model update

README.md

@@ -29,19 +29,19 @@ model-index:
     metrics:
     - name: BLEU4
       type: bleu4
-      value: 5.960482240567237e-10
+      value: 0.13755949895011021
     - name: ROUGE-L
       type: rouge-l
-      value: 0.0054045507102811466
+      value: 0.3723510278895709
     - name: METEOR
       type: meteor
-      value: 0.0029513976825252613
+      value: 0.31606923044567353
     - name: BERTScore
       type: bertscore
-      value: 0.03922946683914634
+      value: 0.9109018614729723
     - name: MoverScore
       type: moverscore
-      value: 0.45608571714273055
+      value: 0.6276807689001792
 ---
 
 # Model Card of `lmqg/t5-large-tweetqa-qag`
@@ -105,7 +105,7 @@ question = pipe('generate question and answer:  Beyonce further expanded her act
 
 | Dataset | Type | BLEU4 | ROUGE-L | METEOR | BERTScore | MoverScore | Link |
 |:--------|:-----|------:|--------:|-------:|----------:|-----------:|-----:|
-| [lmqg/qag_tweetqa](https://huggingface.co/datasets/lmqg/qag_tweetqa) | default | 0.0 | 0.005 | 0.003 | 0.039 | 0.456 | [link](https://huggingface.co/lmqg/t5-large-tweetqa-qag/raw/main/eval/metric.first.sentence.paragraph.questions_answers.lmqg_qag_tweetqa.default.json) | 
+| [lmqg/qag_tweetqa](https://huggingface.co/datasets/lmqg/qag_tweetqa) | default | 0.138 | 0.372 | 0.316 | 0.911 | 0.628 | [link](https://huggingface.co/lmqg/t5-large-tweetqa-qag/raw/main/eval/metric.first.answer.paragraph.questions_answers.lmqg_qag_tweetqa.default.json) | 
 
 
 
@@ -121,13 +121,13 @@ The following hyperparameters were used during fine-tuning:
  - model: t5-large
  - max_length: 256
  - max_length_output: 128
- - epoch: 15
+ - epoch: 16
  - batch: 16
- - lr: 5e-05
+ - lr: 0.0001
  - fp16: False
  - random_seed: 1
  - gradient_accumulation_steps: 4
- - label_smoothing: 0.15
+ - label_smoothing: 0.0
 
 The full configuration can be found at [fine-tuning config file](https://huggingface.co/lmqg/t5-large-tweetqa-qag/raw/main/trainer_config.json).
 

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "lmqg_output/t5_large_tweetqa/best_model",
+  "_name_or_path": "lmqg_output/t5_large_tweetqa/model_mzgdpa/epoch_15",
   "add_prefix": true,
   "architectures": [
     "T5ForConditionalGeneration"

eval/metric.first.answer.paragraph.questions_answers.lmqg_qag_tweetqa.default.json

@@ -1 +1 @@
-{"validation": {"Bleu_1": 0.00024686972216758125, "Bleu_2": 4.168876740936257e-05, "Bleu_3": 1.1308637528972774e-05, "Bleu_4": 4.20122611344465e-06}, "test": {"Bleu_1": 9.788893461144265e-05, "Bleu_2": 1.4497461639428926e-05, "Bleu_3": 3.341798200517125e-06, "Bleu_4": 2.4198848871473664e-10}}
+{"validation": {"Bleu_1": 0.394207422823425, "Bleu_2": 0.27038964481113453, "Bleu_3": 0.18739277300825702, "Bleu_4": 0.13114698367378638, "METEOR": 0.34585457496727634, "ROUGE_L": 0.382560092903778, "BERTScore": 0.9066984992902438, "MoverScore": 0.6275517569967117}, "test": {"Bleu_1": 0.4133416813705724, "Bleu_2": 0.2836783148838816, "Bleu_3": 0.19681581918083613, "Bleu_4": 0.13755949895011021, "METEOR": 0.31606923044567353, "ROUGE_L": 0.3723510278895709, "BERTScore": 0.9109018614729723, "MoverScore": 0.6276807689001792}}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2e44f0123ea97366277358f859e565b30b7863d5c7ec6b23ed706cfb5f768b6b
-size 2950727111
+oid sha256:afba7fd4c15282caf436b4f53864293edad36eb3ae5441b4d7366b001be69d6d
+size 2950734215

tokenizer_config.json

@@ -104,7 +104,7 @@
   "eos_token": "</s>",
   "extra_ids": 100,
   "model_max_length": 512,
-  "name_or_path": "lmqg_output/t5_large_tweetqa/best_model",
+  "name_or_path": "lmqg_output/t5_large_tweetqa/model_mzgdpa/epoch_15",
   "pad_token": "<pad>",
   "special_tokens_map_file": null,
   "tokenizer_class": "T5Tokenizer",

trainer_config.json

@@ -1 +1 @@
-{"dataset_path": "lmqg/qag_tweetqa", "dataset_name": "default", "input_types": ["paragraph"], "output_types": ["questions_answers"], "prefix_types": ["qag"], "model": "t5-large", "max_length": 256, "max_length_output": 128, "epoch": 15, "batch": 16, "lr": 5e-05, "fp16": false, "random_seed": 1, "gradient_accumulation_steps": 4, "label_smoothing": 0.15}
+{"dataset_path": "lmqg/qag_tweetqa", "dataset_name": "default", "input_types": ["paragraph"], "output_types": ["questions_answers"], "prefix_types": ["qag"], "model": "t5-large", "max_length": 256, "max_length_output": 128, "epoch": 16, "batch": 16, "lr": 0.0001, "fp16": false, "random_seed": 1, "gradient_accumulation_steps": 4, "label_smoothing": 0.0}