init

experiments/huggingface_ops.py

@@ -3,7 +3,7 @@ from pprint import pprint
 
 api = HfApi()
 models = api.list_models(filter=ModelFilter(author='vocabtrimmer'))
-models_filtered = [i.modelId for i in models if 'qg-' in i.modelId]
+models_filtered = [i.modelId for i in models if 'cardiffnlp/twitter-roberta-base-jun2020' in i.modelId]
 pprint(sorted(models_filtered))
 # models = api.list_models(filter=ModelFilter(author='tweettemposhift'))
 # models_filtered = [i.modelId for i in models if 'topic-' in i.modelId]

experiments/main.sh

@@ -10,12 +10,6 @@ MODEL="cardiffnlp/twitter-roberta-base-2021-124m"
 MODEL="cardiffnlp/twitter-roberta-base-2022-154m"
 MODEL="cardiffnlp/twitter-roberta-large-2022-154m"
 
-# ABLATION (TimeLMs)
-## Topic & NER
-MODEL="cardiffnlp/twitter-roberta-base-jun2020"
-MODEL="cardiffnlp/twitter-roberta-base-sep2021"
-## NERD
-MODEL="cardiffnlp/twitter-roberta-base-jun2021"
 
 # SENTIMENT
 python model_finetuning_sentiment.py -m "${MODEL}" -d "sentiment_small_temporal"
@@ -78,34 +72,49 @@ python model_finetuning_topic.py -m "${MODEL}" -d "topic_random2_seed2"
 python model_finetuning_topic.py -m "${MODEL}" -d "topic_random3_seed2"
 
 download () {
-  git clone "https://huggingface.co/tweettemposhift/ner-${1}-${MODEL##*/}"
-  mkdir "ckpt/ner-${1}-${MODEL##*/}"
-  mv "ner-${1}-${MODEL##*/}" "ckpt/ner-${1}-${MODEL##*/}/"
-  mv "ckpt/ner-${1}-${MODEL##*/}/ner-${1}-${MODEL##*/}" "ckpt/ner-${1}-${MODEL##*/}/best_model"
+  git clone "https://huggingface.co/tweettemposhift/ner-${1}-${2##*/}"
+  mkdir "ckpt/ner-${1}-${2##*/}"
+  mv "ner-${1}-${2##*/}" "ckpt/ner-${1}-${2##*/}/"
+  mv "ckpt/ner-${1}-${2##*/}/ner-${1}-${2##*/}" "ckpt/ner-${1}-${2##*/}/best_model"
 }
-download "ner_temporal"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_temporal" --skip-train
-download "ner_random0_seed0"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed0" --skip-train
-download "ner_random1_seed0"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed0" --skip-train
-download "ner_random2_seed0"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed0" --skip-train
-download "ner_random3_seed0"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed0" --skip-train
-download "ner_random0_seed1"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed1" --skip-train
-download "ner_random1_seed1"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed1" --skip-train
-download "ner_random2_seed1"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed1" --skip-train
-download "ner_random3_seed1"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed1" --skip-train
-download "ner_random0_seed2"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random0_seed2" --skip-train
-download "ner_random1_seed2"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random1_seed2" --skip-train
-download "ner_random2_seed2"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random2_seed2" --skip-train
-download "ner_random3_seed2"
-python model_finetuning_ner.py -m "${MODEL}" -d "ner_random3_seed2" --skip-train
+
+fix () {
+  download "ner_temporal" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_temporal" --skip-train
+  download "ner_random0_seed0" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random0_seed0" --skip-train
+  download "ner_random1_seed0" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random1_seed0" --skip-train
+  download "ner_random2_seed0" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random2_seed0" --skip-train
+  download "ner_random3_seed0" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random3_seed0" --skip-train
+  download "ner_random0_seed1" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random0_seed1" --skip-train
+  download "ner_random1_seed1" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random1_seed1" --skip-train
+  download "ner_random2_seed1" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random2_seed1" --skip-train
+  download "ner_random3_seed1" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random3_seed1" --skip-train
+  download "ner_random0_seed2" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random0_seed2" --skip-train
+  download "ner_random1_seed2" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random1_seed2" --skip-train
+  download "ner_random2_seed2" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random2_seed2" --skip-train
+  download "ner_random3_seed2" "${1}"
+  python model_finetuning_ner.py -m "${1}" -d "ner_random3_seed2" --skip-train
+}
+
+fix "roberta-base"
+fix "vinai/bertweet-base"
+fix "jhu-clsp/bernice"
+fix "roberta-large"
+fix "vinai/bertweet-large"
+fix "cardiffnlp/twitter-roberta-base-2019-90m"
+fix "cardiffnlp/twitter-roberta-base-dec2020"
+fix "cardiffnlp/twitter-roberta-base-2021-124m"
+fix "cardiffnlp/twitter-roberta-base-2022-154m"
+fix "cardiffnlp/twitter-roberta-large-2022-154m"
+

experiments/model_finetuning_ner.py

@@ -21,6 +21,8 @@ from transformers import AutoTokenizer, AutoModelForTokenClassification, Trainin
 from huggingface_hub import Repository
 
 logging.basicConfig(format="%(asctime)s %(levelname)-8s %(message)s", level=logging.INFO, datefmt="%Y-%m-%d %H:%M:%S")
+os.environ["WANDB_DISABLED"] = "true"
+
 EVAL_STEP = 500
 RANDOM_SEED = 42
 N_TRIALS = 10

experiments/model_finetuning_nerd.py

@@ -21,6 +21,8 @@ from transformers import AutoTokenizer, AutoModelForSequenceClassification, Trai
 from huggingface_hub import Repository
 
 logging.basicConfig(format="%(asctime)s %(levelname)-8s %(message)s", level=logging.INFO, datefmt="%Y-%m-%d %H:%M:%S")
+os.environ["WANDB_DISABLED"] = "true"
+
 EVAL_STEP = 500
 RANDOM_SEED = 42
 N_TRIALS = 10

experiments/model_finetuning_sentiment.py

@@ -21,6 +21,8 @@ from transformers import AutoTokenizer, AutoModelForSequenceClassification, Trai
 from huggingface_hub import Repository
 
 logging.basicConfig(format="%(asctime)s %(levelname)-8s %(message)s", level=logging.INFO, datefmt="%Y-%m-%d %H:%M:%S")
+os.environ["WANDB_DISABLED"] = "true"
+
 EVAL_STEP = 500
 RANDOM_SEED = 42
 N_TRIALS = 10

experiments/requirements.txt

@@ -10,4 +10,5 @@ accelerate==0.23.0
 evaluate==0.4.1
 sentencepiece==0.1.99
 protobuf==4.24.4
-seqeval==1.2.2
+seqeval==1.2.2
+wandb==0.16.0