Skip to content

Commit

Permalink
Merge pull request #68 from boun-tabi-LMG/patch-anonymize
Browse files Browse the repository at this point in the history 
Anonymize config files and scripts
  • Loading branch information
@zeynepyirmibes
zeynepyirmibes authored Feb 10, 2024
2 parents f09d8a9 + db0d240 commit 8700994
Show file tree
Hide file tree
Showing 39 changed files with 84 additions and 86 deletions.
2 changes: 1 addition & 1 deletion experiments/conf/class_prod_reviews.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -12,5 +12,5 @@ training_params:
per_device_eval_batch_size: 32 # for ultramarine
# Conditional generation mode: classification/cg/tr_product_reviews
# Classification mode: classification/tc/tr_product_reviews
output_dir: /stratch/bounllm/finetuned-models/ul2tr/classification/cg/tr_product_reviews
output_dir: /finetuned-models/turna/classification/cg/tr_product_reviews

2 changes: 1 addition & 1 deletion experiments/conf/class_ttc4900.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,6 @@ training_params:
per_device_eval_batch_size: 2 # for ultramarine
# Conditional generation mode: classification/cg/ttc4900
# Classification mode: classification/tc/ttc4900
output_dir: /stratch/bounllm/finetuned-models/ul2tr/classification/cg/ttc4900
output_dir: /finetuned-models/turna/classification/cg/ttc4900


4 changes: 2 additions & 2 deletions experiments/conf/class_tweet_sentiment.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@ defaults:

dataset_name: 17bintweet_sentiment
task: classification
dataset_loc: "/stratch/bounllm/supervised/17bintweet_sentiment/"
dataset_loc: "/supervised/17bintweet_sentiment/"
max_input_length: 160
max_target_length: 4
num_labels: 3
Expand All @@ -13,5 +13,5 @@ training_params:
per_device_eval_batch_size: 32 # for ultramarine
# Conditional generation mode: classification/cg/17bintweet_sentiment
# Classification mode: classification/tc/17bintweet_sentiment
output_dir: /stratch/bounllm/finetuned-models/ul2tr/classification/cg/17bintweet_sentiment
output_dir: /finetuned-models/turna/classification/cg/17bintweet_sentiment

2 changes: 1 addition & 1 deletion experiments/conf/debug.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
model_name: /stratch/bounllm/ckpt-500K
model_name: /pretrained_checkpoints/ckpt-1.74M
task_format: classification
task_mode: '' # '[S2S]: ', '[NLU]: ', '[NLG]: '
adafactor_scheduler: True
Expand Down
2 changes: 1 addition & 1 deletion experiments/conf/default.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
model_name: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_name: /pretrained_checkpoints/ckpt-1.74M
task_mode: '' # '[S2S]: ', '[NLU]: ', '[NLG]: '
training_params:
num_train_epochs: 10
Expand Down
4 changes: 2 additions & 2 deletions experiments/conf/ner_milliyet.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
defaults:
- default_cond_gen

dataset_loc: "/stratch/bounllm/supervised/ner-milliyet"
dataset_loc: "/supervised/ner-milliyet"
dataset_name: milliyet
task: ner
max_input_length: 380
Expand All @@ -10,5 +10,5 @@ training_params:
num_train_epochs: 10
per_device_train_batch_size: 8 # worked in ultramarine
per_device_eval_batch_size: 8 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/ner/milliyet
output_dir: /finetuned-models/turna/ner/milliyet
include_inputs_for_metrics: true
4 changes: 2 additions & 2 deletions experiments/conf/ner_milliyet_token.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
defaults:
- default_classification

dataset_loc: "/stratch/bounllm/supervised/ner-milliyet2"
dataset_loc: "/supervised/ner-milliyet2"
dataset_name: milliyet
task: ner
max_input_length: 380
Expand All @@ -10,4 +10,4 @@ num_labels: 7
training_params:
# per_device_train_batch_size: 64 # worked in ultramarine
# per_device_eval_batch_size: 64 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/turna-encoder/ner/milliyet
output_dir: /finetuned-models/turna-encoder/ner/milliyet
2 changes: 1 addition & 1 deletion experiments/conf/ner_wikiann.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,5 +8,5 @@ max_target_length: 40
training_params:
per_device_train_batch_size: 64 # worked in ultramarine
per_device_eval_batch_size: 64 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/ner/wikiann
output_dir: /finetuned-models/turna/ner/wikiann
include_inputs_for_metrics: true
2 changes: 1 addition & 1 deletion experiments/conf/ner_wikiann_token.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,4 +9,4 @@ num_labels: 7
training_params:
# per_device_train_batch_size: 64 # worked in ultramarine
# per_device_eval_batch_size: 64 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/turna-encoder/ner/wikiann
output_dir: /finetuned-models/turna-encoder/ner/wikiann
2 changes: 1 addition & 1 deletion experiments/conf/nli_multinli_tr.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,4 +10,4 @@ num_labels: 3
training_params:
per_device_train_batch_size: 32 # for ultramarine
per_device_eval_batch_size: 32 # for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/nli/nli_multinli_tr
output_dir: /finetuned-models/turna/nli/nli_multinli_tr
2 changes: 1 addition & 1 deletion experiments/conf/nli_snli_tr.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,4 +10,4 @@ num_labels: 3
training_params:
per_device_train_batch_size: 32 # for ultramarine
per_device_eval_batch_size: 32 # for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/nli/nli_snli_tr
output_dir: /finetuned-models/turna/nli/nli_snli_tr
2 changes: 1 addition & 1 deletion experiments/conf/nli_tr.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,4 +10,4 @@ num_labels: 3
training_params:
per_device_train_batch_size: 32 # for ultramarine
per_device_eval_batch_size: 32 # for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/nli/nli_tr
output_dir: /finetuned-models/turna/nli/nli_tr
2 changes: 1 addition & 1 deletion experiments/conf/paraphrasing.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,4 +8,4 @@ max_target_length: 20
training_params:
per_device_train_batch_size: 128
per_device_eval_batch_size: 128
output_dir: /stratch/bounllm/finetuned-models/paraphrasing/tatoeba
output_dir: /finetuned-models/paraphrasing/tatoeba
4 changes: 2 additions & 2 deletions experiments/conf/pos_boun.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
defaults:
- default_cond_gen

dataset_loc: "/stratch/bounllm/supervised/tr_boun-ud"
dataset_loc: "/supervised/tr_boun-ud"
dataset_name: boun
task: pos_tagging
max_input_length: 90
Expand All @@ -10,5 +10,5 @@ training_params:
num_train_epochs: 10
per_device_train_batch_size: 8 # worked in ultramarine
per_device_eval_batch_size: 8 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/pos/boun
output_dir: /finetuned-models/turna/pos/boun
include_inputs_for_metrics: true
4 changes: 2 additions & 2 deletions experiments/conf/pos_boun_token.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,11 +1,11 @@
defaults:
- default_classification

dataset_loc: "/stratch/bounllm/supervised/tr_boun-ud"
dataset_loc: "/supervised/tr_boun-ud"
dataset_name: boun
task: pos_tagging
max_input_length: 90
max_target_length: 300
num_labels: 17
training_params:
output_dir: /stratch/bounllm/finetuned-models/turna-encoder/pos/boun
output_dir: /finetuned-models/turna-encoder/pos/boun
4 changes: 2 additions & 2 deletions experiments/conf/pos_imst.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
defaults:
- default_cond_gen

dataset_loc: "/stratch/bounllm/supervised/tr_imst-ud"
dataset_loc: "/supervised/tr_imst-ud"
dataset_name: imst
task: pos_tagging
max_input_length: 60
Expand All @@ -10,5 +10,5 @@ training_params:
num_train_epochs: 10
per_device_train_batch_size: 16 # worked in ultramarine
per_device_eval_batch_size: 16 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/pos/imst
output_dir: /finetuned-models/turna/pos/imst
include_inputs_for_metrics: true
4 changes: 2 additions & 2 deletions experiments/conf/pos_imst_token.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,11 +1,11 @@
defaults:
- default_classification

dataset_loc: "/stratch/bounllm/supervised/tr_imst-ud"
dataset_loc: "/supervised/tr_imst-ud"
dataset_name: imst
task: pos_tagging
max_input_length: 60
max_target_length: 210
num_labels: 17
training_params:
output_dir: /stratch/bounllm/finetuned-models/turna-encoder/pos/imst
output_dir: /finetuned-models/turna-encoder/pos/imst
2 changes: 1 addition & 1 deletion experiments/conf/qa_exams.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,4 +9,4 @@ training_params:
num_train_epochs: 10
per_device_train_batch_size: 16 # worked in ultramarine
per_device_eval_batch_size: 16 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/qa/exams
output_dir: /finetuned-models/turna/qa/exams
2 changes: 1 addition & 1 deletion experiments/conf/qa_mkqa.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,4 +9,4 @@ training_params:
num_train_epochs: 10
per_device_train_batch_size: 64 # worked in ultramarine
per_device_eval_batch_size: 64 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/qa/mkqa
output_dir: /finetuned-models/turna/qa/mkqa
4 changes: 2 additions & 2 deletions experiments/conf/qa_tquad.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
defaults:
- default_cond_gen

dataset_loc: "/stratch/bounllm/supervised/tquad"
dataset_loc: "/supervised/tquad"
dataset_name: tquad
task: question_answering
max_input_length: 1350
Expand All @@ -10,4 +10,4 @@ training_params:
num_train_epochs: 10
per_device_train_batch_size: 2 # worked in ultramarine
per_device_eval_batch_size: 2 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/qa/tquad
output_dir: /finetuned-models/turna/qa/tquad
4 changes: 2 additions & 2 deletions experiments/conf/semantic_similarity.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -4,13 +4,13 @@ defaults:

dataset_name: stsb_tr
task: semantic_similarity
dataset_loc: "/stratch/bounllm/supervised/STSb-TR-corrected/"
dataset_loc: "/supervised/STSb-TR-corrected/"
max_input_length: 140
max_target_length: 10
num_labels: 1
training_params:
per_device_train_batch_size: 32 # for ultramarine
per_device_eval_batch_size: 32 # for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/semantic_similarity/stsb_tr/
output_dir: /finetuned-models/turna/semantic_similarity/stsb_tr/


2 changes: 1 addition & 1 deletion experiments/conf/summarization.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@ max_target_length: 128
training_params:
per_device_train_batch_size: 4 # 4 for ultramarine
per_device_eval_batch_size: 4 # 4 for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/summarization/tr_news
output_dir: /finetuned-models/turna/summarization/tr_news
bf16: true


2 changes: 1 addition & 1 deletion experiments/conf/title_generation.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,6 @@ max_target_length: 64
training_params:
per_device_train_batch_size: 8 # 8 for ultramarine
per_device_eval_batch_size: 8 # 8 for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/title_generation/tr_news
output_dir: /finetuned-models/turna/title_generation/tr_news


6 changes: 3 additions & 3 deletions experiments/generation_conf/class_prod_reviews.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,9 +7,9 @@ task: classification
max_input_length: 20
max_target_length: 4
num_labels: 2
model_path: /stratch/bounllm/finetuned-models/ul2tr/classification/cg/tr_product_reviews/
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/classification/cg/tr_product_reviews/
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
test_params:
per_device_eval_batch_size: 32 # for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/classification/cg/tr_product_reviews/
output_dir: /finetuned-models/turna/classification/cg/tr_product_reviews/

6 changes: 3 additions & 3 deletions experiments/generation_conf/class_ttc4900.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,10 +7,10 @@ task: classification
max_input_length: 1450
max_target_length: 8
num_labels: 7
model_path: /stratch/bounllm/finetuned-models/ul2tr/classification/cg/ttc4900
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/classification/cg/ttc4900
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
test_params:
per_device_eval_batch_size: 2 # for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/classification/cg/ttc4900
output_dir: /finetuned-models/turna/classification/cg/ttc4900


8 changes: 4 additions & 4 deletions experiments/generation_conf/ner_milliyet.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,14 +1,14 @@
defaults:
- default_cond_gen

dataset_loc: "/stratch/bounllm/supervised/ner-milliyet"
dataset_loc: "/supervised/ner-milliyet"
dataset_name: milliyet
task: ner
max_input_length: 380
max_target_length: 60
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /stratch/bounllm/finetuned-models/ul2tr/ner/milliyet
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/ner/milliyet
test_params:
per_device_eval_batch_size: 4
output_dir: /stratch/bounllm/finetuned-models/ul2tr/ner/milliyet
output_dir: /finetuned-models/turna/ner/milliyet
include_inputs_for_metrics: true
6 changes: 3 additions & 3 deletions experiments/generation_conf/ner_wikiann.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,9 +5,9 @@ dataset_name: wikiann
task: ner
max_input_length: 60
max_target_length: 40
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /stratch/bounllm/finetuned-models/ul2tr/ner/wikiann
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/ner/wikiann
test_params:
per_device_eval_batch_size: 4
output_dir: /stratch/bounllm/finetuned-models/ul2tr/ner/wikiann
output_dir: /finetuned-models/turna/ner/wikiann
include_inputs_for_metrics: true
4 changes: 2 additions & 2 deletions experiments/generation_conf/nli_snli_tr.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -6,8 +6,8 @@ dataset_name: snli_tr
task: nli
max_input_length: 128
max_target_length: 8
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /stratch/bounllm/finetuned-models/ul2tr/summarization/tr_news/
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/summarization/tr_news/
test_params:
per_device_eval_batch_size: 32 # for ultramarine
output_dir: ./t5_nli_tr
6 changes: 3 additions & 3 deletions experiments/generation_conf/nli_tr.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,8 +7,8 @@ task: nli
max_input_length: 128
max_target_length: 8
num_labels: 3
tokenizer_path: '/stratch/bounllm/pretrained_checkpoints/ckpt-1.74M'
model_path: /stratch/bounllm/finetuned-models/ul2tr/nli/nli_tr/
tokenizer_path: '/pretrained_checkpoints/ckpt-1.74M'
model_path: /finetuned-models/turna/nli/nli_tr/
test_params:
per_device_eval_batch_size: 32 # for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/nli/nli_tr/
output_dir: /finetuned-models/turna/nli/nli_tr/
6 changes: 3 additions & 3 deletions experiments/generation_conf/paraphrasing.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,8 +5,8 @@ dataset_name: tatoeba
task: paraphrasing
max_input_length: 20
max_target_length: 20
model_path: /stratch/bounllm/finetuned-models/ul2tr/paraphrasing/tatoeba/
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/paraphrasing/tatoeba/
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
test_params:
per_device_eval_batch_size: 128 # 4 for ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/paraphrasing/tatoeba/
output_dir: /finetuned-models/turna/paraphrasing/tatoeba/
8 changes: 4 additions & 4 deletions experiments/generation_conf/pos_boun.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,14 +1,14 @@
defaults:
- default_cond_gen

dataset_loc: "/stratch/bounllm/supervised/tr_boun-ud"
dataset_loc: "/supervised/tr_boun-ud"
dataset_name: boun
task: pos_tagging
max_input_length: 90
max_target_length: 300
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /stratch/bounllm/finetuned-models/ul2tr/pos/boun
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/pos/boun
test_params:
per_device_eval_batch_size: 4
output_dir: /stratch/bounllm/finetuned-models/ul2tr/pos/boun
output_dir: /finetuned-models/turna/pos/boun
include_inputs_for_metrics: true
8 changes: 4 additions & 4 deletions experiments/generation_conf/pos_imst.yaml
View file
Original file line number Diff line number Diff line change
@@ -1,14 +1,14 @@
defaults:
- default_cond_gen

dataset_loc: "/stratch/bounllm/supervised/tr_imst-ud"
dataset_loc: "/supervised/tr_imst-ud"
dataset_name: imst
task: pos_tagging
max_input_length: 60
max_target_length: 210
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /stratch/bounllm/finetuned-models/ul2tr/pos/imst
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/pos/imst
test_params:
per_device_eval_batch_size: 4
output_dir: /stratch/bounllm/finetuned-models/ul2tr/pos/imst
output_dir: /finetuned-models/turna/pos/imst
include_inputs_for_metrics: true
6 changes: 3 additions & 3 deletions experiments/generation_conf/qa_exams.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,8 +5,8 @@ dataset_name: exams
task: question_answering
max_input_length: 200
max_target_length: 50
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /stratch/bounllm/finetuned-models/ul2tr/qa/exams
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/qa/exams
test_params:
per_device_eval_batch_size: 16 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/qa/exams
output_dir: /finetuned-models/turna/qa/exams
6 changes: 3 additions & 3 deletions experiments/generation_conf/qa_mkqa.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,8 +5,8 @@ dataset_name: mkqa
task: question_answering
max_input_length: 40
max_target_length: 20
tokenizer_path: /stratch/bounllm/pretrained_checkpoints/ckpt-1.74M
model_path: /stratch/bounllm/finetuned-models/ul2tr/qa/mkqa
tokenizer_path: /pretrained_checkpoints/ckpt-1.74M
model_path: /finetuned-models/turna/qa/mkqa
test_params:
per_device_eval_batch_size: 64 # worked in ultramarine
output_dir: /stratch/bounllm/finetuned-models/ul2tr/qa/mkqa
output_dir: /finetuned-models/turna/qa/mkqa
Loading

0 comments on commit 8700994

Please sign in to comment.