuning.py \\\\\\\\
26
+ --model_name $MODEL_NAME \\\\\\\\
27
+ --dataset_name iapp_thaiqa \\\\\\\\
28
+ --output_dir $MODEL_NAME-finetune-iapp_thaiqa-model \\\\\\\\
29
+ --log_dir $MODEL_NAME-finetune-iapp_thaiqa-log \\\\\\\\
30
+ --pad_on_right \\\\\\\\
31
  --fp16
32
 
33
  export MODEL_NAME=bert-base-multilingual-cased
34
+ python train_question_answering_lm_finetuning.py \\\\\\\\
35
+ --model_name $MODEL_NAME \\\\\\\\
36
+ --dataset_name iapp_thaiqa \\\\\\\\
37
+ --output_dir $MODEL_NAME-finetune-iapp_thaiqa-model \\\\\\\\
38
+ --log_dir $MODEL_NAME-finetune-iapp_thaiqa-log \\\\\\\\
39
+ --pad_on_right \\\\\\\\
 
 
 
 
 
 
 
 
 
40
  --fp16
41
  ```