init

experiments/main.sh

@@ -1,56 +1,40 @@
-# topic [hawk]
-MODEL="cardiffnlp/twitter-roberta-base-2022-154m"
-# topic [hawk]
+# topic, ner
 MODEL="cardiffnlp/twitter-roberta-base-2021-124m"
-# topic [hawk]
+# topic, ner
+MODEL="roberta-base"
+# topic, ner [ukri]
+MODEL="cardiffnlp/twitter-roberta-base-2022-154m"
+# topic [hawk], ner [ukri]
 MODEL="jhu-clsp/bernice"
-# topic [hawk]
+# topic, ner [as it's not fasttokenizer, a bit tricky...]
 MODEL="vinai/bertweet-base"
-# topic [hawk], ner [stone]
-MODEL="roberta-base"
 
 
 # NER
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_temporal"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed0"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed0"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed0"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed0"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed1"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed1"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed1"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed1"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed2"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed2"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed2"
-rm -rf ckpt
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed2"
-rm -rf ckpt
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_temporal" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed0" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed0" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed0" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed0" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed1" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed1" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed1" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed1" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed2" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed2" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed2" --skip-train --skip-test
+python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed2" --skip-train --skip-test
 
 # NERD
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_temporal"
-
-python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random0_seed0" --skip-train --skip-test
+python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random0_seed0"
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random1_seed0"
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random2_seed0"
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random3_seed0"
-
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random0_seed1"
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random1_seed1"
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random2_seed1"
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random3_seed1"
-
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random0_seed2"
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random1_seed2"
 python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random2_seed2"
@@ -58,19 +42,16 @@ python model_finetuning_nerd.py -m "${MODEL}" -d "nerd_random3_seed2"
 
 
 # TOPIC
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_temporal"
-
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random0_seed0"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random1_seed0"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random2_seed0"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random3_seed0"
-
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random0_seed1"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random1_seed1"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random2_seed1"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random3_seed1"
-
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random0_seed2"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random1_seed2"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random2_seed2"
-python model_finetuning_topic.py -m "${MODEL}" -d "topic_random3_seed2"
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_temporal" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random0_seed0" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random1_seed0" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random2_seed0" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random3_seed0" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random0_seed1" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random1_seed1" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random2_seed1" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random3_seed1" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random0_seed2" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random1_seed2" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random2_seed2" --skip-train --skip-test
+python model_finetuning_topic.py -m "${MODEL}" -d "topic_random3_seed2" --skip-train --skip-test

experiments/model_finetuning_ner.py

@@ -17,7 +17,7 @@ from glob import glob
 import numpy as np
 import evaluate
 from datasets import load_dataset
-from transformers import convert_slow_tokenizer, AutoTokenizer, AutoModelForTokenClassification, TrainingArguments, Trainer
+from transformers import AutoTokenizer, AutoModelForTokenClassification, TrainingArguments, Trainer
 from huggingface_hub import Repository
 
 logging.basicConfig(format="%(asctime)s %(levelname)-8s %(message)s", level=logging.INFO, datefmt="%Y-%m-%d %H:%M:%S")
@@ -102,7 +102,6 @@ def main(
 
     def tokenize_and_align_labels(examples):
         tokens = [[preprocess(model, w) for w in t] for t in examples["text_tokenized"]]
-        # if tokenizer.is_fast:
         tokenized_inputs = tokenizer(
             tokens,
             truncation=True,
@@ -111,23 +110,20 @@ def main(
             max_length=128)
         all_labels = examples["gold_label_sequence"]
         new_labels = []
-        for ind, labels in enumerate(all_labels):
-            word_ids = tokenized_inputs.word_ids(ind)
-            new_labels.append(align_labels_with_tokens(labels, word_ids))
+        if tokenizer.is_fast:
+            for ind, labels in enumerate(all_labels):
+                word_ids = tokenized_inputs.word_ids(ind)
+                new_labels.append(align_labels_with_tokens(labels, word_ids))
+        else:
+            for token, label in zip(tokens, all_labels):
+                tmp_labels = [-100]
+                for to, la in enumerate(token, label):
+                    to_tokenized = tokenizer.tokenize(to)
+                    tmp_labels += [la] * len(to_tokenized)
+                tmp_labels = tmp_labels + [-100] * (128 - len(tmp_labels))
+                new_labels.append(tmp_labels)
         tokenized_inputs["labels"] = new_labels
         return tokenized_inputs
-        # else:
-        #     tokenized_inputs = tokenizer(
-        #         tokens)
-        #     all_labels = examples["gold_label_sequence"]
-        #     new_labels = []
-        #     for ind, labels in enumerate(all_labels):
-        #         word_ids = tokenized_inputs.word_ids(ind)
-        #         new_labels.append(align_labels_with_tokens(labels, word_ids))
-        #     tokenized_inputs["labels"] = new_labels
-        #     return tokenized_inputs
-
-
 
     dataset = load_dataset(dataset, dataset_type)
     tokenized_datasets = dataset.map(lambda x: tokenize_and_align_labels(x), batched=True)